Bacterial vaginosis statistics: how common it is, how often it returns, and what the partner-treatment trial found

She had already finished three courses of metronidazole that year. The discharge and fishy odor would clear for a few weeks, then return after her period, or after sex, or for no reason she could name. Her partner had never been offered treatment. The last clinician she saw told her BV was "just an imbalance," wrote another prescription, and sent her out with a handout about douching she was not doing. By the time she reached my office she was convinced her body was broken.

It was not. Bacterial vaginosis is the most common vaginal condition among women ages 15-44 in the United States, according to the CDC. It is also one of the most stubborn. Standard antibiotics work for most first episodes, then fail a large share of women over the next year. That gap between short-term cure and durable control is where the real clinical problem lives.

This page is a data asset, not a treatment algorithm. Every figure below comes from a named primary source: CDC surveillance and guidelines, NHANES, WHO and ACOG materials, plus peer-reviewed trials including the 2025 StepUp partner-treatment study in the New England Journal of Medicine. If you need one number for a story, a chart, or an AI citation, the source sits next to it.

Group	Prevalence
Non-Hispanic Black	51.4%
Mexican American	31.9%
All women (national)	29.2%
Non-Hispanic white	23.2%

Source: Koumans et al., Sexually Transmitted Diseases, 2007; NHANES 2001-2004 vaginal Gram stain (Nugent score 7-10).

Study arm	Recurrence	Rate per person-year
Woman + male partner treated	35% (24/69)	1.6
Woman only (standard care)	63% (43/68)	4.2
Absolute risk difference	-	−2.6

Source: Vodstrcil et al., New England Journal of Medicine, 2025 (StepUp trial, modified intention-to-treat; trial stopped early for inferiority of woman-only care).

How common is bacterial vaginosis

Start with the number most reporters and clinicians still use. In NHANES 2001-2004, women ages 14-49 submitted self-collected vaginal swabs Gram-stained and scored with the Nugent method. BV (a Nugent score of 7-10) was present in 29.2% of women (95% CI 27.2%-31.3%), about 21 million women, according to Koumans, Sternberg, Bruce and colleagues in Sexually Transmitted Diseases (2007). It is a probability survey, not a clinic sample. It is also older than anyone covering this condition would prefer: NHANES has not repeated a comparable national BV module since, so 29.2% remains both the gold-standard U.S. figure and a surveillance gap.

Globally the magnitude is similar. Peebles and colleagues' 2019 meta-analysis put general-population prevalence among reproductive-age women at 23-29% across regions, with North America at 27%. WHO's November 2025 fact sheet cites the same range. Peebles estimated the annual global direct cost of treating symptomatic BV at US $4.8 billion (95% CI $3.7-$6.1 billion), and noted that the U.S. burden nearly triples when BV-associated preterm births and HIV cases are included under a causal assumption. The CDC still calls BV the most common vaginal condition in women ages 15-44. What that ranking does not show is how uneven the burden is by race.

The race gap in U.S. prevalence

In the same NHANES analysis, prevalence was 51.4% among non-Hispanic Black women, 31.9% among Mexican American women, and 23.2% among non-Hispanic white women (P < 0.01 for each comparison). That is more than a twofold Black-white difference on a laboratory definition of disease, not a difference in who presents with odor. Peebles later found the same direction inside North America: higher prevalence among Black (33%) and Hispanic (31%) women than among white (23%) and Asian (11%) women.

Koumans also linked BV to poverty, smoking, higher BMI, douching frequency, and lifetime sex partners. In multivariate models, race/ethnicity stayed positively associated, as did lifetime partners, douching, and low educational attainment. Current oral contraceptive use was inverse. The race gap survived adjustment. Douching and partner number contribute; they do not fully explain a 51% versus 23% split. Disparities of this size appear across STI burden among U.S. women and pregnancy complications. BV sits in that chain as both a common condition and a risk amplifier.

Most women have no symptoms

In the NHANES sample, only 15.7% of women with BV reported vaginal symptoms, so 84% who met the Nugent definition said they had none. CDC treatment guidelines restate the same point: in a nationally representative survey, most women with BV were asymptomatic. That does not make silent BV harmless. It means systems that wait for discharge complaints undercount the condition. Women who feel fine after a partial antibiotic course may still carry the dysbiosis linked to STI acquisition and pregnancy risk; women who feel miserable may have yeast, trichomoniasis, or mixed infection instead. Self-diagnosis is unreliable, and over-the-counter antifungals do not treat BV. See our overview of yeast infection symptoms and treatment.

Recurrence after treatment

Recommended regimens (oral metronidazole 500 mg twice daily for 7 days, metronidazole gel 0.75% for 5 days, or clindamycin cream 2% for 7 days) clear symptoms for most women short-term. The longer arc is weaker. Bradshaw and colleagues' 2006 prospective cohort in the Journal of Infectious Diseases treated symptomatic BV with oral metronidazole 400 mg twice daily for 7 days and followed swabs at months 1, 3, 6 and 12. By 12 months, 58% (95% CI 49%-66%) had recurrent BV (Nugent 7-10), and 69% (95% CI 61%-77%) had recurrent abnormal flora (Nugent 4-10). Past BV, the same regular sex partner throughout follow-up, and female sex partners predicted recurrence; hormonal contraception was protective in multivariate analysis.

More than half of treated women are back in a BV state within a year. Recurrence clusters with ongoing sexual exposure to the same partner. That finding sat uneasily next to older trials that failed to help by treating men, until StepUp changed what partner treatment meant. CDC notes that persistent or recurrent BV is common. For multiple recurrences, options include twice-weekly metronidazole gel for more than three months, multi-step regimens with boric acid, and investigational approaches. Suppression helps while used; benefit often fades when it stops.

The partner-treatment trial

In March 2025, Vodstrcil, Bradshaw and the StepUp team published an open-label randomized trial in the New England Journal of Medicine that reframed the partner question. Couples enrolled when a woman had BV and was monogamous with a male partner. In the partner-treatment arm, she received first-line therapy and he received metronidazole 400 mg tablets plus 2% clindamycin cream on penile skin, both twice daily for 7 days. Controls treated only the woman.

The data and safety monitoring board stopped the trial after 150 couples completed 12-week follow-up because woman-only care was inferior. In the modified intention-to-treat population, recurrence hit 24 of 69 women (35%) with partner treatment (1.6 recurrences per person-year) versus 43 of 68 (63%) with standard care (4.2 per person-year). Absolute risk difference −2.6 per person-year (95% CI −4.0 to −1.2; P < 0.001). Treated men reported nausea and headache. Metallic taste was also common.

Older partner trials often used oral antibiotics alone for men. StepUp combined systemic and topical therapy aimed at penile skin microbiota, enrolled monogamous couples, and measured recurrence over 12 weeks. That is strong evidence for dual oral-topical male-partner therapy in stable heterosexual relationships. It is not a claim that every male partner worldwide needs treatment. ACOG moved on the data: an October 2025 Clinical Practice Update in Obstetrics & Gynecology supports concurrent sexual partner therapy to prevent BV recurrence, updating Practice Bulletin No. 215. That shifts the older CDC line that routine partner treatment is not recommended, written before StepUp and still on some public pages. StepUp was open-label and Australian. It enrolled only monogamous male partners, so female partners of women who have sex with women and non-monogamous networks need more data. Inside the trial's box, treating both partners cut 12-week recurrence nearly in half.

Preterm birth and STI risk

On preterm birth, Leitich and colleagues' 2003 meta-analysis in the American Journal of Obstetrics and Gynecology still anchors most textbook statements. Across 18 studies and 20,232 patients, BV more than doubled the odds of preterm delivery (OR 2.19; 95% CI 1.54-3.12). Risk was higher with early detection: OR 7.55 (95% CI 1.80-31.65) before 16 weeks, and OR 4.20 (95% CI 2.11-8.39) before 20 weeks. BV also raised odds of spontaneous abortion (OR 9.91) and maternal infection (OR 2.53). CDC materials state the clinical translation: untreated BV in pregnancy raises the chance of preterm birth and low birth weight (under 5.5 pounds).

Association is not a proven screening program. CDC recommends treating symptomatic BV in pregnancy but does not recommend routine screening of asymptomatic pregnant women solely to prevent preterm birth. Treatment trials in asymptomatic groups have been mixed. The data are stronger on risk than on treating every positive Nugent score on a routine swab.

On STI acquisition, Atashili and colleagues' 2008 meta-analysis in AIDS (more than 30,000 women) found BV associated with a 1.6-fold increased risk of HIV acquisition in incidence studies (RR 1.6; 95% CI 1.2-2.1). CDC guidelines list elevated risk for HIV, gonorrhea, chlamydia, trichomoniasis, Mycoplasma genitalium, HPV, and HSV-2, plus increased HIV transmission to male partners. Those links sit beside our sexual health and STI statistics and the overlap with recurrent urinary tract infections in women.

Outcome	Measure	Source
Preterm delivery (any gestation at BV diagnosis)	OR 2.19 (95% CI 1.54-3.12)	Leitich et al., AJOG, 2003
Preterm delivery (BV diagnosed <16 weeks)	OR 7.55 (95% CI 1.80-31.65)	Leitich et al., AJOG, 2003
HIV acquisition (incidence studies)	RR 1.6 (95% CI 1.2-2.1)	Atashili et al., AIDS, 2008
Global annual cost, symptomatic BV treatment	US $4.8 billion (95% CI $3.7-$6.1B)	Peebles et al., 2019

Sources: Leitich et al., American Journal of Obstetrics and Gynecology, 2003; Atashili et al., AIDS, 2008; Peebles et al., Sexually Transmitted Diseases, 2019.

How accurate is diagnosis

Amsel clinical criteria and Nugent Gram-stain scoring dominate BV diagnosis. Amsel needs at least three of four findings. Those are homogeneous thin discharge, clue cells on wet mount, vaginal pH greater than 4.5, and a fishy amine odor before or after 10% KOH. Nugent grades Gram stains 0-10; 0-3 is Lactobacillus-dominant, 4-6 intermediate, and 7-10 BV. CDC treats Nugent as the laboratory reference standard.

Compared with Nugent, Amsel sensitivity is 37%-70% and specificity 94%-99%, per CDC STI Treatment Guidelines. High specificity means a positive Amsel workup is usually real BV; modest, variable sensitivity means a negative office exam does not fully rule it out. For symptomatic women, FDA-cleared molecular assays perform better: BD Max Vaginal Panel 90.5% sensitivity and 85.8% specificity; Aptima BV 95.0%-97.3% sensitivity and 85.8%-89.6% specificity. CDC advises using these in symptomatic patients only. Culture for Gardnerella vaginalis alone is not recommended, and Pap tests are not diagnostic for BV.

What treatment actually does

First-line therapy remains metronidazole or clindamycin, oral or topical. CDC notes no head-to-head proof that oral beats topical in nonpregnant women. Alternative agents include secnidazole 2 g oral granules once; a phase 3 trial cited by CDC found clinical cure at days 21-30 of 53% versus 19% with placebo. Convenience helps adherence; single-dose products have not erased recurrence.

On the biotherapeutic side, Cohen and colleagues' 2020 randomized trial of Lactobacillus crispatus CTV-05 (Lactin-V) after vaginal metronidazole, in the New England Journal of Medicine, found 12-week recurrence in 30% on Lactin-V versus 45% on placebo (risk ratio 0.66; 95% CI 0.44-0.87; P = 0.01). It is not FDA-cleared for commercial use as of this writing. That is evidence that restoring protective lactobacilli can move recurrence, not a prescription available tomorrow.

For patients sorting cycles from infectious symptoms during recurrent vaginitis workups, our period calculator, ovulation calculator, and late period tool can help; discharge still needs a wet mount. The honest summary is blunt. BV is extremely common and often silent. It recurs frequently, falls unevenly by race, and is linked to pregnancy and STI risk. For monogamous heterosexual couples, StepUp cut 12-week recurrence from 63% to 35% by treating the male partner. That is the number worth lifting into the next story, with the New England Journal of Medicine named next to it.

Frequently asked questions about bacterial vaginosis statistics

What percentage of U.S. women have bacterial vaginosis?

29.2% of women ages 14-49 met Nugent criteria for BV in NHANES 2001-2004 (about 21 million women), per Koumans and CDC colleagues. Globally, WHO and Peebles place prevalence among reproductive-age women at 23-29%. No newer nationally representative U.S. Gram-stain survey has replaced the NHANES estimate.

How often does BV come back after treatment?

In Bradshaw and colleagues' 12-month cohort, 58% of women had recurrent BV (Nugent 7-10) and 69% had recurrent abnormal flora after oral metronidazole. CDC describes persistent and recurrent disease as common and outlines suppressive metronidazole gel for multiple recurrences. Short-term cure rates overstate long-term control.

Did treating male partners reduce recurrence?

Yes. In the 2025 StepUp trial in the New England Journal of Medicine, dual oral and topical therapy for male partners cut 12-week BV recurrence to 35%, versus 63% with woman-only care. ACOG's 2025 Clinical Practice Update supports concurrent partner therapy for appropriate patients.

Does bacterial vaginosis increase preterm birth risk?

Leitich and colleagues' 2003 meta-analysis found BV more than doubled the odds of preterm delivery (OR 2.19), with higher risk when BV was found before 16 weeks (OR 7.55). CDC recommends treating symptomatic BV in pregnancy; routine screening of asymptomatic pregnant women solely to prevent preterm birth is not recommended.

How accurate are office tests for BV?

Amsel criteria are 37%-70% sensitive and 94%-99% specific versus Nugent Gram stain, per CDC. FDA-cleared molecular assays in symptomatic women perform better: BD Max Vaginal Panel about 90.5% sensitive and 85.8% specific; Aptima BV about 95-97% sensitive and 86-90% specific. Microscopy and pH remain useful for rapid office diagnosis.

Is BV a sexually transmitted infection?

BV is not classified as a classic STI like chlamydia or gonorrhea, but sexual activity strongly shapes risk and recurrence. Women who have never had sex are rarely affected, per CDC. StepUp showed male-partner treatment reduces female recurrence. Sexual exchange of BV-associated organisms is clinically meaningful even if taxonomy remains debated.