Recurrent Yeast Infections: A Treatment Evidence Vacuum

Recurrent yeast infections have a treatment problem—and a research problem

For people with recurrent vulvovaginal candidiasis (RVVC), the usual medical plan is familiar: confirm the diagnosis, treat the current infection, then take an antifungal maintenance regimen for months. The regimen can reduce recurrences while treatment continues. Yet the first comprehensive Cochrane review of RVVC treatment, published in 2022, found low- or very-low-certainty evidence behind every treatment approach it assessed.

That does not mean antifungals are useless. It means the field has a real effect signal wrapped in studies too small, inconsistent, and vulnerable to bias to support much confidence about the exact size of benefit, the best regimen, or what happens after treatment stops. For a chronic condition affecting a large number of women, that is an uncomfortable mismatch between clinical need and clinical evidence.

RVVC generally means at least three symptomatic, culture-confirmed episodes in a year. It deserves that diagnostic care. Symptoms such as itching, discharge, burning, or pain can have causes other than Candida, and repeated treatment without confirmation can turn a difficult condition into a guessing game with pharmacy receipts.

The best evidence supports suppression, not cure

The Cochrane review, led by Georga Cooke and colleagues, searched through October 2021 and included 23 randomized controlled trials with 2,212 participants. It assessed drug and non-drug strategies in people with RVVC. Despite the review design sitting near the top of the evidence hierarchy, the underlying trials set a low ceiling: every included trial enrolled fewer than 100 participants.

The most useful pooled result compared antifungal maintenance therapy with placebo. Across six trials and 607 participants, maintenance antifungals reduced recurrence at six months: risk ratio (RR) 0.36, 95% confidence interval (CI) 0.21 to 0.63; number needed to treat for benefit (NNTB) 2. The confidence interval excludes 1.0, so this pooled result was statistically significant. In practical terms, fewer participants had a recurrence during the measured period in the maintenance-therapy groups.

At 12 months, the pooled estimate was RR 0.80, 95% CI 0.72 to 0.89, with an NNTB of 6. That result also favored maintenance therapy and was statistically significant. These are relative effects, however; the review's summary does not provide a single absolute recurrence rate that would tell every patient how likely recurrence is on a particular regimen.

The review rated both findings low certainty. GRADE uses that label when the estimated effect may differ substantially from the true effect. Low certainty is not a polite way of calling an answer definitive. It is a warning that the direction of the result is useful, while the precision and durability of the conclusion remain limited.

Maintenance antifungals appear to reduce recurrence during treatment, but no intervention in the review demonstrated a cure for RVVC.

The distinction matters. The review did not provide quantified relapse rates after maintenance therapy ends. Clinicians and patients often describe RVVC care as suppression for precisely this reason: the evidence documents fewer recurrences during treatment, not lasting eradication after it.

Why a favorable average does not settle the question

At the six-month antifungal-versus-placebo endpoint, results varied sharply across trials: I² was 82%. I² estimates how much variation between study results exceeds what chance alone might explain. An I² of 82% means the pooled RR of 0.36 combines studies with substantially different results.

Different azoles, schedules, populations, and recurrence definitions may explain some of that variation, but the available evidence could not sort out how much each factor mattered. By 12 months, heterogeneity was lower at I²=21%, yet the certainty rating remained low.

Risk of bias also pulled the evidence down. Many trials had inadequate or unclear allocation concealment and insufficient blinding of participants or outcome assessors. Those flaws matter especially when the outcome includes symptoms that participants report. A trial can be randomized and still overestimate a benefit if expectations differ between groups.

Funding and reporting leave further uncertainty. Eleven of the 23 trials did not report funding, while nine had industry funding. The review could not formally assess publication bias with funnel plots because no comparison had 10 or more trials. Small positive studies are easier to publish than small negative studies; with a sparse literature, that possibility deserves caution rather than a shrug.

Oral versus topical treatment: the evidence cannot pick a winner

Weekly oral fluconazole for six months is a common maintenance approach, but the review did not establish that the oral route works better than topical therapy. Three studies with 206 participants compared oral and topical antifungals. At six months, the RR was 1.66 (95% CI 0.83 to 3.31). At 12 months, it was 0.95 (95% CI 0.71 to 1.27).

Both confidence intervals cross 1.0, so neither estimate showed a statistically significant difference between routes. The review rated this evidence very low certainty. Those numbers allow for oral treatment being better, worse, or similar to topical treatment. Route choice may still matter for convenience, tolerability, drug interactions, pregnancy considerations, and an individual Candida species' susceptibility. This evidence simply cannot settle comparative effectiveness.

Probiotics and partner treatment have not earned a recommendation

Probiotics are often marketed as though a Lactobacillus capsule can solve recurrent yeast infections by restoring microbial order. The RVVC-specific trial evidence is much smaller than that story. The Cochrane review found one probiotic trial with 34 participants. That sample cannot reliably confirm or rule out a clinically meaningful effect.

Partner treatment rests on similarly thin ground. Two trials included a combined 98 participants, with no poolable evidence of benefit. Three complementary-medicine trials, totaling 354 participants, were too clinically heterogeneous to combine. Heterogeneous studies do not become a coherent result merely because they share a wellness aisle.

These findings do not prove probiotics, partner treatment, or complementary approaches never help anyone. They mean current trials do not support recommending them as primary or adjunctive RVVC treatments. Dietary interventions and immunotherapy likewise lack sufficient RVVC trial evidence for guideline-level conclusions.

Research is moving, but early signals are not treatment plans

A newer randomized noninferiority trial by Alvendal et al. compared vaginal chlorhexidine gluconate with oral fluconazole maintenance in women with RVVC. Chlorhexidine is a broad-spectrum antiseptic with reported in-vitro activity against Candida biofilm. The trial addresses a reasonable clinical question: can a non-azole option perform no worse than fluconazole when azole resistance creates pressure for alternatives? Its existence updates the field beyond the Cochrane search cutoff, but one recent trial does not establish chlorhexidine as standard care.

A 2025 Brazilian blue-light LED phototherapy report points in an even earlier direction. Blue light may affect photosensitive fungal cells, and laboratory findings can justify clinical research. This pilot report had no control arm, however. It can generate a hypothesis; it cannot show that the intervention prevents recurrences, cures RVVC, or outperforms established antifungal suppression.

Oteseconazole received FDA approval for RVVC after the Cochrane review's October 2021 search cutoff. Its phase 3 VIOLET trials are an important post-review development, but the evidence summarized here does not provide their efficacy or safety results. It would be irresponsible to fill that gap with promotional shorthand.

What the evidence supports now

For confirmed RVVC, maintenance azole therapy remains the best-supported suppressive strategy in this evidence base, with the clearest pooled comparison favoring antifungals over placebo during follow-up. The support is low certainty, not absent. It is also not proof of cure, proof that oral therapy beats topical therapy, or proof that any popular adjunct works.

Pregnant and immunocompromised people need especially careful individualized care. The Cochrane review excluded both groups, even though they may face higher burden and fewer acceptable treatment options. Fluconazole also raises pregnancy safety concerns, particularly in the first trimester, so findings from the reviewed population should not be casually applied to pregnancy.

The fairest conclusion is less satisfying than a supplement-stack answer: RVVC management can suppress recurrence, but research has not yet shown how to reliably end the condition or which maintenance strategy best fits each patient. Patients deserve diagnosis by culture when appropriate, a discussion of drug risks and options, and better trials than the field has historically funded.

The Cochrane review did not reveal that clinicians have nothing to offer. It revealed how little certainty supports what they offer—and how much work remains.