Bacterial Vaginosis: Why Does It Keep Coming Back Despite Antibiotic Treatment?

Table of Contents

• Key Points
• What Is Bacterial Vaginosis (BV)?
• Why This Research Matters
• Inside the Study: How the Research Was Conducted
• Key Findings: What the Study Revealed
• Why Have Cure Rates Declined Over Time?
• Resistance, Reinfection, or Both?
• What This Means for Patients
• Study Limitations
• Recommendations for Patients and Researchers
• Frequently Asked Questions
• Source Information

Key Points

• Cure rates for bacterial vaginosis have declined from about 90% to 50–80% since metronidazole was introduced.
• A 14-day metronidazole course cured 62% at one week versus 43% for 7 days, but the benefit disappeared by day 21.
• Even after a full 14-day antibiotic course, 38% of women failed treatment at one week and 57% by three weeks.
• Women who used condoms or abstained from sex had a 50% higher cure rate; avoiding douching raised cure rates by 80%.
• Adding azithromycin to metronidazole provided little added benefit, so the solution likely requires better understanding of BV causes.

What Is Bacterial Vaginosis (BV)?

Bacterial vaginosis (BV) is a condition in which the normal balance of bacteria in the vagina is disrupted. Instead of being dominated by healthy Lactobacillus bacteria, the vagina becomes overgrown with a mixture of other organisms, including Gardnerella vaginalis and various anaerobic bacteria (bacteria that thrive in environments without oxygen).

BV is remarkably common. At least 15% of sexually active women have BV at any given time. To put that number in perspective, this makes BV 3–4 times more common than urinary tract infections, and it is far more common than Trichomonas vaginalis infection (a sexually transmitted parasite) — even among patients visiting sexually transmitted disease clinics. BV is also much more common than vulvovaginal candidiasis (a yeast infection).

One of the key challenges in understanding BV is that scientists have not been able to identify a single bacterium that causes it. DNA technology has revealed the presence of a large number of bacteria that cannot be cultured (grown in a laboratory) or are very difficult to grow. No single organism has been proven to be the definitive cause. One leading theory suggests that the destabilization of the healthy Lactobacillus species is the crucial event, and that G. vaginalis and the anaerobic bacteria are simply secondary invaders that move in after the protective bacteria are gone.

Why This Research Matters

BV is not just an inconvenience or a nuisance discharge. It is linked to a wide range of serious upper genital tract infections, which is why successful treatment is so important. The complications associated with BV include:

• Amniotic fluid infection (infection of the fluid surrounding the baby during pregnancy)
• Chorioamnionitis (infection of the fetal membranes)
• Preterm delivery (giving birth before 37 weeks)
• Postpartum endometritis (infection of the uterine lining after childbirth)
• Post-Cesarean delivery wound infections
• Post-hysterectomy infections (infections after removal of the uterus)
• Post-abortion endometritis (uterine infection following an abortion)
• Pelvic inflammatory disease (infection of the female reproductive organs), though this is linked to a lesser degree

These serious consequences make research into BV treatment a compelling public health priority — especially because the infection often goes untreated in women who have no symptoms.

Yet, as Dr. Eschenbach points out, this article was published in 2006 — 51 years after Dr. Gardner first reported the condition we now call BV. Despite half a century of research, only limited and inconclusive evidence exists on three fundamental questions: whether BV is truly an infection, what microbial cause (or causes) are responsible, and what the most effective treatment actually is.

Inside the Study: How the Research Was Conducted

The study that this editorial reviews, conducted by Schwebke and Desmond, was a randomized, double-blind clinical trial — meaning neither the patients nor the doctors knew who was receiving which treatment. This design is considered the "gold standard" in medical research because it eliminates bias.

The researchers enrolled a large group of 420 women with symptomatic BV who returned for their first follow-up examination. The study examined three key questions about BV treatment:

1. Duration of therapy: Does taking the antibiotic for a longer period (14 days) work better than the standard shorter course (7 days)?
2. Definition of BV: Does it matter whether "cure" is defined by clinical symptoms or by laboratory Gram stain criteria (a method of staining and examining bacteria under a microscope)?
3. Reinfection or suboptimal resolution: Does BV come back after treatment because women are being reinfected by a sexual partner, or because the original infection was never fully cleared?

The women were randomly assigned to receive either a 7-day or a 14-day regimen of metronidazole (the most commonly used antibiotic for BV). Some women also received azithromycin, a second antibiotic, to test whether adding it would improve outcomes. Cure rates were measured at both 7 days and 21 days after the last dose of medication.

Key Findings: What the Study Revealed

The results were striking on several fronts. Let's walk through the key numbers:

Longer treatment produced better short-term results. One week after receiving the last dose of medication, the pooled cure rate was significantly higher among women who received the 14-day regimen of metronidazole than among those who received the 7-day regimen:

• 14-day regimen: 62% cure rate at 1 week after treatment
• 7-day regimen: 43% cure rate at 1 week after treatment

This difference was highly statistically significant (P<.001), meaning there is less than a 0.1% chance that this result was due to random chance. For patients, this means the longer antibiotic course genuinely did produce better short-term outcomes.

The advantage did not last. However, when cure rates were measured 21 days after the last dose of medication, the picture changed dramatically. The two groups now had similar cure rates:

• 14-day regimen: 43% cure rate at 21 days after treatment
• 7-day regimen: 51% cure rate at 21 days after treatment

In other words, the longer 14-day treatment produced a definite but temporary benefit. The women who took the longer course were more likely to be cured initially, but by three weeks after finishing the antibiotics, their advantage had evaporated. This raises an important question: was the temporary improvement due to undertreatment, antibiotic resistance, or reinfection — or some combination of all three?

A significant number of women failed treatment entirely. The 14-day metronidazole regimen failed to cure 38% of women at the 7-day follow-up visit. Two weeks later, the failure rate had climbed to 57%. This means that even with a full two-week course of the standard antibiotic, more than half of women eventually experienced a return of their BV.

Azithromycin did not help. The study conclusively demonstrated that adding azithromycin to the metronidazole treatment added little to the overall effect. The combination was not meaningfully better than metronidazole alone, so adding this second antibiotic is not an effective solution.

Gram stain vs. clinical criteria. The study also compared two different ways of defining "cure." Whether the researchers used clinical criteria (symptoms and physical exam) or Gram stain criteria (microscopic examination of vaginal bacteria), the 14-day regimen was consistently superior to the 7-day regimen at the early follow-up point. The definition of cure did not change the overall conclusion about treatment duration.

Why Have Cure Rates Declined Over Time?

One of the most puzzling observations highlighted in this editorial is that the reported cure rate for BV has actually decreased over the past three decades. When metronidazole was first introduced, antibiotics "cured" about 90% of cases. Today, cure rates range from 50% to 80%. The question is: how can cure rates steadily decline over 30 years?

Dr. Eschenbach offers several possible explanations:

1. Changing definitions: The criteria used to define BV have changed over the years, moving from clinical criteria to Gram stain–based criteria. The definition of "cure" has also varied between studies. However, the author notes this is probably a "theoretical" rather than a real explanation for the decline.
2. Increasing antibiotic resistance: The bacteria that cause BV — or the anaerobic bacteria responsible for the overgrowth — may have become relatively more resistant to metronidazole (the most common treatment) or to clindamycin (the other common treatment). This would explain why the same drugs that used to cure 90% of cases now cure far fewer.
3. Reinfection: The infection may be reintroduced back into the vagina after treatment as a result of sexual reinfection from a partner.

Importantly, the failure of 38% of women to be cured even after a 14-day course of metronidazole strongly suggests that resistance to therapy is playing a role. Further evidence comes from the finding that the women with the most severe BV (as measured by Gram stain criteria) responded most poorly to treatment. Direct testing of antibiotic resistance has been difficult, however, because scientists still have not definitively identified the microbe (or combination of microbes) that causes BV.

Resistance, Reinfection, or Both?

This editorial weighs the evidence for resistance versus reinfection, and the study provided important clues pointing to both possibilities.

Evidence for reinfection: The study found that women who abstained from sex or consistently used condoms had a 50% higher cure rate than those who did not. This is a powerful signal that sexual activity — and possibly reinfection from a partner — plays a role in treatment failure. Interestingly, there is also evidence from earlier research that BV occurs in concordance among lesbian couples, meaning if one partner has BV, the other is likely to have it too, further supporting the idea that it can be transmitted sexually.

Evidence related to douching: A curious finding was that women who did not douche had an 80% higher cure rate compared with those who did. This could be explained in two ways. First, douching may disturb the normal vaginal flora (the community of healthy bacteria), making it easier for BV-associated bacteria to thrive. Second, douching might reintroduce bacteria into the vagina if the same douching equipment is used repetitively, essentially self-reinfecting.

Dr. Eschenbach notes that to determine which explanation is correct, future treatment studies need to be designed so that no exposure occurs — neither to a sexual partner nor to douching — to eliminate, at least on a short-term basis, the possibility of reinfection.

What This Means for Patients

So what does all of this mean for women dealing with BV? There are several important takeaways.

First, the standard treatment may not be enough. The conclusion of this large, double-blind trial is that metronidazole — even when administered for 14 days — has limited effect on BV. A limited effect such as this would not be expected if BV were caused by a single bacterium that is susceptible to metronidazole. The infection appears to be more complex, likely involving multiple types of bacteria with varying degrees of antibiotic susceptibility.

Second, a longer course of antibiotics might help short-term but is not a cure-all. The 14-day regimen did produce significantly better cure rates at one week (62% vs. 43%), which could be helpful for women who need a quick resolution, such as those about to undergo a surgical procedure. However, by three weeks after treatment, the benefit had disappeared, meaning patients should not expect that simply taking antibiotics longer will solve the problem permanently.

Third, lifestyle factors may matter more than previously thought. The finding that condom use or sexual abstinence was associated with a 50% higher cure rate suggests that reinfection from a partner may be a real phenomenon. For women who experience recurrent BV, this raises the question of whether male partners should be evaluated or treated as well. The finding about douching — an 80% higher cure rate among women who did not douche — suggests that women with BV should strongly consider stopping douching to give their treatment the best chance of working.

Fourth, adding other antibiotics may not be the answer. The study showed that azithromycin added little to the effect of metronidazole. This means the solution to BV treatment failure probably won't come from simply stacking more antibiotics on top of the standard regimen. Instead, a more fundamental understanding of BV's cause is needed.

Study Limitations

The authors of the study appropriately pointed out one major limitation: the high number of subjects who dropped out of the study. However, Dr. Eschenbach notes that this dropout rate was unlikely to change the overall conclusions.

There are also broader limitations that apply to BV research in general. Because the microbial cause of BV has not been definitively established, direct testing of antibiotic resistance has been impossible. Researchers cannot yet culture every bacterium involved or test each one for susceptibility to antibiotics. The study also could not fully separate the effects of resistance from those of reinfection; it remains unclear from the data whether the temporary effect of the 14-day regimen was related to undertreatment, resistance, reinfection, or a combination of these factors.

Additionally, this was a study of women with symptomatic BV. The findings may not apply fully to the large number of women who have BV without symptoms, and the question of whether to treat asymptomatic women was not directly addressed by this trial.

Recommendations for Patients and Researchers

Based on the findings discussed in this editorial, here are practical recommendations for patients and the scientific community:

For patients with BV:

• Complete the full course of antibiotics exactly as prescribed, even if symptoms improve. The study shows that a longer course (14 days) may provide better short-term results than a 7-day course.
• Consider condom use or sexual abstinence during treatment. Women who abstained from sex or used condoms consistently had a 50% higher cure rate. This may help prevent reinfection.
• Avoid douching. Women who did not douche had an 80% higher cure rate. Douching may disturb healthy vaginal bacteria or reintroduce harmful bacteria. In fact, the American College of Obstetricians and Gynecologists already recommends against routine douching for all women.
• Talk to your doctor about persistent symptoms. If BV returns after treatment, you are not alone — this is a very common problem. Ask your provider whether a longer course of antibiotics might be appropriate for your situation, and be aware that even this approach has limitations.
• Do not expect that additional antibiotics like azithromycin will solve the problem. The study showed this combination added little benefit.

For researchers:

• Much additional and difficult work is required to define the causative bacteria of BV and determine their susceptibility to antibiotics other than metronidazole and azithromycin. The identification of the specific microbes responsible is critical to advancing treatment.
• Future treatment studies should be designed to eliminate exposure to both sexual partners and douching, at least in the short term, to separate the effects of reinfection from treatment failure.
• Basic research is needed on the role of sexual transmission and on the connection between BV and the disruption of healthy Lactobacillus bacteria. Almost everyone would have to conclude that further basic research is required on BV before much effect can be expected on the rather dismal cure rates for this common and important infection.

Frequently Asked Questions

What is bacterial vaginosis and how common is it?

Bacterial vaginosis is a disruption of the normal balance of bacteria in the vagina, with overgrowth of organisms like Gardnerella vaginalis. It is very common: at least 15% of sexually active women have it at any given time. It is more common than urinary tract infections, trichomoniasis, and yeast infections.

Why did the study compare 7-day and 14-day antibiotic treatment?

The study tested whether a longer 14-day course of metronidazole worked better than the standard 7-day course. Investigators wanted to see if a longer treatment could improve cure rates, which have declined over time. They also tested adding azithromycin to see if a second antibiotic helped, and it did not.

What were the cure rates for the 7-day versus 14-day metronidazole regimen?

One week after treatment, the 14-day regimen cured 62% of women, while the 7-day regimen cured 43%. However, at 21 days after treatment, the cure rates were similar: 43% for the 14-day group and 51% for the 7-day group. So the longer course only gave a temporary benefit.

Why does bacterial vaginosis keep coming back after antibiotics?

The editorial suggests several reasons: antibiotic resistance in the bacteria, reinfection from a sexual partner, or undertreatment. The infection likely involves a complex mix of bacteria that antibiotics cannot fully eliminate. Even after a 14-day metronidazole course, more than half of women experienced a return of BV by three weeks.

Does douching affect the chance of curing bacterial vaginosis?

Yes. Women who did not douche had an 80% higher cure rate compared with those who did. Douching may disturb healthy vaginal bacteria or reintroduce harmful bacteria. The article recommends avoiding douching during BV treatment, and the American College of Obstetricians and Gynecologists already advises against routine douching for all women.

Is adding another antibiotic like azithromycin helpful for BV?

No. The study found that adding azithromycin to metronidazole added little to the overall effect. The combination was not meaningfully better than metronidazole alone. The article concludes that simply stacking more antibiotics is not an effective solution, and a more fundamental understanding of BV's cause is needed.

Should I seek a second opinion if my bacterial vaginosis keeps coming back after antibiotic treatment?

Bacterial vaginosis has cure rates of only 50–80%, even with standard antibiotic treatment. A 14-day metronidazole course improved short-term cure to 62% versus 43% for 7 days, but by 21 days the advantage disappeared. Adding azithromycin did not help. Condom use or sexual abstinence and avoiding douching significantly improve cure rates. If BV persists or recurs despite a full antibiotic course, a second opinion can help assess whether the diagnosis is complete, whether lifestyle factors such as reinfection or douching need addressing, and whether a longer regimen or other strategies are appropriate. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: E D I T O R I A L C O M M E N TA R Y

Author: David A. Eschenbach, Department of Obstetrics and Gynecology, University of Washington, Seattle

Publication: Clinical Infectious Diseases 2007; 44:220–221, published by the Infectious Diseases Society of America. Received 8 September 2006; accepted 11 September 2006; electronically published 13 December 2006.

Accompanying study reviewed: Schwebke JR, Desmond RA. "A randomized trial of length of therapy with metronidazole plus or minus azithromycin for treatment of symptomatic bacterial vaginosis." Clinical Infectious Diseases 2006; 43:213–9.

The editorial also references the following prior research: Criswell et al. (1969) on inoculation of Haemophilus vaginalis; Marrazzo et al. (2002) on vaginal flora in women who have sex with women; Fredericks et al. (2005) on molecular identification of bacteria associated with BV; Joesoef et al. (1999) on review of BV treatment options; Eschenbach et al. (1983) on a dose-duration study of metronidazole; and Sanchez et al. (2004) on intravaginal metronidazole gel versus metronidazole plus nystatin ovules.

Note: This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes and should not replace personalized medical advice from a qualified healthcare provider. This article was written more than 15 years after the original publication; treatment recommendations for bacterial vaginosis may have evolved since that time.