Local interventions for the management of alveolar osteitis (dry socket)

Abstract

Background

Alveolar osteitis (dry socket) is a complication of dental extractions more often involving mandibular molar teeth. It is associated with severe pain developing 2 to 3 days postoperatively with or without halitosis, a socket that may be partially or totally devoid of a blood clot, and increased postoperative visits. This is an update of the Cochrane Review first published in 2012. 

Objectives

To assess the effects of local interventions used for the prevention and treatment of alveolar osteitis (dry socket) following tooth extraction.

Search methods

An Information Specialist searched four bibliographic databases up to 28 September 2021 and used additional search methods to identify published, unpublished, and ongoing studies.

Selection criteria

We included randomised controlled trials of adults over 18 years of age who were having permanent teeth extracted or who had developed dry socket postextraction. We included studies with any type of local intervention used for the prevention or treatment of dry socket, compared to a different local intervention, placebo or no treatment. We excluded studies reporting on systemic use of antibiotics or the use of surgical techniques because these interventions are evaluated in separate Cochrane Reviews.

Data collection and analysis

We used standard methodological procedures expected by Cochrane. We followed Cochrane statistical guidelines and reported dichotomous outcomes as risk ratios (RR) and calculated 95% confidence intervals (CI) using random‐effects models. For some of the split‐mouth studies with sparse data, it was not possible to calculate RR so we calculated the exact odds ratio (OR) instead. We used GRADE to assess the certainty of the body of evidence.

Main results

We included 49 trials with 6771 participants; 39 trials (with 6219 participants) investigated prevention of dry socket and 10 studies (with 552 participants) looked at the treatment of dry socket. 16 studies were at high risk of bias, 30 studies at unclear risk of bias, and 3 studies at low risk of bias.

Chlorhexidine in the prevention of dry socket

When compared to placebo, rinsing with chlorhexidine mouthrinses (0.12% and 0.2% concentrations) both before and 24 hours after extraction(s) substantially reduced the risk of developing dry socket with an OR of 0.38 (95% CI 0.25 to 0.58; P < 0.00001; 6 trials, 1547 participants; moderate‐certainty evidence). The prevalence of dry socket varies from 1% to 5% in routine dental extractions to upwards of 30% in surgically extracted third molars. The number of patients needed to be treated (NNT) with chlorhexidine rinse to prevent one patient having dry socket was 162 (95% CI 155 to 240), 33 (95% CI 27 to 49), and 7 (95% CI 5 to 10) for control prevalence of dry socket 0.01, 0.05, and 0.30 respectively. 

Compared to placebo, placing chlorhexidine gel intrasocket after extractions reduced the odds of developing a dry socket by 58% with an OR of 0.44 (95% CI 0.27 to 0.71; P = 0.0008; 7 trials, 753 participants; moderate‐certainty evidence). The NNT with chlorhexidine gel (0.2%) to prevent one patient developing dry socket was 180 (95% CI 137 to 347), 37 (95% CI 28 to 72), and 7 (95% CI 5 to 15) for control prevalence of dry socket of 0.01, 0.05, and 0.30 respectively.

Compared to chlorhexidine rinse (0.12%), placing chlorhexidine gel (0.2%) intrasocket after extractions was not superior in reducing the risk of dry socket (RR 0.74, 95% CI 0.46 to 1.20; P = 0.22; 2 trials, 383 participants; low‐certainty evidence). 

The present review found some evidence for the association of minor adverse reactions with use of 0.12%, 0.2% chlorhexidine mouthrinses (alteration in taste, staining of teeth, stomatitis) though most studies were not designed explicitly to detect the presence of hypersensitivity reactions to mouthwash as part of the study protocol. No adverse events were reported in relation to the use of 0.2% chlorhexidine gel placed directly into a socket.

Platelet rich plasma in the prevention of dry socket 

Compared to placebo, placing platelet rich plasma after extractions was not superior in reducing the risk of having a dry socket (RR 0.51, 95% CI 0.19 to 1.33; P = 0.17; 2 studies, 127 participants; very low‐certainty evidence). 

A further 21 intrasocket interventions to prevent dry socket were each evaluated in single studies, and there is insufficient evidence to determine their effects.

Zinc oxide eugenol versus Alvogyl in the treatment of dry socket

Two studies, with 80 participants, showed that Alvogyl (old formulation) is more effective than zinc oxide eugenol at reducing pain at day 7 (mean difference (MD) ‐1.40, 95% CI ‐1.75 to ‐1.04; P < 0.00001; 2 studies, 80 participants; very low‐certainty evidence)

A further nine interventions for the treatment of dry socket were evaluated in single studies, providing insufficient evidence to determine their effects.

Authors' conclusions

Tooth extractions are generally undertaken by dentists for a variety of reasons, however, all but five studies included in the present review included participants undergoing extraction of third molars, most of which were undertaken by oral surgeons. There is moderate‐certainty evidence that rinsing with chlorhexidine (0.12% and 0.2%) or placing chlorhexidine gel (0.2%) in the sockets of extracted teeth, probably results in a reduction in dry socket. There was insufficient evidence to determine the effects of the other 21 preventative interventions each evaluated in single studies. There was limited evidence of very low certainty that Alvogyl (old formulation) may reduce pain at day 7 in patients with dry socket when compared to zinc oxide eugenol. 

Plain language summary

What treatments can be used to prevent and treat alveolar osteitis (dry socket)? 

Key messages

‐ Rinsing with chlorhexidine mouthwash before a dental extraction or beginning 24 hours after may help to prevent a dry socket.
‐ Placing a chlorhexidine gel directly into the socket immediately after tooth extraction may help to prevent a dry socket.
‐ Chlorhexidine rinses cause some minor adverse (unwanted) effects; chlorhexidine intrasocket gels do not appear to cause adverse effects.
‐ Alvogyl reduces the pain of a dry socket when compared to zinc oxide eugenol, but the evidence for this is very uncertain.
‐ Alvogyl does not appear to cause unwanted effects.
‐ We need future studies to strengthen the evidence and investigate the best ways to prevent and treat dry socket in all teeth.

What is dry socket?

Dry socket is a painful condition that sometimes arises after a tooth has been extracted and is more likely to occur following extraction of wisdom teeth in the lower jaw.

What causes dry socket?

It is thought to be linked to the loss of some or all of the blood clot that forms at the bottom of a tooth socket after a tooth is taken out.

How can we prevent dry socket?

An option for prevention of dry socket is to reduce debris and the bacterial load in the mouth, though dry socket is not bacterial in origin. People with poor oral hygiene (food debris and plaque) are at greater risk of dry socket. Improved oral hygiene and rinsing before a dental extraction or beginning 24 hours after may reduce the likelihood of dry socket.

How can we treat dry socket ?

Options for treating dry socket largely focus on reducing pain locally around the tooth extraction site by placing an obtundent (a soothing medicated dressing).

What did we want to find out ?

We wanted to find out if antiseptic rinses, gels, or healing patches could help to prevent dry socket. We also wanted to find out if placing a medicated dressing could treat a dry socket and whether any unwanted side effects were produced.

What did we do?

We searched for studies that compared antiseptic rinses or intrasocket gels with a placebo (dummy) rinse or nothing and a placebo (dummy) intrasocket gel or nothing.

To find the best way to treat dry socket we searched for studies that compared different soothing agents with a placebo (dummy) with other soothing agents or with nothing.

We compared the results of the studies and summarised the findings. We made an assessment of our confidence in the evidence based on the design of the study and the number of patients recruited.

What did we find?

We identified 49 trials; 39 trials (6219 participants) investigated prevention of dry socket and 10 trials (552 participants) investigated the treatment of dry socket.

We found that:

‐ rinsing both before and after tooth extraction (commencing 24 hours after extraction) with chlorhexidine gluconate rinse (at 0.12% and 0.2% strength) probably results in a reduction in dry socket;
‐ placing chlorhexidine gel (0.2% strength) in the socket of an extracted tooth probably results in a reduction in dry socket;
‐ chlorhexidine rinses and gels are equally effective at reducing dry socket, but the evidence for this comparison is very uncertain;
‐ chlorhexidine rinses produced some minor unwanted effects; chlorhexidine intrasocket gels appeared to produce no unwanted effects;
‐ there was a small amount of evidence of very low certainty from two studies that Alvogyl (old formulation) may reduce pain at day 7 in patients with dry socket when compared to zinc oxide eugenol. This evidence relates to the old formulation of Alvogyl which is no longer available. It should be noted that the formulation of Alvogyl has changed, it is now called Alveogyl. 

Tooth extractions are generally undertaken by dentists for a variety of reasons, however, all but five studies included in the present review included participants undergoing extraction of third molars, most of which were undertaken by oral surgeons. 

What are the limitations of the evidence?

The main limitation of the evidence are that studies:

‐ reported mostly on tooth extractions involving lower wisdom teeth;
‐ were undertaken in ways that introduced errors in the conduct of the study leading to errors in the results; and
‐ produced imprecise results when they were combined.

Due to these errors we have some confidence in the results relating to chlorhexidine rinses and gels but further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

How up to date is this evidence?

The search for existing studies was completed by 28 September 2021.

Summary of findings

Summary of findings 1. Chlorhexidine rinse versus placebo/no treatment for the prevention of dry socket.

Chlorhexidine rinse versus placebo/no treatment for the prevention of dry socket
Patient or population: patients having a tooth extraction Setting: primary or secondary care/home use Intervention: chlorhexidine rinse Comparison: placebo/no treatment control
Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo/no treatment controlb	Chlorhexidine rinse
Presence of dry socket (agreed diagnostic criteria. Follow‐up: median 7 days)	Study population		OR 0.38 (0.25 to 0.58)	1547 (6 studies)	⊕⊕⊕⊝ moderatea	‐
	Low
	10 per 1000	4 per 1000 (3 to 6)
	Moderate
	50 per 1000	20 per 1000 (13 to 30)
	High
	300 per 1000	140 per 1000 (97 to 199)
Adverse events	3 trials reported no adverse events. 1 study reported adverse events (alterations in taste, bad taste, staining) in 24/62 in chlorhexidine group versus 0/40 in placebo group, Peto OR = 8.42 (95% CI 3.31 to 21.39). A further study reported adverse events (bad taste, stomach upset) in 4/40 in chlorhexidine group versus 0/40 in placebo group, Peto OR = 20.87 (95% CI 8.38 to 52.02). Another study reported many adverse events in both experimental and placebo groups but only attributes 1 (stomatitis) to use of chlorhexidine
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds 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.

^aDowngraded once as 3 trials were assessed as being at high risk of bias and 1 at unclear risk of bias.
^bThe assumed risk values were set to reflect prevalence rates for routine dental extractions (1% and 5%) and for extraction of mandibular third molars (30%).

Summary of findings 2. Chlorhexidine gel versus placebo/no treatment for the prevention of dry socket.

Chlorhexidine gel versus placebo/no treatment for the prevention of dry socket
Patient or population: patients having a tooth extraction Setting: primary or secondary care/home use Intervention: chlorhexidine gel Comparison: placebo/no treatment control
Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed riskb	Corresponding risk
	Placebo/no treatment control	Chlorhexidine gel
Presence of dry socket (agreed diagnostic criteria. Follow‐up: median 7 days)	Study population		OR 0.44  (0.27 to 0.71)	753 (7 studies)	⊕⊕⊕⊝ moderatea	‐
	Low
	10 per 1000	5 per 1000 (3 to 7)
	Moderate
	50 per 1000	23 per 1000 (14 to 36)
	High
	300 per 1000	159 per 1000 (104 to 234)
Adverse events	No adverse events were reported
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds 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.

^aDowngraded once as 2 trials were assessed as being at low risk of bias, 2 at unclear risk, and the other 3 at high risk of bias.
^bThe assumed risk values were set to reflect prevalence rates for routine dental extractions (1% and 5%) and for extraction of mandibular third molars (30%).

Summary of findings 3. Chlorhexidine gel versus chlorhexidine rinse for the prevention of dry socket.

Chlorhexidine gel versus chlorhexidine rinse for the prevention of dry socket
Patient or population: patients having a tooth extraction Setting: primary or secondary care/home use Intervention: chlorhexidine gel  Comparison: chlorhexidine rinse
Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed riskb	Corresponding risk
	Chlorhexidine rinse	Chlorhexidine gel
Presence of dry socket (agreed diagnostic criteria. Follow‐up: median 7 days)	Study population		RR 0.74  (0.46 to 1.20)	383 (2 studies)	⊕⊕⊝⊝ lowa	‐
	Low
	10 per 1000	8 per 1000 (5 to 12)
	Moderate
	50 per 1000	37 per 1000 (23 to 60)
	High
	300 per 1000	222 per 1000 (138 to 360)
Adverse events	No adverse events were reported
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; 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.

^aDowngraded twice: once for 2 studies at high risk of bias and once for imprecision of estimate due to the CI including the possibility of a benefit for both treatment groups. 
^bThe assumed risk values were set to reflect prevalence rates for routine dental extractions (1% and 5%) and for extraction of mandibular third molars (30%).

Summary of findings 4. Platelet rich plasma versus placebo/no treatment for the prevention of dry socket.

Platelet rich plasma versus placebo/no treatment for the prevention of dry socket
Patient or population: patients having a tooth extraction Setting: primary or secondary care Intervention: platelet rich plasma Comparison: placebo/no treatment control
Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed riskb	Corresponding risk
	Placebo/no treatment control	Platelet rich plasma
Presence of dry socket (agreed diagnostic criteria. Follow‐up: median 7 days)	Study population		RR 0.51 (0.19 to 1.33)	127 (2 studies)	⊕⊝⊝⊝ very lowa	‐
	Low
	10 per 1000	6 per 1000 (2 to 14)
	Moderate
	50 per 1000	26 per 1000 (10 to 67)
	High
	300 per 1000	153 per 1000 (57 to 399)
Adverse events	No adverse events were reported
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; 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.

^aDowngraded 3 times: once for 2 studies at high risk of bias, once for low event rate, and once for the CI including the possibility of benefit for both treatment groups.
^bThe assumed risk values were set to reflect prevalence rates for routine dental extractions (1% and 5%) and for extraction of mandibular third molars (30%).

Summary of findings 5. Zinc oxide eugenol versus Alvogyl for the treatment of dry socket.

Zinc oxide eugenol versus Alvogyl for the treatment of dry socket
Population: adults with dry socket Setting: primary and secondary care Intervention: zinc oxide eugenol Comparison: Alvogyl
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Alvogyl	Zinc oxide eugenol
Healing	2 studies used different healing outcomes and we were unable to combine the data
Pain VAS (0‐10) Follow‐up: 7 days	The median score in the Alvogyl group was 1.7	The mean score in the zinc oxide eugenol group was 1.04 higher to 1.75 higher	‐	80 (2 studies)	⊕⊝⊝⊝ very lowa	This result was inconsistent with the 5‐day data
Swelling	No studies reported this outcome
Limitation of chewing or swallowing	No studies reported this outcome
Fever	No studies reported this outcome
Adverse effects	1 study (Supe 2018) reported 2 patients (8%) in the Alvogyl group and 9 patients (36%) in the zinc oxide eugenol group had delayed healing, classed as non‐healed sockets after 10 days
*The basis for the assumed risk is median score. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval.
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.

^aDowngraded 3 times as 2 studies at unclear risk of bias, heterogeneity, and differences in the results for day 5 versus day 7.  

Background

Teeth are routinely extracted in general dental practice because they are affected by tooth decay or periodontal disease. In spite of generalised overall improvements in oral health, it is estimated that European dentists in general dental practice extract up to seven teeth per week (McCaul 2001).  Alveolar osteitis (dry socket) is a complication that may follow tooth extraction, and is known to cause severe pain and repeated dental visits (Noroozi 2009; Veale 2015). The most widely used definition of alveolar osteitis was proposed by Blum, who defined it as postoperative pain in and around an extraction site, which increases in severity between 1 and 3 days after the extraction, accompanied by a partially or totally disintegrated blood clot within the alveolar socket, with or without halitosis (Blum 2002)

Description of the condition

Aetiology and incidence

There is variation in the reported incidence of dry socket; for routine extractions the incidence has been reported to range from 0.5% to 7% (Halabi 2018; Vezeau 2000). For surgical removal of impacted third molars the incidence has been reported to be anywhere from 1% to 37.5% (Caso 2005; Fridrich 1990; Haraji 2013; Oyri 2019; Vezeau 2000). Many studies have discussed risk factors for dry socket, but the exact underlying pathogenesis is not fully understood. There is an acceptance of Birn’s proposed theory of increased localised fibrinolysis leading to clot breakdown within the socket and consequently leading to dry socket (Birn 1973; Blum 2002). But the factors causing fibrinolysis are more ambiguous (Kolokythas 2010). Birn proposed that prolonged trauma or infection of the socket increases localised inflammation within the bone which triggers local release of plasminogen activators (Birn 1973). Nitzan showed in in vitro studies that certain bacteria, in particular Treponema denticola, demonstrate plasmin like fibrinolytic activity and can independently increase fibrinolysis (Nitzan 1983) though this has not been proven clinically. Birn's theory is given weight by many studies which have shown a clear link between prolonged or difficult extractions and an increased incidence of dry socket (Noroozi 2009). Surgical extractions, particularly of third molars, show a significant increase in incidence of dry socket over non‐surgical extractions (Chow 2020). A correlation has also been noted between operator inexperience and an increased incidence of dry socket (Christensen 2012; Larsen 1991). Additionally, bacteria seem to have a potential role to play with the incidence of dry socket shown to increase in patients with poor oral hygiene (Peñarrocha 2001) and pre‐existing  pericoronitis (Rud 1970). While differences in the microbial profiles between sockets with dry sockets and normal healing sockets have been shown (Aguilar‐Durán 2019; Shen 2019) the importance and relevance of this is unclear. 

Female gender and taking the oral contraceptive pill have been shown to increase incidence of alveolar osteitis, possibly due to the effect of oestrogen on the fibrinolytic system (Cohen 1995; Garcia 2003; Øyri 2020). The relationship between smoking and development of alveolar osteitis is also supported by the literature (Bortoluzzi 2012; Sweet 1978). Disruption of the blood clot through the use of straws is often mentioned, but there is no clear evidence to support this (Bloomer 2012).

Symptoms and diagnosis

Two of the key challenges when conducting this review were the multitude of terminologies used for dry socket and the classification of signs and symptoms that were accepted as determining the presence of dry socket. Dry socket was frequently conflated with an infected socket in some studies. A continuous throbbing pain that radiates to the ear, temple, and neck is the most common symptom of dry socket (Swanson 1989). Classically, this starts 1 to 3 days postextraction and may be accompanied by other signs and symptoms (e.g. foul taste, bad breath, localised swelling, and lymph‐node involvement) (Blum 2002; Chow 2020; Noroozi 2009; Vezeau 2000). The symptoms can persist for up to 10 days after extraction and may include pain so severe that it is not relieved by even the strongest of analgesic medications (Vezeau 2000).

Clinical history and examination are the principal methods of reaching a diagnosis. The clinical picture is of an extraction socket that is visually devoid of a blood clot exposing the bone within the socket, though there may be a greyish debris. But other causative factors for severe postoperative pain should be excluded such as infected retained roots or a surgical site infection (Blum 2002; Kolokythas 2010).

Prevention

There have been a number of theories as to the aetiology of dry socket and a range of preventative agents have been advocated according to the prevailing theory of causation at the time including: plaque control, antiseptic rinses, preoperative systemic antibiotics, and direct placement of medicaments into the socket (Caso 2005; Goldman 1973; Hall 1971; Hedstrom 2007; Kolokythas 2010; Noroozi 2009; Vezeau 2000). Several studies have reported that preoperative and postoperative antiseptic chlorhexidine rinses can be effective in reducing the incidence of dry socket (Berwick 1990; Halabi 2018; Hermesch 1998; Karabit 2019; Larsen 1991; Tjernberg 1979). Other studies have reported on the use of intrasocket antibiotic medicaments (Mitchell 1984; Reekie 2006; Torres‐Lagares 2006a; Trieger 1991; van Eeden 2006), low‐level laser therapy (Shafaee 2020), and intrasocket antifibrinolytic agents (Gersel‐Pedersen 1979; Ritzau 1977). Two Norwegian studies (Akota 1998; Oyri 2019) report significant reductions in the incidence of dry socket when using tetracycline drains postoperatively. Anecdotally, use of such drains are standard practice following removal of wisdom teeth in Norway. The latter studies have not been included in the present review as they do not meet the criteria for an intrasocket intervention and are reported in a separate Cochrane Review (Bailey 2020). Studies have also looked at the use of honey (Abu‐Mostafa 2019), herbal mouthwash (Divya 2019), and gaseous ozone (Ahmedi 2016). More recently the use of autologous platelet concentrates to improve postoperative healing has been advocated (Del Fabbro 2019) in particular platelet rich fibrin (PRF) (Unsal 2018). PRF is a second generation platelet concentrate, which is extracted from plasma through centrifugation of autologous blood, producing a fibrin matrix which contains concentrations of platelets, leucocytes, and growth factors (Unsal 2018). Two systematic reviews have concluded that PRF can reduce the incidence of alveolar osteitis (Xiang 2019; Zhu 2020). Studies have also reported on the use of flap design to minimise trauma and risk of dry socket (Bello 2011; Haraji 2010; Kirk 2007). 

The prophylactic use of systemic antibiotics is not generally advocated and there is a consensus that the latter should be reserved for individual patients reporting a history of multiple incidents of dry socket or for the immunocompromised patient (Epstein 2000; Fazakerley 1991; Lodi 2021).

Treatment

Forty‐five per cent of patients with dry socket require multiple postoperative visits, which could have significant consequences for the individual patient as well as societal costs including time off work (Nusair 2007; Vezeau 2000). One study found that patients with dry socket required up to four visits for management of their symptoms (Oyri 2019). Treatment options tend to focus on symptomatic relief, which may include the removal of debris from the socket by irrigation with saline or sterile local anaesthetic, and the use of analgesic medication (Blum 2002). Alternative options include the placement of intrasocket medicaments including antibacterials, topical anaesthetics and obtundents, or combinations of all three (Blum 2002). These intrasocket medications include zinc oxide and eugenol impregnated cotton pellets (Bloomer 2000; Chaurasia 2017), Alvogyl (eugenol, iodoform, and butamben) (Kaya 2011; Supe 2018), Alveogyl (eugenol only), Dentalone, bismuth subnitrate and iodoform paste (BIPP) on ribbon gauze, and metronidazole and lidocaine ointment (Silva 2006). Intrasocket dressings appear to provide an obtundent effect, but there has also been some suggestion that they may cause foreign body reactions and delay healing (Syrjänen 1979). Some studies have also reported the use of lasers for the treatment of dry socket (Jovanović 2011; Kaya 2011). More recently several studies have looked at the use of autologous platelet concentrates, such as plasma rich in growth factors (PRGF) (King 2018) and platelet rich fibrin (Yuce 2019) to promote healing.  

It should be noted that Alvogyl (Septodont) has been reformulated and rebranded as Alveogyl (Septodont). The newer version, Alveogyl, no longer contains iodoform (antimicrobial) or butamben (anaesthetic) (Kalsi 2020). It is difficult to establish the exact date when this product was reformulated, but it appears to have been prior to 2014. It is possible that some researchers using this product may have been unaware of the ingredient change as the packaging is almost identical (Kalsi 2020). In view of this, we only included studies in the analysis where the formulation of the product as either Alvogyl or Alveogyl could be confirmed with the authors and matched the timeframe for when these products were in use.

Description of the intervention

Prevention

There are two main approaches taken for prevention of dry socket. One is perioperative rinsing with a variety of mouthwashes from tap water to chlorhexidine and the other involves placement of intrasocket medicaments or autologous platelet concentrates.  

Treatment

There are also two approaches taken for treatment of dry socket. One is placement of obtundent or medicated dressings and the other is the use of autologous platelet concentrates.  

How the intervention might work

Prevention

There are two main approaches proposed to prevent development of dry socket. The first is to improve oral hygiene and reduce plaque, food debris, and bacterial load around extraction sockets by the use of antiseptic mouthrinses. Most studies focus on chlorhexidine which acts against a broad spectrum of aerobic and anaerobic bacteria, and has been shown to have an immediate bactericidal effect and prolonged bacteriostatic action (Larsen 1991). Chlorhexidine may also be applied to the socket in the form of a bioadhesive gel. Betadine, herbal mouthwashes, honey, and ozone gas have also been variously recommended for their antibacterial properties. Rinsing with tap water has also been suggested. 

The second approach is to promote socket healing. Autologous platelet concentrates release growth factors which are thought to improve the healing process. Various types of different tissue patches have also been suggested to support socket healing. 

Treatment 

Alveolar osteitis is generally managed rather than treated, with pain reduction as the primary aim, though promoting healing and reducing time taken for the patient to return to normal function such as eating and chewing is also important. There are a number of different intrasocket dressings such as zinc oxide eugenol (ZOE), Alvogyl/Alveogyl and Neocone. ZOE and Alvogyl/Alveogyl both contain eugenol which has an obtundent effect. Alvogyl also contains butamben, an anaesthetic and Neocone contains tetracaine. Some intrasocket dressings contain antimicrobial agents such as iodoform (Alvogyl) and polymyxin B (Neocone). The new formulation Alveogyl no longer includes iodoform or butamben (Kalsi 2020). These dressings may also act as a physical barrier against the entry of food debris into the socket. 

Autologous platelet concentrates, such as platelet rich growth factor (PRGF), release growth factors that may promote epithelisation of extraction sockets and therefore improve bone coverage. They may also suppress inflammation and have an antimicrobial effect (King 2018). PRGF has been shown to reduce pain scores possibly as a consequence of improved healing, or possibly due to its anti‐inflammatory effects (King 2018). 

Why it is important to do this review

Dry socket is a complication of dental extractions that is associated with severe pain and can result in an increase in postoperative visits. Prevention of dry socket as well as the effective management of its sequelae can help in reducing postoperative morbidity for the individual as well as societal costs, for example, lost time from work and healthcare costs. A systematic review of the current best evidence for the effects of the available interventions could help to inform clinical decision‐making for the prevention and management of dry socket.

The review question has previously been identified as clinically important by the public and key stakeholders (Worthington 2015), and is relevant as dental extractions remain an important part of dental practice (McCaul 2001). 

This systematic review will summarise the evidence of local interventions for the prevention and management of dry socket. Another Cochrane Review has summarised the evidence of the effects of systemic antibiotics prescribed to prevent infectious complications following tooth extraction and includes dry socket as one of the primary outcomes (Lodi 2021).  A further Cochrane Review has evaluated the evidence for surgical techniques (such as surgical drains, wound irrigation and different flap designs) for the removal of mandibular wisdom teeth, which also includes dry socket as a primary outcome (Bailey 2020). In order to avoid duplication, this review evaluates other 'local' interventions for the prevention and treatment of dry socket. This version is an update of the Cochrane Review first published in 2012 (Daly 2012).

Objectives

To assess the effects of local interventions used for the prevention and treatment of alveolar osteitis (dry socket) following tooth extraction.

Methods

Criteria for considering studies for this review

Types of studies

We only considered randomised controlled trials (RCTs) for inclusion in this review. RCTs looking at prevention could have employed a split‐mouth or cross‐over design. All studies included in this review utilised and reported explicit and validated criteria that were used in the diagnosis of dry socket. The diagnosis of dry socket was based on the Blum 2002 criteria (case definition) i.e. a continuous throbbing pain starting 1 to 3 days postextraction, a socket that may be partially or totally devoid of blood clot and which may be accompanied by other signs and symptoms such as foul taste, bad breath, localised swelling, and lymph‐node involvement. 

Types of participants

We considered studies that included adults over the age of 18 years who had undergone an extraction (routine or more complex surgical) of one or more permanent teeth under local anaesthesia with or without sedation or under general anaesthesia. We included studies that included participants who were smokers. We excluded participants who were immunocompromised, had any co‐morbidities or medical conditions that might influence the healing of oral tissues.

Types of interventions

We considered studies that included any type of local intervention used for the prevention and treatment of dry socket compared to a different local intervention, placebo or no treatment. We included studies that permitted the use of concomitant pain medication provided it was made available equally to both groups. We excluded studies that examined the effectiveness of local interventions to prevent dry socket and then subsequently to treat dry sockets. 

We excluded studies which reported upon the use of systemic antibiotics because these are covered in a separate Cochrane Review (Lodi 2021). Likewise, we excluded studies evaluating the use of different surgical procedures (including drains and lavage volume) to manage dry socket because these interventions are part of another Cochrane Review (Bailey 2020). 

Types of outcome measures

We considered studies that included outcome measures that were reported according to clinically important time points i.e. at the end of the intervention and during a follow‐up period of up to 2 weeks.

Adverse effects  

Any specific adverse effects related to any clinically diagnosed reactions to any of the active interventions were noted by the review team and reported as an additional table. 

Primary outcomes

For prevention of dry socket.

1. Proportion of participants presenting with a dry socket within 1 week post‐treatment.

For treatment of established dry socket.

1. Time to heal dry socket and socket healing indices.

2. Pain: its severity and duration from time of administration of intervention to relief of pain assessed using any patient‐reported validated pain scale.

3. Swelling: assessed using photography or digital morphometry.

4. Limitation of chewing or swallowing and time to resumption of normal feeding.

5. Fever.

Secondary outcomes

Secondary outcomes assessed were for the treatment of dry sockets.

1. Quality of life as assessed by a validated questionnaire.

2. Patient satisfaction assessed by any validated measure.

3. Costs.

Search methods for identification of studies

Electronic searches

Cochrane Oral Health’s Information Specialist conducted systematic searches in the following databases for RCTs and controlled clinical trials. There were no language, publication year, or publication status restrictions:

• Cochrane Oral Health’s Trials Register (searched 28 September 2021) (Appendix 1);

• Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 8) in the Cochrane Library (searched 28 September 2021) (Appendix 2);

• MEDLINE Ovid (1946 to 28 September 2021) (Appendix 3);

• Embase Ovid (1980 to 28 September 2021) (Appendix 4).

Subject strategies were modelled on the search strategy designed for MEDLINE Ovid. Where appropriate, they were combined with subject strategy adaptations of the highly sensitive search strategies designed by Cochrane for identifying RCTs and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions, Version 6.1 (Lefebvre 2021)).

Searching other resources

The following trial registries were searched for ongoing studies:

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov; searched 28 September 2021) (Appendix 5);

• World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 28 September 2021) (Appendix 6).

We searched the reference lists of included studies and relevant systematic reviews for further studies.

We checked that none of the included studies in this review were retracted due to error or fraud.

We did not perform a separate search for adverse effects of interventions used, we considered adverse effects described in included studies only.

Data collection and analysis

Data collection and analysis was carried out by two review authors working independently. Any disagreements were resolved by discussion.

Selection of studies

Two review authors independently assessed the abstracts of retrieved studies. The search was designed to be sensitive and include controlled clinical trials, these were filtered out early in the selection process if they were not randomised. We obtained full copies of any studies deemed to be relevant or potentially relevant i.e. those appearing to meet the inclusion criteria, or for which there was insufficient information in the title and abstract to make a clear decision. Two review authors then assessed full‐text papers independently and any disagreements on the eligibility of included studies were resolved through discussion and consensus. If necessary, a third review author was consulted.
We excluded any studies that did not match the inclusion criteria at this stage or at subsequent stages, and noted the reasons for exclusion in the section.

Data extraction and management

Two review authors collected study details and outcome data independently and in duplicate using a predetermined form designed for this purpose. These were entered into RevMan (RevMan Web 2022) (characteristics of included studies, forest plots, and additional tables). Discrepancies in data were discussed and only included if there was an independently reached consensus. If necessary a third review author was consulted to resolve inconsistencies. 

We extracted the following details.

1. Trial methods: method of allocation; masking of participants, operators, and outcomes; exclusion of participants after randomisation and proportion of losses at follow‐up and number analysed.

2. Participants: country of origin; sample size and sample size calculation; age; gender; inclusion and exclusion criteria.

3. Intervention and control: type and procedural information including dose; mode of local use; time of administration relative to extraction details of any other concomitant medication.

4. Outcomes: primary and secondary outcomes; methods of assessment and completeness of reporting as outlined in the Types of outcome measures section of this review.

This information was used to help assess the clinical diversity and generalisability of any included trials.

If stated, we recorded the sources of funding of any of the included studies.

Assessment of risk of bias in included studies

Studies identified for inclusion in this review were assessed independently by two review authors who graded them using the Cochrane risk of bias tool described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The independent evaluations were compared and discussed, and any disagreements were resolved. For cross‐over trials we referred to Higgins 2021 (Table 23.2.a) for guidance on managing assessment of risk of bias.

We assessed each trial for the following domains.

1. Random sequence generation (selection bias).

2. Allocation concealment (selection bias).

3. Blinding of participants and personnel (performance bias).

4. Blinding of outcome assessors (performance bias).

5. Completeness of outcome data.

6. Risk of selective outcome reporting.

7. Risk of other bias

For each domain a description of what occurred as reported in the journal article was reviewed and a judgement made on the risk of bias: high, unclear, or low risk of bias. The judgement was determined using guidance as described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021).

The assessments for each included study are reported in the corresponding section of the risk of bias tables in RevMan (RevMan Web 2022).

Overall risk of bias

• Low risk of bias: all domains are judged to be at low risk of bias.

• Unclear risk of bias: one or more domains judged to be at unclear risk of bias.

• High risk of bias: one or more domains judged to be at high risk of bias.

The overall risk of bias assessment was undertaken without blinding of review authors to the study authors' names or organisations or the journal type. The independent evaluations were compared and discussed, and any disagreements were resolved.

The results of the risk of bias assessment are presented graphically both by domain and by study.

Measures of treatment effect

The primary measure of intervention effect for the prevention of dry socket was the reduction in incidence of dry socket between the control and intervention groups i.e. proportion of participants presenting with a dry socket within 1 week post‐treatment. For the treatment of dry socket the primary measure of intervention effect was the reduction in the time to heal of the socket and reduction in the incidence of pain, swelling, functional limitation (chewing, swallowing, and time to resumption of normal feeding), and fever.

Secondary measures of intervention effect for the prevention and treatment of dry socket were the continuous outcomes: quality of life, patient satisfaction, and costs between the intervention group and the control.

For dichotomous data, we calculated the risk ratio (RR), together with the 95% confidence interval (CI). For split‐mouth/cross‐over studies, we calculated odds ratios (OR) using the Becker‐Balagtas method (BB OR) outlined in Curtin 2002. We chose the Becker‐Balagtas method because we intended to pool data from split‐mouth studies and parallel‐group studies in the same meta‐analyses, and the Becker‐Balagtas method facilitates this data synthesis (as outlined by Stedman 2011). The split‐mouth studies included in the review did not present the paired data by tooth pairs, only as marginals (as parallel‐group studies, not as cross‐classification), so we chose the conservative intraclass correlation coefficient (ICC) of 0.5 for the split‐mouth studies and 0 for the parallel‐group studies. For continuous outcomes, we used the mean differences (MD) and 95% CIs to summarise the data for each group where the mean difference and standard deviations were calculable from the data presented. 

Unit of analysis issues

In parallel‐group studies, we chose the individual to be the unit of analysis.

In split‐mouth studies, we chose the tooth pair within an individual to be the unit of analysis.

In cross‐over and split‐mouth studies, we planned to choose the tooth pair within an individual to be the unit of analysis.

Dealing with missing data

Where possible we calculated missing data from tables and graphs. We contacted authors of included studies to obtain missing trial details and data from the reports.

Assessment of heterogeneity

If a sufficient number of studies had been included in any meta‐analyses, we would have assessed clinical heterogeneity by examining the characteristics of studies and the similarity between types of participants, interventions, and outcomes as specified in the criteria for included studies.

We assessed the significance of discrepancies in estimates of treatment effects from various studies using Cochran's test for heterogeneity and the I² statistic. The I² statistic describes the percentage of variability in effect estimates that is due to heterogeneity rather than to sampling error. A value greater than 50% may represent substantial heterogeneity (Higgins 2003). The Cochrane Handbook for Systematic Reviews of Interventions also gives a rough guide to heterogeneity measured by I² as follows: 0% to 40% may not be important, 30% to 60% represents moderate heterogeneity, 50% to 90% may be classified as substantial heterogeneity, and 75% to 100% represents considerable heterogeneity (Higgins 2021).

In the event that there were insufficient clinically homogeneous trials for any specific intervention or insufficient study data that could be pooled, a narrative synthesis was presented.

Assessment of reporting biases

If sufficient trials had been identified for inclusion in this review, we would have assessed publication bias according to the recommendations on testing for funnel plot asymmetry (as described in Section 10.4.3.1 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021), and if asymmetry was identified, other possible causes would have been assessed. 

Data synthesis

Two review authors analysed the data and reported them as specified in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). Analysis was conducted at the same level as the allocation. The data for effects related to prevention were analysed and presented separately to those which considered treatment only.

The outcomes specified for this review necessitated repeated observations on the participants over a comparatively short period of time during and after the interventions. Therefore, depending on sufficient data being available, we grouped the outcomes and analysed them according to clinically important time points; at the end of the intervention and during the follow‐up period.

We undertook pooling of data to provide estimates of the efficacy of the interventions if included studies were clinically and statistically homogeneous. We used RRs to pool the dichotomous outcomes where possible. If split‐mouth studies were included in the meta‐analysis, then ORs were calculated as described by Stedman 2011 (Stedman 2011).

In general, for the synthesis of any quantitative data, we used the random‐effects model unless there were fewer than four studies, where we used the fixed‐effect model.

We calculated number needed to treat (NNT) for the pooled estimates using control prevalence rates for dry socket. Dry socket is a complication associated with 0.5% to 7% of routine extraction of teeth affected by periodontal disease and dental decay (Field 1988; Halabi 2018; Vezeau 2000), however, the prevalence of dry socket postextraction of mandibular molars, especially impacted wisdom teeth is much higher (1% to 37.5%) (Caso 2005; Fridrich 1990; Haraji 2013; Oyri 2019; Vezeau 2000). For the assumed risk of dry socket in the control group for the summary of findings table, we set prevalence rates to reflect prevalence rates for routine dental extractions (1% and 5%) and for extraction of mandibular third molars (30%).

Subgroup analysis and investigation of heterogeneity

The following subgroup analyses were planned if sufficient studies were identified: complexity of the extraction (surgical removal or simple extraction), and different types of teeth (third molars, molars, premolars). We also planned to undertake a subgroup analysis for chlorhexidine dose. There were insufficient studies to undertake subgroup analyses, however the data are presented for chlorhexidine dose. 

Sensitivity analysis

If a sufficient number of studies with similar characteristics had been included in the review, we would have undertaken sensitivity analyses to assess the robustness of the results by excluding studies at high risk of bias.

Summary of findings and assessment of the certainty of the evidence

We constructed summary of findings (SoF) tables for comparisons including more than one trial. Separate SoF tables were undertaken for the prevention and treatment of dry socket, with the primary outcome, the proportion of participants presenting with a dry socket within 1 week post‐treatment for prevention, and the following outcomes for treatment: time to heal dry socket, pain, swelling, limitation of chewing or swallowing, and fever. Adverse events were included as an outcome for all comparisons.

We followed GRADE methods to provide overall grading of the certainty of the evidence, with reference to overall risk of bias of included studies at each outcome, directness of the evidence, consistency of the results, precision of estimates, and risk of publication bias, as described in Section 14.2 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). The certainty of the evidence was assessed as high, moderate, low, or very low. 

Results

Description of studies

Results of the search

The search identified 1371 references through database searching and a further 11 identified through other sources. Figure 1 presents the PRISMA flow diagram for the review. After examination of the titles and abstracts, duplicates were removed and the remaining references eliminated when they did not match the inclusion criteria. Full‐text copies of the remaining 181 studies were obtained and these were then subjected to further evaluation. A number of studies were translated: three were in German (Birke 1970; Neugebauer 2004; Neuner 1969), two were in Russian (Butylin 1977; Zorina 2019), and one each was in Japanese (Anonymous 1966), Serbian (Jovanovic 2011), French (Turcotte 1997), and Polish (Banach 1973). There were seven in Chinese (Bai 2011; Feng 2009; Hu 2005; Huang 2011; Sun 2007; Wen 2004; Xue 2013). The bibliographical references were examined for all potentially eligible studies and six potentially relevant additional citations were identified (Al‐Hamed 2017; Babar 2012; Baslarli 2015; Delilbasi 2002; Haraji 2012b; Kirk 2007).

1.

Study flow diagram.

We included 49 studies in this review. See Characteristics of included studies. The majority (39 studies) evaluated interventions for the prevention of dry sockets (Abu‐Mostafa 2015; Abu‐Mostafa 2019; Ahmedi 2016; Alissa 2010; Bai 2011; Babar 2012; Cho 2018; Delilbasi 2002; Divya 2019; Feng 2009; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Halabi 2018; Haraji 2013; Hasheminia 2018; Hermesch 1998; Hita‐Iglesias 2008; Hu 2005; Huang 2011; Karabit 2019; Kjellman 1973; Larsen 1991; Metin 2006; Ragno 1991; Reekie 2006; Ritzau 1977; Rodriguez‐Perez 2013; Rubio‐Palau 2015; Shad 2018; Shi 2003; Sun 2007; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Trieger 1991; Tuk 2019; Unsal 2018; van Eeden 2006; Xue 2013). All the prevention trials comprehensively addressed the single (primary) outcome of whether a dry socket occurred or not.

The remaining 10 studies examined treatment strategies for dry sockets occurring after dental extraction (Burgoyne 2010; Chaurasia 2017; Faizel 2015; Kaya 2011; Keshini 2020; King 2018; Lenka 2019; Mitchell 1984; Supe 2018; Yuce 2019). The 10 treatment trials mostly looked at duration and severity of pain, and the time to heal dry sockets.

The search also retrieved 22 reviews: 14 with meta‐analysis (Barona‐Dorado 2014; Canellas 2020; Canellas 2019; Canellas 2017; Canullo 2020; Caso 2005; Del Fabbro 2019; Hedstrom 2007; Rodríguez Sánchez 2017; Shafaee 2020; Xiang 2019; Xu 2019; Yengopal 2012; Zhu 2020) and the remaining seven were narrative reviews (Chow 2020; Kolokythas 2010; Neuner 1969; Noroozi 2009; Turcotte 1997; Veale 2015; Vezeau 2000). All reviews were examined for potentially eligible studies.

Included studies

Forty‐nine studies were considered eligible for inclusion (Figure 1). There were 39 studies on prevention and 10 on the treatment of dry socket. Seven of the prevention studies were designed as split‐mouth studies (Ahmedi 2016; Gersel‐Pedersen 1979; Haraji 2013; Karabit 2019; Trieger 1991; Unsal 2018; van Eeden 2006), one was designed as a cross‐over study (Tuk 2019), and the remainder were of parallel‐group design. Trials in the treatment review will always be of parallel‐group design as there would be insufficient participants with more than one dry socket to undertake this type of study design.

Characteristics of the trial setting and investigators

Thirty‐three of the studies had been conducted in college/university/hospital oral surgery or dentistry departments: Abu‐Mostafa 2015 (Saudi Arabia); Abu‐Mostafa 2019 (Saudi Arabia); Ahmedi 2016 (Kosovo); Alissa 2010 (UK); Bai 2011 (China); Burgoyne 2010 (USA); Chaurasia 2017 (Nepal); Divya 2019 (India); Faizel 2015 (India); Feng 2009 (China); Freudenthal 2015 (Sweden); Gersel‐Pedersen 1979 (Denmark); Hita‐Iglesias 2008 (Spain); Huang 2011 (China); Hu 2005 (China); Karabit 2019 (Syria); Kaya 2011 (Turkey); King 2018 (UK); Kjellman 1973 (Sweden); Metin 2006 (Turkey); Mitchell 1984 (UK); Ritzau 1977 (Denmark); Rodriguez‐Perez 2013 (Spain); Rubio‐Palau 2015 (Spain); Shad 2018 (Pakistan); Sun 2007 (China); Supe 2018 (India); Torres‐Lagares 2006a (Spain); Torres‐Lagares 2006b (Spain); Tuk 2019 (Netherlands); Unsal 2018 (Turkey); Xue 2013 (China); and Yuce 2019 (Turkey). Military clinics were the settings for four of the studies (Babar 2012 (Pakistan); Hermesch 1998; Ragno 1991; van Eeden 2006). Two studies were conducted in private maxillofacial clinics (Haraji 2013 (Tehran) and Hasheminia 2018 (Iran)). Four multicentre studies were conducted, one was in three dental practices in the UK (Reekie 2006), one was conducted in three private dental practices in Australia (Cho 2018), one was conducted in two public dental clinics in Chile (Halabi 2018), and one was conducted in three oral and maxillofacial surgery departments of a university hospital, a hospital, and a private clinic in the Netherlands (Ghaeminia 2017). The settings of the remaining six studies were not stated (Delilbasi 2002; Keshini 2020; Larsen 1991; Lenka 2019; Shi 2003; Trieger 1991).

The skill level and number of operators providing care where stated varied. Single operators provided care in 12 of the included studies (Ahmedi 2016; Bai 2011; Babar 2012; Gersel‐Pedersen 1979; Hu 2005; Karabit 2019; Ragno 1991; Shad 2018; Sun 2007; Tuk 2019; Unsal 2018; Xue 2013). Either oral and maxillofacial surgeons, general dentists, general dentistry residents or senior house officers provided care for participants in 10 studies (Cho 2018; Freudenthal 2015; Ghaeminia 2017; Halabi 2018; Hermesch 1998; Kaya 2011; King 2018; Metin 2006; Rodriguez‐Perez 2013; Yuce 2019). Abu‐Mostafa 2015 and Abu‐Mostafa 2019 stated that dental interns or dental students under supervision of surgery instructors provided the care. Larsen 1991 stated that "multiple surgeons with varying levels of experience" with "formal training in third molar removal" were responsible for provision of care. Reekie 2006 stated that four general dental practitioners were providers of care. The providers of care for 23 of the included studies were not stated (Alissa 2010; Burgoyne 2010; Chaurasia 2017; Delilbasi 2002; Divya 2019; Faizel 2015; Feng 2009; Haraji 2013; Hasheminia 2018; Hita‐Iglesias 2008; Huang 2011; Keshini 2020; Kjellman 1973; Lenka 2019; Mitchell 1984; Ritzau 1977; Rubio‐Palau 2015; Shi 2003; Supe 2018; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Trieger 1991; van Eeden 2006).

Characteristics of the participants

Most prevention studies (29 out of 39) involved sockets of mandibular third molar teeth in adults. 21 studies reported on mandibular third molars which were extracted under local anaesthesia with/without intravenous sedation (Ahmedi 2016; Bai 2011; Babar 2012; Cho 2018; Feng 2009; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Hermesch 1998; Hita‐Iglesias 2008; Hu 2005; Huang 2011; Karabit 2019; Kjellman 1973; Larsen 1991; Rodriguez‐Perez 2013; Sun 2007; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Tuk 2019; Xue 2013). Three studies involved third molar teeth where the participants' ages were specified (Divya 2019; Hasheminia 2018; Shad 2018). A further three studies involved third molar teeth where the participants' ages were not specified (Ragno 1991; Shi 2003; Trieger 1991). Five studies investigated prevention of dry socket after mandibular third molar extraction, however, participants below the age of 18 years were enrolled in: Haraji 2013 (age 13 to 71 years); Metin 2006 (age 17 to 46 years); Ritzau 1977 (age 17 to 61 years); Unsal 2018 (age 15 to 43 years); and van Eeden 2006 (age 16 to 32 years). Two further studies investigated prevention of dry socket after mandibular third molar extraction, however, participant age was not stated (Delilbasi 2002; Rubio‐Palau 2015).

Abu‐Mostafa 2019 investigated prevention of dry socket after extraction of a single molar tooth in patients (age 17 to 69 years). Abu‐Mostafa 2015 also reported on prevention of dry socket following a single molar extraction. Halabi 2018 looked at patients aged 18 years and older having routine dental extractions. Reekie 2006 investigated prevention of dry socket after non‐surgical extraction of one or more molar/premolar teeth under local anaesthetic in adult participants (age 18 to 90 years), while Alissa 2010 also reported on teeth other than third molars.

In the treatment studies, Kaya 2011 reported on treatment of dry socket after mandibular third molar extraction, Yuce 2019 reported on third molars, Burgoyne 2010 reported on premolars and molars, Chaurasia 2017 reported on molars and mandibular canines, King 2018 reported on incisors, premolars and molars, Supe 2018 reported on all teeth, and no specific tooth was identified in Faizel 2015; Lenka 2019; and Mitchell 1984. Burgoyne 2010 investigated the treatment of diagnosed dry socket in 17‐ to 58‐year olds, Kaya 2011 investigated the treatment of diagnosed dry sockets in adults over 18 years of age, Supe 2018 investigated the treatment of dry socket in people aged 18 to 51 years, Yuce 2019 investigated the treatment of dry socket in people aged 18 to 40 years. Chaurasia 2017; Faizel 2015; and King 2018 investigated the treatment of dry sockets in males and females of unstated age, and Lenka 2019 and Mitchell 1984 investigated the treatment of diagnosed dry socket, however, age and gender were unspecified. Keshini 2020 reports on 30 participants presenting with dry socket, but the teeth were not identified.

The number of participants in the prevention studies ranged from 19 to 744 with a median of 100. The number of participants in the treatment studies ranged from 30 to 117 with a median of 47.

Characteristics of the interventions

This section is divided into two main parts: the characteristics of the interventions for the 39 studies reporting on the prevention of dry socket and the characteristics of the 10 trials reporting on the treatment of dry socket.

Prevention

Interventions in this section (39 studies) have been divided into two broad categories: rinses and intrasocket interventions.

Chlorhexidine (rinses or intrasocket gels)

• Chlorhexidine rinse (pre and postoperatively) versus placebo or saline (postoperatively) (Delilbasi 2002; Halabi 2018; Hermesch 1998; Karabit 2019; Larsen 1991; Ragno 1991).

• Chlorhexidine gel (placed in socket at time of surgery) versus placebo/no treatment (Babar 2012; Freudenthal 2015; Haraji 2013; Rubio‐Palau 2015; Shad 2018; Torres‐Lagares 2006a; Torres‐Lagares 2006b).

• Chlorhexidine rinse with monoject (postoperatively) versus chlorhexidine rinse (postoperatively) (Cho 2018).

• Chlorhexidine rinse (preoperatively) versus chlorhexidine rinse (postoperatively) (Metin 2006).

• Chlorhexidine gel versus chlorhexidine rinse (both postoperatively) (Abu‐Mostafa 2015; Hita‐Iglesias 2008).

• 0.2% chlorhexidine gel versus 1% chlorhexidine gel (Rodriguez‐Perez 2013).

Other (rinses or intrasocket interventions)

Two studies looked at platelet rich plasma and two studies at acellular dermal matrix patches. The remainder of the following interventions were evaluated by one trial only.

• Acellular dermal matrix patch versus no treatment (Bai 2011; Sun 2007).

• Apernyl versus placebo (Kjellman 1973).

• Artemisia desertorum Spreng (Shahaosan or Yunnan) versus placebo control (Shi 2003).

• Bovine fibroblastic growth factor versus control (Xue 2013).

• Clindamycin phosphate antibiotic solution patch versus saline patch (Trieger 1991).

• Gaseous ozone versus control (Ahmedi 2016).

• Glucocorticosteroid antibiotic agent versus normal saline (van Eeden 2006).

• Herbal mouthwash versus chlorhexidine rinse (Divya 2019).

• Heal‐all tissue patch (2 x 2.5 cm) versus no treatment (Huang 2011).

• Honey versus chlorhexidine rinse (Abu‐Mostafa 2019).

• Iodine (1%) (pre and intraoperatively) versus control (Hasheminia 2018).

• Iodine tampon versus monoject irrigation with saline (Tuk 2019).

• Irrigation with tap water in monoject versus control (Ghaeminia 2017).

• Metronidazole gel versus placebo gel (Reekie 2006).

• Oral tissue patch versus control (Hu 2005).

• P‐hydroxybenzoic acid versus placebo (Ritzau 1977).

• Platelet rich fibrin versus control (Unsal 2018).

• Platelet rich plasma versus control (Alissa 2010; Feng 2009).

• Tranexamic acid versus placebo (Gersel‐Pedersen 1979).

Treatment

Ten studies reported on the treatment of dry socket. Four studies compared management of dry socket with zinc oxide eugenol and Alvogyl/Alveogyl. It should be noted that Alvogyl (Septodont) has been reformulated and rebranded as Alveogyl (Septodont 2021). Alvogyl is described explicitly as containing iodoform 15.8 gm (antimicrobial) and butamben 25.7 gm (anaesthetic) and eugenol 13.7 gm by Faizel 2015; Kaya 2011; and Supe 2018. The manufacturers have since removed iodoform and butamben leaving eugenol (obtundent) as the only active ingredient. The new product is branded as Alveogyl. It is difficult to clarify exactly when this reformulation occurred but it seems to have been prior to 2014 (Kalsi 2020).

All the studies included in this review (except for Chaurasia 2017) state that they have used Alvogyl and clearly list the ingredients of this product with the exception of Lenka 2019 and King 2018. All included studies were published after 2014, but they may have been using existing stock of Alvogyl or possibly the trial subjects were recruited prior to or during 2014 (Faizel 2015; Kaya 2011; Supe 2018). One study refers in the text to "Alveogyl" but lists the ingredients of Alvogyl (Chaurasia 2017). It is possible that some clinicians may have been unaware of the change in formulation as the packaging of Alveogyl is almost identical to Alvogyl (Kalsi 2020). We have contacted the authors of all the included studies using these products to clarify but we have only received a response from the authors of King 2018.

• Zinc oxide eugenol versus Alvogyl (Faizel 2015; Lenka 2019; Supe 2018).

• Zinc oxide eugenol versus Alveogyl (Chaurasia 2017).

• Neocone versus Alvogyl versus zinc oxide eugenol (Faizel 2015).

• Advanced platelet rich fibrin versus saline (Yuce 2019).

• Alvogyl versus no treatment (Kaya 2011).

• Alvogyl versus plasma rich in growth factors (King 2018).

• Alvogyl versus plasma rich fibrin (Keshini 2020).

• Alvogyl versus SaliCept (Kaya 2011).

• Metronidazole versus placebo (Mitchell 1984).

• SaliCept versus no treatment (Kaya 2011).

• Topical anaesthetic gel (prilocaine‐lidocaine) versus eugenol (Burgoyne 2010).

Characteristics of the outcome measures

Prevention

The primary (and only) outcome measure for prevention was the presence/absence of a dry socket. This was clearly reported in all 39 studies for prevention. Minor adverse events were reported in eight of the prevention studies (Delilbasi 2002; Gersel‐Pedersen 1979; Hermesch 1998; Kjellman 1973; Metin 2006; Ragno 1991; Ritzau 1977; van Eeden 2006).

Treatment

Ten trials investigated the treatment of dry socket (Burgoyne 2010; Chaurasia 2017; Faizel 2015; Kaya 2011; Keshini 2020; King 2018; Lenka 2019; Mitchell 1984; Supe 2018; Yuce 2019) with the following: topical anaesthetic gel (prilocaine‐lidocaine), Alvogyl/Alveogyl, SaliCept, metronidazole, zinc oxide eugenol, Neocone, plasma rich in growth factors, platelet rich fibrin, and advanced platelet rich fibrin.

Three trials compared the use of Alvogyl and zinc oxide eugenol for the management of dry socket (Faizel 2015; Lenka 2019; Supe 2018). Supe 2018 and Faizel 2015 reported on experience of pain at day 5, 7, and 10 while Lenka reported pain at day 7 only (Lenka 2019). One trial (Chaurasia 2017) compared Alveogyl and zinc oxide eugenol, but data could not be included in the meta‐analysis as the authors appeared to have been using the newer formulation Alveogyl. One trial (Keshini 2020) compared Alvogyl versus platelet rich fibrin examining pain visual analogue scale (VAS) scores on day 1, 3, and 10. Although published in 2020 they report a formulation for Alvogyl postdating formula change.

A number of different outcome measures were used in the different studies. In all the studies pain was one of the main outcome measures, but the method of measuring pain varied. One study (Faizel 2015) measured time (in minutes and days) to achieve initial pain relief and complete resolution of pain following placement of a medicament. One study (Mitchell 1984) looked at duration of treatment. Seven studies (Burgoyne 2010; Chaurasia 2017; Keshini 2020; King 2018; Lenka 2019; Supe 2018; Yuce 2019) measured pain on subsequent days, following placement of a medicament using a VAS scale. The VAS scores were taken on different days in different studies.

Four studies (Faizel 2015; Keshini 2020; King 2018; Supe 2018) used different healing outcomes, including changes to signs of dry socket such as empty sockets, exposed sockets, number of socket walls that were exposed, redness around sockets, and also inflammation and healing scores.

The data from Kaya 2011 were unusable as medians and error bars for pain were presented in graphs. There was no evidence of a difference in pain at 48 hours in Burgoyne 2010 (Additional Table 6). Kaya 2011 (page 1574) stated in the text: "The differences in the changes in the clinical signs and symptoms between the control group and all 3 treatment groups were statistically significant (P < 0.05) on the third day after treatment" and "Regardless of the treatment the VAS scores changed during the follow‐up period (P < 0.001); however the intensity of the pain decreased more rapidly in all the treatment groups than for the control group (P < 0.05)."

1. Results for treatment of dry socket (single studies).

Comparison	Data	Effect (95% CI)	P value
Anaesthetic gel versus eugenol (Burgoyne 2010) Pain at 48 hours (VAS 0 to 10)	Anaesthetic gel n = 15, mean = 2.49, SD = 2.51 Placebo n = 20, mean = 2.69, SD = 2.46	Pain at 48 hours VAS (0 to 10) MD ‐0.20 (‐1.87 to 1.47)	0.81
Neocone versus Alvogyl (Faizel 2015) Mean time for initial pain relief (minutes) Mean time for complete resolution of pain (days)	Initial pain relief (minutes) Neocone n = 39, 17.23 minutes, SD = 1.6 Alvogyl n = 39, 7.358 minutes, SD = 1.79 Complete resolution of pain (days) Neocone n = 39, 4.85 days, SD = 0.63 Alvogyl n = 39, 6.47 days, SD = 0.45	Initial pain relief (minutes) 9.87 (9.12 to 10.63) Complete resolution of pain (days) ‐1.62 (‐1.86 to ‐1.38)	< 0.00001 < 0.00001
Neocone versus zinc oxide eugenol (Faizel 2015) Mean time for initial pain relief (minutes)  Mean time for complete resolution of pain (days)	Initial pain relief (minutes) Neocone n = 39, 17.23 minutes, SD = 1.6 Zinc oxide eugenol n = 39, 25 minutes, SD = 3.08 Complete resolution of pain (days) Neocone n = 39, 4.85 days, SD = 0.63 Zinc oxide eugenol n = 39, 8.64 days, SD = 0.47	Initial pain relief (minutes) ‐7.77 (‐8.86 to ‐6.68) Complete resolution of pain (days) ‐3.79 (‐4.04 to ‐3.54)	< 0.00001 < 0.00001
Alvogyl versus zinc oxide eugenol (Faizel 2015) Mean time for initial pain relief (minutes) Mean time for complete resolution of pain (days)	Initial pain relief (minutes) Alvogyl n = 39, 7.358 minutes, SD = 1.79 Zinc oxide eugenol n = 39, 25 minutes, SD = 3.08 Complete resolution of pain (days) Alvogyl n = 39, 6.47 days, SD = 0.45 Zinc oxide eugenol n = 39, 8.64 days, SD = 0.47	Initial pain relief (minutes) ‐17.64 (‐18.76 to ‐16.52) Complete resolution of pain (days) ‐2.17 (‐2.37 to ‐1.97)	< 0.00001 < 0.00001
Plasma rich in growth factors (PRGF) versus Alvogyl (King 2018) VAS pain score (cm)	1st review appointment day 3 (+/‐ 1 day) PRGF n = 22, 4.0 +/‐ 2.7 Alvogyl n = 22, 4.3 +/‐ 2.9 2nd review appointment day 7 (+/‐ 1 day) PRGF n = 22, 2.0 +/‐ 2.0 Alvogyl n = 22, 2.4 +/‐ 2.6	VAS pain score (cm) 1st review MD ‐0.30 (‐1.96 to 1.36) 2nd review MD ‐0.40 (‐1.77 to 0.97)	0.72 0.57
Plasma rich in growth factors (PRGF) versus Alvogyl (King 2018) Exposed bone (present versus absent) (percentage of patients)	1st review (day 3 +/‐ 1) PRGF n = 22, 9.1% Alvogyl n = 22, 9.1% 2nd review (day 7 +/‐ 1) PRGF n = 22, 0% Alvogyl n = 22, 22.7%	‐	1st review Not specified 2nd review < 0.05
Metronidazole versus placebo (Mitchell 1984) Duration of treatment (days)	Metronidazole n = 26, mean = 5.35, SD = 3.52 Placebo n = 29, mean = 8.52, SD = 8.52	Duration of treatment (days) MD ‐3.17 (‐1.04 to ‐5.30)	0.004
Advanced platelet rich fibrin (A‐PRF) versus control (Yuce 2019) Mean VAS pain scores (cm) Post‐operatively day 1	A‐PRF n = 20, 5.2 +/‐ 1.06 Control n = 20, 7.25 +/‐ 1.02	VAS pain score (cm) MD ‐2.05 (‐2.69 to ‐1.41)	< 0.00001
Advanced platelet rich fibrin (A‐PRF) versus control (Yuce 2019) Mean VAS pain scores (cm) Post‐operatively day 3	A‐PRF n = 20, 2.25 +/‐ 0.64 Control n = 20, 7.05 +/‐ 1.23	VAS pain score (cm) MD ‐4.80 (‐5.41 to ‐4.19)	< 0.00001
Advanced platelet rich fibrin (A‐PRF) versus control (Yuce 2019) Mean VAS pain scores (cm) Post‐operatively day 5	A‐PRF n = 20, 0.8 +/‐ 0.62 Control n = 20, 5.9 +/‐ 0.91	VAS pain score (cm) MD ‐5.10 (‐5.58 to ‐4.62)	< 0.00001
Advanced platelet rich fibrin (A‐PRF) versus control (Yuce 2019) Mean VAS pain scores (cm) Post‐operatively day 7	A‐PRF n = 20, 0.45 +/‐ 0.51 Control n = 20, 4.05 +/‐ 0.76	VAS pain score (cm) MD ‐3.60 (‐4.00 to ‐3.20)	< 0.00001

CI = confidence interval; MD = mean difference; n = number; SD = standard deviation; VAS = visual analogue scale.

Mitchell 1984 provided raw data on the duration of treatment and there was a statistically significant reduction in duration of treatment in the metronidazole group. 

Excluded studies

One hundred and thirty‐two studies were excluded from this review. The reasons for their exclusion are summarised below.

• Not a randomised controlled trial/inadequate method of sequence generation (55 studies): Annibali 2012; Anonymous 1966; Banach 1973; Bloomer 2000; Bloomer 2012; Birke 1970; Brignardello 2012; Butylin 1977; Cebi 2020; Christensen 2012; Cooper 2012; Field 1988; Fotos 1992; Garibaldi 1995; Goldman 1973; Goldsmith 2012; Goyal 2012; Hall 1971; Haraji 2012b; Johnson 1988; Jovanovic 2011; Julius 1982; Kamal 2020a; Kamal 2020b; Keskitalo 1973; Krekmanov 1986; Kudiyirickal 2012; Lao 2012; Liu 2011; Long 2012; Malkawi 2011; MacGregor 1975; Mishra 2012; Mitchell 1986a; Neugebauer 2004; Neuner 1969; Prataap 2017; Qi 2012; Rastogi 2018; Reeshma 2021; Ritzau 1978; Sanchis 2004; Sharma 2017; Sorensen 1987; Swanson 1989; Sweet 1985; Syrjanen 1981a; Tek 2014; Torres Lagares 2006; Tjernberg 1979; Tong 2012; Vedtofte 1974; Wen 2004; Yue 2012; Zanetta‐Barbosa 1994. 

• Dry socket not defined (18 studies): Anand 2015; Arakeri 2011; Arenaz‐Bua 2010; Asutay 2017; Daniels 2011; Dubovina 2016; Guazzo 2018; Haupt 2015; Hooley 1995; Kaplan 2020; Kirk 2007; MacGregor 1973; Majid 2010; Mehlisch 2010a; Mehlisch 2010b; Nordenram 1973; Sarkar 2019; Zuniga 2011.

• Irrelevant outcomes (1 study): Betts 1995.

• Wrong interventions, comparisons or other inclusion criteria (43 studies): Afat 2018; Ahmed 2020; Akota 1998; Al‐Hamed 2017; Altman 2011; Al‐Sukhun 2011; Baqain 2012; Baslarli 2015; Bello 2011; Berwick 1990; Bezerra 2011; Butler 1977; Bystedt 1980; Daugela 2018; Dutta 2016; Eshghpour 2014; Eshghpour 2018; Farooq 2019; Gonzalez‐Serrano 2021; Haraji 2010; Hill 2006; Jadhao 2018; Jesudasan 2015; Karthik 2021; Kilinc 2017; Kim 2020; Krekmanov 1981; Lopez‐Cedrun 2011; Mitchell 1986b; Olusanya 2011; Osunde 2014; Osunde 2015; Osunde 2017; Oyri 2019; Ozveri 2020; Rani 2016; Saez‐Alcaide 2020; Schatz 1987; Seethamsetty 2019; Syrjanen 1981b; Torres‐Lagares 2010; Wang 2013; Yuan 2006.

• Study not related to dry socket (5 studies): Jolley 1972; Krishnan 2020; Moberly 2007; Scopp 1967; Vu 2021.

• Study unavailable or abstract available with insufficient information (5 studies): study unobtainable: Nentwig 1985; Schlund 1968. Abstracts available with insufficient information: Njokanma 2019; Olson 1987; Pichler 2001.

• Study awaiting classification (1 study): Zorina 2019.

• Duplicate reporting (4 studies): Haraji 2012a; Haraji 2014; Haraji 2015; Paul 2019.

Further information about the reasons for exclusion of these studies is available in the Characteristics of excluded studies tables.

Risk of bias in included studies

Risk of bias assessment was undertaken for all included studies. For the prevention trials, the risk of bias assessment was undertaken for the primary outcome (whether or not the patient had a dry socket).

Allocation

Sequence generation 

Random sequence generation was assessed at low risk of bias in 24 studies (49%) (Abu‐Mostafa 2015; Abu‐Mostafa 2019; Alissa 2010; Burgoyne 2010; Chaurasia 2017; Cho 2018; Faizel 2015; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Haraji 2013; Hita‐Iglesias 2008; Karabit 2019; King 2018; Larsen 1991; Mitchell 1984; Reekie 2006; Rodriguez‐Perez 2013; Rubio‐Palau 2015; Sun 2007; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Trieger 1991; van Eeden 2006), and unclear in the remainder (25 studies; 51%).

Allocation concealment

Allocation concealment was considered to be low risk of bias in 16 trials (33%) (Alissa 2010; Cho 2018; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Halabi 2018; Hermesch 1998; Hita‐Iglesias 2008; King 2018; Mitchell 1984; Ragno 1991; Reekie 2006; Ritzau 1977; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Tuk 2019) and for the remainder of the studies it was deemed as either unclear (29 trials; 32%) or at high risk of bias (4 trials; 8%) (Abu‐Mostafa 2015; Abu‐Mostafa 2019; Haraji 2013; Karabit 2019).

Blinding

Blinding in relation to outcomes

As both the participants and outcome assessors' assessment form part of the diagnosis of dry socket, the blinding assessment was planned to include both (i.e. both groups must be blinded for this category to be assessed as being at low risk of bias). In studies where the operator was not blinded to group allocation, but the participant was blinded, it was difficult to assess the impact on performance bias. If this was unlikely to have resulted in a deviation from the intervention, it was likely to have a small impact on study outcome. 

Blinding was considered to be at low risk of bias for both performance and detection bias in 8 trials (17%) (Freudenthal 2015; Haraji 2013; Hermesch 1998; Mitchell 1984; Reekie 2006; Ritzau 1977; Torres‐Lagares 2006b; van Eeden 2006). Blinding (performance bias) was judged as being unclear in 33 trials (67%), at high risk of bias in 2 trials (4%), and at low risk of bias for 14 trials (29%). Blinding (detection bias) was judged as being unclear in 30 trials (61%), at high risk of bias in 3 trials (6%) and at low risk of bias in 16 trials (33%).

Incomplete outcome data

We assumed that dropouts in prevention of dry socket studies probably do not have dry socket or they would be returning for treatment. 35 (72%) of the trials were considered to be at low risk of bias with respect to incomplete outcome data (Ahmedi 2016; Bai 2011; Babar 2012; Delilbasi 2002; Divya 2019; Feng 2009; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Halabi 2018; Haraji 2013; Hasheminia 2018; Hermesch 1998; Hita‐Iglesias 2008; Hu 2005; Huang 2011; Karabit 2019; Kaya 2011; Keshini 2020; King 2018; Lenka 2019; Metin 2006; Mitchell 1984; Ragno 1991; Ritzau 1977; Rodriguez‐Perez 2013; Rubio‐Palau 2015; Shad 2018; Sun 2007; Torres‐Lagares 2006a; Torres‐Lagares 2006b; Tuk 2019; van Eeden 2006; Xue 2013; Yuce 2019). 10 trials (20%) were considered to be at unclear risk with respect to incomplete outcomes (Abu‐Mostafa 2019; Burgoyne 2010; Chaurasia 2017; Cho 2018; Faizel 2015; Kjellman 1973; Shi 2003; Supe 2018; Trieger 1991; Unsal 2018), and 4 studies (6%) were considered at high risk (Abu‐Mostafa 2015; Alissa 2010; Larsen 1991; Reekie 2006). 31 studies analysed the same number of participants as were enrolled (Abu‐Mostafa 2019; Ahmedi 2016; Bai 2011; Babar 2012; Burgoyne 2010; Delilbasi 2002; Divya 2019; Feng 2009; Freudenthal 2015; Gersel‐Pedersen 1979; Ghaeminia 2017; Halabi 2018; Haraji 2013; Hasheminia 2018; Hu 2005; Huang 2011; Karabit 2019; Kaya 2011; King 2018; Kjellman 1973; Lenka 2019; Metin 2006; Ragno 1991; Ritzau 1977; Sun 2007; Supe 2018; Torres‐Lagares 2006a; Tuk 2019; van Eeden 2006; Xue 2013; Yuce 2019). Where reported, the range in the number of dropouts was from 3 (Hita‐Iglesias 2008) to 11 (Larsen 1991). One study only assessed participants who returned with pain, so it was assumed that the other participants who did not return did not have dry socket (Reekie 2006).

Selective reporting

Only the reporting of dry socket was considered for this item for the prevention trials. Most trials reported this well and were considered at low risk of bias. Six trials were unclear in their reporting (Abu‐Mostafa 2015; Faizel 2015; Reekie 2006; Shi 2003; Supe 2018; Trieger 1991) and three were considered to be at high risk of reporting bias (Ahmedi 2016; Hermesch 1998; Ragno 1991). Hermesch 1998 only reported dry socket for extracted mandibular third molars although non‐mandibular third molars were also extracted concurrently. Ragno 1991 did not report any data from the questionnaire completed by participants on day 7.

Other potential sources of bias

Thirty‐five of the trials (71%) were considered to be at low risk of bias from other sources. 6 trials (12%) were deemed unclear in this respect (Faizel 2015; Keshini 2020; King 2018; Reekie 2006; Shi 2003; Trieger 1991). The reporting in Trieger 1991 in general was very poor and it was not possible to make a clear judgement in many domains. 8 of the trials (16%) in this review were deemed to be at a high risk of bias from other sources. Of the high risk trials, investigators in Hermesch 1998 randomised at an individual participant level but subsequently analysed participants at an extraction site level; similarly Larsen 1991 randomised individuals, however, subsequent analyses were at tooth level.

Overall assessment of bias

All domains had to be assessed as being at low risk of bias for a study to be considered low risk of bias. If any domain was assessed as being at high risk of bias, the study was assessed as high risk of bias, the remainder were assessed as unclear. Figure 2 presents the review authors' judgements about each risk of bias item presented as percentages across all included studies and Figure 3 presents review authors' judgements about each risk of bias item for each included study.

2.

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

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Three studies were assessed as being at low risk of bias overall (Freudenthal 2015; Mitchell 1984; Torres‐Lagares 2006b), 30 were deemed unclear (Burgoyne 2010; Chaurasia 2017; Cho 2018; Delilbasi 2002; Divya 2019; Faizel 2015; Feng 2009; Gersel‐Pedersen 1979; Ghaeminia 2017; Halabi 2018; Hasheminia 2018; Hu 2005; Kaya 2011; Keshini 2020; King 2018; Kjellman 1973; Lenka 2019; Ritzau 1977; Rodriguez‐Perez 2013; Rubio‐Palau 2015; Shad 2018; Shi 2003; Sun 2007; Supe 2018; Trieger 1991; Tuk 2019; Unsal 2018; van Eeden 2006; Xue 2013; Yuce 2019), and the 16 remaining studies were deemed as being at high risk of bias overall (Abu‐Mostafa 2015; Abu‐Mostafa 2019; Ahmedi 2016; Alissa 2010; Babar 2012; Bai 2011; Haraji 2013; Hermesch 1998; Hita‐Iglesias 2008; Huang 2011; Karabit 2019; Larsen 1991; Metin 2006; Ragno 1991; Reekie 2006; Torres‐Lagares 2006a). 

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5

Prevention

Forest plots and summary of findings tables have only been included when there was more than one study for a specific comparison.

Comparison 1: chlorhexidine rinse versus placebo 

Table 1.

Primary outcome: presence of dry socket

Six trials involving 1547 participants compared rinsing with chlorhexidine at 0.12% concentration (Halabi 2018; Hermesch 1998; Karabit 2019; Larsen 1991; Ragno 1991) and 0.2% concentration (Delilbasi 2002), both pre and postextraction, to rinsing with a placebo for the prevention of dry socket. One trial (Karabit 2019) was of a split‐mouth design, so generic inverse variance was used to calculate the odds ratio (OR) as described in Stedman 2011. The overall pooled estimate of the six trials showed a benefit in rinsing with chlorhexidine to prevent dry socket with an overall OR of 0.38 (95% confidence interval (Cl) 0.25 to 0.58; 6 studies, 1547 participants; moderate‐certainty evidence). There was moderate heterogeneity (I² = 48%). As there is only a single study using 0.2% chlorhexidine, the comparison between subgroups is inappropriate. The number of patients needing to use chlorhexidine rinse to prevent one patient developing dry socket (number needed to treat (NNT)) was 162 (95% CI 155 to 240), 33 (95% CI 27 to 49), and 7 (95% CI 5 to 10) for control prevalence of dry socket 0.01, 0.05, and 0.30 respectively (Analysis 1.1).

1.1. Analysis.

Comparison 1: Prevention of dry socket, Outcome 1: Chlorhexidine rinse versus placebo

Adverse events 

Additional Table 7 presents reports of any adverse events observed. Three trials, with 1137 participants, reported no adverse events (Halabi 2018; Karabit 2019; Larsen 1991). Two studies reported alteration in taste and bad taste (Delilbasi 2002; Ragno 1991). Delilbasi 2002 found 15% (18/118) of participants using chlorhexidine reported alterations in taste. Ragno 1991 reported that 7.5% (3/40) of participants using chlorhexidine complained of a bad taste from the rinse. One study also reported that 10% of participants (10/118) had staining of their teeth/dentures (Delilbasi 2002). One study (Hermesch 1998) reported many adverse events in both the experimental group and the control group, with 16.2% of the subjects in the chlorhexidine group and 23% of the placebo group reporting a range of adverse events, including paraesthesia, headaches, gingivitis, pain, and pharyngitis (see Additional Table 7 for full details). The study authors can only attribute a single case of stomatitis to the use of chlorhexidine (Hermesch 1998). 

2. Adverse events reported in included studies for prevention of dry socket.

Author/study	Intervention	Adverse events
Abu‐Mostafa 2015	0.2% chlorhexidine gel versus 0.12% chlorhexidine rinse. 0.12% chlorhexidine rinse group rinsed twice daily for 1 week. Chlorhexidine gel group had intrasocket application of 0.2% chlorhexidine gel immediately postoperatively and also on day 3. All patients given ibuprofen 600 mg every 8 hours for 3 days	No adverse events reported
Abu‐Mostafa 2019	Honey versus 0.2% chlorhexidine rinse. Patients in the honey group had Manuka honey applied topically to the socket immediately postoperatively and also on day 3. Patients in the chlorhexidine group rinsed with 0.2% chlorhexidine twice daily for 7 days. All patients instructed to take 400 mg ibuprofen every 8 hours on 1st and 2nd day postoperatively	No adverse events reported
Ahmedi 2016	Gaseous ozone. In the experimental group intra‐alveolar O3 was used to irrigate the socket for 12 seconds prior to suturing. All patients given 400 mg ibuprofen	No adverse events reported
Alissa 2010	Platelet rich plasma. Patients all given Co‐codamol to take postoperatively if needed	No adverse events reported
Bai 2011	Acellular dermal matrix	No adverse events reported
Babar 2012	0.2% chlorhexidine gel. In experimental group 0.2% chlorhexidine gel was placed into the socket immediately postoperatively. All patients received 400 mg ibuprofen	No adverse events reported
Cho 2018	Irrigation of the surgical site with 0.2% chlorhexidine versus 0.2% chlorhexidine mouthrinse. The irrigation group used 0.2% chlorhexidine in a monoject syringe to irrigate the socket twice daily for 7 days. The control group used 0.2% chlorhexidine as a mouthrinse twice daily for 7 days. All patients prescribed paracetamol/codeine 500/15 mg 1 to 2 tablets every 4 to 6 hours and 200 mg ibuprofen 1 to 2 tablets every 4 to 6 hours	No adverse events reported
Delilbasi 2002	0.2% chlorhexidine rinse. Paracetamol for postoperative pain relief. Patients rinsed with 15 ml of chlorhexidine solution "just before tooth removal" (page 302). Intraoperatively the surgical site was irrigated with 15 ml of chlorhexidine solution diluted with 15 ml of saline. The day after surgery patients began rinsing with 15 ml chlorhexidine, twice daily for 7 days	Adverse events reported for Group 1 n = 62 Allergy: n = 0 Staining of teeth: n = 4 Mucosal irritation: n = 0 Alteration in taste: n = 12 Gastrointestinal complaints: n = 0 Bad taste: n = 8 No adverse reactions: n = 38 (From Table IV, page 303)
Divya 2019	Herbal mouthwash versus 0.2% chlorhexidine mouthrinse. Each group rinsed with their allotted mouthrinse twice daily for 7 days	No adverse events reported
Feng 2009	Oral tissue patch versus platelet rich plasma versus oral tissue patch + platelet rich plasma versus control group	No adverse events reported
Freudenthal 2015	0.2% chlorhexidine gel versus placebo gel. Patients had the allocated gel placed in the socket prior to suturing. All patients prescribed paracetamol and paracetamol/codeine	No adverse events reported
Gersel‐Pedersen 1979	Trans‐4‐amino‐methyl‐cyclohexane acid (AMCA) cones versus lactose cones. All had 0.2% chlorhexidine 3 times a day	Foreign body reaction to the vehicle delivery system in the cones
Ghaeminia 2017	Tap water. Experimental group was instructed to irrigate extraction socket with tap water in a monoject syringe 4 times daily. All patients 1 g paracetamol 4 times daily and ibuprofen 600 mg 3 times daily	No adverse events
Halabi 2018	0.12% chlorhexidine mouthrinse. Experimental group rinsed with 0.12% chlorhexidine mouthrinse twice daily for 7 days. The control group rinsed with sterile water	No adverse events reported
Haraji 2013	0.2% chlorhexidine gel versus dry dressing. Experimental group had a gelatin sponge saturated in 0.2% chlorhexidine gel packed into the socket prior to suturing	No adverse events reported
Hasheminia 2018	Povidone iodine 1% mouthwash. Experimental group had gauze soaked in 1% iodine placed over teeth preoperatively and also 1% iodine was used intraoperatively. All patients 400 mg ibuprofen every 6 hours for 3 days	No adverse events reported
Hermesch 1998	0.12% chlorhexidine rinse. Participants in the experimental group rinsed for 30 seconds twice per day with 15 ml chlorhexidine 7 days preoperatively. On day of surgery supervised rinse before anaesthesia and surgery. Participants suspended rinsing for the remainder of the day and recommenced the next day	Adverse events: Group 1 chlorhexidine n = 136 and Group 2  placebo n = 135 Paraesthesia reported: chlorhexidine 9, placebo 5 Infection: chlorhexidine 4, placebo 3 Trismus: chlorhexidine 0, placebo 5 Gingivitis: chlorhexidine 1, placebo 3 Glossitis: chlorhexidine 2, placebo 2 Abnormal healing: chlorhexidine 1, placebo 3 Nausea: chlorhexidine 0, placebo 4 Sinusitis: chlorhexidine 1, placebo 3 Headache: chlorhexidine 1, placebo 2 Dysphagia: chlorhexidine 0, placebo 2 Oedema (head and neck): chlorhexidine 1, placebo 1 Haemorrhage (prolonged): chlorhexidine 1, placebo 1 Pain: chlorhexidine 1, placebo 1 Pharyngitis: chlorhexidine 1, placebo 1 Rash: chlorhexidine 1, placebo 1  There were single site observations of each of the following: asthenia, bronchitis, cyst, depression, contact dermatitis, dyspepsia, ecchymosis, fever, herpes simplex, hypalgesia, back pain, rhinitis, stomatitis, and tenosynovitis. All but 2 cases of paraesthesia had resolved by end of study, 1 in each treatment group (From Table VII, page 384)
Hita‐Iglesias 2008	Chlorhexidine 0.2% gel versus 0.12% chlorhexidine mouthrinse. Participants in the 0.2% chlorhexidine gel group had the gel placed in the socket during surgery and then they were required to apply the gel to the socket twice a day (morning and night‐time) for 7 days beginning on the same day as the surgery. Patients in the rinse group rinsed twice a day (morning and night‐time) for 7 days beginning on the same day as the surgery	No adverse events reported
Hu 2005	Oral tissue patch	No adverse events reported
Huang 2011	Intradermal matrix	No adverse events reported
Karabit 2019	0.12% chlorhexidine mouthrinse versus aqua distillate with mint flavour. Patients rinsed with their allocated mouthrinse on twice daily on the day before the extraction and for 7 days postoperatively	No adverse events reported
Kjellman 1973	Apernyl as alveolar	Pain and burning sensation (page 200)
)Larsen 1991	0.12% chlorhexidine rinse versus placebo. All patients received 8 mg dexamethasone (glucocorticoid) IV prior to surgery. Participants were required to rinse twice per day for 30 seconds using 15 ml of the solution for 7 days prior to the surgery. On the day of the surgery they rinsed with the solution immediately prior to surgery (using 15 ml) and postoperatively patients were instructed to begin rinsing the day following surgery	No adverse reactions (page 954)
Metin 2006	0.2% chlorhexidine rinse 7 days preoperatively and 7 days postoperatively (group I) versus 7 days postoperatively only (group II)	Altered taste and numbness (page 3). Numbness in the tongue reported in group I and group II 45.6% and 13.2%. However, disturbance of taste sensation was seen in 56.5% of the patients in group I and in 11.3% of the patients in group II
Ragno 1991	0.12% chlorhexidine rinse. Participants rinsed immediately before surgery, the surgical was irrigated intraoperatively and starting the day after the surgery participants were asked to rinse with solution twice daily (15 ml) for 7 days postoperatively	There were no allergic reactions to the chlorhexidine rinse. 3 participants reported bad taste, 1 reported stomach upset (page 526) but no staining noted. 1 person in the control had a severe surgical reaction which in the author's opinion was not attributable to the medication
Reekie 2006	0.25 ml of 25% metronidazole gel intrasocket intervention	1 participant with nausea and vomiting, 2 complained of a bitter taste
Ritzau 1977	PEPH (p‐hydroxybenzoic acid)	Haematoma and rash
Rodriguez‐Perez 2013	1% chlorhexidine gel versus 0.2% chlorhexidine gel. All patients had 0.2% chlorhexidine gel placed into the socket prior to suturing. The patients also applied either 0.2% or 1% chlorhexidine gel to the wound twice daily for 7 days postoperatively. All patients 600 mg ibuprofen every 8 hours and 1 g paracetamol every 12 hours	No adverse events reported
Rubio‐Palau 2015	0.2% chlorhexidine gel versus placebo gel. Experimental group had 0.2% chlorhexidine gel placed in extraction socket prior to suturing. Placebo gel was placed in socket in control group. All patients given diclofenac 50 mg every 8 hours alternated with metamizole 575 mg every 8 hours and omprazole 20 mg per day	30 (18%) patients gastrointestinal discomfort 10 (6.2%) patients dizziness 2 (1.2%) patients skin rash  Study authors attribute these adverse events to the use of the analgesic metamizole (dipyrone) No adverse events reported in relation to chlorhexidine gel
Shad 2018	0.2% chlorhexidine gel versus placebo gel. Either 0.2% chlorhexidine gel or placebo gel was placed in socket prior to suturing	No adverse events reported
Shi 2003	Artemisia desertorum Spreng (Shahaosan, Yunnan)	No adverse events reported
Sun 2007	Acellular dermal matrix	No adverse events reported
Torres‐Lagares 2006a	0.2% chlorhexidine gel intrasocket versus placebo. Participants in the 0.2% chlorhexidine gel group had the gel placed in the socket during surgery. Note only 1 application of bioadhesive gel during surgery	No adverse events reported
Torres‐Lagares 2006b	0.2% chlorhexidine gel intrasocket versus placebo. Participants in the 0.2% chlorhexidine gel group had the gel placed in the socket during surgery. Note only 1 application of bioadhesive gel during surgery	No adverse events reported
Trieger 1991	Gelfoam clindamycin	No adverse events reported
Tuk 2019	1 x 2 cam iodine tampon placed in surgical site in experimental group. All patients used 0.12% chlorhexidine mouthrinse twice daily for 7 days postoperatively	No adverse events reported
Unsal 2018	Platelet rich fibrin	No adverse events reported
van Eeden 2006	Gelfoam covomycin. 0.2% chlorhexidine advised postoperatively every 6 hours for 5 days	Some events may be attributable to intervention or could be normal sequelae of operation
Xue 2013	Recombinant bovine fibroblast growth factor gel	No adverse events reported

n = number.

Secondary outcomes

Quality of life 

No studies reported this outcome.

Patient satisfaction 

No studies reported this outcome.

Costs

No studies reported this outcome.

Comparison 2: chlorhexidine gel versus placebo/no treatment 

Table 2.

Primary outcome: presence of dry socket

Seven trials compared placing chlorhexidine gel in the extraction socket versus placebo dressing or no treatment (Babar 2012; Freudenthal 2015; Haraji 2013; Rubio‐Palau 2015; Shad 2018; Torres‐Lagares 2006a; Torres‐Lagares 2006b). One trial (Haraji 2013) was of a split‐mouth design, so generic inverse variance was used to calculate the OR as described in Stedman 2011. The meta‐analysis showed the use of chlorhexidine gel in extraction sockets reduced the odds of dry socket (OR 0.44, 95% CI 0.27 to 0.71; P = 0.0008; 7 studies, 753 participants; moderate‐certainty evidence). There was moderate heterogeneity within the data (I² = 44%). The number of patients needing to have chlorhexidine gel placed in an extraction socket to prevent one dry socket (NNT) was 180 (95% CI 137 to 347), 37 (95% CI 28 to 72), and 7 (95% CI 5 to 15) for control prevalences of dry socket of 0.01, 0.05, and 0.30 respectively (Analysis 1.2). 

1.2. Analysis.

Comparison 1: Prevention of dry socket, Outcome 2: Chlorhexidine gel versus placebo/no treatment

Adverse events 

Six trials, with 673 participants reported no adverse events from the use of chlorhexidine gel. One study reported that 30 (18%) patients experienced gastrointestinal discomfort, 10 (6.2%) patients dizziness, and 2 (1.2%) patients had a skin rash (Rubio‐Palau 2015). The study authors attribute these adverse events to the use of the analgesic metamizole (dipyrone) and they report no adverse events in relation to the use of chlorhexidine gel. 

Secondary outcomes

Quality of life 

No studies reported this outcome.

Patient satisfaction 

No studies reported this outcome.

Costs

No studies reported this outcome.

Comparison 3: chlorhexidine gel versus chlorhexidine rinse 

Table 3.

Primary outcome: presence of dry socket

Two studies, with 383 participants, compared postoperative rinsing with chlorhexidine (0.12%) with application of chlorhexidine gel (0.2%) both intraoperatively and also postoperatively (Abu‐Mostafa 2015; Hita‐Iglesias 2008). The meta‐analysis shows no benefit in using chlorhexidine gel over chlorhexidine rinse. The risk ratio (RR) was 0.74 (95% CI 0.46 to 1.20; P = 0.22; 2 studies, 383 participants; low‐certainty evidence), with substantial heterogeneity (Chi² = 2.39, degrees of freedom (df) = 1; P = 0.12; I² = 58%) (Analysis 1.3).

1.3. Analysis.

Comparison 1: Prevention of dry socket, Outcome 3: Chlorhexidine gel versus chlorhexidine rinse

Adverse events

No adverse events were reported in these two trials (Abu‐Mostafa 2015; Hita‐Iglesias 2008).

Secondary outcomes

Quality of life 

No studies reported this outcome.

Patient satisfaction 

No studies reported this outcome.

Costs

No studies reported this outcome.

Comparison 4: platelet rich plasma versus placebo

Table 4.

Primary outcome: presence of dry socket

Two studies compared placement of platelet rich plasma in extraction sockets following extractions compared with leaving the sockets empty (Alissa 2010; Feng 2009). Platelet rich plasma was not superior (or inferior) to placebo (RR 0.51, 95% CI 0.19 to 1.33; P = 0.17; 2 studies, 127 participants; very low‐certainty evidence). There was no evidence of any heterogeneity within this sample (I² = 0%) (Analysis 1.4).

1.4. Analysis.

Comparison 1: Prevention of dry socket, Outcome 4: Platelet rich plasma versus placebo

Adverse events

No adverse events were reported in these two trials.

Secondary outcomes

Quality of life 

No studies reported this outcome.

Patient satisfaction 

No studies reported this outcome.

Costs

No studies reported this outcome.

Comparisons for prevention of dry socket, only including a single study

The remaining 21 interventions to prevent dry socket were each evaluated in single studies as described below, and there is insufficient evidence to determine their effects. The results from these studies are shown in Additional Table 8. 

3. Results for prevention of dry socket (single studies).

Comparison	Data	RR/OR (95% CI)	P value
Honey versus 0.2% chlorhexidine rinse (Abu‐Mostafa 2019)	Honey 7/57 0.2% chlorhexidine rinse 4/43	RR 1.32 (0.41 to 4.22)	0.64
Gaseous ozone versus placebo (Ahmedi 2016)	Gaseous O3 1/30 Placebo 5/30	RR 0.20 (0.02 to 1.61)	0.13
Acellular dermal matrix patch versus no treatment (Bai 2011)	Patch 1/200 No patch 15/200	RR 0.07 (0.01 to 0.50)	0.008
Irrigation of the surgical site with 0.2% chlorhexidine in a monoject versus 0.2% chlorhexidine mouthrinse (Cho 2018)	Monoject with chlorhexidine 4/47 Chlorhexidine rinse 16/48	RR 0.26 (0.09 to 0.71)	0.009
Herbal mouthrinse versus 0.2% chlorhexidine mouthrinse (Divya 2019)	Herbal 1/25 Chlorhexidine 2/25	RR 0.50 (0.05 to 5.17)	0.56
Tranexamic acid versus placebo (Gersel‐Pedersen 1979)	Split‐mouth AMCA Y, placebo Y = 3 AMCA Y, placebo N = 6 AMCA N, placebo Y = 3 AMCA N, placebo N = 108	RR 0.67 (0.30 to 1.48)	0.51
Irrigating with monoject syringe filled with tap water versus no rinsing (Ghaeminia 2017)	Monoject with tap water 9/158 No rinsing 28/178	RR 0.36 (0.18 to 0.74)	0.006
Povidone iodine 1% versus control (Hasheminia 2018)	Iodine 5/97 Control 13/92	RR 0.36 (0.14 to 0.98)	0.05
Oral tissue patch versus placebo (Hu 2005)	Tissue patch 1/100 Placebo 8/100	RR 0.13 (0.02 to 0.98)	0.05
Heal‐all tissue patch versus no treatment (Huang 2011)	Patch 0/40 No patch 6/40	RR 0.08 (0.00 to 1.32)	0.08
Apernyl versus placebo (Kjellman 1973)	Apernyl 1/50 Placebo 4/50	RR 0.25 (0.03 to 2.16)	0.21
Rinsing with chlorhexidine both pre and postextraction versus rinsing just postextraction (Metin 2006)	Pre & post 3/46 Post 6/53	RR 0.58 (0.15 to 2.17)	0.42
1% chlorhexidine gel versus 0.2% chlorhexidine gel (Rodriguez‐Perez 2013)	1% chlorhexidine gel 3/42 0.2% chlorhexidine gel 6/46	RR 0.55 (0.15 to 2.05)	0.37
Metronidazole gel (postoperatively) versus placebo gel (Reekie 2006)	Metronidazole 8/152 Placebo 15/150	RR 0.53 (0.23 to 1.20)	0.13
P‐hydroxybenzoic acid versus placebo (Ritzau 1977)	P‐hydro 0/24 Placebo 5/21	RR 0.08 (0.00 to 1.37)	0.08
Chinese herbs Shahaosan versus placebo (Shi 2003)	Shahaosan 1/92 Placebo 8/96	RR 0.13 (0.02 to 1.02)	0.05
Chinese herbs Yunnan versus placebo (Shi 2003)	Yunnan 2/86 Placebo 8/96	RR 0.28 (0.06 to 1.28)	0.10
Chinese herbs Shahaosan versus Yunnan (Shi 2003)	Shahaosan 1/92 Yunnan 2/86	RR 0.47 (0.04 to 5.06)	0.53
Acellular dermal matrix versus control (Sun 2007)	Dermal matrix 2/150 Control 13/150	RR 0.15 (0.04 to 0.67)	0.01
Iodine tampon versus rinsing with saline in a syringe (Tuk 2019)	Iodine tampon 0/54 Monoject with saline 0/54	No events	‐
Clindamycin phosphate antibiotic solution patch versus saline patch (Trieger 1991)	Split‐mouth Clindamycin only = 0 Solution only = 7 Both = 0 Neither = 79 Total N = 86	OR (exact) 0 (0 to 0.69)	0.016
Platelet rich fibrin versus control (Unsal 2018)	Platelet rich fibrin 4/50 Control 9/50	RR 0.44 (0.15 to 1.35)	0.15
Glucocorticosteroid antibiotic agent (postoperatively) versus normal saline (van Eeden 2006)	Split‐mouth Glucocorticosteroid only = 0 Normal saline only = 3 Both = 0 Neither = 16 Total N = 19	OR (exact) 0 (0 to 2.42)	0.25
Bovine fibroblastic growth factor versus control (Xue 2013)	Growth factor 1/80 Control 10/80	RR 0.10 (0.01 to 0.76)	0.03

CI = confidence interval; N = number; RR/OR = risk ratio/odds ratio.

• Acellular dermal matrix patch versus no treatment (Bai 2011; Sun 2007).

• Apernyl versus placebo (Kjellman 1973).

• Artemisia desertorum Spreng (Shahaosan or Yunnan) versus placebo control (Shi 2003).

• Bovine fibroblastic growth factor versus control (Xue 2013).

• Chlorhexidine gel 0.2% versus chlorhexidine gel 1% (Rodriguez‐Perez 2013).

• Chlorhexidine 0.2% rinse both pre and postoperatively versus postoperatively only (Metin 2006).

• Chlorhexidine 0.2% in a monoject syringe versus plain rinsing with chlorhexidine 0.2% (Cho 2018).

• Clindamycin phosphate antibiotic solution patch versus saline patch (Trieger 1991).

• Gaseous ozone versus control (Ahmedi 2016).

• Glucocorticosteroid antibiotic agent versus normal saline (van Eeden 2006).

• Herbal mouthwash versus chlorhexidine rinse (Divya 2019).

• Heal‐all tissue patch (2 x 2.5 cm) versus no treatment (Huang 2011).

• Honey versus chlorhexidine rinse (Abu‐Mostafa 2019).

• Iodine (1%) (pre and intraoperatively) versus control (Hasheminia 2018).

• Iodine tampon versus monoject irrigation with saline (Tuk 2019).

• Irrigation with tap water in monoject versus control (Ghaeminia 2017).

• Metronidazole gel versus placebo gel (Reekie 2006).

• Oral tissue patch versus control (Hu 2005).

• P‐hydroxybenzoic acid versus placebo (Ritzau 1977).

• Platelet rich fibrin versus control (Unsal 2018).

• Tranexamic acid versus placebo (Gersel‐Pedersen 1979).

Adverse events

Additional Table 7 summarises the adverse events for all included studies. 

Treatment

Comparison 5: zinc oxide eugenol versus Alvogyl/Alveogyl 

Table 5.

Primary outcomes

Healing 

Two studies (Faizel 2015; Supe 2018) looked at healing outcomes but used different outcome measures so they were not comparable. The results were consistent with sockets treated with Alvogyl appearing to heal faster than sockets treated with zinc oxide eugenol, but the variation in the outcome measures used did not allow a meta‐analysis to be carried out. 

Supe 2018 reported 2/25 patients (8%) in the Alvogyl group and 9/25 patients (36%) in the zinc oxide eugenol group had delayed healing, classed as non‐healed sockets after 10 days indicating reduced adverse events for Alvogyl (RR 4.50, 95% CI 1.08 to 18.77; P = 0.04).

Pain ‐ VAS

Two studies (Lenka 2019; Supe 2018), with 80 participants, compared Alvogyl and zinc oxide eugenol for treatment of dry socket with visual analogue scores (VAS) (0 to 10) on day 7. The meta‐analysis shows that Alvogyl is more effective than zinc oxide eugenol at reducing pain at day 7 (mean difference (MD) ‐1.40, 95% CI ‐1.75 to ‐1.04; P < 0.00001; 2 studies, 80 participants; very low‐certainty evidence). The pooled data showed considerable heterogeneity (df = 1 (P = 0.05); I² = 75%) (Analysis 2.1). These studies also provided VAS scores for other days which gave inconsistent results.

2.1. Analysis.

Comparison 2: Treatment of dry socket, Outcome 1: Alvogyl versus zinc oxide eugenol day 7 VAS score (0‐10)

Swelling

No studies reported this outcome.

Limitation of chewing or swallowing

No studies reported this outcome.

Fever

No studies reported this outcome. 

Adverse events

No adverse events were reported. 

Secondary outcomes

Quality of life 

No studies reported this outcome.

Patient satisfaction 

No studies reported this outcome.

Costs

No studies reported this outcome.

Comparisons for treatment of dry socket, only including a single study

The following comparisons to treat dry socket were evaluated in single studies as described below, and therefore there is insufficient evidence to determine their effects. The results from these studies are shown in Additional Table 6. Kaya 2011 has four arms. The data reported in Kaya 2011 cannot be used as medians are presented with error bars. The treatment comparisons were.

• Topical anaesthetic gel (prilocaine‐lidocaine) versus eugenol (Burgoyne 2010).

• Neocone versus zinc oxide eugenol versus Alvogyl (Faizel 2015).

• Alvogyl versus no treatment (Kaya 2011).

• SaliCept versus no treatment (Kaya 2011).

• Alvogyl versus SaliCept (Kaya 2011).

• Alvogyl versus plasma rich in growth factors (PRGF) (King 2018).

• Alvogyl versus platelet rich fibrin (Keshini 2020).

• Metronidazole versus placebo (Mitchell 1984).

• Advanced platelet rich fibrin versus placebo (Yuce 2019).

Adverse events

No adverse events were reported for any of these studies. 

Discussion

Summary of main results

There were 49 studies included in this review involving 6771 participants, which reported on 35 different interventions. Of the 49 included studies, 39 looked at prevention and 10 looked at treatment of dry socket. The quality of the trials was mixed. 

There was moderate‐certainty evidence from six trials, with 1547 participants, comparing chlorhexidine rinse with placebo that rinsing with chlorhexidine perioperatively reduces the risk of developing dry socket. In all the included trials the participants rinsed with chlorhexidine for 7 days postoperatively, starting 24 hours after the extraction. In five out of the six studies the participants also rinsed preoperatively. As the 0.2% chlorhexidine subgroup included only one study, comparisons between subgroups are inappropriate. The number of patients needing to be treated to prevent one patient having dry socket (NNT) varied considerably depending on the prevalence of dry socket. For a 30% prevalence, as might be anticipated following surgical removal of a mandibular third molar, the NNT was 7 (95% confidence interval (CI) 5 to 10) but if the control prevalence is 5%, in line with commonly predicted risk following routine extractions, then the NNT increased to 33 (95% CI 27 to 49) (Table 1). This suggests that rinsing with chlorhexidine may have more of an impact following surgical removal of impacted wisdom teeth. 

There was also moderate‐certainty evidence from seven studies, with 753 participants, that placing 0.2% chlorhexidine gel in a socket following extraction is an effective preventative therapy and will reduce the risk of a dry socket. In all these studies the chlorhexidine gel was only applied once, at the end of surgery, prior to suturing. Again the NNT varied considerably from 7 (95% CI 5 to 15) for control prevalence of 30% to 37 (95% CI 28 to 72) for control prevalence of 5% (Table 2).

There is low‐certainty evidence from two studies, with 383 participants, that rinsing with chlorhexidine mouthrinse and placement of 0.2% chlorhexidine gel are equally effective at reducing the risk of developing a dry socket (Table 3).

Placement of platelet rich plasma in extraction sockets was not superior in reducing the risk of having a dry socket (risk ratio (RR) 0.51, 95% CI 0.19 to 1.33; P = 0.17; 2 studies, 127 participants; very low‐certainty of evidence) (Table 4). 

The studies reporting on the treatment of dry socket presented a number of different outcomes which made comparison of the results difficult. Pain was an outcome in all the treatment studies but it was measured in different ways and at different time points. There was very low‐certainty evidence from two studies involving 80 participants that placing Alvogyl in a dry socket was more effective than zinc oxide eugenol at reducing pain at day 7, but inconsistent results were found for other time periods (Table 5). It appears that these studies were all using the old formulation of Alvogyl, which contained iodoform, butamben, and eugenol. It should be noted that the newer formulation, branded as Alveogyl, contains eugenol only so these results may not be comparable. 

The remaining interventions for the prevention of dry socket and the interventions for the treatment of dry socket were only looked at in single studies and therefore there is insufficient evidence to determine their effects. 

Overall completeness and applicability of evidence

A comprehensive search strategy was employed and it is likely that the majority of published trials are included in this review. It is uncertain how many unpublished trials there are as some of the interventions identified were developed by industry and findings from these trials were not always reported in the past.

Most studies included in this review reported on prevention of dry socket after surgical removal of third molars. Only five studies reported on the prevention of dry socket after routine dental extractions (Abu‐Mostafa 2015; Abu‐Mostafa 2019; Alissa 2010; Halabi 2018, Reekie 2006). Two of the included studies were conducted in general dental practice settings (Cho 2018; Reekie 2006) but the majority involved extractions undertaken by experienced oral surgeons in hospital or minor oral surgery clinics. However, most extractions in dentistry are routine extractions undertaken in primary dental care, by general dental practitioners on teeth other than lower third molars. Most extractions do not involve surgical removal of the tooth. It is important that future well‐designed trials of interventions to treat and prevent dry socket are conducted in primary dental care settings. Such studies should recruit patients who are having a range of tooth type extractions including molar and premolar teeth. It is also important that power calculations inform the sample size of the proposed study to ensure trials are large enough to detect clinically important effects of interventions including hypersensitivity reactions and adverse events.

Seven of the trials in this review employed a split‐mouth design, and one trial (Tuk 2019) employed a cross‐over design. Split‐mouth designs are appropriate when the disease is stable and uniformly distributed and the effects of the intervention are short‐lived or reversible (Antczak‐Bouckoms 1990). There is evidence from one trial that development of dry socket increases the risk of developing dry socket in the future (Reekie 2006), therefore this could compromise the use of cross‐over studies, however, split‐mouth studies appear to be appropriate for looking at the prevention of dry socket. It is also important that participants and observers are blinded to the intervention allocation if possible.

There was considerable variation in the design of trials, inclusion and exclusion criteria, definition of dry socket, presurgical regimens, intraoperative procedures, and postoperative medications. The generalisability of the trials was compromised in some trials by excluding smokers and women on oral contraceptive therapy. In many trials, the existence of pericoronitis and infection was a cause for exclusion. These exclusion criteria have the potential for studies to include patient groups where the risk factors for dry socket are reduced so dry socket is less likely to happen.This effect was balanced by the predominance of studies that largely reported on extraction of third molars where the prevalence of dry socket is higher compared to extraction of other teeth. Also, two studies included participants considered at high risk to alveolar osteitis (smokers and those experiencing alveolar osteitis previously) (Halabi 2018; Karabit 2019).

Dry socket is a common consequence of tooth extraction and it is important that this review presents the latest evidence for prevention and treatment. A range of interventions reported in this review reflect prevailing theories of dry socket causation at the time the study was undertaken. 

Quality of the evidence

The criteria for assessing overall risk of bias were strict (all domains assessed had to be at low risk of bias for the trial to be deemed at low risk of bias) and only 3 studies (6%) were assessed as being at low risk of bias. Of the remainder, 30 (61%) were assessed as being at unclear risk of bias, and 16 (33%) were assessed at high risk of bias. The quality of reporting could be improved by authors reporting their studies in line with CONSORT and, where possible, undertaking double‐blind trials with adequate outcome assessment. Some studies had to be excluded as they did not have an adequate definition of dry socket and some appeared to conflate infection with dry socket. There was also variation in the secondary outcomes reported in the trials and many did not match the ones included in the protocol for the present review. This was a particular issue with outcomes related the treatment of dry socket specifically in relation to healing and pain. There was also generally inadequate reporting of adverse events.

Twenty‐four studies, with 3609 participants, using chlorhexidine in either mouthrinse or gel form were included in this review. Four studies reported some adverse events associated with the use of chlorhexidine mouthrinse. No studies reported adverse events associated with the use of chlorhexidine gel. Two of the studies reporting adverse events were using 0.12% chlorhexidine mouthrinse (Hermesch 1998; Ragno 1991) and two studies were using 0.2% chlorhexidine mouthrinse (Delilbasi 2002; Metin 2006). Adverse events included: staining of teeth, altered taste and bad taste, gastrointestinal upsets, numbness and paraesthesia. Altered taste and bad taste were the most commonly reported adverse events (Delilbasi 2002; Metin 2006; Ragno 1991). Rubio‐Palau 2015 reported a number of adverse reactions including a skin rash in two patients but the authors attributed these to the use of the analgesic metamizole (dipyrone). None of the included studies reported an allergic reaction to chlorhexidine‐based products, though it should be noted that most of the studies had allergy to chlorhexidine as exclusion criteria. 

A Cochrane Review looking at chlorhexidine for reduction of gingivitis (James 2017) found that, apart from staining and calculus formation, the most commonly reported adverse events in the chlorhexidine rinse arms of the included studies were taste disturbance/alteration (reported in 11 studies), effects on the oral mucosa including mucosal irritation, soreness, mild desquamation, mucosal ulceration/erosions and oral mucosal lesions (reported in 13 studies), and a general burning sensation or a burning tongue or both (reported in nine studies). They include no reports of allergic reaction to chlorhexidine (James 2017). There is some evidence from in vitro studies that chlorhexidine may cause a dose‐dependent inhibition of gingival fibroblast proliferation and collagen production, which could potentially negatively affect wound healing (Mariotti 1999), though there is a lack of evidence from clinical studies to support this. The studies included in this review, which looked at chlorhexidine use, did not report on socket healing. 

While serious adverse effects and events are rare, it is recommended that all members of the dental team prescribing chlorhexidine mouthwashes and gels for the management of dry socket are aware of the potential for side effects, are competent to manage a medical emergency associated with anaphylaxis, and warn their patients of the potential for adverse events.

Potential biases in the review process

Reporting inconsistency and missing data meant that some studies, particularly the older ones, could not be entered into the meta‐analysis. Also, there were too few studies in some interventions (e.g. platelet rich plasma (only two studies reported)). Both issues could have introduced bias. Additionally, we had too few studies for both prevention and treatment studies to assess publication bias.

Agreements and disagreements with other studies or reviews

Four systematic reviews concluded that rinsing perioperatively with chlorhexidine mouthrinse was effective in preventing dry socket after extractions (Canullo 2020; Caso 2005; Hedstrom 2007; Rodríguez Sánchez 2017). Three of these reviews restricted inclusion to studies reporting on surgical removal of mandibular third molars (Caso 2005; Hedstrom 2007; Rodríguez Sánchez 2017). The findings from the present review are mostly consistent with the findings reported in these reviews. Contrary to our results, Canullo 2020 carried out a subgroup analysis and found that 0.2% chlorhexidine mouthrinse did not significantly reduce the risk of dry socket. Hedstrom 2007 concluded that while it could not be determined that perioperative rinsing with 0.12% chlorhexidine prevented dry socket in all extractions, there was evidence that perioperative rinsing with chlorhexidine could prevent dry sockets in lower third molar extractions. 

Canullo 2020; Caso 2005; and Hedstrom 2007 included trials were systemic antibiotics were prescribed, provided both the experimental and control groups received the same regimen. As in this review, Rodríguez Sánchez 2017 excluded any trials prescribing systemic antibiotics. The NNT from the latter's review was eight. Yengopal 2012 included a different group of studies to the present review. In contrast to our review, Yengopal 2012 concluded that chlorhexidine rinse had not been conclusively shown to be significantly better than placebo for reducing the incidence of alveolar osteitis (dry socket) after tooth extraction.

Two systematic reviews looking at surgical removal of mandibular third molars (Canellas 2020; Rodríguez Sánchez 2017) and one systematic review included all tooth types extractions (Canullo 2020) with the latter concluding that 0.2% chlorhexidine intrasocket gel significantly reduced the risk of developing a dry socket after extractions. Two of the reviews (Canullo 2020; Rodríguez Sánchez 2017) concluded, in line with our results, that chlorhexidine gel was as effective as chlorhexidine mouthrinse in reducing the risk of dry socket. 

Two systematic reviews looking at placement of platelet rich plasma to aid healing of extraction sockets concluding that there was insufficient data to produce a meta‐analysis with regard to dry socket (Barona‐Dorado 2014; Franchini 2019). 

Four systematic reviews (Canellas 2020; Canellas 2017; Xiang 2019; Zhu 2020) concluded that placement of platelet rich fibrin in extraction sockets significantly reduced the risk of dry socket, this is in contrast to our findings. The inclusion criteria for these four reviews differed from the present review as they included trials where systemic antibiotics had been prescribed. Two systematic reviews looked at plasma rich in growth factor with one review, Del Fabbro 2019, concluding that plasma rich in growth factor reduces the risk of dry socket, and one review (Xu 2019) concluding that there was insufficient data for a meta‐analysis. 

Authors' conclusions

Implications for practice.

There is moderate‐certainty evidence that rinsing perioperatively with 0.12% and 0.2% chlorhexidine gluconate, probably results in a reduction in dry socket. There is moderate‐certainty evidence that chlorhexidine gel placed in the socket postextraction probably results in a reduction in dry socket. The decision to recommend rinsing with chlorhexidine should be balanced against reported adverse effects of tooth staining, taste alteration, and nausea. Adverse events attributable to the rinses were rare, but patients need to be aware and informed of the potential for adverse events associated with the use of chlorhexidine. There was limited evidence of very low certainty that Alvogyl (old formulation) may reduce pain at day 7 in patients with dry socket when compared to zinc oxide eugenol.

Implications for research.

More well‐designed trials in general dental practice settings with teeth other than third molars and including non‐surgical extractions are needed. Further studies comparing rinsing with chlorhexidine with intrasocket chlorhexidine gel to prevent dry socket would help to determine whether one intervention is better than another. Clinicians and researchers in this area need to decide collectively what outcomes should be measured in both prevention and treatment studies. All studies should specifically enquire into and carefully present data on any adverse events. More research is required into the effectiveness of treatment of dry sockets.

What's new

Date	Event	Description
4 May 2022	New citation required and conclusions have changed	28 new studies included in this update bringing the total to 49 included studies
28 September 2021	New search has been performed	Search updated

History

Protocol first published: Issue 1, 2008
Review first published: Issue 12, 2012

Appendices

Appendix 1. Cochrane Oral Health’s Trials Register search strategy 

Cochrane Oral Health’s Trials Register is available via the Cochrane Register of Studies. For information on how the register is compiled, see oralhealth.cochrane.org/trials.

#1 (("dry socket*" or "alveolar osteitis" or "alveolar ostitis" or alveolagia or "alveolar periostitis" or "alveolitis sicca dolorosa" or dry socket*  or "septic socket*" or "necrotic socket osteomylitis" or "fibrinolytic alveolitis")) AND (INREGISTER)
#2 ((infect* and socket*)) AND (INREGISTER)
#3 (#1 or #2) AND (INREGISTER)
#4 (extract* or remov* or surg*) AND (INREGISTER)
#5 (#3 and #4) AND (INREGISTER)

Searches to January 2012 were conducted using the following strategy on the ProCite version of the register:

(("dry socket*" or "alveolar osteitis" or "alveolar ostitis" or alveolagia or "alveolar periostitis" or "alveolitis sicca dolorosa" or dry‐socket*  or "septic socket*" or "necrotic socket osteomylitis" or "fibrinolytic alveolitis" or (infect* and socket*)) AND (extract* or remov* or surg*))

Appendix 2. Cochrane Central Register of Controlled Trials (CENTRAL) search strategy

#1        MeSH descriptor Dry Socket this term only 
#2        ("alveolar osteitis" in All Text or "alveolar ostitis" in All Text) 
#3        ((septic in All Text near/4 socket* in All Text) or "necrotic socket osteomylitis" in All Text or "fibrinolytic alveolitis" in All Text) 
#4        (alveolalgia in All Text or "alveolar periostitis" in All Text or "alveolitis sicca dolorosa" in All Text) 
#5        ("dry socket*" in All Text or dry‐socket* in All Text or (infect* in All Text  near/5 socket* in All Text)) 
#6        (#1 or #2 or #3 or #4 or #5) 
#7        MeSH descriptor Tooth Extraction explode all trees 
#8        ("dental extraction*" in All Text or (tooth in All Text near/3 extract* in All Text) or (teeth in All Text near/3 extract* in All Text) or (teeth in All Text near/3 remov* in All Text) or (tooth in All Text near/3 remov* in All Text)) 
#9        ((dental in All Text or oral in All Text) and (surgery in All Text or surgical  in All Text)) 
#10      (#7 or #8 or #9) 
#11      (#6 and #10) 

Appendix 3. MEDLINE Ovid search strategy 

1. Dry socket/

2. ("alveolar osteitis" or "alveolar ostitis").mp. 

3. (alveolalgia or "alveolar periostitis" or "alveolitis sicca dolorosa").mp. 

4. ((septic adj4 socket*) or "necrotic socket osteomylitis" or "fibrinolytic alveolitis").mp.

5. ("dry socket$" or dry‐socket$ or (infect$ adj5 socket$)).mp.

6. or/1‐5

7. exp Tooth Extraction/

8. ("dental extraction$" or (tooth adj3 extract$) or (teeth adj3 extract$) or (teeth adj3 remov$) or (tooth adj3 remov$)).mp.

9. ((dental or oral) and (surgery or surgical)).mp.

10. or/7‐9

11. 6 and 10

The above subject search was linked with the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials in MEDLINE (as described in Lefebvre 2021, box 3.b). 

1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
3. randomized.ab.
4. placebo.ab.
5. drug therapy.fs.
6. randomly.ab.
7. trial.ab.
8. groups.ab.
9. or/1‐8
10. exp animals/ not humans.sh.
11. 9 not 10

Appendix 4. Embase Ovid search strategy  

1.    Dry socket/
2.    ("alveolar osteitis" or "alveolar ostitis").mp.
3.    (alveolalgia or "alveolar periostitis" or "alveolitis sicca dolorosa").mp.
4.    ((septic adj4 socket*) or "necrotic socket osteomylitis" or "fibrinolytic alveolitis").mp.
5.    ("dry socket$" or dry‐socket$ or (infect$ adj5 socket$)).mp.
6.    or/1‐5
7.    exp Tooth Extraction/
8.    ("dental extraction$" or (tooth adj3 extract$) or (teeth adj3 extract$) or (teeth adj3 remov$) or (tooth adj3 remov$)).mp.
9.    ((dental or oral) and (surgery or surgical)).mp.
10.  or/7‐9
11.  6 and 10

The above subject search was linked with the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials in Embase (as described in Lefebvre 2021, box 3.e). 

1. Randomized controlled trial/

2. Controlled clinical study/

3. random$.ti,ab.

4. randomization/

5. intermethod comparison/

6. placebo.ti,ab.

7. (compare or compared or comparison).ti.

8. ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab.

9. (open adj label).ti,ab.

10. ((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab.

11. double blind procedure/

12. parallel group$1.ti,ab.

13. (crossover or cross over).ti,ab.

14. ((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.

15. (assigned  or allocated).ti,ab.

16. (controlled adj7 (study or design or trial)).ti,ab.

17. (volunteer or volunteers).ti,ab.

18. human experiment/

19. trial.ti.

20. or/1‐19

21. random$ adj sampl$ adj7 ("cross section$" or questionnaire$1 or survey$ or database$1)).ti,ab. not (comparative study/ or controlled study/ or randomi?ed controlled.ti,ab. or randomly assigned.ti,ab.)

22. Cross‐sectional study/ not (randomized controlled trial/ or controlled clinical study/ or controlled study/ or randomi?ed controlled.ti,ab. or control group$1.ti,ab.)

23. (((case adj control$) and random$) not randomi?ed controlled).ti,ab.

24. (Systematic review not (trial or study)).ti.

25. (nonrandom$ not random$).ti,ab.

26. "Random field$".ti,ab.

27. (random cluster adj3 sampl$).ti,ab.

28. (review.ab. and review.pt.) not trial.ti.

29. "we searched".ab. and (review.ti. or review.pt.)

30. "update review".ab.

31. (databases adj4 searched).ab.

32. (rat or rats or mouse or mice or swine or porcine or murine or sheep or lambs or pigs or piglets or rabbit or rabbits or cat or cats or dog or dogs or cattle or bovine or monkey or monkeys or trout or marmoset$1).ti. and animal experiment/

33. Animal experiment/ not (human experiment/ or human/)

34. or/21‐33

35. 20 not 34

Appendix 5. US National Institutes of Health Ongoing Trials Register (ClinicalTrials.gov) search strategy

“dry socket”

Appendix 6. World Health Organization International Clinical Trials Registry Platform search strategy

“dry socket” OR “alveolar osteitis”

Data and analyses

Comparison 1. Prevention of dry socket.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Chlorhexidine rinse versus placebo	6	1600	Odds Ratio (IV, Random, 95% CI)	0.38 [0.25, 0.58]
1.1.1 0.2% chlorhexidine rinse	1	121	Odds Ratio (IV, Random, 95% CI)	0.85 [0.36, 2.01]
1.1.2 0.12% chlorhexidine rinse	5	1479	Odds Ratio (IV, Random, 95% CI)	0.34 [0.23, 0.49]
1.2 Chlorhexidine gel versus placebo/no treatment	7	833	Odds Ratio (IV, Random, 95% CI)	0.44 [0.27, 0.71]
1.3 Chlorhexidine gel versus chlorhexidine rinse	2	383	Risk Ratio (IV, Fixed, 95% CI)	0.74 [0.46, 1.20]
1.4 Platelet rich plasma versus placebo	2	127	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.19, 1.33]

Comparison 2. Treatment of dry socket.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Alvogyl versus zinc oxide eugenol day 7 VAS score (0‐10)	2	80	Mean Difference (IV, Fixed, 95% CI)	‐1.40 [‐1.75, ‐1.04]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abu‐Mostafa 2015.

Study characteristics
Methods	Study design: randomised, controlled, split‐mouth trial Sample size calculation: not stated Sample size: 311 recruited but 10 excluded, therefore 301 reported Setting: not stated Number of centres: 1 Operators: dental interns or dental students under supervision of surgery instructors in the Colleges Clinics Prevention or treatment of dry socket: prevention Type of teeth: "single extraction was done of upper or lower molar tooth" Recruitment period: November 2012 to April 2013 Funding source: Alhokail Orthodontic Center, Dammam City, Saudi Arabia provided 0.12% CHX mouthwash and 0.2% CHX gel for use in this study Declarations of interest: none stated
Participants	Inclusion criteria: upper or lower molar teeth indicated for extraction Exclusion criteria: uncontrolled systemic diseases epinephrine contraindications allergy to CHX, lidocaine and ibuprofen pregnant women breastfeeding women those who were using oral contraceptives presence of an acute infection cystic lesions traumatic extraction with fractured alveolar bone extraction requiring bone reduction extractions that lasted more than 30 minutes Gender: 65 females, 236 males Mean age: not stated Age range: not stated Number randomised: Intervention Group: 160. Control Group: 151 Number analysed: Intervention Group: 160. Control Group: 141 Lost to follow‐up/dropouts: 10 excluded (reasons: "…because they did not use CHX rinse twice daily for a week")
Interventions	All patients: postoperative instructions were given for all patients in addition to the prescription of ibuprofen (Brufen, Hamol Limited, Nottingham, England) 600 mg every 8 hours for 3 days Control group: all patients received a bottle of 0.12% CHX mouthwash (Peridex, Oral Rinse, 3M, ESPE, USA) to start using it on the 2nd day of extraction twice daily for 7 days Intervention group: after performing tooth extraction, CHX 0.2% bioadhesive gel (Elugel, 40 mL gel tube, Pierre Fabre Oral Care, Boulogne, Paris, France) was applied into the extraction socket At the 3rd postoperative day CHX 0.2% bioadhesive gel was applied again into the extraction socket
Outcomes	Primary outcome measures: presence of a dry socket Diagnosis: "Acute alveolar osteitis, (dry socket) was diagnosed if the patient presented between the 2nd and 4th days with pain or tenderness in the socket with probing, empty socket and food debris with or without halitosis" Secondary outcome measures: none reported Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were divided randomly into two parallel groups by asking them to choice one card out of two in which the group number was written on its back" Comment: there is an element of randomisation in the patient's choice
Allocation concealment (selection bias)	High risk	Quote: "Patients were divided randomly into two parallel groups by asking them to choice one card out of two in which the group number was written on its back" Comment: allocation is not concealed with the group number being written on the back, could easily be seen by patient or researcher
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It would not be possible to blind participants or personnel as this is a study comparing the professional application of a gel versus patient utilised mouthwash. Participants knowing the intervention would probably not impact outcomes being assessed but this is unlikely to have led to deviations from the intended intervention. Operators knowing the interventions are unlikely to have led to deviations from the intended intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	There is no mention of outcome assessor blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Three hundred‐eleven patients participated in the study; ten of whom were eventually excluded because they did not use CHX rinse twice daily for a week" Comment: there are no data available for 10 patients who were enrolled in the trial
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a judgement
Other bias	High risk	Quotes: "Extractions were performed by dental interns or dental students under supervision of surgery instructors in the Colleges Clinics"  "Simple extractions were done by elevators and forceps, while root separations were done using a surgical handpiece" Comment: a variety in operator experience may introduce bias especially if this is not equally distributed between the groups. Also the rates of AO were higher (as to be expected from the literature) in surgical extractions. It is unclear whether an attempt was made to ensure the type of extraction was equally represented in both groups. Table 3 page 85 shows 19 of the 29 AO observed in the study were related to extractions via root separation

Abu‐Mostafa 2019.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 100 Setting: university clinics Number of centres: 1 Operators: dental interns Prevention or treatment of dry socket: prevention Type of teeth: single molar tooth Recruitment period: October 2016 to August 2017 Funding source: not reported Declarations of interest: not reported
Participants	Inclusion criteria: patients with upper or lower molar teeth indicated for extraction Exclusion criteria: patients with uncontrolled systemic diseases, epinephrine contraindications, pregnant women, breastfeeding women, and women who were using oral contraceptives. Patients with allergy to CHX, honey, lidocaine, and ibuprofen were also excluded. Other exclusion criteria included smoking, presence of acute infection, cystic lesions, traumatic extraction with fractured alveolar bone, extraction requiring bone reduction or root separation, and extractions that lasted more than 30 minutes Gender: 48 males, 52 females Mean age: 38.13 years Age range: 17 to 69 years Number randomised: 100 Number analysed: 100 Lost to follow‐up/dropouts: 0
Interventions	All patients: regular postoperative care and verbal instructions, instructed to take 400 mg ibuprofen every 8 hours on 1st and 2nd day of the tooth extraction Intervention group 1: 0.2% CHX mouthwash twice daily for 7 days Intervention group 2: Manuka honey topically to extraction socket immediately and on 3rd postoperative day
Outcomes	Primary outcome measures: pain, assessment of socket for food debris and postextraction halitosis Diagnosis: VAS, clinical examination of socket, fetid odour from patients mouth during speech Secondary outcome measures: none reported Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random allocation of the patients into the two parallel groups was done by asking them to choose 1 of the 2 colored cards. The green card indicated Group 1 while the blue indicated Group 2" Comment: there is a random element to this
Allocation concealment (selection bias)	High risk	Quote: "Random allocation of the patients into the two parallel groups was done by asking them to choose 1 of the 2 colored cards" Comment: allocation is not concealed with colour of card denoting group allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It would not be possible to blind participants or personnel as this is a study comparing the professional application of a honey intrasocket versus patient rinsing with CHX. Participants knowing the intervention would probably not impact outcomes being assessed, but this is unlikely to have led to deviations from the intended intervention. Operators knowing the interventions are unlikely to have led to deviations from the intended intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	There is no mention of outcome assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All participants enrolled are accounted for but groups noticeably different in size. No details provided
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	High risk	Quotes: "Extractions were performed by dental interns under close supervision of surgery instructors in the University Clinics" "Additionally, the patients were instructed to take 400 mg of ibuprofen every 8 hours on the 1st and 2nd day of the tooth extraction" Comment: a variety in operator experience may introduce bias especially if this is not equally distributed between the groups Adherence with pain relief regimen was not reported/recorded and could have impacted subjected reporting of pain on VAS

Ahmedi 2016.

Study characteristics
Methods	Study design: randomised, controlled, split‐mouth trial Sample size calculation: not stated Sample size: 30 Setting: Department of Oral Surgery in the University Dental Clinical Centre of Kosovo (UDCCK) Number of centres: 1 Operators: 1 oral surgeon Prevention or treatment of dry socket: prevention Type of teeth: bilateral impacted lower 3rd molars in similar position Recruitment period: January 2014 to June 2014 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: 18 to 30 years bilateral impacted lower 3rd molars in similar position verified by orthopantomography an indication for surgical extraction (pericoronitis, caries, pulpitis, or orthodontic indication) Exclusion criteria: any systemic disease smokers immunocompromised patients pregnant women those who were taking contraceptives Gender:  not reported Mean age: 27.87 ± 4.11years Age range: 18 to 30 years Number randomised: 30 Number analysed: 30 Lost to follow‐up/dropouts: 0
Interventions	All: postoperative instructions and prescribed ibuprofen 400 mg provided Control group: 5 mL of 0.9% saline solution was used to irrigate the socket Intervention group: ozone gas was used to irrigate the socket "We used Prozone equipment to supply the O3 gas, which enabled us to introduce the gas into the socket using plastic attachments for 12 s calibrating the therapeutic dose. This procedure was performed using a surgical suction unit to avoid respiratory aspiration and related complication. Suturing was performed with 3/0 absorbable suture (Ethicon, Somerville, NJ, USA)"
Outcomes	Primary outcome measures: dry socket Diagnosis: "Absence of a blood clot, bone exposure, or necrotic blood clot" Secondary outcome measures: none reported Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Randomisation was used to determine which side would comprise the control group (gr1) and which would comprise the experimental (O3) group (gr2)" Comment: method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "The patients were not aware of these designations" and its unlikely during the procedure (due to anesthesia) that they would have known which intervention they received. All the surgeries were performed by the same oral surgeon who would have known the intervention, but this lack of masking is unlikely to have led to a deviation in the intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	1 surgeon did start and end assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All enrolled participants reported on
Selective reporting (reporting bias)	High risk	All outcomes were not reported i.e. indication for extraction and difficulty of extraction as per objectives 2 and 3 of the study
Other bias	High risk	The indication and difficulty for extraction was not reported, and could have been different in each quadrant i.e. a patient may have had caries/infection in 1 tooth which was not present in the other. These could have influenced the primary outcome of the study (AO)

Alissa 2010.

Study characteristics
Methods	Study design: randomised, controlled, parallel‐group (pilot) study Sample size calculation: a priori sample size calculation suggested needed 34 in each group Sample size: 23, suggesting 60% power Conducted in: University Dental Hospital of Manchester, Manchester, UK. Patients recruited from consultation clinic Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: molars, premolars, canines and incisors, 17 (60.9%) extracted for dental caries and 5 (21.7%) endodontic failure. Extracted under intravenous sedation Recruitment period: January 2005 to April 2008 Funding source: not stated
Participants	Inclusion criteria: healthy adults over 18 years of age patients with a demonstrable need for removal of at least 1 tooth, treated with local anaesthesia and intravenous sedation patients able to consent (written) and willing to complete the requirements of the study protocol Exclusion criteria: pregnancy/nursing mothers/having childbearing potential and not using birth control measures platelet dysfunction syndrome or critical thrombocytopenia under treatment with NSAIDs including aspirin, antibiotics, systemic corticosteroids, anticoagulants or immunosuppressive drugs diabetes mellitus cardiovascular disease including a history of rheumatic fever, or other conditions requiring antibiotic prophylaxis neoplasia or haematological malignancy renal, hepatic or endocrine diseases metabolic bone disease such as osteomalacia, hypocalcaemia or hypercalcaemia participation in another trial Age group: mean = 30.5 Intervention group: number randomised 12; analysed 12 Control group: number randomised 11; analysed 7 Number evaluated: 19
Interventions	Comparison: platelet rich plasma (PRP) placed in extraction sockets versus control All patients were treated under intravenous sedation and blood drawn to manufacture the PRP produce before the surgical procedure and before intravenous sedation administered. All patients had a mucoperiosteal flap raised with 2 releasing incisions. Extraction with forceps to minimise trauma, elevators used as appropriate. All sockets carefully curetted to remove granulation tissue and/or periapical infections PRP group (n = 12): platelet rich plasma placed in extraction sockets after extraction and curettage Control group (n = 11): nothing placed in control sockets Co‐interventions and concomitant medication: none stated. Co‐codamol analgesic tablets (codeine phosphate 30 mg and paracetamol 500 mg)
Outcomes	Primary outcome measures: alveolar osteitis: assessed 1 week postoperatively using Cheung et al 2001 (range) Secondary outcome measures: quality of life. A statistically significant improvement associated with use of PRP was observed for 2 variables: bad taste in mouth (P = 0.03), and food stagnation (P = 0.03), and 'borderline' for change in diet (P = 0.06) as reported by authors patient satisfaction: a high level of patient satisfaction in both groups, and no difference between groups Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A computer‐generated randomization schedule was created by a statistician" page 127 Comment: sequence generation was randomised
Allocation concealment (selection bias)	Low risk	Quote: "The randomization codes were enclosed in sealed, opaque and sequentially numbered envelopes. The patients allocation to either group was revealed by the investigator just before venous cannulation on the day of the patient’s appointment for the extraction" Comment: efforts to conceal allocation done
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It was probably not possible to blind personnel and participants to group allocation. Strict study criteria required that participants would be blinded to group allocation as they will be involved in reporting of subjective patients symptoms. Operators knowing the interventions are unlikely to have led to deviations from the intended intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	While the criteria for dry socket are clear, it is unclear who and how they made the judgement Insufficient information to permit judgement of 'high' or 'low' risk, therefore unclear risk
Incomplete outcome data (attrition bias) All outcomes	High risk	In relation to AO, 2 patients from the control group did not attend any of the scheduled appointments following tooth extraction, and given the low event rate it  may have introduced a bias. There was no loss to follow‐up in the intervention group
Selective reporting (reporting bias)	Low risk	All outcomes reported. The report appears to be free of selective reporting
Other bias	Low risk	The study appears to be free of other sources of bias

Babar 2012.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 100 Setting: Department of Oral and Maxillofacial Surgery, Armed Forces Institute of Dentistry, Pakistan Number of centres: 1 Operators: 1 surgeon Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: 1 January to 31 December 2007 Funding source: not reported Declarations of interest:  not reported
Participants	Inclusion criteria: requiring removal impacted mandibular 3rd molar Exclusion criteria: acute pericoronitis, taking antibiotics, history of smoking, pregnancy, any other bone pathology, taking immunosuppressants Gender: 65 males, 35 females Mean age: 29 Age range: 18 to 40 years Number randomised: 100 Number analysed:  100 Lost to follow‐up/dropouts: 0
Interventions	All patients: 400 mg ibuprofen for postoperative pain Intervention group 1: none Intervention group 2: 0.2% CHX gel placed into alveolus
Outcomes	Primary outcome measures: acute AO diagnosis Diagnosis: AO determined by presence of pain, blood clot disintegration, halitosis Secondary outcome measures: none reported Adverse outcomes: no adverse outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "divided randomly into two groups" page 92 Comment: a random sequence generation was obtained, method was not described
Allocation concealment (selection bias)	Unclear risk	There is no reporting of methods/process to conceal allocation just the statement "100 patients.... were examined clinically and radiologically and then divided randomly into two equal groups"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It would not be possible to blind participants or personnel as this is a study comparing the professional application of a 0.2% CHX intrasocket versus no intervention, but this lack of masking is unlikely to have led to a deviation in the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	There is no mention of outcome assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants enrolled are accounted for
Selective reporting (reporting bias)	Low risk	All  outcomes reported
Other bias	High risk	Quote: "Extractions were performed by dental interns under close supervision of surgery instructors in the University Clinics" Comment: a variety in operator experience may introduce bias especially if this is not equally distributed between the groups

Bai 2011.

Study characteristics
Methods	Study design: parallel RCT Sample size calculation: not clear that this was undertaken Study size: 400 participants Conducted in: Stomatological Center of Chinese PLA, the 306 Hospital of Chinese PLA, Beijing, China Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molar, local anaesthesia Recruitment period: March 2009 to June 2010 Providers of care: 1 surgeon Funding source: no funding
Participants	Inclusion criteria: participants required to have their impacted mandibular 3rd molar extracted aged 18 to 50 years no acute pericoronitis within a week no antibiotics within 3 days before the extraction no contraindications Exclusion criteria: acute pericoronitis within a week before the extraction antibiotics administration within 3 days before the extraction Age group: mean 27.9 Group 1: number randomised: 200; number analysed: 200 Group 2: number randomised: 200; number analysed: 200 Number evaluated: 200
Interventions	Comparison: acellular dermis matrix versus control Group 1 (n = 200): acellular dermis matrix (1×1 cm) embedded into the extraction sockets Group 2 (n = 200): nothing embedded into the extraction sockets Co‐interventions: extraction of the teeth (the detailed information was not reported) Concomitant medication: not reported
Outcomes	When measured: 1 week after the extraction Primary outcome measures: presence of postoperative alveolitis; disintegration of blood clot; haemorrhage rate; swelling rate (the exact measurement method was not reported); food residue rate; rate of red swollen gingiva Secondary outcome measures: none reported Adverse outcomes: none reported
Notes	Author contact failed
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The participants were randomized into intervention group (n = 200) and control group (n = 200)" Comment: the detailed methods of randomisation were not clearly reported
Allocation concealment (selection bias)	Unclear risk	Insufficient reporting to make a judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient reporting to make a judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient reporting to make a judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	None of the participants were reported to have dropped out
Selective reporting (reporting bias)	Low risk	All the outcomes in relation to dry socket were reported clearly
Other bias	High risk	Although the authors reported the gender and the age, the difficulties of the extraction and the detailed information of the tooth impaction were not reported

Burgoyne 2010.

Study characteristics
Methods	Study design: randomised parallel‐group controlled trial Study size: 35 Sample size calculation: not stated Prevention or treatment of dry socket: treatment Type of teeth: mandibular 3rd molars, mandibular and maxillary molars and premolars Conducted in: Department of Oral and Maxillofacial Surgery, School of Dentistry, Virginia Commonwealth University (Richmond, VA), USA Number of centres: 1 Recruitment period: not stated Funding source: not stated
Participants	Inclusion criteria: presenting with AO; diagnosed as pain of increasing severity 2 to 3 days postextraction, absence of a clot in the socket Exclusion criteria: immunocompromised or on immunosuppressant drugs steroids, non‐steroidal anti‐inflammatory drugs within 4 hours of examination allergy to eugenol, lidocaine, prilocaine, acetaminophen or codeine medical conditions: type 1/2 diabetes, glucose 6‐phosphate dehydrogenase deficiency pregnancy drugs associated with drug induced methaemoglobinaemia Age: gauze strip group: mean = 33, range 19 to 53 years; topical anaesthetic gel: mean = 27, range 17 to 58 years Number randomised: 35 Number analysed: 35
Interventions	Comparison: topical anaesthetic gel versus eugenol gauze strip Group 1 (n = 15): 2.5% prilocaine, 2.5% lidocaine (Oraqix; Dentsply Pharmaceutical, York, PA) thermosetting gel syringed into socket Group 2 (n = 20): eugenol on plain gauze into socket Co‐interventions: extraction site irrigated with normal saline prescription for 24 tablets of acetaminophen and codeine, 30 mg, taking 1 or 2 tablets every 4 hours if needed for pain. Details recorded quantity and frequency
Outcomes	Self‐assessment 5, 10, 15 minutes after 1st treatment, and then hourly whilst awake over 48 hours. Assessed at 48 hours and if still in pain re‐treated Primary outcome measures: relief of pain associated with localised AO. Pain measured by VAS: no pain (0) to severe (10) compared at 24 and 48 hours Secondary outcome measures: none assessed Adverse effects: none reported
Notes	The pain outcomes sought were to be assessed during waking hours but it was unclear why the investigators considered pain assessments not recorded while the patients slept as "missing data." Quote: "Two analyses were performed, one treating the missing data as ignorable and the other inputting pain scores of 0 during sleep." We used the one which ignored the missing data
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "control or experimental group by use of a randomization table" page 145  Comment: randomisation method described
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a clear judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants blinded to allocation and to what they received. Personnel were not blinded to interventions. Strict study criteria required that personnel would be blinded to group allocation as they will be involved in reporting of patients symptoms, operator not masked is unlikely to have led to deviation in intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear who made outcome assessment, if it was the operator given the low numbers in the trial s/he could remember group allocation, but insufficient information to form a judgement of high risk
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Outcomes assessments not carried out during periods of sleep were considered as "missing data"
Selective reporting (reporting bias)	Low risk	Although the study protocol was unavailable, the report appears to be free of selective reporting
Other bias	Low risk	Appears to be free of other bias

Chaurasia 2017.

Study characteristics
Methods	Study design: RCT Sample size calculation: not stated Sample size: 88 patients Setting: Department of Oral and Maxillofacial Surgery of Dhulikhel Hospital, Nepal Number of centres: 1 Operators: not stated Prevention or treatment of dry socket: treatment Type of teeth: mandibular and maxillary 2nd and 3rd molars, mandibular canines Recruitment period: not stated Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: pain in and around the extraction socket with or without radiation that increases in severity for some period from 1st to 3rd day after extraction partial or total clot loss in the interior of the alveolus with or without halitosis Exclusion criteria: not stated Gender: 36 females, 52 males Mean age: not stated Age range: not stated Number randomised: 88 Number analysed: not clear Lost to follow‐up/dropouts: not stated
Interventions	All: in both groups irrigation with normal saline was done to remove debris or infected clot Group A: placement of zinc oxide eugenol paste mixed with cotton pellet (manufacturer‐ Septodent India) Group B: placement of Alveogyl TM (manufacturer‐ Septodent India) The medication was changed every day until the pain subsided
Outcomes	Primary outcome measures: intensity of pain: recorded on VAS of 0 to 10. Pain score was recorded at the time of diagnosis, 30 and 60 minutes after placement of medication, on 2nd, 5th, 7th and 10th day. The medication was changed every day until the pain subsided Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The dry socket patients were randomly divided into two groups... They were randomly assigned using randomization table to group A and group B"
Allocation concealment (selection bias)	Unclear risk	This was not mentioned, and there is insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	This was not mentioned, and there is insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	his was not mentioned, and there is insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It is unclear whether all patients were followed up at all time points
Selective reporting (reporting bias)	Low risk	Pain was the main outcome measure and this appears to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Cho 2018.

Study characteristics
Methods	Study design: single‐blind, RCT Sample size calculation: not stated Sample size: 100 patients (120 teeth as 10 bilateral in each group) Setting: 3 private dental clinics in South East Queensland, Australia Number of centres: 3 Operators: 3 surgeons (all with at least 5 years experience of dentoalveolar surgery) did the operations Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: 2014 to 2016 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: age: 18 and 80 years removal of 1 or 2 lower 3rd molars under local anaesthesia, involved the raising of a flap and removal of bone with or without sectioning of the tooth the adjacent tooth was present Exclusion criteria: previous radiotherapy to the head and neck diabetes mellitus organ transplants use of bisphosphonates or steroids smokers pregnancy allergy to CHX use of a mouthwash preoperatively those who did not attend follow‐up appointments Gender: 44 females, 56 males Mean age: control group: 35 years; intervention group: 36 years Age range: control group: 18 to 72 years; intervention group: 19 to 76 years Number randomised: 100 Number analysed: 95 Lost to follow‐up/dropouts: 5 (unable to contact)
Interventions	All patients were given a bottle of CHX mouthwash (Savacol® alcohol‐free antiseptic mouth and throat rinse, Colgate‐Palmolive) with verbal and written instructions on how to use it. Patients were prescribed oral analgesia with paracetamol/codeine 500/15 mg 1 or 2 tablets every 4 to 6 hours as needed, and ibuprofen 200 mg 1 or 2 tablets every 4 to 6 hours as needed, both with a maximum of 8 tablets/day Control group: "On the day after the surgery, start mouthwash irrigation twice a day" – rinse only Intervention group: "On the day after the surgery, start mouthwash irrigation twice a day" – use of a plastic syringe (Monoject 412, Covidien) (to irrigate the socket using the CHX)
Outcomes	Primary outcome measures: pain Diagnosis: VAS used Secondary outcome measures: AO acute infection wound dehiscence food packing trismus facial swelling Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Using a sealed envelope method, we randomly allocated the patients in a 1:1 ratio into the irrigation (intervention) or rinse (control) group (n = 50 in each). This was done postoperatively to minimise operator bias"
Allocation concealment (selection bias)	Low risk	Quote: "Using a sealed envelope method, we randomly allocated the patients in a 1:1 ratio into the irrigation (intervention) or rinse (control) group (n = 50 in each)"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Randomisation "…was done postoperatively to minimise operator bias" Comment: given the nature of the interventions it is not possible to blind the patient. Insufficient data to make a judgement
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The telephone call 48 hours after the operation included a pain score using a visual analogue scale that had been given out after the operation. The interviewer was blinded to the intervention used. During their appointment at the clinic on the seventh postoperative day (acceptable range 5–10 days), the independent observer reviewed the postoperative course, examined the extraction site"
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	5 patients dropped out and there appears to have been a per‐protocol analysis
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Delilbasi 2002.

Study characteristics
Methods	Study design: randomised, placebo‐controlled, parallel‐group study Sample size: 177 Power calculation: not reported Conducted in: not stated (authors based in Turkey and Japan) Number of centres: not stated Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: not stated Funding source: not stated
Participants	Inclusion criteria: at least 1 impacted mandibular 3rd molar patients in good health Exclusion criteria: patients who had pericoronitis or were taking antibiotics for other infections women who were pregnant, breastfeeding, or using oral contraceptives Age: Group 1: 24.1 years; Group 2: 24.7 years; Group 3: 24.2 years Number randomised: 177 Number analysed: 177
Interventions	Group 1 (n = 62): rinse with 15 mL of CHX solution (Klorhex; Drogsan) for 30 seconds just before tooth removal. Intraoperatively, 15 mL of CHX diluted with 15 mL of sterile saline was used as irrigation. The soft tissue was closed with 3/0 silk suture for transalveolar procedures. The day after surgery, the patients began home use of the CHX solution (15 mL for 30 seconds) twice daily for 7 days Group 2 (n = 56): similarly to Group 1. However, in addition to CHX solution, the patients in Group 2 were prescribed Augmentin (500 mg amoxicillin trihydrate, 125 mg clavulanic acid; SmithKline Beecham) twice daily for 5 days postoperatively Group 3 (n = 59): similarly to Group 1, except for the substitution of sterile saline solution (0.09 % NaCl) for CHX Co‐interventions: all 3 groups were instructed to use only 500 mg paracetamol (Minoset; Roche) for postoperative pain relief
Outcomes	Assessment days 3 and 7 postoperatively Primary outcome measures: diagnosis of AO Secondary outcome measures: none Adverse outcomes: adverse events reported for Group 1 n = 62. Allergy: n = 0. Staining of teeth: n = 4. Mucosal irritation: n = 0. Alteration in taste: n = 12. Gastrointestinal complaints: n = 0. Bad taste: n = 8. No adverse reactions: n = 38. From Table IV page 303
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "...were randomly allocated into 3 groups" Comment: no information relating to randomisation process
Allocation concealment (selection bias)	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants in antibiotic group would not have been blinded to intervention (though this arm is not included in the present reviews meta‐analysis). Personnel were blinded to group allocation. Interventions were in identical bottles, a 3rd operator separate to clinicians undertaking the surgery and doing the assessment. Study criteria required both participant and personnel to be masked to intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting
Other bias	Low risk	The study appears to be free of other sources of bias

Divya 2019.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 50 Setting: Oral and Maxillofacial Surgery, Chennai, India Number of centres: 1 Operators: not reported Prevention or treatment of dry socket: prevention Type of teeth: impacted 3rd molars Recruitment period: not reported Funding source: none received Declarations of interest: reported as none
Participants	Inclusion criteria: patients needing surgical removal of impacted 3rd molar teeth Exclusion criteria: patients with any history of systemic illness or allergy to the components of the mouthwash were excluded from the study Gender: 298 males, 22 females Mean age: 23.5 Age range: 18 to 36 years Number randomised: 50 Number analysed: not reported Lost to follow‐up/dropouts: not reported
Interventions	All patients: none Intervention group 1: 0.2% CHX twice daily Intervention group 2: herbal mouthrinse (hioria)
Outcomes	Primary outcome measures: pain, dry socket, wound healing Diagnosis: VAS, persistent pain postextraction, bone exposure with loss of blood clot and halitosis Secondary outcome measures: none Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were randomly assigned to Group A (control) and Group B (test), with 25 patients in each group)" Comment: method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	All enrolled participants reported on
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	No other bias observed

Faizel 2015.

Study characteristics
Methods	Study design: randomised, controlled, parallel‐group trial Sample size calculation: not stated Sample size: 117 sockets Setting: Department of Oral and Maxillofacial Surgery, Peoples College of Dental Sciences and Research Centre, Bhopal, India Number of centres: 1 Operators: not stated Prevention or treatment of dry socket: treatment Type of teeth: any tooth Recruitment period: 1 January 2012 to 28 February 2013 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: all patients who reported to the department for dental extraction Exclusion criteria: pregnancy history of radiotherapy coexisting cellulitis/fascial space infections Gender: 69 females, 48 males Mean age: 34.6 years Age range: not stated Number randomised: 117. Intervention group 1: 39. Intervention group 2: 39. Control group: 39 Number analysed: 39 Lost to follow‐up/dropouts: 0
Interventions	All patients: "The socket was irrigated with warm sterile saline solution. Curettage was avoided. Loose debris was removed, taking care to avoid" Control group: "A few fibers of Alvogyl were placed with the help of a sterile instrument deep into the socket ensuring that the denuded bone was completely covered followed by the placement of sterile gauze. The gauze was removed after 5 min" Intervention group 1: "A gauze piece soaked with freshly prepared ZOE paste was placed in the extraction socket under aseptic precautions" Intervention group 2: "A single pellet of Neocone was placed inside the socket followed by the placement of a sterile gauze piece to cover the socket. The gauze was removed after 5 min"
Outcomes	Primary outcome measures: dry socket Diagnosis: "…clinically established on the basis of the following features by a blinded assessor: 1. Pain in and around the extraction socket with or without radiation that increases in severity for some period from 1 and 3 days after extraction. 2. Partial or total clot loss in the interior of the alveolus with or without halitosis" Secondary outcome measures: "Any other associated findings such as halitosis, lymphadenopathy, etc." Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly assigned using randomisation table to one of the three groups"
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a judgement in relation to participant blinding
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quotes: "The dressings were evaluated by a blinded assessor at every follow‐up visit and were changed in case of persistence of pain" "Clinical examinations for the signs of healing of dry socket were performed on 1st, 3rd, 5th, 7th, and 10th day by a blinded assessor" Comment: outcome assessor blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Data not presented clearly enough to make decision
Selective reporting (reporting bias)	Unclear risk	Data not presented clearly enough to make decision
Other bias	Unclear risk	No other bias noted

Feng 2009.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 200 patients (200 teeth) Setting: outpatients Number of centres: 1 Operators: not reported Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular molar Recruitment period: June 2007 to May 2008 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria:  agree to accept the test and sign the informed consent x‐rays showed impacted mandibular wisdom teeth Exclusion criteria:  patients with a history of acute pericoronitis in the past 10 days patients younger than 18 years old or older than 65 years old patients who smoke more than 20 cigarettes a day patients with oral contraceptives recently patients with tooth extraction contraindications patients with oral administration of any antibiotic 3 days before surgery Gender: 116 males and 84 females Mean age: not reported Age range: 20 to 65 years Number randomised: 200 Number analysed:  200 Lost to follow‐up/dropouts: 0
Interventions	All patients: the patient to return to the clinic 1 week after extraction, and may return to the clinic at any time if discomfort occurs. Record the occurrence of postoperative bleeding, swelling and dry socket. If not returned in time, contact for inquiry by phone  Control group: blank control group. Under routine local anaesthesia, the impacted teeth were removed and the alveolar bone was reduced  Intervention group 1: oral tissue patch alone. Under conventional local anaesthesia, the impacted teeth were extracted, and the oral tissue patches were cut into small pieces according to the shape of the alveolar fossa, and then placed in the root fossa of the human teeth and fitted to the bone wall  Intervention group 2: platelet rich plasma (PRP) alone. Under routine local anaesthesia, the impacted teeth were extracted, and the prepared PRP was directly injected into the alveolar socket  Intervention group 3: oral tissue patch combined with PRP group. Under routine local anaesthesia, the impacted teeth were extracted, the patch was trimmed according to the shape of the alveolar fossa, and the PRP was placed on the patch and then placed in the alveolar fossa to make the patch fit with the wall of the alveolar fossa
Outcomes	Note: authors did not specify which was the primary outcome Outcome measures:  bleeding rate after tooth extraction operation the occurrence rate of dry socket after tooth extraction tooth extraction posterior alveolar bone density Diagnosis criteria:  2 to 3 days after tooth extraction, obvious spontaneous pain began to occur and spread to the auriculotemporal region, and general analgesics did not work there was no normal blood clot in tooth extraction site, and the alveolar bone wall was exposed the wound is gray and the pain is obvious Diagnostic criteria for postoperative bleeding after tooth extraction: local bleeding on the day after tooth extraction surgery or significant local bleeding 24 hours after surgery  At 1 week and 3 months after surgery, the location of the surgical area was radiographs and analysed by Digora. The bone density of the root and canine alveolar fossa was determined, and the growth of alveolar bone in the operative area was compared  All the above tests were performed by another doctor who was not the surgeon Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Mentioned "randomly assigned to 4 groups, 50 in each group" with no further details
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported, no details about who the assessors were
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients were followed up
Selective reporting (reporting bias)	Low risk	The incidence of dry socket was reported fully in Table 2
Other bias	Low risk	No other bias noted

Freudenthal 2015.

Study characteristics
Methods	Study design: double‐blind, RCT Sample size calculation: "A power calculation showed that 100 patients were required to obtain a power of 80% and a significance level of 5% if the hypothesis was a difference of 20% between groups according to AO frequency (5% and 25%, respectively)" Sample size: 100 molars (95 patients) Setting: Oral and Maxillofacial Surgery Clinic at the South Hospital in Stockholm, Sweden Number of centres: 1 Operators: 3 senior house officers Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: February 2011 to March 2012 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: all teeth had a diagnosed pathology, such as pericoronitis, suspected cyst formation, or resorption of adjacent teeth Exclusion criteria: immunodepressive conditions pregnancy unable to follow the study because of severe illness or functional disability patients with bisphosphonate or corticosteroid intake or antibiotic treatment those who declared being allergic to CHX, paracetamol, or codeine Gender:  females (105), males (95) Mean age: 33 ± 10.03 years Age range: 19 to 65 years Number randomised: 100 teeth Number analysed:  95 teeth Lost to follow‐up/dropouts: 5 teeth (reasons: "Five teeth had to be excluded from the study owing to medical complications during surgery or the immediate postoperative period. In 2 cases, the surgery could not be completed owing to the patients’ anxiety. In 1 case, it was discovered immediately after surgery that the patient had started corticoid medication. In another 2 cases, the surgery was complicated by a postoperative purulent infection disabling a correct AO diagnosis")
Interventions	All patients: "All patients were given Alvedon (paracetamol, [Famar A.V.E, Anthoussa Plant, Greece]) 1 g before the surgical procedure… The socket was rinsed with saline" "Postoperatively, the patients were prescribed Alvedon (paracetamol 1 g) and Citodon (paracetamol 500 mg and codeine 30 mg, [Orion, Espoo, Finland]). They were informed to start using only Alvedon and to add Citodon when needed" Control group: a placebo gel (APL dentalgel, APL, Stockholm, Sweden) containing 0.2% NaF was applied Intervention group: Cervitec gel (Ivoclar Vivadent AG, Schaan, Lichtenstein) 10 mL, containing 0.2% CHX and 0.2% NaF, was placed into the alveolus
Outcomes	Primary outcome measures: dry socket Diagnosis: "postoperative pain that increased in severity with time and was accompanied by a partially or totally disintegrated blood cloth within the alveolar socket with or without halitosis" Secondary outcome measures: the intake of analgesic tablets (type and frequency) 1 to 7 days after surgery Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The code list (active gel or placebo) was developed in advance through a randomizing process"
Allocation concealment (selection bias)	Low risk	Quote: "The code list was locked away and was not unlocked until the study was completed"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Neither the surgical team nor the patients knew whether active gel or placebo was administered"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Measurements were registered by one of the surgeons not taking part in the in the actual surgical procedure"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All eligible participants reported
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	The study appears to be free of other sources of bias

Gersel‐Pedersen 1979.

Study characteristics
Methods	Study design: randomised, controlled, double‐blind, split‐mouth trial Sample size: 120 Sample size calculation: not reported Prevention or treatment of dry socket: prevention Type of teeth: bilaterally impacted mandibular 3rd molars – under local anaesthetic Conducted in: Department of Oral Surgery, Royal Dental College, Copenhagen, Denmark Number of centres: 1 Recruitment period: not stated Funding source: no financial support received
Participants	Inclusion criteria: bilateral impacted mandibular 3rd molars without acute pericoronitis Exclusion criteria: not reported Mean age: 23.2 years, range 14 to 58 years Number randomised: 120 Number analysed: 120
Interventions	Intervention (n = 120): 4 AMCA (tranexamic acid 40.0 mg) cones Control (n = 120): 4 placebo (lactose (n = 30) or sorbitol (n = 90) 40 mg) Co‐interventions: the surgical areas were irrigated with 50 mL sterile saline and closure of the wounds was accomplished with 2 silk sutures "Oral hygiene was secured by mouthrinses with chlorhexidine gluconate 0.2%, 3 times a day. The pains were alleviated with tablets containing 500 mg acetylsalicylic acid and 10 mg codeine phosphate." Frequency and quantity not reported
Outcomes	Assessment carried out ''...4, 5 or 6 days (mean = 5 days) after the operation...'' Primary outcome measures: presence of alveolitis sicca dolorosa Secondary outcome measures: none stated Adverse effects: foreign body reaction to the vehicle delivery system in the cones
Notes	2 control groups, 2 teeth per patient
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "An equal distribution of AMCA/Tplacebo between right and left side was ensured by randomization corresponding to 120 consecutive numbers"
Allocation concealment (selection bias)	Low risk	Quote: "The code was unknown to the author during the study and first broken when all the patients had been treated" Comment: allocation concealment reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for in analysis
Selective reporting (reporting bias)	Low risk	Although the study protocol was unavailable, the report appears to be free of selective reporting
Other bias	Low risk	Appears to be free of other bias

Ghaeminia 2017.

Study characteristics
Methods	Study design: RCT Sample size calculation: "calculation of the sample size was based on a randomized clinical trial investigating the usefulness of cone beam CT (CBCT) in patients with an increased risk for inferior alveolar nerve injury following the removal of mandibular third molars..." Sample size: not stated Setting: the study was performed in 3 participating departments of oral and maxillofacial surgery of Radboud University Medical Centre Nijmegen (RUN), Rijnstate Hospital Arnhem (RHA), and a private clinic in Nijmegen (PCN), the Netherlands Number of centres: 3 Operators: senior surgeons and residents Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: 2013 to 2014 stated in paper referenced CBCT Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: adults (> 18 years of age) having 1 or 2 lower 3rd molars with a close relationship with the mandibular canal Exclusion criteria: pregnancy radiological evidence of a cyst or tumour indication for removal under general anaesthesia preoperative neurosensory alterations the existence of an external CBCT Gender:  220 females, 112 males ‐ from table 3 only adds up to 332, but 33 randomised Mean age: not stated Age range: not stated Number randomised: 333 teeth (280 patients) Number analysed: intention‐to‐treat = 333, but also performed a per‐protocol (PP) analysis = 249 Lost to follow‐up/dropouts: 27 lost to follow‐up. A further 57 excluded in the PP analysis due to failure in following the protocol
Interventions	Control group: standard postoperative care instructions, without the use of a Monoject® syringe. The standard postoperative instructions were biting on a gauze for 30 minutes, no rinsing and spitting for the first 24 hours, and starting the regular toothbrushing the day after surgery. Paracetamol (4 times a day 1000 mg) in combination with ibuprofen (3 times a day 600 mg) were prescribed postoperatively Intervention group: Monoject® syringe group. After surgery, a curved tip Monoject® syringe (12 cm3) was provided to the patient. In addition to the standard postoperative care instructions, the participants received instructions with regard to the use of Monoject® syringe (by bringing the tip at the distal side of the 2nd molar in or above the tooth socket and irrigate 4 times a day with plain tap water). To avoid early removal of the blood clot, patients were instructed to start irrigating the wound 48 hours after surgery until the 1st postoperative visit 7 days after surgery
Outcomes	Primary outcome measures: the number of lower 3rd molars with postoperative inflammatory complications, which included surgical wound infection and AO Diagnosis: "Surgical wound infection was defined as the presence of a local abscess, onset of facial or cervical abscess/cellulitis, and other signs suggesting an infection (redness, swelling, purulent discharge, fever). The diagnosis of AO was based on the Blum criteria: postoperative pain in and around the extractions site, which increased in severity at any time between 1 and 3 days after the extraction, accompanied by a partially or totally disintegrated blood clot within the alveolar socket with or without halitosis" Secondary outcome measures: pain (VAS score) trismus (change in maximum interincisal distance) quality of life (Dutch version of Oral Health Related Quality of Life (OHIP‐14) number of emergency visits missed days of work or study Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "a surgical assistant assigned the patients randomly through a computer random generator after logging in the secured website"
Allocation concealment (selection bias)	Low risk	Quote: "The allocation concealment was guaranteed through the Web‐based central concealment"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "One blinded investigator per center assessed the primary and secondary outcome measures"
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were dropouts, an intention‐to‐treat analysis was performed
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Halabi 2018.

Study characteristics
Methods	Study design: randomised, double‐blind, parallel‐group clinical trial Sample size calculation: the sample size was estimated using data published previously be Halabi et al (2012), expecting an incidence reduction of 2‐3rds. The PA expected incidence of disease (AO) in the placebo group was 6.14%, while the PB expected incidence of disease (AO) in the CHX group was 2.05%. The power of the study was set at 80% (β = 0.20), with a = 0.05 as the significance level Sample size: 822 patients assessed for eligibility, 744 met inclusion criteria and completed follow‐up Setting: public community dental clinics Number of centres: 2 Operators: dental surgery team from the clinics’ emergency department Prevention or treatment of dry socket: prevention Type of teeth: any maxillary or mandibular tooth Recruitment period: April 2013 to December 2015 Funding source: self‐funded by the authors Declarations of interest: none stated
Participants	Inclusion criteria: adults aged 18 or over with clinical indications for tooth extraction and, who presented at least 1 of the following risk factors for AO: tobacco smoker (consumption of ≥ 5 cigarettes 24 hours before extraction) previous surgical site infection (clinical diagnosis of chronic periodontitis, pericoronitis, fungal infections or dental pulp gangrene) and/or traumatic extraction (lifting flap, use of elevators for > 4 minutes and/or rotary instruments) Exclusion criteria: patients requiring extraction in the operating theatre residents of rural areas with difficulty returning for follow‐up patients allergic to CHX patients under antimicrobial therapy, antibiotic prophylaxis, or antibiotic therapy after extraction Gender: 381 females, 363 males Mean age: 43.43 years Age range: not stated Number randomised: intervention group: 372; control group: 372 Number analysed: intervention group: 372; control group: 372 Lost to follow‐up/dropouts: 0
Interventions	Control group: (n = 372) rinse with 15 mL placebo (sterile water in similar bottle) for 30 seconds, twice a day for 7 days, starting 24 hours after extraction Intervention group: (n = 372) rinse with 15 mL CHX 0.12% (Oralgene mouthwash 0.12%, Maver, Chile) for 30 seconds, twice a day for 7 days, starting 24 hours after extraction
Outcomes	Primary outcome measures: presence of a dry socket Diagnosis: increasing postoperative pain intensity for 4 days within and around the socket total or partial breakdown of the blood clot in the socket with or without bone exposure Secondary outcome measures: none reported Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "To guarantee that in both groups (treatment and placebo) the risk of alveolar osteitis was similar and comparable, the assignment was performed by randomization, stratified by risk factors" Comment: method of randomisation not described
Allocation concealment (selection bias)	Low risk	Quote: "We stored black envelopes in a box containing a paper with the letter C for chlorhexidine or P for placebo (half of each). The envelopes were chosen for each patient after the extraction and transported to another room (without opening them)..."
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no dropouts reported
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Haraji 2013.

Study characteristics
Methods	Study design: randomised, split‐mouth study Sample size calculation: "A pilot study of 45 individuals was undertaken. The proportions of dry socket in the control and experimental sides were 35.6% and 13.3% respectively. On the basis of this pilot study, a sample size of 76 patients (76 x 2 sockets) was needed to obtain a 0.9 power (α = 0.5, β = 0.1). Thus it was prospectively determined as 160 extraction sites in 80 patients to gain test powers ≥ 0.9" Sample size: 80 patients (160 teeth) Setting: a private maxillofacial clinic in Tehran, Iran Number of centres: 1 Operators: not stated Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: 2010 to 2011 Funding source: none stated  Declarations of interest: none stated
Participants	Inclusion criteria: not specifically stated Exclusion criteria: unwillingness to participate unwillingness/refusal to avoid consuming analgesics/antibiotics after surgery failure to attend follow‐up sessions (day 1 and 3 postoperatively) presence of systemic disease/infection ingestion of any medication (including antibiotics and anticoagulants) during the 4 days before the operation existence of any condition that contraindicated surgery presence of any psychological conditions receiving sedatives/analgesics/psychiatric medications presence of pain inducing conditions such as aching teeth, recurrent apthous stomatitis or cheek biting the need for antibiotic prophylaxis pregnancy being immunodepressed having allergy to any medications Gender: 41 females, 39 males Mean age: 21.6 ± 2.5 years Age range: 17 to 31 years Number randomised: 80 Number analysed: 80 Lost to follow‐up/dropouts: 4 (failure to attend follow‐up) – unclear whether this is from the 200 screened or 80 included
Interventions	Control group: gelatin sponge with colloidal silver (Gelatamp, Rocko) Intervention group: gelatin sponge with colloidal silver (Gelatamp, Rocko) impregnated with 0.2% CHX bioadhesive gel (1,6‐bis|N‐p‐chlorphenyl‐biguanidol hexane digluconate|, Kimia)
Outcomes	Primary outcome measures: dry socket Diagnosis: "Diagnosis of dry socket was regarded as positive when the patient experienced postoperative pain that intensified sometime between the first and third days, with total or partial loss of the blood clot" Secondary outcome measures: pain. "..recorded on a visual analogue scale (VAS). The end points were considered 'no pain' and 'intolerable pain'. The results on the VAS were later converted to 10 ordered ranks (0‐9)" Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation via a coin toss used to allocate to groups
Allocation concealment (selection bias)	High risk	Allocation based on a coin toss by person performing intervention
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Both the surgeon and patients were blinded to the assignment orders"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Subjects were evaluated in the first and third post‐operative days. Clinical assessments were performed by the blinded maxillofacial surgeon…"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All eligible participants reported
Selective reporting (reporting bias)	Low risk	All outcomes were reported
Other bias	Low risk	The study appears to be free of other sources of bias

Hasheminia 2018.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 189 patients Setting: private dental clinic Number of centres: 1 Operators: not reported Prevention or treatment of dry socket: prevention Type of teeth: impacted 3rd molars Recruitment period: not reported Funding source: not reported Declarations of interest: none reported
Participants	Inclusion criteria: patients having impacted 3rd molars removed teeth categorised as class A and B in Pell and Gregory classification Exclusion criteria: smokers patients who are taking antibiotics or oral contraceptives women in the first 22 days of menstrual cycle Gender: 80 men, 109 women Mean age: 31.3 years (intervention group), 30.4 years (control group) Age range: 22 to 46 years Number randomised: 189 Number analysed: 189 Lost to follow‐up/dropouts: 0
Interventions	All patients: 400 mg ibuprofen given 1 hour before the procedure. All patients received 400 mg ibuprofen 6 hourly for 3 days Control group: (n = 92) no intervention Intervention group: (n = 97) gauzes soaked in 1% povidone iodine oral antiseptic solution (Betadine, Purdue Pharma LP, CT, USA) were placed on the 3rd molars and adjacent teeth for 2 to 5 minutes. During the procedure povidone iodine 1% mouthwash was also used
Outcomes	Primary outcome measures: presence of dry socket Diagnosis: criteria used not detailed Secondary outcome measures: none Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were grouped into the experimental and control groups using random odd and even number" Comment: method of randomisation is not clearly reported
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no dropouts reported
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Hermesch 1998.

Study characteristics
Methods	Study design: randomised, double‐blind, placebo‐controlled, parallel‐group study Sample size: 279 Sample size calculation: not reported Conducted in: a military dental clinic (authors from San Antonio, Texas USA) Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molar ‐ under local anaesthesia with optional intravenous sedation. However, other 3rd molars extracted concurrently Recruitment period: not stated Funding source: not reported
Participants	Inclusion criteria: at least 1 impacted mandibular 3rd molar over 18 years of age in good health Exclusion criteria: pericoronitis medical conditions requiring antibiotic prophylaxis participants taking antibiotics 2 weeks prior pregnant or lactating females allergy to CHX Number randomised: 279; 0.12% CHX n = 140; placebo n = 139 Number analysed: 271; 0.12% CHX n = 136; placebo n = 135 Lost to follow‐up/dropouts: 8, reasons for exclusion provided
Interventions	Comparison: Group 1 (n = 140): 0.12% CHX gluconate rinse in 11.6% alcohol Group 2 (n = 139): placebo rinse containing 11.6% alcohol Co‐interventions: prior to surgery: daily rinse for 7 days, 15 mL of test product for 30 seconds twice a day postoperative: Daily rinse for 7 days,15 mL of test product 30 seconds twice a day
Outcomes	3 to 4 days postoperatively telephone contact. Day 7 ‐ clinical evaluation Primary outcome measures: presence or absence of alveolar osteitis pain not relieved by analgesics Secondary outcome measures: not evaluated Adverse effects: collected and reported for both intervention groups, but unclear to what extent these were intervention related or essentially postoperative complications (Additional Table 7)
Notes	NB: number above = patients, extractions: CHX 239, placebo 240. Randomisation at individual participant level but analysis of data at extraction site level
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Each subject was then randomized to a treatment group within these strata" Comment: method of sequence generation not reported
Allocation concealment (selection bias)	Low risk	Quote: "Each bottle was labelled with a unique subject number, which allowed for blinding of treatment assignment"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Each bottle was labelled with a unique subject number, which allowed for blinding of treatment assignment" Comment: both the placebo and active interventions were identical in appearance and therefore participants were blinded to group allocation, however it is not clear how bottle labelling ensured blinding of personnel. Strict study criteria required that participants and personnel would be blinded to group allocation. Lack of masking of operator is unlikely to have led to deviation in intervention
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All diagnoses were made by 1 of 2 attending clinical examiners involved in this study, without knowledge of treatment assignment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Participants not included in final analysis reported (n = 8) and reasons for exclusion provided. Equal numbers in both intervention groups
Selective reporting (reporting bias)	High risk	271 participants, 973 third molars extracted, 51.7% (239 CHX, 240 placebo) were mandibular molars, no data were reported for the non‐mandibular molars
Other bias	High risk	Randomisation at individual participant level but analysis of data at extraction site level

Hita‐Iglesias 2008.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled trial Sample size: 73 Sample size calculation: not reported Conducted in: Faculty of Odontology at the University of Seville and the Oral and Maxillofacial Surgery service of the Virgen de Rocio University Hospitals, Seville, Spain Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars, under local anaesthesia. Difficulty index 4‐7 Koerner scale Recruitment period: from June to November 2005 Funding source: Laboratorios Lacer Barcelona España, donated the CHX samples used in this study
Participants	Inclusion criteria: mandibular 3rd molars; difficulty index 4‐7 Koerner scale (Koerner 1994), surgical extraction and suture Exclusion criteria: antibiotic or analgesic 4 days prior AIDs or any type of immunosuppression pregnant/lactating women allergy to: CHX, articaine, paracetamol or ibuprofen patients requiring simultaneous extraction of 2 third molars any jaw associated pathology operation times > 30 minutes non‐co‐operative patients (psychic‐motor dysfunction) Age: 18 to 60 years, mean 29 years Gel group 1: randomised 41; analysed 39 Mouthrinse group 2: randomised 32; analysed 31 Number randomised: 73 Number analysed: 70 Lost to follow‐up/dropouts: 3
Interventions	Comparisons: CHX bioadhesive gel versus mouthrinse Group 1 (n = 41): 0.2% CHX bioadhesive gel, placed in wound after extraction then self‐applied to the wound twice a day (day 1 to 7) beginning same day as intervention Group 2 (n = 32): 0.12% CHX mouthrinse twice a day (day 1 to 7) beginning same day as intervention Co‐interventions: ibuprofen 600 mg 3 times a day, paracetamol 500 mg with codeine 14.05 mg 3 times a day
Outcomes	Subjects evaluated on the 3rd and 7th day postoperatively Primary outcome measures: incidence of AO day 1 to day 3 Secondary outcome measures: not reported Adverse effects: Quote: "tolerance to the treatment defined as the frequency that patients developed one or more adverse effects .. assessed on a verbal score 1 (max tolerance) to 5 (max intolerance) during the 3rd and 7th postoperative day" page 443 Comment: "tolerance" was not defined in the report nor the type of "adverse effects" which might constitute maximum intolerance, and the relevant data reported in Table 1 were not readily interpretable
Notes	November 2012: communication from Dr Torres‐Lagares confirmed that Hita‐Iglesias 2008; Torres‐Lagares 2006a and Torres‐Lagares 2006b were separate studies each involving a different group of participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomly classified into two groups by means of a simple allocation system using a computer program" Comment: probably done
Allocation concealment (selection bias)	Low risk	Quote: "Having carried out the procedure the envelope corresponding to the patient code was opened" Comment: allocation concealment described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants could not be blinded to group allocation, nor could operators. Strict study criteria required that participants and personnel would be blinded to group allocation as they will be involved in reporting of patients symptoms
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "All clinical assessment was carried out by a single blind investigator previously trained on diagnostic criteria"
Incomplete outcome data (attrition bias) All outcomes	Low risk	The dropouts/losses to follow‐up were reported. Similar numbers in both groups, 2 in gel group, 1 in rinse group
Selective reporting (reporting bias)	Low risk	Although the study protocol was unavailable, the report appears to be free of selective reporting
Other bias	Low risk	Appears to be free of other potential sources of bias

Hu 2005.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 200 patients (200 teeth) Setting: outpatient Number of centres: 1 Operators: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular molar Recruitment period: May to September 2004 Funding source: none stated Declarations of interest: none stated
Participants	Inclusion criteria: agree to accept the test and sign the informed consent x‐rays showed impacted mandibular wisdom teeth Exclusion criteria: patients with a history of acute pericoronitis in the past 10 days patients younger than 18 years old or older than 65 years old patients who smoke more than 20 cigarettes a day patients with oral contraceptives recently patients with tooth extraction contraindications patients with oral administration of any antibiotic 3 days before surgery Gender: 97 males and 103 females Mean age: 30 Age range: 18 to 64 years Number randomised: 200 Number analysed: 200 Lost to follow‐up/dropouts: 0
Interventions	All patients: postoperative medical advice: spit out all the tampons after 30 minutes; not to gargle on that day; return visit at any time if there is anything unusual, otherwise, a week later. Patients who smoked were asked not to smoke for 3 days after surgery Control group: blank control group (no description) Intervention group 1: oral tissue patch The impacted tooth was removed by the same surgeon under routine local block anaesthesia. After the reduction of the alveolar bone, the oral tissue patch was trimmed into the shape of the alveolar fossa and implanted into the alveolar fossa. (If the alveolar interval is too high, in order to avoid the falling off of the implanted oral tissue patch, the oral tissue patch should be trimmed into several parts according to the number of tooth roots pulled out and placed into the root fossa respectively, and fitted to the bone wall of the alveolar fossa. After the patch is covered with blood, a cotton strip should be placed on the wound). All processes are performed by the same person
Outcomes	Note: authors did not specify which was the primary outcome Outcome measures: the occurrence rate of dry slot after tooth extraction bleeding rate after tooth extraction operation Diagnosis criteria: 2 to 3 days after tooth extraction, obvious spontaneous pain began to occur and spread to the auriculotemporal region, and general analgesics did not work there was no normal blood clot in tooth extraction trauma, and the alveolar bone wall was exposed the wound is gray and the pain is obvious Diagnostic criteria for postoperative bleeding after tooth extraction: local bleeding on the day after tooth extraction surgery or significant local bleeding 24 hours after surgery Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "200 random numbers were read from any column or any row in the random number table, and then divided each number by 2. Patients were assigned to the two groups depended on the remainder being odd or even" Comment: randomisation method not clear
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The extractions were done by 1 operator who was not involved in the assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients were followed up
Selective reporting (reporting bias)	Low risk	The incidence of dry socket was reported fully in Table 1
Other bias	Low risk	No other bias noted

Huang 2011.

Study characteristics
Methods	Study design: parallel‐group RCT Sample size: 80 Sample size calculation: not reported Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molar, local anaesthesia with lidocaine Conducted in: Department of Stomatology, the First Affiliated Hospital of Sun Yat‐sen University, Guangzhou, China Number of centres: 1 Recruitment period: May 2008 to July 2010 Providers of care: not reported Funding source: scientific plan of Cantoon Province
Participants	Inclusion criteria: participants required to have their impacted mandibular 3rd molar extracted aged 18 to 45 years horizontal impacted 3rd mandibular molar on x‐ray agreed to participate and could be followed‐up Exclusion criteria: have contraindication and could not be followed‐up Number randomised: 80 Number evaluated: 80 Lost to follow‐up/dropouts: 0
Interventions	Group 1 (n = 40): acellular dermis matrix (2 × 2.5 cm) embedded into the extraction sockets Group 2 (n = 40): nothing embedded into the extraction sockets Co‐interventions: extraction of the teeth via flap elevation, bone removing and tooth splitting
Outcomes	When measured: 1 week after the extraction Primary outcome measures: presence of postoperative alveolitis Secondary outcome measures: none Adverse outcomes: none reported
Notes	Author contact failed
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The participants were randomized into two groups with 40 in each group" Comment: the detailed methods of randomisation were not clearly reported
Allocation concealment (selection bias)	Unclear risk	It is not clear from the report how allocation concealment was achieved
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were not blinded to group allocation (placebo received no intrasocket intervention), personnel were not blinded to group allocation. Strict study criteria required that participants and personnel would be blinded to group allocation as they will be involved in reporting of patients symptoms
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear whether assessors were blinded to group allocation
Incomplete outcome data (attrition bias) All outcomes	Low risk	None of the participants dropped out
Selective reporting (reporting bias)	Low risk	All the outcomes were reported clearly
Other bias	High risk	Authors stated that the gender and the age in each group were similar at baseline, but the difficulties of the extraction and the detailed information of the tooth impaction were not reported

Karabit 2019.

Study characteristics
Methods	Study design: randomised, split‐mouth study Sample size calculation: not reported Sample size: 53 Setting: Department of Oral Maxillofacial Surgery, Syria Number of centres: 1 Operators: 1 surgeon Prevention or treatment of dry socket: prevention Type of teeth: bilateral mesially impacted mandibular molars Recruitment period: 1 August 2016 to 5 January 2017 Funding source: reported as none Declarations of interest: not reported
Participants	Inclusion criteria: smoker Exclusion criteria: patient on antibiotic prophylaxis, known to be allergic to CHX, with systemic disorders, epinephrine contraindications or breastfeeding women and who are on oral contraceptives Gender:  31 males, 22 females Mean age: not reported Age range: 18 to 26 years Number randomised: 53 Number analysed: 53 Lost to follow‐up/dropouts: 0
Interventions	All patients: none Intervention group 1: 0.12% CHX gluconate Intervention group 2: mint flavoured aqua distillate
Outcomes	Primary outcome measures: pain Diagnosis: VAS Secondary outcome measures: presence of clot, exposed alveolar bone Adverse outcomes: visual examination
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Surgical site was chosen randomly by asking the patients to choose one card out of two in which site was written on its back (left or right)." Group allocation patients "divided randomly into two parallel group and we asked the patients to choose one card out of two in which the bottle number was written on its back"
Allocation concealment (selection bias)	High risk	Quote: "Surgical site was chosen randomly by asking the patients to choose one card out of two in which site was written on its back (left or right). ....we asked the patients to choose one card out of two in which the bottle number was written on its back" Comment: unlikely to be concealed to operator
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants appeared blinded Quote: (page 30) "Group A: the patients received a bottle of 0.12% CHXG mouthwash with mint flavor to start using it on the day before and the day after the extraction twice daily for 7 days. Group B: the patients received a bottle of Aqua distillate with mint flavor to start using as directed in bottle A" Operators appear blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear who performed clinical assessments and whether blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All eligible participants reported
Selective reporting (reporting bias)	Low risk	All planned outcome observations reported
Other bias	Low risk	The study appears to be free of other sources of bias

Kaya 2011.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled trial Study size: 104 Sample size calculation: not reported Conducted in: Department of Oral and Maxillofacial Surgery, Atatürk University Faculty of Dentistry, Turkey Number of centres: 1 Prevention or treatment of dry socket: treatment Type of teeth: mandibular permanent 1st molar Recruitment period: 18 months, January 2008 to July 2009 Funding source: not reported
Participants	Inclusion criteria: 18 years or older the ability to understand verbal and written instructions previously diagnosed, but untreated AO in the mandibular permanent 1st molar extraction socket Exclusion criteria: previous radiotherapy any medical condition that could affect AO treatment (e.g. bone pathologic features, vascular or haematologic disorders, diabetes mellitus) the use of antibiotics pregnancy or lactation an allergy to iodine, eugenol, or paracetamol patients who smoked, used oral contraceptives, were menstruating, or would require a surgical flap to remove the tooth were excluded Number randomised: 104 Number evaluated: 104 Lost to follow‐up/dropouts: 0
Interventions	Group 1 (n = 26): curettage and irrigation alone Group 2 (n = 26): curettage and irrigation followed by Alvogyl applied directly to the socket Group 3 (n = 26): curettage and irrigation followed by SaliCept patch applied directly to the socket Group 4 (n = 26): curettage and irrigation followed by low‐level laser therapy irradiation Co‐interventions: patients were allowed 500 mg of acetaminophen as a rescue medication, as required, and were instructed to record how many times daily the medication was used. Additional follow‐up visits were organised through the department, as necessary
Outcomes	Primary outcome measures: pain (VAS) ‐ assessed each morning for 1 week AO – days 3 and 7 Secondary outcome measures: not stated Adverse outcomes: not stated
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were randomly assigned…" Comment: method not clear
Allocation concealment (selection bias)	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for in analysis
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting
Other bias	Low risk	The study appears to be free of other sources of bias

Keshini 2020.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 30 Setting: not reported Number of centres: not reported Operators: not reported Prevention or treatment of dry socket: treatment Type of teeth: not reported Recruitment period: not reported Funding source: not detailed Declarations of interest: none
Participants	Inclusion criteria: patients reported with postextraction pain, who fit criteria for dry socket Exclusion criteria: patients under 14 and over 60 years of age Gender: not given Mean age: not given Age range: not given Number randomised: 30 Number analysed: 30 Lost to follow‐up/dropouts: none reported
Interventions	All patients: gentle irrigation with warm saline and debridement Intervention group 1:. n = 15, socket packed with Alvogyl and sutured with 3‐0 mersilk Intervention group 2: n = 15, platelet rich fibrin membrane placed in socket and sutured with 3‐0 mersilk
Outcomes	Primary outcome measures: pain and wound healing  Diagnosis: VAS score, wound healing assessed by number of socket walls exposed Secondary outcome measures: none Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: (page 321) "A total of thirty patients were included in the study ....were randomly divided into two Groups (Group A and Group B), with 15 patients in each group using simple randomization" Comment: method not described
Allocation concealment (selection bias)	Unclear risk	Quote: (page 321) "A total of thirty patients were included in the study ....were randomly divided into two Groups (Group A and Group B), with 15 patients in each group using simple randomization" Comment: not mentioned and no information to suggest this was done
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear who operator is who performed assessments and whether operator was blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	None of the participants dropped out
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Unclear risk	Sparse reporting of details in relation to participants, setting, teeth extracted, operators and comparison between the 2 groups

King 2018.

Study characteristics
Methods	Study design: RCT Sample size calculation: "The data obtained by Kaya et al were used as reference to calculate the statistical power in the present study. It was assumed that a sample size of 40 sockets (20 per each group), with a standard deviation of 2 and a decrease in pain of 6 points in the VAS, would give a statistical power of 86.9%. Therefore, the aim was to recruit 20 sockets per group with the expectation that obtaining data from 17 per group would give a statistical power of 80% with the P value set at 0.05" Sample size: 38 patients (44 sockets) Setting: UK dental hospital Number of centres: 1 Operators: all initial treatment was conducted by the lead clinician, a qualified dentist Prevention or treatment of dry socket: treatment Type of teeth: incisors, premolars and molars Recruitment period: from the beginning of December 2014 to the end of February 2015 Funding source: "the PRGF consumables used in this study were supplied by the BTI Biotechnology Institute (San Antonio, Spain)" Declarations of interest: "the PRGF consumables used in this study were supplied by the BTI Biotechnology Institute (San Antonio, Spain)"
Participants	Inclusion criteria: at least 18 years old diagnosed with AO after tooth extraction by dental clinicians in the oral surgery department of the dental hospital AO diagnosis was based on patients having pain and exposed bone; if either of these parameters was missing, patients were not accepted for the study Exclusion criteria: breastfeeding bleeding disorders immunocompromised patients current or recurrent disease or dental pathology that could affect the assessments severe or unstable physical or psychiatric illness allergy or intolerance to study materials a recent history of alcohol or substance abuse Gender: 18 females, 20 males Mean age: 40.7 ± 17.3 years Age range: not stated Number randomised: 44 Number analysed:  44 Lost to follow‐up/dropouts: 1 patient did not attend the 1st review appointment and another patient did not attend the 2nd review appointment. The data from the review appointments when the patients did attend were included in the statistical analysis
Interventions	Control group: the socket was irrigated with sterile saline and packed with Alvogyl approximately 0.20 g to fill the defect according to the manufacturer’s instructions Intervention group: plasma rich in growth factors (PRGF) 4.9 mL tubes of venous blood containing 3.8% sodium citrate were collected. The blood was centrifuged at 580g for 8 minutes at room temperature (PRGF Model System Centrifuge IV, BTI Biotechnology Institute). The 2 plasma fractions, F1 and F2, were separated using a PRGF Plasma Transfer Device supplied with the PRGF‐Endoret kit. Care was taken not to include leukocytes (the buffy coat) when extracting F2. F1 was poured into a flat glass dish and placed in a PlasmaTherm H Oven heated to 37°C for 15 minutes. The F1 clot underwent retraction to produce the F1 membrane and the liquid supernatant. Then, F2 was activated to induce fibrin clot formation. Before placement of the fibrin clot, PRGF liquid supernatant was used to irrigate the socket. Then, the F1 membrane was applied to cover the surgical area and the socket margins were sutured
Outcomes	Outcome measures: exposed alveolar bone halitosis and dysgeusia inflammation was assessed on a scale of 0 to 3 as presented pain ‐ patients were asked to complete a VAS on a 100‐mm line representing the spectrum of pain from no pain at all to the worst pain imaginable patient rated swelling, bruising, bleeding, and speed of healing using the VAS Adverse outcomes: no adverse events reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "computer‐generated, block randomization schedule"
Allocation concealment (selection bias)	Low risk	Quote: "Randomization numbers were assigned by study staff at the study site in ascending numerical order as patients were determined to be fully eligible to participate in the study"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Given the nature of the interventions it would not have been possible to blind the operator and patients. Lack of masking of participant and operator is unlikely to have led to deviation in intervention. Strict study criteria required that participants and personnel would be blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "A second clinician, a qualified dentist, who was blinded to the original treatment, conducted the review appointments" Comment: pre‐treatment assessments were reported to have been undertaken by the treating clinician before randomisation and treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "One patient did not attend the first review appointment and another patient did not attend the second review appointment. The data from the review appointments when the patients did attend was included in the statistical analysis" Comment: this suggest an intention‐to‐treat analysis was performed
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Unclear risk	Analgesia use was not reported and could have had an impact on pain reported by participants

Kjellman 1973.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled trial Sample size: 100 Sample size calculation: not reported Conducted in: Department of Oral Surgery at Södersjukhuset Stockholm, Sweden Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molars, degree of difficulty unspecified, under local anaesthesia Recruitment period: not stated Funding source: not stated
Participants	Inclusion criteria: impacted mandibular 3rd molars Exclusion criteria: cases that have required relatively major operations on bone tissue Number randomised: 100 Number evaluated: 100
Interventions	Group 1 (n = 50): Apernyl cone, consisting of acetylsalicylic acid and p‐oxybenzoic acid (preservative) inserted immediately after extraction Group 2 (n = 50): placebo without acetylsalicylic acid inserted immediately after extraction Co‐interventions: concomitant pain medication unreported
Outcomes	Assessments performed on days 1, 2 and 4 postoperatively Primary outcome measures: the occurrence of postoperative periosteitis "dry socket" Secondary outcome measures: not evaluated Adverse effects: pain and burning sensation
Notes	Inadequately reported, sparse trial details and incomplete and largely unusable data
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "...with the help of a randomising procedure..." Comment: method not described
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a clear judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not mentioned and no information to suggest this was done
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Very limited data, reported as graph plots. Insufficient information to make a clear judgement
Selective reporting (reporting bias)	Low risk	Although the protocol was unavailable all of the outcomes specified in the methods section appear to have been reported
Other bias	Low risk	Appears to be free of other potential sources of bias

Larsen 1991.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled trial Sample size: 150 Sample size calculation: none reported Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molars Conducted in: not stated Number of centres: not stated Recruitment period: not stated Funding source: Procter and Gamble Co, Cincinnati, USA
Participants	Inclusion criteria: bilaterally impacted mandibular 3rd molars Exclusion criteria: acute infection participants on antibiotics those requiring prophylaxis Number randomised: 150 Number analysed: 139 Lost to follow‐up/dropouts: 11
Interventions	Group 1 (n = 72): 0.12% CHX gluconate 15 mL for 30 seconds a day for 1 week prior to surgery and 1 week after Group 2 (n = 67): placebo rinse (identical appearance) Concomitant medication: preoperative: 8 mg dexamethasone intravenous postoperative: 325 mg acetaminophen plus 30 mg codeine when necessary
Outcomes	Assessment 1 week postoperatively or earlier if pain Primary outcome measures: presence of dry socket Secondary outcome measures: none stated Adverse effects: none stated
Notes	Sponsored by Procter and Gamble Co, Cincinatti, USA Randomisation was by patient but analysis by sites which will narrow the confidence intervals slightly
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: "using a computer software routine supplied by the sponsor ..." " Within strata subjects were randomly assigned into two groups"
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a judgement
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Patients were unaware of group allocation, personnel were unaware of group allocation. Active and placebo were identical in appearance
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear who made the assessment of outcomes, while operators were blinded to allocation and appearance of placebo was identical, unlikely that they would have known group allocation when making assessment, however insufficient information to make a judgement
Incomplete outcome data (attrition bias) All outcomes	High risk	No data available for dropouts (11): unreported from which groups, when and the reasons
Selective reporting (reporting bias)	Low risk	Although the study protocol was unavailable, the report appears to be free of selective reporting of the outcomes
Other bias	High risk	Participants were stratified then randomised but subsequent analyses were at tooth level. Although we know how many sites were available for analysis, we only know the baseline number of patients

Lenka 2019.

Study characteristics
Methods	Study design: RCT Sample size calculation: not stated Sample size: 30 Setting: not stated Number of centres: 1 Operators: not stated Prevention or treatment of dry socket: treatment Type of teeth: not stated Recruitment period: not stated Funding source: self‐funded Declarations of interest: stated as none
Participants	Inclusion criteria: pain around the extraction socket after 1 day, foul smell, total or partial loss of clot Exclusion criteria: not stated Gender: not stated Mean age: not stated Age range: not stated Number randomised: 30 Number analysed: not stated Lost to follow‐up/dropouts: not stated
Interventions	All patients: local anaesthesia, saline irrigation Intervention group 1:. zinc oxide eugenol paste/gauze pieces dressing Intervention group 2: Alvogyl
Outcomes	Primary outcome measures: pain Diagnosis: VAS Secondary outcome measures: none Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patient selected for the study was those who have severe pain after one day of extraction and proper examination the diagnosis was dry socket are equally and randomly to control and test groups" Comment: no information to make a decision
Allocation concealment (selection bias)	Unclear risk	No information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no dropouts reported
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	No other bias observed

Metin 2006.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled, double‐blind, single‐centre trial Sample size: 99 Sample size calculation: not reported Conducted in: Department of Oral and Maxillofacial Surgery, Faculty of Dentistry, Ondokuz, Mayis, University of Samsun, Turkey Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molars, under local anaesthesia Recruitment period: not stated Funding source: not reported
Participants	Inclusion criteria: patients requiring surgical removal of impacted mandibular 3rd molars Exclusion criteria: acute infection pericoronitis patients using antibiotics or requiring antibiotic before treatment Number randomised: 99 Number analysed: 99 Lost to follow‐up/dropouts: 0
Interventions	Comparison: pre and postoperative CHX rinse versus postoperative only Group 1 (n = 46): rinsed with CHX 0.2% 15 mL for 30 seconds twice per day for both 1 week prior to and 1 week after surgery Group 2 (n = 53): rinsed with CHX 0.2% 15 mL for 30 seconds twice per day for 1 week after surgery Co‐interventions: after removal of the teeth, the surgical sites were rinsed with 10 mL of sterile saline solution, and the soft tissue was closed and sutured with 3‐0 silk suture
Outcomes	On the 7th day (or on preceding days if pain was present), the extraction sites were evaluated Primary outcome measures: incidence of AO Secondary outcome measures: not evaluated Adverse effects: altered taste and numbness page 3. Numbness in the tongue reported in Group 1 and Group 2 45.6% and 13.2%. However, disturbance of taste sensation was seen in 56.5% of the patients in Group 1 and in 11.3% of the patients in Group 2
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were randomly assigned into two groups" Comment: no information provided on how
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a clear judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants may have known their group allocation, it is not clear if personnel knew group allocation. A strict study criteria require that participants and personnel are blinded to group allocation but unclear reporting and unclear impact on performance bias. Lack of masking of patient and operator is unlikely to have led to deviation in intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "The same examiners made all the diagnoses" Comment: It would appear that operators made the assessment of the outcome
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomised participants included in the outcome evaluation
Selective reporting (reporting bias)	Low risk	Although the protocol was unavailable, all of the outcomes specified in the methods section appear to have been reported
Other bias	Low risk	Appears to be free of other potential sources of bias

Mitchell 1984.

Study characteristics
Methods	Study design: randomised, parallel‐group, placebo‐controlled, double‐blind trial Sample: 64 Conducted in: Oral Surgery Department, Newcastle Dental Hospital, UK Number of centres: 1 Prevention or treatment of dry socket: treatment Type of teeth: no tooth specified Recruitment period: not stated Funding source: not reported
Participants	Inclusion criteria: a diagnosed dry socket Exclusion criteria: none stated Number randomised: 64, 1 lost to follow‐up but group allocation not stated Group 1 (metronidazole): randomised not reported; 6 lost to follow‐up; analysed 26 (18%) Group 2 (Orabase placebo): randomised 32; 3 lost to follow‐up; analysed 29 (10%) Number analysed: 55
Interventions	Comparison: metronidazole (10%) versus Orabase alone Group 1 (n = 26): metronidazole (10%) in carboxymethylcellulose gelatin (Orabase) paste Group 2 (n = 29): Orabase paste alone Dressing syringed into the sockets Co‐interventions: warm saline irrigation at presentation
Outcomes	Review at 2 days post‐intervention initially and then re‐application dressing where necessary and review until cure Primary outcome measures: absence of pain Secondary outcome measures: treatment time in days number of visits to effect resolution extraction time to start of treatment Adverse effects: not reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were randomly allocated to the test or placebo group and the operator was unaware of the code breaker"
Allocation concealment (selection bias)	Low risk	Quote: "The patients were randomly allocated to the test or placebo group and the operator was unaware of the code breaker" Comment: probable allocation concealed
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The test and control pastes were packaged in identical 2 mL syringes. Participants and personnel are blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Operator unaware of group allocation and code only broken after study was complete
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 dropout reported but not clear from which group, all other dropouts fully reported. Unlikely to have had large impact on data analysis
Selective reporting (reporting bias)	Low risk	Although the study protocol was unavailable, the report appears to be free of selective reporting of the outcomes
Other bias	Low risk	Appears to be free of other potential sources of bias

Ragno 1991.

Study characteristics
Methods	Study design: randomised, controlled, double‐blind, parallel‐group study Sample size calculation: not reported Conducted in: Department of Oral and Maxillofacial Surgery, Walter Reed Army Medical Centre, Washington DC, USA Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars, under intravenous sedation and local anaesthesia Operator: resident surgeon for 8 years Recruitment period: July 1987 to April 1989 Funding source: not stated
Participants	Inclusion criteria: > 18 years of age negative medical history not taking any medication apart from birth control pill 2 mandibular 3rd molars for extraction Exclusion criteria: not stated Mean age: not reported Intervention Group: number randomised 40; analysed 40 Control Group: number randomised 40; analysed 40 Number randomised: 40 patients, but sites not (80) Number analysed: 40 patients, but sites not (80)
Interventions	Comparisons: on the day of surgery patients first rinsed with 15 mL of designated solution, teeth were then extracted, surgical sites irrigated with 15 mL of designated solution and soft tissue closed and sutured. The day after surgery patients began home use of solutions Group 1 (n = 40; 80 surgical sites): 0.12% CHX rinse, 15 mL twice daily for 7 days postoperatively Group 2 (n = 40; 80 surgical sites): placebo rinse, 15 mL twice daily for 7 days postoperatively Co‐interventions: none stated Concomitant interventions: none stated
Outcomes	Postoperative examination on days 3 and 7 Primary outcome measures: AO Secondary outcome measures: postoperative questionnaire day 7, though not reported but presume it relates to adverse events associated with mouthrinse Adverse effects: there were no allergic reactions to the CHX rinse. 3 participants reported bad taste, 1 reported stomach upset (page 526) but no staining noted. 1 person in the control had a severe surgical reaction which in the author's opinion was not attributable to the medication
Notes	On the day of the procedure patients first rinsed with 15 mL of their assigned solution for 30 seconds, after the procedure the sites were rinsed with 15 mL of the same solution
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: ''The patients within each group were randomly assigned" Comment: no mention of process
Allocation concealment (selection bias)	Low risk	Quotes: "The pharmacist manufactured the placebo and maintained the code for patient assignments"...."Decoding of the patient assignments revealed..." Comment: probably done
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants masked to group intervention. Not clear that personnel were blinded to group allocation. It is unlikely that lack of masking of operator would have led to deviation in intervention. Strict study criteria required that participants and personnel would be blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not clear who assessor was, not clear blinded to group allocation
Incomplete outcome data (attrition bias) All outcomes	Low risk	No patient lost to follow‐up
Selective reporting (reporting bias)	High risk	Postoperative questionnaire completed day 7, no data reported
Other bias	Low risk	Appears to be free of other potential sources of bias

Reekie 2006.

Study characteristics
Methods	Study design: randomised, parallel‐group, double‐blind, placebo‐controlled study Sample size: 302 Sample size calculation: undertaken Conducted in: 3 dental practices in the UK Number of centres: 3 Prevention or treatment of dry socket: prevention Type of teeth: routine non‐surgical extraction of 1 or more molar or premolar tooth under local anaesthetic Recruitment period: 2000 to 2003 Funding source: a grant from the Oral and Dental Research Trust
Participants	Inclusion criteria: non‐surgical extraction of 1 or more molar/premolar tooth under local anaesthetic Exclusion criteria: grade 3 mobile teeth participants on warfarin, nicoumalone, phenytoin, fluocil, lithium and cimetidine pregnant/breastfeeding women participants with intellectual impairment Number randomised: 302 Number analysed: 27 Only patients calling back with pain were examined and assessed. It was assumed all others did not have dry socket
Interventions	Comparison: metronidazole 25% gel versus placebo gel Group 1 (n = 152): 0.25 ml of 25% metronidazole gel (62.5 mg) Group 2 (n = 150): placebo gel (KY jelly) Gel syringed into socket immediately after dental extraction Mean age: 49.5 years, SD 14.77 (range 19 to 93) Co‐interventions: not stated
Outcomes	Assessment made only on participants with severe pain and who phoned the surgery. Only patients with the classical signs of dry socket were reported Primary outcome measures: presence of dry socket Secondary outcome measures: none assessed or reported Adverse effects: not selected as an outcome but were reported (see Notes)
Notes	Adverse events: reported: 1 participant with nausea and vomiting, 2 complained of a bitter taste
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Blinding and randomisation was achieved by the manufacturer ....The syringes were allocated a code number derived from a random number sequence" Comment: probably done
Allocation concealment (selection bias)	Low risk	Quote: "Blinding and randomisation was achieved by the manufacturer ..The code constructed by the manufacturer and not broken until the end of the study" Comment: probably done
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and personnel were blinded to group allocation. Placebo and intervention gel had identical appearance
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Code was not broken until after the study, assessors were blinded to group allocation
Incomplete outcome data (attrition bias) All outcomes	High risk	Only patients with pain and who phoned the surgery were "offered an appointment the same day to see a dentist" Comment: incomplete outcome data and judged as potentially at high risk of bias
Selective reporting (reporting bias)	Unclear risk	Not enough data to make a decision
Other bias	Unclear risk	Quote: "Where more than one extraction was needed only one was randomly chosen to be included in the study" Comment: unclear if this would constitute an element of selection bias

Ritzau 1977.

Study characteristics
Methods	Study design: randomised, parallel‐group, double‐blind, placebo‐controlled study Sample size: 45 Sample size calculation: not reported Conducted in: Royal Dental College in Aarhus, Denmark Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molars Recruitment period: February to May 1974 Funding source: not stated
Participants	Inclusion criteria: partially and totally impacted mandibular 3rd molars Exclusion criteria: none stated Mean age: 27 (range 17 to 61 years) Number randomised: 45 Number analysed: 45 Lost to follow‐up/dropouts: 0
Interventions	Comparison: propylic ester of p‐hydroxybenzoic acid versus placebo Group 1 (n = 24): propylic ester of p‐hydroxybenzoic acid inserted in each socket Group 2 (n = 21): placebo tablet gel inserted in each socket Co‐interventions: postoperative analgesic tablets containing 10 mg of codeine phosphate, 500 mg of acetyle acid, and 70 mg of magnesium oxide were prescribed, and the number of tablets consumed was recorded Experimental substance or placebo inserted into postoperative socket immediately after removal of impacted 3rd molar
Outcomes	Assessment day 7 postoperatively. Patients also assessed if they returned in pain at any point ‐ the day of return was noted Primary outcome measures: occurrence of alveolitis sicca dolorosa Secondary outcome measures: pain Adverse outcomes: haematoma and rash
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "PEPH or placebo were inserted at random selection...." Comment: method not specified
Allocation concealment (selection bias)	Low risk	Quote: '"...The code was unknown to the investigators" Comment: probably done
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Patients and personnel were blinded to group allocation. Intervention and placebo were identical in appearance
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Patients and personnel were blinded to group allocation. Intervention and placebo were identical in appearance, code for group allocation not broken until after the study completed
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses to follow‐up
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting of the outcomes
Other bias	Low risk	Appears to be free of other potential sources of bias

Rodriguez‐Perez 2013.

Study characteristics
Methods	Study design: RCT with 2 parallel groups Sample size calculation: not reported Sample size: 88 patients Setting: School of Dentistry at the University of Granada, and Oral and Maxillofacial Surgery Service of Virgen de las Nieves Hospital of Granada, Spain Number of centres: 2 Operators: 2 surgeons Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: January 2009 to January 2011 Funding source: not reported Declarations of interest: none reported
Participants	Inclusion criteria: patients needing surgical removal of 1 3rd molar with a difficulty rating of between 4 and 7 on the Koerner scale Exclusion criteria: taking antibiotics or analgesics in the 4 days before the procedure other disease contraindicating oral surgery AIDS or immunosuppression pregnancy or breastfeeding allergy to CHX, articaine, paracetamol or ibuprofen epinephrine contraindication the simultaneous extraction of 2 3rd molars any jawbone‐associated pathology uncooperative patients (psychic‐motor dysfunction and behaviour disorders) extractions lasting over 30 minutes Gender: 46 female, 42 male Mean age: 26 years Age range: 18 to 44 years Number randomised: 88 Number analysed:  88 Lost to follow‐up/dropouts: 0
Interventions	All patients: bioadhesive 0.2% CHX gel applied to the socket postextraction. All patients took 1 g paracetamol every 12 hours and 600 mg ibuprofen every 8 hours Intervention group 1: (n = 46) apply 0.2% CHX gel to the socket twice daily for 7 days Intervention group 2:  (n = 42) apply 1% CHX gel to the socket twice daily for 7 days
Outcomes	Primary outcome measures: presence of dry socket Diagnosis: Blum’s criteria for AO Secondary outcome measures: Oral Health Related Quality of Life Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were randomly classified into two groups, using a simple allocation using a computer program"
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	All eligible participants reported
Selective reporting (reporting bias)	Low risk	All outcomes specified were reported
Other bias	Low risk	The study appears to be free of other sources of bias

Rubio‐Palau 2015.

Study characteristics
Methods	Study design: randomised, double‐blind, clinical trial Sample size calculation: "The sample size was calculated with 80 patients treated with chlorhexidine gel and 80 with placebo, with a significance level of 5% and a statistical power of 80% to detect as significant a difference corresponding to an incidence of 11% in the chlorhexidine group and 30% in the placebo group" Sample size: 160 patients Setting: Minor Outpatient Surgery Unit of Hospital Vall d’Hebron, Spain Number of centres: 1 Operators: not reported Prevention or treatment of dry socket: prevention Type of teeth: mandibular 3rd molars Recruitment period: April 2008 to November 2010 Funding source: CHX and placebo gels provided by Laboratorios Lacer (C/ Sardenya 350, Barcelona, Spain) Declarations of interest: none reported
Participants	Inclusion criteria: patients needing surgical removal of 1 3rd molar with a difficulty rating of between 4 and 7 on the Koerner scale Exclusion criteria: unwillingness to participate patients with pericoronitis, active infection or antibiotic treatment within 2 weeks of the surgery patients with significant systemic disease, pregnancy, immunocompromise, AIDS or associated bone pathology Gender: 86 females, 74 males Mean age: 25.04 years Age range: not reported Number randomised: 160 Number analysed: 160 Lost to follow‐up/dropouts: 0
Interventions	All patients: diclofenac 50 mg every 8 hours alternated with metamizole 575 mg every 8 hours and omeprazole 20 mg per day Control group: (n = 80) a single dose of bioadhesive placebo was placed in the socket postextraction Intervention group: (n = 80) a single dose of 0.2% CHX bioadhesive gel was placed in the socket postextraction
Outcomes	Primary outcome measures: presence of dry socket Diagnosis: uncontrolled pain between 1 and 3 days postextraction with 1 or more of the following: partial or total disintegration of the clot, detritus, empty socket/exposed bone +/‐ halitosis Secondary outcome measures: none Adverse outcomes: 30 patients reported gastrointestinal discomfort, dizziness in 10 patients, skin rash in 2 patients (these patients were in both groups ‐ they report gastrointestinal upset was due to metamizole)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization of patients was performed by the Department of Statistics by a random list grouping the total of 160 patients in groups of 4 so that the distribution of the two groups (chlorhexidine and placebo) were homogeneous throughout the sample"
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "neither the surgeon nor the patient knew the substance"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Low risk	All eligible participants reported
Selective reporting (reporting bias)	Low risk	All specified outcomes were reported
Other bias	Low risk	The study appears to be free of other sources of bias

Shad 2018.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 180 patients Setting: Dentistry Department of Ayub Medical College, Abbottabad, Pakistan Number of centres: 1 Operators: experienced surgeon with 5 years experience Prevention or treatment of dry socket: prevention Type of teeth: impacted 3rd molars Recruitment period: January 2015 to July 2017 Funding source: not reported Declarations of interest: none reported
Participants	Inclusion criteria: patient requiring removal of an impacted lower 3rd molar, classed as of mild or moderate difficulty on Pedersen Difficulty Index Exclusion criteria: uncontrolled metabolic disease taking oral contraceptives pregnant patients patients who had received an antibiotic within 2 weeks prior to the extraction patients who received more than 2 cartridges of local anaesthetic Gender: male 83, female 97 Mean age: 27.47 years Age range: 21 to 35 years Number randomised: 180 patients Number analysed: 180 patients Lost to follow‐up/dropouts: 0
Interventions	Control group: 0.2% bioadhesive CHX gel placed in socket after extraction Intervention group: placebo gel placed in socket after extraction
Outcomes	Primary outcome measures: presence of dry socket Diagnosis: postoperative pain which increased in intensity 2 to 3 days after extraction and there was partial/total disintegration of clot resulting in an empty socket and denuded bone with or without halitosis Secondary outcome measures: none Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to make a decision
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients accounted for in analysis
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

Shi 2003.

Study characteristics
Methods	Study design: randomised, parallel‐group, controlled trial Sample size: 274 Sample size calculation: not reported Prevention or treatment of dry socket: prevention Type of teeth: impacted wisdom teeth Conducted in: not stated (appears to be 4th Medical University, Xi, Shaanxi China) Number of centres: not stated Recruitment period: November to April 2001 Funding source: not stated
Participants	Inclusion criteria: impacted wisdom tooth Exclusion criteria: history of acute pericoronitis in 10 days < 18 and > 65 years of age smoking more than 3 cigarettes oral contraceptives contraindication of tooth extraction Number randomised: 274 Number analysed: not fully reported
Interventions	Group 1 (n = 92): Shahaosan Group 2 (n = 86): Yunnan white drug Group 3 (n = 96): blank control Co‐interventions: none stated
Outcomes	Assessments 5 to 7 days postextraction. Quote: "incidence and intensity of dry socket in each group were observed and evaluated by scores" Primary outcome measures: incidence of dry socket Secondary outcome measures: self‐assessed POSSE (Postoperative Symptom Severity Scale) global assessment to include pain and "influence of daily life." Time and frequency of assessment unreported Adverse effects: no report of any assessment
Notes	3 different colour capsules, unclear if both A and B are active interventions. Unclear if this was systemic or topical and method and timing of administration not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "patients with extraction of impacted tooth were randomly divided into 3 groups" Comment: method of sequence generation not clear
Allocation concealment (selection bias)	Unclear risk	Nothing reported, unable to make a clear judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Selective reporting (reporting bias)	Unclear risk	Very limited data reported, unable to make a clear judgement
Other bias	Unclear risk	Very limited data reported, unable to make a clear judgement

Sun 2007.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 300 patients Setting: outpatient Number of centres: 1 Operators: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molar Recruitment period: March 2004 to December 2005 Funding source: none stated Declarations of interest: none stated
Participants	Inclusion criteria: age 18 to 50 years no outbreak of acute pericoronitis occurred in nearly 1 week no antibiotics were taken orally 3 days before surgery agree to accept the experiment and sign the informed consent no tooth extraction contraindication Exclusion criteria: not reported Gender: 140 males, 160 females Mean age: 30 years Age range: 18 to 49 years Number randomised: 300 Number analysed: 300 Lost to follow‐up/dropouts: 0
Interventions	All patients: the patient to return to the clinic 1 week after extraction, and may return to the clinic at any time if discomfort occurs. Inquiring the patients about postoperative bleeding and swelling in detail, the patients with suspected dry socket must be confirmed after examination by 2 attending doctors, and the patients who failed to return in time should be followed up by telephone Control group: blank control group (no description) Intervention group 1: oral tissue patch. Under conventional local anaesthesia, the impacted teeth were pulled out, the oral tissue patch was placed into the alveolar fossa, and the wound was sutured in counterposition. A yarn ball was placed on top to bite tightly, and the yarn ball was spat out after 0.5 hours. All operations are performed by the same doctor
Outcomes	Note: authors did not specify which was the primary outcome Outcome measures: composition ratio of impacted teeth type between the 2 groups bleeding rate after tooth extraction operation the occurrence rate of dry slot after tooth extraction incidence of swelling after tooth extraction Diagnosis criteria: 2 to 3 days after tooth extraction, obvious spontaneous pain began to occur and spread to the auriculotemporal region there was no normal blood clot in tooth extraction trauma, and the alveolar bone wall was exposed the wound is gray and the pain is obvious Diagnostic criteria for postoperative bleeding after tooth extraction: local bleeding on the day after tooth extraction surgery or significant local bleeding 24 hours after surgery. Diagnostic criteria of swelling after tooth extraction: local swelling on the day after tooth extraction or cheek swelling after 24 hours Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The patients were randomly divided into two groups by using the randomised table, 150 people in each group (with no further details being reported)
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported, no details about who the assessors were
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients were followed up
Selective reporting (reporting bias)	Low risk	The outcome (incidence of dry socket) was reported fully in Table 2
Other bias	Low risk	No other bias reported

Supe 2018.

Study characteristics
Methods	Study design: single‐blinded, prospective study Sample size calculation: not reported Sample size: 50 patients Setting: department of oral and maxillofacial surgery Number of centres: not stated Operators: not reported Prevention or treatment of dry socket: treatment Type of teeth: all teeth Recruitment period: September 2013 to June 2015 Funding source: none Declarations of interest: none
Participants	Inclusion criteria: patients presenting to the department of oral and maxillofacial surgery who required treatment for dry socket aged between 14 and 70 years Exclusion criteria: patients who were allergic to Alvogel and/or to other medications prescribed in study pregnant patients medically compromised patients, for example, diabetes in which healing is delayed patients having fractured root or foreign body in the dry socket Gender: 29 females, 21 males Mean age: 32.32 years Age range: 18 to 51 years Number randomised: 50 Number analysed: 50 Lost to follow‐up/dropouts: 0
Interventions	All patients: all patients received diclomol 50 mg Intervention group 1: patients received Alvogel (combination of iodoform and butylparaminobenzoate) paste as an intrasocket medication Intervention group 2: patients received zinc oxide eugenol as an intrasocket medication
Outcomes	Primary outcome measures: pain (VAS scale) Diagnosis: not clearly stated Secondary outcome measures: time required for complete pain relief, healing (measured by amount of socket covered by initial granulation tissue) Adverse outcomes: 2 patients in group 1 (Alvogyl), and 9 patients in group 2 (zinc oxide eugenol) showed delayed healing
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to make a decision
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to make a decision
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient information to make a decision. All patients are accounted for in terms of healing scores at day 14 (final follow‐up). It should be noted that pain scores have only been reported for 4/5 days
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a decision
Other bias	Low risk	The study appears to be free of other sources of bias

Torres‐Lagares 2006a.

Study characteristics
Methods	Study design: randomised, controlled "...prospective, parallel, single‐blind clinical trial" Sample size: 30 Sample size calculation: not reported Conducted in: Faculty of Odontology of the University of Seville, Spain Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: 1 or 2 impacted mandibular 3rd molars, under local anaesthesia Recruitment period: September to December 2001 Funding source: not stated
Participants	Inclusion criteria: 1 or 2 mandibular impacted 3rd molars; difficulty index 4‐7 Koerner scale (Koerner 1994) no symptoms 10 days pre‐surgery Exclusion criteria: contraindications for the intervention AIDS smokers immunodepressed patients pregnant/lactating women women taking oral contraceptives allergies to: CHX, lidocaine or paracetamol extraction of both wisdom teeth at 1 visit bone pathology having ingested any medication 4 days preoperatively Number randomised: 30 Number analysed: 30 Lost to follow‐up/dropouts: 0
Interventions	Group 1 (n = 17): 10 mL of 0.2% CHX bioadhesive gel applied intra‐alveolar postextraction Group 2 (n = 13): no intrasocket medication Co‐interventions: ''All the patients took, as postoperative treatment, 14.05 mg codeine phosphate and 500 mg of paracetamol on demand...''
Outcomes	Unclear if assessed on days 3 and 7 postoperatively Primary outcome measures: development of AO Secondary outcome measures: none stated Adverse effects: none stated
Notes	November 2012: communication from Dr Torres‐Lagares confirmed that Hita‐Iglesias 2008; Torres‐Lagares 2006a and Torres‐Lagares 2006b were separate studies each involving a different group of participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The aforementioned allocation into one group or another was carried out by computer before the start of the study" Comment: probably done
Allocation concealment (selection bias)	Low risk	Quote: ''...an envelope was opened, in which it indicated whether the patient should receive the bio‐adhesive gel or not"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Patients did not know allocation but operators would have. It is unlikely that lack of masking of operator would lead to deviation in intervention. Strict study criteria required that participants and personnel would be blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	High risk	Operators detecting presence of dry socket could have remembered group allocation potential for high risk
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients completed the protocol and analysed
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting of the outcomes
Other bias	Low risk	Appears to be free of other potential sources of bias

Torres‐Lagares 2006b.

Study characteristics
Methods	Study design: randomised, controlled, parallel‐group, double‐blind study Sample size: 103 Sample size calculation: not reported Conducted in: Faculty of Dentistry of the University of Seville, Spain Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular 3rd molars ‐ under local anaesthesia Recruitment period: January to June 2003 Funding source: Laboratorios Lacer, Barcelona, Spain, donated the medication used in this study
Participants	Inclusion criteria: 1 or 2 mandibular impacted 3rd molars; difficulty index 4‐7 Koerner scale (Koerner 1994) no symptoms 10 days pre‐surgery aged 18 to 60 years of age Exclusion criteria: contraindications for the intervention AIDS smokers immunodepressed patients pregnant/lactating women women taking oral contraceptives allergies to: CHX, lidocaine or paracetamol extraction of 2 molars at once bone pathology having ingested any medication 4 days preoperatively Number randomised: 103 Number analysed: 94
Interventions	Comparisons: 0.2% CHX bioadhesive gel versus no application Group 1 (n = 49): 10 mL of 0.2% CHX bioadhesive gel applied intra‐alveolar postextraction Group 2 (n = 45): placebo gel ‐ excipient containing Co‐interventions: all the patients took, as postoperative treatment, 14.05 mg codeine phosphate and 500 mg of paracetamol on demand
Outcomes	Primary outcome measures: development of AO Secondary outcome measures: none stated Adverse effects: none stated
Notes	No clear statement of when follow‐up was done November 2012: communication from Dr Torres‐Lagares confirmed that Hita‐Iglesias 2008; Torres‐Lagares 2006a and Torres‐Lagares 2006b were separate studies each involving a different group of participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: ''The random assignment was carried out by means of a random number list"
Allocation concealment (selection bias)	Low risk	Quote: ''The allocation of patients into one group or the other was carried out by computer before the start of the study. The gel was identifiable by a patient inclusion code number"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Neither the patient nor the operator knew group allocation, the code list was kept in a sealed enveloped and was not opened until after the study. Placebo did not contain active ingredient"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Operator did not know group allocation. The code list was kept in a sealed enveloped and was not opened until after the study"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Number of patients not included in final check‐up visit stated, not included in analysis: 5 patients (1 in the control group and 4 in the test group) did not have their final check‐up visit
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting of the outcomes
Other bias	Low risk	Appears to be free of other potential sources of bias

Trieger 1991.

Study characteristics
Methods	Study design: randomised, split‐mouth study Prevention or treatment of dry socket: prevention Type of teeth: mandibular bony 3rd molar impactions Conducted in: setting unreported Number of centres: not stated Recruitment period: not stated Funding source: "This study was made possible with a grant from Upjohn, Kalamazoo, Michigan"
Participants	Inclusion criteria: bilateral mandibular 3rd molar bony impactions Exclusion criteria: no antibiotics 2 weeks prior Number randomised: 172 sites in 86 patients Number analysed: unclear
Interventions	Group 1 (n = 86): gelfoam square saturated with 1 mL clindamycin phosphate solution (150 mg/mL) Group 2 (n = 86): gelfoam square saturated with saline placebo Co‐interventions: concomitant analgesic medication allowed but no details reported
Outcomes	Not clear when assessments were made and by whom Primary outcome measures: dry socket Secondary outcome measures: not reported Adverse events: unreported and unclear if assessed
Notes	We assume there were no dropouts but study poorly reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "...according to a randomized distribution"
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a clear judgement if adequate
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Very limited data reported, unable to make a clear judgement
Selective reporting (reporting bias)	Unclear risk	Very limited data reported, unable to make a clear judgement
Other bias	Unclear risk	Very limited data reported, unable to make a clear judgement

Tuk 2019.

Study characteristics
Methods	Study design: cross‐over RCT Sample size calculation: yes, using a paired‐samples t test, an a of 5%, a power of 90%, 2‐tailed testing, and an effect size of 0.5 (a moderate effect size was used because the effect should be clinically relevant) results in a total sample size of 44. Using a paired‐samples t test, an a of 5%, a power of 90%, 2‐tailed testing, and an effect size of 0.5 (a moderate effect size was used because the effect should be clinically relevant) results in a total sample size of 44 Sample size: 54 patients Setting: Department of Oral and Maxillofacial Surgery, the Netherlands Number of centres: 1 Operators: 1 Prevention or treatment of dry socket: prevention Type of teeth: bilateral symmetrically, horizontally impacted mandibular 3rd molars Recruitment period: 1 January 2016 to 31 July 2017 Funding source: not reported Declarations of interest: reported as none
Participants	Inclusion criteria: native Dutch speakers with bilateral horizontally impacted mandibular 3rd molars, with a Gregory‐Pell grade 3B impaction, who were planning to undergo bilateral mandibular 3rd molar surgery, age of 18 years or older; American Society of Anesthesiologists status 1; patients with no systemic diseases or medical conditions, no discernible active pathology associated with the 3rd molars, and no acute pericoronitis; and patients without periodontal disease Exclusion criteria: smoking, allergy to ibuprofen, presence of systemic disease, history of a recent and/or symptomatic peptic ulcer, antiplatelet or anticoagulant therapy, pregnancy or lactation, recent (< 15 days) local infection before surgery, previous radiation therapy to the maxillofacial region, local pathology (e.g. cyst or tumour) associated with the 3rd molars, or lack of consent to undergo the procedure or participate in the study Gender:  25 male, 29 female Mean age: 25.1 years Age range: not reported Number randomised: 54 Number analysed:  54 Lost to follow‐up/dropouts: 0
Interventions	All patients: surgical site closure, ibuprofen 600 mg 3 times a day, 0.12% CHX mouthwash Intervention group 1: 1 x 2 cm iodine tampon Intervention group 2: saline rinsing using a monoject syringe 4 times daily 48 hours postoperatively
Outcomes	Primary outcome measures: oral health related quality of life, pain, swelling, limited mouth opening, postoperative infection and AO Diagnosis: OHIP‐14, VAS, questionnaire, the presence of purulent discharge in the extraction socket and/or excessive swelling with fluctuation, with or without pain; presence of a local abscess; or onset of facial or cervical cellulitis plus other signs suggesting infection, such as pain, increased heat, erythema, and/or fever. A diagnosis of AO was established if there was new onset or increasing pain more than 36 hours after removal of the mandibular 3rd molar in combination with an absence of the hematic clot in the socket, resulting in exposed bone. A putrid smell may be present in combination with intense neuralgia‐type pain. Gentle probing or irrigation of the wound aggravated the pain. All elements needed to be present for the diagnosis of AO Secondary outcome measures: self‐care: pain medication, ice cooling Adverse outcomes: not reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "a surgical assistant allocated the patient to either trial condition using a computer random generator... After the allocation to a trial condition, another surgical assistant assigned a side to either the experimental treatment or the control treatment, again using a computer random generator"
Allocation concealment (selection bias)	Low risk	Quote: "The concealment of allocation was guaranteed through a sealed envelope" Comment: suitable washout period of 8 weeks
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants likely to be aware of procedure. Operator would be aware of procedure. Unlikely that lack of masking would have caused deviation in intervention. Strict study criteria required that participants and personnel would be blinded to group allocation, but it is likely that they were aware as experimental was very different from control
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "One independent assessor evaluated the presence of wound infection or AO. Infection was defined by any of the following"
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no dropouts reported
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias. There was a suitable washout period to eliminate carry out

Unsal 2018.

Study characteristics
Methods	Study design: split‐mouth randomised study Sample size calculation: not given Sample size: 50 patients Setting: Department of Oral and Maxillofacial Surgery, Ankara University, Turkey Number of centres: 1 Operators: 1 surgeon Prevention or treatment of dry socket: prevention Type of teeth: partially erupted lower 3rd molars Recruitment period: February 2013 to February 2014 Funding source: none Declarations of interest: none
Participants	Inclusion criteria: presence of bilateral, symmetrically orientated, partially erupted lower 3rd molars requiring removal for prophylactic reasons absence of pathology associated with the 3rd molars no pre‐existing systemic diseases no chronic opioid use Exclusion criteria: patients with no 2nd molars pregnant or lactating women Gender: 17 males, 33 females Mean age: 23.96 years Age range: 15 to 43 years Number randomised: 50 Number analysed:  50 Lost to follow‐up/dropouts: none reported
Interventions	All patients: rinsed with 0.2% CHX 3 times daily postoperatively Control group: no intervention Intervention group: platelet rich fibrin inserted into socket
Outcomes	Primary outcome measures: presence of AO Diagnosis: patients experiencing “severe” level of pain for at least 4 days postoperatively, with the occurrence of situations outside of the normal healthy healing tissue Secondary outcome measures: periodontal probing depths on the distal surface of the mandibular 2nd molar at 3 months Adverse outcomes: none reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to make a decision
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to make a decision
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All postoperative evaluations were performed by the surgeon blinded to treatment assignment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient information to make a decision
Selective reporting (reporting bias)	Low risk	All outcome measures appear to have been reported
Other bias	Low risk	The study appears to be free of other sources of bias

van Eeden 2006.

Study characteristics
Methods	Study design: randomised controlled, split‐mouth trial Sample size: 19 Sample size calculation: not reported Setting: Military Hospital in South Africa Number of centres: 1 Prevention or treatment of dry socket: prevention Type of teeth: bilateral impacted mandibular 3rd molars of similar difficulty assessed clinically and radiologically. Carried out under general anaesthesia Recruitment period: not stated Funding source: not stated
Participants	Inclusion criteria: bilateral impactions of similar levels as determined by clinical and radiographic examination no pericoronal infection no antibiotics no anti‐inflammatory medication prior to surgery no associated comorbidity Exclusion criteria: none stated Mean age: 21.4 years (16 to 32 range) Number randomised: 19 Number analysed: 19 Lost to follow‐up/dropouts: 0
Interventions	Comparison: covomycin D versus inert gel foam carrier Group 1 (n = 19): 1 mm covomycin D (2.0 mg chloramphenicol, 5.0 mg neomycin sulphate and 0.5 mg dexamethasone) delivered within an inert gel foam carrier Group 2 (n = 19): inert gel foam carrier ‐ 1 mL normal saline Co‐interventions: analgesic/anti‐inflammatory medication 6 hourly when necessary (Myprodol®) oral antibiotic medication (amoxicillin 500 mg 8 hourly or in penicillin allergic patients erythromycin 500 mg 6 hourly) for 5 days a 0.2% CHX gluconate mouthrinse 6 hourly for 5 days
Outcomes	Pain scores were recorded at 6‐hour intervals from the day of surgery until day 6. The patients were examined clinically on day 6 Primary outcome measures: development of AO, diagnostic criteria Secondary outcome measures: none stated Adverse effects: some of the events may be attributable to intervention or could be normal sequelae of operation
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "...preselected on a random basis by the flip of a coin"
Allocation concealment (selection bias)	Unclear risk	No mention of this. Therefore, insufficient information to make a clear judgement if adequate
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants were blinded to group allocation. Personnel were blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "..patients were examined clinically by an independent surgeon blinded to the site of intrasocket medication"
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients were accounted for at day 6
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting of the outcomes
Other bias	Low risk	Appears to be free of other potential sources of bias

Xue 2013.

Study characteristics
Methods	Study design: RCT Sample size: not stated Sample size calculation: 160 Setting: outpatient clinic Number of centres: 1 Operators: 1, associate chief physician with 10 years experience Prevention or treatment of dry socket: prevention Type of teeth: impacted mandibular molar Recruitment period: June 2011 to June 2012 Funding source: no direct or indirect financial or benefit sponsorship from any manufacturer, relevant employer or other economic organisation: none stated
Participants	Inclusion criteria: aged 20 to 40 years oral pantomogram shows impacted wisdom teeth in mandibular level there was no history of acute pericoronitis in the past 2 weeks, and no purulent secretions around the crown all patients gave informed consent to the trial and treatment plan Exclusion criteria: patients have systemic diseases such as severe heart disease, high blood pressure, diabetes, and malnutrition Gender: 72 males, 88 females Mean age: 28 years Age range: 20 to 40 years Number randomised: 160 Number analysed: 160 Lost to follow‐up/dropouts: 0
Interventions	All patients: preoperative preparation: detailed inquiries were made about the patient's chief complaint, history of present disease, and past history, and tooth extraction contraindication was excluded. Good communication with patients, so that patients have an understanding of the necessity of the operation and the possible complications during and after the operation, so that patients can relax, avoid tension and fear, and can cooperate well during the operation. The situation of the impacted mandibular teeth was examined and the curved oral pantogram was taken to understand the resistance of the impacted teeth and the relationship between the impacted teeth and the adjacent teeth and the mandibular inferior dental nerve. Before operation, 75% ethanol was used to disinfect the perioral and facial skin, and then the face was covered with a sterile towel. The tooth extraction operative area was disinfected with 1% iodine tincture. Routine preparation scalpel, gingival separator, turbine, split drill, tooth brace, tooth forceps, needle holder, suture, aspirator. Impacted teeth were extracted: the experimental group was treated with 2% lidocaine and 1% adrenaline hydrochloride at 1:200000 to block the inferior alveolar nerve, lingual nerve and buccal nerve, and then infiltrated with anaesthesia at the medial and distal buccal angles of the impacted teeth to enhance the anaesthetic effect and reduce the intraoperative and postoperative bleeding. Buccal and distal incisions were made with a number 11 sharp knife blade, and the buccal and distal bone surfaces were exposed after flaps were turned over. According to the bone coverage, buccal, distal bone and accretion were removed with a fast turbine, and the impacted crowns were ground vertically, then the crowns and roots were removed. The tooth extraction sockets were carefully examined and the residual bone or tooth fragments were removed. Reposition was performed in patients with displacement of bone plates, and immediate trimming was performed in those with sharp bone processes. The doctors' advice after tooth extraction was the same in both groups. Patients were required not to gargle or brush their teeth within 24 hours after tooth extraction surgery, not to accept any antibiotics 1 week after the surgery, not to smoke, and to return to the doctor in time with abnormal conditions, such as bleeding Control group: the control group did not intervene after pulling out the cotton ball for 30 minutes Intervention group 1: recombinant bovine basic fibroblast growth factor gel intervention: after the extraction of the affected teeth, 0.1 mL recombinant bovine basic fibroblast growth factor gel was dropped into the bottom of the tooth extraction socket with a 1 mL injector immediately after tooth extraction was extracted, and the cotton ball was bit for 30 minutes after the blood filling of the tooth extraction fossus
Outcomes	Note: authors did not specify which was the primary outcome Outcome measures: the general condition of the tooth extraction wound of the patient (diagnosis was made according to the diagnostic criteria of dry slot) the occurrence rate of dry socket after tooth extraction Diagnostic criteria for dry socket syndrome: diagnosis mainly depends on clinical manifestations: tooth extraction acute pain in 2 or 3 days, and spread to the auriculotemporal or lower anterior teeth, general analgesics are ineffective the tooth extraction wound was empty and covered with a layer of rotten matter, which had a special stench. The internal alveolar bone of tooth extraction wound was exposed, which was extremely sensitive, and severe pain could be caused by light rubbing or cold water stimulation Adverse outcomes: no local or systemic reactions, such as hypersensitivity or tissue hyperplasia, occurred in the experimental group after the application of recombinant bovine basic fibroblast growth factor gel
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	160 random numbers were read from any column or any row in the random number table, and then divided each number by 2. Patients were assigned to the two groups depended on the remainder being odd or even. Insufficient clarity to understand the randomisation process
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported, no details about who the assessors were
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients were followed up
Selective reporting (reporting bias)	Low risk	All outcomes specified were reported, the incidence of dry socket was reported in Table 2
Other bias	Low risk	No other bias observed

Yuce 2019.

Study characteristics
Methods	Study design: RCT Sample size calculation: not reported Sample size: 40 Setting: university oral and maxillofacial surgery department Number of centres: 1 Operators: 1 surgeon for initial treatment; 3 surgeons for assessments Prevention or treatment of dry socket: treatment Type of teeth: 3rd molar Recruitment period: not reported Funding source: no funding received Declarations of interest: no conflicts of interest
Participants	Inclusion criteria: between 18 and 40 years of age; no systemic disease; positively diagnosed and untreated AO in 3 days after extraction of mandibular 3rd molars of class A and 1 according to the Pell–Gregory classification Exclusion criteria: the use of medications that may interfere with the healing process; smoking; pregnancy or lactation; were menstruating; presence of any conditions such as inflammation, periodontitis, gingivitis and dental abscess in the area of the teeth; undergoing antibiotic or anti‐inflammatory drug therapies in the 7 days before extraction; using oral contraceptives and radiation therapy or chemotherapy in the 12 months before extraction and would require raising a flap and/or removing bone to remove the tooth Gender: 55% female, 45% male Mean age: 31.2 years Age range: not reported Number randomised: 40 Number analysed: 40 Lost to follow‐up/dropouts: 0
Interventions	All patients: extraction sockets were curetted to remove debris and gently cleaned by irrigation with saline only, diclofenac 50 mg every 12 hours as needed Intervention group 1: saline irrigation postoperatively on day 1, 3, 5 and 7 Intervention group 2: extraction sockets were packed with A‐PRF + (advanced platelet rich fibrin) and stitched; suture removed at day 7
Outcomes	Primary outcome measures: pain, soft tissue healing, bone density Diagnosis: VAS; index of Landry, Turnbull and Howley, OPT i‐Dixel 2.1.8.2 software by the average gray level values Secondary outcome measures: none reported Adverse outcomes: reported as none
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to make a decision
Allocation concealment (selection bias)	Unclear risk	Insufficient information to make a decision
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Given the nature of the interventions it would not have been possible to blind the operator and patients. Lack of masking unlikely to have led to deviation in the intervention. Strict study criteria required that participants and personnel would be blinded to group allocation
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information
Incomplete outcome data (attrition bias) All outcomes	Low risk	No sample sizes numbers presented in results tables or in text; flow chart would suggest no loss to follow‐up
Selective reporting (reporting bias)	Low risk	The report appears to be free of selective reporting bias
Other bias	Low risk	No other observed bias

AO = alveolar osteitis; CHX = chlorhexidine; NaCl = sodium chloride; RCT = randomised controlled trial; SD = standard deviation; VAS = visual analogue scale.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Afat 2018	Antibiotics prescribed postoperatively. All patients received amoxicillin 1 g twice daily for 1 week
Ahmed 2020	Details of intervention unclear
Akota 1998	The intervention used was an impregnated gauze drain which was specifically excluded from this review
Al‐Hamed 2017	Antibiotics prescribed postoperatively. All patients received amoxicillin 500 mg 4 times daily for 5 days
Al‐Sukhun 2011	3 analgesics taken systemically to assess their ability to control pain after extraction. They were not local measures
Altman 2011	Not dry socket
Anand 2015	Dry socket not clearly defined; page 26: “for the purpose of this investigation, the diagnosis of dry socket was based on clinical symptoms rather than the appearance of dry socket”
Annibali 2012	Not an RCT. Consensus statement on management of third molars
Anonymous 1966	Not an RCT. Translated from the Japanese by Professor Ken Yaegaki, Nippon Dental University Department of Oral Health, Tokyo, Japan
Arakeri 2011	Dry socket not defined
Arenaz‐Bua 2010	Dry socket not defined, general comments on sequelae
Asutay 2017	Unclear criteria for diagnosis of dry socket; page 1533: “if patients experienced persistent and progressive pain, it considered to dry socket”
Banach 1973	Not an RCT (translated from Polish)
Baqain 2012	The intervention was surgical (flap design) which was specifically excluded from the review
Baslarli 2015	Antibiotics prescribed postoperatively; all patients received amoxicillin 1000 mg twice for 5 days
Bello 2011	The intervention was wound closure techniques excluded from this review
Berwick 1990	Dropouts substituted making trial results invalid
Betts 1995	Evaluated the efficacy of lidocaine jelly for the alleviation of pain experienced during the instrumentation of extraction sites diagnosed with alveolar osteitis. Assessment of pain related to instrumentation only
Bezerra 2011	Systemic amoxicillin antibiotic, not local intervention
Birke 1970	Not an RCT (translated from German)
Bloomer 2000	Quote: "In a consecutive manner, 1 of the 2 lower third molar sockets was packed on each patient. Sockets were packed in a series of 25 on 1 side then changed to a series of 25 on the opposite side throughout the series, totaling 100 patients" Comment: CCT quasi‐randomised
Bloomer 2012	Not an RCT/inadequate method of sequence generation
Brignardello 2012	Study about techniques, not dry socket
Butler 1977	The interventions were lavage techniques which are excluded from this review
Butylin 1977	Cohort study (translated from Russian)
Bystedt 1980	Interventions were systemic antibiotics
Cebi 2020	Randomisation unclear; page 681: "patients were assigned to groups A,B,C." Diagnostic criteria for dry socket also unclear
Christensen 2012	Not an RCT. Expert opinion/commentary
Cooper 2012	Pain study, not dry socket
Daniels 2011	Dry socket not defined. Pain management study
Daugela 2018	Antibiotics prescribed; all patients received clindamycin 600 mg 1 hour before and 6 hours after surgery
Dubovina 2016	Criteria for diagnosis and follow‐up of dry socket unclear
Dutta 2016	Antibiotics prescribed; page 47: "Each patient received identical postoperative antibiotics"
Eshghpour 2014	Antibiotics prescribed postoperatively; all patients received amoxicillin 500 mg 3 times daily for 7 days
Eshghpour 2018	Antibiotics prescribed postoperatively; all patients received amoxicillin 500 mg 3 times daily for 7 days
Farooq 2019	It does not appear from the text that a clinician diagnosed the dry socket and while they mention Blum’s criteria they have conflated this with abscess in an “alveolitics” score they have developed where main items are not shown
Field 1988	Quote: "The trial was 'open' and on arrival extraction cases were consecutively allocated to one of three groups" Comment: CCT quasi‐randomised
Fotos 1992	Quote: "70 randomly selected healthy patients" page 383, "each subject was treated at one extraction site with a saline solution whereas the other site received CHX" page 384 Comment: no evidence of randomisation of participants to intervention
Garibaldi 1995	Quote: "Patients were assigned in sequential (A, B, C, A, B, C, etc) order" Comment: CCT quasi‐randomised
Goldman 1973	CCT, non‐randomised study
Goldsmith 2012	Intervention involved surgical techniques which are excluded from the review
Gonzalez‐Serrano 2021	Antibiotics administered as part of intervention; amoxicillin 750 mg 3 times a day for 7 days
Goyal 2012	Not dry socket, instrument evaluation
Guazzo 2018	No criteria given for “alveolitis”
Hall 1971	Not an RCT. Expert opinion/commentary
Haraji 2010	Interventions were envelope versus modified triangular flap designs, which are excluded from this review
Haraji 2012a	Duplicate reporting
Haraji 2012b	Criteria for diagnosis of dry socket is unclear – they use a scale that includes the presence of pus
Haraji 2014	Duplicate reporting: "The authors used some information collected during a double‐blinded randomised control trial reported previously." Some of these data previously reported in Haraji 2013
Haraji 2015	Duplicate reporting
Haupt 2015	No definition or criteria given for diagnosis of dry socket
Hill 2006	Pain study, not dry socket
Hooley 1995	No clinical diagnosis of dry socket made, only used pain
Jadhao 2018	Antibiotics were prescribed; all patients received co‐amoxiclav 625 mg twice a day for 7 days
Jesudasan 2015	Antibiotics prescribed postoperatively; all patients received metronidazole 400 mg 3 times daily for 3 days
Johnson 1988	CCT, non‐randomised study
Jolley 1972	Quote: "The purpose of this was to determine the effectiveness of the gel in controlling pain from ill‐fitting dentures after extractions and in other situations" page 72 Comment: not dry socket
Jovanovic 2011	Not an RCT (after contact with author)
Julius 1982	Not an RCT. Split‐mouth where all left‐hand sides received intervention and all right‐hand sides received the control
Kamal 2020a	Not an RCT. Quote page 2: "selected patients were divided into two groups based on their choice to be treated with either conventional treatment (group I) or CGF treatment (group II)"
Kamal 2020b	Not an RCT. Quote page 613: "Sixty patients with one dry socket each, at University Dental Hospital Sharjah, were divided into three treatment groups based on their choice"
Kaplan 2020	Presence of dry socket not one of the main outcome measures. No criteria given for diagnosis of dry socket
Karthik 2021	Antibiotics administered as part of intervention
Keskitalo 1973	Quote: "Alternate patients were treated with Apernyl cones" Comment: CCT quasi‐randomised
Kilinc 2017	Antibiotics prescribed; all patients given amoxicillin + clavulanic acid 2 g/day for 5 days
Kim 2020	Presence of dry socket is not an outcome measure
Kirk 2007	Definition of dry socket not provided
Krekmanov 1981	Only systemic interventions
Krekmanov 1986	Unsure if randomised
Krishnan 2020	Presence of dry socket is not an outcome measure
Kudiyirickal 2012	Not an RCT, not dry socket, looks at oral facial infections
Lao 2012	Not dry socket
Liu 2011	Not dry socket
Long 2012	The intervention was a surgical technique which is excluded from this review
Lopez‐Cedrun 2011	Systemic antibiotics, not a local intervention
MacGregor 1973	Outcomes specified were pain and swelling, no clinical diagnosis of dry socket
MacGregor 1975	Quote: "Successive patients were entered in to the trials... arrangements were made that there was an equal distribution of experiments and controls" Comment: no evidence of randomisation
Majid 2010	No definition of dry socket in report. Study refers to generalised sequelae
Malkawi 2011	Cross‐sectional study, not an RCT
Mehlisch 2010a	No definition of dry socket. Study refers to pain management
Mehlisch 2010b	No definition of dry socket. Study refers to pain management
Mishra 2012	Not dry socket, dental pain
Mitchell 1986a	Quote: "They were allocated into groups according to a predetermined protocol" Comment: inadequate method of sequence generation, quasi‐randomised CCT
Mitchell 1986b	Systemic intervention, not a local intervention
Moberly 2007	Wrong outcome; this study is comparing efficacy of different analgesics
Nentwig 1985	Study unobtainable
Neugebauer 2004	Participants "randomised into two groups." Randomisation at participant level but allocation of intervention 'split‐mouth' at extraction socket level Comment: open allocation. CCT (German translation)
Neuner 1969	Study to treat pain and there is no mention that it is an RCT (German translation)
Njokanma 2019	Only abstract available with insufficient information
Nordenram 1973	Outcomes reported were "postoperative complications": pain/swelling/infection but with no independent assessments of classical 'dry socket'
Olson 1987	Abstract only, insufficient information for inclusion
Olusanya 2011	Systemic antibiotics, not a local intervention
Osunde 2014	Antibiotics prescribed postoperatively; all patients received amoxicillin 500 mg 8 hourly and metronidazole 200 mg 8 hourly for 5 days
Osunde 2015	Antibiotics prescribed; quote: “all patients received the same antibiotics”
Osunde 2017	Antibiotics prescribed; all patients received amoxicillin 500 mg 8 hourly and metronidazole 200 mg 8 hourly for 5 days
Oyri 2019	The intervention used was an impregnated gauze drain which was specifically excluded from this review
Ozveri 2020	Antibiotics prescribed; all patients received amoxicillin + clavulanic acid 625 mg every 12 hours for 7 days
Paul 2019	Retracted by the editorial board of the journal due to “significant double publication”
Pichler 2001	Abstract only, insufficient information for inclusion
Prataap 2017	Not an RCT
Qi 2012	Not dry socket, pain management study
Rani 2016	Postoperative antibiotics prescribed
Rastogi 2018	Not an RCT
Reeshma 2021	Not an RCT
Ritzau 1978	Unclear if RCT
Saez‐Alcaide 2020	Antibiotics prescribed; all patients were prescribed amoxicillin 750 mg every 8 hours for 7 days
Sanchis 2004	Quote: "We divided the cases into a group of 100 patients who underwent extraction...." Comment: non‐randomised controlled (no treatment) study
Sarkar 2019	Criteria for diagnosis of dry socket not provided
Schatz 1987	Age range of a group of participants outside the inclusion criteria and no subgroup data reported. Open allocation sequence
Schlund 1968	Study unobtainable
Scopp 1967	Wrong outcome; this study evaluates the clinical efficacy of different analgesics
Seethamsetty 2019	Antibiotics prescribed; all patients were given amoxicillin
Sharma 2017	Not an RCT
Sorensen 1987	Quote: "Patients were randomly selected and divided so that approximately half would receive...." Comment: non‐RCT
Swanson 1989	Quote: "One hundred impacted lower third molars were to be operated. They were to be included in the study in the random order in which they came to the surgery by routine booking methods" Comment: CCT quasi‐randomised
Sweet 1985	Quote: "The rinses were chosen by a random‐selection technique" Comment: CCT quasi‐randomised
Syrjanen 1981a	Controlled study, non‐RCT
Syrjanen 1981b	Does not measure dry socket as an outcome
Tek 2014	Consecutive rather than randomisation of site
Tjernberg 1979	Quote: "Patients referred to the department for the surgical removal of a partially erupted, lower third molar were randomly distributed into two pools" Comment: CCT quasi‐randomised
Tong 2012	Not dry socket
Torres Lagares 2006	Not an RCT/inadequate method of sequence generation
Torres‐Lagares 2010	Patients in the study have bleeding disorders
Vedtofte 1974	Quote: "The cones were packed in randomly numbered packets and the code was unknown to the investigators" Comment: inadequate method of randomisation
Vu 2021	Dry socket not a primary outcome measure
Wang 2013	Antibiotics prescribed; all patients in the study received amoxicillin 500 mg 3 times daily and tinidazole 1 g once daily for 5 days
Wen 2004	Quote: "The method of random‐digit dialling was adopted to divide groups" Translated from the Chinese by Dr Nian Fang. "Patients were divided into 2 groups Group A and Group B, with complete randomization according to the visit order....." Comment: suggestive of quasi‐randomisation Inconsistencies between the English abstract and the translated version of the Chinese paper and lack of clarity in the methodology as reported did not provide any degree of confidence that adequate measures had been taken to satisfactorily randomised the participants or to conceal the allocation sequence in this study
Yuan 2006	Antibiotics prescribed; all patients received cephalosporin 4 g intravenous and metronidazole 250 ml infusion for 3 days
Yue 2012	No dry socket, pain management study
Zanetta‐Barbosa 1994	Following email communication with the principal investigator, "the first patient was decided by a coin toss and the following was always allocated by alternation in the other group" Comment: CCT quasi‐randomised
Zuniga 2011	Dry socket not defined. Pain management study

CCT= controlled clinical trial; RCT = randomised controlled trial.

Characteristics of studies awaiting classification [ordered by study ID]

Zorina 2019.

Methods	No details
Participants	No details
Interventions	No details
Outcomes	No details
Notes

Differences between protocol and review

• The original protocol was only for the treatment of dry socket. We have expanded the scope of the review to incorporate the prevention of dry socket.

• We added the primary outcome for treatment of dry socket: time to heal.

• We reworded 'type of participant' to clarify the inclusion of any extracted teeth.

• We changed analyses from fixed‐effect to random‐effects as requested by peer reviewer.

• We have added undertaking a subgroup analysis for chlorhexidine dose in the methods. 

Contributions of authors

• Conceiving the idea: Blánaid Daly (BD).

• Writing the protocol: BD, Mohammad O Sharif (MOS), Tim Newton (TN), and Kate Jones (KJ). 

• Organising retrieval of papers: BD, MOS, Anna Beattie (AB).

• Writing to authors of papers for additional information: BD, MOS, AB.

• Providing additional data about papers: BD, AB.

• Data collection for the review: BD, MOS, AB.

• Screening search results: BD, AB, MOS, Helen Worthington (HW).

• Screening retrieved papers against inclusion criteria: BD, AB, MOS, HW.

• Appraising quality of papers: BD, AB, MOS, KJ.

• Extracting data from papers: BD, MOS, AB, HW, KJ.

• Obtaining and screening data on unpublished studies: AB, MOS, HW.

• Entering data into Review Manager: BD, AB, HW, KJ, MOS.

• Analysis of data: BD, AB, HW.

• Writing the review: BD, AB, KJ, HW, MOS.

Sources of support

Internal sources

• The University of Manchester, UK

  Cochrane Oral Health is supported by The University of Manchester and the NIHR Manchester Biomedical Research Centre.

• Manchester Academic Health Sciences Centre (MAHSC), UK

  Cochrane Oral Health is supported by the MAHSC and the NIHR Manchester Biomedical Research Centre.

External sources

• Cochrane Oral Health Global Alliance, UK

  The production of Cochrane Oral Health reviews has been supported financially by our Global Alliance since 2011 (oralhealth.cochrane.org/partnerships-alliances). Contributors in recent years have been the American Association of Public Health Dentistry, USA; AS‐Akademie, Germany; the British Association for the Study of Community Dentistry, UK; the British Society of Paediatric Dentistry, UK; the Canadian Dental Hygienists Association, Canada; the Centre for Dental Education and Research at All India Institute of Medical Sciences, India; the National Center for Dental Hygiene Research & Practice, USA; New York University College of Dentistry, USA; and Swiss Society of Endodontology, Switzerland.

• National Institute for Health and Care Research (NIHR), UK

  This project was supported by the NIHR, via Cochrane Infrastructure funding to Cochrane Oral Health. The views and opinions expressed herein are those of the review authors and do not necessarily reflect those of the Evidence Synthesis Programme, the NIHR, the NHS, or the Department of Health and Social Care.

Declarations of interest

There are no financial conflicts of interest and the review authors declare that they do not have any associations with any parties who may have vested interests in the results of this review. Helen V Worthington is an Editor with Cochrane Oral Health, and was previously Co‐ordinating Editor. She was not involved in conducting the editorial process for the review.

New search for studies and content updated (conclusions changed)