Injection sclerotherapy for varicose veins

Abstract

Background

Varicose veins are enlarged and tortuous veins, affecting up to one‐third of the world's population. They can be a cause of chronic venous insufficiency, which is characterised by oedema, pigmentation, eczema, lipodermatosclerosis, atrophie blanche, and healed or active venous ulcers. Injection sclerotherapy (liquid or foam) is widely used for treatment of varicose veins aiming to transform the varicose veins into a fibrous cord. However, there is limited evidence regarding its effectiveness and safety, especially in patients with more severe disease. This is the second update of the review first published in 2002.

Objectives

To assess the effectiveness and safety of injection sclerotherapy for the treatment of varicose veins.

Search methods

For this update, the Cochrane Vascular Information Specialist searched the Cochrane Vascular Specialised Register, CENTRAL, MEDLINE, Embase, AMED, CINAHL, and LILACS databases, and the World Health Organization International Clinical Trials Registry Platform and ClinicalTrials.gov trials registries, on 20 July 2021.

Selection criteria

We included all randomised controlled trials (RCTs) (including cluster‐randomised trials and first phase cross‐over studies) that used injection sclerotherapy for the treatment of varicose veins.

Data collection and analysis

Two review authors independently assessed, selected and extracted data. Disagreements were cross‐checked by a third review author. We used Cochrane's Risk of bias tool to assess the risk of bias. The outcomes of interest were cosmetic appearance, complications, residual varicose veins, quality of life (QoL), persistence of symptoms, and recurrent varicose veins. We calculated risk ratios (RRs) or mean difference (MD) with 95% confidence intervals (CIs). We used the worst‐case‐scenario for dichotomous data imputation for intention‐to‐treat analyses. For continuous outcomes, we used the 'last‐observation‐carried‐forward' for data imputation if there was balanced loss to follow‐up. We assessed the certainty of the evidence using the GRADE approach.

Main results

We included 23 new RCTs for this update, bringing the total to 28 studies involving 4278 participants. The studies differed in their design, and in which sclerotherapy method, agent or concentration was used. None of the included RCTs compared sclerotherapy to no intervention or to any pharmacological therapy. The certainty of the evidence was downgraded for risk of bias, low number of studies providing information for each outcome, low number of participants, clinical differences between the study participants, and wide CIs.

Sclerotherapy versus placebo

Foam sclerotherapy may improve cosmetic appearance as measured by IPR‐V (independent photography review ‐ visible varicose veins scores) compared to placebo (polidocanol 1%: mean difference (MD) ‐0.76, 95% CI ‐0.91 to ‐0.60; 2 studies, 223 participants; very low‐certainty evidence); however, deep vein thrombosis (DVT) rates may be slightly increased in this intervention group (RR 5.10, 95% CI 1.30 to 20.01; 3 studies, 302 participants; very low‐certainty evidence). Residual varicose vein rates may be decreased following polidocanol 1% compared to placebo (RR 0.19, 95% CI 0.13 to 0.29; 2 studies, 225 participants; very low‐certainty evidence). Following polidocanol 1% use, there may be a possible improvement in QoL as assessed using the VEINES‐QOL/Sym questionnaire (MD 12.41, 95% CI 9.56 to 15.26; 3 studies, 299 participants; very low‐certainty evidence), and possible improvement in varicose vein symptoms as assessed using the Venous Clinical Severity Score (VCSS) (MD ‐3.25, 95% CI ‐3.90 to ‐2.60; 2 studies, 223 participants; low‐certainty evidence). Recurrent varicose veins were not reported for this comparison.

Foam sclerotherapy versus foam sclerotherapy with different concentrations

Three individual RCTs reported no evidence of a difference in cosmetic appearance after comparing different concentrations of the intervention; data could not be pooled for two of the three studies (RR 1.11, 95% CI 0.84 to 1.47; 1 study, 80 participants; very low‐certainty evidence). Similarly, there was no clear difference in rates of thromboembolic complications when comparing one foam concentration with another (RR 1.47, 95% CI 0.41 to 5.33; 3 studies, 371 participants; very low‐certainty evidence). Three RCTs investigating higher concentrations of polidocanol foam indicated the rate of residual varicose veins may be slightly decreased in the polidocanol 3% foam group compared to 1% (RR 0.67, 95% CI 0.43 to 1.04; 3 studies, 371 participants; moderate‐certainty evidence). No clear improvement in QoL was detected. Two RCTs reported improved VCSS scores with increasing concentrations of foam. Persistence of symptoms were not reported for this comparison. There was no clear difference in recurrent varicose vein rates (RR 0.91, 95% CI 0.62 to 1.32; 1 study, 148 participants; low‐certainty evidence).

Foam sclerotherapy versus liquid sclerotherapy

One RCT reported on cosmetic appearance with no evidence of a difference between foam or liquid sclerotherapy (patient satisfaction scale MD 0.2, 95% CI ‐0.27 to 0.67; 1 study, 126 participants; very low‐certainty evidence). None of the RCTs investigated thromboembolic complications, QoL or persistence of symptoms. Six studies individually showed there may be a benefit to polidocanol 3% foam over liquid sclerotherapy in reducing residual varicose vein rate; pooling data from two studies showed a RR of 0.51, with 95% CI 0.41 to 0.65; 203 participants; very low‐certainty evidence. One study reported no clear difference in recurrent varicose vein rates when comparing sodium tetradecyl sulphate (STS) foam or liquid (RR 1.10, 95% CI 0.86 to 1.42; 1 study, 286 participants; very low‐certainty evidence).

Sclerotherapy versus sclerotherapy with different substances

Four RCTs compared sclerotherapy versus sclerotherapy with any other substance. We were unable to combine the data due to heterogeneity or assess the certainty of the evidence due to insufficient data.

Authors' conclusions

There is a very low to low‐certainty evidence that, compared to placebo, sclerotherapy is an effective and safe treatment for varicose veins concerning cosmetic appearance, residual varicose veins, QoL, and persistence of symptoms. Rates of DVT may be slightly increased and there were no data concerning recurrent varicose veins. There was limited or no evidence for one concentration of foam compared to another; foam compared to liquid sclerotherapy; foam compared to any other substance; or one technique compared to another. There is a need for high‐quality trials using standardised sclerosant doses, with clearly defined core outcome sets, and measurement time points to increase the certainty of the evidence.

Plain language summary

Injection sclerotherapy for varicose veins

Background

Varicose veins are enlarged, visibly lumpy knotted veins, usually found in the legs. They can cause pain, burning discomfort, aching, and itching as well as generalised aching, heaviness or swelling in the legs. However, some people do not attribute all symptoms to varicose veins, as there is little correlation between these symptoms and the extent or size of the varicose veins. Varicose vein disease can affect quality of life as it is cosmetically unattractive, can cause phlebitis, bleeding, skin pigmentation, and ulcers.

Injection sclerotherapy can be used to treat varicose veins. Sclerotherapy is achieved by injection of an irritating liquid or foam into the blood vessels, leading to transformation of varicose veins into a fibrous cord. Possible complications include the formation of blood clots, skin pigmentation, inflammation, ulceration with tissue damage, and reactions to the sclerosing agent.

This review aimed to investigate how effective and safe sclerotherapy is by comparing it to no treatment or placebo (sham) treatment; comparing different ways to perform sclerotherapy (including different doses, liquid versus foam, injection techniques, different substances) or comparing sclerotherapy to compression.

Study characteristics and key findings

We searched for relevant studies and identified 28 randomised controlled trials, involving over 4278 participants (search current to 20 July 2021). We looked at cosmetic appearance, complications, residual varicose vein rates, quality of life, and symptomatic improvement related to sclerotherapy. There were no trials assessing treatment versus no interventions, sclerotherapy at different time intervals, or sclerotherapy compared with pharmacological therapy.

Our results showed that, compared to placebo, sclerotherapy may improve cosmetic appearance and quality of life, and may reduce residual varicose veins and persistence of symptoms but the evidence is uncertain. Rates of deep vein thrombosis may be slightly increased but the evidence is uncertain. There were no data for recurrent varicose veins. There was not enough Information to conclude if one concentration of foam was better than another, or if foam was better than liquid or any other substance; or if there was any benefit of one injection technique compared to another.

Reliability of the evidence

The evidence was graded as very low to moderate for all outcomes. The reliability of the evidence was downgraded due to concerns about how the studies were designed (risk of bias), the low number of studies providing information for each outcome, the low number of participants, clinical differences between the study participants, and wide confidence intervals. There is a need for more high‐quality trials using standardised methods to increase our confidence in the results.

Summary of findings

Summary of findings 1. Foam sclerotherapy versus placebo for treating varicose veins.

Foam sclerotherapy (polidocanol 1%) compared to placebo for varicose veins
Patient or population: participants with varicose veinsa Setting: clinical or hospital‐based study Intervention: polidocanol 1% foam sclerotherapy Comparison: placebo
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo	Risk with polidocanol 1% foam sclerotherapy
Cosmetic appearance (IPR‐V) (Scale from: 0 to 4; 0 being the best appearance (0: none, 1: mild, 2: moderate, 3: severe, 4: very severe)) follow‐up: 8 weeks	The adjusted mean change from baseline in cosmetic appearance (IPR‐V) was ‐0.04	The MD was 0.76 lower (0.91 lower to 0.6 lower)	‐	223 (2 RCTs)	⊕⊝⊝⊝ VERY LOW b,c
Complications ‐ DVT follow‐up: 8 weeks	Study population		RR 5.10 (1.30 to 20.01)	302 (3 RCTs)	⊕⊝⊝⊝ VERY LOW b,c,d,e,f
	13 per 1000	68 per 1000 (17 to 265)
Residual varicose veins follow‐up: 8 weeks	Study population		RR 0.19 (0.13 to 0.29)	225 (2 RCTs)	⊕⊝⊝⊝ VERY LOW b,c
	973 per 1000	185 per 1000 (127 to 282)
QoL (VEINES‐QOL/Sym; scale from: 0 to 50 with 0 being the worst and 50 better) follow‐up: 8 weeks	The mean change from baseline in QoL (VEINES‐QOL/Sym score) was 9.60	The MD was 12.41 higher (9.56 higher to 15.26 higher)	‐	299 (3 RCTs)	⊕⊝⊝⊝ VERY LOWf,g,h
Persistence of symptoms (VCSS) follow‐up: 8 weeks	The mean change from baseline in persistence of symptoms (VCSS score) was ‐1.14	The MD was 3.25 lower (3.9 lower to 2.6 lower)	‐	223 (2 RCTs)	⊕⊕⊝⊝ LOW g,h
Recurrent varicose veins	See comment		‐	‐	‐	No study evaluated recurrent varicose veins in this comparison
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; DVT: deep vein thrombosis; IPR‐V: independent photography review ‐ visible varicose veins scores;MD: mean difference; QoL: quality of life; RCT: randomised controlled trial; RR: risk ratio; VCSS: venous clinical severity score; VEINES‐QOL/Sym: disease‐specific QoL instrument for chronic venous disorders of the leg
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

^a There were CEAP C1 (11.63% in Ceulen 2007), C5 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), and C6 (Hamel‐Desnos 2007) participants included but there was no description of the proportion of these participants in the groups. Ceulen 2007 used the lower leg as the unit of analysis. In the remaining studies, the participant was the unit of analysis.

^b We downgraded the certainty of the evidence two levels as neither personnel nor participants were blinded (risk of bias concerns).

^c We downgraded one level for indirectness as Todd 2014 included 8.8% CEAP C5 and C6 in the placebo group and 3.4% CEAP C5 and C6 in the polidocanol 1% foam group.

^d We downgraded one level due to inconsistency ‐ clinical heterogeneity among studies (severity of disease).

^e We downgraded one level for imprecision due to low number of events; CI included both 'no difference' and a 'clinically important difference'.

^f We downgraded one level due to a wide CI.

^g We downgraded one level for imprecision ‐ low number of participants.

^h We downgraded one level for indirectness (8.8% CEAP C5 and C6 participants in the placebo group).

Summary of findings 2. Foam sclerotherapy versus foam sclerotherapy for treating varicose veins.

Foam sclerotherapy (polidocanol 3%) compared to foam sclerotherapy (polidocanol 1%) for treating varicose veins
Patient or population: participants with varicose veinsa Setting: hospital or clinical setting Intervention: polidocanol 3% foam sclerotherapy Comparison: polidocanol 1% foam sclerotherapy
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with polidocanol 1% foam sclerotherapy	Risk with polidocanol 3% foam sclerotherapy
Cosmetic appearance (self‐described cosmetic improvement) follow‐up: mean 1 year	Study population		RR 1.11 (0.84 to 1.47)	80 (1 RCT)	⊕⊝⊝⊝ VERY LOW b,c,d,e
	675 per 1000	749 per 1000 (567 to 992)
Complications ‐ thromboembolic follow‐up: range 3 weeks to 6 months	Study population		RR 1.47 (0.41 to 5.33)	371 (3 RCTs)	⊕⊝⊝⊝ VERY LOW d,f,g
	16 per 1000	24 per 1000 (7 to 86)
Residual varicose veins follow‐up: range 6 months to 2 years	Study population		RR 0.67 (0.43 to 1.04)	371 (3 RCTs)	⊕⊕⊕⊝ MODERATE f
	364 per 1000	244 per 1000 (156 to 378)
QoL	See comment		‐	‐	‐	No study evaluated QoL for this comparison
Persistence of symptoms	See comment		‐	‐	‐	No study evaluated persistence of symptoms for this comparison
Recurrent varicose veins follow‐up: mean 2 years	Study population		RR 0.91 (0.62 to 1.32)	148 (1 RCT)	⊕⊕⊝⊝ LOW c,h
	446 per 1000	406 per 1000 (276 to 589)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; QoL: quality of life; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

^aCeulen 2007 included 12.5% C1 participants. The studies included participants from C2 to C6 clinical classes without description of the proportions of C5 and C6 participants.

^b We downgraded by one level as most risk of bias domains were judged as "unclear".

^c We downgraded by one level due to imprecision ‐ single study.

^d We downgraded by one level due to imprecision ‐ wide confidence interval.

^e We downgraded by one level due to imprecision ‐ low number of participants.

^f There were CEAP C1 (11.63% in Ceulen 2007), C5 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), and C6 (Hamel‐Desnos 2007) participants included but there was no description of the proportion of these participants in the groups. Ceulen 2007) used the lower leg as the unit of analysis. In the remaining studies, the participant was the unit of analysis. We downgraded by one level for indirectness.

^g We downgraded by one level due to imprecision ‐ low number of events.

^h There were participants from C5 and C6 clinical classes included (Hamel‐Desnos 2007) but there was no description of the proportion of these participants in the groups. We downgraded by one level for indirectness.

Summary of findings 3. Foam sclerotherapy versus liquid sclerotherapy for treating varicose veins.

Foam sclerotherapy compared to liquid sclerotherapy for varicose treating veins
Patient or population: participants with varicose veinsa Setting: hospital or clinic‐based Intervention: foam sclerotherapy using the same agent and concentration of the control Comparison: liquid sclerotherapy using the same agent and concentration of the intervention
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with liquid sclerotherapy	Risk with foam sclerotherapy
Cosmetic appearance (patient satisfaction scale, the higher the better) follow‐up: long‐term (365 days)	The mean cosmetic appearance of the patient satisfaction scale was 7.2	The MD was 0.2 higher (0.27 lower to 0.67 higher)	‐	126 (1 RCT)	⊕⊝⊝⊝ VERY LOW a,b,c	Alos 2006 reported no difference when using polidocanol 0.75% to 1.25% foam versus polidocanol 1.5% to 2.5% liquid. Different concentrations and different vein diameters were compared, reducing our confidence in these results
Complications ‐ DVT	See comment		‐	‐	‐	No study evaluated DVT or thromboembolic events for this comparison
Residual varicose veins follow‐up: range 3 months to 2 years	Study population		RR 0.51 (0.41 to 0.65)	203 (2 RCTs)	⊕⊝⊝⊝ VERY LOW b,c,d
	842 per 1000	429 per 1000 (345 to 547)
QoL	See comment		‐	‐	‐	No study evaluated QoLfor this comparison
Persistence of symptoms	See comment		‐	‐	‐	No study evaluated persistence of symptoms for this comparison
Recurrent varicose veins follow‐up: mean 5 years	Study population		RR 1.10 (0.86 to 1.42)	286 (1 RCT)	⊕⊝⊝⊝ VERY LOW a,e,h
	433 per 1000	477 per 1000 (373 to 615)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; DVT; deep vein thrombosis; MD: mean difference; QoL: quality of life;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

^a Did not describe randomisation method. Due to the nature of the intervention was probably impossible to blind interventionist. We did not find a protocol. Risk of bias concerns

^b Single study. We downgraded by one level for imprecision.

^c Compared different concentration of the same substances for different vein diameter. Used different concentrations of the drug if liquid or foam

^d We downgraded by one level for risk of bias concerns (query about randomisation method).

^e We downgraded by one level for risk of bias concerns (intervention was presumably not blinded (foam vs liquid) due to the nature of the intervention).

^f One study (Ouvry 2008) included participants from CEAP C2 to C6 without discriminating between them. Rabe 2008 included 1 CEAP C5 participant in the foam group (1.85%) and 4 CEAP C5 participants in the liquid group (7.69%). We downgraded by one level for indirectness.

^hBelcaro 2003b used a tensioactive substance (ultrasound contrast agent) to produce foam. This is not a standard way to produce foam. We downgraded by one level for indirectness.

Summary of findings 4. Sclerotherapy versus sclerotherapy with different substances for treating varicose veins.

Sclerotherapy with STS 1% compared to sclerotherapy with polidocanol 1%a
Patient or population: participants with varicose veinsb Setting: clinical or hospital‐based study Intervention: STS 1% sclerotherapy Comparison: polidocanol 1% sclerotherapy
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with polidocanol 1% sclerotherapy	Risk with STS 1% sclerotherapy
Cosmetic appearance follow‐up: 12 to 16 weeks	See comment		‐	‐	‐	Two RCTs compared sclerotherapy with sclerotherapy with different substances but used a range of concentrations, formulations and drug volumes. We were unable to pool the data or draw conclusions
Complications ‐ DVT follow‐up: 16 weeks to 12 months	See comment		‐	‐	‐	Two RCTs reported on DVT but we were unable to pool the data or draw conclusions
Residual varicose veins follow‐up: 12 months	See comment		‐	‐	‐	One RCT reported on residual varicose veins but we were unable to draw conclusions
QoL	See comment		‐	‐	‐	No study evaluated QoL in this comparison
Persistence of symptoms	See comment		‐	‐	‐	No study evaluated persistence of symptoms in this comparison
Recurrent varicose veins	See comment		‐	‐	‐	No study evaluated recurrent varicose veins in this comparison
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; DVT: deep vein thrombosis; QoL: quality of life; RCT: randomised controlled trial: RR: risk ratio; STS: sodium tetradecyl sulphate
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

^a Four RCTs compared sclerotherapy with sclerotherapy with different substances but used a range of concentrations, formulations and drug volumes so we were not able to pool the data (Goldman 2002; Rao 2005; Schadeck 1995a; Wright 2006). None of these compared STS 1% sclerotherapy versus sclerotherapy with polidocanol 1% which we had planned to present in this table.

^b Three studies did not describe CEAP classification of participants (Goldman 2002; Rao 2005; Schadeck 1995a). One study compared Varisolve with polidocanol 1% to any marketed sclerosant at any marketed concentration liquid or foam (Wright 2006)

Background

Description of the condition

Varicose veins are enlarged and tortuous veins. They are subcutaneous and at least 3 mm in diameter (Eklöf 2004; Hamdan 2012). It is difficult to find a satisfactory definition of varicose veins upon which consensus has been reached since reticular varicose veins can also be enlarged and tortuous. In 2012, the Vein Consult program cohort reported the epidemiology of chronic venous disease of a large international population of 91,545 adult patients' cohort across a diverse geographic area, including Europe, Latin America, the Middle East, and the Far East (Rabe 2012). Rabe and colleagues reported that 41.4% of these patients had early stage chronic venous disease (CEAP class C0s–C1) and 84% had any class of chronic venous disease greater than C0s (Rabe 2012).

There are different prevalences of varicose vein disease described around the world, depending on factors such as personal lifestyle, obesity, and age (Davies 2019).

The symptoms attributable to varicose veins, and their correlation with the extent of venous reflux (reversal of blood flow through the valves of veins during standing or sitting), are not clearly defined. Epidemiological evidence suggests that even in the presence of 'trunk' varicose veins, defined as varicose from great or short saphenous veins or their first or second order branches, most lower‐limb symptoms have a non‐venous cause (Bradbury 1999). The Edinburgh Vein Study has demonstrated superficial venous reflux in 9% of randomly selected men and 15% of women as well as deep vein reflux in 22% of men and 11% of women (Allan 2000). Subjects with visible varicose veins have a higher incidence of reflux on Duplex ultrasound, although in many cases of documented reflux, there are no visible varicose veins. There is also little correlation between symptoms of varicose veins and their extent or size on examination. However, commonly reported symptoms include local discomfort of the varicose veins (pain, burning discomfort, aching and itching), generalised lower limb symptoms (aching, heaviness, and swelling), as well as complaints about cosmetic appearance.

Swelling and night cramps are commonly reported signs and symptoms of varicose veins in pregnancy and the treatment options for pregnant women have been evaluated in another Cochrane Review (Smyth 2015; Young 1998).

Clinicians use the classification system based on clinical manifestations (C), aetiologic factors (E), anatomic distribution (A), and underlying pathophysiology (P) (Eklöf 2004).

The majority of outcome measures, following treatment of varicose veins, are assessed subjectively, i.e. symptomatic improvement, cosmetic appearance, and quality of life measures. Outcomes that can be objectively evaluated include varicose vein recurrence and complication rates.

Varicose vein treatment options include conservative treatments (stockings, lifestyle modifications, sick leave, and medications) (Hamdan 2012), and a variety of surgical interventions. Traditionally, surgery is commonly used to treat 'main stem' varicose veins (where the vein is stripped). Endovascular procedures (where the vein is treated but not stripped) include endovenous laser therapy, radiofrequency ablation therapy, and sclerotherapy. Endovenous laser therapy and radiofrequency ablation therapy use specific catheter devices to heat and cause thrombotic and fibrotic closure (Piazza 2014) damaging the endothelium and the inner media. Sclerotherapy is another endovascular procedure that aims to obliterate the varicose veins using liquid or foam preparations injected inside the lumen of the varicose vein. The option of treatment to use is driven by the patient preference for less invasive procedures, the location of the vein, the size, aetiology, and the costs of the treatment options (Carrol 2013; Hamdan 2012; Piazza 2014).

Description of the intervention

Traditionally, sclerotherapy is defined as hardening of the vein. It is carried out by a needle puncture and injection of sclerosant (a substance that cause thrombosis and fibroses of the vein leading to reabsorption and resolution of the varicose vein). Vascular surgeons and dermatologists also use sclerotherapy to treat thread veins, telangiectasias, and reticular veins. In larger veins, it involves injecting liquid or foam medicines to cause sclerosing. Foam sclerotherapy is performed by mixing the liquid sclerosant agent with air. Despite presenting similar endpoints of other treatment options (for instance, phlebectomy, laser, radiofrequency), injection sclerotherapy can treat collateral veins which are sometimes difficult to treat with laser or radiofrequency and, probably, with lower costs (Carrol 2013).

How the intervention might work

The procedure of injecting sclerosants directly into the vessel damages the varicose vein without taking it out. This damage directly affects the endothelium and media wall causing inflammation and apoptosis (cell death) in up to 300 μm of the media vein wall (Whiteley 2016). This produces a fibrous cord known as sclerosis with a functional result similar to surgery (Rabe 2010), that is, occluding the varicose vein and eliminating the reflux (O'Hare 2010).

Why it is important to do this review

Currently, varicose veins may be treated by either conventional surgery, or newer techniques such as foam sclerotherapy, endovenous laser ablation (EVLT), and radiofrequency ablation (RFA). There is limited evidence to aid choice of which treatment is more effective. This update will provide the current evidence regarding injection sclerotherapy for varicose veins to inform health policymakers, health professional and consumers.

This review is the second update of a systematic review first published in 2002 and updated in 2006.

Objectives

To assess the effectiveness and safety of injection sclerotherapy for the treatment of varicose veins.

Methods

Criteria for considering studies for this review

Types of studies

We included all randomised controlled trials (RCTs) ‐ including cluster‐randomised trials and first phase cross‐over studies ‐ that used injection sclerotherapy for the treatment of varicose veins. Quasi‐randomised controlled trials, where the method of allocation was known but not considered truly random (e.g. alternation, date of birth), were excluded (Higgins 2017a).

Types of participants

We included all participants referred to a surgical outpatient clinic or primary care practitioner with symptomatic or cosmetic varicose veins, defined by clinical and/or ultrasound criteria as follows.

• Clinical criteria: CEAP C2, C3, and C4. CEAP C2 was defined as varicose veins more than 3 mm in diameter, C3 as a limb with oedema related to venous disease, and C4 as a limb with skin and subcutaneous changes (including pigmentation or eczema, and lipodermatosclerosis or atrophie blanche) related to venous disease (Eklöf 2004).

• Ultrasound criteria: handheld Doppler or duplex ultrasound defining varicose veins with or without the description of the depth of the vein related to the skin.

We excluded people diagnosed as having CEAP C1 (including telangiectasias and reticular varicose veins), people with a healed venous leg ulcer (CEAP C5), and people with a venous leg ulcer (CEAP C6).

Types of interventions

We included studies which used any injection sclerotherapy technique to treat varicose veins. We considered:

• Any sclerosant (e.g. sodium tetradecyl sulphate (STS), ethanolamine, polidocanol (Sclerovein, Aetoxysclerol, Aethoxysklerol, Aethoxysclerol, Atoxisclerol, Sotrauerix, Laureth 9), chrome alum (Scleremo), hypertonic saline (NaCl 17.55%);

• Any formulation (e.g. liquid, foam, and gas in foam (ambient air, O₂, CO₂));

• Any dosage (e.g. STS 0.2%, 0.5%, 1%, 3%, polidocanol 1%, 3%);

• Using any technique (e.g. injection in empty veins, bandaging and compression techniques and repeat treatment intervals).

We included studies comparing sclerotherapy with nonsurgical interventions (no intervention, conservative treatment including pharmacological therapy and/or graduated compression stockings, or sclerotherapy). Studies which compared sclerotherapy with an open surgical procedure (e.g. stripping), other endovascular approaches (laser or radiofrequency), and studies which investigated compression after sclerotherapy were not relevant for this review. They will be covered in other Cochrane Reviews.

We included the following comparisons:

• Sclerotherapy versus no intervention or no active intervention (placebo);

• Sclerotherapy versus sclerotherapy (sclerosant versus different sclerosant/formulation/dose/technique);

• Sclerotherapy versus different conservative treatment (including pharmacological therapy and/or graduated compression stockings when sclerotherapy was the intervention being tested). We excluded studies investigating compression after sclerotherapy treatment.

Types of outcome measures

We considered RCTs that assessed at least one of the following outcomes.

Primary outcomes

• Cosmetic appearance (assessed by any validated tool)

• Complications (including haematoma formation, skin pigmentation, ulceration and necrosis, superficial thrombophlebitis, thromboembolic complications (deep vein thrombosis and pulmonary embolism as defined by the authors of primary studies), and anaphylactic reaction)

Secondary outcomes

• Frequency of residual varicose veins (number of veins that failed obliteration, assessed clinically and/or by ultrasound examination)

• Quality of life measures (measured as any specific and validated questionnaires such as Assessment of Burden in Chronic disease – Venous (ABC‐V), VEnous INsufficiency Epidemiological and Economic Study on Quality of Life (VEINES‐QOL), Freiburg Life Quality Assessment (FLQA), Specific Quality of life and Outcomes Response – Venous (SQOR‐V), ChronIc Venous Insufficiency Questionnaire (CIVIQ), and Aberdeen Varicose Vein Questionnaire (AVVQ) (Launois 2015))

• Frequency of persistence of symptoms (such as pain, burning discomfort, aching, itching, limb heaviness, oedema and nocturnal cramps)

• Frequency of recurrent varicose veins and venous flare formation

We were guided by the time points reported in the individual studies.

Search methods for identification of studies

There were no restrictions on language, date, or status of publication.

Electronic searches

For this update, the Cochrane Vascular Information Specialist (CIS) first searched the following databases for relevant trials (20 July 2021):

• Cochrane Vascular Specialised Register (20 July 2021);

• Cochrane Central Register of Controlled Trials (20 July 2021) via the Cochrane Register of Studies Online.

See Appendix 1 for details of the search strategy used to search CENTRAL.

The CIS subsequently conducted systematic top‐up searches of the following databases for RCTs and CCTs without language, publication year or publication status restrictions:

• Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web) (20 July 2021);

• Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Register of Studies Online (20 July 2021);

• MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE) (20 July 2021);

• Embase via Ovid (20 July 2021);

• CINAHL via EBSCO (20 July 2021);

• AMED via Ovid (20 July 2021);

• LILACS (20 July 2021).

The CIS modelled search strategies for the listed databases on the search strategy designed for CENTRAL. Where appropriate, they were combined with adaptations of the Cochrane highly sensitive search strategy for identifying reports of randomised controlled trials and controlled clinical trials (Lefebvre 2011). Search strategies for major databases are provided in Appendix 2.

The Information Specialist also performed searches of the following trials registries on 20 July 2021:

• ClinicalTrials.gov (clinicaltrials.gov);

• World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch).

Searching other resources

• We searched the list of references from relevant studies to identify additional studies.

• We contacted pharmaceutical companies for information on any unpublished trials: Kreusller Pharma, Germany (polidocanol); BTG Pharma (polidocanol); LGM Pharma, USA (polidocanol); Shanghai T&W Pharmaceutical Co. Ltd., China (polidocanol); Omega Laboratories Limited, Montreal, Canada (hypertonic saline, polidocanol); Medeva Pharma Limited, Leatherhead, UK (ethanolamine); STD Pharmaceutical, Hereford, UK (STD).

• We contacted experts in this area about further unpublished or ongoing trials.

Data collection and analysis

Selection of studies

Two review authors (RAO and VV) independently assessed the search results. First, we screened titles and abstracts (if available). Additional review authors (RR and JCBS) resolved disagreements when necessary. We read the selected, potentially relevant trials in totality to check eligibility with the inclusion criteria. We recorded explanations/justifications for studies excluded at this second phase. If necessary, we contacted study authors for further information. To facilitate transparency, we presented a flow diagram showing the full list of included and excluded studies (PRISMA), as described by Liberati 2009.

Data extraction and management

Two review authors (RAO and VV) extracted data from the included studies. We double‐checked that data were inserted correctly into the form. We used a standard form to collect the following information on study characteristics and outcome data.

• Publication details (i.e. year, country, authors)

• Study design and methods (inclusion/exclusion criteria, randomisation method, allocation concealment, blinding)

• Setting

• Population data (i.e. age, the severity of the disease, type of varicose vein)

• Details of intervention (i.e. the dose of the sclerosant, technique, duration)

• Outcome measures (including effectiveness and adverse effects)

• The number of dropouts

• Length of follow‐up

• Types of data analyses (e.g. intention‐to‐treat, modified intention‐to‐treat)

• Any potential risk of bias

We discussed disagreements with the involvement of a third review author (RR or JCBS) until consensus was reached. Two review authors (RAO and VV) entered data into the Review Manager 5.3 software (Review Manager 2014).

Assessment of risk of bias in included studies

We used Cochrane's Risk of bias tool for assessing the risk of bias, by evaluating the following criteria.

• Random sequence generation (selection bias)

• Allocation concealment (selection bias)

• Blinding of participants and personnel (performance bias)

• Blinding of outcome assessment (detection bias)

• Incomplete outcome data (attrition bias)

• Selective reporting (reporting bias)

• Other bias (bias due to problems not covered in the other domains)

Two review authors (RAO and VV) independently evaluated the risk of bias for each included study. We resolved disagreements by discussion and consultation with a third review author (RR). We categorised the domains as 'high risk', 'low risk', or 'unclear risk' of bias according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017b). For the domains 'Blinding of participants and personnel,' 'Blinding of outcome assessment' and 'Incomplete outcome data,' we judged each outcome separately. Where studies did not report on an outcome, this section remained blank in the risk of bias table.

We considered the overall methodological quality of the studies as low risk if the sequence generation, allocation concealment and blinding (participants and personnel, and assessors) domains were all at low risk of bias. We considered the overall methodological quality of the studies as moderate if one of these domains was judged as being at unclear risk. Finally, we considered the overall methodological quality as 'high risk' if at least one of these domains was judged as 'high risk' of bias.

In any case of 'unclear risk' of bias, we emailed the study contact author and asked to clarify any questions. When no answer was received from the study author, the judgement was assessed as 'unclear risk' of bias.

Measures of treatment effect

We reported the results as risk ratios (RRs) with 95% confidence intervals (CIs) for dichotomous outcomes. For continuous outcomes, we expressed the results as mean differences (MDs) with 95% CIs or as the standardised mean difference (SMD) with 95% CIs if different assessment tools were used by the included studies to report the same outcome.

Unit of analysis issues

The unit of randomisation analysed in the included studies was likely to be the individual patient. In some studies, for outcomes such as complications, residual varicose veins, recurrent varicose veins, for which bilateral procedures are possible, each treated limb or varicose vein was considered as an individual unit of analysis. We presented the data for such studies only where the disparity between the units of analysis and number of randomised participants was small. If we pooled data, we performed a sensitivity analysis to examine the effects of excluding any trials with analyses that had not been corrected for this unit of analysis issue. We avoided unit of analysis issues related to repeated observations of the same outcome, such as results presented for several periods of follow‐up. If encountered, we planned to evaluate each outcome at different follow‐up periods: at the end of the treatment, at short‐term and at long‐term follow‐up (Deeks 2017).

Dealing with missing data

We emailed the study authors for additional information when there were missing, incomplete or unavailable data. We performed a search for protocols and/or further articles related to the included trials. When relevant data were unavailable, we presented and discussed the results in the main text of the review. In the case of missing data or dropouts, intention‐to‐treat analyses were conducted by imputing data which considered the worst‐case scenario for dichotomous data. We also undertook an intention‐to‐treat analysis for continuous outcomes, by imputing the 'last‐observation‐carried‐forward'. This action is justified due to the balanced loss of follow‐up that did not change results with data imputation tested by sensitivity analyses. We followed the methods described in Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions using the average of two units greater than the observed data in the intervention arm and decreasing two units in the control arm (Higgins 2017a). We did not find discrepancies in the trend of effect direction. Standard deviation imputation was used in one study (King 2015) using data from another included study (Todd 2014) since there was no other source for data imputation.

Assessment of heterogeneity

We considered heterogeneity using the Chi² test and the I² statistic, which indicates the degree of variation across studies related to heterogeneity rather than related to chance. We considered an I² value greater than 50% as substantial heterogeneity (Deeks 2017). We explored clinical and methodological differences as potential causes of heterogeneity.

Assessment of reporting biases

We planned to use funnel plots to explore possible reporting biases only if there were 10 or more included studies.

Data synthesis

We synthesised data using Review Manager 5.3 software (Review Manager 2014). Varicose vein conditions are heterogeneous by nature. Considering the heterogeneity of the disease and interventions used, we have used a random‐effects model for all meta‐analyses in the review. When a meta‐analysis was not possible, data were reported to measure effect size whenever possible. In all cases, we did a descriptive analysis of the study.

As there is considerable variation in the time point measures of outcomes, a stratification into three time intervals was used to conduct a meta‐analysis. We considered a short‐term period from the intervention to up to one month, an intermediate‐term period from one month to three months, and a long‐term period that included any outcome measured after three months. This alternative was considered because some outcomes could not be appropriately measured at the short‐term period (for instance, staining) and others continued to be a risk (for instance, risks of pulmonary embolism (PE) in people with deep vein thrombosis (DVT) and superficial thrombophlebitis) (Kearon 2016).

We intended to present the results from the following comparisons:

• Sclerotherapy versus no intervention or no active intervention (placebo)

• Sclerotherapy versus sclerotherapy (sclerosant versus different sclerosant/formulation/dose/technique)

• Sclerotherapy versus different conservative treatment (including pharmacological therapy and/or graduated compression stockings when sclerotherapy was the intervention being tested). We excluded studies investigating compression sclerotherapy after treatment.

Subgroup analysis and investigation of heterogeneity

We planned to evaluate the following subgroups if sufficient data had been available.

• CEAP classification: CEAP C2 (varicose veins with no oedema or staining), CEAP C3 (varicose veins with oedema and no staining), and CEAP C4 (a limb with skin and subcutaneous changes (including pigmentation or eczema, and lipodermatosclerosis or atrophie blanche) related to venous disease) (Eklöf 2004), since we expected worse results for the higher categories.

• Presence or absence of truncal incompetence: superficial venous incompetence demonstrated on Duplex ultrasound scanning, (i.e. great saphenous vein, short saphenous vein, and thigh and calf vein perforators (veins connecting superficial and deep venous systems)) and varicose veins with no evidence of superficial venous incompetence, since there is different prognosis and we expect different effects for bigger diameters.

• The depth of the vein analysed assessed by Duplex ultrasound scanning: veins below 5 mm deep, and veins more than 5 mm deep, since we expected less pigmentation in deeper veins.

• Symptomatic versus nonsymptomatic participants, since we expected fewer symptoms after treatment.

• Sclerosant concentration for truncal incompetence (e.g. polidocanol 1% and 3%), since we expected to define the best concentration to use.

• Comparison of the gas used when foam sclerotherapy (e.g. CO₂ versus ambient air), since we expected to assess if a gas was better than another one.

• Previous complications (thrombophlebitis, DVT), since we expected different clinical evolution in participants with previous complications.

Sensitivity analysis

We planned the following sensitivity analyses to investigate the robustness of our findings by:

• excluding studies with high risk of bias. The high risk of bias studies were defined as those presenting at least one of the following domains classified as 'high risk': selection, performance or detection;

• excluding studies funded by pharmaceuticals;

• excluding studies with data imputation (analysing by per‐protocol analysis).

Summary of findings and assessment of the certainty of the evidence

We created Summary of findings (SoF) tables to present the main findings of this review using GRADEpro GDT software (GRADEpro GDT 2015). We used the five GRADE criteria (risk of bias, inconsistency, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence as it related to the outcomes that contributed to the meta‐analyses. We used the methods and recommendations described in Section 8.5 (Higgins 2017b) and Chapter 12 (Schünemann 2017) of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017a). We explained all judgements to downgrade the certainty of the evidence in the footnotes, and we added comments to help readers understand the judgements of the review if necessary. We included the outcomes considered to be most clinically relevant in the Summary of findings tables.

• Cosmetic appearance

• Thromboembolic complications

• Residual varicose veins

• Quality of life

• Persistence of symptoms

• Recurrent varicose veins

We considered the following comparisons to be of the most clinical relevance and created a SoF table for each where data were available.

• Foam sclerotherapy (polidocanol 1%) versus placebo for treating varicose veins (Table 1)

• Foam sclerotherapy (polidocanol 3%) versus foam sclerotherapy (polidocanol 1%) for treating varicose veins (Table 2)

• Foam sclerotherapy versus liquid sclerotherapy for treating varicose veins (Table 3)

• Sclerotherapy with STS 1% versus sclerotherapy with polidocanol 1% for treating varicose veins (Table 4)

Results

Description of studies

Please see Figure 1 for the search results.

1.

Study flow diagram.

Results of the search

We assessed the reports identified by the searches and cross‐checked these with the reports provided by BTG Pharma. No additional studies were identified in the reports provided by BTG Pharma.

We reassessed the included and excluded studies from the previous version (Tisi 2006), as changes have been made to the inclusion criteria. See Differences between protocol and review. Four studies excluded in the previous version of the review are now included as they met the inclusion criteria (Martimbeau 2003a; Martimbeau 2003b; Wright 2006; Zeh 2003). Studies which were included in the previous version but which compared sclerotherapy with open surgery (e.g. stripping) or investigated compression after sclerotherapy were reassessed as not relevant for this update.

Included studies

We identified 23 new studies for this update (Alos 2006; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Hamel‐Desnos 2007; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014). This review includes a total of 28 RCTs (43 records), and they are described in the Characteristics of included studies tables (Abramowitz 1973; Alos 2006; Belcaro 2003b; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Schadeck 1995a; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014). These studies have the following characteristics.

• Design: all included studies were randomised controlled trials (RCTs) with a parallel‐design intervention.

• Setting

  • Location of participants

    • Inpatients: seven RCTs (Abramowitz 1973; Belcaro 2003b; Chleir 1997; Demagny 2002; Kahle 2004; Wright 2006; Zhang 2014);

    • Outpatients: 10 RCTs (Alos 2006; Ceulen 2007; Goldman 2002; Hamel‐Desnos 2007; King 2015; Ouvry 2008; Rabe 2008; Santos 2019; Todd 2014; Ukritmanoroat 2011);

    • Not specified: 11 RCTs (Blaise 2010; Gibson 2010; Martimbeau 2003a; Martimbeau 2003b; Ragg 2015; Rao 2005; Schadeck 1995a; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009).

• Country

  • Brazil: one RCT (Santos 2019);

  • Chile: one RCT (Zeh 2003);

  • China: two RCTs (Zhang 2009; Zhang 2014);

  • France: six RCTs (Blaise 2010; Chleir 1997; Demagny 2002; Hamel‐Desnos 2007; Ouvry 2008; Schadeck 1995a);

  • Germany: three RCTs (Kahle 2004; Rabe 2008; Ragg 2015);

  • Japan: two RCTs (Yamaki 2009; Yamaki 2012);

  • South Africa: one RCT (Abramowitz 1973);

  • Spain: one RCT (Alos 2006);

  • Thailand: one RCT (Ukritmanoroat 2011);

  • The Netherlands: one RCT (Ceulen 2007);

  • United States of America: seven RCTs (Gibson 2010; Goldman 2002; King 2015; Martimbeau 2003a; Martimbeau 2003b; Rao 2005; Todd 2014);

  • Multicentre trials involving Belgium, France, Greece, Italy, South Africa, Sweden, The Netherlands, and United Kingdom: two RCTs (Belcaro 2003b; Wright 2006).

• Study duration (calculated by adding the time from start of recruitment to follow‐up duration), when described, ranged from 12 months (Yamaki 2012) to 10 years (Belcaro 2003b). The follow‐up ranged from three minutes (Schadeck 1995a) to 10 years (Belcaro 2003b).

• Characteristics of the participants:

  • Age

    • Nine studies did not state the age of the participants (Abramowitz 1973; Goldman 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Rao 2005; Schadeck 1995a; Zeh 2003; Zhang 2009);

    • Six described the mean age ranging from 41.9 to 59 years of age (Alos 2006; Belcaro 2003b; Chleir 1997; Demagny 2002; Gibson 2010; Ukritmanoroat 2011);

    • Twelve studies described the mean age for each intervention and control group, ranging from a minimum of 28 years to a maximum of 72 years of age (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Ouvry 2008; Rabe 2008; Todd 2014; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2014);

    • One study included people older than 18 years (group A: 44.8 (SD ± 11.33), group B: 47.5 (SD ± 13.47)) (Santos 2019).

  • Gender

    • Nine studies did not state the gender of the participants (Chleir 1997; Gibson 2010; Goldman 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Schadeck 1995a; Zeh 2003; Zhang 2009);

    • One study included only women (Abramowitz 1973);

    • All other studies had groups with higher proportions of women than men. The lowest proportion of women included was 60% in two groups (Ragg 2015; Zhang 2014), and the highest was 95% (Rao 2005).

  • CEAP classification

    • Two studies were published before or up to the first CEAP publication (Abramowitz 1973; Schadeck 1995a). As the description of the studies matched with CEAP C2 or C3, both studies were included.

    • 13 studies did not clearly state the CEAP classification (Alos 2006; Belcaro 2003b; Chleir 1997; Demagny 2002; Gibson 2010; Goldman 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Ragg 2015; Rao 2005; Ukritmanoroat 2011; Zeh 2003). All of these studies suggested the inclusion of people with CEAP C2 and/or C3 and so were included in the review.

    • 12 studies stated the CEAP classification without discriminating between subgroups ranging from C1 to C6 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Ouvry 2008; Rabe 2008; Todd 2014; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2009; Zhang 2014).

    • Two studies also allocated people with CEAP C1 (Ceulen 2007; Zhang 2009), one study included people with CEAP C2 to C5 (Ceulen 2007), and one study included people with CEAP C2 (Zhang 2009).

    • One study included only people with CEAP C3 (Santos 2019).

    • Three studies included people with CEAP C2 to C4 (Wright 2006; Yamaki 2009; Yamaki 2012).

    • The remaining seven studies included people in CEAP groups C2 to C5 or C6 (Blaise 2010; Hamel‐Desnos 2007; King 2015; Ouvry 2008; Rabe 2008; Todd 2014; Zhang 2014).

  • Sample size ranged from 10 participants (Chleir 1997) to 877 participants (Belcaro 2003b).

• Unit of analysis:

  • 23 studies considered the individual as the unit of analysis (Abramowitz 1973; Belcaro 2003b; Blaise 2010; Gibson 2010; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Schadeck 1995a; Todd 2014; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014);

  • Two considered the lower leg (Ceulen 2007; Chleir 1997);

  • Three considered the vein as the unit of analysis (Alos 2006; Demagny 2002; Ukritmanoroat 2011).

• Characteristics of the interventions/comparator and comparisons:

  • Injection technique: different techniques were used to inject the sclerosant. Most of them were described as traditional injection sclerotherapy where a traditional needle and syringe is used to inject the intravenous medicine. One study described the technique as one injection sclerotherapy (Gibson 2010), and four studies used a catheter to deliver the medicine (Ceulen 2007; Ragg 2015; Santos 2019; Zhang 2014). We do not believe this intervention can be considered pharmacomechanical treatment since the catheter had the sole function to deliver the drug in this case.

  • Interventions and comparisons:

    • Sclerotherapy versus compression: one RCT (Abramowitz 1973);

    • Sclerosant versus placebo: five RCTs (Gibson 2010; Kahle 2004; King 2015; Todd 2014; Zhang 2009);

    • Sclerosant versus sclerosant: 23 RCTs (Alos 2006; Belcaro 2003b; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Goldman 2002; Hamel‐Desnos 2007; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Schadeck 1995a; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2014). Two of these studies (King 2015; Todd 2014) also compared sclerosant to placebo and were included in both comparator groups;

    • Foam versus liquid sclerotherapy: six RCTs (Alos 2006; Demagny 2002; Martimbeau 2003a; Ouvry 2008; Rabe 2008; Ukritmanoroat 2011);

    • Sodium tetradecyl sulphate (STS) liquid versus liquid plus foam: one RCT (Belcaro 2003b);

    • Polidocanol foam in different concentrations: five RCTs (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Todd 2014);

    • Polidocanol foam 1% versus any other sclerosant liquid or foam in any concentration: one RCT (Wright 2006);

    • Single versus fractionated injection for sclerotherapy: two RCTs (Chleir 1997; Yamaki 2009);

    • Ultrasound‐guided foam sclerotherapy versus visual‐foam sclerotherapy: one RCT (Yamaki 2012);

    • Catheter directed sclerotherapy using antegrade versus retrograde approach to inject sclerosant: one RCT (Zhang 2014);

    • Injection by needle versus microcatheter and catheter: one RCT (Ragg 2015);

    • Injection by needle versus catheter deliver of the foam associated with tumescent technique:one RCT (Santos 2019);

    • STS versus polidocanol: three RCTs (Goldman 2002; Rao 2005; Schadeck 1995a);

    • Addition of perfluropropane‐filled albumin microspheres of foam STS with air filled foam STS versus polidocanol foam sclerotherapy alone: one RCT (Martimbeau 2003b);

    • Addition of low molecular weight synthetic protein (Gelufusine) to polidocanol foam versus polidocanol foam sclerotherapy alone: one RCT (Zeh 2003).

  • Sclerotherapy agent and dose:

    • Four studies described the sclerotherapy agent but not the volume of sclerosant administered using polidocanol (Ceulen 2007; Zhang 2009), or polidocanol and/or STS (Goldman 2002; Schadeck 1995a);

    • 19 studies described the use of polidocanol (Alos 2006; Blaise 2010; Ceulen 2007; Gibson 2010; Hamel‐Desnos 2007; Kahle 2004; King 2015; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019, Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014) ranging from 0.5 mL liquid (Alos 2006; Yamaki 2009), to 60 mL foam sclerotherapy (Wright 2006);

    • Seven studies used STS (Abramowitz 1973; Belcaro 2003b; Chleir 1997; Demagny 2002; Goldman 2002; Martimbeau 2003a; Martimbeau 2003b) ranging from 0.5 mL liquid (Abramowitz 1973) to up to 6 mL of foam (Belcaro 2003b);

    • One study (Rao 2005) described the use of 1 mL from the liquid form of polidocanol or STS to produce the foam in a ratio 1 to 4 to be injected.

• Outcomes ‐ time points used were variable depending on the outcome to be measured. These ranged from three minutes for venous spasm (Schadeck 1995a) to 10 years for recurrent varicose veins (Belcaro 2003b). Outcomes measured are described:

  • Cosmetic appearance was assessed in seven RCTs (Abramowitz 1973; Alos 2006; Ceulen 2007; Goldman 2002; King 2015; Rao 2005; Todd 2014). Abramowitz 1973 described combined outcomes of symptomatic improvement and cosmetic results quoted as good, fair, and poor results. Alos 2006 described cosmetic appearance as patient satisfaction using a scale from zero (minimum satisfaction) to ten (maximum satisfaction). Ceulen 2007 described cosmetic appearance as cosmetic improvement described by participants. Goldman 2002 and Rao 2005 described cosmetic appearance as disappearance scale and satisfaction scale ranging from 1 (worse than before treatment) to 5 (complete disappearance). Rao 2005 pooled results as tolerability and satisfaction. King 2015 and Todd 2014 described by IPR‐V (Independent photography review: Visible Varicose veins; scale: 0 to 4; 0 being the best appearance (0: none, 1: mild, 2: moderate, 3: severe, 4: very severe)) and PA‐V (patient self‐assessment of visible varicose veins);

  • At least one complication was described in 23 studies (Abramowitz 1973; Alos 2006; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2014);

  • Residual varicose veins were assessed in 25 studies (Alos 2006; Belcaro 2003b; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Schadeck 1995a; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014);

  • Recurrent varicose veins were described in six studies (Belcaro 2003b; Blaise 2010; Martimbeau 2003a; Martimbeau 2003b; Yamaki 2009; Yamaki 2012);

  • Quality of life (QoL) was measured in six studies (Blaise 2010; Gibson 2010; King 2015; Rabe 2008; Santos 2019; Todd 2014). Blaise 2010 and Rabe 2008 used the CIVIQ questionnaire (using a five‐point Likert scale in the items) to score QoL. Gibson 2010 used the modified venous insufficiency epidemiological and economic study QoL/symptoms (m‐VEINES‐QOL/Sym), Heavy legs, Achy legs, Swelling, Throbbing, and Itching (HASTI) questionnaire, and the chronic venous insufficiency questionnaire 2 (CIVIQ‐2). King 2015 used the modified venous insufficiency epidemiological and economic study QoL/symptoms (m‐VEINES‐QOL/Sym). Todd 2014 used the venous insufficiency epidemiological and economic study score (VEINES‐QOL). Santos 2019 used Aberdeen Varicose Vein Questionnaire (AVVQ);

  • Persistence of symptoms were described in seven studies (Abramowitz 1973; Ceulen 2007; Chleir 1997; Gibson 2010; King 2015; Todd 2014; Yamaki 2012). King 2015, Todd 2014 and Yamaki 2012 used venous clinical severity score (VCSS). Gibson 2010 used Veines‐QOL/Sym and Abramowitz 1973, Ceulen 2007 and Chleir 1997 reported this as 'clinical improvements'; and

  • 18 other studies described sclerosis or occlusion or venous shrinkage or venous spasm (Alos 2006; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Hamel‐Desnos 2007; Kahle 2004; Ouvry 2008; Rabe 2008; Ragg 2015; Schadeck 1995a; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2014). The outcomes were evaluated by duplex ultrasound.

Excluded studies

We excluded a total of 13 studies from this review update as they investigated sclerotherapy versus an intervention other than nonsurgical interventions (that is, no intervention, conservative treatment including pharmacological therapy and/or graduated compression stockings, or sclerotherapy) (Ariyoshi 1996; Biemans 2013; Brittenden 2015; Lattimer 2012; Leung 2016; Mishra 2016; NCT02462720; Ragg 2015a; Rasmussen 2011; Theivacumar 2008; Vahaaho 2015; Vernermo 2016; Zafarghandi 2017). See the Characteristics of excluded studies tables.

Studies which investigated compression after sclerotherapy or compared sclerotherapy with open surgery (e.g. stripping), other endovascular techniques (endolaser or radiofrequency), and compression therapy were assessed as not relevant as they clearly did not meet the inclusion criteria (Criteria for considering studies for this review).

Ongoing studies

We did not identify any ongoing studies.

Awaiting classification

Five studies were assessed as 'awaiting classification' as in all cases we were unable to contact authors and there was a lack of information to include or exclude them (Labas 2003; Satokawa 2003; Schadeck 1995b; SLCTR/2008/014; Varnagy 1985). In two of these studies, there was limited information (intervention groups were not clear) (Satokawa 2003; Schadeck 1995b). Labas 2003 was included in the previous version of the review, and Schadeck 1995b was excluded in the previous version of the review. On reassessment, we did not find sufficient information to either include or exclude these two studies. We are not sure if one study is a randomised or quasi‐randomised trial (Varnagy 1985). Should further information be provided, we will include these studies in future updates.

Risk of bias in included studies

Two review authors (RAO, VV) independently assessed the included studies for risk of bias following the recommendations in the Cochrane Handbook of Systematic Reviews of Interventions (Higgins 2017a). Any disagreements were resolved by a third author (RR) in consensus with the group. Details of the risk of bias judgements can be found in the Risk of bias tables, within the Characteristics of included studies tables. A graphical presentation of the risk of bias is shown in Figure 2. For the domains of 'Blinding of participants and personnel', 'Blinding of outcome assessment', and 'Incomplete outcome data' we have also assessed each study by outcome. We have also distinguished between studies comparing intervention versus intervention and intervention versus placebo as we believe this can impact the risk of bias. Not all studies reported on all outcomes and those sections were left blank where appropriate in the tables.

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

The randomisation sequence generation was rated as low risk in 10 RCTs (Belcaro 2003b; Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Rabe 2008; Santos 2019; Todd 2014; Yamaki 2009; Yamaki 2012) and unclear risk for 18 RCTs (Abramowitz 1973; Alos 2006; Chleir 1997; Demagny 2002; Gibson 2010; Goldman 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Ragg 2015; Rao 2005; Schadeck 1995a; Ukritmanoroat 2011; Wright 2006; Zeh 2003; Zhang 2009; Zhang 2014).

Allocation concealment

One RCT described opening the study randomisation code after deciding the intervention and was judged as having high risk of allocation concealment (Belcaro 2003b). None of the remaining RCTs described allocation concealment, so all were judged as having unclear risk of bias.

Blinding

Performance bias

Cosmetic appearance (assessed by any tool)

Seven RCTs reported on this outcome (Abramowitz 1973; Alos 2006; Ceulen 2007; Goldman 2002; King 2015; Rao 2005; Todd 2014). Four were judged to be at low risk of bias because the blinding of participants and personnel was described and considered appropriate (Goldman 2002; King 2015; Rao 2005; Todd 2014). Two were judged to be at unclear risk because they did not describe the blinding method making judgement impossible (Abramowitz 1973; Ceulen 2007), and one was judged to be at high risk of bias because the nature of the intervention precluded blinding (Alos 2006). Two RCTs compared sclerosant with sclerosant and with placebo (King 2015; Todd 2014). They were considered as being at high risk of bias when any comparison was made with placebo since, due to the nature of the intervention, it was impossible to blind the placebo group (placebos had different physical appearance). We added columns in the Risk of bias table to describe the risk of bias of these RCTs comparing intervention versus intervention and intervention versus placebo.

Complications (at least one of the following complications: haematoma formation, skin pigmentation, ulceration and necrosis, superficial thrombophlebitis, deep vein thrombosis and anaphylactic reaction)

Twenty‐three RCTs reported on this outcome (Abramowitz 1973; Alos 2006; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2014). Seven were judged as being at low risk of bias since blinding of participants and personnel was described, and we judged it as appropriate (Blaise 2010; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Rao 2005; Todd 2014). Seven RCTs were judged to be at unclear risk of bias because they did not describe the blinding method (Abramowitz 1973; Ceulen 2007; Chleir 1997; Demagny 2002; Martimbeau 2003a; Martimbeau 2003b; Zhang 2014). Nine RCTs were judged as being at high risk of bias as the nature of the intervention precluded blinding (Alos 2006; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012). As above, two RCTs were considered as having high risk of bias when any comparison was made with placebo (King 2015; Todd 2014).

Residual varicose veins (veins that failed obliteration, assessed clinically and/or by ultrasound examination)

Twenty‐five RCTs reported on this outcome (Alos 2006; Belcaro 2003b; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Schadeck 1995a; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014). Five studies described properly the blinding of participants and personnel and were judged to be at low risk of bias (Blaise 2010; Hamel‐Desnos 2007; Kahle 2004; King 2015; Todd 2014). Nine studies were considered to be at unclear risk of bias because they did not describe the blinding method despite it being possible to blind due to the nature of the procedure (Belcaro 2003b; Ceulen 2007; Chleir 1997; Demagny 2002; Martimbeau 2003a; Martimbeau 2003b; Schadeck 1995a; Zeh 2003; Zhang 2014). Eleven RCTs were judged as being at high risk of bias (Alos 2006; Gibson 2010; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2009). For ten of these RCTs, it was considered impossible to blind due to the nature of the intervention (Gibson 2010; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2009). One RCT was described as blinded but the method was not described. We believe that the nature of the intervention precluded blinding when comparing foam versus liquid sclerotherapy since even a small amount of the sclerosant may flow back to the injection site (Ragg 2015). One RCT described blinding of the patient and ultrasonograph professional. We believe that the nature of the intervention (foam versus liquid) precluded blinding of the interventionist and the ultrasonography professional (Alos 2006). As before, King 2015 and Todd 2014 were considered as being at high risk of bias when any comparison was made with placebo.

Recurrent varicose veins and venous flare formation

Six RCTs reported on this outcome (Belcaro 2003b; Blaise 2010; Martimbeau 2003a; Martimbeau 2003b; Yamaki 2009; Yamaki 2012). One study was judged to be at low risk of bias because it described blinding of participants and personnel and it seemed to be appropriate (Blaise 2010). Three were judged to be at unclear risk of bias because they did not state the blinding method (Belcaro 2003b; Martimbeau 2003a; Martimbeau 2003b). Two RCTs were considered to be at high risk of bias since the nature of the intervention precluded any blinding (Yamaki 2009; Yamaki 2012).

Quality of life (using any specific, validated questionnaire)

Six RCTs reported on this outcome (Blaise 2010; Gibson 2010; King 2015; Rabe 2008; Santos 2019; Todd 2014). Three of them were judged to be at low risk of bias because they described blinding of participants and personnel and it seemed to be appropriate (Blaise 2010; King 2015; Todd 2014). As before, King 2015 and Todd 2014 were considered as being at high risk of bias when any comparison was made with placebo. Three RCTs were considered to be at high risk of bias since it was impossible to blind personnel due to the nature of the procedure (Gibson 2010; Rabe 2008; Santos 2019).

Frequency of persistence of symptoms

Seven RCTs reported on this outcome (Abramowitz 1973; Ceulen 2007; Chleir 1997; Gibson 2010; King 2015; Todd 2014; Yamaki 2012). Two of these studies were considered to be at low risk of bias since blinding of participants and personnel was described and seemed to be appropriate (King 2015; Todd 2014). Three studies were considered to be at unclear risk of bias since they did not state who was blinded or how the blinding was carried out (Abramowitz 1973; Ceulen 2007; Chleir 1997). Two studies were judged to be at high risk of bias due to the nature of the intervention preventing blinding of personnel (Gibson 2010; Yamaki 2012).

Detection bias

Cosmetic appearance (assessed by any tool)

From the seven RCTs reporting on this outcome (Abramowitz 1973; Alos 2006; Ceulen 2007; Goldman 2002; King 2015; Rao 2005; Todd 2014), five RCTs were judged as being at low risk of bias since they described blinding the outcome assessment, and it seemed to be appropriate (Alos 2006; Goldman 2002; King 2015; Rao 2005; Todd 2014). Two RCTs did not describe the blinding of the assessors of the outcomes. We had no data to judge if they did or not. So, we judged the RCTs as having unclear risk of bias (Abramowitz 1973; Ceulen 2007).

Complications (at least one of the following complications: haematoma formation, skin pigmentation, ulceration and necrosis, superficial thrombophlebitis, deep vein thrombosis and anaphylactic reaction)

Twenty‐three RCTs reported on this outcome, and ten studies were judged as having low risk of bias since they described the blinding of the outcome assessment and it seemed to be appropriate (Alos 2006; Blaise 2010; Goldman 2002; Hamel‐Desnos 2007; King 2015; Rao 2005; Todd 2014; Ukritmanoroat 2011; Yamaki 2009; Yamaki 2012). Ten studies were judged as having unclear risk of bias since the authors did not include any text which enabled us to make an assessment. We preferred to judge them as having unclear risk of bias since some studies could have undertaken this task (Abramowitz 1973; Ceulen 2007; Chleir 1997; Demagny 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Ragg 2015; Zhang 2014). Three studies were declared as non‐blinded studies and were judged as having high risk of bias (Rabe 2008; Santos 2019; Wright 2006).

Residual varicose veins (veins that failed obliteration, assessed clinically and/or by ultrasound examination)

Twenty‐five studies reported on this outcome. Eight studies were judged as having low risk of bias since they described blinding of the outcome assessment and it seemed to be appropriate (Alos 2006; Blaise 2010; Hamel‐Desnos 2007; King 2015; Todd 2014; Ukritmanoroat 2011; Yamaki 2009; Yamaki 2012). Fourteen studies did not describe any kind of the blinding of assessment (Belcaro 2003b; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Ragg 2015; Schadeck 1995a; Zeh 2003; Zhang 2009; Zhang 2014). As we were not sure about the method, we judged these studies as having unclear risk of bias. Three studies were judged as having high risk of bias (Rabe 2008; Santos 2019; Wright 2006). One study stated that it was impossible to perform a 'double‐blind' study (Rabe 2008). The second study was described as an 'open‐label' study (Wright 2006). We decided to judge these as having high risk of bias as we had no response from the authors for further information. One study was reported as an unblinded study for the outcome assessment (Santos 2019).

Recurrent varicose veins and venous flare formation

Six studies reported on this outcome; three studies were judged as having low risk of bias since they described the blinding of outcome assessors and it seemed to be appropriate (Blaise 2010; Yamaki 2009; Yamaki 2012), and three studies were judged as having unclear risk of bias since they did not state any blinding for outcome assessment (Belcaro 2003b; Martimbeau 2003a; Martimbeau 2003b).

Quality of life (using any specific, validated questionnaire)

Six RCTs reported on this outcome. Three studies were judged as having low risk of bias as the authors described assessment blinding, and it seemed to be appropriate (Blaise 2010; King 2015; Todd 2014), and two RCTs were judged as having unclear risk of bias since they did not mention blinding of the assessors of the outcomes (Gibson 2010; Rabe 2008). One study was described as not being able to undertake double‐blinding (Rabe 2008). We believe that, even in this case, the blinding of outcome assessment was possible and we judged it as having high risk of bias. One RCT reported outcome assessors were unblinded and was judged as having high risk of bias (Santos 2019).

Frequency of persistence of symptoms

Seven RCTs reported on this outcome. Three studies were judged as having low risk of bias since the authors described blinding of outcome assessment and it seemed to be appropriate (King 2015; Todd 2014; Yamaki 2012), and four RCTs were judged as having unclear risk of bias since they did not describe any blinding of the assessors (Abramowitz 1973; Ceulen 2007; Chleir 1997; Gibson 2010).

Incomplete outcome data

Cosmetic appearance

Seven studies reported on this outcome (Abramowitz 1973; Alos 2006; Ceulen 2007; Goldman 2002; King 2015; Rao 2005; Todd 2014). Five were judged to be at low risk of bias because they described all proposed outcomes and there was no indication of incomplete outcome data (Alos 2006; Goldman 2002; King 2015; Rao 2005; Todd 2014). One study had 50% of dropouts in the control group and was judged as having unclear risk of bias (Abramowitz 1973). One study was judged to be at unclear risk because there was no description of the dropouts (Ceulen 2007). No study was judged to be at high risk of bias.

Complications

Twenty‐three RCTs reported on this outcome (Abramowitz 1973; Alos 2006; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Goldman 2002; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Santos 2019; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2014). Thirteen were considered to be at low risk of bias because they reported all the proposed outcomes, there were a low number of dropouts (less than 20%), and so were considered as appropriate (Alos 2006; Blaise 2010; Chleir 1997; Kahle 2004; King 2015; Ouvry 2008; Rabe 2008; Rao 2005; Santos 2019; Todd 2014; Yamaki 2009; Yamaki 2012; Zhang 2014). Eight RCTs were judged to be at unclear risk of bias due to lack of dropout data to judge (Ceulen 2007; Demagny 2002; Goldman 2002; Hamel‐Desnos 2007; Martimbeau 2003a; Martimbeau 2003b; Ragg 2015; Ukritmanoroat 2011), one RCT had 50% of dropouts in the control group and was judged as having unclear risk of bias (Abramowitz 1973), and one RCT was considered to be at high risk of bias because six participants in the intervention group were removed by the sponsor and reasons for that were not clear to us (Wright 2006).

Residual varicose veins

Twenty‐five RCTs reported on this outcome (Alos 2006; Belcaro 2003b; Blaise 2010; Ceulen 2007; Chleir 1997; Demagny 2002; Gibson 2010; Hamel‐Desnos 2007; Kahle 2004; King 2015; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Santos 2019; Schadeck 1995a; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014). Sixteen RCTs were considered to be at low risk of bias because they described all data and dropouts and were judged as having appropriate completeness of data (Alos 2006; Belcaro 2003b; Blaise 2010; Chleir 1997; Gibson 2010; Kahle 2004; King 2015; Ouvry 2008; Rabe 2008; Santos 2019; Schadeck 1995a; Todd 2014; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2014). Eight studies were judged to be at unclear risk of bias because they did not describe dropouts, and we did not know to what extent this could impact the results (Ceulen 2007; Demagny 2002; Hamel‐Desnos 2007; Martimbeau 2003a; Martimbeau 2003b; Ragg 2015; Ukritmanoroat 2011; Zhang 2009). One RCT was judged as having high risk of bias as there was an intervention by the sponsor (six participants removed and reasons for that were unclear) (Wright 2006).

Recurrent varicose veins and venous flare formation

Six RCTs reported on this outcome (Belcaro 2003b; Blaise 2010; Martimbeau 2003a; Martimbeau 2003b; Yamaki 2009; Yamaki 2012). Four RCTs were considered to be at low risk of bias since participants were described and reasons for dropouts reported (Belcaro 2003b; Blaise 2010; Yamaki 2009; Yamaki 2012). Two RCTs provided data in pooled percentages, we were not able to use it to determine the impact of data in the study, and they did not describe dropouts, so were considered to be at unclear risk of bias (Martimbeau 2003a; Martimbeau 2003b).

Quality of life

Six RCTs reported on this outcome and all of them were judged as being at low risk of bias because all studies reported numbers and reasons for dropouts (Blaise 2010; Gibson 2010; King 2015; Rabe 2008; Santos 2019; Todd 2014).

Frequency of persistence of symptoms:

Seven RCTs reported on this outcome (Abramowitz 1973; Ceulen 2007; Chleir 1997; Gibson 2010; King 2015; Todd 2014; Yamaki 2012). Five RCTs were considered to be at low risk of bias since they described numbers and reasons for dropouts (Abramowitz 1973; Gibson 2010; King 2015; Todd 2014; Yamaki 2012). Two RCTs were judged to be at high risk of bias because they did not describe dropouts, and we did not know how this could impact results (Ceulen 2007; Chleir 1997).

Selective reporting

Seven RCTs were judged as being at low risk of reporting bias since the protocol was available and reporting was done according to what was planned (Belcaro 2003b; Blaise 2010; Gibson 2010; Hamel‐Desnos 2007; King 2015; Santos 2019; Todd 2014). None of the remaining RCTs provided a protocol. These RCTs were judged as being at unclear risk of bias (Abramowitz 1973; Alos 2006; Ceulen 2007; Chleir 1997; Demagny 2002; Goldman 2002; Kahle 2004; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Rao 2005; Schadeck 1995a; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zeh 2003; Zhang 2009; Zhang 2014).

Other potential sources of bias

It is estimated that an increase of 48% of size effect occurs in RCTs with fewer than 50 participants (Dechartres 2013). So we graded included studies with fewer than 50 participants as having unclear risk of bias since we could not estimate the impact of such studies in a systematic review and a meta‐analysis. We also considered the diameter chosen to be sclerosed. If there was restriction in the vein diameter, external validity would not be applicable to all varicose vein diameters. When there was no imbalance between intervention and control groups and these previous aspects were observed, the studies were judged as having low risk of other bias. Nine RCTs were therefore classified as having low risk (Alos 2006; Belcaro 2003b; Blaise 2010; Hamel‐Desnos 2007; King 2015; Rabe 2008; Todd 2014; Yamaki 2009; Yamaki 2012). All remaining RCTs were judged as having unclear risk as 12 studies had fewer than 50 participants (Abramowitz 1973; Ceulen 2007; Gibson 2010; Goldman 2002; Kahle 2004; Ouvry 2008; Ragg 2015; Rao 2005; Santos 2019; Schadeck 1995a; Zeh 2003; Zhang 2014). Nine RCTs did not describe baseline characteristics (Abramowitz 1973; Chleir 1997; Demagny 2002; Gibson 2010; Goldman 2002; Ragg 2015; Rao 2005; Schadeck 1995a; Ukritmanoroat 2011); one RCT described imbalances between groups (Ceulen 2007); two RCTs provided insufficient data to judge any imbalance between groups (Wright 2006; Zhang 2014); one RCT provided insufficient data to evaluate (Zhang 2009); and two RCTs did not provide any data regarding the baseline characteristics (Martimbeau 2003a; Martimbeau 2003b).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Sclerotherapy versus no intervention or placebo

Comparison 1: Sclerotherapy versus no intervention

None of the included RCTs compared sclerotherapy with no intervention.

Comparison 2: Foam sclerotherapy versus placebo

Five RCTs compared sclerosant with placebo (Gibson 2010; Kahle 2004; King 2015; Todd 2014; Zhang 2009). One study did not provide sufficient data regarding the number of participants included in each group or details on the concentration of polidocanol used, so we could not include it in the analysis (Zhang 2009). One study referred to the use of polidocanol 2% or 3% (Kahle 2004). As the concentrations of polidocanol were different from the other studies, the data were not used in meta‐analysis. We have chosen polidocanol 1% foam to compare with placebo because one study considered this concentration to be necessary to achieve good sclerotherapy results (King 2015). We understand this concentration is clinically relevant for foam sclerotherapy. As described in Data synthesis, we decided to use short‐, intermediate‐, and long‐term time periods according to the moment of the outcome measurement (up to one month, one to three months, and after three months, respectively). There were no available data to perform subgroup analysis by CEAP classification, presence or absence of truncal incompetence, depth of the vein, symptomatic versus nonsymptomatic participants, and previous complications (thrombophlebitis, DVT). Subgroup analysis from different foam concentrations was performed in the comparator 'Sclerotherapy versus sclerotherapy' in the subgroup of 'Foam sclerotherapy versus foam sclerotherapy with different concentrations' due to the importance of the issue. Care was taken to prevent double counting. Due to limitations in the blinding due to the nature of the intervention, all studies were at high risk of bias for participants and personnel. See Table 1.

Cosmetic appearance

Two RCTs reported on this outcome, both at intermediate‐term periods (King 2015; Todd 2014). Cosmetic appearance was reported using independent photography review ‐ visible varicose veins scores (IPR‐V) or using patient self‐assessment of visible varicose veins scores (PA‐V). In both cases, there was a possible benefit in favour of foam sclerotherapy if polidocanol 1% or 2% foam was used: polidocanol 1%: IPR‐V mean difference (MD) ‐0.76, 95% CI ‐0.91 to ‐0.60; 2 studies, 223 participants; very low‐certainty evidence; polidocanol 2%: IPR‐V MD ‐0.90, 95% CI ‐1.11 to ‐0.69; 1 study, 119 participants; (Analysis 1.1); polidocanol 1%: PA‐V MD ‐1.46, 95% CI ‐1.72 to ‐1.20; 2 studies, 223 participants; and polidocanol 2%: PA‐V MD ‐1.60, 95% CI ‐1.95 to ‐1.25; 1 study, 119 participants; (Analysis 1.2) (see Table 1). We have chosen IPR for grading the evidence since it is independent of participants and personnel. We downgraded the certainty of the evidence due to risk of bias concerns and indirectness (Todd 2014 included 8.8% CEAP C5 and C6 participants in the placebo group compared to 3.4% CEAP C5 and C6 participants in the polidocanol 1% group).

1.1. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 1: Cosmetic appearance: medical related IPR‐V adjusted mean change from baseline (intermediate term)

1.2. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 2: Cosmetic appearance: participant self assessment PA‐V adjusted mean change from baseline (intermediate term)

Complications

Four RCTs reported on complications including DVT, phlebitis or thrombophlebitis, haemorrhagic and neurologic complications (Gibson 2010; Kahle 2004; King 2015; Todd 2014). We assessed the outcome of DVT using the GRADE approach due to severity of the conditions related to this event.

Deep vein thrombosis rates may be increased following polidocanol 1% in the short and intermediate term (RR 5.1, 95% CI 1.3 to 20.01; 3 studies, 302 participants; very low‐certainty evidence) (Analysis 1.3). We downgraded the certainty of the evidence due to risk of bias concerns, indirectness, inconsistency and imprecision (see Table 1).

1.3. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 3: DVT (short and intermediate term)

Polidocanol foam may increase phlebitis or thrombophlebitis rates. The lowest number of cases of phlebitis or thrombophlebitis were found in the placebo group compared with the polidocanol foam group during the intermediate‐term periods (e.g. for polidocanol ≥ 1% foam: RR 3.12, 95% CI 1.1 to 8.83; 3 studies, 302 participants) (Analysis 1.4; Analysis 1.5).

1.4. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 4: Phlebitis or thrombophlebitis rates (intermediate term)

1.5. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 5: Phlebitis or thrombophlebitis rates (intermediate term)

Haematoma formation was described as a haemorrhagic complication. There were no clear differences in haemorrhagic outcomes between polidocanol foam and placebo in the short term (polidocanol 1%: RR 1.83, 95 CI 0.75 to 4.47; 3 studies, 302 participants) (Analysis 1.6).

1.6. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 6: Haemorrhagic complications (short term)

There was no evidence of a difference in neurological complication rates between 1% or more concentrated polidocanol foam and placebo at the short term (polidocanol 1%: RR 1.03, 95 CI 0.22 to 4.91; 3 studies, 302 participants) (Analysis 1.7). Gibson 2010 used a different gas mixture (Varithena) in which there is a low proportion of nitrogen aiming to reduce cerebral side effects. A case of transient ischaemic attack was reported on day 13 and could not be related to the intervention. As the gas concentrations were different, we conducted a sensitivity analysis that did not change the effect direction. Aiming to evaluate whether the polidocanol foam concentration could change complication rates, we performed subgroup analysis comparing different concentrations of polidocanol foam (see the comparison 'Sclerotherapy versus sclerotherapy' in the analysis of 'Foam sclerotherapy versus foam sclerotherapy with different concentrations').

1.7. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 7: Neurologic complications (dizziness, TIA) (short term)

Residual varicose veins

Five RCTs reported on this outcome (Gibson 2010; Kahle 2004; King 2015; Todd 2014; Zhang 2009). Residual varicose vein rates may be decreased in the 1% or more polidocanol foam group than in the placebo group (for polidocanol 1% foam: RR 0.19, 95% CI 0.13 to 0.29; 2 studies, 225 participants; very low‐certainty evidence) (Analysis 1.8) at an intermediate‐term period. We downgraded the certainty of the evidence due to risk of bias concerns and indirectness (Todd 2014 included 8.8% CEAP C5 and C6 participants in the placebo group and 3.4% CEAP C5 and C6 participants in the polidocanol 1% group).

1.8. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 8: Residual varicose veins (intermediate term)

Quality of life

Three RCTs reported on this outcome (Gibson 2010; King 2015; Todd 2014). Quality of life (QoL) was measured using QoL adjusted mean change VEINES‐QOL/Sym (Gibson 2010; King 2015; Todd 2014) in polidocanol foam versus placebo (for polidocanol 1%: MD 12.41; 95% CI 9.56 to 15.26; 3 studies, 299 participants; very low‐certainty evidence) (Analysis 1.9), showing a possible improvement in the QoL in the foam group. Although King 2015 did not describe standard deviations, we used standard deviations from Todd 2014 to perform the meta‐analysis following recommendations for standard deviation data imputation from Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017c). The certainty of the evidence was downgraded due to wide confidence intervals, low number of participants, and indirectness (participants with CEAP C5 and C6 included in the studies). Gibson 2010 used a different gas mixture (Varithena) in which there is a low proportion of nitrogen aiming to reduce cerebral side effects. A sensitivity analysis did not show any difference in effect.

1.9. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 9: Quality of life: VEINES‐QOL score change from baseline (intermediate term)

Frequency of persistence of symptoms

Three RCTs reported on this outcome (Gibson 2010; King 2015; Todd 2014). Persistence of symptoms was evaluated by Gibson 2010 using the VEINES‐QOL/Sym score to evaluate the patient perspective using polidocanol 1% foam versus placebo. In this case, polidocanol foam was superior to placebo in improving symptoms (MD ‐14.00; 95% CI ‐15.39 to ‐12.61; 1 study, 77 participants) (Analysis 1.10). The remaining two RCTs also evaluated VEINES‐QOL/Sym but we did not have access to sufficient data to conduct a meta‐analysis (missing standard deviations). These RCTs (King 2015; Todd 2014) also described VCSS; we used this outcome to grade the certainty of the evidence. Both studies reported an improvement in symptoms at intermediate‐term periods after sclerotherapy (eight weeks), and pooling the data for polidocanol 1% versus placebo showed: MD ‐3.25, 95% CI ‐3.90 to ‐2.60; 2 studies, 223 participants; low‐certainty evidence (Analysis 1.11). We downgraded the certainty of the evidence due to the low number of participants and indirectness (participants with CEAP C5 and C6 included in the RCTs).

1.10. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 10: Persistence of symptoms: absolute change from baseline score for the VVSymQ (total score) (intermediate term)

1.11. Analysis.

Comparison 1: Foam sclerotherapy versus placebo, Outcome 11: Persistence of symptoms: mean change from baseline score for the VCSS (total score) (intermediate term)

Recurrent varicose veins

None of the RCTs reporting on this comparison described recurrent varicose veins.

Sclerotherapy versus sclerotherapy

Twenty‐two RCTs investigated this comparison. We were able to pool results from the comparison 'foam sclerotherapy versus foam sclerotherapy with different concentrations'; 'Foam versus liquid sclerotherapy'; and from the comparison 'different techniques'. Three RCTs compared 'different substances' (Goldman 2002; Schadeck 1995a; Wright 2006), two RCTs compared addition of substances including perfluropropane‐filled albumin microspheres of foam STS with air‐filled foam STS (Martimbeau 2003b) or a low molecular weight synthetic protein (Gelufusine) to polidocanol foam versus polidocanol foam sclerotherapy alone (Zeh 2003), and we considered it inappropriate to pool results due to indirect comparisons (different concentrations of the substances). Three RCTs did not supply data to measure the size effect (Martimbeau 2003a; Martimbeau 2003b; Zeh 2003).

No RCT compared different time intervals as part of their injection technique.

Comparison 3: Foam sclerotherapy versus foam sclerotherapy with different concentrations

Five RCTs reported on this comparison (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Todd 2014). Different concentrations of polidocanol foam were used for comparison and they were categorised and described as polidocanol < 1% versus ≥ 1% or the most concentrated polidocanol foam described. We adopted this approach based on data representing an increased efficacy of polidocanol foam to occlude varicose veins at 1% concentration (King 2015).

For polidocanol foam concentrations at 1% to 2%, there was no clear difference in venous occlusion rates (80.4% for polidocanol 1% foam and 82.5% for polidocanol 2% foam) (King 2015). On the other hand, we are not confident that more concentrated foams could show an increase in success rates. So polidocanol 1% was used as a comparator against more concentrated foam preparations whenever there was a possibility to compare in order to investigate if concentrations greater than 1% could improve sclerotherapy effects. When there was an uncertainty of effect size, a subgroup analysis was performed. Care was taken to avoid double counting of participants. Although two authors (King 2015; Todd 2014) used polidocanol 0.125% foam as the smallest concentration permitting a stable foam as a control with other polidocanol concentrations, this sclerosant concentration showed a pharmacologic action, capable of producing objective outcomes (e.g. common femoral vein thrombus extension) as well as subjective outcomes (e.g. QoL improvements); and so is described in this session as an intervention group and not as the inactive placebo group.

We used polidocanol 3% versus 1% to grade the certainty of the evidence and to look for any additional effect with the increase of the concentration of the sclerosant. See Table 2.

Cosmetic appearance

Three RCTs reported on this outcome (Ceulen 2007; King 2015; Todd 2014).

One RCT described self‐related cosmetic improvement over a long‐term period (complete cosmetic satisfaction) when comparing polidocanol 3% foam with polidocanol 1% foam (Ceulen 2007). There were no clear differences in cosmetic appearance in this analysis at a short‐term period (RR 1.31, 95% CI 0.81 to 2.12; 1 study, 80 participants) (Analysis 2.1), or at long‐term period (RR 1.11, 95% CI 0.84 to 1.47; 1 study, 80 participants; very low‐certainty evidence) (Analysis 2.2). The certainty of the evidence was downgraded due to concerns over risk of bias, indirectness (study authors included participants with CEAP C1, C5 and C6), and imprecision (low number of participants, wide confidence intervals, and single study).

2.1. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 1: Cosmetic appearance: number of participants with cosmetic improvement (short term) ‐ polidocanol 3% foam versus polidocanol 1% foam

2.2. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 2: Cosmetic appearance: number of participants with cosmetic improvement (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam

Two RCTs evaluated the PA‐V (patient‐visible varicose vein instrument) and the IPR‐V (independent expert panel rating scale) of photographs according to Kingsley 2012 (King 2015; Todd 2014). Increased concentrations of polidocanol possibly increased the improvement in cosmetic appearance at intermediate‐term periods as seen with PA‐V: polidocanol 0.125% versus 1% (MD 0.45, 95% CI 0.03 to 0.87; 2 studies, 224 participants). Polidocanol 0.5% versus 1% showed no clear difference (MD 0.06, 95% CI ‐0.21 to 0.32; 2 studies, 221 participants; Analysis 2.3). A similar effect was seen with the IPR‐V (independent expert panel rating scale) of photographs: polidocanol 0.125% versus 1% (MD 0.19, 95% CI ‐0.01 to 0.40; 2 studies, 224 participants); polidocanol 0.5% versus 1% (MD ‐0.04, 95% CI ‐0.19 to 0.12; 2 studies, 221 participants; Analysis 2.4). If the concentration of test polidocanol foam was similar to the comparator (1%), there was no clear difference detected in cosmetic improvement, as is expected. We used the standard deviation from Todd 2014 to perform the meta‐analysis. Data imputation followed recommendations from the Cochrane Handbook to deal with missing data (Higgins 2017c).

2.3. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 3: Cosmetic appearance: adjusted mean PA‐V (intermediate term) ‐ polidocanol foam versus polidocanol foam

2.4. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 4: Cosmetic appearance: adjusted mean IPR‐V (intermediate term) ‐ polidocanol foam versus polidocanol foam

Complications

Five RCTs reported on this outcome (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Todd 2014). Complications were described as DVT, skin pigmentation, phlebitis or thrombophlebitis rates, matting, thromboembolic complications, haemorrhagic complications, and neurologic complications.

Thromboembolic complications (including DVT and PE) were described by three RCTs (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), and there were no clear differences between polidocanol 3% versus 1% in foam (RR 1.47, 95% CI 0.41 to 5.33; 3 studies, 371 participants; very low‐certainty evidence Analysis 2.5). Ceulen 2007 used the lower leg as the unit of analysis and included 11.63% of participants with CEAP C1. Blaise 2010 and Hamel‐Desnos 2007 used the individual as the unit of analysis. A sensitivity analysis was performed and it did not change the effect direction (Analysis 2.6). The certainty of the evidence was downgraded due to indirectness (inclusion of participants with CEAP C1, C5, and C6), and imprecision (low number of events).

2.5. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 5: Thromboembolic complications (intermediate term) ‐ polidocanol 3% foam versus polidocanol 1% foam

2.6. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 6: Thromboembolic complications (intermediate term) ‐ polidocanol 3% foam versus polidocanol 1% foam (sensitivity analysis)

From polidocanol 0.125% to 2% foam, there was no evidence of increased DVT for the evaluated subgroups at short or intermediate‐term periods (for polidocanol foam 0.125% versus 2%: RR 1.81, 95% CI 0.17 to 19.43; 1 study, 120 participants) (Analysis 2.7) (for polidocanol 1% versus 2%: RR 0.83, 95% CI 0.12 to 5.66; 1 study, 115 participants) (Analysis 2.7). Analyses were carried out separately to dichotomise polidocanol concentrations lower than 1% versus higher or equal to 1%, and 1% compared to the higher concentrations of polidocanol foam. No differences were detected by the test for subgroup differences (P = 0.73).

2.7. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 7: DVT (short and intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration

Two RCTs (Ceulen 2007; Hamel‐Desnos 2007), investigated the frequency of PE at the short term comparing polidocanol 3% with 1% foam and there was no clear difference seen, with only one case of PE described in the 3% group (RR 3.00, 95% CI 0.13 to 71.51; 2 studies, 228 participants) (Analysis 2.8).

2.8. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 8: PE (short term) ‐ polidocanol 1% foam versus polidocanol 3% foam

Two RCTs (Blaise 2010; Ceulen 2007) described skin pigmentation at long‐term periods comparing polidocanol 3% with 1% foam and meta‐analysis showed no clear difference in skin pigmentation rates (RR 1.61, 95% CI 0.82 to 3.17; 2 studies, 223 participants) (Analysis 2.9). As Ceulen 2007 used the lower leg as the unit of analysis, a sensitivity analysis was conducted and there were no differences in the overall effect.

2.9. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 9: Skin pigmentation rates (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam

A subgroup analysis was carried out to show the results of phlebitis or thrombophlebitis rates in two RCTs with a concentration of polidocanol ranging from 0.125% to 2% (King 2015; Todd 2014); no clear differences in phlebitis and thrombophlebitis rates were seen in any subgroup despite the variation of sclerosant concentrations (polidocanol 0.5% versus 2%: RR 0.77, 95% CI 0.27 to 2.22; 1 study, 114 participants) (Analysis 2.10). No differences were detected by the test for subgroup differences (P = 0.3). Low thrombophlebitis rates were reported with polidocanol 0.125% in the King 2015 study (Analysis 2.10).

2.10. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 10: Phlebitis or thrombophlebitis rates (intermediate term) ‐ polidocanol < 1% foam versus polidocanol ≥ 1% foam

Phlebitis and thrombophlebitis were also described by two other RCTs comparing polidocanol 3% with polidocanol 1% foam (Blaise 2010; Ceulen 2007). These studies were not included in the subgroup analysis above because we dichotomised data to polidocanol foam concentration less than 1% versus polidocanol foam concentration equal to or more than 1%. We assessed the effect size of the higher polidocanol foam concentration separately. There were no clear differences in thrombophlebitis rates in these foam concentrations (RR 1.46, 95% CI 0.88 to 2.41; 2 studies, 223 participants) (Analysis 2.11).

2.11. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 11: Phlebitis or thrombophlebitis ‐ polidocanol 3% foam versus polidocanol 1% foam

Haematoma complications were described in two RCTs (King 2015; Todd 2014), and no clear differences were detected between any investigated subgroups. For polidocanol 0.125% versus 1%: RR 0.61, 95% CI 0.20 to 1.83; 2 studies, 228 participants) (test for subgroup differences P = 0.18) (Analysis 2.12).

2.12. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 12: Haemorrhagic complications (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration

Neurological complications were described as mild in two RCTs (King 2015; Todd 2014), and no clear differences were detected between any of the polidocanol concentrations investigated (0.125% and 0.5% versus 1% and 2%) (Analysis 2.13); or polidocanol 1% versus 2% (RR 0.28, 95% CI 0.01 to 6.64; 1 study, 115 participants) (Analysis 2.13).

2.13. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 13: Neurologic complications (dizziness, TIA) (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration

Residual varicose veins

Five RCTs reported on this outcome (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007; King 2015; Todd 2014). Residual varicose veins were described by two studies at an intermediate‐term period (King 2015; Todd 2014), and three RCTs reported at a long‐term period (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007). Subgroup analysis was used to compare concentrations of the foam sclerotherapy agent in reducing residual varicose veins (Analysis 2.14). Increasing the polidocanol foam concentration reduced the number of participants with residual varicose veins (for polidocanol 0.5% versus 2%: RR 2.36, 95% CI 1.26 to 4.42; 1 study, 114 participants) (Analysis 2.14). King 2015 compared polidocanol 1% versus 2% and there was no clear difference in residual varicose vein rates (RR 1.21, 95% CI 0.57 to 2.57; 1 study, 115 participants) (Analysis 2.14). Three studies evaluated higher concentrations of polidocanol foam at a long‐term period (3% versus 1% at 6, 12, and 24 months) (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007). The rate of residual varicose veins may be slightly decreased in the polidocanol 3% foam group from six to 24 months (RR 0.67, 95% CI 0.43 to 1.04; 3 studies, 371 participants; moderate‐certainty evidence) (Analysis 2.15). The effect size was not changed following sensitivity analysis to exclude a study that included participants with CEAP C1(Ceulen 2007; Analysis 2.16).

2.14. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 14: Residual varicose veins (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration

2.15. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 15: Residual varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam

2.16. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 16: Residual varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam (sensitivity analysis)

The evidence was downgraded due to indirectness (inclusion of participants with CEAP C1 (11.63% in Ceulen 2007), C5 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), and C6 (Hamel‐Desnos 2007) but there was no description of the proportion of these participants in the groups). One study used the lower leg as the unit of analysis (Ceulen 2007).

Quality of life

Two RCTs reported on this outcome (King 2015; Todd 2014).

Quality of life analysis was possible with data imputation using the standard deviations from Todd 2014 when using the VEINES‐QOL/Sym scores. The study authors reported a tendency for improvement of quality of life if more concentrated polidocanol foam was used. We were only able to conduct meta‐analysis comparing 0.125%, 0.5%, and 1% polidocanol foam (Analysis 2.17). Our analysis did not find a clear improvement in the quality of life when increasing polidocanol foam concentration up to 1% polidocanol foam (test for subgroup differences P = 0.11).

2.17. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 17: Quality of life: adjusted mean change VEINES‐QOL (intermediate term) ‐ polidocanol foam versus polidocanol foam

No study evaluated polidocanol 3% versus 1%, so the certainty of the evidence was not graded.

Frequency of persistence of symptoms

Two RCTs reported the frequency of persistence of symptoms using the VCSS score (King 2015; Todd 2014). Both reported an improvement in VCSS scores with increasing concentrations of polidocanol foam and this was confirmed by meta‐analysis: polidocanol 0.125% versus 0.5% (MD ‐0.87, 95% CI ‐1.52 to ‐0.21; 2 studies, 225 participants) (Analysis 2.18). Only polidocanol 0.5% versus 1% did not show a difference in VCSS scores (MD 0.10, 95% CI ‐0.56 to 0.75; 2 studies, 221 participants) (test for subgroup differences P = 0.08) (Analysis 2.18).

2.18. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 18: Persistence of symptoms: adjusted mean change from baseline VCSS score (intermediate term) ‐ polidocanol foam versus polidocanol foam

The certainty of the evidence was not graded since no study compared polidocanol 3% versus 1%.

Recurrent varicose veins and venous flare formation

One RCT reported on this outcome (Hamel‐Desnos 2007). There was no clear difference in the number of recurrent varicose veins when comparing polidocanol 3% versus 1% foam concentrations at a long‐term period (RR 0.91, 95% CI 0.62 to 1.32; 1 study, 148 participants; low‐certainty evidence) (Analysis 2.19).

2.19. Analysis.

Comparison 2: Foam sclerotherapy versus foam sclerotherapy with different concentrations, Outcome 19: Recurrent varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam

The certainty of the evidence was graded as low since data were from a single study with an indirect population (participants with CEAP C5 and C6).

Comparison 4: Foam sclerotherapy versus liquid sclerotherapy

Nine RCTs reported on this comparison (Alos 2006; Belcaro 2003b; Demagny 2002; Martimbeau 2003a; Martimbeau 2003b; Rabe 2008; Ouvry 2008; Ukritmanoroat 2011; Zeh 2003). Three studies did not provide any usable data (Martimbeau 2003a; Martimbeau 2003b; Zeh 2003). One RCT used an ultrasound contrast agent to produce a foam (Belcaro 2003b). As this is not a standard way to produce a foam sclerotherapy agent, we believe it cannot be compared to any conventional foam in which the medicine is agitated with a gas. To grade the certainty of the evidence, we have tried to use similar or equal foam concentrations since it is not known if there is an ideal concentration. Using the same concentration will be more useful for clinicians. See Table 3.

Cosmetic appearance

One RCT reported on this outcome. Long‐term cosmetic appearance was reported by Alos 2006 as patient satisfaction and there was no difference when using polidocanol 0.75% to 1.25% foam versus polidocanol 1.5% to 2.5% liquid (MD 0.20, 95% CI ‐0.27 to 0.67; 1 study, 126 participants; very low‐certainty evidence) (Analysis 3.1). The unit of analysis was the vein. No other RCT included in this comparison reported on this outcome and there were differences in the concentrations compared. We downgraded the certainty of the evidence to very low due to risk of bias concerns, and indirectness (different concentrations of the polidocanol used). As the authors did not use similar or equal concentrations of sclerosant, we undertook a descriptive analysis and used this study to grade the certainty of the evidence in Table 3.

3.1. Analysis.

Comparison 3: Foam sclerotherapy versus liquid sclerotherapy, Outcome 1: Cosmetic appearance: participant satisfaction (long term) ‐ polidocanol foam versus polidocanol liquid

Complications

Seven RCTs reported on this outcome (Alos 2006; Demagny 2002; Martimbeau 2003a; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ukritmanoroat 2011). Two studies did not report complications as an outcome (Belcaro 2003b; Zeh 2003). There was a lack of reported information and interest in adverse events and complications, including thromboembolic events. Overall, complications were described as mild, and these included pain, phlebitis, haematomas, and skin pigmentation. One study described a case of allergy in the liquid intervention group and five cases of amaurosis in the foam intervention group using STS sclerosant agent (Demagny 2002). Ouvry 2008 described complications as pooled results including an event of vaso‐vagal response, one of sensation of heat in the mouth immediately after injection, persistent inflammation in the thigh, two cases of bruising, two of thrombophlebitis of the leg, one of asthenia, one patient with headache, and two cases of pain in the thigh and knee; with the study authors reporting that there was no difference between the two treatment groups. Alos 2006 described four participants in the foam group as having severe pain as well as 25.3% of the participants from the foam group and 9.5% from the liquid group with local inflammation at 15 days post‐treatment. The RCT also described a greater number of pigmentations in the foam group (33%) compared to the liquid group (6.3%) at one year. However, this RCT included participants with CEAP C1 and we were unable to access individual patient data. Also, the sclerosant concentration was stratified for subgroups of vein diameter. For veins with 1 mm to 2 mm diameter, 1% liquid polidocanol or 0.5% for the foam group was used; veins with 1 mm to 3 mm used 1.25% liquid polidocanol versus 0.5% for foam; varicose veins with 3.1 mm to 4 mm used 1.5% liquid polidocanol and 0.75% for foam; 4.1 mm to 5 mm in diameter varicose veins used 2% liquid polidocanol versus 1% for foam; 5.1 mm to 6 mm in diameter veins used 2.5% liquid polidocanol versus 1.25% for foam sclerotherapy. As we had no access to subgroups of the participants, we decided not to calculate the effect size because it could result in bias. Rabe 2008 and Ukritmanoroat 2011 described skin pigmentation as mild or moderate, with no clear difference detected between the foam and liquid sclerotherapy intervention groups (RR 1.75, 95% CI 0.81 to 3.80; 2 studies, 208 participants) (Analysis 3.2). One RCT described 0.2% of side effects for the foam group versus 2.4% for the liquid group but the study authors did not describe which side effects occurred (Martimbeau 2003a). One RCT described an increase in phlebitis rates in the air‐filled foam compared to microspheres filled (P = 0.01) (Martimbeau 2003b).

3.2. Analysis.

Comparison 3: Foam sclerotherapy versus liquid sclerotherapy, Outcome 2: Skin pigmentation (long term) ‐ polidocanol foam versus polidocanol liquid

Residual varicose veins

Six RCTs reported on this outcome (Alos 2006; Belcaro 2003b; Demagny 2002; Ouvry 2008; Rabe 2008; Ukritmanoroat 2011). One study investigated the use of a tensioactive substance to produce a foam with STS 2% or 3% (Belcaro 2003b). The authors presented results as the reduction of ambulatory venous pressure, refilling time, and number of major sites of incompetence on duplex ultrasound, and using the duplex ultrasound demonstrated no differences in residual varicose veins after 10 years of follow‐up (1.1 points of major site incompetence in each treatment option). As the RCT used different medicine concentrations, different outcome measures, and time points to report the outcomes, we decided not to include it in the analysis. There was a benefit for the use of foam seen in all RCTs for occluding the varicose veins over a long time (Analysis 3.3). We did not think it was appropriate to pool overall, as different concentrations and sclerosants were used, so data have been presented as a subgroup analysis (Analysis 3.3). As Alos 2006 used different concentrations of sclerosant, a sensitivity analysis was conducted and it did not change the effect direction. We decided to report and assess results from polidocanol 3%, as this was most clinically relevant. For polidocanol 3%, there was an RR of 0.51 (95% CI 0.41 to 0.65; 2 studies, 203 participants; very low‐certainty evidence) (Analysis 3.3). Two sclerotherapy products were included in this analysis: polidocanol and STS. There was also heterogeneity in the population due to differences in varicose vein diameter, type of vein treated (saphenous or collaterals), number of sessions necessary to treat, and the CEAP inclusion. Some of these data were difficult or impossible to recover and there were differences in results between different centres of the same RCTs (Rabe 2008). Despite these limitations, sensitivity analysis showed a consistent benefit for the foam groups over the liquid groups. The certainty of the evidence was downgraded due to risk of bias concerns and indirectness: Ouvry 2008 included participants with CEAP C2 to C6 without discriminating between them; Rabe 2008 included one participant with CEAP C5 in the foam group (1.9%) and four participants with CEAP C5 in the liquid group (7.7%).

3.3. Analysis.

Comparison 3: Foam sclerotherapy versus liquid sclerotherapy, Outcome 3: Residual varicose veins (long term) ‐ polidocanol foam versus polidocanol liquid

Residual varicose veins were also investigated following STS 3%, and at a long‐term period showed a benefit of foam over liquid sclerotherapy (RR: 0.63, 95% CI 0.48 to 0.82; 1 study, 130 participants) (Analysis 3.4).

3.4. Analysis.

Comparison 3: Foam sclerotherapy versus liquid sclerotherapy, Outcome 4: Residual varicose veins (long term) ‐ polidocanol foam versus polidocanol liquid (sensitivity analysis)

Quality of life

None of the RCTs reporting on this comparison reported QoL.

Frequency of persistence of symptoms

None of the RCTs reporting on this comparison reported persistence of symptoms.

Recurrent varicose veins

Two RCTs reported on recurrent varicose veins (Belcaro 2003b; Martimbeau 2003a). Martimbeau 2003a did not provide any data, but described no difference in recurrent varicose veins at 12 months (1.2% in foam group and 1.7% in liquid group). Belcaro 2003b evaluated this outcome at five years comparing STS foam with liquid and reported recurrence as similar in both groups (RR 1.1, 95% CI 0.86 to 1.42; 1 study, 286 participants; very low‐certainty evidence) (Analysis 3.5).

3.5. Analysis.

Comparison 3: Foam sclerotherapy versus liquid sclerotherapy, Outcome 5: Recurrent varicose veins (long term) ‐ STS 3% liquid versus STS 3% foam

The certainty of the evidence was downgraded due to the risk of bias (intervention probably impossible to blind due to the nature of the intervention), imprecision (single study), and indirectness (did not describe CEAP of the participants and used a tensioactive substance to produce foam ‐ not the standard way to use gas or room air). The study author described the participants as people with varicose veins with diameters larger than 3 mm.

Comparison 5: Sclerotherapy versus sclerotherapy with different substances

Four RCTs compared sclerotherapy versus sclerotherapy with any other substance (Goldman 2002; Rao 2005; Schadeck 1995a; Wright 2006). These studies investigated a range of concentrations, formulations, and drug volumes, so we were unable to pool all the data. Goldman 2002 studied STS 1.5% versus polidocanol 3%. The same substances were investigated by Rao 2005 but using STS 0.5% versus polidocanol 1%, and Schadeck 1995a with STS 3% versus polidocanol 4%. Wright 2006 compared polidocanol 1% with any market sclerosant in any presentation or concentration. Although we had planned to present a SoF table for the comparison sclerotherapy with STS 1% versus sclerotherapy with polidocanol 1% (see Methods section), no included RCT considered this comparison. See Table 4.

Cosmetic appearance

Two RCTs reported on this outcome (Goldman 2002; Rao 2005). Cosmetic appearance was described by Goldman 2002 as the disappearance of the varicose veins after treating the veins (diameter between 3 mm and 6 mm) using polidocanol 3% versus STS 1.5%. The scale to describe disappearance of the varicose veins ranged from 1 to 5: 1 = worse than before treatment; 2 = no change; 3 = minor disappearance; 4 = moderate disappearance; 5 = complete disappearance (MD 0.20, 95% CI ‐0.01 to 0.41; 1 study, 54 participants) (Analysis 4.1). As different polidocanol and STS concentrations were used, we were not able to draw conclusions about the benefit of one product over the other. Rao 2005 described outcomes as clearance scores, tolerability, and patient satisfaction. It was not possible to estimate the effect size from Rao 2005 since there were pooled data from CEAP C1 and C2 and different interventions (STS 0.25% or polidocanol 0.5% solution for vein diameters lower than 1 mm; STS 0.5% or polidocanol 1% solution for 1 mm to 3 mm veins, and STS 0.5% or polidocanol 1% foam for veins with diameter between 3 mm and 6 mm). We believe a sclerosant could reach veins with small or larger diameters once injected in a varicose vein, since reflux could drive sclerosant agent to that site, adding a confounding factor to analysis. We did not pool results to grade the certainty of the evidence. We decided not to grade the certainty of the evidence due to discrepancy in concentration of sclerosants and no discrimination of CEAP population.

4.1. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 1: Cosmetic appearance: disappearance of varicose veins (long term) ‐ polidocanol 3% versus STS 1.5%

Complications

Two RCTs reported on complications but we were unable to pool the data due to differences in sclerosant concentrations used (Goldman 2002; Wright 2006). One RCT compared a commercial polidocanol 1% foam (Varisolve) versus any market sclerosant liquid or foam and reported on skin pigmentation, DVT and haematoma at short‐ or intermediate‐term periods (Wright 2006). No differences were seen in the occurrence of DVT (RR 0.97, 95% CI 0.09 to 10.54; 1 study, 384 participants) (Analysis 4.2); skin pigmentation (RR 1.01, 95% CI 0.80 to 1.27; 1 study, 384 participants; Analysis 4.3); or haematoma formation (RR 1.56, 95% CI 0.76 to 3.18; 1 study, 384 participants) (Analysis 4.4). Goldman 2002 described complications without discriminating between participants with CEAP C1 and C2. Otherwise, intervention and control groups appeared to be balanced. The number of skin pigmentation events were not different between STS 1.5% versus polidocanol 3% (RR 1.21, 95% CI 0.95 to 1.55; 1 study, 167 participants) (Analysis 4.5). Skin necrosis numbers were not clearly different in the STS 1.5% group and in the polidocanol 3% group but no events were reported in the polidocanol 3% group resulting in a very wide CI (RR 14.12, 95% CI 0.81 to 246.76; 1 study, 167 participants) (Analysis 4.6). Goldman 2002 also described thrombosis of the vein, probably related to thrombophlebitis or venous occlusion (superficial venous system). As the RCT compared different sclerosant concentrations, we considered it inappropriate to present the data in a SOF table. We decided not to grade the certainty of the evidence due to discrepancy in concentrations of sclerosants and no discrimination of CEAP population.

4.2. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 2: DVT (short and intermediate term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)

4.3. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 3: Skin pigmentation (intermediate term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)

4.4. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 4: Haematoma (short term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)

4.5. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 5: Skin pigmentation (long term) ‐ STS 1.5% versus polidocanol 3%

4.6. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 6: Skin necrosis (long term) ‐ STS 1.5% versus polidocanol 3%

Frequency of residual varicose veins

Wright 2006 investigated polidocanol 1% (Varisolve) versus any market sclerosant liquid or foam. We considered the veins that had no occlusion or elimination of reflux as residual varicose veins. At 12 months, the rates of residual varicose veins were greater in any market sclerosant than in the intervention group (RR 0.64, 95% CI 0.42 to 0.97; 1 study, 373 participants) (Analysis 4.7). As the comparators were so discrepant (any concentrations and forms of liquid or foam), we decided not to grade the certainty of the evidence.

4.7. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 7: Residual varicose veins (long term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)

Quality of life

None of the RCTs reporting on this comparison reported QoL.

Frequency of persistence of symptoms

None of the RCTs reporting on this comparison reported on persistence of symptoms.

Frequency of recurrent varicose veins and venous flare formation

Goldman 2002 investigated matting and there was no clinical difference in the STS 1.5% intervention group compared to polidocanol 3% (RR 1.53, 95% CI 0.58 to 4.02; 1 study, 167 participants) (Analysis 4.8). We did not grade the certainty of the evidence due to the different sclerosant concentrations. We believe that this concentration comparison does not represent a relevant clinical situation.

4.8. Analysis.

Comparison 4: Sclerotherapy versus sclerotherapy with different substances, Outcome 8: Matting (long term) ‐ STS 1.5% versus polidocanol 3%

Comparison 6: Sclerotherapy versus sclerotherapy with different techniques

Six RCTs evaluated different techniques for injection sclerotherapy (Chleir 1997; Ragg 2015; Santos 2019; Yamaki 2009; Yamaki 2012; Zhang 2014). There was important heterogeneity between comparisons due to the use of different interventions and time points assessed. These techniques included single versus multiple injections (with the same product and the same volume) (Chleir 1997); polidocanol 1% to 3% foam injected in few injections versus in multiple injections (Yamaki 2009); and ultrasound‐guided foam sclerotherapy with visual foam sclerotherapy using polidocanol 1% foam (Yamaki 2012); polidocanol 1% foam sclerotherapy injected by a needle compared to polidocanol 1% foam catheter‐directed sclerotherapy, and with polidocanol 1% foam microcatheter‐directed sclerotherapy (Ragg 2015); polidocanol 3% foam injected by needle versus delivered by a catheter associated with tumescent technique (Santos 2019); and antegrade versus retrograde technique with polidocanol 1% foam (Zhang 2014).

Cosmetic appearance

None of the RCTs reporting on this comparison reported cosmetic appearance.

Complications

Zhang 2014 compared antegrade versus retrograde techniques using polidocanol 1% foam and no DVT events were reported in either group. Similarly, no clear differences were detected in short‐term period thrombophlebitis rates (RR 2.00, 95% CI 0.54 to 7.45; 1 study, 80 participants) (Analysis 5.1); or long‐term period skin pigmentation rates (RR 1.5, 95% CI 0.26 to 8.50; 1 study, 80 participants) (Analysis 5.2); or in thrombophlebitis rates using catheter versus needle injection (RR 1.00, 95% CI 0.07 to 15.12; 1 study, 50 participants) (Analysis 5.3).

5.1. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 1: Phlebitis or thrombophlebitis rates (short term) ‐ polidocanol 1% foam ‐ antegrade versus retrograde technique

5.2. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 2: Skin pigmentation rates (long term) ‐ polidocanol 1% foam ‐ antegrade versus retrograde technique

5.3. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 3: Thrombophlebitis rates (short term) ‐ polidocanol 3% foam injected by catheter associated to tumescent technique versus polidocanol 3% foam injected by needle

There was more foam agent detected in the deep venous system by ultrasound when using fewer injections compared to more injections in Yamaki 2012 (RR 2.12, 95% CI 1.36 to 3.30; 1 study, 112 participants) (Analysis 5.4). Chleir 1997 reported seven participants with pain, nine with phlebitis (described as "venulites"), and two with oedema. Ragg 2015 reported no adverse reaction, especially bleeding. Yamaki 2009 described two cases of migraine in the few injections group, two cases of superficial thrombophlebitis in each group, and one case of pain in each group. Yamaki 2012 reported one case of thrombophlebitis in each group, and one case of migraine in the visual foam sclerotherapy group. One RCT reported skin irritation in two participants, one case of migraine in the catheter foam‐delivered group, and hyperpigmentation in eight participants from the needle injection group, compared to four participants from the catheter group (Santos 2019). As there was a co‐intervention (tumescent technique), we did not include these data in a meta‐analysis.

5.4. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 4: Detection of polidocanol foam in deep veins ‐ few injections versus multiple injections

Frequency of residual varicose veins

Six studies reported on this outcome but we were not able to pool data due to heterogeneity (Chleir 1997; Ragg 2015; Santos 2019; Yamaki 2009; Yamaki 2012; Zhang 2014). Yamaki 2012 compared visual foam sclerotherapy with ultrasound‐guided foam sclerotherapy using polidocanol 1% foam and there were no clear differences in residual varicose vein rates in the long‐term period (RR 1.05, 95% CI 0.75 to 1.48; 1 study, 103 participants) (Analysis 5.5). Two studies reported on this outcome questioning if an increase in puncture number could impact residual varicose vein rates (Chleir 1997; Yamaki 2009). Chleir 1997 evaluated STS 3% injected in fractionated versus single injections with no differences detected in residual varicose vein rates (RR 1.08, 95% CI 0.79 to 1.49; 1 study, 150 participants) (Analysis 5.6). Yamaki 2009 compared polidocanol 1% or 3% foam in multiple injections versus few injections showing no differences in the residual varicose vein rates (RR 1.17, CI 0.78 to 1.78; 1 study, 112 participants) (Analysis 5.6). The injection technique (retrograde versus antegrade) made no difference to the number of residual varicose veins at the intermediate‐term period (Zhang 2014) (RR 1.80, 95% CI 0.66 to 4.90; 1 study, 80 participants) (Analysis 5.7).

5.5. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 5: Residual varicose veins (long term) ‐ visual foam sclerotherapy versus ultrasound guided foam + visual foam sclerotherapy

5.6. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 6: Residual varicose veins (long term) ‐ multiple/fractionated injections versus few/single injections

5.7. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 7: Residual varicose veins (intermediate term) ‐ polidocanol 1% foam retrograde versus antegrade technique

Ragg 2015 compared polidocanol 1% foam sclerotherapy injected by a needle with polidocanol 1% foam catheter‐directed sclerotherapy, and with polidocanol 1% foam microcatheter‐directed sclerotherapy. There were low numbers of participants in each group (10), and there was no difference in residual varicose veins at the intermediate‐term period between all comparable groups: microcatheter versus catheter (RR 5.00, 95% CI 0.27 to 92.62; 1 study, 20 participants; Analysis 5.8); or needle versus catheter with zero events in the catheter group leading to very wide CIs (RR 9.00, 95% CI 0.55 to 147.95; 1 study, 20 participants; Analysis 5.9); and needle sclerotherapy versus microcatheter (RR 2.00, 95% CI 0.47 to 8.56; 1 study, 20 participants; Analysis 5.10). Santos 2019 compared polidocanol 3% foam injected by a catheter associated with tumescent technique versus needle and reported lower rates of residual varicose veins in the catheter‐delivered group at a short‐term period (RR 0.31, 95% CI 0.14 to 0.72; 1 study, 50 participants; Analysis 5.11). At a long‐term period, both groups had a similar number of participants requiring retreatments (RR 0.92, 95% CI 0.50 to 1.67; 1 study, 50 participants; Analysis 5.12).

5.8. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 8: Residual varicose veins (intermediate term) ‐ microcatheter directed foam sclerotherapy versus catheter directed foam sclerotherapy

5.9. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 9: Residual varicose veins (intermediate term) ‐ needle foam sclerotherapy versus catheter directed foam sclerotherapy

5.10. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 10: Residual varicose veins (intermediate term) ‐ needle foam sclerotherapy versus microcatheter directed foam sclerotherapy

5.11. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 11: Residual varicose veins (short term) ‐ polidocanol 3% foam injected by catheter versus polidocanol 3% foam injected by needle

5.12. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 12: Participants not requiring retreatment (long term) ‐ polidocanol 3% foam injected by catheter versus polidocanol 3% foam injected by needle

Frequency of recurrent varicose veins and venous flare formation

None of the RCTs reporting on this comparison reported on recurrent varicose veins and venous flare formation.

Quality of life

One RCT reported QoL in this comparison (Santos 2019). The authors reported an improvement in QoL without inter‐group differences but with intra‐group differences before and after six months of follow‐up. They used AVVQ (Aberdeen Varicose Vein Questionnaire) with the most favourable result being the lowest score. Our analysis did not show a clear improvement for either catheter‐directed or ultrasound‐guided sclerotherapy (MD ‐2.71, 95% CI ‐9.33 to 3.91; 1 study, 50 participants; Analysis 5.13).

5.13. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 13: QoL: AVVQ (long term) ‐ catheter‐directed foam sclerotherapy with tumescence versus ultrasound‐guided foam sclerotherapy

Frequency of persistence of symptoms

One RCT reported on this outcome. Yamaki 2012 did not find clinical differences in VCSS in the long‐term period (MD 0.00, 95% CI ‐0.77 to 0.77; 1 study, 103 participants; Analysis 5.14).

5.14. Analysis.

Comparison 5: Sclerotherapy versus sclerotherapy with different techniques, Outcome 14: VCSS (long term) ‐ ultrasound guided foam + visual foam sclerotherapy versus visual foam sclerotherapy

Sclerotherapy versus different conservative treatment

Comparison 7: Sclerotherapy versus pharmacological therapy

No study compared sclerotherapy with any pharmacological therapy.

Comparison 8: Sclerotherapy versus conservative compression treatment

One study compared sclerotherapy versus conservative management at six months to two years of follow‐up (Abramowitz 1973). This study was conducted in pregnant women and reported a benefit with injection sclerotherapy compared to conservative management (compression) on the outcome of symptomatic improvement and cosmetic appearance (RR 7.96, 95% CI 3.38 to 18.76; 1 study, 101 participants) (Analysis 6.1). However, an increased number of ulcers were reported after sclerotherapy (RR 13.63, 95% CI 0.77 to 240.13; 1 study, 101 participants) (Analysis 6.2). It is also important to note that the outcomes, cosmetic appearance and symptomatic improvement, were described as pooled data and we were unable to discriminate data for each outcome. There were no other outcomes evaluated by the RCT.

6.1. Analysis.

Comparison 6: Sclerotherapy versus compression conservative treatment, Outcome 1: Cosmetic appearance: good symptomatic improvement and cosmetic result (long term) liquid STS versus compression

6.2. Analysis.

Comparison 6: Sclerotherapy versus compression conservative treatment, Outcome 2: Ulcer (short, intermediate and long term) ‐ liquid STS versus compression

Sensitivity analysis

We performed a sensitivity analysis excluding studies which had data imputation. As the lost of follow‐up was balanced in all groups, this action did not change the effect direction. We preferred to present results using intention‐to‐treat analysis.

Fourteen RCTs were judged as being at high risk of bias for at least one of the following domains: selection, performance, or detection (Alos 2006; Belcaro 2003b; Gibson 2010; King 2015; Martimbeau 2003b; Ouvry 2008; Rabe 2008; Ragg 2015; Todd 2014; Ukritmanoroat 2011; Wright 2006; Yamaki 2009; Yamaki 2012; Zhang 2009). Two of these RCTs (King 2015; Todd 2014) were considered as having high risk of bias when placebo was the comparator, as it was impossible to blind placebo due to the nature of the intervention. Belcaro 2003b declared that investigators opened the randomisation code after randomisation. Martimbeau 2003b did not describe the outcome assessment. All other studies were deemed as being at high risk of detection bias as it was not possible to blind participants and personnel. As planned, we conducted a sensitivity analysis

any of these studies were included in a meta‐analysis. High risk of bias studies were combined in most meta‐analyses (e.g. foam versus placebo). One meta‐analysis evaluated residual varicose veins at a long‐term period comparing foam versus liquid sclerotherapy (Analysis 3.3). Only one RCT was not judged as being at high risk of bias (Demagny 2002). A sensitivity analysis was performed and the effect direction was not changed (RR 0.63, 95% CI 0.48 to 0.82; 1 study, 300 participants; Analysis 3.4). To perform this analysis we evaluated only great saphenous veins (300 included, and 50% in each group) and considered recanalisation as residual varicose veins, since a duplex ultrasound image was presented in the paper showing the recanalised vein as an unresolved varicose vein, with sinals of thrombus in the lumen. So, with sensitivity analysis, foam appeared better than liquid to avoid residual varicose veins. Another analysis compared sclerotherapy versus sclerotherapy with different techniques and two studies in a subgroup analysis had different judgements for the risk of bias (Analysis 5.6). There was no change in the effect direction following sensitivity analysis, indicating residual varicose veins were not different with single versus fractionated doses of injection sclerotherapy.

Six RCTs were funded by pharmaceutical companies (Blaise 2010; Gibson 2010; King 2015; Rabe 2008; Todd 2014; Wright 2006). These RCTs were involved in six meta‐analyses (Analysis 2.5; Analysis 2.6; Analysis 2.9; Analysis 2.11; Analysis 2.15; Analysis 3.3). We performed a sensitivity analysis for all of these analyses and we did not find changes in the direction of the effect for any of the meta‐analyses. So, RCTs funded by industry did not appear to change the effect of the intervention.

We expected to find the participant as the unit of analysis. However, in five RCTs, the unit of analysis was the vein (Alos 2006; Demagny 2002; Rao 2005; Ukritmanoroat 2011; Zhang 2014) and, in four RCTs, the lower leg (Ceulen 2007; Chleir 1997; Yamaki 2009; Yamaki 2012). In these cases, whenever a meta‐analysis was performed, we did a sensitivity analysis. We did not find any change in effect direction in any of the analyses evaluated (Analysis 2.5; Analysis 2.8; Analysis 2.9; Analysis 2.11; Analysis 2.12; Analysis 3.2; Analysis 3.3; Analysis 5.6).

Subgroup analysis

Planned subgroup analysis was not possible for CEAP clinical classes since we did not have data to discriminate participants from each clinical class. No RCT cited the depth of the vein or discriminated between symptomatic versus asymptomatic participants.

RCTs comparing foam versus foam sclerotherapy evaluated only truncal incompetence, precluding subgroup analysis.

Previous complications (e.g. thrombophlebitis) were exclusion criteria of the included studies, making subgroup analysis impossible.

We used many comparators in the foam versus foam sclerotherapy dealing with subgroups. So, the 'Foam versus foam sclerotherapy' comparison was presented as subgroup analysis. King 2015 demonstrated that the best foam concentration to reduce residual varicose veins was at least polidocanol 1%. But polidocanol 3% foam was more effective at reducing residual varicose veins than 1% foam (RR 0.67, 95% CI 0.43 to 1.04; 3 studies, 371 participants; moderate‐certainty evidence) (Analysis 2.15). Polidocanol 1% foam versus 3% did not change cosmetic appearance, complications, QoL or symptoms (low‐ to moderate‐certainty evidence).

Four RCTs investigated different gases (called physiologic gases including O₂ and CO₂ in Varithena and Varisolve) in the foam preparation (Gibson 2010; King 2015; Todd 2014; Wright 2006). The study authors compared Varisolve with any market sclerosant liquid or foam (Wright 2006), precluding subgroup analysis. Gibson 2010 compared foam sclerotherapy with placebo. The other proposed subgroup analysis involved the other three studies in the same meta‐analysis.

We also performed subgroup analyses of the different sclerosant concentrations to simplify the data interpretation (Analysis 1.1; Analysis 1.2; Analysis 1.4; Analysis 1.5; Analysis 1.3; Analysis 1.6; Analysis 1.7; Analysis 1.8; Analysis 1.9; Analysis 1.11; Analysis 2.3; Analysis 2.4; Analysis 2.7; Analysis 2.9; Analysis 2.10; Analysis 2.12; Analysis 2.13; Analysis 2.14; Analysis 2.17; Analysis 2.18; Analysis 3.2; Analysis 3.3; Analysis 5.6).

Discussion

Summary of main results

We included 28 RCTs involving 4278 participants. The studies varied in the interventions and comparisons used.

Sclerotherapy versus no intervention or placebo

Comparison 1: Sclerotherapy versus no intervention

The review did not find any RCT comparing sclerotherapy with no intervention. This lack of information highlights the continuance of the disappointing gap of knowledge for this comparison, mainly for CEAP C2 varicose veins.

Comparison 2: Sclerotherapy versus placebo

Table 1

Foam sclerotherapy may improve cosmetic appearance compared to placebo (polidocanol 1%: independent photography review ‐ visible varicose veins scores (IPR‐V) mean difference (MD) ‐0.76, 95% CI ‐0.91 to ‐0.60; 2 studies, 223 participants; very low‐certainty evidence (Analysis 1.1)).

Mild to moderate rates of complications were higher in the sclerotherapy group, including deep vein thrombosis (DVT) rates (polidocanol 1% foam: RR 5.10, CI 1.30 to 20.01; 3 studies, 302 participants; very low‐certainty evidence) (Analysis 1.3); and thrombophlebitis (polidocanol 1% foam: RR 3.12, 95% CI 1.10 to 8.83; 3 studies, 302 participants) (Analysis 1.5). DVT occurrence is rare in sclerotherapy and can also occur in other modalities of treatment (Jacobs 2014; Paravastu 2016). In this case, the occurrence of DVT is derived from usually small asymptomatic thrombus, and did not cause an increased rate in pulmonary embolism (PE) in most participants (no symptomatic PE events in any of the three studies were included in meta‐analysis) (Gibson 2010; King 2015; Todd 2014). Most thrombi originated in the saphenous veins and extended to the femoral vein and there were no participants with signs or symptoms of DVT (King 2015).

Sclerotherapy did not increase the rates of haemorrhagic (polidocanol 1%: RR 1.83, 95 CI 0.75 to 4.47; 3 studies, 302 participants) (Analysis 1.6) or neurologic complications (polidocanol 1%: RR 1.03, 95 CI 0.22 to 4.91; 3 studies, 302 participants) (Analysis 1.7).

Sclerotherapy may reduce the rate of residual varicose veins (polidocanol 1% foam: RR 0.19, 95% CI 0.13 to 0.29; 2 studies, 225 participants; very low‐certainty evidence) (Analysis 1.8); may improve quality of life (polidocanol 1%: MD 12.41, 95% CI 9.56 to 15.26; 2 studies, 222 participants; low‐certainty evidence) (Analysis 1.9); and may reduce varicose vein symptoms (MD ‐3.25, 95% CI ‐3.90 to ‐2.60; 2 studies, 223 participants; low‐certainty evidence) (Analysis 1.11).

Sclerotherapy versus sclerotherapy

Twenty‐two RCTs investigated sclerotherapy versus sclerotherapy, with four comparisons involved:

• Foam versus foam sclerotherapy with different concentrations

• Foam versus liquid sclerotherapy

• Sclerotherapy versus sclerotherapy with different substances

• Sclerotherapy versus sclerotherapy with different techniques

Comparison 3: Foam sclerotherapy versus foam sclerotherapy with different concentrations

Five RCTs reported on this comparison. See Table 2.

Increasing concentrations of polidocanol foam to 3% showed no clear differences in cosmetic appearance compared to polidocanol 1% foam in a single study at one month (RR 1.31, 95% CI 0.81 to 2.12; 1 study, 80 participants; Analysis 2.1); or one year (RR 1.11, 95% CI 0.84 to 1.47; 1 study, 80 participants; very low‐certainty evidence; Analysis 2.2). Lower concentrations of polidocanol foam sclerotherapy (0.125%) had a similar effect to placebo. Higher concentrations of polidocanol possibly increased the improvement in cosmetic appearance as assessed by the adjusted mean for patient visible varicose vein instrument (PA‐V) (polidocanol 0.125% versus 1% (MD 0.45, 95% CI 0.03 to 0.87; 2 studies, 224 participants).Polidocanol 0.5% versus 1% showed no clear difference (MD 0.06, 95% CI ‐0.21 to 0.32; 2 studies, 221 participants; Analysis 2.3). A similar effect was seen with IPR‐V (independent expert panel rating scale) of photographs:polidocanol 0.125% versus 1% (MD 0.19, 95% CI ‐0.01 to 0.40; 2 studies, 221 participants); polidocanol 0.5% versus 1% (MD ‐0.04, 95% CI ‐0.20 to 0.12; 2 studies, 224 participants; Analysis 2.4). However, improvement in cosmetic appearance was still uncertain for concentrations higher than 1%.

No clear differences in the rates of thromboembolic complications were detected between polidocanol 3% and 1% (RR 1.47, 95% CI 0.41 to 5.33; 3 studies, 371 participants; very low‐certainty evidence; Analysis 2.5). Whenever there was a thrombus detected, it was clinically well manageable, and there were no severe clinical endpoints described after the event (Todd 2014). Pulmonary embolism was diagnosed despite no detection of DVT in one study (Ceulen 2007). It was impossible to predict if the embolus could have derived from the superficial or deep venous system. No deaths or cardiac complications were described in the pooled studies. However, the number of events was small, with a need for more studies to better answer this concern. There was still imprecision due to the low number of events and participants. Skin pigmentation also impacts on worsening cosmetic appearance. Skin pigmentation rates did not increase when using polidocanol foam 3% compared to 1% (Analysis 2.9). Aspects associated with the relationship between the vein and the skin still need to be studied (vein depth, for example). We are not sure if the depth of the vein correlates with more pigmentation rates.

One RCT demonstrated that a concentration of at least 1% polidocanol foam is necessary to reach the best effect for occluding veins (King 2015). There was no other similar finding for any other sclerotherapy agent, and the ideal polidocanol foam concentration is still not established. If the concentration of polidocanol foam was increased, the residual varicose vein rates may decrease (RR 0.67, 95% CI 0.43 to 1.04; 3 studies, 371 participants; moderate‐certainty evidence) (Analysis 2.15).

Two RCTs investigated the improvements in QoL (King 2015; Todd 2014), and no clear improvement was detected when increasing polidocanol foam concentration up to 1% polidocanol foam (Analysis 2.17).

Increasing the foam concentration for sclerotherapy improved symptoms measured using change in VCSS from baseline (polidocanol 0.125% versus 0.5%, MD ‐0.87, 95% CI ‐1.52 to ‐0.21; 2 studies, 225 participants; Analysis 2.18). When polidocanol 0.5% foam was compared to 1% foam, there was no improvement in symptoms. There is a reasonable chance to improve the effect with increasing foam concentration if symptoms could be attributed to varicose vein disease. Therefore, there is a need to define which symptoms we could attribute to the varicose veins and which symptoms cannot be attributed to varicose veins.

Only one RCT investigated the recurrence of varicose vein rates (Hamel‐Desnos 2007). No differences were detected (RR 0.91, 95% CI 0.62 to 1.32; 1 study, 148 participants; low‐certainty evidence; Analysis 2.19). Despite the low number of participants, recurrent varicose veins could have other factors driving the formation of new varicose veins. The recanalisation rate is another important issue to be studied. As described by Hamel‐Desnos 2007, the recanalisation event in sclerotherapy evolves into a smaller vein diameter. Those recanalised veins were successfully treated with a new sclerotherapy session. It is not clear if the recanalisation and reflux into a smaller vein diameter is a disease capable of producing chronic venous insufficiency in the long run since fibrous tissues in venous walls could limit newer diameter increments. Labropoulos 1995 demonstrated that saphenous veins were developing reflux after a working day in normal medical subjects (Labropoulos 1995). Maybe, in this case, reflux is not the best outcome to be evaluated. The possible confusion sources drove the authors to consider technical success as occluded, fibrotic, or reabsorbed veins.

Comparison 4: Foam sclerotherapy versus liquid sclerotherapy

Nine RCTs reported on this comparison. See Table 3.

The use of liquid or foam sclerotherapy for better cosmetic appearance results still needs to be answered since there were conflicting data due to the heterogeneity of sclerosant concentrations (polidocanol liquid concentrations ranged from 0.75% to 1.25% versus polidocanol foam 1.5% to 2.5%). In one study in the long‐term period, there was no difference in patient satisfaction (MD 0.20, 95% CI ‐0.27 to 0.67; 1 study, 126 participants; very low‐certainty evidence) (Analysis 3.1). It is necessary to consider that a liquid volume is converted to a higher volume in foam presentation, possibly being more efficient to occupy vessel lumen and to sclerose varicose veins using a smaller overall volume of sclerosant agent (Pollak 2008). Gas also causes floating phenomena that boost foam to the top part of the target vessels. In a lying patient, this floating phenomenon could cause damage only in the top portion of the venous lumen. These factors could explain more spasm in foam than in liquid intervention (Ouvry 2008), but it could drive sclerotherapy results in any effect direction since it can produce recanalisation with no success of occluding the varicose veins.

Complications were described again as mild to moderate in foam versus liquid sclerotherapy and there were no DVTs described in the participants treated. An active search for DVT was not reported by the included studies. There were no differences in skin pigmentation rates between both interventions (RR 1.75, 95% CI 0.81 to 3.80; 2 studies, 208 participants) (Analysis 3.2).

There was a consistent reduction in residual varicose vein rates at a long‐term period in the foam intervention group. We did not think it was appropriate to pool overall as different concentrations and sclerosants were used, so data have been presented as subgroup analyses. We decided to report and assess results from polidocanol 3% foam versus 3% liquid as this is the most clinically relevant: RR 0.51, 95% CI 0.41 to 0.65; 2 studies, 203 participants; very low‐certainty evidence (Analysis 3.3), indicating foam is more effective than liquid sclerotherapy in resolving residual varicose veins.

No RCT reported on quality of life or persistence of symptoms.

One additional RCT reported similar recurrence rates between the foam and liquid groups (Belcaro 2003b) (RR 1.10, 95% CI 0.86 to 1.42; 1 study, 286 participants; very low‐certainty evidence, Analysis 3.5). Differently from conventional foams, the foam produced in this RCT used a tensioactive substance to enhance ultrasound vision (ultrasound contrast agent). Otherwise, factors that drive recurrence are related to the formation of new varicose veins and maybe not be influenced by the sclerosant used.

Comparison 5: Sclerotherapy versus sclerotherapy with different substances

Four RCTs compared sclerotherapy versus sclerotherapy with any other substance. See Table 4.

Cosmetic appearance was evaluated by two RCTs but it was not possible to draw conclusions as different polidocanol and STS concentrations were used (Goldman 2002), and different vein diameters for different concentrations compared (Rao 2005), so we were not able to draw conclusions about the benefit of one product over the other (Analysis 4.1). No other sclerosant was investigated, so, the best sclerotherapy agent still needs to be determined.

King 2015 and Todd 2014 described rare neurologic events as mild and completely recovered with observation in all participants using foam with room air or with CO₂ and O₂ gas mixture (Varisolve) (Wright 2006); but there was insufficient information to indicate any ideal gas mixture for sclerotherapy. The same RCT lacked congruency since comparison was with any market sclerosant liquid or foam and did not discriminate between participants with CEAP C1 and C2 (Wright 2006). However, the RCTs found no differences in the occurrence of DVT (RR 0.97, 95% CI 0.09 to 10.54; 1 study, 384 participants; Analysis 4.2), skin pigmentation (RR 1.01, 95% CI 0.80 to 1.27; 1 study, 384 participants; Analysis 4.3), or haematoma formation (RR 1.56, 95% CI 0.76 to 3.18; 1 study, 384 participants; Analysis 4.4). Goldman 2002 found no differences in skin pigmentation rates (RR 1.21, 95% CI 0.95 to 1.55; 1 study, 167 participants) (Analysis 4.5) and skin necrosis numbers (RR 14.12, 95% CI 0.81 to 246.76; 1 study, 167 participants) (Analysis 4.6). The study also described thrombosis of the vein, probably related to thrombophlebitis or venous occlusion (superficial venous system) and did not describe DVT.

Wright 2006 also investigated the rates of residual varicose veins at 12 months and found higher rates in any market sclerosant (RR 0.64, 95% CI 0.42 to 0.97; 1 study, 373 participants; Analysis 4.7). The incongruence in the comparator made it difficult to interpret results.

No RCT investigated quality of life or persistence of symptoms. It is a disappointing finding for the systematic review since varicose vein treatment aims to improve quality of life and to reduce complication rates.

Matting also was investigated by Goldman 2002 and there were no clinical differences in the number of matting rates between the STS 1.5% group compared to polidocanol 3% (RR 1.53, 95% CI 0.58 to 4.02; 1 study, 167 participants; Analysis 4.8). There is also a need for congruency in objectives and outcomes in this comparison since varicose vein disease is probably incurable, and treatment is not directed to the aetiology of the disease (Labas 2003). Many factors that can produce varicose veins will drive the recurrence rates. Mainly, sclerotherapy alone or in association with any other intervention is essential for varicose vein treatment. Sclerotherapy is a unique and effective treatment in any varicose vein and with any diameter (Labas 2003). Sometimes, matting can be confused with a flare formation. Matting is formed by tiny vessels that usually disappear by themselves from weeks to months. As Goldman 2002 described the outcome at 16 weeks, we decided to include these results as residual varicose veins in this section (considering, possibly, a flare formation or a residual telangiectasia).

Comparison 6: Sclerotherapy versus sclerotherapy with different techniques

Six RCTs evaluated different techniques for injection sclerotherapy (Chleir 1997; Ragg 2015; Santos 2019; Yamaki 2009; Yamaki 2012; Zhang 2014).

Surprisingly, no RCT investigated cosmetic appearance using different techniques.

There were no thromboembolic complications detected in the antegrade versus retrograde sclerotherapy comparison (Analysis 5.1) (Zhang 2014). Phlebitis and thrombophlebitis, and skin pigmentation rates were also not different (Analysis 5.1; Analysis 5.2). It is important to report that, in a study comparing few versus multiple injections for sclerotherapy, less foam was detected in the deep veins when using multiple injections (Yamaki 2009) (Analysis 5.4). If an embolus originated from the superficial venous system, there is a better prognosis due to the smaller diameter of the veins (Wu 2017). On the other hand, the deep venous system has the potential to produce a bigger embolus increasing the risk of death to patients. Thromboembolic events were related to a proximal extension of a superficial thrombus usually derived from saphenous veins, or the extension of a perforator vein. DVT can also occur without an extension thrombus from the superficial venous system. So it is essential to investigate if foam in higher concentrations in the deep vein could cause any hazard to patients.

Injection by a catheter associated with tumescent technique reduced residual varicose veins in the short‐term period compared to needle injection (Santos 2019). However, other different sclerotherapy techniques did not show any benefit in reducing residual varicose vein rates, clinical severity scores, and complications (Analysis 5.5; Analysis 5.6; Analysis 5.7; Analysis 5.8; Analysis 5.9; Analysis 5.10; Analysis 5.14). After six months, participants from both groups (catheter‐delivered associated with tumescent technique versus needle‐delivered sclerotherapy) needed retreatment (Santos 2019). The studies need to be more congruent and with a larger number of participants to increase certainty. Despite incongruence, it does not appear that different techniques produce different outcomes. Concerns about outcomes remain related to proximal versus distal technique for sclerotherapy (if treating proximal saphenous vein versus distal saphenous vein).

Sclerotherapy versus different conservative treatment

Comparison 7: Sclerotherapy versus pharmacological therapy

No RCT evaluated sclerotherapy with pharmacological therapy. It is important to note that this comparison involves expensive medication for some people. So, a study including a cost‐effectiveness analysis should be carried out to ensure there are more benefit than harms to patients since varicose veins impact quality of life more than they risk life.

Comparison 8: Sclerotherapy versus conservative compression treatment

Only one RCT reported on this comparison (Abramowitz 1973). The RCT was conducted in 1973 in South Africa and enrolled only pregnant women. It is of some importance since pregnant women are a subgroup of vulnerable patients at risk of complications (phlebitis, deep vein thrombosis, bleeding) (Smyth 2015). In the RCT, there was an improvement of pooled data from symptomatic and cosmetic improvement outcomes (RR 7.96, 95% CI 3.38 to 18.76; 1 study, 101 participants; Analysis 6.1), at the same time, producing more leg ulcer rates in the sclerotherapy group (RR 13.63, 95% CI 0.77 to 240.13; 1 study, 101 participants; Analysis 6.2). Despite the need to treat pregnant women, there are concerns related to the risks of DVT and PE produced by the procedure. As described by Smyth 2015, there is still a paucity of data in the literature. Ethical concerns will drive future research.

Overall completeness and applicability of evidence

Overall, 4278 participants were included in the RCTs, increasing the confidence that the procedure is a safe and an effective treatment. The aim of the review was to include CEAP C2, C3, and C4 criteria varicose veins. Due to differences in the vein diameter, the effects presented in this review are likely not relevant to people with CEAP C1, but possibly some results could also be applied to people with CEAP C5 and C6 (e.g. occlusion rates). Even in the cases of residual varicose veins, new sessions of sclerotherapy are feasible and acceptable procedures while other more invasive methods are related to patients' costs and risks.

RCTs described cosmetic appearance and complications (including staining, DVT and PE). However, there is a lack of consistency between interventions, time points, and outcomes evaluated. There was variability and heterogeneity in participants, drugs, drug concentrations, varicose veins classification, and interventions (e.g. usage or non‐usage of stockings after sclerotherapy). The present systematic review did not investigate the role of co‐interventions.

Overall, varicose veins are heterogeneous and, even when not described, there are likely to be smaller varicose veins (including CEAP C1) associated with the clinical classes described. Sclerosant agents may reach these veins, as noted in clinical practice. So it is important to have similar proportions of participants with C1, C5, and C6 between intervention and control groups. We described these proportions of CEAP C1, C5, and C6 in the footnotes of the meta‐analyses to assist with a critical analysis. It was not possible to estimate if the depth of the vein could result in different ratings of skin pigmentation due to the lack of information provided.

Quality of life and symptoms were poorly reported. Some studies used VCSS to report symptoms; however, for VCSS, there is one attribute for symptoms and eight attributes for signs. So, there is a lack of data to increase confidence in the estimate of the effect. For symptoms, there is a need for establishing a cause‐effect relationship.

It should be taken into account that it is not the aim of this review to evaluate injection sclerotherapy versus varicose vein surgery methods. This subject is the scope of another Cochrane Review. Applicability of the evidence is limited to injection sclerotherapy and their comparators with other types of sclerotherapy including other techniques of performing injection sclerotherapy.

There was no evidence found for three interventions due to a lack of RCTs investigating these.

• Sclerotherapy versus no interventions for varicose veins

• Sclerotherapy versus sclerotherapy at different time intervals

• Sclerotherapy versus pharmacological therapy

Quality of the evidence

Overall, certainty of the evidence was very low to moderate. This is due to the heterogeneity of the interventions and population (including participants with CEAP class C1, C5, and C6 in the studies), variation in concentration of medicines, different time point measures, lack of core outcome set definitions (for instance: recurrent varicose veins and residual varicose veins), some outcomes evaluated by single RCTs, low number of events (e.g. DVT), wide confidence intervals including the line of no effect, most risk of bias domains judged as unclear, and different units of analysis studied. We decided to use the GRADE approach to assess our certainty of the evidence for the outcomes cosmetic appearance, complications (thromboembolic complications), residual varicose veins, quality of life, persistence of symptoms, and recurrent varicose veins. We created a summary of findings table for each of the four main comparisons.

• Foam sclerotherapy (1% polidocanol) versus placebo for treating varicose veins (Table 1)

• Foam sclerotherapy (3% polidocanol) versus foam sclerotherapy (polidocanol 1%) for treating varicose veins (Table 2)

• Foam sclerotherapy versus liquid sclerotherapy for treating varicose veins (Table 3)

• Sclerotherapy with STS 1% versus sclerotherapy with polidocanol 1% (Table 4)

Foam sclerotherapy (polidocanol 1%) versus placebo

See Table 1.

Five RCTs compared polidocanol foam sclerotherapy versus placebo but only three provided results from polidocanol 1% foam versus placebo to grade the certainty of the evidence (Gibson 2010; King 2015; Todd 2014). There was a considerable variation of polidocanol concentration used, ranging from 0.125% (King 2015; Todd 2014) to 2% (King 2015). For grading certainty of the evidence, polidocanol 1% foam was chosen as the interventional group since there is a trend to increase occluded veins rates with up to 1% polidocanol foam (King 2015), and 1% concentration is clinically relevant.

We understand that is impossible to blind personnel and participants for this comparison due to the nature of the intervention but we downgraded two levels for risk of bias concerns.

For cosmetic appearance, the evidence was downgraded to very low‐certainty for risk of bias concerns and indirectness (one RCT (Todd 2014) included 8.8% participants with CEAP C5 and C6 in the placebo group and 3.4% participants with CEAP C5 and C6 in the 1% polidocanol foam group ‐ we planned to include CEAP C2, C3 or C4 participants. As the study authors included participants with other CEAP classes and we were unable to obtain information to exclude these data, we determined the population in the study as an indirect population).

The evidence for thromboembolic complications was downgraded to very low‐certainty for risk of bias concerns and due to the indirectness, imprecision (low number of events, wide confidence intervals), and both clinical heterogeneity among studies (severity of the disease) and statistical heterogeneity (I² = 53%).

The outcome, residual varicose veins, was downgraded to very low‐certainty evidence due to the same limitations observed in the cosmetic appearance outcome.

Quality of life and the persistence of symptoms were both downgraded to very low‐certainty evidence due to indirectness (8.8% of participants in the placebo group had CEAP C5 or C6), imprecision (low number of participants, and wide confidence intervals).

No RCT investigated recurrent varicose veins.

Foam sclerotherapy (polidocanol 3%) versus foam sclerotherapy (polidocanol 1%)

See Table 2.

Polidocanol foam 3% versus 1% concentrations were chosen to evaluate the certainty of the evidence due to clinical relevance.

Only Ceulen 2007 evaluated cosmetic appearance. As most of the risk of bias domains were unclear, inclusion of only a single study with a wide confidence interval (imprecision), and a low number of participants (imprecision); the evidence was downgraded to very low‐certainty.

Thromboembolic complications were measured by three RCTs (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007) and the evidence was judged as very low‐certainty due to indirectness (there were participants with CEAP C1 (11.63% in Ceulen 2007), C5 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007) and C6 (Hamel‐Desnos 2007) but the lower legs were equal in each group), imprecision (low number of events) and wide confidence intervals. Ceulen 2007 used the lower leg as the unit of analysis and there was a heterogeneous population (11.64% with CEAP C1), so a sensitivity analysis was carried out (Analysis 2.6). The effect direction was not changed.

Residual varicose veins were evaluated by three RCTs (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), with indirectness (there were participants with CEAP C1 (11.63% in Ceulen 2007), C5 (Blaise 2010; Ceulen 2007; Hamel‐Desnos 2007), and C6 (Hamel‐Desnos 2007) but these were equal in each group). Therefore, the evidence was rated as moderate‐certainty evidence.

No RCT described the quality of life or the persistence of symptoms for this comparison.

Recurrent varicose veins were investigated by one RCT (Hamel‐Desnos 2007). We detected imprecision (single study), and indirectness (participants with CEAP C5 and C6 (Hamel‐Desnos 2007) without describing the proportion of these participants in the study arms). We downgraded by two levels and the evidence was judged as low‐certainty evidence.

Foam sclerotherapy versus liquid sclerotherapy

See Table 3.

Ten RCTs reported on the comparison foam versus liquid sclerotherapy. There were outcomes that may not have correlated with clinical implications (immediate spasm after sclerotherapy) (Ouvry 2008) and substantial differences in sclerosant concentrations (ranging from 1.5% to 2.5% in liquid versus 0.75% to 1.25% in the foam group (Alos 2006)), lacking congruency.

For cosmetic appearance, the certainty of the evidence was downgraded to very low due to lack of description of the randomisation method, risk of bias concerns (the RCT probably not blinded due to the nature of the intervention), indirect population, single study, and comparing different concentrations of the same substances in different vein diameters in both arms of the study.

No RCT investigated DVT, quality of life, or persistence of symptoms.

There was a lack of data for the complications outcome, despite some RCTs describing complications such as thrombophlebitis and skin pigmentation as mild. No DVT events were described and we are not sure if the outcome was systematically investigated in the RCTs.

Six RCTs reported on residual varicose veins and two of them compared polidocanol 3% liquid versus foam (Ouvry 2008; Rabe 2008). We assessed this outcome as having very low‐certainty evidence. The quality was downgraded since there was doubt about the randomisation method, the intervention was presumably not blinded due to the nature of the intervention, and one study (Ouvry 2008) included participants from CEAP C2 to C6 without discriminating between then. The other study (Rabe 2008) included one participant with CEAP C5 in the foam group (1.85%) and four participants with CEAP C5 in the liquid group intervention (7.69%).

Recurrent varicose veins and flare formation was investigated by one RCT (Belcaro 2003b). There was some doubt about the randomisation method used. The intervention was presumably not blinded due to the nature of the intervention. There was a single RCT, and the author did not describe CEAP participants. Belcaro 2003b also used a tensioactive substance to enhance ultrasound vision (ultrasound contrast agent) to produce foam, which is not the usual way to produce foam. We judged the evidence as very low‐certainty.

Sclerotherapy with STS 1% versus sclerotherapy with polidocanol 1%

Data were insufficient to allow assessment using the GRADE criteria. See Table 4.

Potential biases in the review process

We were unable to carry out a funnel plot analysis to investigate publication bias since there were fewer than 10 RCTs for each meta‐analysis. Due to unavailable data, it was impossible to perform a subgroup analysis between CEAP C2, C3, and C4 groups and symptomatic versus nonsymptomatic participants. Indeed, no RCTs evaluated the depth of the vein as a possible issue influencing pigmentation. Also, previous complications were exclusion criteria for the studies, limiting subgroup analysis.

Authors were contacted to reduce bias in the review process. Most of them did not reply. Some RCTs were old, and others had no contact authors available (Abramowitz 1973; Chleir 1997; Demagny 2002; Ragg 2015; Schadeck 1995a; Zeh 2003; Zhang 2009; Zhang 2014).

Two RCTs had heterogeneous populations including some participants with CEAP C1 (Ceulen 2007; Zhang 2009). It was expected that there would be fewer thromboembolic and pigmentation rates in some of these subgroups. Differences from protocols were noted where participants with CEAP C1 should not be included. Ceulen 2007 included 12.8% of participants with CEAP C1, and Zhang 2009 did not discriminate between the proportions of the C1 class. Given this consideration, we included Ceulen 2007 in the meta‐analysis but Zhang 2009 was not included since the author did not describe the proportion of CEAP C1 included in each group. A sensitivity analysis was conducted to check the robustness of the results. Similarly, other RCTs considered CEAP C5 and C6 inclusions. All of them were described, with the proportions of the non‐proposed clinical classes included in each analysis and taken into consideration when interpreting the results.

Three studies were assessed as 'awaiting classification'; there was a lack of data and we were unable to contact the authors for two of the studies (Labas 2003; Satokawa 2003; Schadeck 1995b); and we were unable to determine if one study was a randomised or quasi‐randomised controlled trial (Varnagy 1985).

Co‐interventions were used during and after injection sclerotherapy (e.g. walking after injection sclerotherapy) (King 2015; Todd 2014). These co‐interventions may have influenced the outcomes and there could be other confounders to consider. One crucial co‐intervention is the use or not of medical stockings after injection sclerotherapy. We suggest that this issue could be better evaluated in a further systematic review to evaluate sclerotherapy versus compression sclerotherapy for varicose veins. Lattimer 2012 described a lack of criteria between side effects and treatment effects, making it difficult to quantify in a meaningful way an effect size in a multi‐ethnic and multicultural population. Lattimer 2012 also concluded that pigmentation, inflammation, pain, and other side effects could be effects from the stockings rather than from sclerotherapy. These arguments contributed to deciding not to include compression stockings in this review. Despite that, the co‐intervention was common in the included studies, and the real effect of compression was not evaluated.

Despite a proposed subgroup analysis to evaluate the different gases for foam sclerotherapy, there was no study comparing them. So, the issue of different gases to reduce incidence of neurological complications is still to be resolved (Gibson 2010).

We considered it important to make a distinction between 'recurrent varicose veins' and 'residual varicose veins' because of the clinical relevance. The previous version of this review considered the time to measure the outcome and 'recurrent varicose veins' was defined as an outcome related to the time of the detection (described as 'long‐term patient follow‐up'). The previous version of the review included 'flare formation' in the same outcome. We believe these outcomes are quite different and they should not be measured together. Differences in definitions used by the study authors led us to consider the long‐term outcome as a 'recurrent varicose vein' and short‐term outcome as 'residual varicose vein'. This definition carries bias but we believe we need to be pragmatic. We will consider the subject again in a future update of this review.

As with many other interventions, age is likely to increase heterogeneity between study groups. Despite balanced groups there could be different results in older and younger patients. This Cochrane review did not address this topic since we did not have data to investigate this. If possible, we will address this concern in a future update.

Agreements and disagreements with other studies or reviews

One systematic review described foam sclerotherapy as an effective and safe procedure similar to the findings of this review (Rathbun 2012). It also described rare serious events like myocardial infarction and death related to sclerotherapy detected in case reports. Our update did not find any of these severe events. Death rates could be a result of the amount of sclerotherapy agent administered as well as anaphylactic reaction and the RCTs were perhaps aware of the injection volume. Rabe 2014 recommended the maximum injection foam volume of 10 mL since complications like thrombosis and neurological symptoms tend to occur in higher volume injections. This updated review considered only randomised clinical trials while Rathbun 2012 considered other study types, including case reports, making it difficult to directly compare findings. However, it is important to report that most neurologic events are described in case reports since they are rare (Sarvananthan 2012; Willemberg 2013). So, systematic reviews involving case reports could find more events than those described in the present update.

A systematic review by Hammel‐Desnos 2009 found foam better than liquid for reducing rates of residual varicose veins. This review update confirmed this finding with very low‐certainty evidence. There is still a need for congruence in comparators and doses.

No other systematic review involved only sclerotherapy, highlighting the importance of this update to answer clinical questions.

Authors' conclusions

Implications for practice.

There is a very low‐ to low‐certainty evidence that, compared to placebo, sclerotherapy is an effective and safe treatment for varicose veins concerning cosmetic appearance, residual varicose veins, QoL, and persistence of symptoms. Rates of DVT may be slightly increased and there were no data concerning recurrent varicose veins. There was limited or no evidence for one concentration of foam compared to another; foam compared to liquid sclerotherapy; foam compared to any other substance; or one technique compared to another.

Implications for research.

The cost‐effectiveness of the treatment of varicose veins has significant implications for healthcare. Although there are some undesirable residual varicose veins, additional sclerotherapy sessions could be performed to resolve them (Hamel‐Desnos 2007). Sclerotherapy is a cheaper procedure when compared to any other surgical intervention for varicose veins, despite the issue of increasing costs with newer sessions; it might be the most cost‐effective treatment available (Carrol 2013). Sclerotherapy reaches veins which are impossible for treatment by other methods, so it could always be used as the main or the complementary choice of treatment. Possibly there is a dose‐response effect, and an increase in the concentrations of polidocanol foam decreased residual varicose veins up to polidocanol 3%. We do not know if increasing the polidocanol concentration even more could produce better results or if it could increase risks to patients. This question still needs to be answered with new studies.

There is a need for high‐quality RCTs with a large number of participants, with comparable CEAP clinical classification, standardised sclerosant dose, clearly defined outcomes and time points to measure. A core outcome set for varicose veins would be helpful in defining the best outcomes to be studied. Adverse events should be better evaluated, mainly for higher concentrations of polidocanol foam (e.g. 3%). There is a need for studies comparing varicose vein interventions versus no interventions to understand if the natural history of the disease is worse than the risks of the treatment.

There is a lack of RCTs using the same time point to reproduce results from cosmetic appearance after sclerotherapy. Ideally, cosmetic appearance should be evaluated after three to six months since this period is sufficient for phlebitis to disappear and to reabsorb the sclerosis material. One RCT described the same success rates at six months, with a need for new sclerotherapy sessions (Santos 2019). Probably, as described in another study, recanalisation occurrence is high and should be reviewed at six months (Hamel‐Desnos 2007). It is essential to avoid limiting the diameter of the vein studied. Instead, there should be strict criteria for varicose vein definition. We propose that only dilated, tortuous veins with loss of function are included. It is possible that a considerable number of veins treated with a low diameter and with reflux detected are not varicose veins. Greater diameter varicose veins are subject to events influencing sclerosing: 1) effect of Archimedes (greater veins depend on higher volumes to occupy the lumen); and 2) floating phenomenon (foam tends to fluctuate and treat the upper wall of the vein leaving the base untreated). These effects should be considered for a subgroup analysis using foam sclerotherapy in larger varicose veins.

Complications are rare events associated with sclerotherapy. The low number of events contributed to downgrading the certainty of the evidence. Therefore, RCTs focussing on complications should include a larger number of participants. There are complications limited to a short period of follow‐up (e.g. DVT and PE), and time points should have a rationale for those events. We propose time points for evaluating thromboembolic events up to 30 days after the last injection and between 30 days to six months. One other single source of bias is the amount of sclerosant injected in a single puncture or multiple ones. There is a lack of RCTs evaluating if the amount of sclerosant is important in ensuring the sclerosant agent reaches the deep venous system, providing a source of deep vein thrombosis. Comparing different volumes injected in each puncture could reduce the risk of including another source of bias for DVT, thrombophlebitis, and cerebrovascular complications. Skin pigmentation, on the other hand, should be better described at three, six, and 24 months since it appears later in the process and could also disappear later on.

Evidence from residual varicose veins and recurrent varicose veins is confused due to differences in definitions and time point measures. To date, the spasm frequently does not lead to venous occlusion and recanalisation is not recurrent varicose veins. So, we propose one week to one month ideally to verify residual varicose veins and occluded veins after a sclerotherapy session. After this period, it is possible to develop recurrent varicose veins. Evidence of occlusion should lead to the diagnosis of recanalisation. These aspects introduce a necessity to clarify outcomes and time points to measure them. We also understand that time intervals between sclerotherapy sessions could have a substantial impact on residual and recurrent varicose veins. Time‐interval sclerotherapy is a particularly important issue since every treatment for varicose veins may have an influence on the rate of recanalisation and pigmentation. If necessary, new sessions of sclerotherapy should be performed without increasing risk but with increments in effectiveness (Hamel‐Desnos 2007).

Despite the low number of RCTs describing quality of life, this is the critical factor for considering a varicose vein treatment. Quality of life is affected in many ways in the course of the disease. Indeed, we still need to know if different gases, substances, presentations (liquid versus foam), and many other comparisons could be the best way to treat varicose veins (e.g. using the same concentration of sclerosants which one would be the most effective to treat varicose veins?). Finally, further studies are needed to increase the number of events and so our confidence in the effect.

Different techniques, including proximal versus distal treatment sclerotherapy, need to be studied since proximal sclerotherapy could have other effects (e.g. reducing the diameter of distal varicose veins for future new sessions) and risks (e.g. DVT and PE).

There is a need for definitions of core outcome sets for varicose veins since some outcomes and the best way to measure them are still to be defined.

Feedback

Forest plots errors, 6 April 2011

Summary

I noticed a few errors:
1. Analysis 7.1 ‐ the "Favours sclero" and "favours stockings" headings under the Forest plot are the wrong way round, which is very misleading (many reader will not read the text fully and will only look at the plots)
2. Several of the Forest plots have non‐sensical results ‐ SDs are zero, and MD and CIs are also zero ‐ for example, analysis 6.5.

Reply

The forest plot labels for Analysis 7.1 have been corrected. Comment 2 was addressed in the update of this review.

Contributors

Feedback: Mark Perry, Occupation: Research Fellow
Response: Marlene Stewart, Managing Editor on behalf of the authors.

What's new

Date	Event	Description
21 July 2021	New citation required but conclusions have not changed	New search run; 23 new studies included and 12 new studies excluded. New author team taken over review, inclusion criteria amended and text updated to reflect current Cochrane standards. 'Summary of findings' tables added. No change to conclusions.
21 July 2021	New search has been performed	New search run; 23 new studies included and 12 new studies excluded.

History

Protocol first published: Issue 3, 1999
Review first published: Issue 1, 2002

Date	Event	Description
6 April 2011	Feedback has been incorporated	Feedback received and incorporated
17 October 2008	Amended	Converted to new review format.
6 November 2006	New search has been performed	Searches re‐run and no new studies found. Search dates amended. Citation error and copy editing errors corrected.
21 August 2006	New citation required but conclusions have not changed	Five further RCTs included and nine RCTs excluded. Plain language summary added. Contributions of Reviewers updated. Overall conclusions unchanged.
26 May 2004	New search has been performed	Two further RCTs included. Overall conclusions unchanged.

Appendices

Appendix 1. CRS search strategy

Search run on Tue Jan 19 2016
#1	MESH DESCRIPTOR Varicose Veins EXPLODE ALL TREES	744
#2	(varicos* near3 (vein* or veno*)):TI,AB,KY	750
#3	(tortu* near3 (vein* or veno*)):TI,AB,KY	6
#4	(incomp* near3 (vein* or veno* or saphenous or valv*)):TI,AB,KY	83
#5	(insuffic* near3 (vein* or veno* or saphenous)):TI,AB,KY	127
#6	(((saphenous or vein* or veno*) near3 reflux)):TI,AB,KY	119
#7	MESH DESCRIPTOR Saphenous Vein EXPLODE ALL TREES WITH QUALIFIERS SU	194
#8	(GSV or CVI or SSV):TI,AB,KY	271
#9	MESH DESCRIPTOR Venous Insufficiency EXPLODE ALL TREES	375
#10	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9	1593
#11	MESH DESCRIPTOR Sclerotherapy EXPLODE ALL TREES	429
#12	MESH DESCRIPTOR Sclerosing Solutions EXPLODE ALL TREES	370
#13	sclero*:TI,AB,KY	7871
#14	(tetradecyl near2 (sulfate or sulphate)):TI,AB,KY	57
#15	MESH DESCRIPTOR Sodium Tetradecyl Sulfate EXPLODE ALL TREES	36
#16	MESH DESCRIPTOR Saline Solution, Hypertonic EXPLODE ALL TREES	405
#17	MESH DESCRIPTOR Ethanolamines	1455
#18	(polydocanol or polidocanol):TI,AB,KY	193
#19	saline:TI,AB,KY	17786
#20	(ethanolamine near2 oleate):TI,AB,KY	63
#21	(sodium near/ morrhuate):TI,AB,KY	0
#22	sotradecol:TI,AB,KY	6
#23	(aetoxisclerol or aethoxysclerol):TI,AB,KY	15
#24	(aetoxiskerol or aethoxyskerol):TI,AB,KY	1
#25	Turbofoam:TI,AB,KY	2
#26	(foam* or microfoam*):TI,AB,KY	1200
#27	varisolve:TI,AB,KY	2
#28	((atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo)):TI,AB,KY	3
#29	#11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28	28196
#30	#10 AND #29	243

Appendix 2. Sources searched and search strategies

Source	Search strategy	Hits retrieved
VASCULAR REGISTER IN CRSW (Date of most recent search: 20 July 2021)	#1 Varicose Veins OR Venous Insufficiency OR VVEINS* AND INREGISTER #2 Sclero* OR Sodium Tetradecyl Sulfate OR Saline Solution, Hypertonic OR Ethanolamines AND INREGISTER #3 #1 AND #2	Jan 2019: 15 Jan 2020: 14 July 2021: 11
CENTRAL via CRSO (Date of most recent search: 20 July 2021)	#1 MESH DESCRIPTOR Varicose Veins EXPLODE ALL TREES #2 (varicos* near3 (vein* or veno*)):TI,AB,KY #3 (tortu* near3 (vein* or veno*)):TI,AB,KY #4 (incomp* near3 (vein* or veno* or saphenous or valv*)):TI,AB,KY #5 (Insuffic* near3 (vein* or veno* or saphenous)):TI,AB,KY #6 ((saphenous or vein* or veno*) near3 reflux):TI,AB,KY #7 MESH DESCRIPTOR Saphenous Vein EXPLODE ALL TREES #8 (GSV or CVI or SSV):TI,AB,KY #9 MESH DESCRIPTOR Venous Insufficiency EXPLODE ALL TREES #10 #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 #11 MESH DESCRIPTOR Sclerotherapy EXPLODE ALL TREES #12 MESH DESCRIPTOR Sclerosing Solutions EXPLODE ALL TREES #13 sclero*:TI,AB,KY 12042 #14 (tetradecyl near2 (sulfate or sulphate)):TI,AB,KY #15 MESH DESCRIPTOR Sodium Tetradecyl Sulfate EXPLODE ALL TREES #16 MESH DESCRIPTOR Saline Solution, Hypertonic EXPLODE ALL TREES #17 MESH DESCRIPTOR Ethanolamines EXPLODE ALL TREES #18 (polydocanol or polidocanol):TI,AB,KY #19 saline:TI,AB,KY 21058 #20 (ethanolamine near2 oleate):TI,AB,KY 68 #21 (sodium near morrhuate):TI,AB,KY #22 sotradecol:TI,AB,KY #23 (aetoxisclerol or aethoxysclerol):TI,AB,KY #24 (aetoxiskerol or aethoxyskerol):TI,AB,KY 1 #25 Turbofoam:TI,AB,KY #26 (foam* or microfoam*):TI,AB,KY 1724 #27 varisolve:TI,AB,KY #28 (atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo):TI,AB,KY #29 #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 #30 #10 AND #29 #31 01/01/2016 TO 14/01/2019:CD #32 #30 AND #31	Jan 2019: 120 Jan 2020: 125 July 2021: 73
Clinicaltrials.gov	Varicose Veins OR Venous Insufficiency | Sclerotherapy OR Sclerosing Solutions OR Sodium Tetradecyl Sulfate OR Saline Solution, Hypertonic OR Ethanolamines	Jan 2019: 16 Jan 2020: 1 July 2021: 8
ICTRP Search Portal (Date of most recent search: 20 July 2021)	Varicose Veins OR Venous Insufficiency | Sclerotherapy OR Sclerosing Solutions OR Sodium Tetradecyl Sulfate OR Saline Solution, Hypertonic OR Ethanolamines	Jan 2019: 15 Jan 2020: 4 July 2021: N/A
MEDLINE (Ovid MEDLINE® Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE® Daily and Ovid MEDLINE®) 1946 to present (Date of most recent search: 20 July 2021)	1 exp Varicose Veins/ 2 (varicos* adj3 (vein* or veno*)).ti,ab. 3 (tortu* adj3 (vein* or veno*)).ti,ab. 4 (incomp* adj3 (vein* or veno* or saphenous or valv*)).ti,ab. 5 (insuffic* adj3 (vein* or veno* or saphenous)).ti,ab. 6 ((saphenous or vein* or veno*) adj3 reflux).ti,ab. 7 exp Saphenous Vein/su [Surgery] 8 (GSV or CVI or SSV).ti,ab. 9 exp Venous Insufficiency/ 10 or/1‐9 11 exp SCLEROTHERAPY/ 12 exp Sclerosing Solutions/ 13 sclero*.ti,ab. 14 (tetradecyl adj2 (sulfate or sulphate)).ti,ab. 15 exp Sodium Tetradecyl Sulfate/ 16 exp Saline Solution, Hypertonic/ 17 exp ETHANOLAMINES/ 18 (polydocanol or polidocanol).ti,ab. 19 saline.ti,ab. 20 (ethanolamine adj2 oleate).ti,ab. 21 (sodium adj morrhuate).ti,ab. 22 sotradecol.ti,ab. 23 (aetoxisclerol or aethoxysclerol).ti,ab. 24 (aetoxiskerol or aethoxyskerol).ti,ab. 25 Turbofoam.ti,ab. 26 (foam* or microfoam*).ti,ab. 27 varisolve.ti,ab. 28 (atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo).ti,ab. 29 or/11‐28 30 10 and 29 31 randomized controlled trial.pt. 32 controlled clinical trial.pt. 33 randomized.ab. 34 placebo.ab. 35 drug therapy.fs. 36 randomly.ab. 37 trial.ab. 38 groups.ab. 39 or/31‐38 40 exp animals/ not humans.sh. 41 39 not 40 42 30 and 41 43 (2017* or 2018* or 2019*).ed. 44 42 and 43 45 from 44 keep 1‐60	Jan 2019: 60 Jan 2020: 60 July 2021: 43
Embase via OVID (Date of most recent search: 20 July 2021)	1 exp varicosis/ 2 (varicos* adj3 (vein* or veno*)).ti,ab. 3 (tortu* adj3 (vein* or veno*)).ti,ab. 4 (incomp* adj3 (vein* or veno* or saphenous or valv*)).ti,ab. 5 (insuffic* adj3 (vein* or veno* or saphenous)).ti,ab. 6 ((saphenous or vein* or veno*) adj3 reflux).ti,ab. 7 exp saphenous vein/ 8 (GSV or CVI or SSV).ti,ab. 9 exp vein insufficiency/ 10 or/1‐9 11 exp endoscopic sclerotherapy/ or exp sclerotherapy/ 12 exp sclerosing agent/ 13 sclero*.ti,ab. 14 exp tetradecyl sulfate sodium/ 15 exp ethanolamine derivative/ 16 (polydocanol or polidocanol).ti,ab. 17 saline.ti,ab. 18 (ethanolamine adj2 oleate).ti,ab. 19 sotradecol.ti,ab. 20 (aetoxisclerol or aethoxysclerol).ti,ab. 21 (aetoxiskerol or aethoxyskerol).ti,ab. 22 Turbofoam.ti,ab. 23 (foam* or microfoam*).ti,ab. 24 varisolve.ti,ab. 25 (atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo).ti,ab. 26 (sodium adj morrhuate).ti,ab. 27 or/11‐26 28 10 and 27 29 randomized controlled trial/ 30 controlled clinical trial/ 31 random$.ti,ab. 32 randomization/ 33 intermethod comparison/ 34 placebo.ti,ab. 35 (compare or compared or comparison).ti. 36 ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab. 37 (open adj label).ti,ab. 38 ((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab. 39 double blind procedure/ 40 parallel group$1.ti,ab. 41 (crossover or cross over).ti,ab. 42 ((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)).ti,ab. 43 (assigned or allocated).ti,ab. 44 (controlled adj7 (study or design or trial)).ti,ab. 45 (volunteer or volunteers).ti,ab. 46 trial.ti. 47 or/29‐46 48 28 and 47 49 (2017* or 2018* or 2019*).em. 50 48 and 49 51 from 50 keep 1‐297	Jan 2019: 297 Jan 2020: 243 July 2021: 150
CINAHL via EBSCO (Date of most recent search: 20 July 2021)	S42 S28 AND S41 S41 S29 OR S30 OR S31 OR S32 OR S33 OR S34 OR S35 OR S36 OR S37 OR S38 OR S39 OR S40 S40 MH "Random Assignment" S39 MH "Single‐Blind Studies" or MH "Double‐Blind Studies" or MH "Triple‐Blind Studies" S38 MH "Crossover Design" S37 MH "Factorial Design" S36 MH "Placebos" S35 MH "Clinical Trials" S34 TX "multi‐centre study" OR "multi‐center study" OR "multicentre study" OR "multicenter study" OR "multi‐site study" S33 TX crossover OR "cross‐over" S32 AB placebo* S31 TX random* S30 TX trial* S29 TX "latin square" S28 S10 AND S27 S27 S11 OR S12 OR S13 OR S14 OR S15 OR S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 S26 TX atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo S25 TX varisolve S24 TX foam* or microfoam* S23 TX Turbofoam S22 TX aetoxisclerol or aethoxysclerol S21 TX sotradecol S20 TX sodium n2 morrhuate S19 TX ethanolamine n2 oleate S18 TX saline S17 TX polydocanol or polidocanol S16 MH "Ethanolamines+" S15 MH "Saline Solution, Hypertonic" S14 TX tetradecyl n2 (sulfate or sulphate) S13 TX sclero* S12 MH "Sclerosing Solutions" S11 MH "Sclerotherapy" S10 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 OR S8 OR S9 S9 MH "Venous Insufficiency+" S8 TX GSV or CVI or SSV S7 TX Saphenous Vein S6 TX ((saphenous or vein* or veno*) n3 reflux) S5 TX insuffic* n3 (vein* or veno* or saphenous) S4 TX tortu* n3 (vein* or veno*) S3 TX incomp* n3 (vein* or veno* or saphenous or valv*) S2 TX varicos* n3 (vein* or veno*) S1 MH "Varicose Veins+"	Jan 2019: 34 Jan 2020: 9 July 2021: 15
AMED via OVID (Date of most recent search: 20 July 2021)	1 (varicos* adj3 (vein* or veno*)).ti,ab. 2 (tortu* adj3 (vein* or veno*)).ti,ab. 3 (incomp* adj3 (vein* or veno* or saphenous or valv*)).ti,ab. 4 (insuffic* adj3 (vein* or veno* or saphenous)).ti,ab. 5 ((saphenous or vein* or veno*) adj3 reflux).ti,ab. 6 (GSV or CVI or SSV).ti,ab. 7 or/1‐6 8 sclero*.ti,ab. 9 (tetradecyl adj2 (sulfate or sulphate)).ti,ab. 10 (polydocanol or polidocanol).ti,ab. 11 saline.ti,ab. 12 (ethanolamine adj2 oleate).ti,ab. 13 (sodium adj morrhuate).ti,ab. 14 sotradecol.ti,ab. 15 (aetoxisclerol or aethoxysclerol).ti,ab. 16 (aetoxiskerol or aethoxyskerol).ti,ab. 17 Turbofoam.ti,ab. 18 (foam* or microfoam*).ti,ab. 19 varisolve.ti,ab. 20 (atoxisclerol or Sotrauerix or Laureth or chrome alum or Scleremo).ti,ab. 21 or/8‐20 22 7 and 21 23 exp CLINICAL TRIALS/ 24 RANDOM ALLOCATION/ 25 DOUBLE BLIND METHOD/ 26 Clinical trial.pt. 27 (clinic* adj trial*).tw. 28 ((singl* or doubl* or trebl* or tripl*) adj (blind* or mask*)).tw. 29 PLACEBOS/ 30 placebo*.tw. 31 random*.tw. 32 PROSPECTIVE STUDIES/ 33 or/23‐32 34 22 and 33 35 ("2017" or "2018" or "2019").yr. 36 34 and 35	Jan 2019: 0 Jan 2020: 0 July 2021: 0
LILACS (Date of most recent search: 20 July 2021)	Varicose OR Venous Insufficiency [Words] and Sclerotherapy OR Sclerosing Solutions OR Sodium Tetradecyl Sulfate OR Saline Solution, Hypertonic OR Ethanolamines	Jan 2019: 6 Jan 2020: 1 July 2021: 3
TOTAL before de‐duplication		Jan 2019: 570 Jan 2020: 457 July 2021: 301
TOTAL after de‐duplication		Jan 2019: 431 Jan 2020: 365 July 2021: 238

Data and analyses

Comparison 1. Foam sclerotherapy versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Cosmetic appearance: medical related IPR‐V adjusted mean change from baseline (intermediate term)	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1.1 Polidocanol 1% foam versus placebo	2	223	Mean Difference (IV, Random, 95% CI)	‐0.76 [‐0.91, ‐0.60]
1.1.2 Polidocanol 2% foam versus placebo	1	119	Mean Difference (IV, Random, 95% CI)	‐0.90 [‐1.11, ‐0.69]
1.2 Cosmetic appearance: participant self assessment PA‐V adjusted mean change from baseline (intermediate term)	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.2.1 Polidocanol 1% foam versus placebo	2	223	Mean Difference (IV, Random, 95% CI)	‐1.46 [‐1.72, ‐1.20]
1.2.2 Polidocanol 2% foam versus placebo	1	119	Mean Difference (IV, Random, 95% CI)	‐1.60 [‐1.95, ‐1.25]
1.3 DVT (short and intermediate term)	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.3.1 Polidocanol 0.125% foam versus placebo	2	227	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.08, 4.95]
1.3.2 Polidocanol 0.5% foam versus placebo	2	224	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.08, 5.16]
1.3.3 Polidocanol 1% foam versus placebo	3	302	Risk Ratio (M‐H, Random, 95% CI)	5.10 [1.30, 20.01]
1.3.4 Polidocanol 2% foam versus placebo	1	119	Risk Ratio (M‐H, Random, 95% CI)	1.78 [0.17, 19.08]
1.4 Phlebitis or thrombophlebitis rates (intermediate term)	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.4.1 Polidocanol 0.125% foam versus placebo	2	227	Risk Ratio (M‐H, Random, 95% CI)	4.48 [1.02, 19.63]
1.4.2 Polidocanol 0.5% foam versus placebo	2	224	Risk Ratio (M‐H, Random, 95% CI)	3.26 [1.09, 9.75]
1.5 Phlebitis or thrombophlebitis rates (intermediate term)	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.5.1 Polidocanol 1% foam versus placebo	3	302	Risk Ratio (M‐H, Random, 95% CI)	3.12 [1.10, 8.83]
1.5.2 Polidocanol 2% foam versus placebo	1	119	Risk Ratio (M‐H, Random, 95% CI)	3.56 [0.79, 16.05]
1.6 Haemorrhagic complications (short term)	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.6.1 Polidocanol 0.125% foam versus placebo	2	227	Risk Ratio (M‐H, Random, 95% CI)	1.47 [0.42, 5.18]
1.6.2 Polidocanol 0.5% foam versus placebo	2	224	Risk Ratio (M‐H, Random, 95% CI)	2.73 [0.87, 8.53]
1.6.3 Polidocanol 1% foam versus placebo	3	302	Risk Ratio (M‐H, Random, 95% CI)	1.83 [0.75, 4.47]
1.6.4 Polidocanol 2% foam versus placebo	1	119	Risk Ratio (M‐H, Random, 95% CI)	1.33 [0.23, 7.69]
1.7 Neurologic complications (dizziness, TIA) (short term)	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.7.1 Polidocanol 0.125% foam versus placebo	2	227	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.08, 4.95]
1.7.2 Polidocanol 0.5% foam versus placebo	2	224	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.16, 7.36]
1.7.3 Polidocanol 1% foam versus placebo	3	302	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.22, 4.91]
1.7.4 Polidocanol 2% foam versus placebo	1	119	Risk Ratio (M‐H, Random, 95% CI)	0.30 [0.01, 7.14]
1.8 Residual varicose veins (intermediate term)	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.8.1 Polidocanol 0.125% foam versus placebo	2	227	Risk Ratio (M‐H, Random, 95% CI)	0.51 [0.34, 0.78]
1.8.2 Polidocanol 0.5% foam versus placebo	2	224	Risk Ratio (M‐H, Random, 95% CI)	0.28 [0.11, 0.72]
1.8.3 Polidocanol 1% foam versus placebo	2	225	Risk Ratio (M‐H, Random, 95% CI)	0.19 [0.13, 0.29]
1.8.4 Polidocanol 2% foam versus placebo	1	119	Risk Ratio (M‐H, Random, 95% CI)	0.18 [0.11, 0.31]
1.9 Quality of life: VEINES‐QOL score change from baseline (intermediate term)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.9.1 Polidocanol 0.125% foam versus placebo	2	227	Mean Difference (IV, Random, 95% CI)	11.16 [6.17, 16.15]
1.9.2 Polidocanol 0.5% foam versus placebo	2	223	Mean Difference (IV, Random, 95% CI)	14.14 [11.02, 17.25]
1.9.3 Polidocanol 1% foam versus placebo	3	299	Mean Difference (IV, Random, 95% CI)	12.41 [9.56, 15.26]
1.10 Persistence of symptoms: absolute change from baseline score for the VVSymQ (total score) (intermediate term)	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.10.1 Polidocanol 1% foam versus placebo	1	77	Mean Difference (IV, Random, 95% CI)	‐14.00 [‐15.39, ‐12.61]
1.11 Persistence of symptoms: mean change from baseline score for the VCSS (total score) (intermediate term)	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.11.1 Polidocanol 0.125% foam versus placebo	2	227	Mean Difference (IV, Random, 95% CI)	‐2.47 [‐3.12, ‐1.82]
1.11.2 Polidocanol 0.5% foam versus placebo	2	224	Mean Difference (IV, Random, 95% CI)	‐3.35 [‐4.00, ‐2.70]
1.11.3 Polidocanol 1% foam versus placebo	2	223	Mean Difference (IV, Random, 95% CI)	‐3.25 [‐3.90, ‐2.60]

Comparison 2. Foam sclerotherapy versus foam sclerotherapy with different concentrations.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Cosmetic appearance: number of participants with cosmetic improvement (short term) ‐ polidocanol 3% foam versus polidocanol 1% foam	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.2 Cosmetic appearance: number of participants with cosmetic improvement (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2.3 Cosmetic appearance: adjusted mean PA‐V (intermediate term) ‐ polidocanol foam versus polidocanol foam	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
2.3.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	224	Mean Difference (IV, Random, 95% CI)	0.45 [0.03, 0.87]
2.3.2 Polidocanol 0.5% foam versus polidocanol 1% foam	2	221	Mean Difference (IV, Random, 95% CI)	0.06 [‐0.21, 0.32]
2.4 Cosmetic appearance: adjusted mean IPR‐V (intermediate term) ‐ polidocanol foam versus polidocanol foam	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
2.4.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	224	Mean Difference (IV, Random, 95% CI)	0.19 [‐0.01, 0.40]
2.4.2 Polidocanol 0.5% foam versus polidocanol 1% foam	2	221	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.19, 0.12]
2.5 Thromboembolic complications (intermediate term) ‐ polidocanol 3% foam versus polidocanol 1% foam	3	371	Risk Ratio (M‐H, Random, 95% CI)	1.47 [0.41, 5.33]
2.6 Thromboembolic complications (intermediate term) ‐ polidocanol 3% foam versus polidocanol 1% foam (sensitivity analysis)	2	291	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.31, 5.22]
2.7 DVT (short and intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.7.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	226	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.01, 14.53]
2.7.2 Polidocanol 0.125% foam versus polidocanol 2% foam	1	120	Risk Ratio (M‐H, Random, 95% CI)	1.81 [0.17, 19.43]
2.7.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	4.71 [0.59, 37.61]
2.7.4 Polidocanol 0.5% foam versus polidocanol 2% foam	1	114	Risk Ratio (M‐H, Random, 95% CI)	1.62 [0.15, 17.35]
2.7.5 Polidocanol 1% foam versus polidocanol 2% foam	1	115	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.12, 5.66]
2.8 PE (short term) ‐ polidocanol 1% foam versus polidocanol 3% foam	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.9 Skin pigmentation rates (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	1.61 [0.82, 3.17]
2.10 Phlebitis or thrombophlebitis rates (intermediate term) ‐ polidocanol < 1% foam versus polidocanol ≥ 1% foam	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.10.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	226	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.54, 5.23]
2.10.2 Polidocanol 0.125% foam versus polidocanol 2% foam	1	120	Risk Ratio (M‐H, Random, 95% CI)	2.49 [1.17, 5.27]
2.10.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	1.19 [0.55, 2.57]
2.10.4 Polidocanol 0.5% foam versus polidocanol 2% foam	1	114	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.27, 2.22]
2.11 Phlebitis or thrombophlebitis ‐ polidocanol 3% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	1.46 [0.88, 2.41]
2.12 Haemorrhagic complications (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.12.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	226	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.20, 1.83]
2.12.2 Polidocanol 0.125% foam versus polidocanol 2% foam	1	120	Risk Ratio (M‐H, Random, 95% CI)	1.47 [0.34, 6.30]
2.12.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	1.25 [0.53, 2.93]
2.12.4 Polidocanol 0.5% foam versus polidocanol 2% foam	1	114	Risk Ratio (M‐H, Random, 95% CI)	3.29 [0.92, 11.78]
2.12.5 Polidocanol 1% foam versus polidocanol 2% foam	1	115	Risk Ratio (M‐H, Random, 95% CI)	0.41 [0.11, 1.57]
2.13 Neurologic complications (dizziness, TIA) (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.13.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	226	Risk Ratio (M‐H, Random, 95% CI)	0.47 [0.06, 3.57]
2.13.2 Polidocanol 0.5% foam versus polidocanol 1% foam	2	223	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.08, 7.45]
2.13.3 Polidocanol 0.125% foam versus polidocanol 2% foam	1	120	Risk Ratio (M‐H, Random, 95% CI)	3.31 [0.14, 79.67]
2.13.4 Polidocanol 0.5% foam versus polidocanol 2% foam	1	114	Risk Ratio (M‐H, Random, 95% CI)	6.15 [0.30, 125.38]
2.13.5 Polidocanol 1% foam versus polidocanol 2% foam	1	115	Risk Ratio (M‐H, Random, 95% CI)	0.28 [0.01, 6.64]
2.14 Residual varicose veins (intermediate term) ‐ polidocanol foam lower concentration versus polidocanol foam higher concentration	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.14.1 Polidocanol 0.125% foam versus polidocanol 1% foam	2	219	Risk Ratio (M‐H, Random, 95% CI)	2.69 [1.76, 4.09]
2.14.2 Polidocanol 0.125% foam versus polidocanol 2% foam	1	120	Risk Ratio (M‐H, Random, 95% CI)	3.42 [1.92, 6.09]
2.14.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	221	Risk Ratio (M‐H, Random, 95% CI)	1.50 [0.93, 2.43]
2.14.4 Polidocanol 0.5% foam versus polidocanol 2% foam	1	114	Risk Ratio (M‐H, Random, 95% CI)	2.36 [1.26, 4.42]
2.14.5 Polidocanol 1% foam versus polidocanol 2% foam	1	115	Risk Ratio (M‐H, Random, 95% CI)	1.21 [0.57, 2.57]
2.15 Residual varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam	3	371	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.43, 1.04]
2.16 Residual varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam (sensitivity analysis)	2	291	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.35, 1.28]
2.17 Quality of life: adjusted mean change VEINES‐QOL (intermediate term) ‐ polidocanol foam versus polidocanol foam	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
2.17.1 Polidocanol 0.125% foam versus polidocanol 0.5% foam	2	225	Mean Difference (IV, Random, 95% CI)	2.86 [‐0.24, 5.97]
2.17.2 Polidocanol 0.125% foam versus polidocanol 1% foam	2	224	Mean Difference (IV, Random, 95% CI)	1.10 [‐2.50, 4.69]
2.17.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	221	Mean Difference (IV, Random, 95% CI)	‐1.82 [‐4.96, 1.32]
2.18 Persistence of symptoms: adjusted mean change from baseline VCSS score (intermediate term) ‐ polidocanol foam versus polidocanol foam	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
2.18.1 Polidocanol 0.125% foam versus polidocanol 0.5% foam	2	225	Mean Difference (IV, Random, 95% CI)	‐0.87 [‐1.52, ‐0.21]
2.18.2 Polidocanol 0.125% foam versus polidocanol 1% foam	2	224	Mean Difference (IV, Random, 95% CI)	‐0.77 [‐1.42, ‐0.12]
2.18.3 Polidocanol 0.5% foam versus polidocanol 1% foam	2	221	Mean Difference (IV, Random, 95% CI)	0.10 [‐0.56, 0.75]
2.19 Recurrent varicose veins (long term) ‐ polidocanol 3% foam versus polidocanol 1% foam	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

Comparison 3. Foam sclerotherapy versus liquid sclerotherapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Cosmetic appearance: participant satisfaction (long term) ‐ polidocanol foam versus polidocanol liquid	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
3.2 Skin pigmentation (long term) ‐ polidocanol foam versus polidocanol liquid	2	208	Risk Ratio (M‐H, Random, 95% CI)	1.75 [0.81, 3.80]
3.2.1 Polidocanol 3%	1	108	Risk Ratio (M‐H, Random, 95% CI)	1.20 [0.34, 4.24]
3.2.2 Polidocanol liquid versus foam (unknown concentration)	1	100	Risk Ratio (M‐H, Random, 95% CI)	2.20 [0.82, 5.87]
3.3 Residual varicose veins (long term) ‐ polidocanol foam versus polidocanol liquid	5	753	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.33, 0.64]
3.3.1 Polidocanol foam (0.75 to 1.25%) versus liquid (1.25 to 2.5%)	1	150	Risk Ratio (M‐H, Random, 95% CI)	0.14 [0.06, 0.34]
3.3.2 Polidocanol 3%	2	203	Risk Ratio (M‐H, Random, 95% CI)	0.51 [0.41, 0.65]
3.3.3 Polidocanol (unknown concentration)	1	100	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.12, 0.96]
3.3.4 STS 3%	1	300	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.48, 0.82]
3.4 Residual varicose veins (long term) ‐ polidocanol foam versus polidocanol liquid (sensitivity analysis)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
3.4.1 STS 3%	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
3.5 Recurrent varicose veins (long term) ‐ STS 3% liquid versus STS 3% foam	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 4. Sclerotherapy versus sclerotherapy with different substances.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Cosmetic appearance: disappearance of varicose veins (long term) ‐ polidocanol 3% versus STS 1.5%	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
4.2 DVT (short and intermediate term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.3 Skin pigmentation (intermediate term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.4 Haematoma (short term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.5 Skin pigmentation (long term) ‐ STS 1.5% versus polidocanol 3%	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.6 Skin necrosis (long term) ‐ STS 1.5% versus polidocanol 3%	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.7 Residual varicose veins (long term) ‐ Varisolve polidocanol foam versus any market sclerosant (liquid or foam)	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
4.8 Matting (long term) ‐ STS 1.5% versus polidocanol 3%	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 5. Sclerotherapy versus sclerotherapy with different techniques.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Phlebitis or thrombophlebitis rates (short term) ‐ polidocanol 1% foam ‐ antegrade versus retrograde technique	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.2 Skin pigmentation rates (long term) ‐ polidocanol 1% foam ‐ antegrade versus retrograde technique	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.3 Thrombophlebitis rates (short term) ‐ polidocanol 3% foam injected by catheter associated to tumescent technique versus polidocanol 3% foam injected by needle	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.4 Detection of polidocanol foam in deep veins ‐ few injections versus multiple injections	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.5 Residual varicose veins (long term) ‐ visual foam sclerotherapy versus ultrasound guided foam + visual foam sclerotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.6 Residual varicose veins (long term) ‐ multiple/fractionated injections versus few/single injections	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.6.1 STS 3% single versus fractionated injections	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.6.2 Polidocanol 1% or 3% low versus fractionated injections	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.7 Residual varicose veins (intermediate term) ‐ polidocanol 1% foam retrograde versus antegrade technique	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.8 Residual varicose veins (intermediate term) ‐ microcatheter directed foam sclerotherapy versus catheter directed foam sclerotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.9 Residual varicose veins (intermediate term) ‐ needle foam sclerotherapy versus catheter directed foam sclerotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.10 Residual varicose veins (intermediate term) ‐ needle foam sclerotherapy versus microcatheter directed foam sclerotherapy	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.11 Residual varicose veins (short term) ‐ polidocanol 3% foam injected by catheter versus polidocanol 3% foam injected by needle	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.12 Participants not requiring retreatment (long term) ‐ polidocanol 3% foam injected by catheter versus polidocanol 3% foam injected by needle	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
5.13 QoL: AVVQ (long term) ‐ catheter‐directed foam sclerotherapy with tumescence versus ultrasound‐guided foam sclerotherapy	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
5.14 VCSS (long term) ‐ ultrasound guided foam + visual foam sclerotherapy versus visual foam sclerotherapy	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

Comparison 6. Sclerotherapy versus compression conservative treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Cosmetic appearance: good symptomatic improvement and cosmetic result (long term) liquid STS versus compression	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
6.2 Ulcer (short, intermediate and long term) ‐ liquid STS versus compression	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abramowitz 1973.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: hospital‐based study Duration of recruitment to study: not stated Duration of follow‐up: 6 months to 2 years Unit of analysis: participants
Participants	Number: 101 patients Age (mean): not stated Gender: female Inclusion criteria: pregnant, primary or recurrent varicose veins CEAP: not stated Exclusion criteria: patients seen during the last 6 weeks of pregnancy Dropouts: 29 patients at study end; 1 in sclerotherapy group and 28 in GCS group
Interventions	Group 1 (n = 56): sclerotherapy and GCS Group 2 (n = 45): GCS Group 3 (n = 39): compressive sclerotherapy (nonrandom). This group was not included in this review. Sclerosant: STS Dose: 0.5 mL (concentration not stated) Type of compression: "Sorbo rubber pad and elastic crepe bandages" Duration of compression: 5 to 6 weeks Number of sites: not stated Veins condition: empty veins Additional treatment: not stated
Outcomes	Primary outcomes: symptomatic improvement and cosmetic result Secondary outcomes: superficial sloughs
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: did not state the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Unclear risk	Quote: "Response to treatment was gauged on an assessment of two parameters, namely symptomatic improvement and cosmetic result." Pg. 609 Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	Unclear risk	Quote: "Response to treatment was gauged on an assessment of two parameters, namely symptomatic improvement and cosmetic result." Pg. 609 Comment: not stated
Blinding of outcome assessment ‐ cosmetic appearance	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Unclear risk	Comment: not stated
Incomplete outcome data ‐ cosmetic appearance	Unclear risk	Quote: "Three classes of response were considered, as detailed in Table Ill." Quote: "Table III indicates that 28 of the 56 cases treated conservatively were lost to late follow‐up." Comment: there was 50% of loss of follow‐up in the control group. We did not know how it could influence results.
Incomplete outcome data ‐ complications	Unclear risk	Quote: "No serious complications were encountered apart from minimal superficial sloughs in 5 cases in the early stages of this study." Quote: "Table III indicates that 28 of the 56 cases treated conservatively were lost to late follow‐up." Comment: there was 50% of loss of follow‐up in the GCS group. We did not know how it could influence results.
Incomplete outcome data ‐ frequency of persistence of symptoms	Low risk	Quote: "Three classes of response were considered, as detailed in Table Ill." Comment: described pooled results as proposed
Selective reporting (reporting bias)	Unclear risk	Comment: the proposed outcomes were described. There was no protocol; this is not unusual given the year.
Other bias	Unclear risk	Included fewer than 50 participants in one group (45 in the GCS group)

Alos 2006.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: outpatient clinic of a general hospital Duration of recruitment to study: not stated Duration of follow‐up: 365 days Unit of analysis: varicose vein
Participants	Number: 75 patients Age (mean): 59 years (range: 23‐78) Gender: 69 women (92%) and 6 men (8%) Inclusion criteria: "postoperative varices". Patients with primary reticular varices (those of more than 2 mm diameter) or postoperative varices in more than one region that did not involve the saphenous femoral junction. Observation: there was a group of reticular veins treated. CEAP: not stated Exclusion criteria: "patients with truncal varices with junctional (terminal valve) and extrajunctional incompetence, postoperative varices that involved the saphenofemoral junction, post‐thrombotic varices with occluded deep veins, varices secondary to arteriovenous fistulas, bilateral varices of asymmetric calibre, unilateral varices with asymmetric calibre between regions, chronic ischemia of the lower limbs, severe arterial hypertension (blood pressure greater than 180/95 mmHg), and patients being treated with anticoagulants and anti‐inflammatories and/or diuretics for other pathologies to avoid these affecting the appearance or degree of possible secondary effects." Dropouts: "..at 15 days: 1 patient in liquid group. At 30 days 1 patient in foam group and 1 patient in liquid group. At 90 days 4 patients in foam group and 4 patients in liquid group. At 365 days 12 patients in each group"
Interventions	Group 1 (n = 75): sclerotherapy with foam Group 2 (n = 75): liquid sclerotherapy Sclerosant: polidocanol Dose: "0.5 mL of polidocanol in each injection in a volume of 0.5 mL in liquid injection and 2 mL in foam injections. The concentration ranged with vein diameter: for veins with diameter ranging from 3.1 to 4.0 mm in diameter polidocanol was used 1.5% if liquid and 0.75% in foam. For diameters between 4.1 to 5.0 mm polidocanol had 2% if liquid and 1% if foam was done. For diameters between 5.1 to 6.0 mm polidocanol was used in 2.5% if liquid and 1.25% if foam was done." Type of compression: Sorbo rubber pad and elastic crepe bandages Duration of compression: 48 hours with stockings at a pressure of 25‐35 mmHg Number of sites: not stated Vein condition: not stated Additional treatment: heparinoid ointment (3 times a day)
Outcomes	Primary outcomes: efficacy: "according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb. Sclerosis was considered complete when the lumen of the vein was sealed and the vein occluded." Secondary outcomes: During injection: "pain, dizziness, blurred vision, or other complication that could be related to the injection were recorded. During follow‐up pain, skin pigmentation, formation of bulla, skin necrosis were recorded if any had occurred."
Notes	Corresponding author: Dr Jesus Alos Villacrosa, Servicio de Angiologia y Cirugia Vascular, Hospital de Mataro, Crta. Cirera s/n, 08304 Mataro, Barcelona, Spain Conflict of interest: not stated Funding: not stated Contact with author made: (24/11/2016) CEAP classification, blinding of outcome assessment and any published protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Regions were randomly assigned one or other procedure as follows; the right limb (if both legs were involved) or the upper region (if only one leg was involved) were always treated first. The assignment of the first region to be treated to liquid or foam was performed according to a list of 38 random numbers from 0 to 75 created by a specific software. Each patient had an identification number according strict chronological recruitment order, if this number was in the list of random numbers then the patient received first foam and if not received liquid sclerosant first." Pg. 102 Comment: there was no information whether the 38 numbers were visible.
Allocation concealment (selection bias)	Unclear risk	Comment: details not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	High risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." Comment: described as done. Did not describe blinding of the person who injected sclerotherapy agent. Due to the nature of the intervention, probably not done
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." Comment: described as done. Did not describe blinding of the person who injected sclerotherapy agent. Due to the nature of the intervention, probably not done
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." Comment: described as done. Did not describe blinding of the person who injected sclerotherapy agent. Due to the nature of the intervention, probably not done
Blinding of outcome assessment ‐ cosmetic appearance	Low risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." Comment: described as done
Blinding of outcome assessment ‐ complications	Low risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." Comment: described as done
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "Both the patient and the research assistant who assessed the clinical and ultrasound results were blind to the type of treatment applied in each area." "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb." Comment: described as done
Incomplete outcome data ‐ cosmetic appearance	Low risk	Quote: "...at 15 days: 1 patient in liquid group. At 30 days 1 patient in foam group and 1 patient in liquid group. At 90 days 4 patients in foam group and 4 patients in liquid group. At 365 days 12 patients in each group." Comment: proposed outcomes were described and appeared appropriate. Described dropouts up to 20%
Incomplete outcome data ‐ complications	Low risk	Quote: "...at 15 days: 1 patient in liquid group. At 30 days 1 patient in foam group and 1 patient in liquid group. At 90 days 4 patients in foam group and 4 patients in liquid group. At 365 days 12 patients in each group." Comment: proposed outcomes were described and appeared appropriate. Described dropouts up to 20%
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Quote: "...at 15 days: 1 patient in liquid group. At 30 days 1 patient in foam group and 1 patient in liquid group. At 90 days 4 patients in foam group and 4 patients in liquid group. At 365 days 12 patients in each group." Comment: proposed outcomes were described and appeared appropriate. Described dropouts up to 20%
Selective reporting (reporting bias)	Unclear risk	Comment: we did not find a protocol
Other bias	Low risk	Comment: no other source of bias detected

Belcaro 2003b.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: multicentre hospital‐based study Duration of recruitment to study: not stated Duration of follow‐up: 5 years and 10 years Unit of analysis: participants
Participants	Number: 877 patients Age (mean): for sclerotherapy low‐dose STD: 44.3; for sclerotherapy high‐dose STD: 45.5; for sclerotherapy STD + foam J&J: 42.6 mean years of age Gender: for sclerotherapy low‐dose STD: 33 men; for sclerotherapy high‐dose STD: 31 men; for sclerotherapy STD + foam J&J: 31 menu Inclusion criteria: "...patients (age range 25‐65 years) with uncomplicated (no thrombosis/phlebitis, hemorrhage, or skin changes due to prolonged chronic venous insufficiency) primary varicose veins." CEAP: not stated Exclusion criteria: "...pregnancy, obesity, post‐thrombotic occlusion, and history of previous thrombosis, coagulation disorders, any cardiovascular or systemic disease requiring treatment, tumors, bone and joint problems, diabetes, and any possible cause of venous obstruction. Also patients with severe venous insufficiency, lipodermatosclerosis, and ulcerations were excluded. No patient was included in a period of 12 months after pregnancy." Dropouts: in 5 years: sclerotherapy low‐dose STD: 48 patients; sclerotherapy high‐dose STD: 41 patients; sclerotherapy STD + foam J&J: 44 patients; at 10 years: sclerotherapy low‐dose STD: 56 patients; sclerotherapy high‐dose STD: 49 patients; sclerotherapy STD + foam J&J: 51 patients. Also described lost patients: sclerotherapy low‐dose STD: 25 patients; sclerotherapy high‐dose STD: 24 patients; sclerotherapy STD + foam J&J: 21 patients
Interventions	Group 1 (n = 148): sclerotherapy low‐dose STD Group 2 (n = 136): sclerotherapy high‐dose STD Group 3 (n = 150): sclerotherapy STD + foam J&J There were other 3 groups in this study: surgery and ligation; stab‐avulsion; surgery + sclerotherapy Sclerosant: described as STD (possibly sotradecol). Foam produced by using 0.1 to 0.2 mL J&J‐93FA with STD 3% Dose: "Veins larger than 3 mm in diameter were treated with 1‐2 mL injections of 3% sclerosing agent, veins between 2 and 3 mm, with 2% solution. High‐dose sclerotherapy was used with the same indications and procedures of "standard" sclero, but the dosage of injected sclerosing agent was between 3 and 6 mL of 3% sclero‐agent in larger veins (> 3 mm in diameter)." Type of compression: not stated Duration of compression: "Compression was used for 10‐30 days after sclerotherapy on the basis of the dimension of the veins (3 mm in diameter or more: 3 weeks; 2 mm or less: 2 weeks). ..The idea behind high dose was that injecting more sclerosing agent into the vein produced a better displacement of blood and a higher level of vein wall inflammation and could be associated with a less prolonged period of compression (1‐2 weeks for the larger veins)." Number of sites: not stated Vein condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: "variations in AVP, RT" Secondary outcomes: "presence of duplex reflux; number of recurrent or new incompetent venous sites"
Notes	Corresponding author: The VEDICO Study, G Belcaro, Via Vespucci 65, 65100 Pescara, Italy Conflict of interest: Declared no interest conflict: "There is no conflict of interest. The study was organized as an independent trial. No commercial sponsor was included in the study." Funding: Declared that "This independent study was supported only by institutional grants". Contact with author made: (24/11/2016) allocation concealment, blinding of participants and personnel, and blinding of outcome assessors
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomized according to a random code into different groups." Comment: described and seemed to be done
Allocation concealment (selection bias)	High risk	Quote: "The code was opened after intervention had been decided." Comment: open study
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: did not describe the blinding method
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Comment: did not describe the blinding method
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: did not describe assessment blinding
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Comment: did not describe assessment blinding
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Quote: "AVP: At inclusion AVP and RT were comparable in the different groups. At 10 years the decrease in AVP and the increase in refilling time (RT), indicating the decrease in venous reflux, were generally comparable in the different groups (Table III)." Pg. 311 Comment: Described all dropouts and outcomes and seemed to be appropriate
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: "The occurrence of new varicose veins at 5 years varied from 34% for group F (surgery + sclero) and ligation (C) to 44% for the foam + sclero group (E) and 48% for group A (dose 1 sclero). At 10 years the occurrence of new veins varied from 37% in F to 56% in A (Table II)." Pg. 311 Comment: Described all dropouts and outcomes and seemed to be appropriate
Selective reporting (reporting bias)	Low risk	Comment: protocol appeared to have been followed.
Other bias	Low risk	Comment: no other source of bias detected

Blaise 2010.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and tolerance Setting: not stated; a multicentre study Duration of recruitment to study: 1 year Duration of follow‐up: 3 years Unit of analysis: participants Quote: "During the follow‐up period, patients were assessed clinically and by duplex ultrasound imaging at 8 days, 6 weeks, 3 months, 6 months, 1 year, 2 years and 3 years. Further ultrasound‐guided foam sclerotherapy was allowed by the protocol at 6 weeks, 3 and 6 months." Comment: As there was a new intervention at 6 weeks, only the first phase (up to 6 weeks) was considered in this review.
Participants	Number: 143 patients Age (mean): polidocanol 1% (53 years); polidocanol 3% (52 years) Gender: polidocanol 1% (men 18); polidocanol 3% (men 19) Inclusion criteria: "Age between 25 and 75 years. Incompetence of the great saphenous vein on Doppler ultrasound (reflux > 1 s). Classification CEAP C2, C3, C4, C5; maximum diameter of great saphenous vein at the thigh = (supine position) 8 mm. Giving of written informed consent." CEAP: stated as C2, C3, C4 and C5. Did not describe the proportion of participants in each CEAP clinical class Exclusion criteria: "Post‐thrombotic disease. Incompetence of the small saphenous vein or a non saphenous vein (CEAP: A4‐5). Classification CEAP: C0, C1 or C6. Recurrent varicose of the great saphenous vein after stripping. Known thrombophilia or a history of thromboembolic disease. Psychiatric disorders. Known allergy to polidocanol or to one of its constituents. Arterial disease of the lower limbs (IPS < 0.8). Post‐phlebitic disease. Chronic liver disease. Renal failure (creatine > 150 micromol/L). Pregnancy or breastfeeding. Women without effective contraception with risk of pregnancy. Progressing malignancy. Uncontrolled hypertension. Cardiac or respiratory insufficiency. Intolerance to alcohol acquired or induced by a treatment. History of migraine, particularly ophthalmic." Dropouts: 6 weeks polidocanol 1% 1; polidocanol 3% 2. At 3 months, polidocanol 1% 1; polidocanol 3% 2. At 1 year, polidocanol 1% 5; polidocanol 3% 4. At 2 years, polidocanol 1% 7; polidocanol 3% 4. At 3 years, polidocanol 1% 6; polidocanol 3% 6
Interventions	Group 1 (n = 73): polidocanol 1% Group 2 (n = 70): polidocanol 3% Sclerosant: polidocanol Dose: "Sclerosant foam was obtained using a Turbofoam system (Kreussler Pharma, Wiesbaden, Germany). The sclerosant liquid was mixed with sterile air to obtain a liquid‐to‐air ratio of 1:5. The foam was injected immediately after production or within 170 s. The volume of foam to be injected was estimated from the quantity that produced venous spasm, but did not exceed 10 mL." Type of compression: "Compression of 34‐‐49 mmHg was achieved by applying an adhesive foam bandage (Elastomousse, BSNmedical, LeMans, France) and followed by a further layer of Elastoplast (BSNmedical, LeMans, France). Bandages were applied immediately after the treatment to be worn continuously for 3 days. They were then removed and a class 2 elastic compression stocking (OedemaTwin 25, Kreussler Pharma, Wiesbaden, Germany; 23 ‐ 32 mmHg compression) was worn during the day for 15 days." Duration of compression: 15 days Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: "abolition of venous reflux in the GSV at 6 months assessed by colour duplex ultrasonography (absence of saphenous reflux in the thigh lasting more than 1 s)". Secondary outcomes: "clinical evaluation using the VCSS, the quality of life score using CIVIQ both at 6 months and 3 years, and absence reflux in the GSV at 3 years. Tolerance to treatment was assessed from local and systemic side effects, which occurred immediately and after 6 months and 3 years of follow‐up."
Notes	Corresponding author: JL Bosson Conflict of interest: none to declare Funding: "Grant from the French association for vascular medicine, donation of Turbo foam machines for each investigating centre (8) and batches of polidocanol at 1% and 3% without commercial labels by Kreussler Pharma." Contact with author made: (01/12/2016) subgroup data (CEAP C2, C3, and C4) and protocol information. Contact author answered that they had no data from baseline CEAP clinical classes.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The randomisation list was computer centralised, without stratification, equilibrated and with blocks of variable size. Envelopes were distributed to the investigating centres with numbers corresponding to the sclerosant lots." Pg. 780 Comment: seemed to be adequate
Allocation concealment (selection bias)	Unclear risk	Quote: "Envelopes were distributed to the investigating centres with numbers corresponding to the sclerosant lots." Pg. 780 Comment: allocation concealment not described
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "The investigator and patient remained blinded to the treatment that was administered through out the 3‐year study period. In the event of a side effect (including thromboembolic events), there was no need for unblinding as knowledge of the concentration used would in no respect change what is to be done." Pg. 782 Comment: described and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Low risk	Quote: "Randomisation and the management of the stocks for double‐blind use were performed by the Clinical Research Centre in collaboration with Kreussler Pharma who supplied the ampoules of polidocanol without commercial labels." Pg. 780 Comment: seemed to be appropriate
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: considering recanalisation at 3 years: "Randomisation and the management of the stocks for double‐blind use were performed by the Clinical Research Centre in collaboration with Kreussler Pharma who supplied the ampoules of polidocanol without commercial labels." Pg. 780 Comment: seemed to be adequate
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	Low risk	Quote: "Randomisation and the management of the stocks for double‐blind use were performed by the Clinical Research Centre in collaboration with Kreussler Pharma who supplied the ampoules of polidocanol without commercial labels." Pg. 780 Comment: seemed to be adequate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "A critical events committee made up of three non‐investigator physicians classified the adverse events with respect to the treatment received. Every reported event was examined by the committee. Events were classified as expected local side effects or systemic (visual symptoms, etc.) and serious adverse events (deep vein thrombosis, pulmonary embolism, stroke, etc.). Following consultation of the patient’s medical records, the committee decided on the imputability." Pg. 783 Comment: seemed to be adequate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "The investigator and patient remained blinded to the treatment that was administered throughout the 3‐year study period." Comment: seemed to be adequate
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: "The investigator and patient remained blinded to the treatment that was administered throughout the 3‐year study period." Comment: seemed to be adequate
Blinding of outcome assessment ‐ quality of life	Low risk	Quote: "The investigator and patient remained blinded to the treatment that was administered throughout the 3‐year study period." Comment: seemed to be adequate
Incomplete outcome data ‐ complications	Low risk	Quote: "At 3 years, 3.5% were lost to follow‐up." Pg. 783 Comment: described on page 782 and 783 and seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Quote: "At 3 years, 3.5% were lost to follow‐up." Pg. 783 Comment: described on page 783 and seemed to be appropriate
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: "At 3 years, 3.5% were lost to follow‐up." Pg. 783 Comment: described on page 783 and seemed to be appropriate
Incomplete outcome data ‐ quality of life measures	Low risk	Quote: "At 3 years, 3.5% were lost to follow‐up." Pg. 783 Comment: described on page 783 and seemed to be appropriate
Selective reporting (reporting bias)	Low risk	Comment: there was a protocol available and proposed outcomes were described.
Other bias	Low risk	Comment: no other bias detected. Baseline characteristics were described on Pg 782 and appeared to be similar between groups.

Ceulen 2007.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: outpatient clinic Duration of recruitment to study: 6 months Duration of follow‐up: 1 year Unit of analysis: lower leg
Participants	Number: 80 limbs of 80 consecutive patients Age (mean): polidocanol 1% (50.7 years); polidocanol 3% (53.6 years) Gender: male:female (n): polidocanol 1% 6:34; polidocanol 3% 16:24 Inclusion criteria: "Patients with primary incompetence of the GSV and saphenofemoral junction insufficiency with a reflux time 0.5 seconds measured over a traject of 20 cm in the upper leg." CEAP: C1 (n = 10), C2 (n = 52), or C3 to C5 (n = 18). Did not describe a baseline characteristic to discriminate between intervention groups. Did not describe the proportion of CEAP C5 legs Exclusion criteria: "pregnancy, active thrombophlebitis, clotting disturbances or coagulation disorders, a history of deep vein thrombosis, and malignancies in the patient’s medical history." Dropouts: not stated
Interventions	Group 1 (n = 40): polidocanol 1% Group 2 (n = 40): polidocanol 3% Sclerosant: polidocanol Dose: volume injected not stated Type of compression: "local compression pad and an anti‐embolism (10 mmHg) stocking, during day and night for the first week. For the first 6 weeks, patients were advised to wear Class II (30 mmHg) elastic stockings (Mediven Plus, Medi, Bayreuth, Germany), during daytime." Duration of compression: 6 weeks Number of sites: 1 site (single injection) Veins condition: full vein (patient in supine position) Additional treatment: not stated
Outcomes	Primary outcomes: "presence of reflux, measured with duplex imaging. Success of treatment was defined as complete occlusion of the treated vein. The current definition of incompetent saphenous veins is reflux with an extent of 0.5 seconds." Secondary outcomes: "hyperpigmentation, deep vein thrombosis, thrombophlebitis, fibrosis, and persistence of venous complaints. Venous complaints are defined as presence of at least one of the following complaints: cramps, pain, and restless legs."
Notes	Corresponding author: RPM Ceulen, MD, Department of Dermatology, University Hospital Maastricht, PO Box 5800, 6202 AZ Maastricht, The Netherlands Conflict of interest: none Funding: not stated Contact with author made: (01/12/2016) allocation concealment, blinding of participants and personnel, blinding of outcome assessors, published protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random study group assignment was based on a computer‐generated list." Pg. 277 Comment: adequately described
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Unclear risk	Quote: "Both sclerosing foams were prepared with the double‐syringe system technique applying a sclerosant air ratio of 1:5." Comment: did not state the blinding method (e.g. no labels)
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ cosmetic appearance	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Unclear risk	Comment: not stated
Incomplete outcome data ‐ cosmetic appearance	Unclear risk	Comment: outcomes were described on pages 278 and 279 and seemed to be appropriate. Did not describe the dropouts. We considered no dropouts, but we did not know if it could be some factor affecting results.
Incomplete outcome data ‐ complications	Unclear risk	Comment: outcomes were described on pages 278 and 279 and seemed to be appropriate. Did not describe the dropouts. We considered no dropouts, but we did not know if it could be some factor affecting results.
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: outcomes were described on pages 278 and 279 and seemed to be appropriate Did not describe the dropouts. We considered no dropouts, but we did not know if it could be some factor affecting results.
Incomplete outcome data ‐ frequency of persistence of symptoms	High risk	Comment: did not describe persistence of venous complaints
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol described
Other bias	Unclear risk	There were some imbalances at baseline between the groups related to diameter of the veins and number of men/women. Sample size fewer than 50 in both groups

Chleir 1997.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: multicentre, hospital‐based study Duration of recruitment to study: 6 months Duration of follow‐up: 5 years Unit of analysis: lower leg
Participants	Number: 223 patients (226 internal saphenous vein) Age (mean): 18 to 81 years (57.36 years) Gender: women/men: 170/53 Inclusion criteria: any patient with "single trunk saphenous reflux the calibre of which lies between 4 and 8 mm to 3 cm from the sapheno‐femoral junction with a reflux greater than 2 seconds, no preliminary treatment having been administered to this internal saphenous vein." CEAP: not stated Exclusion criteria: STS "known allergy, atopy, pregnancy, age inferior to 18 years, previous deep venous thrombosis, arteriovenous malformation." Dropouts: not stated
Interventions	Group 1 (n = 73): single dose Group 2 (n = 77): fractionated dose Sclerosant: STS Dose: "the number of sessions required to achieve effective sclerosis was 2 on average, with a total dose of 3% STS 3 ml and a mean of 1.5 ml per session" Type of compression: not stated Duration of compression: not stated Number of sites: one versus two sites Veins condition: not reported Additional treatment: not reported
Outcomes	Primary outcomes: "the patient who will appreciate leg heaviness: present or absent; pulsed Doppler at the injected area with criteria: Doppler: present or absent reflux, ultrasound: free vein or sclerotic." Secondary outcomes: "pigmentation, inflammation, sores, blood or neurological injury"
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "once the patient included, was assigned a numbered file 1 to 10 in the order of inclusion of patients. This file we have previously determined by random choice, the letter "U" for a single injection or an "F" for fractional injection." Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Low risk	Comment: described proposed outcomes in pages 278 and 279
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described proposed outcomes in pages 278 and 279
Incomplete outcome data ‐ frequency of persistence of symptoms	High risk	Comment: not stated
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol described
Other bias	Unclear risk	Comment: did not describe balance in baseline characteristics

Demagny 2002.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: hospital‐based Duration of recruitment to study: 26 months Duration of follow‐up: 30, 60 and 180 days Unit of analysis: the vein
Participants	Number: 254 patients. 400 incompetent saphenous veins (300 great saphenous veins and 100 short saphenous veins) Age (mean): 18 to 80 years (52.2 years) Gender: 210 women and 44 men Inclusion criteria: long or short saphenous incompetent veins, less than 9 mm in diameter measured 30 mm distant from saphenous‐femoral or saphenous‐popliteal junction CEAP: not stated Exclusion criteria: "patients with anterior or posterior collateral flowing into the last 10 cm of the trunk before saphenofemoral junction were excluded from the study." Dropouts: not stated
Interventions	Group 1 (n = 200): liquid form sclerosant Group 2 (n = 200): foam in echo‐guided sclerosant Sclerosant: STS Dose: to great saphenous veins ‐ liquid form: 2 cc nondiluted STS; foam form: 2 cc foam 3% STS. To short saphenous veins ‐ liquid form: 2 cc STS 1.5% or 3% in second section; foam form: 2 cc foam STS 1.5% or 3% "we used disposable plastic syringes of 2.5 cc. The mousse was obtained by mixing 0.5 cc of STS variable concentration (in our study 1.5 to 3%). Two methods are possible and allow the production of a more stable foam during 30 minutes: ‐ Exerting a series of pulls on the plunger of a syringe after its end closed by a sterile hard plastic tip; ‐ By a two‐sided coupling connecting two syringes of 2.5 cc. An alternating pressure on each plunger of the syringe mixing the volume with the other, transforming it into a compact foam." Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: full veins: "seated position" Additional treatment: "In all the cases, the patient is lengthened for 10 minutes to increase the efficiency of the act of multiple and avoid lipothymias". This means that participants were kept lying down for 10 minutes
Outcomes	Primary outcomes: presence or absence of reflux at duplex‐scan Secondary outcomes: collateral effects
Notes	Corresponding author: A Demagny. 29, rue Lord Kitchener 76600 Le Harve Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Unclear risk	Comment: the outcome complications were described on page 135 as allergic reaction (one case), and five cases of amaurosis fugax, and seemed to be appropriate. Did not describe dropouts
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: described on page 135 as success and seemed to be appropriate (we considered the non success as residual varicose veins) Did not describe dropouts
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Comment: the vein calibre did not exceed 9 mm. Did not describe baseline characteristics and we did not know to which extent this could impact results

Gibson 2010.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: safety/efficacy Setting: not stated, multicentre Duration of recruitment to study: 12 months Duration of follow‐up: four, 8 weeks and 12 weeks Unit of analysis: participants
Participants	Number: 77 patients Age (mean): polidocanol (45.3 years); placebo (44.9 years) Gender: (female/male): polidocanol (30/9); placebo (32/6) Inclusion criteria: "Males and females 18‐65 years old; VEINES Sym Score less than 75 points; Varicose Vein clinical classification CEAP 2, 3, 4, or 5; Incompetence of SFJ associated with incompetence of the GSV or other major accessory vein; superficial venous disease manifested by both symptoms and visible varicosities; ability to comprehend and sign an informed consent document and completed study questionnaires in English" CEAP: polidocanol C3: 35, C4: 2 and C5: 2. This implied 5.128% of C5 CEAP in the group. For placebo: C3: 34, C4: 4 and C5: 0 Exclusion criteria: "Incompetence of the SSV which substantially contributes to the filling of visible varicose veins; Current or previous DVT; leg obesity; PAD in the leg to be treated; reduced mobility; planned prolonged travel with limited mobility with in 4 weeks of treatment; history of PE or stroke; major surgery, prolonged hospitalisation or pregnancy within 3 months; current anticoagulation therapy (within 7 days of enrolment); participation in a clinical study involving a investigational product within 3 months; major co‐existing disease or clinically significant laboratory abnormalities; known allergic response to polidocanol or heparin or severe and/or multiple allergic reactions; women of childbearing potential not using effective contraception one month prior to enrolment and/or unwilling to continue while on study; pregnant or lactating women; current alcohol or drug abuse" Dropouts: no dropouts
Interventions	Group 1 (n = 39): polidocanol 1% endovenous microfoam ablation Group 2 (n = 38): placebo (agitated saline) Sclerosant: Varisolve (Polidocanol Endovenous Microfoam) Dose: "Varisolve (Polidocanol Endovenous Microfoam) 1% polidocanol, up to 15 mL, one treatment session (initially up to 30 mL, reduced to up to 15 mL in Amendment #2) compared to drug: agitated saline 10 u/mL normal heparinized saline solution, up to 20 mL, one treatment session." Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: "The Absolute Change From Baseline Score for the VVSymQ (total Score) at 8 weeks (time frame: baseline to 8 weeks)" "CIVIQ 2 and VCSS were completed before, and then eight and 12 weeks after treatment." Secondary outcomes: "elimination of reflux or occlusion of the great saphenous vein"
Notes	Corresponding author: David Wright MBBS. Organization: BTG International Inc. Conflict of interest: BTG International Inc. Funding: BTG International Inc. Contact with author made: (27/09/2016): we asked if Gibson 2010 and Anon 2008a could be the same study. The answer was yes. It was also clarified that "Microfoam sclerosant, (Polidocanol Endovenous Microfoam, Varisolve ™ Varithena®) Catheter directed, ultrasound guided treatment for major (trunk) vein incompetence [and] should not be confused with injection sclerotherapy." ‐ Nunzia Sposito (personal communication with Dr Wright) (12/01/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Quote: "Patients were blinded to treatment assignment." Pg. 309 Comment: impossible to blind personnel. Did not describe the blinding method
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	High risk	Quote: "Patients were blinded to treatment assignment." Pg. 309 Comment: impossible to blind personnel. Did not describe the blinding method
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	High risk	Quote: "Patients were blinded to treatment assignment." Pg. 309 Comment: impossible to blind personnel. Did not describe the blinding method
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ quality of life	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Unclear risk	Comment: not stated
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described dropouts and causes
Incomplete outcome data ‐ quality of life measures	Low risk	Comment: described dropouts and causes
Incomplete outcome data ‐ frequency of persistence of symptoms	Low risk	Comment: described dropouts and causes
Selective reporting (reporting bias)	Low risk	Comment: there was a protocol available and the study paper described the proposed outcomes.
Other bias	Unclear risk	Comment: sample size less than 50 in each group. Did not describe baseline characteristics

Goldman 2002.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: safety and efficacy Setting: clinic‐based study Duration of recruitment to study: not stated Duration of follow‐up: 1, 4 and 16 weeks Unit of analysis: participants
Participants	Number: 46 patients. 129 patients were randomised (42 with less than 1 mm veins, 41 with veins 1‐3 mm in diameter and 46 with 3‐6 mm in diameter) Age (mean): not stated Gender: not stated Inclusion criteria: "telangiectasia < 1 mm, reticular veins 1 to 3 mm diameter or varicose veins 3 to 6 mm diameter." CEAP: not stated Exclusion criteria: "each patient" 'leg veins that did not have incompetence from the saphenofemoral or saphenopopliteal junctions..." Dropouts: none reported
Interventions	Group 1 (n = 27): STS Group 2 (n = 27): polidocanol Sclerosant: STS 1.5% or polidocanol 3% Dose: did not state the volume Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: "Photographic score: appearance of veins (range from 1 (worse) to 5 (complete disappearance)." Secondary outcomes: "Complications: skin necrosis, hyperpigmentation, matting, local urticaria."
Notes	Corresponding author: Mitchel P Goldman, MD, Dermatology/Cosmetic Laser Associates of San Diego, Inc., 850 Prospect St., La Jolla, CA 92037 Conflict of interest: declared no significant interest conflict with commercial supporters Funding: not stated Contact with author made: (12/01/2016): data from CEAP C2, C3, and C4; random sequence generation; allocation concealment; blinding of participants and personnel; blinding of outcome assessment; any published protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: dId not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Low risk	Quote: "the treating physician was blinded as to the agent being injected". Pg. 2 Comment: probably done
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "the treating physician was blinded as to the agent being injected". Pg. 2 Comment: probably done
Blinding of outcome assessment ‐ cosmetic appearance	Low risk	Quote: "Three vascular surgeons blinded to treatment and study center evaluated pre‐ and posttreatment photographs to determine overall disappearance on a scale of 1‐5". Pg. 4 Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "the treating physician was blinded as to the agent being injected". Pg. 2 Comment: probably done
Incomplete outcome data ‐ cosmetic appearance	Low risk	Comment: cosmetic appearance described as "disappearance of varicosities" on page 54 and seemed to be appropriate
Incomplete outcome data ‐ complications	Unclear risk	Comment: described on page 54. Did not discriminate between subgroups of complications and there was different prognosis for these outcomes (e.g.; DVT, pigmentation)
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol provided
Other bias	Unclear risk	Comment: did not describe the baseline characteristics or balance between study groups Sample size less than 50 in each group

Hamel‐Desnos 2007.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: multicentre clinical‐based study Duration of recruitment to study: 4 months Duration of follow‐up: 3 months, 6 months, 1 year, 18 months and 2 years Unit of analysis: participants
Participants	Number: 148 patients Age (mean): polidocanol 1%: (56 years); polidocanol 3%: (53 years) Gender: women:men n (%): polidocanol 1% 58:16 (78 %:22 %); polidocanol 3%: 59:15 (80%:20%) Inclusion criteria: "Patient of either sex with an incompetent great saphenous vein in the thigh (reflux > 1 second) measuring 4 to 8 mm in diameter in the standing position." CEAP classifications authorised: C2 to C6 CEAP: referred as CEAP C2‐C6. Did not stratify the number of participants in each CEAP clinical class Exclusion criteria: "Patient with a contraindication to sclerotherapy in general and/or to polidocanol in particular" Dropouts: not stated
Interventions	Group 1 (n = 74): polidocanol 1% foam Group 2 (n = 74): polidocanol 3% foam Sclerosant: polidocanol Dose: "....a sclerosant liquid‐air mixture of 1 + 4." "The first injection (2.5 ml foam) was given at the junction between the upper and middle‐third of the thigh. If necessary, up to 2 more injections were allowed into the GSV in the thigh (total foam volume not exceeding 7.5 ml)." Type of compression: "Compression bandages or stockings were not applied to the limb following treatment, in keeping with our usual practice for saphenous trunks. Elastic compression, analgesic or anti‐inflammatory drugs were only prescribed if secondary inflammation or painful reactions occurred." Duration of compression: not used Number of sites: not stated Veins condition: full veins: "Patients lay supine during treatment." Additional treatment: "Additional or alternative treatment was offered to patients in whom the single treatment session failed to obliterate the GSV."
Outcomes	Primary outcomes: abolition of venous reflux at 2 years Secondary outcomes: "...length of occluded vein assessed by duplex ultrasonography. This was only measured at the 3 week assessment. The local and systemic complication rates were also recorded as part of the study, including visual disturbance and chest symptoms."
Notes	Corresponding author: Dr C Hamel‐Desnos, MD, Centre Hospitalier Prive Saint Martin, 18 rue des Roquemonts, 14050 Caen, France Conflict of interest: "...Kreussler Pharma for their support of this study in providing the study drugs." Funding: "sponsored by the Société Française de Phlébologie" Contact with author made: (12/08/2016) random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, any published protocol, data from C2, C3 and C4. (31/12/2016) Dr Claudine suggested asking information from Professor Allaert after 15/01/2017. Contact made (16/01/2017)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "In each centre, a randomization list was created by the statistician and a series of numbered ampoules of Aethoxysklerol (concentration not indicated) was provided by Kreussler Laboratories (Photo 1). Each treatment was assigned to the patients according to the randomisation list." Pg. 724 Comment: seemed to be adequate
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "...and a series of numbered ampoules of Aethoxysklerol (concentration not indicated) was provided by Kreussler Laboratories (Photo 1). Each treatment was assigned to the patients according to the randomisation list. Statistical analysis was conducted on blinded data and the unblinding was done after writing of the statistical report." Pg. 724 Comment: done and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Low risk	Quote: "...and a series of numbered ampoules of Aethoxysklerol (concentration not indicated) was provided by Kreussler Laboratories (Photo 1). Each treatment was assigned to the patients according to the randomisation list. Statistical analysis was conducted on blinded data and the unblinding was done after writing of the statistical report." Pg. 724 Comment: done and seemed to be appropriate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "Each treatment was assigned to the patients according to the randomisation list. Statistical analysis was conducted on blinded data and the unblinding was done after writing of the statistical report." Pg. 724 Comment: done and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "Each treatment was assigned to the patients according to the randomisation list. Statistical analysis was conducted on blinded data and the unblinding was done after writing of the statistical report." Pg. 724 Comment: done and seemed to be appropriate
Incomplete outcome data ‐ complications	Unclear risk	Comment: described as side effects including chest symptoms, cough, neurologic complications, malaise, pain, allergy, and flushing on page 727 and seemed to be appropriate. Did not describe dropouts. Did not describe DVT outcomes
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: described as elimination of reflux and disappearance of the great saphenous vein on page 727 and seemed to be appropriate. Did not describe dropouts. Did not describe DVT outcomes. Described GSV diameter between 4 and 8 mm (Pg. 724)
Selective reporting (reporting bias)	Low risk	Comment: there was a published protocol that seemed to describe the proposed outcomes.
Other bias	Low risk	Comment: described baseline characteristics as similar on Pg. 726. There was no other source of bias.

Kahle 2004.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: safety and efficacy Setting: hospital‐based study Duration of recruitment to study: not stated Duration of follow‐up: 4 weeks Unit of analysis: participants
Participants	Number: 25 patients Age (mean): not stated Gender: not stated Inclusion criteria: superficial varicose veins of 3 to 6 mm diameter CEAP: not stated. Exclusion criteria: "...hypersensitivity to polidocanol (generalized urticaria, i.e., immediate‐type hypersensitivity, or eczema, i.e., delayed‐type hypersensitivity), pregnancy, acute thrombosis/phlebitis, thrombophilia, and peripheral arterial occlusive disease (ankle‐brachial index<0.9)" Dropouts: none
Interventions	Group 1 (n = 14): polidocanol Group 2 (n = 11): saline Sclerosant: polidocanol 2 or 3% Dose: "injection sclerotherapy using 2 or 3% polidocanol without preservatives (Aethoxysklerol, Chemische Fabrik Kreussler, Wiesbaden, Germany) adapted to the caliber of the varicose veins. Veins of 3 to 4 mm in diameter were treated with 2% sclerosing liquid and in those of 5‐ to 6‐mm sclerosants in a concentration of 3% were injected." Type of compression: "After the procedure, external adhesive compression bandages were applied for 2 to 3 days (2 days in case 2% liquid was used and 3 days if a patient received 3% liquid). Moreover, the patients were treated with knee‐length compression stockings for 1 week after the sclerotherapy." Duration of compression: 1 week Number of sites: not stated Veins condition: not stated Additional treatment: no concomitant treatment
Outcomes	Primary outcomes: vein obliteration Secondary outcomes: "haemodynamic improvement measured by venoarterial flow index."
Notes	Corresponding author: Birgit Kahle, Department of Dermatology, University of Heidelberg, D‐69117 Heidelberg, Germany Conflict of interest: not stated Funding: not stated Contact with author made: (12/01/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of outcome assessment, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not stated
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "The procedure was double‐blinded: neither the patients nor the physician performing the procedure knew whether polidocanol (Aethoxysklerol) or placebo was injected." Pg. 724 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Low risk	Quote: "The procedure was double‐blinded: neither the patients nor the physician performing the procedure knew whether polidocanol (Aethoxysklerol) or placebo was injected." Pg. 724 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Low risk	Comment: described the outcomes on pages 725 and 726 and seemed to be appropriate. Described no dropouts
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described the outcomes on pages 725 and 726 and seemed to be appropriate. Described no dropouts
Selective reporting (reporting bias)	Unclear risk	Comment: we had no access to protocol.
Other bias	Unclear risk	Comment: no other source of bias detected. Did not describe baseline characteristics. Sample size less than 50 in each group

King 2015.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: safety/efficacy Setting: clinical‐based study Duration of recruitment to study: not stated Duration of follow‐up: 1 week, 8 week Unit of analysis: participants
Participants	Number: 279 patients Age (mean): placebo: 46 years; polidocanol 0.125%: 51.6 years; polidocanol 0.5%: 48.2 years; polidocanol 1%: 48.8 years; polidocanol 2%: 49.7 years Gender: female n (%): placebo: 44 (78.6); polidocanol 0.125%: 42 (73.7); polidocanol 0.5%: 37 (72.5); polidocanol 1%: 38 (73.1); polidocanol 2%: 47 (74.6) Inclusion criteria: "Ages Eligible for Study: 18 years to 75 years (adult, senior). Genders eligible for study: both. Incompetence of SFJ associated with incompetence of the GSV or other major accessory vein. Ability to comprehend and sign an informed consent document and complete study questionnaires in English. Ability to record symptoms in accordance with the protocol. Symptomatic varicose veins. Visible varicose veins" CEAP: placebo: C2: 22; C3: 24; C4: 10; C5‐C6: 0; polidocanol 0.125%: C2: 32; C3: 14; C4: 11; C5‐C6: 0; polidocanol 0.5%: C2: 25; C3: 13; C4: 10; C5‐C6: 3. This implied 5.882% of C5 to C6 included in clinical classes. Polidocanol 1%: C2: 26; C3: 15; C4: 9; C5‐C6: 2. This implied 3.773% of C5 to C6 clinical classes. Polidocanol 2%: C2: 32; C3: 13; C4: 17; C5‐C6: 1. This accounted for 1.587% of C5 to C6 included participants in this group. Exclusion criteria: "Patients who only have telangiectatic or reticular veins (Clinical Finding C1), as assessed by CEAP. Leg obesity impairing the ability to access the vein to be treated and/or to apply post‐procedure compression bandaging and stockings. Ultrasonographic or other evidence of current or previous deep vein thrombosis or occlusion. Deep vein reflux unless clinically insignificant in comparison to superficial reflux. Peripheral arterial disease precluding the wearing of post‐procedure compression bandaging and stockings. Reduced mobility. Major surgery, prolonged hospitalization or pregnancy within 3 months of screening. Major co‐existing disease (e.g. malignancy; pulmonary disease; renal or hepatic insufficiency; serious skin disease/condition that may compromise the ability of the patient to comply with the compression protocol, etc.). Known allergic response to polidocanol or heparin, including history of heparin‐induced thrombocytopenia, and/or multiple allergic reactions. Current alcohol or drug abuse. Pregnant or lactating women. Women of childbearing potential not using effective contraception. History of DVT, pulmonary embolism or stroke." "small saphenous and deep vein incompetence, history of or active deep vein thrombosis (DVT), pulmonary embolism (PE), or stroke; inability to comply with post‐treatment compression or walk unaided. There were no exclusions for medications, including anticoagulants. Importantly, there were no restrictions on vein diameter or tortuosity of veins to be treated." Dropouts: placebo: 1; polidocanol 0.125%: 1; polidocanol 0.5%: 0; polidocanol 1%: 2; polidocanol 2%: 0 Of the 284 participants randomised, five withdrew before treatment.
Interventions	Group 1 (n = 57): polidocanol 0.125% Group 2 (n = 51): polidocanol 0.5% Group 3 (n = 52): polidocanol 1% Group 4 (n = 63): polidocanol 2% Group 5 (n = 56): placebo comparator Sclerosant: polidocanol Dose: maximum volume of 15 mL of study drug (in 5 mL aliquots) was allowed regardless of treatment assignment. Type of compression: "The treated leg was wrapped in a short‐stretch bandage with compression pads that provided eccentric pressure over the treated venous segments. An overstocking and thigh‐length 30‐40 mmHg compression stocking with waist band (Venosan North America, Asheboro, NC, USA) were placed over the dressing. The compression bandages and stocking were worn continuously for 48 hours. The compression stocking alone was worn continuously for an additional 12 days." Duration of compression: 12 days Number of sites: not stated Veins condition: not stated Additional treatment: "Patients were mobilized when treatment was complete and were encouraged to walk for at least 5 minutes during each waking hour for the week following treatment."
Outcomes	Primary outcomes: change in VVSymQ Score Secondary outcomes: change from baseline to 8 weeks in Appearance as Rated by Patient (PA‐V3). Change from baseline at 8 weeks post‐treatment in IPR‐V3 Score: Physician Photographic Review of Appearance
Notes	Corresponding author: BTG International Inc., Five Tower Bridge, Suite 800, 300 Barr Harbor Drive, West Conshohoken, PA 19428, USA Conflict of interest: BTG International Inc. Funding: BTG International Inc. Contact with author made: (12/01/2016) CEAP C2, C3 and C4, allocation concealment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were equally randomized to PEM 0.125% (control), 0.5%, 1%, 2%, or placebo. Randomization was stratified by baseline symptom score and site using an automated interactive voice recognition system (IVRS, United States, Biosource Corporation, San Francisco, CA, USA)." Pg. 3 Comment: randomisation described and seemed to be appropriate
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Low risk	Quote: "All PEM dose concentrations were in identical canisters identified only by an individual numeric code assigned by the IVRS. Placebo (agitated diluent solution, 5 mL aliquots) was prepared immediately prior to injection." Pg. 3 and 4 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs placebo)	High risk	Quote: "It was not possible to create placebo foam that is indistinguishable from PEM; thus physician and ultrasonographer were unblinded to placebo. The PEM 0.125% arm was added, 0.125% being the lowest concentration of polidocanol that can generate foam, which was presumed to be sub‐therapeutic, to allow blinded evaluation of duplex response results. Placebo solution contained the excipients of PEM, and was visible on ultrasound when agitated using the Tessari method after administration." Pg. 4 Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "All PEM dose concentrations were in identical canisters identified only by an individual numeric code assigned by the IVRS. Placebo (agitated diluent solution, 5 mL aliquots) was prepared immediately prior to injection." Pg. 3 and 4 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ complications (treatment vs placebo)	High risk	Quote: "It was not possible to create placebo foam that is indistinguishable from PEM; thus physician and ultrasonographer were unblinded to placebo. The PEM 0.125% arm was added, 0.125% being the lowest concentration of polidocanol that can generate foam, which was presumed to be sub‐therapeutic, to allow blinded evaluation of duplex response results. Placebo solution contained the excipients of PEM, and was visible on ultrasound when agitated using the Tessari method after administration." Pg. 4 Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Low risk	Quote: "All PEM dose concentrations were in identical canisters identified only by an individual numeric code assigned by the IVRS. Placebo (agitated diluent solution, 5 mL aliquots) was prepared immediately prior to injection." Pg. 3 and 4 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs placebo)	High risk	Quote: "It was not possible to create placebo foam that is indistinguishable from PEM; thus physician and ultrasonographer were unblinded to placebo. The PEM 0.125% arm was added, 0.125% being the lowest concentration of polidocanol that can generate foam, which was presumed to be sub‐therapeutic, to allow blinded evaluation of duplex response results. Placebo solution contained the excipients of PEM, and was visible on ultrasound when agitated using the Tessari method after administration." Pg. 4 Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	Low risk	Quote: "All PEM dose concentrations were in identical canisters identified only by an individual numeric code assigned by the IVRS. Placebo (agitated diluent solution, 5 mL aliquots) was prepared immediately prior to injection." Pg. 3 and 4 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ quality of life (treatment vs placebo)	High risk	Quote: "It was not possible to create placebo foam that is indistinguishable from PEM; thus physician and ultrasonographer were unblinded to placebo. The PEM 0.125% arm was added, 0.125% being the lowest concentration of polidocanol that can generate foam, which was presumed to be sub‐therapeutic, to allow blinded evaluation of duplex response results. Placebo solution contained the excipients of PEM, and was visible on ultrasound when agitated using the Tessari method after administration." Page 4 Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	Low risk	Quote: "All PEM dose concentrations were in identical canisters identified only by an individual numeric code assigned by the IVRS. Placebo (agitated diluent solution, 5 mL aliquots) was prepared immediately prior to injection." Pg. 3 and 4 Comment: blinding of participants and personnel described and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs placebo)	High risk	Quote: "It was not possible to create placebo foam that is indistinguishable from PEM; thus physician and ultrasonographer were unblinded to placebo. The PEM 0.125% arm was added, 0.125% being the lowest concentration of polidocanol that can generate foam, which was presumed to be sub‐therapeutic, to allow blinded evaluation of duplex response results. Placebo solution contained the excipients of PEM, and was visible on ultrasound when agitated using the Tessari method after administration." Pg. 4 Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of outcome assessment ‐ cosmetic appearance	Low risk	Quote: "Importantly, assessors for all primary and secondary endpoints were completely blinded." Comment: described blinding of assessors and seemed to be appropriate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "Importantly, assessors for all primary and secondary endpoints were completely blinded." Comment: described blinding of assessors and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "Importantly, assessors for all primary and secondary endpoints were completely blinded." Comment: described blinding of assessors and seemed to be appropriate
Blinding of outcome assessment ‐ quality of life	Low risk	Quote: "Importantly, assessors for all primary and secondary endpoints were completely blinded." Comment: described blinding of assessors and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Low risk	Quote: "Importantly, assessors for all primary and secondary endpoints were completely blinded." Comment: described blinding of assessors and seemed to be appropriate
Incomplete outcome data ‐ cosmetic appearance	Low risk	Comment: described the proposed outcomes and seemed to be appropriate. Baseline characteristics similar between groups
Incomplete outcome data ‐ complications	Low risk	Comment: described the proposed outcomes and seemed to be appropriate. Described dropouts and seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described the proposed outcomes and seemed to be appropriate. Described dropouts and seemed to be appropriate
Incomplete outcome data ‐ quality of life measures	Low risk	Comment: described the proposed outcomes and seemed to be appropriate. Described dropouts and seemed to be appropriate
Incomplete outcome data ‐ frequency of persistence of symptoms	Low risk	Comment: described the proposed outcomes and seemed to be appropriate. Described dropouts and seemed to be appropriate
Selective reporting (reporting bias)	Low risk	Comment: there was a published protocol, describing outcomes of the article that seemed to be appropriate.
Other bias	Low risk	Comment: no other source of bias identified. Baseline characteristics similar between groups

Martimbeau 2003a.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 12 months Unit of analysis: participants
Participants	Number: 200 patients Age (mean): not stated Gender: not stated Inclusion criteria: not stated CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = 100): STS foam Group 2 (n = 100): STS liquid Sclerosant: STS 1% Dose: 2 mL injection "at the point of reflux at first visit, 4 cm distally at second visit, and 4 cm distally at third visit." Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: not stated Secondary outcomes: not stated Authors did not state outcomes to be studied. They described in results outcomes related to complications, residual varicose veins, and recurrent varicose veins.
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Subjects were randomly assigned..." Comment: did not describe randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Quote: "The rate of adverse events was 0.2% for group F and 2.4% for group L". Commnent: did not describe blinding of participants and personnel
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Quote: "Group F required 1.3 sessions and group L required 2.4 sessions". Commnent: did not describe blinding of participants and personnel
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Quote: "The recurrence rate for group F was 1.2% and for group L was 1.7%". Commnent: did not describe blinding of participants and personnel
Blinding of outcome assessment ‐ complications	Unclear risk	Quote: "The rate of adverse events was 0.2% for group F and 2.4% for group L". Commnent: did not describe blinding of outcome assessment
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Quote: "Group F required 1.3 sessions and group L required 2.4 sessions". Commnent: did not describe blinding of outcome assessment
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Quote: "The recurrence rate for group F was 1.2% and for group L was 1.7%". Comment: did not describe blinding of outcome assessment
Incomplete outcome data ‐ complications	Unclear risk	Comment: described data in pooled percentages that limited our interpretations. Did not discriminate between complications. Did not describe dropouts
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: described data in pooled percentages that limited our interpretations. Did not describe dropouts
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Comment: described data in pooled percentages that limited our interpretations. Did not describe dropouts
Selective reporting (reporting bias)	Unclear risk	Comment: we had no information on protocols (there was an abstract with no specific information on a protocol).
Other bias	Unclear risk	Comment: we had no information on the baseline characteristics or vein diameter to judge.

Martimbeau 2003b.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 2 years Unit of analysis: participants
Participants	Number: 144 patients Age (mean): not stated Gender: not stated Inclusion criteria: not stated CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = 72): foam perfluoropopane‐filled albumin microspheres of STS Group 2 (n = 72): foam air filled STS Sclerosant: STS 1% Dose: 5 mL injection Type of compression: "Postinjection compression with ultrasonic probe was performed". Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: not stated Secondary outcomes: not stated Described five parameters as outcomes but listed six: "patient discomfort, phlebitis, veins induration, incompressibility, diameter, and length of sclerosing effect..."
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were randomly assigned..." Comment: did not describe randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: did not describe blinding of participants and personnel
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: did not describe blinding of participants and personnel
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Comment: did not describe blinding of participants and personnel
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: did not describe blinding of outcome assessment
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: did not describe blinding of outcome assessment
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Comment: did not describe blinding of outcome assessment
Incomplete outcome data ‐ complications	Unclear risk	Quote: "At postsclerotherapy week 2 Group A showed significantly less patient discomfort and higher vein induration, incompressibility, and sclerosing effect than Group B (P = 0.003, P = 0.04, P = 0.02, P = 0.001, respectively). At year 2, Group A continued to show higher reduction in diameter of sclerosing effect (P = 0.02)." Comment: described statistical significance, but there were no data to determine the number of participants developing each outcome. So, we did not know if there were a low or high number of events, and if the effect occurrence was due to chance.
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Quote: "At postsclerotherapy week 2 Group A showed significantly less patient discomfort and higher vein induration, incompressibility, and sclerosing effect than Group B (P = 0.003, P = 0.04, P = 0.02, P = 0.001, respectively). At year 2, Group A continued to show higher reduction in diameter of sclerosing effect (P = 0.02)." Comment: described statistical significance, but there were no data to determine the number of participants developing each outcome. So, we did not know if there was a low or high number of events, and if the effect occurrence was due to chance.
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Unclear risk	Quote: "At postsclerotherapy week 2 Group A showed significantly less patient discomfort and higher vein induration, incompressibility, and sclerosing effect than Group B (P = 0.003, P = 0.04, P = 0.02, P = 0.001, respectively). At year 2, Group A continued to show higher reduction in diameter of sclerosing effect (P = 0.02)." Comment: described statistical significance, but there were no data to determine the number of participants developing each outcome. So, we did not know if there was a low or high number of events, and if the effect occurrence was due to chance.
Selective reporting (reporting bias)	Unclear risk	Comment: did not provide a protocol
Other bias	Unclear risk	Comment: there was no information on the baseline characteristics, and no description of the vein diameter.

Ouvry 2008.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: multicentre clinic‐based study Duration of recruitment to study: 10 months Duration of follow‐up: 3 week and 2 years Unit of analysis: participants
Participants	Number: 95 patients Age (mean): liquid 55 years; foam 54 years Gender: 7 male, 88 female: female %: liquid 94% and foam 92% Inclusion criteria: "18‐80 years old; written consent; GSV insufficiency, truncal diameter of between 4 and 8 mm inclusive (reflux greater than 1 s; patient standing: manual compression of the calf, followed by release); CEAP: C2‐C6." CEAP: did not discriminate between groups C2‐C6 Exclusion criteria: "Mental or psychiatric disturbance; chronic hepatic, cardiac, renal or respiratory insufficiency; pregnancy or breast feeding; location outside the geographic study area; allergy to polidocanol or lauromacrogol; personal history of DVT; progressive malignant disease; constitutional or acquired thrombophilia; absence of effective contraception; intolerance to alcohol." Dropouts: only in foam group: 3 at 6 months, 1 at 1 year and 1 at 2 years
Interventions	Group 1 (n = 47): polidocanol foam Group 2 (n = 48): polidocanol liquid Sclerosant: polidocanol 3% Dose: "patients were treated with a single echo‐guided GSV injection of 3% polidocanol (Aetoxisclerol, Kreussler, Germany) in either foam or liquid form. For veins of 4‐6 mm in truncal diameter, 2 mL were injected; 2.5 mL were injected in veins of 6‐8 mm. The mixture to produce the foam consisted of one part of 3% polidocanol to 4 parts of sterile air and the technique used was the DSS technique." Type of compression: "No compression was applied after the treatment; a class 2 compression stocking (15‐20 mmHg) was only recommended combined with paracetamol if secondary pain or inflammation developed." Duration of compression: no compression Number of sites: not stated (1 dose each) Veins condition: full veins (supine during treatment) Additional treatment: "no special postoperative instruction regarding exercise and/or walking."
Outcomes	Primary outcomes: "duplex ultrasound imaging to demonstrate the elimination of venous reflux in the saphenous trunk." Secondary outcomes: "differences in length of occlusion measured by B mode ultrasonography, the rate of recanalisation and incidence of side effects following treatment."
Notes	Corresponding author: C Hamel‐Desnos Conflict of interest: not stated Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: randomisation not described
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented blinding of the personnel.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented blinding of the personnel.
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Low risk	Comment: described dropouts (10% in foam group) and outcomes. Seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described dropouts (10% in foam group) and outcomes. Seemed to be appropriate
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Comment: none identified. Described baseline characteristics with similar groups. Sample size less than 50 in each group

Rabe 2008.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: clinical‐based study Duration of recruitment to study: 27 months Duration of follow‐up: 3 months Unit of analysis: participants
Participants	Number: 108 patients Age (mean): group 1: 50.5 years; group 2: 49.8 years Gender (female/male): group 1: 35/19; group 2: 39/13 Inclusion criteria: "incompetent GSV with a diameter less than 12 mm measured in upright position 3 cm below the sapheno‐femoral junction (SFJ), incompetent terminal valve, reflux duration of 1 second or more measured 3 cm below the SFJ under Valsalva manoeuvre, and refilling time measured by PPG of less than 25 sec. Other inclusion criteria were: reflux in the GSV from the SFJ at least to the knee, CEAP‐classifications: C2‐C5, age between 18 and 70 years." CEAP: group 1: C2: 26; C3: 15; C4: 12; C5: 1. CEAP C5 accounted for 1.851% of included participants. Group 2: C2: 26; C3: 14; C4: 8; C5: 4. CEAP C5 accounted for 7.692% of participants. Exclusion criteria: "Patients with a history of deep vein thrombosis, superficial thrombosis, major leg oedema, known patent foramen ovale, and known predisposition to migraine were excluded from the trial. Only patients who gave their informed written consent were included in the clinical trial." Dropouts: For group 1: 1; group 2: 1. Both for violation to entry criteria
Interventions	Group 1 (n = 55): polidocanol 3% foam Group 2 (n = 53): polidocanol 3% liquid Sclerosant: polidocanol Dose: "After aspiration of 1.6 mL polidocanol 3% (Aethoxysklerol 3%, Chemische Fabrik Kreussler & Co. GmbH, Wiesbaden, Germany) into the other syringe (1:5.6) standardised movements of the syringe plungers were achieved by using the Turbofoam machine, with a controlled number of movements, speed, and power. In the liquid group Aethoxysklerol 3% was used. The maximum doses per treatment session were limited to 5 mL polidocanol foam and 4 mL liquid." Type of compression: "A German class 2 (30 mm Hg) thigh‐length stocking was then applied. Patients were asked to wear this for at least 8 hours a day for a 14 day period." Duration of compression: 14 days Number of sites: not stated Veins condition: full vein: "During treatment patients lay supine to minimise the risk of syncope." Additional treatment: "Further treatment sessions were arranged 2 weeks and 4 weeks following the first." "Patients were advised to continue their normal daily activities, whilst avoiding strenuous sports, extreme physical activities, sun bathing and saunas."
Outcomes	Primary outcomes: "elimination of venous reflux of duration >0.5 sec at a location 3 cm distal to the SFJ 3 months following completion of treatment." Secondary outcomes: "occlusion of the GSV 3 cm and 25 cm distal to the SFJ as judged by duplex ultrasonography. Patients completed the CIVIQ questionnaire prior to commencing the study and 3 months after last injection in order to record baseline quality of life and possible improvements.". "During the treatment and the follow‐up period any adverse events and adverse drug reactions were recorded."
Notes	Corresponding author: Professor E Rabe, MD, Department of Dermatology, University of Bonn, Sigmund‐Freud‐Str. 25, 53105 Bonn, Germany Conflict of interest: declared no conflict: "E. Rabe and all other investigators of the ESAF‐study have no proprietary, financial, professional or other personal interest of any nature in any kind that could be construed as influencing their position as investigator of this trial or the manuscript entitled." Funding: "The study was funded by Chemische Fabrik Kreussler & Co. GmbH, Wiesbaden, Germany according to GCP rules." Contact with author made: (12/08/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Stratification procedures were not implemented. For the screening procedure, a 4‐digit number, constituting the site number (e.g. 01, 02, 03,..., 12) as the first (two) digits and consecutive 2 digit numbers (e.g. 01, 02, 03, ..., n), were assigned to patients as screening numbers at visit 1 (e.g. 0101 for the first patient in site 1, 1201 for the first patient in site 12). Allocation of patients by randomisation to treatment groups was performed at the sites at Visit 2. Patients were randomised to treatments by a defined randomisation procedure. Sealed envelopes containing the information of therapy to be administered to the patient, were issued to each site before recruitment phase. If a patient was eligible to participate, one of the issued randomisation envelopes was broken at the patient's visit 2, in order to allocate the individual to a treatment arm. The breaking of the envelope was documented in source data and in the case report form (CRF). In order to ensure equal amounts of patient inclusions in both treatment arms at every stage of patient enrolment, a block randomisation was chosen. One block (or four patients) constituted two patients from each treatment arm in random order. This randomisation scheme was applied and documented by a sealed randomisation list." Information obtained by email. Comment: randomisation method described and seemed to be appropriate
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	High risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Blinding of outcome assessment ‐ complications	High risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	High risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Blinding of outcome assessment ‐ quality of life	Unclear risk	Quote: "It was infeasible to perform this study using a ‘double‐blind’ protocol." Comment: not stated
Incomplete outcome data ‐ complications	Low risk	Described proposed data. Described dropouts. Comment: seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Described proposed data. Described dropouts. Comment: seemed to be appropriate
Incomplete outcome data ‐ quality of life measures	Low risk	Described proposed data. Described dropouts Comment: seemed to be appropriate
Selective reporting (reporting bias)	Unclear risk	Comment: protocol not available
Other bias	Low risk	Comment: no other source of bias detected. Described baseline characteristics and these appeared similar but no P value provided

Ragg 2015.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 2 weeks, 8 weeks and 6 months Unit of analysis: participants
Participants	Number: 30 patients Age (mean): 28‐72 years Gender: 12 male and 18 female Inclusion criteria: "patients with great saphenous vein insufficiency." CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = 10): catheter foam sclerotherapy Group 2 (n = 10): microcatheter foam sclerotherapy Group 3 (n = 10): injection sclerotherapy Sclerosant: polidocanol Dose: "Aethoxysclerol 1%, 1 + 4 with air, total amount: 4‐7.5 mL. Mean 5.8 mL" Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: empty veins: "lying patient without incline, after one minute leg elevation" Additional treatment: not stated
Outcomes	Primary outcomes: "venous occlusion" Secondary outcomes: adverse events
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact Abstract only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: did not describe allocation concealment
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Unclear risk	Comment: outcomes described and seemed to be appropriate. Did not describe dropouts. Did not describe DVT
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: outcomes described and seemed to be appropriate. Did not describe dropouts. Did not describe DVT
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Did not describe baseline characteristics. Included 30 participants

Rao 2005.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 12 weeks Unit of analysis: the vein
Participants	Number: 20 patients Age (mean): not stated Gender: 19 female and 1 male Inclusion criteria: "healthy subjects with varicose, reticular, and/or telangiectatic leg veins without incompetence at the saphenofemoral or saphenopopliteal junctions" CEAP: not stated Exclusion criteria: not stated Dropouts: no dropouts
Interventions	Group 1 (n = 10): STS foam Group 2 (n =10): polidocanol foam Sclerosant: polidocanol or STS Dose: "1 mL of sclerosing solution and 4 mL of air, mixed using the double‐syringe system technique." Type of compression: "class II graduated compression stockings (30–40 mm Hg) around the clock for 7 days following treatment." Duration of compression: 7 days Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: cosmetic appearance Secondary outcomes: tolerability, satisfaction
Notes	Corresponding author: Mitchel P Goldman, MD, Dermatology/Cosmetic Laser Associates of La Jolla, Inc., 7630 Fay Avenue, La Jolla, CA 92037 Conflict of interest: not stated Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Subjects were then randomized to have the veins of either the right leg or the left leg treated with STS and the contralateral leg veins treated with POL." Pg. 631 Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Low risk	Quote: "The treating physician was blinded as to the agent being injected." Pg. 631 Comment: described blinding of physician and patients and it seemed to be appropriate
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "The treating physician was blinded as to the agent being injected." Pg. 631 Comment: described blinding of physician and patients and it seemed to be appropriate
Blinding of outcome assessment ‐ cosmetic appearance	Low risk	Quote: "At the end of the study, four independent physicians, who were blinded to the sclerosing agents administered, assessed these images for overall clearance of vessels based on vessel size." Pg. 632 Comment: described blinding of outcome assessment and it seemed to be appropriate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "At the end of the study, four independent physicians, who were blinded to the sclerosing agents administered, assessed these images for overall clearance of vessels based on vessel size." Pg. 632 Comment: described blinding of outcome assessment and it seemed to be appropriate
Incomplete outcome data ‐ cosmetic appearance	Low risk	Quote: "All 20 subjects completed the study", Pg. 632 Comment: described dropouts and outcomes and seemed to be adequate
Incomplete outcome data ‐ complications	Low risk	Quote: "All 20 subjects completed the study". Pg. 632 Comment: described dropouts and outcomes and seemed to be adequate
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Comment: Did not describe baseline characteristics. Sample size less than 50 in each group

Santos 2019.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: outpatient setting Duration of recruitment to study: 28 months Duration of follow‐up: seven and 28 days, six and 12 months Unit of analysis: participants
Participants	Number: 50 patients Age (mean): group A: 44.8 (SD ± 11.33). Group B: 47.5 (SD ± 13.47) Gender (n female/total): group A 22/25. Group B: 21/25 Inclusion criteria: "adults aged 18 years or older, clinical, etiologic, anatomic, pathophysiologic (CEAP) classification C3, incompetent GSV with diameter of 6– 10 mm in the proximal thigh at 3 cm distal from the saphenofemoral junction (SFJ),21 both measured using color duplex ultrasound (CDU)" CEAP: CEAP C3 Exclusion criteria: "allergy to polidocanol (POL), acute deep venous thrombosis (DVT), skin disorder in the puncture area, symptomatic foramen ovale patency, pregnancy, peripheral arterial insufficiency (ankle‐brachial index <0.8), concomitant small saphenous vein incompetence, use of anticoagulant drugs, and change of CEAP class." Dropouts: described 1 loss of follow‐up in each group (analysed 24 participants in each group). The participant from the group A died due to previous heart disease. The participant from group B abandoned the trial due to dissatisfaction.
Interventions	Group A (n = 25): usual ultrasound‐guided foam sclerotherapy Group B (n = 25): catheter‐directed foam sclerotherapy with tumescence Sclerosant: polidocanol 3% Dose: "The volume injected per session in all patients was 10 mL". Type of compression: "Post‐surgical care consisted of a 48‐h eccentric compression with cotton gauze placed along the treated segment of the GSV and the use of a 30–40‐mmHg thigh compression stocking." Duration of compression: 19 days Number of sites: One site of venipuncture "The GSV trunk was punctured at its distal refluxing site with 0.1–0.2 mL of 2% lidocaine without vasoconstrictor". Veins condition: empty veins (Trendelemburg position) Additional treatment: "After compression by stocking, patients were encouraged to walk for 30 min and were discharged 1 h later." "Manual injections of 10–30 mL per puncture for PTLA were applied under CDU guidance within the saphenous compartment in order to squeeze the target segments of the GSV with reflux". "After SF injection, patients in both groups remained in the Trendelenburg position and underwent about 25 hook phlebectomies per leg through cutaneous mini‐incisions".
Outcomes	Primary outcomes: "Great Sapnenous Vein occlusion Rate (time frame: 6 months after the intervention). It will be measured by ultrasound". Secondary outcomes: "the full success rates of the treated GSV after the retreatment sessions at six‐ and 12‐month follow‐up. The secondary outcomes were changes in quality of life (QoL) at six‐month follow‐up and side effects and complications of the intervention during all the follow‐up." "Quality of Life Change (time frame: 6 months follow‐up after the intervention). It will be measured by the application of the Aberdeen varicose vein questionnaire." "Complications (time frame: from the intervention until 6 months follow‐up). [It] will be considered: anaphylaxis; superficial thrombophlebitis; deep venous thrombosis; pulmonary embolism; visual disturbances; migraine; transient ischemic attack; tissue necrosis; intra‐arterial injection; neurological injury; edema; skin pigmentation; skin irritation; burning after the sclerosant injection; retaining clots."
Notes	Corresponding author: Jorgete B dos Santos Conflict of interest: declared no interest conflict Funding: "financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nıvel Superior – Brasil (CAPES) – Finance Code 001". Contact with author made: (04/03/2020) we requested information of allocation concealment, blinding of outcome assessors, and raw data from the QoL.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "was generated by online randomization (https://www.sealedenvelope.com)" Comment: described randomisation and seemed to be appropriate
Allocation concealment (selection bias)	Unclear risk	Comment: did not describe allocation concealment
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Quote: "Clinical evaluation assessed local side effects and complications, and CDU, performed by the same blinded investigator, was used to exclude DVT at seven days and evaluate the GSV treatment, based on morphological (occlusion, partial or complete recanalization) and hemodynamic (no flow, anterograde flow or reflux) findings at the subsequent follow‐up periods". Comment: we believed it was impossible to blind participants due to the nature of the intervention.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Quote: "Clinical evaluation assessed local side effects and complications, and CDU, performed by the same blinded investigator, was used to exclude DVT at seven days and evaluate the GSV treatment, based on morphological (occlusion, partial or complete recanalization) and hemodynamic (no flow, anterograde flow or reflux) findings at the subsequent follow‐up periods". Comment: we believed it was impossible to blind participants due to the nature of the intervention.
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	High risk	Quote: "Clinical evaluation assessed local side effects and complications, and CDU, performed by the same blinded investigator, was used to exclude DVT at seven days and evaluate the GSV treatment, based on morphological (occlusion, partial or complete recanalization) and hemodynamic (no flow, anterograde flow or reflux) findings at the subsequent follow‐up periods". Comment: we believed it was impossible to blind participants due to the nature of the intervention.
Blinding of outcome assessment ‐ complications	High risk	Comment: answered by email as unblinded
Blinding of outcome assessment ‐ frequency of residual varicose veins	High risk	Comment: answered by email as unblinded
Blinding of outcome assessment ‐ quality of life	High risk	Comment: answered by email as unblinded
Incomplete outcome data ‐ complications	Low risk	Comment: described as side effects in table 2 Pg. 5. It seemed to be appropriate.
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described success rates in Pg. 4. We considered the non‐occluded veins as residual varicose veins.
Incomplete outcome data ‐ quality of life measures	Low risk	Comment: described quality of life in Pg. 5. It seemed to be appropriate. The author supplied QoL data by email.
Selective reporting (reporting bias)	Low risk	Comment: The study was conducted according to a previously published protocol and it seemed that there were no other selective reporting biases.
Other bias	Unclear risk	Comment: fewer than 50 participants in each group

Schadeck 1995a.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 3 minutes Unit of analysis: participants
Participants	Number: 30 patients Age (mean): not stated Gender: not stated Inclusion criteria: great saphenous vein with reflux CEAP: not stated Exclusion criteria: "Great saphenous vein with a diameter more than 6 mm." Dropouts: not stated
Interventions	Group 1 (n = 15): sotradecol 3% Group 2 (n = 15): "aetoxisclerol 4%" Sclerosant: polidocanol and STS Dose: did not state the volume Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: venous spasm Secondary outcomes: not stated
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: randomisation method not stated
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: did not describe blinding
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: did not describe blinding
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described data and dropouts which were considered appropriate
Selective reporting (reporting bias)	Unclear risk	Comment: there was no protocol available.
Other bias	Unclear risk	Comment: did not describe baseline characteristics. Fewer than 50 participants in each group

Todd 2014.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: multicentre Duration of recruitment to study: not stated Duration of follow‐up: five years Unit of analysis: participants
Participants	Number: 232 patients Age (mean): 50.8 years Gender: 72.8% female Inclusion criteria: "male and female patients aged 18–75 years who had SFJ incompetence (SFJ reflux > 0.5 s) due to reflux of the great saphenous vein (GSV) or major accessory veins as determined by duplex ultrasound, and superficial venous disease manifested by symptoms and visible varicosities. There was no maximum limit for GSV diameter, which was measured standing. Additionally, patients needed to be able to reliably complete the electronic diary and meet other study requirements". CEAP: C2 (varicose veins): placebo: 35.1%; PEM 0.125%: 22.8%; PEM 0.5%: 35.0%; PEM 1%: 34.5% C3 (oedema): placebo: 38.6%; PEM 0.125%: 52.6%; PEM 0.5%: 31.7%; PEM 1%: 37.9% C4 (skin changes): placebo: 17.5%; PEM 0.125%: 19.3%; PEM 0.5%: 30.0%; PEM 1%: 24.1% C5 and C6 (healed; active ulcers): placebo: 8.8%; PEM 0.125%: 5.3%; PEM 0.5%: 3.3%; PEM 1%: 3.4% Exclusion criteria: "patients with a history of deep vein thrombosis or pulmonary embolism; inability to comply with post‐treatment compression due to severe peripheral arterial disease or leg obesity; incompetence of the small saphenous vein, deep venous reflux, or reduced mobility; or major surgery, pregnancy, or prolonged hospitalization within 3 months. Patients with only telangiectasia or reticular veins (clinical, etiology, anatomy, pathophysiology [CEAP] clinical class C1) were also excluded". Dropouts: one participant in the placebo group and one participant in the polidocanol 1% group
Interventions	Group 1 (n = 57): placebo Group 2 (n = 57): polidocanol 0.125% Group 3 (n = 60): polidocanol 0.5% Group 4 (n = 58): polidocanol 1% Sclerosant: polidocanol Dose: at each treatment session, a maximum volume of 15 mL of study drug (in 5 mL aliquots) was allowed regardless of treatment assignment. Type of compression: "treated leg was wrapped in a limited‐stretch bandage with compression pads placed over the treated venous segments. An overstocking and thigh‐length 30‐ to 40‐mm Hg GCS (Venosan North America, Asheboro, North Carolina, USA) was placed over the dressing. The compression bandages and stocking were worn continuously for the first 48 h and the GCS only was worn day and night for an additional 12 days". Duration of compression: 12 days Number of sites: not stated Veins condition: not stated Additional treatment: participants were mobilised as soon as the treatment was complete and were encouraged to walk for at least 5 min during each waking hour for the week following treatment.
Outcomes	Primary outcomes: "patient‐reported improvement in symptoms, as measured by the change from baseline to week 8 in the 7‐day average electronic daily diary VVSymQ score" Secondary outcomes: "improvement in appearance of visible varicose veins from baseline to week 8, as measured by the IPR‐V3 and PA‐V3 scores, improvement in VCSS, and improvement in the modified VEINES‐QOL score; all were assessed at week 8".
Notes	Corresponding author: David Wright, BTG International Ltd., Five Tower Bridge, Suite 800, 300 Barr Harbor Drive, West Conshohocken, PA 19428, USA Conflict of interest: "D.I.W. is an employee of BTG International. K.L.T. conducts research and is a consultant for BTG International. He is also a consultant and stockholder for VENITI, Inc." Funding: "This study was funded by BTG International Inc." Contact with author made: (12/01/2016) allocation concealment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was stratified by baseline symptom score and by site, using an automated interactive voice recognition system (IVRS; United BioSource Corporation/Bracket, San Francisco, California, USA)." Pg. 610 Comment: described randomisation method and seemed to be appropriate
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs treatment)	Low risk	Quote: "All patients underwent an initial blinded treatment session in which they received their assigned study drug. At the week 1 follow‐up visit, if the initial treatment was unsuccessful or incomplete, patients could receive an optional, additional treatment with blinded study drug from the same PEM canister. After completion of all assessments for the primary endpoint at week 8, patients with unsuccessful or incomplete treatment or residual visible varicosities could receive one or two optional, open‐label treatments with 1.0% PEM, 1 week apart, using the same PEM canister." Pg. 610 "all dose‐concentrations of PEM were in identical canisters identified only by a numeric code; the code for a given patient was assigned by the IVRS. Placebo (agitated diluent solution) was prepared immediately prior to injection. To maintain patient blinding, the study drug was prepared outside of the treatment room; patients were physically blocked from viewing the procedure using screening or drapes and study staff were instructed to use ‘‘study drug’’ terms when referring to either PEM or placebo." Pg. 610 "Blinding was highly effective. On the day of treatment, 85% (196/230) of patients did not think they knew the treatment they had received; of the 34 who thought they knew, 20 (8.7% of all patients treated) correctly guessed their treatment." Pg. 614 Comment: described blinding of participants and personnel and seemed to be appropriate
Blinding of participants and personnel ‐ cosmetic appearance (treatment vs placebo)	High risk	Quote: "It was not possible to blind the study site staff to patient treatment with the placebo due to visual differences with the anticipated therapeutic doses (0.5% and 1.0%) of PEM." Comment: impossible to blind placebo. So, any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Low risk	Quote: "All patients underwent an initial blinded treatment session in which they received their assigned study drug. At the week 1 follow‐up visit, if the initial treatment was unsuccessful or incomplete, patients could receive an optional, additional treatment with blinded study drug from the same PEM canister. After completion of all assessments for the primary endpoint at week 8, patients with unsuccessful or incomplete treatment or residual visible varicosities could receive one or two optional, open‐label treatments with 1.0% PEM, 1 week apart, using the same PEM canister." Pg. 610 "all dose‐concentrations of PEM were in identical canisters identified only by a numeric code; the code for a given patient was assigned by the IVRS. Placebo (agitated diluent solution) was prepared immediately prior to injection. To maintain patient blinding, the study drug was prepared outside of the treatment room; patients were physically blocked from viewing the procedure using screening or drapes and study staff were instructed to use ‘‘study drug’’ terms when referring to either PEM or placebo." Pg 610 "Blinding was highly effective. On the day of treatment, 85% (196/230) of patients did not think they knew the treatment they had received; of the 34 who thought they knew, 20 (8.7% of all patients treated) correctly guessed their treatment." Pg. 614 Comment: described blinding of participants and personnel and seemed to be appropriate.
Blinding of participants and personnel ‐ complications (treatment vs placebo)	High risk	Quote: "It was not possible to blind the study site staff to patient treatment with the placebo due to visual differences with the anticipated therapeutic doses (0.5% and 1.0%) of PEM." Comment: any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Low risk	Quote: "All patients underwent an initial blinded treatment session in which they received their assigned study drug. At the week 1 follow‐up visit, if the initial treatment was unsuccessful or incomplete, patients could receive an optional, additional treatment with blinded study drug from the same PEM canister. After completion of all assessments for the primary endpoint at week 8, patients with unsuccessful or incomplete treatment or residual visible varicosities could receive one or two optional, open‐label treatments with 1.0% PEM, 1 week apart, using the same PEM canister." Pg. 610 "all dose‐concentrations of PEM were in identical canisters identified only by a numeric code; the code for a given patient was assigned by the IVRS. Placebo (agitated diluent solution) was prepared immediately prior to injection. To maintain patient blinding, the study drug was prepared outside of the treatment room; patients were physically blocked from viewing the procedure using screening or drapes and study staff were instructed to use ‘‘study drug’’ terms when referring to either PEM or placebo." Pg. 610 "Blinding was highly effective. On the day of treatment, 85% (196/230) of patients did not think they knew the treatment they had received; of the 34 who thought they knew, 20 (8.7% of all patients treated) correctly guessed their treatment." Pg. 614 Comment: described blinding of participants and personnel and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs placebo)	High risk	Quote: "It was not possible to blind the study site staff to patient treatment with the placebo due to visual differences with the anticipated therapeutic doses (0.5% and 1.0%) of PEM." Comment: any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ quality of life (treatment vs treatment)	Low risk	Quote: "All patients underwent an initial blinded treatment session in which they received their assigned study drug. At the week 1 follow‐up visit, if the initial treatment was unsuccessful or incomplete, patients could receive an optional, additional treatment with blinded study drug from the same PEM canister. After completion of all assessments for the primary endpoint at week 8, patients with unsuccessful or incomplete treatment or residual visible varicosities could receive one or two optional, open‐label treatments with 1.0% PEM, 1 week apart, using the same PEM canister." Pg. 610 "all dose‐concentrations of PEM were in identical canisters identified only by a numeric code; the code for a given patient was assigned by the IVRS. Placebo (agitated diluent solution) was prepared immediately prior to injection. To maintain patient blinding, the study drug was prepared outside of the treatment room; patients were physically blocked from viewing the procedure using screening or drapes and study staff were instructed to use ‘‘study drug’’ terms when referring to either PEM or placebo." Pg. 610 "Blinding was highly effective. On the day of treatment, 85% (196/230) of patients did not think they knew the treatment they had received; of the 34 who thought they knew, 20 (8.7% of all patients treated) correctly guessed their treatment." Pg. 614 Comment: described blinding of participants and personnel and seemed to be appropriate
Blinding of participants and personnel ‐ quality of life (treatment vs placebo)	High risk	Quote: "It was not possible to blind the study site staff to patient treatment with the placebo due to visual differences with the anticipated therapeutic doses (0.5% and 1.0%) of PEM." Comment: any comparison with placebo should be judged as high risk of bias.
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	Low risk	Quote: "All patients underwent an initial blinded treatment session in which they received their assigned study drug. At the week 1 follow‐up visit, if the initial treatment was unsuccessful or incomplete, patients could receive an optional, additional treatment with blinded study drug from the same PEM canister. After completion of all assessments for the primary endpoint at week 8, patients with unsuccessful or incomplete treatment or residual visible varicosities could receive one or two optional, open‐label treatments with 1.0% PEM, 1 week apart, using the same PEM canister." Pg. 610 "all dose‐concentrations of PEM were in identical canisters identified only by a numeric code; the code for a given patient was assigned by the IVRS. Placebo (agitated diluent solution) was prepared immediately prior to injection. To maintain patient blinding, the study drug was prepared outside of the treatment room; patients were physically blocked from viewing the procedure using screening or drapes and study staff were instructed to use ‘‘study drug’’ terms when referring to either PEM or placebo." Pg. 610 "Blinding was highly effective. On the day of treatment, 85% (196/230) of patients did not think they knew the treatment they had received; of the 34 who thought they knew, 20 (8.7% of all patients treated) correctly guessed their treatment." Pg. 614 Comment: described blinding of participants and personnel and seemed to be appropriate
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs placebo)	High risk	Quote: "It was not possible to blind the study site staff to patient treatment with the placebo due to visual differences with the anticipated therapeutic doses (0.5% and 1.0%) of PEM." Comment: any comparison with placebo should be judged as high risk of bias.
Blinding of outcome assessment ‐ cosmetic appearance	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ complications	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ quality of life	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Incomplete outcome data ‐ cosmetic appearance	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Incomplete outcome data ‐ complications	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Incomplete outcome data ‐ quality of life measures	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Incomplete outcome data ‐ frequency of persistence of symptoms	Low risk	Quote: "Each reviewer independently scored the photographs of the treatment leg using a high‐resolution monitor and specialized software that allowed the reviewer to zoom in and out and pan left and right, approximating a live review. The images were presented, in random order, one at a time; the reviewer was blinded to patient, time point, and treatment assignment." Comment: described and seemed to be appropriate
Selective reporting (reporting bias)	Low risk	Quote: "This trial is registered on www.clinicaltrials.gov no. NCT01231373". Comment: there was a protocol available and the reporting was according to what was planned.
Other bias	Low risk	Comment: fewer than 20% of dropouts and no other source of bias detected

Ukritmanoroat 2011.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: outpatient clinic Duration of recruitment to study: not stated Duration of follow‐up: 15, 30 and 90 days Unit of analysis: the vein
Participants	Number: 50 patients (100 sites) Age (mean): 45.5 years (SD 11.3) Gender: male (%): 5 (10); female (%): 45 (90) Inclusion criteria: "Patients with primary reticular varices (more than 2 mm of diameter) or postoperative varices that did not involve the saphenofemoral junction." CEAP: not stated Exclusion criteria: "Patients with truncal varices with junctional (terminal valve) and extra‐junctional incompetence. ‐ Postoperative varices that involved the saphenofemoral junction ‐ Post‐thrombotic varices with occluded deep veins ‐ Chronic ischaemia of the lower limbs ‐ Severe arterial hypertension (blood pressure greater than 180/95 mmHg) ‐ Patients being treated with anticoagulants and anti‐inflammatories and/or diuretics to avoid affecting the appearance of possible secondary effects." Dropouts: not stated
Interventions	100 sites were treated (84 had reticular veins and 16 had postoperative varicose veins) Group 1 (n = 50 sites): liquid Group 2 (n = 50 sites): foam Sclerosant: polidocanol Dose: "The intervention in the study group consisted of injecting 2 mL of foam into only one varicose vein. Foam was obtained from 0.5 mL of liquid Polidocanol mixed with air at a ratio of 1: 4, using the Tessari method which uses a three‐way stopcock to mix the sclerosant." Type of compression: "After the sclerosant was injected, the sclerosed vein was compressed for 48 h with stockings at a pressure of 25‐35 mmHg." Duration of compression: 48 hours Number of sites: not stated Veins condition: not stated Additional treatment: anti‐inflammatory medicines
Outcomes	Primary outcomes: "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound". Secondary outcomes: "Appearance of the following side effects was recorded: pain in the treated region graded on an ordinal scale (absent, mild, moderate, or severe), requirement for analgesic treatment and type of analgesic given, inflammation in the treated region and degree of severity of inflammation (mild, moderate), appearance of skin pigmentation in the sclerosis region, formation of bulla, cutaneous necrosis and other effects."
Notes	Corresponding author: T Ukritmanoroat, Department of Surgery, Rajavithi Hospital, College of Medicine, Rangsit University, 2 Phyathai Road, Ratchathewi, Bangkok 10400, Thailand Conflict of interest: declared no interest conflict Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, allocation concealment, blinding of participants and personnel, protocol, provision to evaluate DVT in risk analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The assignment of the first region to be treated to liquid or foam was performed according to a list of random numbers from 0 to 50." Pg. S36 Comment: described as randomised but did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Quote: "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb." Pg. S36 Comment: did not describe blinding for patient or interventionist (for interventionist analysis, probably impossible to blind)
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Quote: "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb." Pg. S36 Comment: did not describe blinding for patient and interventionist (for interventionist analysis, probably impossible to blind)
Blinding of outcome assessment ‐ complications	Low risk	Quote: "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb." Pg. S36 "Results were assessed by research team other than the doctor who had performed the treatment." Pg. S36 Comment: described blinding of outcome assessment and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "Efficacy was assessed according to whether sclerosis of the vein was complete as shown by duplex ultrasound performed by an observer unaware of the treatment group assigned to each area of the limb." Pg. S36 "Results were assessed by research team other than the doctor who had performed the treatment." Pg. S36 Comment: described blinding of outcome assessment and seemed to be appropriate
Incomplete outcome data ‐ complications	Unclear risk	Comment: described every proposed outcome. Did not describe dropouts
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: described every proposed outcome. Did not describe dropouts
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available. Data were described.
Other bias	Unclear risk	Comment: no description of baseline characteristics

Wright 2006.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: safety and efficacy Setting: hospital‐based Duration of recruitment to study: not stated Duration of follow‐up: 7 day, 28 day, 3 moths, 12 months Unit of analysis: participants
Participants	Number: 399 (274 for Varisolve and 125 for sclerotherapy) Age (mean): Varisolve: 50.7; sclerotherapy: 49.6 Gender: (male/female) Varisolve: 76/183; sclerotherapy: 32/92 Inclusion criteria: "male or female; aged 18–75 years; SFJ or SPJ junction incompetence or both with retrograde blood flow for more than 1 s and less than 7 s demonstrated by duplex scanning; great saphenous incompetence and/or short saphenous incompetence; minimum of 10 cm proximal trunk vein incompetence; normal arterial circulation; ankle: brachial pressure ratio more than 0.9; haematology and blood chemistry profiles within normal arteriosclerosis acceptable limits, i.e. abnormalities not clinically significant; normal deep venous system on duplex scanning, no evidence of occlusion or incompetence. Evidence of reflux was acceptable, however, if it was confined to a limited segment caused by filling of the superficial system." CEAP: considered as treated: Varisolve: 259; sclerotherapy: 125; Varisolve: C2: 187; C3: 58; C4: 18. Sclerotherapy: C2: 79; C3: 35; C4: 11 Exclusion criteria: "History of major superficial thrombophlebitis; venographic or ultrasonographic evidence of current or previous; immobility – unable to walk a minimum of 5 min/h/day; body mass index more than 32 kg/m2; contraindications for polidocanol: severe hypertension, glomerulonephritis and nephroses, hepatopathy, active phlebitis, general febrile disease, diabetes, asthma, advanced arteriosclerosis; serious coexisting illness including carcinoma, cardiac or respiratory disease, past or present venous ulcers; pregnancy or lactation; concomitant treatment with disulfiram, systemic corticosteroids, oestrogen or anticoagulants; known allergic response to polidocanol or severe allergic diathesis; abnormal, clinically significant haematology or biochemistry profiles" Dropouts: 15 did not receive Varisolve. At 3 months: Varisolve: 3; sclerotherapy: 0. At 12 months: Varisolve: 7; sclerotherapy: 1
Interventions	Group 1 (n = 274): Varisolve technique with polidocanol 1% Group 2 (n = 125): any marketed sclerosant at any marketed concentration liquid or foam Sclerosant: polidocanol 1% Dose: "The maximum total dose of microfoam was initially set at 60 mL, and this was subsequently reduced to 30mL." Type of compression: "an under bandage from ankle to groin, foam pads along the treated veins, a second layer of bandage, a thin overstocking, and finally 30–40mmHg thigh‐length compression stocking (Sigvaris) was applied. Compression was worn for 14 days post treatment: day and night for the first seven days." Duration of compression: 14 days Number of sites: not stated Veins condition: not stated Additional treatment: "If additional treatment was necessary, this was performed on days 7 and 14 for patients treated with Varisolves, and between days 7 and 21 for surgery or sclerotherapy patients."
Outcomes	Primary outcomes: "occlusion of incompetent trunk vein(s) from the source of deep to superficial flow and the elimination of reflux greater than 0.5 s in the treated vein(s) 2–5 cm from its point of termination, as determined by duplex scanning." Secondary outcomes: "severity of post‐procedure pain (days 1–6 measured using VAS), time taken to return to normal activities, number of treatment sessions required and Month 12 response rate."
Notes	Corresponding author: Mr David Wright FRCS, Provensis Limited, BTG Plc, 10 Fleet Place, Limeburner Lane, London EC4M 7SB, UK Conflict of interest: Sigvaris Ltd: provision of compression stockings. Nothing more stated Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, any available protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients underwent central randomization to either the Varisolves technique or alternative treatment at a ratio of 2:1 in favour of the Varisolves technique." Pg. 181 Comment: did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Comment: no allocation concealment described
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Described as "open‐label". Pg. 181 Comment: no blinding of participants and personnel
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Described as "open‐label". Pg. 181 Comment: no blinding of participants and personnel
Blinding of outcome assessment ‐ complications	High risk	Described as "open‐label". Pg. 181 Comment: no blinding of outcome assessment
Blinding of outcome assessment ‐ frequency of residual varicose veins	High risk	Described as "open‐label". Pg. 181 Comment: no blinding of outcome assessment
Incomplete outcome data ‐ complications	High risk	Quote: "Figure 1 Patient disposition. M3, Month 3; M12, Month 12. a Study interrupted by sponsor; b Adverse event; c Patient decision; d Other reasons; e Missing; f First treatment (Leg 1)" Comment: legend of figure 1 of study paper declared participants withdrawn as study was interrupted by sponsor. In total 54 participants were withdrawn before treatment as the study was delayed to allow for an investigation of potential causes of DVT.
Incomplete outcome data ‐ frequency of residual varicose veins	High risk	Quote: "Figure 1 Patient disposition. M3, Month 3; M12, Month 12. a Study interrupted by sponsor; b Adverse event; c Patient decision; d Other reasons; e Missing; f First treatment (Leg 1)" Comment: legend of figure 1 of study paper declared participants withdrawn as study was interrupted by sponsor. In total 54 participants were withdrawn before treatment as the study was delayed to allow for an investigation of potential causes of DVT.
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol provided
Other bias	Unclear risk	Comment: insufficient information to judge any imbalance between groups

Yamaki 2009.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: effectiveness Setting: not stated Duration of recruitment to study: 9 months Duration of follow‐up: 3 days, 2 weeks, 1 month, 3 months and 6 months Unit of analysis: lower limb
Participants	Number: 107 patients (112 limbs) Age (mean): multiple injections: 65 years (SD 12.6); few injections: 64.4 years (SD 10) Gender: female n (%): multiple injections: 41 (77.4); few injections: 36 (66.7) female/male: 78/29 Inclusion criteria: "patients with isolated GSV reflux associated with SFJ incompetence" CEAP: multiple injections: C2: 39; C4: 15; few injections: C2: 41; C4: 14. Exclusion criteria: "Patients with myocardial ischaemia, arterial insufficiency with an ankle brachial index of less than 0.9, pregnancy in the first trimester and after the 36th week of gestation, local infection in the area of sclerotherapy, active thrombophlebitis and acute deep vein thrombosis." Dropouts: not stated
Interventions	Group 1 (n = 56): multiple injections (3 or more) less than 0.5 mL/injection Group 2 (n = 56): few injections (less than 3) more than 0.5 mL/injection Sclerosant: polidocanol 3% Dose: "Foam was obtained from 0.5 mL of liquid polidocanol mixed with air at a ratio of 1:4 using a three‐way stopcock to mix the sclerosant. Each visible varicose tributary vein was injected first, with multiple injections or a few injections, using 27‐gauge needles. In patients who received multiple injections, the number of injections depended on the locations of the varicose veins. In contrast, only 1‐3 injections were used to occlude varicose tributaries in patients who received few injections. Then 3% POL‐foam was injected under ultrasound guidance using 21‐gauge needles, starting 3‐4 cm distal to the SFJ, and a second injection was made 5‐10 cm distal to the initial point." Type of compression: not stated Duration of compression: not stated (described doing compression) Number of sites: variable according the group Veins condition: empty veins (affected legs elevated 30 degrees) Additional treatment: not stated
Outcomes	Primary outcomes: obliteration of the GSV at 6 months Secondary outcomes: complications related to foam sclerotherapy
Notes	Corresponding author: T Yamaki Conflict of interest: declared no conflict Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, dropouts, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: described as block randomised (by email)
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of outcome assessment ‐ complications	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Incomplete outcome data ‐ complications	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol provided
Other bias	Low risk	Comment: no other source of bias detected

Yamaki 2012.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: effectiveness Setting: not stated Duration of recruitment to study: 6 months Duration of follow‐up: 2 weeks, one, three and six months Unit of analysis: lower limbs
Participants	Number: 97 patients (103 limbs) (31 males and 66 females) Age (mean): UGFS: 69 years; VFS: 69 years Gender: female n (%): UGFS: 36 (75); VFS: 33 (67.3) Inclusion criteria: "GSV reflux associated with sapheno‐femoral junction (SFJ) incompetence" CEAP: UGFS: C2: 39, C3: 1, C4: 11; VFS: C2: 36, C3: 1, C4: 15 Exclusion criteria: "Patients with myocardial ischaemia, arterial insufficiency with an ankle‐brachial index of less than 0.9, in the first trimester of pregnancy and after the 36th week of gestation, local infection in the area for sclerotherapy, active thrombophlebitis and acute deep vein thrombosis." Dropouts: not stated
Interventions	Group 1 (n = 51): UGFS Group 2 (n = 52): VFS Sclerosant: polidocanol 1% Dose: "Patients who were treated with UGFS combined with VFS received <0.5 ml POL foam per injection to minimise any foam migration beyond the target vein. Subsequently 1% POL foam was injected into the GSV under ultrasound guidance, starting 3‐4 cm distal to the SFJ. A second injection was performed 5‐10 cm distal to the initial point using a 21‐gauge venous catheter. The GSV cannulae were inserted before injection of the tributaries. Patients who were treated with VFS alone received 0.6‐1.0 ml POL foam per injection. Thus, the total amount of injected foam did not exceed 10 ml in any of the cases." Type of compression: "After completion of the DUS examination, compression pads and elastic bandages were applied, and kept on continuously for the first 2 days. All the patients were encouraged to ambulate after the treatment. On post‐sclerotherapy day 3, the elastic bandages and compression pads were removed, and a class II thigh‐high compression stocking was applied." Duration of compression: not stated Number of sites: not stated Veins condition: empty veins ("the affected leg elevated 30 degrees") Additional treatment: "all patients received the same sclerosant and were allowed only one additional treatment session during the follow‐up period of 6 months."
Outcomes	Primary outcomes: "obliteration of the GSV at 6 months" Secondary outcomes: complications related to foam sclerotherapy
Notes	Corresponding author: T Yamaki Conflict of interest: declared no conflict Funding: not stated Contact with author made: (12/08/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, dropouts, protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: described as block randomised (informed by email)
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of recurrent varicose veins and venous flare formation	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of participants and personnel ‐ frequency of persistence of symptoms (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of outcome assessment ‐ complications	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of residual varicose veins	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by email. Comment: described and seemed to be appropriate
Blinding of outcome assessment ‐ frequency of persistence of symptoms	Low risk	Quote: "We managed to minimize the selection bias not to be aware of the upcoming assignment among attending physicians, sonographers and subjects." "We used single blind clinical trial." Informed by e‐mail. Comment: described and seemed to be appropriate
Incomplete outcome data ‐ complications	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Incomplete outcome data ‐ frequency of recurrent varicose veins and venous flare formation	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Incomplete outcome data ‐ frequency of persistence of symptoms	Low risk	Comment: described no dropouts (informed by email). Described proposed outcomes
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Low risk	Comment: no other source of bias detected. Described baseline characteristics, which appeared to be similar

Zeh 2003.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: one month Unit of analysis: participants
Participants	Number: 20 patients Age (mean): not stated Gender: not stated Inclusion criteria: "long saphenous vein incompetence diameters between 3 and 9 mm" CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = 10): polidocanol 3% 0.5 cc + air 2 cc Group 2 (n = 10): polidocanol 3% 0.4 cc + Gelofusine 0.1 cc + air 2 cc Sclerosant: polidocanol 3% Dose: 2.5 mL Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: sclerose of the LSV Secondary outcomes: spasm and reflux
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: randomisation method not stated
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described outcomes. Described no dropouts
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Comment: did not describe baseline balance. Fewer than 50 participants in each group

Zhang 2009.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: not stated Unit of analysis: not stated
Participants	Number: 288 patients Age (mean): not stated Gender: not stated Inclusion criteria: "Patients with reticular veins and/or small‐sized varicose veins (1‐5 mm) Group C: patients with medium‐sized and/or large non‐sapheneous subcutaneous varicose veins with a reflux > 0.5 s ( > 5 mm)." CEAP: C1 and C2. Did not discriminate between CEAP clinical classes Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = not stated): placebo Group 2 (n = not stated): polidocanol Sclerosant: polidocanol Dose: volume injected not described Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: "disappearance of the veins according to a 5‐grade scale using a digital imaging system. For group C occlusion of vein and/or absence of reflux > 0.5 s was measured by duplex scan." Secondary outcomes: not stated
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: randomisation method not stated
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	High risk	Comment: the nature of the intervention prevented the blinding of personnel.
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ frequency of residual varicose veins	Unclear risk	Comment: data not presented
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol available
Other bias	Unclear risk	Comment: the study was an abstract that described the method but without published data to perform analyses.

Zhang 2014.

Study characteristics
Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: hospital‐based study Duration of recruitment to study: 2 years Duration of follow‐up: 3 and 6 months Unit of analysis: lower limb
Participants	Number: 80 patients Age (mean): 34 to 65 years (49 ± 12.5 years) Gender: 32 men and 48 women Inclusion criteria: "moderate or severe lower extremity varicose veins; patients having post‐exercise aching and swollen sensation with weakness; colour Doppler showing moderate or severe regurgitant flow in the GSV." CEAP: C2 to C6. Did not discriminate between CEAP clinical classes Exclusion criteria: "deep‐vein thrombosis, stenosed internal iliac veins." Dropouts: not stated
Interventions	Group 1 (n = 40): antegrade Group 2 (n = 40): retrograde Sclerosant: polidocanol 1% Dose: 12 to 15 mL Type of compression: "postoperative treatment after surgery is completed immediately with elastic bandage bandaged limb. 24 h after replacement of elastic compression stockings (ankle portion pressure 25 ‐30 mmHg, 1 mmHg = 0.133 kPa)." Duration of compression: 3 weeks Number of sites: one site Veins condition: full vein ("a rubber band tied at the knee") Additional treatment: phlebography
Outcomes	Primary outcomes: vessel occlusion (success rate) Secondary outcomes: "operation time, recovery time, the dosage of the sclerosing agent used, the incidence of complications and the use of additional treatment"
Notes	Corresponding author: Zhang Wan‐gao. Intervention Center, the First Affiliated Hospital, Anhui University of Traditional Chinese Medicine and Pharmacology, Hefei, Anhui Province 230031, China Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Individual patients drew a numbered lot themselves, and were separated to either anterograde or retrograde groups based on odd or even numbers." Comment: unclear methodology for randomisation
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ complications (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of participants and personnel ‐ frequency of residual varicose veins (treatment vs treatment)	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ complications	Unclear risk	Comment: not stated
Blinding of outcome assessment ‐ frequency of residual varicose veins	Unclear risk	Comment: not stated
Incomplete outcome data ‐ complications	Low risk	Comment: described and seemed to be appropriate
Incomplete outcome data ‐ frequency of residual varicose veins	Low risk	Comment: described and seemed to be appropriate
Selective reporting (reporting bias)	Unclear risk	Comment: there was no protocol available.
Other bias	Unclear risk	Comment: described participant characteristics but not baseline balance between groups. Unable to judge. Fewer than 50 participants in each group

AVP: ambulatory venous pressure
CEAP: clinical, etiologic, anatomic, and pathophysiologic classification of venous disorders
CIVIQ: the chronic venous insufficiency questionnaire
CDU: colour duplex ultrasound
CRF: care report form
DSS: Double syringe system technique
DUS: duplex ultrasound
DVT: deep vein thrombosis
Foam J&J‐sclerotherapy: sclerotherapy with foam including a tensioactive substance to enhance ultrasound vision and STD (sclerosing agent)
GCS: graduated compression stockings
GSV: great saphenous vein
h: hours
IPR‐V3: independent photography review ‐ visible varicose veins scores
ITT: intention to treat analysis
IVRS: interactive voice recognition system
J&J‐93FA: tensioactive substance to enhance ultrasound vision (ultrasound contrast agent)
LSV: great saphenous vein
n: number of participants
PAD: peripheral arterial disease
PA‐V3: patient self‐assessment of visible varicose veins
PE: pulmonary embolism
PEM: polidocanol endovenous microfoam
POL: polidocanol
PPG: photoplethysmography
PTLA: perivenous tumescent local anaesthesia
QoL: quality of life
RCT: randomised controlled trial
RT: refilling time
s: seconds
SD: standard deviation
SF: sclerosant foam
SFJ: sapheno‐femoral junction
SPJ: sapheno‐popliteal junction
SSV: small saphenous vein
STD: sodium tetradecyl sulphate
STS: sodium tetradecyl sulphate
Sym: symptom
UGFS: ultrasound guided foam sclerotherapy
VAS: visual analogue scale
VCSS: Venous clinical severity score
VEINES(‐QOL): VEnous INsufficiency Epidemiological and Economic Study on Quality of Life
VFS: visual foam sclerotherapy
VV: varicose veins
VVSymQ Score: patient‐reported symptoms of varicose veins

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Ariyoshi 1996	Compared surgery and sclerotherapy versus sclerotherapy alone
Biemans 2013	Compared foam sclerotherapy versus EVLA for varicose veins
Brittenden 2015	Compared foam sclerotherapy with surgery or EVLA
Lattimer 2012	Compared UGFS versus EVLA
Leung 2016	Compared sclerotherapy with MOCA versus EVLA
Mishra 2016	Compared UGFS with RFA
NCT02462720	Compared sclerotherapy with RFA
Ragg 2015a	Compared injection sclerotherapy and EVLA versus no injection sclerotherapy
Rasmussen 2011	Compared EVLA versus RFA versus UGFS versus stripping
Theivacumar 2008	Compared standard EVLA versus EVLA concomitant to sclerotherapy
Vahaaho 2015	Compared UGFS versus EVLA versus varicose vein stripping
Vernermo 2016	Compared UGFS versus EVLA versus surgery
Zafarghandi 2017	Compared UGFS with stab avulsion ambulatory phlebectomy

EVLA: endovenous laser ablation
MOCA: mechanochemical ablation
RFA: radiofrequency ablation
UGFS: ultrasound guided foam sclerotherapy

Characteristics of studies awaiting classification [ordered by study ID]

Labas 2003.

Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: hospital‐based study Duration of recruitment to study: 10 years Duration of follow‐up: 6 months, 5 years Unit of analysis: participants
Participants	Number: 1622 patients Age (mean): not stated Gender: not stated Inclusion criteria: not stated CEAP: not stated Exclusion criteria: not stated Dropouts: none reported
Interventions	Group 1 (n = 454): Siggs technique using polidocanol Group 2 (n = 876): Fegan's technique using STS Group 3 (n = 292): Fegan's procedure with a combination of two sclerosants (polidocanol + STS) Sclerosant: aethoxysclerol (polidocanol), STS Dose: did not state the volume Type of compression: "uninterrupted 6‐week compression (stockings 44 mmHg)" Duration of compression: 6 weeks Number of sites: not stated Veins condition: empty vein Additional treatment: "Forced mobilisation was used".
Outcomes	Primary outcomes: cosmetic appearance, including photographic evidence Secondary outcomes: "Symptoms: cramps, pain, fatigue, heaviness, disappearance of varices"
Notes	Corresponding author: P Labas, MD, PhD, Dept of Surgery, Faculty of Medicine, University hospital, Mickiewiczova 13, SK‐813 69 Bratislava 1, Slovakia Conflict of interest: not stated Funding: not stated Contact with author made: (12/01/2016) CEAP C2, C3 and C4, random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, protocol. Third group treated with both aethoxysclerol and STD (Fegan's method) not discussed further: relative proportion of two sclerosants not stated. Comment: described inclusion of people with CVI. On page 79, described "varices, ulcers, eczemas, pigmentation, and perforators". We did not know if all participants had CEAP C5 and C6 or if there were participants with C2, C3 and C4 clinical classes. The study author used different techniques (Sigg and Fegan) and sclerosants (polidocanol and sodium tetradecyl sulphate) and these interventions were not planned for the review. If details become available we could consider including the study in a future version of the review.

Satokawa 2003.

Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy and safety Setting: not stated Duration of recruitment to study: 6 years Duration of follow‐up: not stated Unit of analysis: participants
Participants	Number: 86 patients Age (mean): not stated Gender: not stated Inclusion criteria: not stated. Described "polidocanol 3% with varicosis greater than 3 mm in diameter" CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = not stated): polidocanol 3% Group 2 (n = not stated): placebo Sclerosant: polidocanol 3% Dose: not stated Type of compression: did use compression, but did not describe the type or grade of compression Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: venous occlusion (efficacy) Secondary outcomes: not stated
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact Comment: no data (intervention groups were not clear), unable to contact author. Unsure if true RCT

Schadeck 1995b.

Methods	Type of study: RCT Intervention model: parallel Primary purpose: efficacy Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: 3 minutes Unit of analysis: participants
Participants	Number: 50 patients Age (mean): not stated Gender: not stated Inclusion criteria: "incompetent great saphenous vein" CEAP: not stated Exclusion criteria: "Great saphenous vein with a diameter more than 6 mm." Dropouts: not stated
Interventions	Group 1 (n = 25): sotradecol 3% sclerotherapy in great saphenous vein + in collaterals Group 2 (n = 25): sotradecol 3% sclerotherapy in great saphenous vein Sclerosant: sotradecol 3% Dose: not stated Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: venous recanalisation Secondary outcomes: not stated
Notes	Corresponding author: not stated Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact Comment: no data (intervention groups were not clear), unable to contact author. Unsure if true RCT

SLCTR/2008/014.

Methods	Type of study: RCT Intervention model: parallel Primary purpose: not stated Setting: not stated Duration of recruitment to study: not stated Duration of follow‐up: not stated Unit of analysis: not stated
Participants	Number: 300 patients Age (mean): not stated Gender: not stated Inclusion criteria: "Patients with varicose veins due to sapheno femoral incompetence and long saphenous vein incompetence. Patients who had given the consent" CEAP: not stated Exclusion criteria: "Proven secondary varicose veins due to deep vein thrombosis, arterio venous fistulae or pregnancy. Recurrent sapheno femoral incompetence after previous sapheno femoral ligation. Significant below knee varicosities outside the long saphenous. Vein territory which would require stab avulsions. Patients with active venous ulcers. Patients with documented allergy to sclerosant" Dropouts: not stated
Interventions	Group 1 (n = not stated): Group 2 (n = not stated): standard therapy/practice Sclerosant: 4% sodium tetradecyl sulphate Dose: 4 mL sodium tetradecyl sulphate, 4 mL of distilled water, and 4 mL of ambient air Type of compression: not stated Duration of compression: not stated Number of sites: not stated Veins condition: not stated Additional treatment: not stated
Outcomes	Primary outcomes: venous occlusion Secondary outcomes: "Post operative pain; Time to discharge after the procedure; Time to return to daily activities; Post operative complications like wound infection, DVT"
Notes	Corresponding author: Mandika Wijeyaratne Conflict of interest: not stated Funding: Faculty of Medicine Contact with author made: emailed 03/04/2020 asking for data from the comparators and interventions, and a complete protocol Comment: unsure if the comparator met inclusion criteria (we were not sure of the meaning of "standard therapy/practice", as described. Unable to contact author

Varnagy 1985.

Methods	Type of study: RCT Intervention model: parallel Primary purpose: effectiveness Setting: outpatient‐based Duration of recruitment to study: not stated Duration of follow‐up: not stated Unit of analysis: participants
Participants	Number: 200 patients Age (mean): 17‐67 years (no description of mean or range) Gender: females Inclusion criteria: varicose veins CEAP: not stated Exclusion criteria: not stated Dropouts: not stated
Interventions	Group 1 (n = 50): 20% sodium chloride Group 2 (n = 50): sodium morrhuate Group 3 (n = 50): aethoxysclerol 3% Group 4 (n = 50): etoxisclerol 3% (national) Sclerosant: aethoxysclerol 3% Dose: not stated Type of compression: no compression Duration of compression: no compression Number of sites: not stated Veins condition: full vein Additional treatment: not stated
Outcomes	Primary outcomes: venous occlusion Secondary outcomes: pain, tattoo, necrosis rate, subjective symptomatic improvement
Notes	Corresponding author: Dr Roberto Varnagy, PO Box 28256, Caracas 1,100‐A, Venezuela Conflict of interest: not stated Funding: not stated Contact with author made: unable to contact Comment: unsure if it was an RCT or a quasi‐randomised controlled study since described randomisation as "allotted randomly (by chronological order)". Unable to contact author

CEAP: clinical, etiologic, anatomic, and pathophysiologic classification of venous disorders
CVI: chronic venous insufficiency
RCT: randomised controlled trial
STD: sodium tetradecyl sulphate
STS: sodium tetradecyl sulphate

Differences between protocol and review

For this update, changes were made to the inclusion criteria which narrowed the scope from the original version of the review. We included only studies which compared sclerotherapy with nonsurgical interventions and included participants with clinical criteria CEAP C2, C3, and C4. Studies comparing sclerotherapy with open surgery (e.g. stripping), other endovascular therapy methods (endolaser or radiofrequency), or which investigated compression after sclerotherapy were assessed as not relevant to this review as they no longer met the inclusion criteria. Studies comparing sclerotherapy to open surgery (e.g. stripping) and investigating compression after sclerotherapy will be covered in separate Cochrane Reviews.

The previous version of the review included participants aged 15 years or older. We included participants of all ages. Should studies include children, we planned to undertake subgroup analysis.

Five studies had the vein as the unit of analysis (Alos 2006; Demagny 2002; Rao 2005; Ukritmanoroat 2011; Zhang 2014). Four studies considered the lower limb as the unit of analysis (Ceulen 2007; Chleir 1997; Yamaki 2009; Yamaki 2012). These studies were included, and if any meta‐analysis could be done, we performed a sensitivity analysis to evaluate the impact of the interventions.

We intended to include only results from CEAP C2, C3 and C4. However, although requested from study authors, we were unable to obtain individual data for those studies which included participants with CEAP C1, C5 or C6. In the tables of Characteristics of included studies, we have reported, where available, the percentages/numbers of participants in the different CEAP classes. We included these studies when we believed that by including the studies with those participants, the interventions would not change the proposed outcomes, nor impact on bias as we were not studying outcomes such as leg ulcer healing after sclerotherapy in participants with CEAP C6. Otherwise, excluding these studies would have significantly reduced the available data, and therefore the usefulness of the review. We performed sensitivity analyses, where possible.

We did not undertake fixed‐effect meta‐analysis as previously planned since there was heterogeneity between interventions, comparators, and CEAP clinical classes.

We considered comparisons that are of clinical relevance. This concern motivated the use of concentrations of sclerosant in the comparisons to grade the certainty of the evidence.

Contributions of authors

RAO: study selection, data extraction, quality assessment, manuscript development
RR: study selection, data extraction, quality assessment, manuscript development, methodology review
VV: study selection, data extraction, quality assessment, manuscript development, and language review
JCBS: cross‐checked data, manuscript development; methodology review

Sources of support

Internal sources

• Cochrane Brazil, Brazil

  The review team includes members of Cochrane Brazil

External sources

• Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK

  The Cochrane Vascular editorial base is supported by the Chief scientist Office.

Declarations of interest

RAO: none known
RR: none known
VV: none known
JCBS: none known

New search for studies and content updated (no change to conclusions)