Treatment for recurrent vulvovaginal candidiasis (thrush)

Georga Cooke; Cathy Watson; Laura Deckx; Marie Pirotta; Jane Smith; Mieke L Driel

doi:10.1002/14651858.CD009151.pub2

. 2022 Jan 10;2022(1):CD009151. doi: 10.1002/14651858.CD009151.pub2

Treatment for recurrent vulvovaginal candidiasis (thrush)

Georga Cooke ¹, Cathy Watson ², Laura Deckx ¹, Marie Pirotta ², Jane Smith ³, Mieke L Driel ^1,^✉

Editor: Cochrane STI Group

PMCID: PMC8744138 PMID: 35005777

Abstract

Background

Recurrent vulvovaginal candidiasis (RVVC) affects up to 5% of women. No comprehensive systematic review of treatments for RVVC has been published.

Objectives

The primary objective was to assess the effectiveness and safety of pharmacological and non‐pharmacological treatments for RVVC. The secondary objective was to assess patient preference of treatment options.

Search methods

We conducted electronic searches of bibliographic databases, including CENTRAL, MEDLINE, Embase, and CINAHL (search date 6 October 2021). We also handsearched reference lists of identified trials and contacted authors of identified trials, experts in RVVC, and manufacturers of products for vulvovaginal candidiasis.

Selection criteria

We considered all published and unpublished randomised controlled trials evaluating RVVC treatments for at least six months, in women with four or more symptomatic episodes of vulvovaginal candidiasis in the past year. We excluded women with immunosuppressive disorders or taking immunosuppressant medication. We included women with diabetes mellitus and pregnant women.

Diagnosis of RVVC must have been confirmed by presence of symptoms and a positive culture and/or microscopy. We included all drug and non‐drug therapies and partner treatment, assessing the following primary outcomes:

• number of clinical recurrences per participant per year (recurrence defined as clinical signs and positive culture/microscopy); • proportion of participants with at least one clinical recurrence during the treatment and follow‐up period; and • adverse events.

Data collection and analysis

Two authors independently reviewed titles and abstracts to identify eligible trials. Duplicate data extraction was completed independently by two authors. We assessed risk of bias as described in the Cochrane Handbook for Systematic Reviews of Interventions. We used the fixed‐effects model for pooling and expressed the results as risk ratio (RR) with 95% confidence intervals (CI). Where important statistical heterogeneity was present we either did not pool data (I² > 70%) or used a random‐effects model (I² 40‐70%). We used the GRADE tool to assess overall certainty of the evidence for the pooled primary outcomes.

Main results

Studies: Twenty‐three studies involving 2212 women aged 17 to 67 years met the inclusion criteria. Most studies excluded pregnant women and women with diabetes or immunosuppression. The predominant species found on culture at study entry was Candida albicans. Overall, the included studies were small (<100 participants). Six studies compared antifungal treatment with placebo (607 participants); four studies compared oral versus topical antifungals (543 participants); one study compared different oral antifungals (45 participants); two studies compared different dosing regimens for antifungals (100 participants); one study compared two different dosing regimens of the same topical agent (23 participants); one study compared short versus longer treatment duration (26 participants); two studies assessed the effect of partner treatment (98 participants); one study compared a complementary treatment (Lactobacillus vaginal tablets and probiotic oral tablets) with placebo (34 participants); three studies compared complementary medicine with antifungals (354 participants); two studies compared 'dermasilk' briefs with cotton briefs (130 participants); one study examined Lactobacillus vaccination versus heliotherapy versus ciclopyroxolamine (90 participants); one study compared CAM treatments to an antifungal treatment combined with CAM treatments (68 participants). We did not find any studies comparing different topical antifungals. Nine studies reported industry funding, three were funded by an independent source and eleven did not report their funding source.

Risk of bias: Overall, the risk of bias was high or unclear due to insufficient blinding of allocation and participants and poor reporting.

Primary outcomes: Meta‐analyses comparing drug treatments (oral and topical) with placebo or no treatment showed there may be a clinically relevant reduction in clinical recurrence at 6 months (RR 0.36, 95% CI 0.21 to 0.63; number needed to treat for an additional beneficial outcome (NNTB) = 2; participants = 607; studies = 6; I² = 82%; low‐certainty evidence) and 12 months (RR 0.80, 95% CI 0.72 to 0.89; NNTB = 6; participants = 585; studies = 6; I² = 21%; low‐certainty evidence). No study reported on the number of clinical recurrences per participant per year.

We are very uncertain whether oral drug treatment compared to topical treatment increases the risk of clinical recurrence at 6 months (RR 1.66, 95% CI 0.83 to 3.31; participants = 206; studies = 3; I² = 0%; very low‐certainty evidence) and reduces the risk of clinical recurrence at 12 months (RR 0.95, 95% CI 0.71 to 1.27; participants = 206; studies = 3; I² = 10%; very low‐certainty evidence). No study reported on the number of clinical recurrences per participant per year.

Adverse events were scarce across both treatment and control groups in both comparisons. The reporting of adverse events varied amongst studies, was generally of very low quality and could not be pooled. Overall the adverse event rate was low for both placebo and treatment arms and ranged from less than 5% to no side effects or complications.

Authors' conclusions

In women with RVVC, treatment with oral or topical antifungals may reduce symptomatic clinical recurrences when compared to placebo or no treatment. We were unable to find clear differences between different treatment options (e.g. oral versus topical treatment, different doses and durations). These findings are not applicable to pregnant or immunocompromised women and women with diabetes as the studies did not include or report on them. More research is needed to determine the optimal medication, dose and frequency.

Plain language summary

The effectiveness and safety of treatments for recurrent vulvovaginal candidiasis (thrush)

Review question

Cochrane authors set out to investigate the effectiveness and safety of drug treatments or non‐drug treatments (such as complementary and alternative medicines) for women with recurrent vulvovaginal candidiasis (thrush).

Background

Thrush of the vaginal/vulvar area is caused by Candida, a form of yeast that is commonly found in the vagina as part of normal flora without causing symptoms. For unknown reasons, Candida can start growing and cause symptoms of vulvovaginal candidiasis (commonly named thrush). Symptoms of vulvovaginal thrush include itching, swelling, and irritation of the vaginal and vulval areas. It is estimated that uncomplicated vulvovaginal thrush affects up to 75% of women at some time during their reproductive years. Recurrent vulvovaginal candidiasis (RVVC) occurs when a woman has four or more fungal infections during a 12‐month period. Up to 5% of women suffer from RVVC. Some doctors advise taking antifungals as a prevention, but there are no clear evidence‐based guidelines.

Study characteristics

We identified 23 studies involving 2212 participants between 17 and 67 years old with a diagnosis of RVVC confirmed by a positive test. The studies compared a range of antifungal medicines (taken by mouth or inserted into the vagina), and some complementary or alternative treatments (such as Lactobacillus vaccines or probiotics and special underwear). The studies reported the effects on recurrence of thrush after 6 and 12 months. Only one study reported the number of clinical recurrences at 12 months. Nine studies reported industry funding, three were funded by an independent source and nine studies did not report their funding source. The evidence is current to October 2021.

Key results

Six studies (607 participants) compared antifungal medication with placebo or no treatment. Based on low certainty evidence antifungal treatments may give a clinically relevant reduction in clinical recurrence at 6 and 12 months. We are uncertain if oral treatments are better than topical treatments (very low certainty evidence for no difference).

Adverse effects of taking antifungal medication to prevent recurrence of thrush were not common. Studies reported on adverse effects differently, making comparisons difficult. More research is needed to determine the optimal medication, dose and frequency. We were unable to determine the effects in women who are pregnant or have diabetes.

Certainty of the evidence

Our confidence in the evidence was low to very low due to serious concerns about risk of bias, unclear reporting and the limited number of studies that could be combined. This means that new studies might change our confidence in the effects of treatments and our conclusions.

Summary of findings

Background

Description of the condition

Candida is a yeast that is commonly found in the vagina as part of normal flora without causing symptoms. For largely unknown reasons, Candida changes from being a commensal organism (i.e. it can live in the environment without causing problems for the host) to a pathogenic one (Sobel 2007), which causes symptoms of vulvovaginal candidiasis (VVC, commonly named thrush) (Sobel 2007). VVC can be diagnosed clinically, with signs and symptoms including vaginal itching or burning with or without redness and swelling of the vulvae, and a white discharge, and stinging or burning when passing urine. Diagnosis can be confirmed with a wet mount preparation (saline with 10% KOH) or Gram stain of the discharge, which will show budding yeasts and hyphae under the microscope, and/or a culture of Candida (CDC 2015).

It is estimated that uncomplicated VVC affects up to 75% of women some time during their reproductive years (Sobel 2007), although these figures have been disputed (Rathod 2014). Predisposing factors have been identified for uncomplicated VVC (Patel 2004), which include the use of antibiotics, hormone replacement therapy, pregnancy, diabetes mellitus, genetic factors, and behavioural factors (Sobel 2006).

Up to 5% of women suffer from recurrent vulvovaginal candidiasis (RVVC), which is commonly defined as four or more episodes of VVC in a 12‐month period (Sobel 2007). On a global scale RVVC affects approximately 138 million women each year (range 103‐172 million). The global prevalence is estimated as 3871 per 100 000 women with 372 million women affected over their lifetime (Denning 2018), specifically women between 25 and 35 years old. Denning calculated that in high‐income countries the economic burden of lost productivity could be as high as US$14.39 billion per year (Denning 2018). The role of predisposing factors for uncomplicated vulvovaginal candidiasis is not certain in RVVC. In approximately half of women with RVVC, no risk factors can be identified (Nyirjesy 2008).

The aetiology of RVVC is unclear. Most cases (85% to 95%) of uncomplicated vulvovaginal candidiasis are caused by Candida albicans. However, less common candidal species such as Candida glabrata may be implicated in RVVC. Vaginitis caused by non‐albicans species tends to be more resistant to treatment (Sobel 2007). Many theories, such as maladaptive immune response and involvement of polymorphonuclear leucocytes, remain controversial (Fidel 1998; Peters 2014; Sobel 2007), and the role of individual and genetic susceptibility has not yet been defined (Sobel 2007).

The effects of RVVC on the intimate relationships and daily living of women can be significant. Ehrstrom 2007 found that women with RVVC were significantly more likely to report signs of burnout, emotional and physical stress, and poor work‐life balance than a control group of women without RVVC. The major impact of the physical symptoms, including discharge, itchiness, pain and psychological effects, often goes unrecognised (Chapple 2000). Additionally, in some countries long‐term treatment can be expensive (Sobel 2014), and approximately 50% of women experience recurrence of symptoms within months of finishing treatment (Sobel 2004b).

Description of the intervention

Management of uncomplicated vulvovaginal candidiasis usually consists of topical or oral antifungal treatments with frequency ranging from a single dose to treatment for up to 14 days (Pappas 2015; Therapeutic Guidelines 2015). Antifungal drugs fall broadly into five classes: polyenes, azoles, allylamines, echinocandins, and other agents, including griseofulvin and flucytosine (Chen 2007). The most commonly recommended antifungals for the treatment of vulvovaginal thrush are azoles (e.g. clotrimazole, fluconazole, itraconazole, miconazole, ketoconazole) and nystatin (a polyene). These treatments have been shown to be effective as topical treatment in pregnant women (Young 2001) and non‐pregnant women (Lopez 2013). Systematic reviews (Lopez 2013; Nurbhai 2007 and Watson 2002) found that oral and intravaginal (topical) treatments were equally effective in women with uncomplicated acute (i.e. less than four episodes in 12 months) vulvovaginal candidiasis, but Nurbhai 2007 found that women generally prefer oral treatment.

The recommended treatment regimen for complicated vulvovaginitis such as RVVC as outlined in clinical guidelines is not effective for all women (Pappas 2015; Therapeutic Guidelines 2015), whether oral or topical (Shahid 2009). Some women experience adverse effects including headache, abdominal pain, and nausea with oral treatment and dyspareunia or irritation with vaginal treatment (Sobel 1995; Stein 1993).

Topical antifungal agents are applied to the vaginal mucosa, and are delivered in the form of creams or pessaries. Oral antifungal treatment exists in the form of tablets or capsules. The currently recommended treatment for RVVC rarely cures this condition but rather aims to suppress symptoms (Pappas 2015). Initially high doses of oral or topical antifungal agents are used for up to two weeks to induce suppression of symptoms. In severe cases, the suppression of symptoms may take up to six months (Sobel 2004b). This suppression period is followed by long‐term regular (weekly or monthly) antifungal treatment to maintain clinical remission (Pappas 2015; Therapeutic Guidelines 2015).

Hormones, such as depo‐medroxyprogesterone (Falagas 2006; Miller 2000) have been suggested as treatment for VVC, but their efficacy has been questioned by some studies (Pirotta 2004), and the lack of supporting evidence has been highlighted by others (Van Kessel 2003).

Other treatment options include changing contraceptive, partner treatment, and use of topical gentian violet. These options have been poorly researched.

Conventional treatment of RVVC is not always successful and often involves long‐term or multiple treatments (Pappas 2015; Therapeutic Guidelines 2015). Significant variation in approach is seen between individual practitioners (Watson 2011). Due either to the limitations of current treatments or patient preference, many women seek treatment alternatives. Complementary and alternative medicine (CAM) is highly acceptable to women and is widely used in managing this condition (Nyirjesy 2001). One study found that up to 40% of women use CAM to treat or prevent vulvovaginal candidiasis despite the wide availability of conventional antifungal agents (Pirotta 2003). Examples of CAM used are herbal preparations such as tea tree oil, garlic, probiotics such as Lactobacillus, and vaginal acidifying agents. Evaluation of the effectiveness and safety of CAM for vulvovaginal candidiasis in the literature has been sparse. A review of the evidence concludes the effectiveness of oral or intravaginal garlic, intravaginal teat tree oil and yoghurt (containing Lactobacillus acidophilus) either oral or intravaginal, is unknown (Lopez 2013). Garlic taken orally can cause gastro‐intestinal symptoms such as heartburn, nausea, diarrhoea and bloating, and prolonged intravaginal use may result in allergic reactions or chemical burns (Lopez 2013). Topical tea tree oil can irritate the skin and cause a severe allergic rash (Lopez 2013). Yoghurt can cause gastrointestinal discomfort in people with lactose intolerance (Lopez 2013).

A fungal immunotherapeutic vaccine (NDV‐3A) has recently been the subject of phase 2 trials and shows potential in reducing recurrences of VVC (Edwards 2018). Given that resistance to conventional antifungal treatment for RVVC has been reported and that conventional therapy does not work for all women (Sobel 2016), it is vital to investigate novel treatments that may increase options for RVVC management.

How the intervention might work

The mode of action of antifungal agents is generally fungistatic, that is they Interfere with biosynthesis or integrity of ergosterol, the major sterol in the fungal cell membrane, and induce breakdown of the fungal cell wall (Chen 2007; Wildfeuer 1997). The most commonly used antifungal class is the azoles, which is made up of the imidazoles (e.g. ketoconazole, miconazole, clotrimazole and econazole) and the triazoles (e.g. fluconazole and itraconazole) (Chen 2007). Imidazole drugs are used in superficial fungal infections, while triazoles have a place in both superficial and invasive fungal infections (Chen 2007). The azoles act by inhibiting the fungal cytochrome P450–Erg11p or cytochrome P51p enzymes, inhibiting 14a‐demethylation of lanosterol in the ergosterol biosynthetic pathway (Odds 2003). This alters the fluidity and permeability of the fungal cell wall.

Alteration of the host environment so that it is less favourable for the proliferation of Candida is the underlying theory supporting the use of other therapies such as probiotics, gels used to restore vaginal acid balance, and treatments such as depo‐medroxyprogesterone (Falagas 2006; Miller 2000).

Some less commonly used treatments such as immunotherapy claim to promote the restoration of the immune system (Moraes 2000).

Why it is important to do this review

A number of systematic reviews address treatment options for uncomplicated acute VVC (Nurbhai 2007; Xie 2017; Young 2001). However, RVVC poses different challenges and there is currently no comprehensive systematic review of the treatment options for RVVC. Available guidelines are based on consensus only. RVVC is a common condition that can severely impact women and their partners both physically and psychologically. Many treatments are expensive, and the evidence for the efficacy of several treatments has not been systematically collated. There are also concerns about safety, especially for the non‐conventional complementary and alternative treatment. Evaluation of the many therapies used to manage this condition is essential to provide information so that women and their healthcare practitioners can make informed decisions about the management of RVVC.

Objectives

The primary objective of this systematic review was to assess the effectiveness and safety of pharmacological and non‐pharmacological treatments for recurrent vulvovaginal candidiasis. The secondary objective of the review was to assess patient preference of these treatment options.

Methods

Criteria for considering studies for this review

Types of studies

We considered all published and unpublished randomised controlled trials evaluating treatments for RVVC for at least six months for inclusion. Outcome measurement at six months was considered meaningful as the focus of this review is on managing a recurring condition rather than managing a single episode. We considered all types of randomised trials (including parallel and cluster randomised trials). We did not identify any cluster‐randomised trials.

Types of participants

We included all women with at least four symptomatic episodes in the past year of vulvovaginal candidiasis, confirmed by the presence of symptoms and a positive culture or symptoms and positive microscopy. We included women with diabetes and women who are pregnant. We excluded women with immunosuppressive disorders or taking immunosuppressant medication.

We included all settings (e.g. family medicine clinic, gynaecology outpatient clinic, sexually transmitted disease or family planning clinic).

Types of interventions

We considered the following interventions.

Antifungal treatments, administered intravaginally or orally
Other treatments for candida vulvovaginitis, such as probiotics, gentian violet, acidifying agents (including cider vinegar, boric acid, and vaginal gels), tea tree oil, douching, garlic and dietary modification
Partner treatment

We made the following comparisons.

Drug treatment (oral and topical) versus placebo/no treatment (excluding partner treatment)
Oral drug treatment versus topical drug treatment
Oral drug treatment versus oral drug treatment
Topical drug treatment versus topical drug treatment
Comparison of different doses of the same agent
Short duration of treatment versus longer duration of treatment
Partner treatment versus placebo/no treatment
Complementary and alternative medicine versus placebo/no treatment
Complementary and alternative medicine versus drug treatments
Complementary and alternative medicine versus non‐drug treatments
Vaccination versus other treatment

Types of outcome measures

We included the following primary and secondary outcome measures.

Primary outcomes

Number of clinical recurrences per participant per year (recurrence defined as clinical features and positive culture or microscopy)
Proportion of participants with at least one clinical recurrence during the treatment and follow‐up period
Adverse events

Secondary outcomes

Time to first recurrence
Number of symptomatic days per year
Number of mycological recurrences per participant per year
Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period
Duration of symptoms after treatment initiation
Patient preference

Search methods for identification of studies

We attempted to identify as many relevant randomised controlled trials (RCTs) as possible of "treatment" for "recurrent vulvovaginal candidiasis", irrespective of language of publication, publication date, and publication status (published, unpublished, in press, and in progress). We used both electronic searching in bibliographic databases and handsearching, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017).

Electronic searches

We contacted the Information Specialist of the Cochrane Sexually Transmitted Infections Group and our local Information Specialist to design a comprehensive search strategy to identify as many relevant RCTs as possible in the electronic databases. We used a combination of controlled vocabulary (MeSH, Emtree, DeCS, including exploded terms) and free‐text terms (considering spelling variants, synonyms, acronyms and truncation) for "vulvovaginal candidiasis" and "recurrence", with field labels, proximity operators, and Boolean operators. There were no date or language restrictions.

We searched the following electronic databases.

Cochrane Central Register of Studies Online (CENTRAL) (inception to 6 October 2021) (Appendix 1)
MEDLINE, Ovid platform (1946 to 6 October 2021) (Appendix 1)
Embase, Embase.com platform (inception to 6 October 2021) (Appendix 1)
CINAHL (Cumulative Index to Nursing and Allied Health Literature) (inception to 11 September 2020) (Appendix 2). Any later CINAHL output is contained in the CENTRAL 6 October 2021 search output.

The following appendices present earlier searches of CENTRAL (Appendix 3; Appendix 4), MEDLINE (Appendix 3; Appendix 5), Embase (Appendix 3; Appendix 6), Web of Science (Appendix 7), and ProQuest (Appendix 8).

For MEDLINE, we used the Cochrane Highly Sensitive Search Strategy for identifying RCTs: sensitivity‐ and precision‐maximising version Ovid format (Deeks 2017).

Searching other resources

We searched the following trial registers on 27 July 2019. Ongoing trials from both trial registries were included in the 2021 CENTRAL search.

US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov/) (Appendix 9)
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch/) (Appendix 10)

We handsearched reference lists of systematic reviews, other relevant publications on the same topic, and the reference lists of all included studies to identify further potentially eligible studies. We also contacted the authors of identified trials, experts in RVVC, and manufacturers producing products for vulvovaginal candidiasis to identify unpublished studies.

Data collection and analysis

Selection of studies

Two review authors (GC and CW) independently reviewed the titles and abstracts identified by the search for potentially relevant studies and then assessed the full texts of those studies deemed potentially relevant for inclusion in the review. We excluded double publications and reports of the same study at different points of time of follow up. Any disagreements were resolved by discussion or by consulting with a third review author (MVD). We contacted trial authors if we needed more information before deciding on inclusion.

Data extraction and management

Two review authors (GC and CW) independently extracted data using a paper data extraction form. Any discrepancies were resolved by discussion or by consulting with a third review author (MVD). The data extraction form included information on study citation; study characteristics (design, sample size, randomisation method, blinding, measurement of outcomes, method of analysis, (intention‐to‐treat, per protocol, etc.), duration of follow‐up; population (age of participants, inclusion and exclusion criteria, setting, number of excluded patients, number lost to follow up, number of analysed patients); interventions (drugs/treatments, dose, route of administration, duration of treatment, comparisons); outcomes (primary and secondary outcomes, including definition of objective and subjective outcomes, attrition); and other issues (including information on funding) (see Characteristics of included studies). We contacted authors for additional clarification or information if necessary. We performed data analysis using Review Manager 5 software (Review Manager 2014).

Assessment of risk of bias in included studies

Four review authors (JS, MP, LD, and MVD) independently evaluated risk of bias in the included trials as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), including assessment of sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. Any discrepancies were resolved by discussion or by consulting with a third review author (MVD or GC). We contacted authors of trials for more information if necessary.

We assessed the risk as 'low' if sufficient information was reported and the method described was adequate. We assessed the risk as 'high' if the described method was inadequate or if there was insufficient information to permit an assessment and it was likely that a rigorous method was not used. We assessed the risk of bias as 'unclear' if insufficient information was reported to permit assessment of the methods used.

Random sequence generation (selection bias)

We assessed the method used to generate the allocation sequence as described in the study publication (and/or protocol if available) as:

'low risk' if a random process was used, e.g. random number table or computer random number generator;
'high risk' if a non‐random process was used, e.g. odd or even date of birth, alternating dates, hospital record number; or
'unclear risk' if the trial is described as randomised, but the method used for the allocation sequence generation is not or is insufficiently described.

Allocation concealment (selection bias)

We assessed the method used to conceal the allocation sequence as described in the study publication (and/or protocol if available) as: • 'low risk' if an appropriately concealed method was used (e.g. central randomisation; sealed, opaque envelopes); • 'high risk' if open random allocation was used or other unblinded methods (e.g. unsealed or non‐opaque envelopes, alternate dates); or • 'unclear risk' if the trial is described as randomised but the method used to conceal sequence allocation is not or is insufficiently described.

Blinding of participants and personnel (performance bias)

We assessed the method used to ensure blinding of participants and personnel as described in the study publication (and/or protocol if available) as:

• 'low risk' if an appropriate method of blinding was used (e.g. identical appearance of treatments and identical administration routines); • 'high risk' if the method used was transparent (e.g. oral versus topical administration); or • 'unclear risk' if the trial is described as blinded but the method used to ensure blinding is not or is insufficiently described.

Blinding of outcome assessment (detection bias)

We assessed the risk of bias related to blinding of outcome assessment as described in the study publication (and/or protocol if available) as:

• 'low risk' if outcome assessors were unable to determine the treatment allocated; • 'high risk' if outcome assessors had knowledge or could have had knowledge of the allocated treatment, and this could have influenced their assessment (e.g. in the case of instruments assessed through interview); or • 'unclear risk' if information about blinding of outcome assessors was insufficient to permit a judgement of low or high risk of bias.

Incomplete outcome data (attrition bias)

For each study, we reported the number of participants randomised and the number for whom outcomes were reported for the different outcomes and at different time points. We assessed the completeness of reporting and attrition as described in the study publication as: • 'low risk' if outcomes are reported for all or more than 80% of randomised participants across groups; • 'high risk' if missing data are not balanced between groups and/or outcome data are missing for > 20% of randomised participants and no explanation is provided; or • 'unclear risk' if this is insufficiently described.

Selective reporting bias

We assessed the risk of reporting bias by comparing the outcomes reported in the study protocol (if available) or the methods section of the published trial report with the outcomes reported in the results section of the published trial report as: • 'low risk' if it is clear that all of the study’s prespecified outcomes have been reported; • 'high risk' if not all of the study’s prespecified outcomes have been reported or one or more reported primary outcomes were not prespecified; or

•'unclear risk' if the information provided in the study report is insufficient.

Other bias

We assessed the risk of other types of bias such as conflicts of interest or funding sources that could have impacted the study results as: • 'low risk' of other bias; • 'high risk' of other bias; or • 'unclear risk' of other bias.

We took risk of bias into account when making a judgement about the overall certainty of the pooled evidence as per the GRADE tool.

Measures of treatment effect

For dichotomous data, we reported the results as risk ratios with 95% confidence intervals. For continuous data, we planned to report mean difference of pre‐ and post‐measurements or weighted mean difference if different scales were used. We also reported standard deviations.

Unit of analysis issues

The unit of analysis in meta‐analysis was the woman. If the woman is not the unit of randomisation, such as is the case in cluster‐randomised trials, we would make adjustments for clustering following the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). In the case of multiple study arms, we combined the active interventions if this was clinically appropriate and compared them to placebo, in order to avoid double counting.

Dealing with missing data

We attempted to obtain information on missing data from study authors. If this was not successful, we performed intention‐to‐treat (ITT) analysis, which considers all missing data as treatment failures. We followed the guidance in the Cochrane Handbook for Systematic Reviews of Interventions to perform a sensitivity analysis that explored the impact of studies with high risk of bias due to large numbers of missing data (Higgins 2011).

Assessment of heterogeneity

We assessed heterogeneity between trials that investigate effects of the same interventions in two ways. Before pooling, we first explored clinical, face value heterogeneity (study populations, interventions, etc.). If we considered the trials to be too different (e.g. clinically different populations, different interventions, clinically non‐comparable outcomes), we did not pool the studies. We then assessed statistical heterogeneity by visual exploration of the results on the forest plot and by performing a Chi² test and calculating the Higgins I² statistic according to the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017). We interpreted the I² value as follows (Higgins 2011):

0% to 40%: may not be important; a fixed effects model is recommended;
40% to 70%: may represent moderate heterogeneity; a random effects model is recommended;
>70%: may represent substantial heterogeneity; pooling is not recommended.

Assessment of reporting biases

If outcome data for relevant outcomes that were mentioned in the study protocol or methods of the study publication were not reported (i.e. risk of reporting bias), we contacted the authors to request information about the missing outcome data. We planned to assess publication bias by generating funnel plots if 10 or more trials were available for the comparison (Higgins 2011).

Data synthesis

We used Review Manager 5 software to perform the statistical analysis (Review Manager 2014). Where we judged the trial populations, methods, and outcome measures to be similar (low clinical heterogeneity), and the I² was less than 40% (low statistical heterogeneity), we pooled the data using a fixed‐effect model (Higgins 2011). We did not pool data when there was obvious clinical heterogeneity and pooling studies did not make clinical sense (see Assessment of heterogeneity). Where statistical heterogeneity was moderate (I² 40% to 70%) or substantial (I² >70%), we pooled the data using a random‐effects model and performed a sensitivity analysis to assess the impact of heterogeneity on the robustness of the overall effect estimate (see Sensitivity analysis) (Higgins 2011).

We calculated a number needed to treat for an additional beneficial outcome (NNTB) and a number needed to treat for an additional harmful outcome (NNTH) where data were available and the results were statistically significant. We used the formulae provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Subgroup analysis and investigation of heterogeneity

We planned to carry out the following subgroup analyses based on relevant variables.

Sexually active women versus non‐sexually active women
Pregnant versus non‐pregnant women
Women with diabetes mellitus versus non‐diabetic women
RVVC caused by Candida albicans versus non‐albicans species

Heterogeneity was assessed as per the guidance in Assessment of heterogeneity.

Sensitivity analysis

We performed the following sensitivity analyses:

To assess the effect of risk of selection bias on the overall estimate of the meta‐analysis we first pooled all studies and then we excluded those studies with high risk of selection bias in order to evaluate the change in the pooled effect estimator.

Summary of findings and assessment of the certainty of the evidence

We generated the Summary of findings tables using GRADEpro and Cochrane methods (Higgins 2011; GRADEpro GDT). These tables evaluate the overall quality of the body of evidence for the main review outcomes (proportion of participants with at least one clinical recurrence during the treatment and follow‐up period, adverse events) for the main review comparisons (Drug treatment (oral and topical) versus placebo/no treatment (excluding partner treatment) and Oral drug treatment versus topical drug treatment). We assessed the certainty of the evidence using GRADE criteria (risk of bias, consistency of effect, imprecision, indirectness and publication bias). Judgements about evidence certainty (high, moderate, low or very low) were made by two review authors working independently (GC and LD), with disagreements resolved by discussion and an arbiter (MVD). We give a narrative description of the effects taking into account the level of certainty as per the handbook recommendations (Higgins 2011).

Results

Description of studies

The results of the searches and included and excluded studies are described below.

Results of the search

The searches up to 6 October 2021 identified 1782 references from electronic databases (trial registries excluded). We identified an additional four publications from contacting experts in the field (including the authors of this review) (Fan 2015; Fardyazar 2007 (2 publications); (Sobel 1985a), and one additional publication from contacting manufacturers (Merkus 1990). Our review of reference lists of included studies and publications selected from the electronic database searches identified one full‐text study for review (Guaschino 2001). We identified a further study from the trial registry searches (Rabiee 2013). After removal of duplicates, 1181 abstracts were available for review. Of these 1070 were excluded as they did not meet our inclusion criteria. We selected a total of 111 abstracts for full‐text review (Figure 1). Eighty‐two studies were excluded because participants did not have RVVC (n = 57), the study was not an RCT (n = 17), the duration of follow‐up was less than 6 months (n = 6), or the study was not a primary study (n = 2). Six full‐text publications contained inadequate information to permit a decision regarding inclusion in the review (Coric 2006; Diba 2010; Fan 2005; Houang 1989; Miller 1984; Shalev 1996); we contacted the authors of these studies, but none of these publications were found to be suitable for inclusion.

Two studies were not published in English and we have placed them in Awaiting classification pending translation into English (Rabiee 2013 ‐ published in Persian; Zivaljevic 2012 ‐ published in Serbian). In one study it was not clear if women met the inclusion criteria for our review (Russo 2019).

Our search in 2020 identified three registered ongoing trials: NCT04029116 is recruiting, NCT04208555 is not yet recruiting. For EUCTR2019‐000925‐27‐SK it is unclear if recruitment has started, however, this trial has also been registered in clinicaltrials.gov which reports the trial is still recruiting (NCT04734405). We identified 10 additional trials from the trial registry searches in 2020, however, no publication was located for these trials. We contacted the primary contact identified in the database record, but received no additional information or trial results for seven trials (ACTRN12614001258640; ChiCTR‐IPR‐15006314; ChiCTR‐TRC‐10000833; EUCTR2013‐002480‐26‐PL; NCT00479947; NCT02251093; RBR‐892mp4). The contact listed in the database record could not be contacted for three trials (EUCTR2010‐021502‐38‐DE; EUCTR2011‐004718‐40‐IT; NCT00915629), either because there were insufficient details in the database or the details were not current. Thus this study has been assessed as awaiting classification.

The search in October 2021 identified three additional trials in clinicaltrials.gov: NCT04292704 is still recruiting; NCT04639544 and NCT04699240 have been assessed as awaiting classification as it is unclear from the information in the trials database whether microbiological confirmation of VVC was an inclusion criterion. NCT04734405 was identified as a duplicate registration of EUCTR2019‐000925‐27‐SK.

Included studies

We included 23 studies involving a total of 2212 participants in the review. We identified 20 studies from the electronic searches (Bolouri 2009; Chopra 2013; Corthay 1988; D'Antuono 2012; D'Antuono 2013; Fong 1992a; Fong 1992b; Fong 1994; Kumari 2011; Li 2018; Lopez‐Olmos 2000; Mendling 2011; Metts 2003; Roth 1990; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Spinillo 1997; Witt 2009). The remaining three studies were identified from the personal libraries of authors of this review (Fan 2015; Fardyazar 2007; Sobel 1985a). However, three studies did not report data for our primary outcomes in a way that could be extracted (Corthay 1988; D'Antuono 2013; Li 2018). We requested further information from D'Antuono 2013; Li 2018, but unfortunately the authors were not able to provide us with the required data. Corthay 1988 was a conference abstract from 1988 and the authors could not be traced.

Participants and setting

Overall, the included studies were small, with 16 of the 23 studies including fewer than 100 participants (Bolouri 2009; Corthay 1988; D'Antuono 2013; Fong 1992a; Fong 1992b; Fong 1994; Kumari 2011; Li 2018; Lopez‐Olmos 2000; Mendling 2011; Metts 2003; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989; Spacek 2005). The largest study included 387 participants (Sobel 2004). Eleven studies excluded participants with diabetes (Bolouri 2009; Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fong 1992a; Fong 1992b; Fong 1994; Kumari 2011; Mendling 2011; Spinillo 1997; Witt 2009). One study had no participants with diabetes (Sobel 1989). Five studies did not explicitly exclude women with diabetes but did not state if any participants had diabetes (Corthay 1988; Metts 2003; Roth 1990; Spacek 2005; Witt 2009). Six studies included some participants with diabetes. One study included a single participant with diabetes (Fan 2015). Almost 7% of participants had diabetes in Lopez‐Olmos 2000, and 11% of participants had diabetes in Fardyazar 2007, one had 3% (Sobel 1985a), one had 2% (Sobel 1986), and one had 2% in the fluconazole group and 5% in the placebo group (Sobel 2004).

We did not find any study including pregnant women. Eighteen studies excluded pregnant women (Bolouri 2009; Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fan 2015; Fong 1992a; Kumari 2011; Li 2018; Mendling 2011; Metts 2003; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Spinillo 1997; Witt 2009). One study did not exclude pregnant women, but none enrolled (Fardyazar 2007). Four further studies also did not explicitly exclude pregnant women, but did not state if any participants were pregnant (Corthay 1988; Fong 1992b; Fong 1994; Lopez‐Olmos 2000).

Five studies required participants to have a culture C albicans at enrolment (Roth 1990; Sobel 1985a; Sobel 1986; Spinillo 1997; Witt 2009). Nine studies did not report the prevalence of C albicans at enrolment (Corthay 1988; Chopra 2013; Fardyazar 2007; Fong 1992a; Fong 1992b; Fong 1994; Li 2018; Mendling 2011; Metts 2003). In the remaining studies, the predominant Candida species at enrolment was C albicans, with between 27% and 100%. Lopez‐Olmos 2000 did not require a positive culture at enrolment; however, if the culture was positive, it was most commonly C albicans.

Five studies required women to have a negative culture for Candida spp prior to commencing the treatment phase (Bolouri 2009; Fan 2015; Roth 1990; Sobel 1989; Sobel 2004). In one study participants were given additional treatment to achieve negative cultures for Candida spp (Spinillo 1997), and in four studies participants did not need to have a negative culture for Candida spp (Fardyazar 2007; Fong 1992a; Li 2018; Sobel 1986).

The most common setting was an outpatient clinic (Bolouri 2009; Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fan 2015; Fardyazar 2007; Fong 1992a; Fong 1992b; Kumari 2011; Li 2018; Spacek 2005; Spinillo 1997; Witt 2009). One study was conducted exclusively in a primary care setting (a university student health centre) (Metts 2003). Two studies included participants from multiple sites (gynaecology/family planning clinics) (Lopez‐Olmos 2000; Mendling 2011). A third study was also multicentre, but the settings were not described (Sobel 2004). The setting was not stated in six studies (Corthay 1988; Fong 1994; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989).

Ten studies were conducted in Europe (Austria, the Czech Republic, Germany, Italy, Spain, and Sweden) (Chopra 2013; D'Antuono 2012; D'Antuono 2013; Kumari 2011; Lopez‐Olmos 2000; Mendling 2011; Roth 1990; Spacek 2005; Spinillo 1997; Witt 2009), and eight in North America (Fong 1992a; Fong 1992b; Fong 1994; Metts 2003; Sobel 1985a; Sobel 1986; Sobel 1989; Sobel 2004). One study did not state the location (Corthay 1988). The remaining four studies were conducted in Asia, two in Iran ( Bolouri 2009; Fardyazar 2007) and two in China (Fan 2015; Li 2018).

Interventions and comparators

The included studies involved a variety of comparisons. Six studies compared drug treatments to placebo or no treatment (Bolouri 2009; Roth 1990; Sobel 1986; Sobel 1989; Sobel 2004; Spinillo 1997). Four studies compared oral drug treatment with topical drug treatment (Fan 2015; Fardyazar 2007; Fong 1992a; Lopez‐Olmos 2000), and one compared two types of oral drugs (Lopez‐Olmos 2000), but none compared different topical agents. Two studies compared different oral doses of the same agent (Sobel 1986; Spacek 2005), and one compared different oral treatment durations (Spacek 2005). One study compared two different dosing regimens of the same topical agent (Fong 1994). Two studies compared partner treatments with placebo or no treatment (Fong 1992b; Sobel 1985a), and two studies compared non‐drug treatments (DermaSilk briefs) with other treatments (cotton briefs) (D'Antuono 2012; D'Antuono 2013). One study compared adenosine triphosphate (ATP)‐infrared bio‐effect treatment and a nursing intervention with usual care (Li 2018). CAM treatments were compared with placebo in one study (Metts 2003), and with drug treatments in three studies (Chopra 2013; Kumari 2011; Witt 2009). Another study compared CAM treatments (HPT with an aqueous Candida albicans extract to an antifungal treatment (local + digestive), then a low‐sugar diet with vitamins, biotin and lyophilized bacillus subfilis supplements (Corthay 1988). One study compared vaccination therapy with drug treatment and alternative medicine (Mendling 2011).

Fifteen studies had an intervention duration of six months (Bolouri 2009; D'Antuono 2012; D'Antuono 2013; Fardyazar 2007; Fong 1992a; Fong 1994; Kumari 2011; Li 2018; Lopez‐Olmos 2000; Metts 2003; Roth 1990; Sobel 1989; Sobel 2004; Spinillo 1997; Witt 2009). One study had an intervention duration of three months (Corthay 1988). The three arms in Sobel 1986 had variable but similar lengths (6 menstrual cycles, 5 months, 6 months). One study had a 24‐month intervention period (Chopra 2013). The two partner treatment studies had shorter durations: Fong 1992b only treated participants once (for a five‐day course), and Sobel 1985a treated partners for one week. The duration of the interventions varied in Mendling 2011. Spacek 2005 randomised participants to one or three days of treatment.

Outcome measurement

Fifteen studies continued outcome measurement beyond the intervention period as per their study protocols (additional ~6 months ‐ Bolouri 2009; Corthay 1988; Chopra 2013; Fan 2015; Fardyazar 2007; Fong 1992a; Kumari 2011; Lopez‐Olmos 2000; Roth 1990; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Spinillo 1997; Witt 2009; two studies additional ~12 months ‐ Fong 1992b; Sobel 1985a; one study was variable due to the variable length of the interventions ‐ Mendling 2011). Eleven studies stated that they assessed participants between planned visits if they developed symptoms in the intervening period (Chopra 2013; Fan 2015; Fong 1992b; Mendling 2011; Metts 2003; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989; Spinillo 1997; Witt 2009).

Reporting of primary outcomes

All included studies, except three studies (Corthay 1988; D'Antuono 2013; Li 2018), provided information for one of our primary outcomes. Only one study presented data for number of clinical recurrences per participant per year (Mendling 2011). Eighteen studies reported the proportion of women with clinical recurrence (Bolouri 2009; Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fardyazar 2007; Fong 1992a; Fong 1992b; Kumari 2011; Lopez‐Olmos 2000; Metts 2003; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Spinillo 1997; Witt 2009), but in three studies the data for this outcome were not presented in a way that allowed for data extraction (D'Antuono 2013; Fong 1994; Metts 2003; Witt 2009). Sixteen studies reported adverse events (Bolouri 2009; Chopra 2013; D'Antuono 2012; Fan 2015; Fardyazar 2007; Fong 1992a; Fong 1992b; Kumari 2011; Mendling 2011; Metts 2003; Roth 1990; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Witt 2009).

Reporting of secondary outcomes

Two studies reported data for time to first recurrence (Sobel 1986; Sobel 2004). No studies reported on number of symptomatic days per year or number of mycological recurrences per participant per year. Eight studies reported data for proportion of participants with at least one mycological recurrence during the treatment and follow‐up period (Bolouri 2009; Fong 1992a; Kumari 2011; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 2004; Spinillo 1997). No studies reported on duration of symptoms after treatment initiation. Four studies reported data for patient preference (Fardyazar 2007; Fong 1994; Mendling 2011; Sobel 1985a).

See the Characteristics of included studies tables for more information on setting, participants, interventions, and outcomes of specific studies.

Excluded studies

After review of full text we excluded 23 studies that on face value appeared to have been eligible for inclusion (Avijgan 2012; Azima 2018; Bartz 2000; Brand 2018; Bushell 1988a; Bushell 1998b; Coric 2006; Davidson 1978; Diba 2010; Donders 2008; Edwards 2018; Fan 2005; Golero 1993; Hilton 1992; Houang 1989; Miller 1984; Rashid 1991; Shalev 1996; Silverman 1971; Sobel 1994; Sobel 2001; Vladareanu 2018; Yang 2000). The reasons excluding these studies are outlined in the Characteristics of excluded studies tables. The most common reason for exclusion was that the case definition of RVVC as reported in our protocol was not met (e.g. a case definition of only three episodes, rather than four episodes, of candidiasis in the past 12 months was used) (Avijgan 2012; Bartz 2000; Brand 2018; Bushell 1988a; Davidson 1978; Edwards 2018; Golero 1993; Silverman 1971; Sobel 1994; Sobel 2001; Yang 2000).

Risk of bias in included studies

Summaries of the 'Risk of bias' assessments for each study are outlined in Figure 2 and Figure 3.

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

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

Allocation

Random sequence generation

We assessed seven studies that used computer‐generated or other adequate random number sequence processes as at low risk of bias (Chopra 2013; Fong 1992a; Fong 1994; Kumari 2011; Metts 2003; Spinillo 1997; Witt 2009). Two studies were judged to be of high risk of bias for this domain. Li 2018 did not describe the sequence generation and method of randomisation. Fan 2015 was also classified as at high risk of bias for random sequence generation as women with RVVC known to be caused by C glabrata at admission (before Candida culture and identification) were not randomised but were enrolled in the nystatin treatment group. We assessed the remaining studies as at unclear risk of bias as the randomisation method was unclear (often claimed to be a randomised study but no description was provided to enable assessment of risk of selection bias).

Allocation concealment

We assessed four studies that used sealed bags or opaque envelopes as low risk of bias for allocation concealment (D'Antuono 2012; D'Antuono 2013; Fong 1992a; Metts 2003). We assessed six studies to be at high risk of bias for this domain (Fan 2015; Fong 1992b; Li 2018 Lopez‐Olmos 2000; Mendling 2011; Sobel 1985a). Lopez‐Olmos 2000 allocated treatment by day of the week, which implies lack of proper concealment and therefore a high risk of bias for this domain. There was no clear description of concealment in the remaining studies, which were therefore assessed as at unclear risk of selection bias (Bolouri 2009; Chopra 2013; Corthay 1988; Fardyazar 2007; Fong 1994; Kumari 2011; Roth 1990; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005; Spinillo 1997; Witt 2009).

Blinding

Performance bias

We assessed blinding against the primary outcomes, which all involved a component of patient reporting/symptoms. We initially considered only two studies as low risk of performance bias for the primary outcomes (Metts 2003; Roth 1990), as both were blinded to both participants and investigators. It was initially unclear if the studies by D'Antuono were adequately blinded (D'Antuono 2012; D'Antuono 2013), but author correspondence reassured us about the difficulty distinguishing between the different underwear fabrics, therefore we rated these studies as at low risk of bias for this domain. Five studies were assessed as having unclear risk of bias; two studies lacked clarity about whether the placebo was identical to the intervention (Bolouri 2009; Sobel 1986), one study did not refer to blinding (Corthay 1988) and the remaining two studies reported blinding but lacked clarity about whether it was participants, clinicians, or laboratory staff being blinded (Sobel 1989; Sobel 2004).

We assessed 14 studies as at high risk of bias for this domain (Chopra 2013; Fan 2015; Fardyazar 2007; Fong 1992a; Fong 1992b; Fong 1994; Kumari 2011; Li 2018; Lopez‐Olmos 2000; Mendling 2011; Sobel 1985a; Spacek 2005; Spinillo 1997; Witt 2009), due predominantly to lack of masking between the intervention and control groups. The three comparators used by Mendling 2011 were quite different, being ultraviolet A and B rays, vaginal pessaries, and intramuscular injections, and were therefore easy to distinguish.

Detection bias

We considered the risk of detection bias to be low in only one study (Roth 1990). Blinding of outcome assessment was unclear for sixteen of the studies (Bolouri 2009; Chopra 2013; Corthay 1988; D'Antuono 2012; D'Antuono 2013; Fan 2015; Fardyazar 2007; Fong 1992b; Fong 1994; Kumari 2011; Li 2018; Lopez‐Olmos 2000; Metts 2003; Sobel 1986; Sobel 1989; Sobel 2004; Spacek 2005). We assessed the risk of detection bias to be high for five studies, usually because outcomes were assessed by participants who were not blinded and the outcome was "soft" (Fong 1992b; Mendling 2011; Sobel 1985a; Spinillo 1997; Witt 2009).

Incomplete outcome data

We assessed eight studies with low attrition rates (15% or less) as well as good explanations for any loss to follow‐up as at low risk of attrition bias (Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fardyazar 2007; Fong 1992a; Fong 1992b; Fong 1994; Roth 1990). Sobel 2004 reported only the per‐protocol analysis but claimed that the results were similar in a modified intention‐to‐treat‐analysis for which the data were not shown, therefore we assessed the risk of attrition bias for this study to be unclear. Other studies that were deemed to have unclear risk of bias for this domain were two studies that described per‐protocol analysis rather than intention‐to‐treat (Lopez‐Olmos 2000; Spinillo 1997), and two studies where withdrawal and loss to follow‐up were not clearly described (Spacek 2005; Sobel 1985a). We considered 10 studies as at high risk of attrition bias due to high attrition rates or incomplete explanations for this, or both (Bolouri 2009; Corthay 1988; Fan 2015; Kumari 2011; Li 2018; Mendling 2011; Metts 2003; Sobel 1986; Sobel 1989; Witt 2009).

Selective reporting

No published protocol was available for any of the studies. We considered all but two studies as at unclear risk of reporting bias as there was insufficient information in the available publications to assess whether all intended outcomes had been reported. We assessed Kumari 2011 as at high risk for selective reporting because outcome measures were not described in sufficient detail. Sobel 2004 was also considered high risk of bias for this domain, as one of the primary outcome (number of clinical recurrences per participant per year) was not reported.

Other potential sources of bias

We considered four studies to a lowrisk of other bias, as the studies and their authors appeared to be independent of pharmaceutical financial support (Bolouri 2009; Fan 2015; Metts 2003; Spacek 2005). Three studies were funded by companies with vested interests such as pharmaceutical manufacturers and were therefore assessed as at high risk of bias (Fardyazar 2007; Sobel 2004; Witt 2009). Furthermore, in Fardyazar 2007 non‐compliant participants were excluded from the analysis. In Sobel 2004, the authors report that "about halfway through the enrolment period (before the data were analysed), the sponsor decided to combine the parallel trials into a single trial". The third study judged to be at high risk of bias did not provide details about similarity between treatment groups at baseline (Witt 2009). One study was considered at high risk of bias as a conference paper abstract only was located, with no full publication identified Corthay 1988. We considered the remaining 14 studies as at unclear risk of bias as the role of the founder was not clear or the funding source was not reported (Chopra 2013; D'Antuono 2012; D'Antuono 2013; Fong 1992a; Fong 1992b; Fong 1994; Kumari 2011; Lopez‐Olmos 2000; Mendling 2011; Roth 1990; Sobel 1985a; Sobel 1986; Sobel 1989; Spinillo 1997).

Effects of interventions

See: Table 1; Table 2

Summary of findings 1. Drug treatment compared to placebo or no treatment for recurrent vulvovaginal candidiasis.

Drug treatment compared to placebo/no treatment (excluding partner treatment) for recurrent vulvovaginal candidiasis (thrush)
Patient or population: non‐pregnant women with a diagnosis of recurrent vulvovaginal candidiasis Setting: outpatients Intervention: antifungals (fluconazole, clotrimazole, ketoconazole, itraconazole) Comparison: placebo or no treatment (excluding partner treatment)
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo/no treatment (excluding partner treatment)	Risk with drug treatment
Primary outcome 1: Number of clinical recurrences per participant per year	‐		‐	(0 studies)	‐	Not reported
Primary outcome 2: Clinical recurrence at 6 months	Study population		RR 0.36 (0.21 to 0.63)	607 (6 RCTs)	⊕⊕⊝⊝ LOW ¹
	647 per 1000	233 per 1000 (136 to 408)
Primary outcome 2: Clinical recurrence at 12 months	Study population		RR 0.80 (0.72 to 0.89)	585 (6 RCTs)	⊕⊕⊝⊝ LOW ¹
	781 per 1000	625 per 1000 (562 to 695)
Primary outcome 3: Adverse events	‐		Not estimable	(5 RCTs)	⊕⊝⊝⊝ VERY LOW ^{2 3}	Pooling of the incidence of adverse events was not possible as the outcome was reported differently across the included trials. Overall the adverse event rate was low for both placebo and treatment arms. Sobel 2004 reported that 2.9% of participants dropped out due to adverse events in the fluconazole group and 1.2% in the placebo group. 2 participants dropped out in Sobel 1986. No participants dropped out in Bolouri 2009. Less than 5% reported adverse effects in Sobel 1989, and no side effects or complications occurred in Roth 1990.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Oral drug treatment compared to topical drug treatment for recurrent vulvovaginal candidiasis
Patient or population: non‐pregnant women with a diagnosis of recurrent vulvovaginal candidiasis Setting: outpatients Intervention: oral drug treatment (fluconazole, itraconazole) Comparison: topical drug treatment (nystatin, clotrimazole)
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with topical drug treatment	Risk with oral drug treatment
Primary outcome 1: Number of clinical recurrences per participant per year	‐		‐	(0 studies)	‐	Not reported
Primary outcome 2: Clinical recurrence at 6 months	Study population		RR 1.66 (0.83 to 3.31)	206 (3 RCTs)	⊕⊝⊝⊝ VERY LOW ^{1 2}
	118 per 1000	196 per 1000 (98 to 392)
Primary outcome 2: Clinical recurrence at 12 months	Study population		RR 0.95 (0.71 to 1.27)	206 (3 RCTs)	⊕⊝⊝⊝ VERY LOW ^{1 2 3}
	495 per 1000	470 per 1000 (351 to 628)
Primary outcome 3: Adverse events	‐		Not estimable	(3 RCTs)	⊕⊝⊝⊝ VERY LOW ^{4 5}	Due to heterogeneity in reporting we were not able to pool the data.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Study ID	Bolouri 2009	Roth 1990	Sobel 1986	Sobel 1989	Sobel 2004	Spinillo 1997
Participants (n)	97	64	74	42	387	114
Candida albicansat enrolment (%)	78	100^a	87	86	94	100^a
Mean episodes in previous 12 months	NS^b	6.2 (clotrimazole), 6.5 (placebo)	NS^b	9.2 (clotrimazole), 8.3 (placebo)	NS^b	NS^b
Withdrawals/loss to follow‐up (%)	34	3	15	36	4	7
Setting	Hospital outpatient clinic	NS^b	NS^b	NS^b	Multicentre	Outpatient clinic
Country	Iran	Sweden	USA	USA	USA	Italy
Intervention 1	Fluconazole (weekly)	Clotrimazole (postmenstrual)*	Ketoconazole (with menses)	Clotrimazole (postmenstrual)*	Fluconazole (weekly)	Itraconazole (with menses)
Intervention 2	Placebo (weekly)	Placebo (postmenstrual)*	Ketoconazole (daily)	Placebo (postmenstrual)*	Placebo (weekly)	No treatment (with menses)
Intervention 3	‐	‐	Placebo (daily)	‐	‐	‐
Duration of intervention	6 months	6 months	6 months	6 months	6 months	6 months
Duration of observation after intervention	6 months	6 months	6 months	6 months	6 months	6 months
Planned assessments	Monthly for 6 months, then 9 months, 12 months	Monthly for 6 months, then 9 months, 12 months	Monthly for 12 months	Monthly for 6 months, then 9 months, 12 months	Monthly for 6 months, then 9 months, 12 months	3, 6, and 12 months
Assessment between planned visits if symptoms developed	‐	Yes	Yes	Yes	‐	Yes

Study ID	Fan 2015	Fardyazar 2007	Fong 1992a	Lopez‐Olmos 2000
Participants (n)	330	124	44	45
Candida albicansat enrolment (%)	77.5	‐	98	27^a
Mean episodes in previous 12 months	NS^b	NS^b	8.8 (itraconazole), 9.7 (clotrimazole)	NS^b
Withdrawals/loss to follow‐up (%)	11	6	14	8
Setting	Outpatient clinic	Outpatient clinic	Hospital outpatient clinic	Multicentre ‐ a gynaecology outpatient clinic and a family planning clinic
Country	China	Iran	Canada	Spain
Intervention 1	Fluconazole (weekly)	Fluconazole (weekly)	Itraconazole (twice weekly)	Fluconazole (with menses)
Intervention 2	Nystatin (with menses)	Clotrimazole (twice weekly)	Clotrimazole (twice weekly)	Itraconazole (with menses)
Intervention 3	‐	‐	‐	Clotrimazole (with menses)
Duration of intervention	6 months	6 months	6 months	6 months
Duration of observation after intervention	6 months	6 months	6 months	6 months
Planned assessments	Monthly for 6 months and at 9 months and 12 months	Monthly for 12 months	Monthly for 12 months	At 6 months and 12 months
Assessment between planned visits if symptoms developed	Yes	‐	‐	‐

Study ID	Fong 1994	Sobel 1986	Spacek 2005
Participants (n)	23	74	26
Candida albicans at enrolment (%)	‐	87	96
Mean episodes in previous 12 months	7.5	NS^a	NS^a
Withdrawals/loss to follow‐up (%)	0	15	4
Setting	NS	NS^a	Outpatient clinic
Country	Canada	USA	Czech Republic
Intervention 1	Clotrimazole (perimenstrually)	Ketoconazole (with menses)	Itraconazole, 400 mg
Intervention 2	Clotrimazole (empirically, with symptoms)	Ketoconazole (daily)	Itraconazole, 600 mg
Intervention 3	‐	Placebo (daily)	‐
Duration of intervention	6 months	6 months	1 to 3 days
Duration of observation after intervention	N/A Cross over trial	6 months	6 months
Planned assessments	Bimonthly for 12 months	Monthly for 12 months	14 days, 1 month, 3 months, 6 months
Assessment between planned visits if symptoms developed	‐	Yes	‐

*Study ID ()**	Fong 1992b	Sobel 1985a
Participants (n)	58	40
Candida albicansat enrolment (%)	‐	100
Mean episodes in previous 12 months	8.4 (ketoconazole), 9.5 (no treatment)	5.1 (miconazole), 5.3 (no treatment)
Withdrawals/loss to follow‐up (%)	7	9
Setting	Hospital outpatient clinic	NS^a
Country	Canada	USA
Intervention 1	Partner treatment with ketoconazole daily for 5 days	1% miconazole cream daily for 1 week
Intervention 2	No treatment	No treatment
Intervention 3	‐	‐
Duration of intervention	Single treatment	1 week
Duration of observation after intervention	12 months	12 months
Planned assessments	Monthly for 12 months	Monthly for 6 months, then 9 months, 12 months
Assessment between planned visits if symptoms developed	Yes	Yes

Study ID	Chopra 2013	Kumari 2011	Witt 2009
Participants (n)	122	82	150
Candida albicansat enrolment (%)	‐	54	100^a
Mean episodes in previous 12 months	‐	NS^b	NS^b
Withdrawals/loss to follow‐up (%)	0	10	53
Setting	Gynaecology outpatients clinic	Hospital outpatient clinic	NS^b
Country	Italy	Italy	Austria
Intervention 1	Itraconazole (weekly)	Itraconazole (weekly)	Itraconazole (weekly)
Intervention 2	Polygodyal/anethole phytocompound (twice daily for 1 week a month)	Polygodyal/anethole phytocompound (twice weekly)	Itraconazole (monthly) and lactobacilli (monthly for 6 days)
Intervention 3	‐	‐	Classic homeopathy
Duration of intervention	24 months	6 months	6 months
Duration of observation after intervention	6 months	6 months	6 months
Planned assessments	Monthly for 24 months, then 6 months	Monthly for 6 months, then 9 months, 12 months	Monthly for 12 months
Assessment between planned visits if symptoms developed	Yes	‐	Yes

*Study ID ()**	D'Antuono 2012	D'Antuono 2013
Participants (n)	100	30
Candida albicansat enrolment (%)	85	100
Mean episodes in previous 12 months	NS^a	NS^a
Withdrawals/loss to follow‐up (%)	4	0
Setting	Hospital outpatient clinic	Hospital outpatient clinic
Country	Italy	Italy
Intervention 1	DermaSilk briefs	DermaSilk briefs
Intervention 2	Cotton briefs	Cotton briefs
Intervention 3	‐	‐
Duration of intervention	6 months	6 months
Duration of observation after intervention	‐	‐
Planned assessments	1 months, 3 months, and 6 months	3 months and 6 months
Assessment between planned visits if symptoms developed	‐	‐

Search electronic report #1
Search type	Update
Databases	MEDLINE
Platform	Ovid
Search date	06/10/2021
Update date	Undefined
Range of search date	2019‐Current
Language restrictions	None
Other limits	None
Search strategy (results)	1. exp Candidiasis, Vulvovaginal/ (3632) 2. exp Candida/ (48200) 3. exp Candidiasis/ (32981) 4. candid.tw. (358579) 5 monilia.tw. (1266) 6. thrush.tw. (1206) 7. or/2‐6 (376084) 8. exp Vaginal Diseases/ (28738) 9. exp Vulvar Diseases/ (18167) 10. vulv.tw. (19712) 11. vagin.tw. (108669) 12. vulvovagin.tw. (3355) 13. vulvo‐vagin.tw. (352) 14. genit.tw. (80077) 15. or/8‐14 (201525) 16. 7 and 15 17. 1 or 16 (8599) 18. exp Recurrence/ (191862) 19. exp Secondary Prevention/ (21647) 20. exp Chemoprevention/ (21467) 21. (recur* or relaps* or prevent* or chemoprev* or chemoprophyla* or prophylax* or prophylac* or refractory or repeat* or repetit* or chronic*).tw. (3634433) 22. exp Chronic Disease/ (271265) 23. treatment failure/ (36424) 24. (treatment$ adj2 fail$).tw. (39494) 25. or/18‐24 (3798342) 26. 17 and 25 (2592) 27. randomized controlled trial.pt. (544063) 28. controlled clinical trial.pt. (94402) 29. randomized.ab. (463771) 30. placebo.ab. (201328) 31. clinical trials as topic.sh. (197484) 32. randomly.ab. (310818) 33. trial.ti. (215023) 34. or/27‐33 (1256039) 35. exp animals/ not humans.sh. (4889865) 36. 34 not 35 (1144770) 37. 26 and 36 (350) 38. limit 37 to yr="2020‐Current" (18)
# of records identified	18

Search electronic report #2
Search type	Update
Databases	EMBASE
Platform	ELSEVIER
Search date	06/10/2021
Update date	Undefined
Range of search date	2018‐Current
Language restrictions	None
Other limits	None
Search strategy (results)	1. vagina candidiasis'/exp 5,796 2. candida'/exp 93,690 3. candidiasis'/exp 55,58 4. candid:ab,ti OR monilia:ab,ti OR thrush:ab,ti 568,224 5. #2 OR #3 OR #4 612,797 6. vagina disease'/exp 70,593 7. vulva disease'/exp 20,455 8. vulv:ab,ti OR vagin:ab,ti OR vulvovagin:ab,ti OR genit:ab,ti OR ((vulvo NEAR/1 vagin):ab,ti) 301,184 9. #6 OR #7 OR #8 329,889 10. #5 AND #9 15,697 11. #1 OR #10 15,697 12. recurrent disease'/de 194,304 13. secondary prevention'/de 31,012 14. prophylaxis'/de OR 'chemoprophylaxis'/de OR 'infection prevention'/de 212,635 15. recur:ab,ti OR relaps:ab,ti OR prevent:ab,ti OR chemoprev:ab,ti OR repeat:ab,ti OR repetit:ab,ti OR chronic:ab,ti OR prophylax:ab,ti OR prophylac:ab,ti OR refractory:ab,ti 5,759,315 16. chronic disease'/de 206,400 17. treatment failure'/exp 182,363 18. (treatment NEAR/2 fail):ab,ti 71,300 19. #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 6,052,729 20. #11 AND #19 5,116 21. randomized controlled trial'/de 680,287 22. 'controlled clinical study'/de 435,176 23. random:ti,ab 1,709,875 24. randomization'/de 91,949 25. intermethod comparison'/de 277,719 26. placebo:ti,ab 330,912 27. compare:ti OR compared:ti OR comparison:ti 570,466 28. (evaluated:ab OR evaluate:ab OR evaluating:ab OR assessed:ab OR assess:ab) AND (compare:ab OR compared:ab OR comparing:ab OR comparison:ab) 2,373,746 29. (open NEAR/1 label):ti,aB 91,224 30. ((double OR single OR doubly OR singly) NEAR/1 (blind OR blinded OR blindly)):ti,ab 251,275 31. double blind procedure'/de 188,893 32. (parallel NEXT/1 group):ti,ab 28,194 33. crossover:ti,ab OR 'cross over':ti,ab 113,165 34. ((assign OR match OR matched OR allocation) NEAR/5 (alternate OR group* OR intervention* OR patient* OR subject* OR participant*)):ti,ab 364,722 35. assigned:ti,ab OR allocated:ti,ab 428,789 36. (controlled NEAR/7 (study OR design OR trial)):ti,ab 391,000 37. volunteer:ti,ab OR volunteers:ti,ab 262,747 38. trial:ti 346,345 39. human experiment'/de 557,732 40. #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 OR #38 OR #39 5,560,462 41. #20 AND #40 1,116 42. #20 AND #40 AND [embase]/lim AND [2020‐2021]/py 98
# of records identified	98

Search electronic report #3
Search type	Update
Databases	The Cochrane Central Register of Controlled Trials (CENTRAL)
Platform	Ovid
Search date	06/10/2021
Update date	Undefined
Range of search date	2019‐Current
Language restrictions	None
Other limits	None
Search strategy (results)	1. exp Candidiasis, Vulvovaginal/ (333) 2. exp Candida/ (444) 3. exp Candidiasis/ (958) 4. candid.tw. (20077) 5. monilia.tw. (31) 6. thrush.tw. (124) 7. or/2‐6 (20406) 8. exp Vaginal Diseases/ (1321) 9. exp Vulvar Diseases/ (668) 10. vulv.tw. (1853) 11. vagin.tw. (19025) 12. vulvovagin.tw. (786) 13. vulvo‐vagin.tw. (77) 14 genit.tw. (6036) 15. or/8‐14 (24627) 16. 7 and 15 (1355) 17. 1 or 16 (1355) 18. exp Recurrence/ (12295) 19. exp Secondary Prevention/ (3106) 20. exp Chemoprevention/ (1612) 21. (recur* or relaps* or prevent* or chemoprev* or chemoprophyla* or prophylax* or prophylac* or refractory or repeat* or repetit* or chronic*).tw. (476046) 22. exp Chronic Disease/ (13339) 23. treatment failure/ (3336) 24. (treatment$ adj2 fail$).tw. (13490) 25. or/18‐24 (489667) 26. 17 and 25 (540) 27. limit 26 to yr="2020‐Current" (51)
# of records identified	51

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Clinical recurrence at 6 months	6	607	Risk Ratio (M‐H, Random, 95% CI)	0.36 [0.21, 0.63]
1.1.1 Oral	4	518	Risk Ratio (M‐H, Random, 95% CI)	0.29 [0.14, 0.62]
1.1.2 Topical	2	89	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.26, 1.14]
1.2 Clinical recurrence at 12 months (6 months active, 6 months observation)	6	585	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.72, 0.89]
1.2.1 Oral	4	496	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.68, 0.88]
1.2.2 Topical	2	89	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.77, 1.13]
1.3 Mycological recurrence at 6 months	5	568	Risk Ratio (M‐H, Random, 95% CI)	0.44 [0.25, 0.78]
1.3.1 Oral	4	506	Risk Ratio (M‐H, Random, 95% CI)	0.37 [0.22, 0.63]
1.3.2 Topical	1	62	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.62, 1.06]
1.4 Mycological recurrence at 12 months	3	189	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.79, 1.03]
1.4.1 Oral	2	127	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.71, 1.04]
1.4.2 Topical	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.82, 1.17]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Clinical recurrence at 6 months	3	206	Risk Ratio (M‐H, Fixed, 95% CI)	1.66 [0.83, 3.31]
2.2 Clinical recurrence at 12 months (6 months active, 6 months observation)	3	206	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.71, 1.27]
2.3 Mycological recurrence at 6 months	1	44	Risk Ratio (M‐H, Fixed, 95% CI)	1.22 [0.64, 2.35]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Clinical recurrence at 6 months	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	3.32 [0.84, 13.02]
3.2 Clinical recurrence at 12 months	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	1.62 [0.73, 3.61]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Clinical recurrence at 6 months	2	68	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.08, 2.69]
4.2 Clinical recurrence at 12 months	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.47, 1.49]
4.3 Mycological recurrence at 6 months	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.14, 1.74]
4.4 Mycological recurrence at 12 months	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.53, 1.38]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Clinical recurrence at 6 months	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.64, 1.32]
6.2 Clinical recurrence at 12 months (6 months active, 6 months observation)	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.81, 1.31]
6.3 Mycological recurrence at 6 months	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.46, 1.17]
6.4 Mycological recurrence at 12 months	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.61, 1.26]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Clinical recurrence at 6 months	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.10, 2.58]
7.2 Clinical recurrence at 12 months	2	189	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.36, 0.80]
7.3 Clinical recurrence at 24 months	1	107	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.36, 0.74]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Mean number of clinical recurrences per participant per year	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
9.1.1 Vaccination versus heliotherapy	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.60 [‐1.61, 0.41]
9.1.2 Vaccination versus ciclopirox olamine	1	46	Mean Difference (IV, Fixed, 95% CI)	‐0.40 [‐1.33, 0.53]

*Study characteristics*
Methods	Randomised controlled trial, placebo controlled Blinding ‐ not clearly stated Withdrawals/loss to follow‐up: 34% (33/97) Intention‐to‐treat analysis: No
Participants	N = 97 enrolled, 64 completed study Inclusion criteria: Female outpatients aged 18 to 45 years with an acute episode of vulvovaginal candidiasis (VVC) and a history of recurrent vulvovaginal candidiasis (RVVC) Definition of RVVC: At least 4 documented episodes of Candida vaginitis in the previous 12 months Exclusion criteria Diabetes HIV Mixed vaginal infections Women who were pregnant or who planned pregnancy during the study Women who had used antifungal agents within the past 4 weeks or vaginal medication/douching in last 48 hours Setting Country: Iran Hospital outpatients department Baseline characteristics Age: 31.9 ± 7.5 (fluconazole), 31.8 ± 6.4 (placebo) Mean episodes of vulvovaginal candidiasis in previous 12 months: Not reported Hormonal contraceptive use: Not reported Candida spp. at enrolment: C albicans (78%), C glabrata (16%) Other health conditions: Not reported
Interventions	Fluconazole ‐ Oral, 150 mg, weekly, 6 months Placebo ‐ Oral, weekly, 6 months
Outcomes	Timing of measurements Assessed at: Monthly for 6 months, 9 months, 12 months (6 months active treatment, 6 months further follow‐up) Assessed if became symptomatic between planned visits: Not reported Primary outcomes Number of clinical recurrences per participant per year: ‐ Proportion of women with clinical recurrence Definition: “symptoms, as well as microscopic or cultural evidence of Candida” Results: 19% (6/32) in fluconazole group vs 50% (16/32) in placebo group at 6 months. 72% (23/32) in fluconazole group vs 81% (26/32) in placebo group at 12 months (6 months active treatment, 6 months observation) Adverse events: Side effects ‐ monitored for at monthly visits. No cessation of therapy due to side effects Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with mycological recurrence: At 6 months: 25% (8/32) in fluconazole group vs 63% (20/32) in placebo group. At 12 months: 78% (25/32) in fluconazole group vs 88% (28/32) in placebo group Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Pharmaceutical Sciences Resource Centre, Tehran University of Medical Sciences
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Sequence generation and method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Placebo used. No comment as to whether similar between treatment and placebo in appearance
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear if the assessors were blinded to the type of treatment (oral fluconazole vs placebo) received by the participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	Withdrawals/loss to follow‐up: 34%. Reasons for this not outlined. Per‐protocol analysis
Selective reporting (reporting bias)	Unclear risk	No protocol available for assessment. Side effects were monitored for during monthly visit but not reported.
Other bias	Low risk	Funding from Pharmaceutical Sciences Resource Centre, Tehran University of Medical Sciences

*Study characteristics*
Methods	Randomised controlled trial, active control Blinding ‐ not stated Withdrawals/loss to follow‐up: 0% (1/122) Intention‐to‐treat analysis: No participants lost to follow‐up or non‐compliant with treatment regimen were excluded from analyses. Participants who relapsed during the 2‐year follow‐up and who were not mycologically cured after prespecified treatment protocol were excluded; 56 of 61 participants in the K‐712 group and 51 of 61 in the itraconazole group were included in the analysis.
Participants	N = 122 enrolled, 121 completed the study Inclusion criteria: Women with overall general good health, presenting to the gynaecology outpatient clinic with a history of RVVC Definition of RVVC: 4 proven cases of VVC in past 12 months Exclusion criteria Pregnant < 6 weeks postabortion or postpartum Diabetic Immunocompromised Chronic drug therapy (e.g. steroids or antibiotics) Pelvic inflammatory disease Diagnosed or suspected genital malignancy Setting Country: Italy Gynaecology outpatients clinic Baseline characteristics Age: 22 to 63 years Mean episodes of vulvovaginal candidiasis in previous 12 months: Not reported Hormonal contraceptive use: Not reported Candida spp. at enrolment: Not reported Other health conditions: Not reported Note: No characteristics provided of individual groups
Interventions	Polygodyal phytocompound (K‐712) ‐ 1 tablet twice a day for 1 week to a month (the 100 mg composition is: 10 mg of oleoresin from Pseudowintera colorata at 30% concentration in polygodyal together with trace amounts of Olea europea) for 24 months. In case of relapse, participants received 1 tablet a day for 4 weeks. Itraconazole ‐ 200 mg once a week (administered as 100 mg tablets twice daily with meals) for 24 months. In case of relapse, participants received 200 mg every day for 7 days.
Outcomes	Timing of measurements Assessed at: Monthly for 2 years (active treatment), then 6‐month follow‐up post‐treatment Assessed if became symptomatic between planned visits: Yes Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of women with clinical recurrence Definition: Clinical symptoms, confirmed by culture Results: 16% (10/56) in K‐712 and 26% (16/51) in itraconazole group at 12 months; 36% (22/56) in K‐712 and 64% (39/51) in itraconazole group at 24 months Adverse events: Major side effect – nil in either group Transient/mild symptoms ‐ 31% (19/56, total of 25 episodes) in itraconazole group (nausea, abdominal discomfort, unpleasant taste) vs 5% (3/51, total of 5 episodes) in K‐712 group (slight dyspepsia) Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: Not reported Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Random numbers were computer generated"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding is not referred to in trial publication. Given the different dosing regimens, it is unlikely that blinding occurred or was preserved.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear if the assessors were blinded to the type of treatment received by the participants.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawals/loss to follow‐up: 0%
Selective reporting (reporting bias)	Unclear risk	No protocol located to permit judgement of low risk or high risk.
Other bias	Unclear risk	Role of industry funding unclear

*Study characteristics*
Methods	Randomised active controlled, open‐label trial Blinding ‐ No, open‐label Withdrawals/loss to follow‐up: 14% (6/44) Intention‐to‐treat analysis: Yes
Participants	N = 44 enrolled, 38 completed study Inclusion criteria: Recurrent, clinical symptomatic vaginal candidiasis, presenting with an acute episode Definition of RVVC: Clinical symptomatic vaginal candidiasis; defined as greater than 4 episodes (proven by culture) in the past year Exclusion criteria Chronic underlying illness, including diabetes mellitus Chronic use of antibiotics or steroids Known immunodeficiency Pregnancy Setting Country: Canada Women’s clinic, university teaching hospital Baseline characteristics Age (mean): 33.6 years (itraconazole group), 31.0 years (clotrimazole group) ‐ not significant Mean episodes of vulvovaginal candidiasis in previous 12 months: 8.8 (itraconazole group), 9.7 (clotrimazole group) ‐ not significant Hormonal contraceptive use: Oral contraceptive pill 23% (itraconazole group), 32% (clotrimazole group) ‐ not significant Candida spp. at enrolment: Not reported Other health conditions: Not reported
Interventions	Itraconazole ‐ 200 mg orally daily (administered as 100 mg twice daily with meals) for 5 days, then 200 mg twice weekly for 6 months Clotrimazole ‐ 200 mg, intravaginally daily for 5 days, then 200 mg twice weekly for 6 months
Outcomes	Timing of measurements Assessed at: Monthly for 12 months (6 months active treatment, 6 months follow‐up) Assessed if became symptomatic between planned visits: Not reported Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of women with clinical recurrence (clinical symptomatic vaginal candidiasis, including failure to respond at all to treatment) Definition: Positive culture for C albicans with signs or symptoms of candidiasis. Symptoms and signs included excessive vaginal discharge, itching, burning, dysuria or dyspareunia, associated with clinical findings of white or creamy vaginal discharge in excessive amount, with or without redness of vulva or vagina. Results: 36% (8/22) in itraconazole group vs 23% (5/22) in clotrimazole group at 6 months; 50% (11/22) in itraconazole group vs 73% (16/22) in clotrimazole group at 12 months (6 months active, 6 months observation) Per‐procotol analysis presented in paper, but intention‐to‐treat numbers inferred. Adverse events: Side effects ‐ 33% (7/21) in itraconazole group vs 0% (0/22) in clotrimazole group at 6 months. No side effects resulted in discontinuation of study medication due to adverse events. Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: 50% (11/22) in itraconazole group vs 41% (9/22) in clotrimazole group at 6 months Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Janssen Pharmaceutica Inc, Canada
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The random numbers were computer generated.
Allocation concealment (selection bias)	Low risk	Numbered, sealed envelopes containing the therapeutic regimen were kept in the Pharmacy Department, ensuring that the investigator remained blinded to the randomisation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Study is open‐label.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome assessment uses culture and clinical symptoms; participants not blinded, lab assessment not described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawals/loss to follow‐up: 14% (6/44). 5 participants dropped out in the clotrimazole group (2 non‐compliance, 3 lost to follow‐up), compared to 1 in the itraconazole group (desired to be pregnant). Both per‐protocol and intention‐to‐treat analysis conducted.
Selective reporting (reporting bias)	Unclear risk	No protocol located.
Other bias	Unclear risk	Funded by Janssen Pharmaceuticals

*Study characteristics*
Methods	Randomised controlled trial, no treatment as control Single‐blind Withdrawal/loss to follow‐up: 7% (4/58)
Participants	N = 58 enrolled, 54 completed study Inclusion criteria: Women with RVVC and their male sexual partners, who were in stable, monogamous relationships Definition of RVVC: 4 or more episodes of candidiasis (proven by culture) in the year preceding the study Exclusion criteria Diabetes mellitus Chronic antibiotic or steroid use Chronic illnesses Setting Country: Canada Outpatient clinic Baseline characteristics Age: 32.0 (SD 7.0) in ketoconazole group, 31.1 (SD 5.4) in no‐treatment group, P = 0.06 Mean episodes of vulvovaginal candidiasis in previous 12 months: 8.4 (SD 3.36) in ketoconazole group, 9.5 (SD 3.28) in no‐treatment group, P = 0.5 Hormonal contraceptive use: Not reported Candida spp. at enrolment: Not reported Other health conditions: Not reported
Interventions	Ketoconazole ‐ 200 mg daily for 5 days for male partners treated at enrolment No treatment
Outcomes	Timing of measurements Assessed: Monthly for 12 months (5 days active treatment, then 12 months of follow‐up) Assessed if became symptomatic between planned visits: Yes Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of women with clinical recurrence Definition: Clinical recurrence of Candida vaginitis with confirmation by smear or culture Results: 65% (17/26) in ketoconazole group vs 71% (20/28) in no‐treatment group at 6 months; 85% (22/26) in ketoconazole group vs 82% (23/28) in no‐treatment group at 12 months Adverse events: In the untreated group, 1 participant was removed from the study due to an adverse reaction. Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: Not reported Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Janssen Pharmaceuticals, Canada
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method unclear
Allocation concealment (selection bias)	High risk	"Numbered envelopes contained the specific regimens". As the envelopes contained the treatment or no tablets, it is likely that the allocation may not have been concealed.
Blinding of participants and personnel (performance bias) All outcomes	High risk	"Single blind" study, assumed that the partners who received the treatment were blinded, but treating doctors were not. No treatment as comparator.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcome assessed by lab culture plus clinical symptoms; not specifically stated that assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawal/loss to follow‐up: 7% (4/58). In the treated group, 2 participants dropped out of the study due to lack of interest. In the untreated group, 1 participant was removed from the study due to an adverse reaction and 1 dropped out due to disinterest.
Selective reporting (reporting bias)	Unclear risk	No protocol located to permit a judgement of low risk or high risk.
Other bias	Unclear risk	Funded by Janssen Pharmaceuticals

*Study characteristics*
Methods	Randomised controlled trial Blinding ‐ not stated Withdrawals/loss to follow‐up: 10% (8/82) Intention‐to‐treat analysis: Unclear ‐ “Only patients adhering to the protocol were analysed for efficacy; patients lost to follow‐up or non‐compliant with the treatment regimen were excluded from analysis". Later in text stated that "patients lost to follow‐up or who had recurrences due to non‐compliance were assessed as failures" (p547).
Participants	N = 109 eligible, 82 enrolled, 76 analysed (the text describes 4 dropouts/exclusions per group, but the analysis describes 38 participants per group, which would equate to 3 dropouts/exclusions per group) Inclusion criteria: Women with overall general good health, presenting to the gynaecology outpatient clinic with abnormal vaginal discharge (confirmed on speculum examination) and with a history of at least 4 proven episodes of VVC in the prior 12 months Definition of RVVC: 4 proven cases of VVC in past 12 months Exclusion criteria Pregnant < 6 weeks postabortal or postpartum Diabetic Immunocompromised Chronic drug therapy (e.g. steroids or antibiotics) Pelvic inflammatory disease Diagnosed or suspected genital malignancy Setting Country: Italy (presumed) Gynaecology outpatients clinic Baseline characteristics Age: 19 to 61 years Mean episodes of vulvovaginal candidiasis in previous 12 months: Not reported Hormonal contraceptive use: Not reported Candida spp. at enrolment: Positive culture not required at enrolment, but if positive ‐ C albicans (54%), C glabrata (28%), C krusei (8%). Other health conditions: Not reported
Interventions	Polygodyal/anethole phytocompound (K‐712) ‐ the 100 mg composition is: 10 mg of oleoresin from Pseudowintera colorata at 30% concentration in polygodyal together with trace amounts of Olea europea. Assumed to be oral and self‐administered. 1 tablet twice a day for 4 weeks and then for the first 2 weeks each month for a total of 6 months. Itraconazole ‐ 100 mg twice daily with meals for 4 days, then 200 mg once weekly for 6 months
Outcomes	Timing of measurements Assessed at: Monthly for 6 months, then at 9 and 12 months (6 months active treatment, 6 months further follow‐up) Assessed if became symptomatic between planned visits: Not reported Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of women with clinical recurrence: Definition: Not clearly defined in text. Contacted author to request clarification, but no additional information received. Results: 5% (2/41) in K‐712 group vs 10% (4/41) in itraconazole group at 6 months (these data were extracted from a graphical figure in the text). 34% (14/41) in K‐712 group vs 66% (27/41) in itraconazole group at 12 months (6 months active treatment, 6 months observation) Adverse events: 0% in K‐712 group compared to 10% (4/41) in itraconazole group (nausea, abdominal discomfort, unpleasant taste) Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: 22% (9/41) in K‐712 group vs 17% (7/41) in itraconazole group at 6 months. 34.2% (14/41) in K‐712 group vs 65.7% (27/41) in itraconazole group at 12 months (6 months active treatment, 6 months observation) Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Not stated
Notes	Author contacted, not able to provide data in format suitable for this review. Polygodyal/anethole phytocompound (K‐712) was supplied by Canova Foundation, Lesmo, Italy.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Random numbers were computer generated"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding is not referred to in the trial publication. Given the different dosing regimens, it is unlikely that blinding occurred or was preserved.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear if the assessors were blinded to the type of treatment received by the participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	It is unclear why 82 participants were randomised from 109 eligible participants. Dropouts and exclusions are described, but these do not align with the figures provided (e.g. the text describes 4 per arm of the trial but figures reported show 3 per arm).
Selective reporting (reporting bias)	High risk	Outcome measures are not described in detail, such that it is unclear what was being measured and reported as a primary outcome measure and in the graphical representations of data. A protocol was not published prior to reporting of the outcomes.
Other bias	Unclear risk	Role of industry funding is not clear.

*Study characteristics*
Methods	Randomised controlled trial Withdrawals/loss to follow‐up: not reported Intention‐to‐treat analysis: Not reported
Participants	N = 68 Inclusion criteria: Meeting diagnostic criteria of RVVC, with the chief complaints of genital and/or vaginal itching, burning sensation, dyspareunia, and increased vaginal secretions Patients with vulvovaginal congestion, oedema or rupture, soybean curd‐like or curd‐like secretions covered on vaginal mucosa and cervical surface, and red and swollen mucosa or mucosal erosion exposed after being wiped Blastospore and pseudohypha could be seen in secretions via 10% potassium hydroxide solution wet film method Gram staining microscopy Patients with the course of disease for more than 1 year, and disease recurrence at least 4 times within 1 year Patients who had not taken other antibacterial drugs within half a month before treatment and who had received no systemic anti‐infective and antifungal treatment during the test Patients who could adhere to the therapeutic regimen and follow‐up on time Definition of RVVC: Mycologically confirmed symptomatic vulvovaginal candidiasis (VVC) 4 times or more within 1 year Exclusion criteria Pregnant women Patients with severe heart, liver, kidney, or blood disease Patients who were allergic to drugs used in this study Patients who breached the therapeutic regimen and still had sexual life during treatment Patients who were unwilling to co‐operate in the follow‐up Setting Country: China Hospital setting – Gynecological Clinic of Women and Children's Health Care Hospital of Linyi Baseline characteristics Age: mean age: 35.25 years (Standard Deviation (SD) 10.12) Mean episodes of vulvovaginal candidiasis in previous 12 months: Not reported Oral contraceptive pill: Not reported
Interventions	Intervention group Adenosine triphosphate (ATP)‐infrared bio‐effect treatment nursing intervention (questionnaire, assessment, inquiry, propaganda and education, individualised nursing, follow‐up) 500,000 units of nystatin tablets and 0.2 g metronidazole tablets for 7 days followed by consolidation therapy for 2 courses of treatment according to the menstrual cycle Control group Routine nursing care 500,000 units of nystatin tablets and 0.2 g metronidazole tablets for 7 days followed by consolidation therapy for 2 courses of treatment according to the menstrual cycle
Outcomes	Timing of measurements: 1 month, 3 months, and 6 months after all courses of treatment Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of participants with at least one clinical recurrence during the treatment and follow‐up period: Not reported Adverse events: Not reported Secondary outcomes: Not reported
Funding	Not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Sequence generation and method of randomisation not described
Allocation concealment (selection bias)	High risk	Allocation concealment not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	The study was not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear if the assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	Withdrawal was not reported but “patients who could adhere to the therapeutic regimen and follow‐up on time” was an inclusion criterion.
Selective reporting (reporting bias)	Unclear risk	No protocol located to permit judgement of low risk or high risk.
Other bias	Low risk	No grant was received for this study; authors declare no conflict of interest.

*Study characteristics*
Methods	Randomised controlled trial, not placebo controlled Blinding ‐ No Withdrawals/loss to follow‐up: 24% (22/90) Intention‐to‐treat analysis: Yes
Participants	N = 90 randomised, divided into 3 groups Withdrawals: 5 participants in vaccine group, 8 in heliotherapy group, 9 in ciclopirox olamine group excluded due to withdrawal (total 22) Inclusion criteria: Women with RVVC presenting to a gynaecology or family planning clinic in Berlin, Germany Definition of RVVC: More than 4 episodes of infection per year; diagnosis confirmed by a positive culture at time of inclusion or (if not a current acute infection) a positive culture within previous 3 months Exclusion criteria Severe illness Diabetes mellitus Antibiotic treatment Immunosuppressant or immunostimulant treatment Non‐co‐operative patients Pregnant or breastfeeding women Alcohol or drug abuse Visiting a solarium or taking photosensitising medication Setting Country: Germany (Berlin) Centre – A gynaecology and family planning clinic Baseline characteristics Age: 18 to 67 years Mean episodes of vulvovaginal candidiasis in previous 12 months: 7.7 in vaccine group, 7.1 in heliotherapy group, 6.5 in ciclopirox olamine group Hormonal contraceptive use: Oral contraceptive pill ‐ all premenopausal women, and 6 menopausal women taking hormone replacement therapy (HRT) Candida spp. at enrolment: Not reported Other health conditions: Not reported
Interventions	Lactobacillus vaccination ‐ 7 x 10⁹ inactivated lactobacilli of 8 species, administered 3 times intramuscular (IM) every 2 weeks Heliotherapy ‐ according to skin type with sub‐erythematous doses of ultraviolet A and B rays, administered 3 times per week for 6 weeks Ciclopyroxilamine ‐ 100 mg vaginal tablets daily for 6 days
Outcomes	Timing of measurements Assessed at baseline, 6 to 12 weeks later, and 1 year later Assessed if became symptomatic between planned visits: Yes (but not clearly mentioned in methods section; in case of recurrence during the study women received 150 mg fluconazole) Also questionnaire developed by researchers assessing sexuality, partner relationship, doctor‐patient relationship, and use of antimycotics Primary outcomes Mean number of clinical recurrences per participant per year(Standard Deviation) vaccine group: 1.4 (1.6) heliotherapy group: 2.0 (1.9) ciclopirox olamine group: 1.8 (1.6) Proportion of women with clinical recurrence: Not reported Adverse events: 2 participants in the vaccine group reported flu‐like symptoms, 1 of whom withdrew from the study. Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: Not reported Duration of symptoms after treatment initiation: Not reported Patient preference: Assessed by percentage of participants satisfied with the treatment vaccine group: 47.7% heliotherapy group: 26.7% ciclopirox olamine group: 20%
Funding	Dr Kern GmbH provided the heliotherapy equipment and funding for laboratory tests and statistical analysis (the latter performed at the Institute of Natural Medicine). Strathmann GmbH provided the vaccine. Taurus Pharma provided the ciclopirox olamine vaginal capsules. Pfizer Pharma provided the fluconazole capsules.
Notes	Publication in German
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The title suggests the study is randomised, but the method of sequence generation is not described anywhere in the paper.
Allocation concealment (selection bias)	High risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	The study is not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Primary outcome includes symptoms as well as lab tests; participants were not blinded and it is not clear if the assessors of lab tests were blinded to the type of treatment received by the participants.
Incomplete outcome data (attrition bias) All outcomes	High risk	22 of 90 randomised participants withdrew before completion of the study at 1‐year follow‐up. The authors performed intention‐to‐treat (ITT) analysis to take this into account.
Selective reporting (reporting bias)	Unclear risk	All outcomes in methods section are reported. However, a protocol was not published prior to reporting of the outcomes.
Other bias	Unclear risk	Industry funding, but the role of industry is not clear. The first author reports receiving honoraria for presentations from Strathmann and Taurus Pharma. The second author does not have any interests to declare.

PERMALINK

Treatment for recurrent vulvovaginal candidiasis (thrush)

Georga Cooke

Cathy Watson

Laura Deckx

Marie Pirotta

Jane Smith

Mieke L Driel

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Random sequence generation (selection bias)

Allocation concealment (selection bias)

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting bias

Other bias

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Results

Description of studies

Results of the search

1.

Included studies

Participants and setting

Interventions and comparators

Outcome measurement

Reporting of primary outcomes

Reporting of secondary outcomes

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Random sequence generation

Allocation concealment

Blinding

Performance bias

Detection bias

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings 1. Drug treatment compared to placebo or no treatment for recurrent vulvovaginal candidiasis.

*Study characteristics*
Methods	Randomised controlled trial, short duration versus long duration of the same treatment Blinding – Not described Withdrawals/loss to follow‐up: Not clearly described, but likely that 1 participant was lost to follow‐up, 4% (1/26) Intention‐to‐treat analysis: Not described
Participants	N = 26 participants with RVVC enrolled, 26 completed. Trial also enrolled 34 participants with acute RVVC. Inclusion criteria: Diagnosis of RVVC confirmed by a positive culture at enrolment. All participants with RVVC had the primary idiopathic form with no known causes. Definition of RVVC: a history of 4 or more culture‐documented attacks within the last year (or at least 3 episodes unrelated to antibiotic therapy) Exclusion criteria Pregnancy Patients with immunoalteration Patients taking medications with a known significant interaction with itraconazole Setting Country: Czech Republic Outpatient department of a gynaecology clinic Baseline characteristics (participants with RVVC only) Age mean: 28.7 years, 7.17 SD, range 18 to 44 Mean episodes of vulvovaginal candidiasis in previous 12 months: Not reported Hormonal contraceptive use: 46% Candida spp. at enrolment: C albicans (82%); C glabrata (3%); other (10%) Other health conditions: Not reported
Interventions	1‐day regimen ‐ itraconazole (oral), 400 mg total, divided into 2 doses of 200 mg on the same day 3‐day regimen ‐ itraconazole (oral), 600 mg total, divided into 3 doses of 200 mg on 3 consecutive days
Outcomes	Timing of measurements Assessed at: 14 days, 1 month, 3 months, and 6 months (mentioned in results not in methods) (1‐ or 3‐day active treatment, followed by 6 months of follow‐up) Assessed if became symptomatic between planned visits: Not reported Primary outcome Number of clinical recurrences per participant per year: Not reported Proportion of participants with at least one clinical recurrence during the treatment and follow‐up period Definition: Not clear Results: After 6 months, 62% (8/13) in 1‐day group versus 77% (10/13) in 3‐day group Adverse events: 1 woman with reversible alopecia areata and 1 woman with mild nausea with diarrhoea, but it is unclear if these women had RVVC or acute VCC or to which group they had been allocated Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: Not reported Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	The study was supported by the research project IGA No. 3694‐3 of the Czech Ministry of Health.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Sequence generation and method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Given the nature of the treatment (1 vs 3 days) and absence of a double‐dummy design, blinding of participants was not possible.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not clear if the assessors were blinded to the treatment regimen (1‐ vs 3‐day itraconazole course) received by the participants.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Withdrawals/loss to follow‐up: not clearly described
Selective reporting (reporting bias)	Unclear risk	No protocol located to permit a judgement of low risk or high risk.
Other bias	Low risk	The study was supported by the research project IGA No. 3694‐3 of the Czech Ministry of Health.

*Study characteristics*
Methods	Randomised controlled trial, no‐treatment control Blinding ‐ Participants not blinded, both nurses and physicians were "unaware of the study", microbiology staff ‐ unclear Withdrawals/loss to follow‐up: 7% (8/114) Intention‐to‐treat analysis: Per‐protocol analysis presented in paper
Participants	N = 114 enrolled, 106 completed study Inclusion criteria: Patients with 4+ documented episodes of vulvovaginal candidiasis, presenting with an acute episode of vaginitis Definition of RVVC: At least 4 documented episodes (positive vaginal culture) of symptomatic vulvovaginal candidiasis in the 12 months preceding enrolment Exclusion criteria Pregnancy Diabetes HIV infection < 18 years or older than 50 years Mixed infection Non‐albicans infection Intended pregnancy Setting Country: Italy Outpatient clinic Baseline characteristics Age mean: 30.2 years (itraconazole), 30.9 (no treatment) Mean episodes of vulvovaginal candidiasis in previous 12 month: Not reported Hormonal contraceptive use: Not reported Candida spp. at enrolment: C albicans 100% (required for enrolment) Other health conditions: Not reported
Interventions	Itraconazole ‐ 400 mg monthly (200 mg, 12 hours apart), during fourth or fifth day of menstrual cycle for 6 months No‐treatment control
Outcomes	Timing of measurements Assessed at: 3, 6, and 12 months (6 months active treatment, 6 months further follow‐up) Assessed if became symptomatic between planned visits: Yes ("If symptom developed between visits, patients were examined within 48 hours.") Primary outcomes Number of clinical recurrences per participant per year: Not reported Proportion of women with clinical recurrence Definition: Clinical symptoms, confirmed by culture Results: 36% (20/55) in itraconazole group vs 64% (34/53) in no‐treatment group at 6 months. 61% (33/54) in itraconazole group vs 71% (37/52) in no‐treatment group at 12 months (6 months active treatment, 6 months observation) Adverse events: Not reported Secondary outcomes Time to first recurrence: Not reported Number of symptomatic days per year: Not reported Number of mycological recurrences per participant per year: Not reported Proportion of participants with at least one mycological recurrence during the treatment and follow‐up period: 42% (23/55) in itraconazole group vs 66% (35/53) in no‐treatment group at 6 months Duration of symptoms after treatment initiation: Not reported Patient preference: Not reported
Funding	Not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Random numbers generated by a personal computer"
Allocation concealment (selection bias)	Unclear risk	"Sealed envelope method", but no comment as to whether envelopes were opaque
Blinding of participants and personnel (performance bias) All outcomes	High risk	"Both nurses and physicians who carried out the visits were unaware of the study being performed." However, women were not blinded as there was a 'no treatment' arm of the trial rather than a placebo.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome includes symptoms and lab tests. It is not clear if the assessors of lab tests were blinded to the type of treatment received by the participants.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Withdrawals/loss to follow‐up: 7% (8/114) Intention‐to‐treat analysis: Per‐protocol analysis presented in paper
Selective reporting (reporting bias)	Unclear risk	No protocol was available for assessment.
Other bias	Unclear risk	Funding not reported

Study	Reason for exclusion
Avijgan 2012	Diagnostic criteria for recurrent vulvovaginal candidiasis (RVVC) did not meet inclusion criteria ‐ RVVC defined as 3 or more attacks of vulvovaginal candidiasis (VVC) in 12 months.
Azima 2018	Diagnostic criteria for RVVC did not meet inclusion criteria (number of previous episodes not documented).
Bartz 2000	Diagnostic criteria for RVVC did not meet inclusion criteria ("two or more episodes of clinician‐identified vaginitis in a 6 month timeframe").
Brand 2018	Diagnostic criteria for RVVC did not meet inclusion criteria ("Subject eligibility required a documented history of RVVC (defined as three episodes of acute VVC in the past 12 months, which included the episode at the screening visit").
Bushell 1988a	Diagnostic criteria for RVVC did not meet inclusion criteria ("at least three documented episodes of vaginal candidosis").
Bushell 1998b	Publication very brief. Diagnostic criteria for RVVC not provided. Unclear if this is a randomised study
Coric 2006	Additional information requested from author but not received.
Davidson 1978	Diagnostic criteria for RVVC did not meet inclusion criteria ("a history of recurrent 'thrush' for at least one year treated elsewhere each time with anti‐fungal therapy, and confirmed where possible, or had had at least three proved treated clinical attacks of candidosis in the previous six months in the clinic").
Diba 2010	Additional information requested from author but not received.
Donders 2008	Not a randomised controlled trial
Edwards 2018	Diagnostic criteria for RVVC did not meet inclusion criteria ("at least two episodes of vaginitis during the12 months prior to the current episode").
Fan 2005	Unable to separate findings for participants who had recurrent vulvovaginal candidiasis from those who just had acute candidiasis. Authors contacted but no additional information provided.
Fong 1993	Not an original randomised controlled trial (RCT). This study uses data of previously published randomised trials to study in vitro resistance of Candida albicans to the imidazoles.
Golero 1993	Diagnostic criteria for RVVC did not meet inclusion criteria ("patients with chronic vaginal candidosis, unresponsive to previous topical antimycotic treatments, were included in this study").
Hilton 1992	Not reported as an RCT ‐ after 8 months trial stopped randomising participants to different arms. The data reported in the publication include participants enrolled both before and after this change in protocol. Author contacted but no additional information provided.
Houang 1989	Inadequate information in publication to extract data. Author contacted but no additional information provided.
Miller 1984	Unable to separate findings for participants who had recurrent vulvovaginal candidiasis from those who just had acute candidiasis. Authors contacted but no additional information provided.
Rashid 1991	Not a randomised controlled trial
Shalev 1996	Unable to separate findings for participants who had recurrent vulvovaginal candidiasis from those who just had acute candidiasis. Authors contacted but no additional information provided.
Silverman 1971	Diagnostic criteria for RVVC did not meet inclusion criteria ("recent (two years) history of two or more attacks of candidal vaginitis").
Sobel 1994	Diagnostic criteria for RVVC did not meet inclusion criteria ("three or more episodes of acute candidal vaginitis in the 12 months prior").
Sobel 2001	Diagnostic criteria for RVVC did not meet inclusion criteria ("three or more episodes of acute candidal vaginitis in the 12 months prior").
Vladareanu 2018	Follow‐up for less than 6 months (90 days)
Yang 2000	Diagnostic criteria for RVVC did not meet inclusion criteria ("three or more episodes of acute candidal vaginitis in the 12 months prior").

Methods	Interventional study
Participants	Inclusion criteria: Female 18 years and older Premenopausal Chronic vulvovaginal candidiasis as per diagnostic criteria
Interventions	Fluconazole ‐ 50 mg orally daily for 12 weeks Fluconazole ‐ 150 mg orally weekly for 12 weeks
Outcomes	Presence of chronic vulvovaginal candidiasis symptoms at 6 weeks, 12 weeks, and 12 months
Notes	In trial database record, diagnostic criteria for this study do not clearly require pathological confirmation of diagnosis and may not meet the inclusion criteria for this review (4+ episodes of vulvovaginal candidiasis (VVC) within the past 12 months). Trial database record states that the trial was "stopped early" after having recruited 75 participants (target sample size 200 participants). We emailed the primary contact but have received no reply.

Methods	Interventional study ‐ randomised parallel controlled trial
Participants	Inclusion criteria: Patients whose VVC is cured Patients whose vaginal discharge shows negative findings by fungal microscopy Patients aged 18 years old and above
Interventions	"JUC Spray Dressing" Miconazole nitrate
Outcomes	Symptoms and signs Fungal microscopy
Notes	Unlikely to meet inclusion criteria for this review as may only be enrolling participants with acute VVC. Trial database record states that the trial is still recruiting, but the last update of the database record was on 27 April 2015. We emailed the primary contact but have received no reply.

Methods	Interventional ‐ randomised controlled trial
Participants	Patients with histologically/cytologically confirmed cervical intraepithelial neoplasia grade 1 or higher Patients with chronic, recurrent vaginal candidiasis with at least 4 infection episodes in the last year and microbiologically confirmed candidal infection at inclusion in the study
Interventions	4 phenyl‐butyrate prolonged release tablet
Outcomes	Measurement of vaginal interleukins Assessment of vaginal candidiasis
Notes	Database entry in German. Trial status in database is "prematurely ended". It is not clear why this occurred, nor is it clear if enrolment of participants had commenced when the study was ceased. No contact details for primary investigator included in record.

Methods	Interventional ‐ randomised trial. Record states that study was randomised, but not controlled.
Participants	Participants were women with vaginal infections by Candida albicans and non‐albicans Candida spp relapsing recurrent (≥ 4 events/year) in the acute phase of the disease.
Interventions	Clotrimazole and metronidazole
Outcomes	Reduction in recurrence
Notes	The trial database stated the recruitment status of this trial is ongoing. The record in the database was created on 8 March 2012. However, the email address of the primary contact was not current and no further details could be located.

Methods	Interventional ‐ randomised placebo‐controlled trial
Participants	Patients with recurrent vulvovaginal candidiasis, defined by the existence of at least 4 episodes of VVC during the previous year including the episode that was the subject of the screening visit (in addition to the current episode, at least 1 episode that occurred in the previous 2 years should also be documented by mycological examination)
Interventions	Lcr Regenerans Placebo
Outcomes	Rate of clinical recurrence bacteriologically confirmed by mycological tests occurring in the treatment period or in the 5 months following discontinuation of the study treatment Time to first recurrence Number of clinical recurrences per year Safety
Notes	Trial database record states that the trial concluded on 22 February 2017. No publication has been located. We emailed the primary contact but have received no reply.

Methods	Interventional ‐ randomised, placebo‐controlled trial
Participants	Premenopausal women suffering from acute or chronic yeast vaginitis and a history of 3 or 4 episodes over the past 12 months were eligible to participate.
Interventions	1 oral dose of fluconazole + capsules of Lactobacillus rhamnosus GR‐1 and L reuteri RC‐14 containing 5000 million live organisms for 3 months 1 oral dose of fluconazole + oral placebo capsules for 3 months
Outcomes	Vaginal symptoms and swabs It is not clear if the outcomes were collected for 3 months or 6 months (active treatment for 3 months, plus a further 3 months follow‐up).
Notes	The trial database states the recruitment status of this trial is unknown. The database record was created on 30 May 2007, and no update has been provided since then. We emailed the primary contact but have received no reply.

Methods	Interventional ‐ randomised trial
Participants	Participants were women (16 to 65 years) with RVVC (4 or more episodes of VVC during the year prior to the survey). Participants were all symptomatic at time of enrolment with microbiological evidence of Candida albicans infection.
Interventions	Lactibiane candisis 5M 2 gelatin capsules per day for 2 months then 1 gelatin capsule per day for 4 months
Outcomes	Outcomes not clear but measured to 9 months.
Notes	The database record for this trial was created on 8 June 2009 and last updated on 6 October 2011. No publication has been located for this trial. The contact details in the database were insufficient to locate a current contact for the primary investigator. The database states that the trial has been terminated.