Zinc-based therapies for prevention and treatment of vulvovaginal candidiasis: a systematic review and meta-analysis

Abstract

Background

Zinc is essential for epithelial integrity and antifungal immunity and may influence host–pathogen interactions in vulvovaginal candidiasis (VVC). However, the relationship between zinc status, zinc-based interventions, and VVC pathogenisis remains unclear. We aimed to assess the clinical efficacy of zinc-based interventions for recurrence prevention and to examine their effects on Candida virulence and inflammatory pathways.

Methods

We systematically searched the MEDLINE, Scopus, the Cochrane Central Register, and trial registries from inception to September 2025. We included observational, interventional, and experimental studies that evaluated zinc status or zinc-based therapies in VVC. Random-effects meta-analyses were used to estimate the pooled standardized mean differences (SMD) for zinc concentrations with subgroup analyses by sample type and pregnancy status. Risk ratio and risk difference for recurrence outcomes were calculated using Fisher’s exact test.

Results

Our search identified 45 records and eleven studies (six observational, two randomised controlled trials (RCTs), and three quasi-experimental) met the inclusion criteria. Six plasma/serum comparisons from observational studies were meta-analyzed. Women with VVC had significantly lower systemic zinc levels than controls (pooled SMD = − 0.72, 95% CI: −1.15 to − 0.28, p = 0.001; I²=73.5%). Pregnancy status was a significant moderator (X² = 9.14, df = 1, p = 0.003), with a much larger effect in pregnant women (SMD − 2.19, 95% CI -3.19 to – 1.18) than in non-pregnant women (SMD − 0.55, 95% CI -0.89 to − 0.21). One RCT (n = 68) showed that zinc supplementation produced a numerically lower 90-day reinfection rate that did not reach statistical significance (28.6% vs. 48.5%, RR = 0.59, 95% CI: 0.31–1.11, p = 0.134). One RCT (n = 192) reported improved pruritus with adjunctive zinc-containing therapy (P < 0.005). Experimental studies demonstrated that zinc restriction upregulates PRA1 expression, enhancing neutrophil-mediated inflammation, while zinc supplementation or zinc oxide nanoparticles downregulate virulence factors (SAP1-3) and attenuate inflammation without reducing fungal burden.

Conclusions

Zinc deficiency is associated with VVC, with a larger effect in pregnant women likely reflecting altered zinc physiology in pregnancy. Lower zinc levels in VVC may reflect a host nutritional immunity rather than a pre-existing deficiency, and causality cannot be established from the available cross-sectional evidence. Clinical trials remain insufficient to support zinc-based therapies for routine use in VVC.

Trial registration

Not applicable.

Registration

The protocol was prospectively registered in PROSPERO (CRD420251152091).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-13520-2.

Introduction

Vulvovaginal candidiasis (VVC) is one of the most common mucosal fungal infections, affecting approximately 75% of women of reproductive age at least once in their lifetime, with an estimated 138 million women worldwide suffering from recurrent VVC (RVVC) (defined as three or more episodes per year) [1, 2]. The symptoms of VVC, including itching, burning, discharge, and dyspareunia, result from hyperinflammatory responses to Candida species rather than fungal colonization alone, which can occur asymptomatically [3]. Current first-line treatments with azole antifungals achieve clinical cure in 80–90% of acute VVC cases, but recurrence rates remain problematically high, with 40–50% of women experiencing reinfection within six months of completing maintenance therapy [2]. Moreover, the emergence of azole resistance in clinical isolates underscores the urgent need for novel therapeutic approaches [3].

Zinc has been recognized as an essential nutritional element for over a century [4, 5]. In the context of VVC, zinc plays multifaceted roles in host immunity, fungal biology, and pathogen-host interactions. Zinc deficiency has been implicated in increased susceptibility to RVVC, with early studies from 1986 demonstrating significantly lower plasma zinc levels in affected women than in controls [6]. However, subsequent investigations have yielded inconsistent findings, with some failing to replicate these observations [7]. In addition to nutritional immunity, zinc critically influences the growth and morphogenesis of Candida albicans. Cultures enter the stationary phase under both zinc-depleted and zinc-excess conditions, with evidence suggesting that growth cessation is specifically linked to zinc availability rather than general nutrient depletion [8, 9]. A recent study demonstrated that zinc restriction in the vaginal environment induced the expression of C. albicans PRA1 (pH-regulated antigen), a zinc-scavenging protein implicated in IL-8-mediated neutrophil infiltration, a hallmark pathology of symptomatic VVC [10]. These findings suggest that zinc supplementation may prevent symptomatic disease not through fungicidal activity but by modulating specific fungal virulence factors, representing a potentially different therapeutic approach from conventional antifungal therapy.

Despite accumulating evidence suggesting the therapeutic potential of zinc, the literature remains fragmented, preventing clinical translation. Critical gaps persist including the consistency of zinc deficiency across VVC populations being unknown, the clinical efficacy of zinc interventions remaining uncertain, and recent mechanistic discoveries identifying PRA1 suppression as a therapeutic mechanism of zinc having not been integrated with clinical outcomes. Previous reviews have broadly discussed the immunological importance of zinc but have not systematically evaluated zinc-based therapies for VVC or integrated observational, interventional, and mechanistic evidence.

This systematic review aimed to quantify the differences in zinc levels between women with VVC and healthy controls, assess the clinical efficacy of zinc-based interventions for recurrence prevention, and examine the effects of zinc on fungal virulence and inflammatory pathways.

Methods

Study design and protocol registration

This was a systematic review and meta-analysis. The protocol was prospectively registered in PROSPERO (CRD420251152091) on September 22, 2025, and the review was conducted according to the PRISMA guidelines. The full protocol is available at https://www.crd.york.ac.uk/PROSPERO/view/CRD420251152091.

Eligibility criteria

Randomized controlled trials, non-randomized comparative studies, prospective cohort studies, and cross-sectional studies were included. We excluded reviews, commentaries, and editorial articles. We included studies investigating women of reproductive age with symptomatic VVC or RVVC, defined as three or more symptomatic episodes within 12 months. We noted that some studies used a threshold of four or more episodes. We included these studies and assessed the impact of this variation in heterogeneity analyses. We included pregnant and non-pregnant women. The interventions/exposures of interest were any zinc-containing formulations for treatment or prevention of VVC including oral zinc supplementation, topical zinc gel, zinc-containing vaginal hydrogel, zinc suppository, or zinc pessary, at any dose or duration. For observational studies, exposure of interest were zinc concentrations measured in biological samples from women with VVC.

We included studies with comparators, including placebo, vehicle control, standard antifungal therapy (e.g., azoles, fluconazole), no treatment, or healthy controls (case-control studies). The primary outcomes of interest were: (1) symptom relief (itching, burning, discharge, pain, and dryness assessed using validated scales or patient reports), (2) recurrence rate (reinfection episodes and time to the next episode), and (3) zinc levels in the biological samples (serum, plasma, erythrocytes, and cervicovaginal fluid). Secondary outcomes included vaginal biomarkers (pH, IL-1β, IL-8, PRA1 expression, and cervicovaginal zinc levels), fungal virulence markers, adverse events, and quality of life. We did not apply restrictions on the publication year, language, geographical region, or publication status.

Information sources

We searched the following databases from inception to September 20, 2025: MEDLINE (via PubMed), Scopus, and the Cochrane Central Register of Controlled Trials (CENTRAL). We also searched trial registries, including ClinicalTrials.gov, the EU Clinical Trials Register, and the WHO International Clinical Trials Registry Platform (ICTRP). Grey literature was searched using Google Scholar (first 200 results). We performed forward citation searching (identifying articles citing the included studies) and backward citation searching (reviewing the reference lists of the included studies) to identify additional relevant studies. We did not contact the study authors for unpublished data.

Search strategy

PubMed Search Strategy (conducted September 20, 2025): (“Zinc” [MeSH Terms] OR “Zinc” [Title/Abstract] OR “zinc based” [Title/Abstract]) AND (“candidiasis, vulvovaginal” [MeSH Terms] OR “candidiasis vulvovaginal” [Title/Abstract] OR “vulvovaginal candidiasis” [Title/Abstract] OR “recurrent vulvovaginal candidiasis” [Title/Abstract] OR “RVVC” [Title/Abstract]). No language, date, or publication-type filters were applied. Complete search strategies for all databases are available at https://www.crd.york.ac.uk/PROSPEROFILES/c801edbf532ac7ae5d32bc602b417a46.pdf.

Review of studies

Two reviewers (BJ and BN) independently screened the titles and abstracts of all retrieved records against the eligibility criteria using a standardized MS Excel-based form. Studies marked as potentially eligible proceeded to full-text screening. The same two reviewers independently assessed the full-text articles for eligibility. To assess inter-rater reliability during title and abstract screening, we calculated Cohen’s kappa (κ) statistic. Discrepancies between the two reviewers were resolved through discussion and consensus agreement. We documented the reasons for the exclusion of full-text articles (S1 Table). The complete selection process is illustrated in the PRISMA flowchart (Fig. 1).

Fig. 1.

PRISMA diagram showing the study selection process

Data collection process

Two reviewers (BJ and BN) independently extracted the data using a standardized data extraction form that had been piloted on three studies and refined prior to use (S2 Table). The extracted data included study design, participant demographics, intervention details, comparator information, outcomes, and statistical results. For continuous outcomes, we extracted the sample sizes, means, and standard deviations (or medians and ranges where means were unavailable). For dichotomous outcomes, we extracted the number of events and total number of participants in each group. Discrepancies between the reviewers were resolved through discussion. We did not contact the study authors to obtain or confirm missing data. When multiple publications reported data from the same study, we treated them as a single study and extracted complementary data from each report.

In particular, the following data were extracted from each study: first author, publication year, country, design, sample size, follow-up duration, age (mean, range), pregnancy status, parity, VVC type (acute vs. recurrent), number of prior episodes, Candida species, inclusion/exclusion criteria, baseline zinc levels, zinc formulation (oral, topical, vaginal), frequency, co-interventions, type (placebo, antifungal, no treatment), dose, duration, and outcomes.

Risk of bias in individual studies

Two reviewers (BJ and BN) jointly appraised the included studies for the risk of bias using the Joanna Briggs Institute (JBI) critical appraisal tools that were appropriate for each study design (S3 Table).

Randomized controlled trials: JBI checklist for RCTs (13 items) assessing random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias.

Case-control studies: JBI checklist for case-control studies (10 items) assessing the comparability of groups, measurement of exposure, confounding factors, and strategies to address confounding.

Quasi-experimental studies: JBI checklist for quasi-experimental studies (nine items) assessing clarity of cause and effect, similarity of participants across groups, presence of a control group, multiple measurements of outcomes, completeness of follow-up, and appropriateness of statistical analysis.

Each item was rated as “yes” (low risk), “no” (high risk), “unclear” (uncertain risk), or “not applicable.” The overall risk of bias for each study was categorized as low, moderate, or high based on the pattern of responses. The certainty of evidence for each primary outcome was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework, considering five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. We did not exclude studies based on the risk of bias alone. However, we performed sensitivity analyses based on the most influential studies after study influence analysis. Additional sensitivity analyses were performed by restricting the meta-analysis to studies using the ≥ 3 episodes threshold (k = 4) and to studies using the ≥ 4 episodes threshold (k = 2), to assess whether the RVVC episode definition influenced the pooled estimate.

Risk of bias across studies

We did not formally assess publication bias using funnel plots or Egger’s test due to an insufficient number of studies per outcome (< 10 studies required for meaningful assessment). However, our comprehensive search strategy, including grey literature and trial registries, was designed to minimize this risk. We assessed selective reporting bias within individual studies as part of the JBI risk of bias assessment (S3 Table).

Summary measures

Standardized mean differences (SMDs) with 95% confidence intervals (CIs) were calculated for case–control studies to account for differences in measurement units (µg/dL versus µmol/L). According to Cohen’s convention, SMDs of 0.2, 0.5, and 0.8 were interpreted as small, medium, and large effects, respectively. In intervention studies assessing recurrence prevention, risk ratios (RRs) with 95% CIs were computed for dichotomous outcomes (reinfection: yes/no). When data were available, risk differences (RDs) and numbers needed to treat (NNT = 1/RD) were also derived. For intervention studies evaluating symptom relief, mean differences (MDs) with 95% CIs were extracted for continuous outcomes, such as the visual analog scale (VAS) scores. All effect estimates are presented with corresponding 95% confidence intervals and two-tailed p-values.

Synthesis of results

Effect sizes were calculated as standardized mean differences (SMDs) using Hedges’ g to adjust for small sample sizes in the studies. Zinc concentrations reported in micromoles per liter (µmol/L) were converted to micrograms per deciliter (µg/dL) using the molecular weight of zinc (65.38 g/mol) according to the following formula:

, to standardize the units across studies. Red blood cell (RBC) zinc measurements were excluded from the meta-analysis because of their distinct cellular matrix, which is different from plasma/serum, and their measurement basis (µg/g hemoglobin). Zinc measured in cervicovaginal secretion was also excluded from the pooled analysis because it represents a distinct biological compartment not directly comparable to systemic zinc levels and was instead reported narratively. Where a single study contributed two comparisons sharing the same control group (attack and remission phases), only the primary comparison (attack phase) was included in the primary meta-analysis to avoid double-counting of control participants. The remission phase was included in a sensitivity analysis.

A random-effects meta-analysis was performed using the restricted maximum likelihood (REML) method to estimate pooled effect sizes, accounting for both within- and between-study variability. Statistical heterogeneity was assessed using the I² statistic, and tau². I² values of 25%, 50%, and 75% were interpreted as low, moderate, and high heterogeneity, respectively.

Subgroup analyses were conducted based on the sample type (plasma versus serum) and pregnancy status as pre-specified moderators. A mixed effects model was applied. For recurrence outcomes, risk ratios (RRs) and risk differences (RDs) were calculated using Fisher’s exact test due to small cell counts. Influence diagnostics and leave-one-out analyses were performed to assess robustness of estimates.

A pre-specified sensitivity analysis was performed excluding Baltaci et al. [11], given the biological rationale that pregnant women have distinct physiology that may not be generalizable to the broader non-pregnant VVC population. Forest plots were generated to visualize the individual and pooled estimates. All analyses were performed in R (version 4.5.3). Effect sizes were calculated using the metafor package (version 4.8.0), meta-analysis pooling, subgroup analyses, and forest plots were performed using the meta package (version 4.8.0), and the epitools package (version 0.5.10.1) was used for recurrence outcome statistics.

In vitro studies

Quasi-experimental and invitro studies were included in the systematic review to capture mechanistic evidence on zinc’s anti-virulence activity. Their findings were reported narratively to contextualize the biological plausibility of zinc-based interventions.

Handling missing data

Where individual data were incompletely reported, including sampling methods, exclusion criteria, or participant characteristics, the available data were extracted as reported and the gaps noted in extraction form (S2 Table). Studies were not excluded solely on the basis of incomplete reporting.

Results

Search results

Our search identified 45 records, of which 32 underwent title and abstract screenings. The two reviewers (BJ and BN) agreed on 24 studies, with Cohen’s kappa indicating moderate inter-rater reliability (75% observed agreement, κ = 0.506, z = 2.95, p = 0.003). Eight studies (25%) had discrepant inclusion/exclusion decisions between reviewers, which were subsequently resolved through discussion and consensus. After resolving eight discrepant decisions through consensus discussion, we included 15 studies for the full-text review and subsequently 11 [6, 7, 10–18] in the systematic review (Fig. 1). Two studies were excluded because of the unavailability of full-text articles (due to paywall restrictions or complete lack of access). One clinical trial was excluded because no results were available, and another study was excluded because the assessment of zinc’s effects on VVC was not its primary objective.

Risk of bias within studies

The risk of bias was assessed for all 11 included studies using JBI checklists, with one (9%) rated as low risk, eight (73%) as moderate risk, and two (18%) as high risk. Case-control studies (n = 6) mostly showed moderate risk, with one low and one high risk. All two RCTs had moderate risk, and among the two quasi-experimental studies, one was moderate and one was of high risk (S3 Table). Common methodological limitations included the lack of blinding across interventional studies, inadequate control of confounding (only 27% addressed confounders), incomplete outcome data with high attrition in RCTs, and design limitations in quasi-experimental studies (e.g., absence of control groups). Strengths across studies included reliable outcome measures, clear case definitions, appropriate randomization in RCTs, and generally suitable statistical analyses.

Study characteristics

Across all 11 included studies [6, 7, 10–18], there were 946 participants from clinical or observational studies and 46 Candida isolates from experimental investigations (Table 1).

Table 1.

Characteristics of included studies

Study, year, country, reference	Design	Population	Sample Size	Age (years)	Intervention / Exposure	Comparator/ Control	Primary Outcome(s)	Main findings
Edman et al. 1986, USA [6]	Cross-sectional case-control	Women with RVVC and healthy matched controls	29 RVVC 20 controls	RVVC: 35 ± 6.7 (22–47); Control: 34 ± 8.1 (20–51)	RVVC (≥ 4 episodes/12 months)	Healthy age-, race-, parity-matched controls	Plasma and RBC zinc levels	Women with RVVC had significantly lower plasma zinc levels than controls (81 ± 11.6 µg/dL versus 91 ± 14.2 µg/dL, p = 0.015), while RBC zinc levels did not differ (14 ± 2.3 µg/g versus 13 ± 1.5 µg/g, p = 0.44). Additionally, 17% (5/29) of RVVC patients had plasma zinc < 70 µg/dL compared to 0% of controls.
Shahverdi et al. 1993, Iran [12]	Case-control	Women with RVVC and volunteers	40 RVVC, 40 controls	Matched between groups (NR)	RVVC (≥ 4 episodes/year despite treatment)	Volunteers matched for age and pregnancies	Plasma zinc levels	RVVC patients had lower plasma zinc levels compared with controls (55 ± 21 µg/dL versus 62 ± 23 µg/dL), though the difference was not statistically significant (p < 0.09). The odds of low zinc status were 1.7-fold higher in RVVC patients than controls
Böhler et al. 1994, Austria [7]	Cross-sectional controlled comparative	Non-pregnant women with regular cycles	21 RVVC 15 controls	20–49 (mean 33.7)	RVVC (≥ 4 episodes/12 months)	Healthy women, culture-negative for Candida and STIs	Plasma and cervicovaginal zinc levels	RVVC women had plasma zinc of 109 ± 17.7 µg/dL and cervicovaginal zinc of 41 ± 30.2 µg/dL, while controls had plasma zinc of 107 ± 12.8 µg/dL and cervicovaginal zinc of 42 ± 29.6 µg/dL, with no significant differences between groups (plasma P = 0.71, cervicovaginal P = 0.80). Additionally, there was no meaningful correlation between plasma and cervicovaginal zinc levels in either group (r = − 0.05 for patients; r = − 0.07 for controls)
Baltaci 1998, Turkey [11]	Case-control	Pregnant women with VVC	15 VVC 10 controls	Not specified	VVC (acute, not recurrent)	Healthy pregnant controls	Serum zinc, leukocytes, lymphocytes	Pregnant women with acute VVC had reduced serum zinc (62.36 ± 8.62 µg/dL) compared with healthy pregnant controls (80.11 ± 7.11 µg/dL) (p < 0.05) and showed altered leukocyte counts (patients: 7300 ± 1927/mm³ versus controls: 6700 ± 1190/mm³).
Spacek et al. 2005, Czech Republic [14]	Cohort multicenter observational	Women with RVVC and healthy controls	44 RVVC 30 controls	RVVC: 31.7 (17–51), Control: 33.7 (21–44)	RVVC (≥ 4 culture-documented episodes/12 months)	Healthy women	Serum zinc, calcium, magnesium, iron	Patients with RVVC had significantly lower serum zinc levels than healthy controls during both the attack phase (13.25 ± 2.69 µmol/L versus 17.52 ± 5.62 µmol/L; p ≤ 0.001) and remission (13.16 ± 2.28 µmol/L vs. 17.52 ± 5.62 µmol/L; p ≤ 0.001), while attack versus remission showed no difference (ns). Approximately 75% of all patient zinc values clustered near the lower limit of normal
Zangeneh et al. 2008, Iran [15]	Case-control	Women with RVVC	53 RVVC 51 controls	Control: 33.12 ± 5.37 (15–50)	RVVC (≥ 4 episodes/12 months)	Healthy age-matched women	Plasma zinc levels	Baseline plasma zinc concentration in the RVVC patient group was 57.33 ± 74.0 µg/dL, compared with 93.38 ± 74.0 µg/dL in the healthy control group (p < 0.05). 94% of patients with RVVC and 88% of controls showed no statistically significant within-group differences in zinc values (p > 0.05).
Baltaci et al. 1999, Turkey [13]	RCT, open-label	Non-pregnant sexually active women with symptomatic VVC	35 intervention 33 control 54 completed 90d	Group I: 24.61 ± 6.44; Group II: 25.87 ± 5.30 (18–37)	Fluconazole 150 mg/day + zinc sulfate 50 mg/day × 30 days	Fluconazole 150 mg/day × 30 days alone	Cumulative reinfection rate	At 10 days, reinfection rates were similar between the fluconazole-only group (Group I: 15.15%) and the fluconazole plus zinc group (Group II: 14.28%) with comparable KOH and culture negativity rates. At 30–35 days, reinfection rates were 21.42% in Group I and 6.66% in Group II. At 90 days, reinfection rates were 22.72% in Group I and 10.71% in Group II. Overall, total reinfection rates across the study period were 48.48% in the fluconazole-only group (Group I) and 28.57% in the fluconazole plus zinc group (Group II) (p < 0.005).
Murina et al. 2023, Italy [17]	Multicenter clustered RCT, open-label	Women with moderate-severe VVC (Sobel score ≥ 2)	220 recruited; 192 completed (97 intervention, 95 control)	Group 1: 27.5 ± 6.3, Group 2: 27.3 ± 6.3 (18–45)	Clotrimazole 2% vaginal cream + acid pH thymol and zinc-containing cleansing wash	Clotrimazole 2% vaginal cream alone	Pruritus and burning (VAS scores)	At Day 10 and Day 15, pruritus (itching) VAS scores were significantly lower in Group 1 (clotrimazole vaginal cream with zinc cleansing wash) compared to Group 2 (clotrimazole vaginal cream alone) (p < 0.005 at both timepoints). The VAS scores for burning sensation improved in both groups, with no significant difference between them. Two patients in Group 1 reported mild burning, which resolved within 2–3 days
Roselletti et al. 2023, Hungary [10]	Retrospective single-arm cohort	Women with recurrent vaginal infections (RVVC or RBV)	8 completed (6 RVVC, 2 RBV), 10 enrolled	Not specified	Zinc vaginal hydrogel (20µM zinc sulphate) 2 ml nightly × 2 weeks, then 2×/week	None (single-arm)	Reinfection within 3 months	Among women with RVVC, 5 of 6 (83%) had no reinfection during the 3-month follow-up. In contrast, both women with RBV (0/2) experienced reinfection
Roselletti et al. 2023, Italy [10]	Cross-sectional observational + mechanistic (ex vivo, in vivo)	Women with and without Candida colonization/infection	29 (17 C. albicans+, 12 negative) for biomarkers; 36 total for neutrophil counts	18–50 (Group 1: 21–40, Group 2: 18–50)	C. albicans colonization/infection	Candida-negative women	IL-1β, IL-8, calprotectin, vaginal pH, neutrophils, PRA1 expression	During vaginal infections, C. albicans strongly upregulates the PRA1 gene, which encodes a zinc-binding molecule (PRA1), and its expression positively correlates with levels of proinflammatory cytokines in women. Genetic deletion of PRA1 in mice prevented vaginal inflammation, while application of a zinc solution downregulated PRA1 expression and blocked immunopathology. Vulvovaginal candidiasis is associated with a vaginal inflammatory environment characterized by elevated pH (mean 4.57 versus 3.75 in uninfected women), increased proinflammatory cytokines (IL-1β and IL-8), and a 9-fold higher neutrophil infiltration. This response is species-specific: C. albicans infections induce robust neutrophil recruitment (9.25 PMN/field), whereas C. glabrata infections show neutrophil counts similar to uninfected women (2.25 vs. 1.08 PMN/field).
Hosseini et al. 2020, Iran [16]	Descriptive-analytic cross-sectional (in vitro)	Women with VVC attending gynecology hospital	280 VVC patients; 196 C. albicans isolates (10 fluconazole-resistant selected)	18–46	ZnO nanoparticles tested against fluconazole-resistant isolates (in vitro)	Standard reference strain ATCC 10,231	MIC, SAP1-3 genes expression, antifungal activity	ERG11 was detected in all fluconazole-resistant isolates, but not in susceptible strains. SAP1-3 genes were expressed in most isolates, with significantly higher expression in the resistant strains (P < 0.001). Treatment with zinc oxide nanoparticles (ZnO-np) at sub-MIC concentrations significantly downregulated SAP gene expression (SAP1: 2.44-fold, SAP2: 5.55-fold, SAP3: 2.6-fold; all P < 0.05), while Nystatin also reduced SAP expression (SAP1: 4.16-fold, SAP2: 3.12-fold, SAP3: 4.8-fold; all P < 0.05). Overall SAP downregulation was significant for ZnO (P = 0.01) and marginal for Nystatin (P = 0.07).
Abid et al. 2024, Iraq [18]	Cross-sectional laboratory (in vitro)	Married women with VVC attending hospital	120 clinical samples; 69 C. albicans positive (10 selected for testing)	20–50	Garlic-synthesized ZnO nanoparticles tested against isolates (in vitro)	No clinical control group	Inhibition zones, virulence gene detection (ALS1, HWP1, ACT1)	Molecular analysis showed 9/10 isolates had ALS1, 10/10 had HWP1, and 7/10 had ACT1 genes. ZnO NPs produced inhibition zones of 8.99 ± 2.24 mm at 250 ppm and 19.57 ± 2.54 mm at 500 ppm, while AgNPs produced 8.04 ± 3.47 mm at 250 ppm and 21.25 ± 3.18 mm at 500 ppm. Both nanoparticles showed concentration-dependent antifungal activity, with silver nanoparticles slightly more effective than zinc at the higher concentration

RVVC = Recurrent Vulvovaginal Candidiasis, VVC = Vulvovaginal Candidiasis, RBV = Recurrent Bacterial Vaginosis, RCT = Randomized Controlled Trial, NR = Not Reported, OCP = Oral Contraceptive Pills, VAS = Visual Analogue Scale, STI = Sexually Transmitted Infection, MIC = Minimum Inhibitory Concentration, SAP = Secreted Aspartyl Proteinase, RBC = Red Blood Cell, CVL = Cervicovaginal Lavage, ppm = parts per million

Zinc levels in VVC

Six plasma/serum comparisons from six observation studies [6, 7, 11, 12, 14, 15] were eligible for the primary meta-analysis, comprising 209 patients with VVC/RVVC and 149 healthy controls. Red blood cell zinc [6], and cervicovaginal secretion zinc [7] were reported narratively. For Spacek et al., (2005) [14], the attack phase comparison was retained for the primary pooled analysis while the remission phase (same control group) was examined in sensitivity analysis. All studies used atomic absorption spectrophotometry for zinc measurements.

Across individual studies, plasma/serum zinc levels in VVC/RVVC patients ranged from 62 to 109 µg/dL compared to 80 - 114 µg/dL in controls. Random-effects meta-analysis of six comparisons (k = 6) revealed significantly lower systemic zinc levels in patients with VVC/RVVC than in healthy controls (pooled SMD =-0.72, 95% CI: -1.15 to -0.28, p = 0.001) (Fig. 2). Substantial variability was observed (I²=73.5%, tau²=0.2096, p = 0.0021).

Fig. 2.

Standardized mean differences in systemic zinc levels between women with VVC/RVVC and healthy controls. (a) Primary random-effects analysis of all six eligible plasma/serum comparison. (b) Sensitivity analysis restricted to non-pregnant women. Square markers represent individual study effect sizes proportional to study weight, horizontal line represent 95% CI, diamonds represent pooled estimates. SMD = standardized mean difference, VVC = vulvovaginal candidiasis, RVVC = recurrent VVC

Sample type

Sample type was examined as a subgroup moderator. Using common effects model, both serum (SMD = − 0.83, 95% CI: -1.19 to -0.46) and plasma (SMD = -0.55, 95% CI: -0.81 to -0.28) showed significant reductions in systemic zinc levels in patients with VVC/RVVC compared to controls. However, the test for subgroup differences between sample types was not significant (p = 0.226). Under random effects model, serum showed a larger but non-significant effect (SMD = -0.98, 95% CI: -2.09 to 0.14) with substantial within-serum sub-group heterogeneity (I2 = 87.1%), while plasma showed a significant consistent effect as with the common effects model (SMD = -0.55, 95% CI: -0.81 to -0.28). After adjusting for sample type, substantial residual heterogeneity remained (I² = 73.5%).

Influence and sensitivity analysis

Influence diagnostics identified Baltaci et al., (1998) [11] as the only statistically influential study (rstudent = -2.66), DFFITS = -0.86, Cook’s distance = 0.54). Pregnancy was a significant moderator of the zinc-VVC association (X² = 9.14, df = 1, p = 0.003). The pregnant cohort showed a larger effect (SMD = -2.19, 95% CI: -3.19 to -1.18, p < 0.001) than the five non-pregnant cohorts (SMD = -0.55, 95% CI: -0.89 to – 0.21, p = 0.002). After adjusting for the pregnancy status, residual heterogeneity (I²) decreased from 73.5% to 57.0% (p = 0.054) (Fig. 3).

Fig. 3.

Pregnancy status as a pre-specified moderator of the zinc-VVC association. Studies are grouped by pregnancy status

RVCC definition

Sensitivity analyses by RVVC episode definition showed that studies using ≥ 4-episode threshold (k = 2) [14, 15] produced a significant and consistent difference in systemic zinc levels (SMD = -0.80, 95% CI: -1.18 to -0.42, I² = 31%, p < 0.001). In contrast, restriction to studies using ≥ 3-episode threshold (k = 4) [6, 7, 11, 12] yielded a non-significant difference in systemic zinc levels (SMD = -0.70, 95% CI: -1.43 to 0.03, I² = 81%, p = 0.061). The lack of statistical significance in the ≥ 3-episode subgroup likely reflected the high heterogeneity within this subgroup driven by the opposing effects of Baltaci (pregnancy status’ large effect) [11] and Bohler’s null result [7].

Certainty of evidence

The evidence for lower systemic zinc levels in women with VVC was rated as low certainty, downgraded for serious risk of bias inherent to the observational cross-sectional design, substantial heterogeneity (I² = 73.5%), and serious imprecision (S4 Table). Evidence for zinc supplementation reducing VVC recurrence was rated as very low certainty, due to the single underpowered trial and wide confidence intervals (RR = 0.59, 95% CI: 0.31–1.11, p = 0.134). Evidence for adjunct zinc-containing therapy improving pruritus was similarly rated as very low certainty, based on single trial with a confounded intervention (zinc combined with thymol and acidic pH) and an outcome data that preclude attribution of benefit specifically to zinc.

Recurrence prevention

The cumulative 90-day reinfection rate was 28.57% (10/35) in the zinc group and 48.48% (16/33) in the fluconazole-only control group (RR = 0.59, 95% CI: 0.31–1.11, p = 0.134; RD = -19.9%, 95% CI: -42.6% to 2.8%; NNT = 5). At individual timepoints, re-infection rates were numerically lower in zinc group at all assessments (Table 2). None of the differences in re-infection rates between the zinc and control groups at any time-point reached statistical significance.

Table 2.

Cumulative incidence of recurrent VVC episodes over 90 days with zinc supplementation

Timepoint	Events Zinc (n/N)	Events Control (n/N)	Reinfection Rate (Zinc) (%)	Reinfection Rate (Control)(%)	RR (95% CI)	RD (95% CI)	NNT	Fisher’s P-value*
10 days	5/35	5/33	14.3%	15.2%	0.94 (0.30–2.96)	−0.9% (− 17.7% to 16.0%)	115	1.000
30 days	2/30	6/28	6.7%	21.4%	0.31 (0.07–1.42)	−14.8% (− 32.4% to 2.9%)	7	0.138
90 days	3/28	5/22	10.7%	22.7%	0.47 (0.13–1.76)	−12.0% (− 32.9% to 8.9%)	8	0.277
Total (0–90 days)	10/35	16/33	28.6%	48.5%	0.59 (0.31–1.11)	−19.9% (− 42.6% to 2.8%)	5	0.134

*All p-values calculated using Fisher’s exact test. RR = Risk Ratio (Zinc/Control), RD = Risk Difference (Zinc − Control), negative RD indicates fewer reinfections in the zinc group, NNT = Number Needed to Treat, reported descriptively given non-significant outcomes. Source: Data from Baltaci et al. 1999 [13]. Denominators decrease across timepoints due to loss to follow-up

A single-arm retrospective cohort study (Roselletti et al. 2023, n = 8) [10] assessed topical zinc hydrogel (20 µM zinc sulfate, applied 2 ml intravaginally nightly for 2 weeks, then twice weekly) in women with recurrent vaginal infections. Among the women with RVVC, 83% (5/6) had no reinfection during the 3-month follow-up period. In contrast, both women treated for recurrent bacterial vaginosis (0/2, 0%) experienced re-infection.

Symptom relief

One multicenter RCT (Murina et al. 2023, n = 192) [17] compared clotrimazole 2% vaginal cream plus acid-pH thymol and zinc-containing cleansing wash (n = 97) with clotrimazole 2% vaginal cream alone (n = 95) over 15 ± 3 days. Participants self-reported daily symptom severity using Visual Analog Scale (VAS) scores (0–10). Pruritus (itching) VAS scores were significantly lower in the zinc combination group than in the clotrimazole alone group on days 10 (p < 0.005) and 15 (p < 0.005). The VAS scores for burning sensation improved in both groups, with no significant difference between them. Two patients in the zinc group reported mild transient burning, which resolved spontaneously within 2–3 days.

Immune and micronutrient alterations

Baltaci (1998) found that pregnant women with VVC had 24% lower leukocyte counts (7300 ± 1927/mm³ vs. 9700 ± 803/mm³, p < 0.05) and 28% lower lymphocyte percentages (16.09 ± 4.43% vs. 22.26 ± 2.40%, p < 0.05) than healthy pregnant controls, indicating immune cell depletion associated with VVC [11].

Spacek et al. (2005) reported that patients with RVVC had significantly lower micronutrient levels than controls: 8% lower calcium levels (2.21–2.23 mmol/L versus 2.38 mmol/L, p ≤ 0.001) and 9–10% lower magnesium levels (0.79–0.82 mmol/L versus 0.87 mmol/L, p ≤ 0.001), with magnesium further declining during symptomatic attack phases compared to remission periods (0.79 versus 0.82 mmol/L, p ≤ 0.01) [14]. Iron showed non-significant trends toward elevation in patients with RVVC (17.53 ± 8.74 µmol/L during attack versus 13.68 ± 5.14 µmol/L in controls, p > 0.05). Notably, all micronutrient values remained within the normal reference ranges.

Böhler et al. (1994) measured both plasma and cervicovaginal zinc in 21 patients with RVVC and 15 controls. They found no significant differences in either compartment (plasma: 109 ± 17.7 versus 107 ± 12.8 µg/dL, p = 0.71; cervicovaginal: 41 ± 30.2 versus 42 ± 29.6, p = 0.80) and no correlation between plasma and cervicovaginal zinc levels (r = -0.05 in VVC affected women, r = -0.07 in controls) [7].

Inflammatory markers and PRA1

Roselletti et al. (2023) measured inflammatory markers in 17 C. albicans-colonized/infected women versus 12 uninfected controls [10]. VVC was characterized by an elevated vaginal inflammatory environment with elevated vaginal pH (mean 4.57 versus 3.75 in uninfected women), increased proinflammatory cytokines (IL-1β and IL-8), elevated calprotectin, and 9-fold higher neutrophil infiltration (9.25 versus 1.08 polymorphonuclear neutrophils per field). This inflammatory response was species specific. C. albicans infections triggered robust neutrophil recruitment (9.25 PMN/field), whereas C. glabrata infections showed neutrophil counts comparable to those in the uninfected states (2.25 versus 1.08 PMN/field).

Among all measured parameters, PRA1 (pH-regulated antigen 1) expression showed the strongest correlation with IL-8 levels, directly linking this fungal zinc-scavenging protein to host inflammatory cytokine production. In experimental models, zinc supplementation at concentrations mimicking therapeutic levels (25 µM zinc sulfate) downregulated PRA1 expression by over 300-fold in vitro in vaginal epithelial cultures maintained at an acidic pH [10]. In mouse VVC models, intravaginal zinc administration reduced PRA1 expression by 2.6-fold in vivo, abolished all measured inflammatory cytokines (IL-1β, CxCL-1, CxCL-2, CxCL-5, and IL-6), and reduced neutrophil infiltration to levels indistinguishable from uninfected controls, without reducing fungal colonization burden, suggesting a mechanism for converting pathogenic Candida into harmless commensals rather than killing fungi.

Genetic manipulation experiments validated PRA1’s causal role. Deletion of PRA1 from C. albicans abolished its ability to recruit human neutrophils in ex vivo assays, whereas genetic complementation restored this ability [10]. Conversely, engineering C. glabrata (which naturally lacks PRA1) to express this gene conferred a neutrophil chemotactic function. In mouse models, PRA1-deleted C. albicans strains colonized vaginal tissue at levels comparable to wild-type strains, but triggered significantly reduced inflammatory cytokine production, neutrophil infiltration, and clinical signs of infection [10].

Fungal virulence genes

Hosseini et al. (2020) tested zinc oxide nanoparticles (ZnO-np) against 10 fluconazole-resistant C. albicans isolates from 280 patients with VVC [16]. At sub-minimum inhibitory concentrations (concentrations too low to kill fungi directly), ZnO-np significantly downregulated the secreted aspartyl proteinase gene SAP1 by 2.44-fold (p < 0.02), SAP2 by 5.55-fold (p < 0.03), and SAP3 by 2.6-fold (p < 0.05), with an overall significant reduction (p = 0.01). These SAP enzymes mediate tissue invasion and immune evasion. Importantly, zinc maintained its anti-virulence activity against azole-resistant strains harboring the ERG11 resistance marker, with minimum inhibitory concentrations of 5–11 µg/mL. Nystatin significantly downregulated each SAP gene (SAP1,4.16-fold; SAP2,3.12-fold; SAP3,4.8-fold; all p < 0.05). However, when SAP1–3 were evaluated collectively, the overall reduction effect of Nystatin was only marginal (p = 0.07) [16].

Abid (2024) evaluated garlic-synthesized ZnO nanoparticles against 10 C. albicans isolates from 69 VVC-positive clinical samples (57.5% positivity rate) [18]. Molecular analysis showed near-universal expression of virulence genes associated with tissue adherence and invasion, where 100% (10/10) carried HWP1 (hyphal wall protein), 90% (9/10) carried ALS1 (adhesin), and 70% (7/10) carried ACT1 [18]. ZnO nanoparticles demonstrated concentration-dependent antifungal activity with inhibition zones of 8.99 ± 2.24 mm at 250 parts per million (ppm) and 19.57 ± 2.54 mm at 500 ppm. For comparison, garlic-synthesized silver nanoparticles produced inhibition zones of 8.04 ± 3.47 mm at 250 ppm and 21.25 ± 3.18 mm at 500 ppm, showing slightly superior efficacy at the higher concentration [18].

Discussion

We systematically synthesized evidence from 11 studies to evaluate the role of zinc in vulvovaginal candidiasis. Our meta-analysis of six plasma/serum comparisons involving 358 participants revealed significantly lower systemic zinc levels in women with VVC than in healthy controls, though the certainty of this evidence was rated as low owing to the observational cross-sectional design and significant heterogeneity between studies. Clinical trial evidence remains sparse and of very low certainty. However, experimental studies have identified PRA1, a zinc-regulated fungal virulence factor, as a candidate molecular link between zinc deficiency and symptomatic inflammation, although this evidence was derived primarily from a single study [10] and requires independent validation.

Serum measurements showed the largest and only statistically significant reduction in zinc level (SMD = -0.94, 95% CI: -1.79 to -0.10), while plasma showed a non-significant trend (SMD = − 0.57, 95% CI: -1.36 to 0.22). Substantial between-study heterogeneity (I² = 73.5%) was partly explained by pregnancy status, which emerged as a highly significant moderator (X² = 9.14, df = 1, p = 0.003). The pregnant cohort [11] showed a substantially larger zinc deficit than non-pregnant women (SMD = -2.19 versus − 0.55) and the residual heterogeneity decreased to I² = 57.0% (p = 0.054) after accounting for this difference. The larger effect in pregnant women most likely reflects hemodilution [19, 20], altered zinc transport [21], and increased fetal demand [21] rather than greater biological susceptibility to VVC. Among non-pregnant women, residual heterogeneity likely reflects unmeasured differences in dietary zinc intake, fasting status at blood collection, genetic variation in zinc transporters, and assay methodology across studies over two decades and five countries.

A single study measuring cervicovaginal zinc found no significant difference between patients and controls, and no correlation with plasma zinc levels [7]. This disconnect between systemic and local vaginal zinc concentrations suggests that systemic measurements may not predict zinc availability at the site of infection. Whether VVC results from systemic zinc deficiency affecting immune function, local vaginal zinc depletion enabling fungal virulence, or both acting synergistically remains unclear.

The broader pattern of micronutrient abnormalities we identified across studies, including reduced calcium, magnesium, and immune cell populations [11, 14], raises the possibility that VVC may be associated with generalized nutritional insufficiency rather than isolated zinc deficiency, although these observations are based on limited data and require further investigations.

Critically, we noted bidirectional nature of zinc-VVC relationship. The observed lower zinc levels in women with VVC may not exclusively represent pre-existing nutritional deficiency predisposing to infection. During active Candida infection, the host actively reduces zinc availability as defense strategy (nutritional immunity), a process mediated in part by calprotectin [22], a zinc-sequestering protein found to be elevated in vaginal environments of women with VVC [10, 23]. Under this model, reduced zinc levels may partly represent a host response to infection rather than a risk factor. Since all six observational studies used a cross-sectional design measuring zinc at the time of active infection, distinguishing cause from consequence was not possible from the available evidence. Longitudinal studies measuring zinc before, during, and after VVC resolution are required to establish the temporal direction of this association. Importantly, this bidirectional relationship does not invalidate the therapeutic rationale for zinc supplementation. Zinc's suppressive effects on PRA1-mediated inflammation remains a plausible target regardless of whether zinc deficiency is primary or secondary.

Experimental evidence suggests a plausible explanation for how zinc deficiency may translate into symptomatic VVC. The proposed pathway centers on PRA1, a fungal protein expressed by C. albicans under zinc-restricted conditions that may trigger an inflammatory response [10]. When zinc is scarce, C. albicans expresses PRA1, which stimulates vaginal epithelial cells to produce IL-8, recruiting neutrophils that generate the burning, itching, and discharge characteristics of VVC, without effectively clearing the infection.

Genetic manipulation experiments have demonstrated the specificity of this pathway. Deletion of PRA1 in C. albicans eliminates inflammatory responses, whereas fungal colonization persists. Conversely, engineering PRA1 into C. glabrata (which naturally lacks it) confers a neutrophil chemotactic potential [10]. Zinc supplementation appears to reverse this process by suppressing PRA1 expression, with concentrations achievable through supplementation (25 µM) reducing PRA1 levels by more than 300-fold in vitro and abolishing inflammatory cytokines in mouse VVC models without reducing fungal colonization [10]. Beyond PRA1, zinc oxide nanoparticles at sub-inhibitory concentrations downregulated secreted aspartyl proteinase genes (SAP1-3) by 2.4- to 5.6-fold in fluconazole-resistant C. albicans isolates [16], suggesting additional anti-virulence mechanisms independent of fungicidal activity. This framework could explain why fungal colonization can be asymptomatic [24, 25], why symptoms correlate poorly with fungal burden [26, 27], and why C. albicans causes inflammatory VVC, while C. glabrata typically does not, despite similar colonization rates [28, 29].

The species specificity of PRA1 expression carries direct clinical implications for the therapeutic role of zinc. Since non-albicans species do not express PRA1 [10], the proposed anti-inflammatory mechanism of zinc is operative only in C. albicans- associated VVC. Given that non-albicans species account for an estimated 30–40% of RVVC cases, including the increasingly prevalent C. glabrata and C. tropicalis, zinc-based therapy targeting PRA1 suppression may be ineffective in a substantial proportion of patients. Future trials should stratify the effect by Candida species and test whether clinical response to zinc differs between C. albicans and non-albicans infections. Beyond species specificity, several limitations constrain the interpretation of the mechanistic evidence more broadly.

All PRA1 mechanistic work is derived from a single study [10], and independent validation in diverse populations has not been reported. The proposed therapeutic model, whereby zinc converts pathogenic Candida into harmless commensals by silencing virulence factors rather than killing fungi, remains a hypothesis that has not yet been demonstrated in human clinical trials. Until replicated and validated, these findings represent a promising mechanistic concept rather than an established biological pathway warranting clinical translation.

This experimental-clinical gap is further illustrated by the available intervention data. The very low certainty of available clinical evidence means that no recommendation for or against zinc-based therapy can be made with the current data. The single adequately controlled trial testing the effect zinc supplementation on preventing VVC recurrence [13], produced a numerically lower 90-day reinfection rate (28.6% versus 48.5%) but did not reach statistical significance. The consistent directional trend across all time points and the descriptive low number needed to treat (NNT = 5) are hypothesis generating, but the study was underpowered to provide definitive evidence. The single-arm retrospective cohort study that reported 83% remission with topical zinc hydrogel [10] lacked a control group and was insufficiently powered to draw any inferential conclusions. It was hypothesis generating.

One multicenter RCT found that adjunctive zinc-containing cleansing wash improved pruritus scores compared to antifungal therapy alone at days 10 and 15 (p < 0.005) [17]. However, the certainty of this finding is very low given the intervention combined zinc with thymol and acidic pH, precluding attribution of benefit to zinc. Additionally, the trial was not designed to isolate the zinc component.

The uncertain relationship between systemic and vaginal zinc concentrations further complicates therapeutic development. If oral supplementation increases blood zinc levels but not vaginal tissue concentrations, the therapeutic rationale based on local PRA1 suppression weakens. No clinical trial has measured vaginal zinc levels after supplementation. Therefore, it is not possible to determine whether any intervention route delivers biologically active concentrations to the infection site (vaginal environment). The numerically lower but non-significant result from the oral zinc trial [13] may reflect inadequate delivery to the vaginal tissue rather than failure of PRA1 pathway. But without pharmacokinetic data, this remains speculative.

Overall, the body of evidence evaluated in this review, rated as low to very low certainty across all primary outcomes, is insufficient to confirm or dismiss the effectiveness of zinc-based therapies in VVC.

Research priorities

Several research priorities could address these uncertainties. Pharmacokinetic studies measuring vaginal zinc concentrations following different supplementation routes (oral, topical, and combination) could establish whether and how zinc reaches the target vaginal tissue at biologically active concentrations. Independent mechanistic validation of findings on PRA1 and SAP1-3 genes in diverse populations and by independent research groups would strengthen the confidence in these biological targets. Species-stratified observational studies could clarify whether the PRA1-inflammation association generalizes beyond the Hungarian and Italian cohorts originally studied and whether the effects differ between VVC caused by C. albicans and non-albicans Candida.

Adequately powered randomized trials should integrate the pharmacokinetic assessment of vaginal zinc levels stratified by Candida species and baseline zinc status, including clinically meaningful endpoints such as symptom-free intervals and quality of life along with recurrence rates at 6–12 months, and compare oral, topical, and combination zinc formulations. These foundational studies would inform the design of definitive efficacy trials, potentially reducing the risk of testing for inappropriate interventions or formulations.

Limitations

This systematic review has several important limitations. The limited number and small size of interventional trials preclude confident therapeutic recommendations, and the certainty of evidence was rated as low to very low across all primary outcomes using the GRADE framework. All experimental evidence linking PRA1 to alleviating VVC symptoms was derived from a single study and requires independent validation across diverse populations. All six observational studies used cross-sectional designs, measuring zinc at the time of active infection which precludes causal inference as lower zinc levels may reflect a host nutritional immunity response rather than a pre-existing deficiency. Substantial unexplained heterogeneity in zinc measurements suggests unmeasured confounders, such as dietary patterns, genetic variations in zinc transporters, or methodological differences in specimen handling. Variability in the RVVC episode threshold across studies (≥ 3 versus ≥ 4 per year) contributed significant heterogeneity. No interventional studies measured vaginal zinc concentrations post-supplementation, leaving uncertainty regarding whether zinc does achieve biologically active tissue levels. The primary interventional trial did not report Candida species distribution, participant baseline zinc status, or adherence data, variables that are critical for interpreting the null result and determining which patients may benefit. The absence of species-stratified outcomes in clinical trials limits our ability to assess whether effects differ between VVC caused by C. albicans and non-albicans Candida. Systematic extraction of adverse event data was limited by sparse reporting in primary studies and all zinc deficiency studies were conducted in high-income countries (USA, Europe, Iran), which may limit generalizability to settings such as sub-Saharan Africa where baseline zinc deficiency prevalence is substantially higher and VVC burden is considerable. We could not formally assess publication bias because of insufficient studies (< 10) per outcome.

Conclusions

Lower systemic zinc levels are observed in women with VVC compared with health controls. However, this finding is based on low-certainty observational evidence and precludes distinguishing cause from effect, including the possibility of infection driven zinc sequestration through nutritional immunity. Preclinical data provide biologically plausible mechanisms linking zinc restriction to inflammation in C. albicans, although these are species-specific and not validated in humans. Clinical evidence for zinc-based interventions remains of very low certainty and is insufficient to support routine use in VVC or RVVC management. Well-powered trials incorporating species stratification, baseline zinc status assessment, and pharmacokinetics endpoints are needed before clinical translation can be considered.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

ALS1

Adhesin-Like Sequence 1

CI

Confidence Interval

CVL

Cervicovaginal Lavage

DFFITS

Difference in Fits

ECE1

Extent of Cell Elongation 1

HWP1

Hyphal Wall Protein 1

IL-1β

Interleukin-1 beta

IL-8

Interleukin-8

JBI

Joanna Briggs Institute

MD

Mean Difference

MIC

Minimum Inhibitory Concentration

NNT

Number Needed to Treat

OCP

Oral Contraceptive Pills

PMN

Polymorphonuclear Neutrophils

PRA1

pH-Regulated Antigen 1

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RBC

Red Blood Cell

RBV

Recurrent Bacterial Vaginosis

RCT

Randomized Controlled Trial

RD

Risk Difference

REML

Restricted Maximum Likelihood

RR

Risk Ratio

RVVC

Recurrent Vulvovaginal Candidiasis

SAP

Secreted Aspartyl Proteinase

SMD

Standardized Mean Difference

STI

Sexually Transmitted Infection

VAS

Visual Analogue Scale

VVC

Vulvovaginal Candidiasis

ZnO

Zinc Oxide

Author contributions

Bwambale Jonani (BJ): Conceptualization, methodology, investigation, data curation, formal analysis, visualisation, validation, writing - original draft. Betty Nakawuka (BN): Methodology, investigation, data curation, writing - review & editing. Gerald Mboowa (GM): Conceptualisation, Methodology, supervision, writing - review & editing. Felix Bongomin (FB): Conceptualization, methodology, supervision, writing - review & editing. Richard Kwizera (RK): Conceptualisation, Methodology, supervision, data validation, writing - review & editing, Project administration, validation.

Funding

None.

Data availability

The search strings, quality appraisal, risk of bias and all data extracted for this systematic review are available in the supplementary files.

Declarations

Ethical approval

This systematic review and meta-analysis utilized only published, de-identified aggregate data from previous studies. As no primary data collection involving human participants was undertaken, institutional review board approval was not required for this study. All the included studies reported ethical approval from their respective institutions.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.