CALLAHAN, DAVID B. MD*†; WEINBERG, MICHELLE MD*†; GUNN, ROBERT A. MD, MPH*†
PREVENTING PRETERM BIRTH (PTB; <37 weeks of gestation) remains a major public health goal. Approximately 300,000 preterm/low weight (<2500 g) births occur per year in the United States. PTB and its sequelae are the leading cause of death among black infants and the second leading cause of infant deaths overall in the United States. 1 Although survival of preterm babies has improved over the previous 15 years, rates of preterm birth have not improved. 2 Furthermore, there is marked racial disparity in rates of PTB: the rate among black women is twice that among white women, despite some narrowing of this disparity since 1993. 1
Genitourinary infections, including sexually transmitted diseases (STDs) due to Neisseria gonorrhoeae and Chlamydia trachomatis, 3 have been demonstrated to increase the risk of PTB, and screening for these infections has been promulgated as a standard part of prenatal care. 4 Bacterial vaginosis (BV) is also associated with PTB, and screening and treating pregnant women with a history of PTB for BV has been shown to decrease the risk of another PTB. The data are mixed for women in other risk subgroups and for screening asymptomatic pregnant women in general. 5,6 Although the efficacy of screening and treatment has been studied, little is known about providers’ practices concerning BV prevention (e.g., having patients avoid douching), diagnosis, screening, and treatment. 7
BV is common in pregnant women (prevalence of 12–20%3) and is associated with a relative risk for PTB of 1.2 to 3.0. 8 Recent studies have indicated that treating symptomatic BV early in pregnancy (before intrauterine infection is established) can substantially decrease the risk of PTB. 9 This research has led to the recommendation by the Centers for Disease Control and Prevention (CDC) that all pregnant, high-risk women (i.e., those with a prior PTB) and women with symptoms of vaginal infection be tested for BV and treated if BV is found. 10 Currently, the CDC has no recommendation for screening asymptomatic pregnant women for BV.
Diagnosis of BV depends on the physician's recognition of certain clinical signs that, taken together, establish the syndromic diagnosis. The presence of three of the following four signs fulfills Amsel's criteria for the diagnosis of BV: (1) increased vaginal pH (>4.5), (2) grayish-white homogenous discharge, (3) amine smell (with or without application of potassium hydroxide), and (4) clue cells (>20%) on examination of a wet-mount slide (“wet-mount”). Wet-mount is the classic bedside test for BV: it is generally familiar to physicians, can be performed rapidly, and has high sensitivity and specificity for the diagnosis of BV. 11 More recently, rapid-test kits for BV (similar to those used to test for pregnancy and strep throat) have been developed, but they are not yet widely used. Gram staining of a vaginal smear is also useful but is not commonly done in clinical practice. Wet-mount, in conjunction with the other Amsel criteria, remains the single most important diagnostic test for BV.
Metronidazole is an effective therapy for BV because of its efficacy against vaginal anaerobic bacteria that predominate in BV. In addition, metronidazole has little activity against the normal lactobacilli, thus helping to restore the normal vaginal flora. 12 Administration of oral metronidazole is the preferred therapy for BV in pregnant women because it not only relieves vaginal symptoms but, more importantly, also is effective against anaerobic infections in the amniotic membranes. However, concern about the use of metronidazole in pregnancy persists because of a theoretical risk of teratogenicity (on the basis of animal study findings), even though clinical studies have demonstrated it is safe. 13 Systemic effects might explain the finding that regimens including oral medications have been found to reduce PTB, whereas studies of vaginal clindamycin have not shown it to be effective 9,14,15; however, no direct comparison of systemic and topical medications for effectiveness in preventing PTB has been published. In general, physicians prefer topical medications to systemic medications in pregnancy because this limits the exposure of the fetus to the medication. However, for pregnant patients with BV, this desire to protect the fetus might actually increase the risk of PTB.
Understanding how physicians manage BV in pregnancy is important for the development of interventions aimed at improving physicians’ diagnosis and treatment practices for BV, which might affect the rates of preterm birth. In 1999, we surveyed physicians in San Diego County, California, who provided gynecologic and/or obstetric care to determine how they approached the screening, diagnosis, and treatment of BV for both symptomatic and asymptomatic, nonpregnant and pregnant patients.
The survey was mailed to all obstetrician/gynecologists (ob/gyns), family practitioners (FPs), general practitioners (GPs), and adolescent medicine physicians in San Diego County. The physicians’ names were obtained from a list of all licensed physicians in San Diego compiled by the San Diego County Medical Society. After initial responses to the mailing were received, nonresponding physicians practicing in the high-risk areas of San Diego County (defined as regions in central and south-central San Diego, which have the highest rates of STDs) who did not respond were contacted by telephone and asked to complete the survey, which was sent to them again in the mail or via fax. Physicians from non-high-risk areas who did not respond were sent a reminder notice approximately 2 to 4 weeks after the initial mailing.
Physicians were asked about their knowledge of BV complications and the association between BV and a variety of conditions/diseases, screening (disease in asymptomatic patients) and diagnostic (disease in symptomatic patients) practices for vaginal infections (including STDs and BV), and treatment of BV for both nonpregnant and pregnant patients. In addition, we asked about their treatment practices for BV by trimester of pregnancy. To assess physicians’ knowledge, we offered them a list of possible sequelae of BV and asked them to indicate if good evidence exists for a possible causal association with BV. For physicians who cared for pregnant patients, we assigned a vaginal infection management score based on five criteria (one point for each): (1) used appropriate diagnostic tests for BV (wet-mount, Gram stain, or rapid diagnostic kit); (2) asked about douching; (3) prescribed oral metronidazole in pregnancy; (4) treated BV in first trimester with systemic medications; and (5) screened pregnant patients for gonorrhea and/or chlamydia.
This score was used to compare reported physicians’ practices to several potential preventive interventions related to vaginal infections in pregnancy. We did not request information concerning the overall prevalence of BV in physicians’ practices but did ask what percentage of patients with symptoms or signs of vaginitis had nonspecific vaginitis (including that due to Gardnerella) or BV diagnosed.
Information regarding demographics, practice location, and medical training dates was also collected from physicians. Physicians who graduated from medical school >20 years ago (before 1979) were compared with more recent graduates to broadly assess changes in physicians’ training (regarding management of vaginal infections) over time. Physicians who reported that they did not provide gynecologic or obstetric care were excluded from the analysis. Data were analyzed with use of Epi Info version 6.04c (CDC, Atlanta, GA).
Surveys were mailed to 603 physicians, and 250 (41%) were returned; 42 surveys were excluded because the physicians did not provide gynecologic or obstetric care. Among ob/gyns, the response rate was 46% (88/193), which was higher than that for FPs/GPs (33%; 132/404). Our study group consisted of the 208 responding physicians who provided gynecologic or obstetric care: 102 (49%) of the study group provided care to pregnant patients. Physicians in the study group were predominantly male (67%) and in private practice (63%), and the majority had graduated from medical school before 1979 (55%). Among respondents, 48% (100/208) reported diagnosing BV or nonspecific vaginitis (including that due to Gardnerella) in one half of patients with vaginitis; another 28% of physicians diagnosed this in 25% of their patients with vaginitis, and 24% diagnosed it in fewer than 25% of their patients with vaginitis.
Overall, physicians were not aware of many of the various outcomes that are associated with BV (Table 1). Fewer than 60% knew that substantial evidence exists that BV causes PTB, pelvic inflammatory disease (PID), and postprocedural infections (after hysterectomy and abortion). Physicians who provided care for pregnant patients were more likely to believe that BV causes PTB than physicians who did not provide such care (74% [75/101] versus 38% [41/107]; RR = 1.9 [95% CI = 1.5–2.5]). Physicians who graduated from medical school after 1979 were more likely to believe that there is substantial evidence that BV plays a role in increased risk of PTB (67% versus 48%), that BV increased risk of HIV transmission (29% versus 19%), and that BV is associated with abnormal Papanicolaou smears (71% versus 56%). However, there was no difference in physicians’ beliefs about other adverse outcomes potentially related to BV (e.g., facilitation of acquisition of bacterial STDs, urinary tract infections, and postsurgical infections;Table 1).
To diagnose BV in both nonpregnant and pregnant symptomatic patients, wet-mount was the most widely used test (used by 73% for nonpregnant patients and by 66% for pregnant patients;Table 2), followed by vaginal culture (used by 18% for nonpregnant patients and 20% for pregnant patients). Physicians who did not use wet-mount to diagnose BV were not substituting other methods with a high sensitivity and specificity (e.g., single-use diagnostic cards [fewer than 5% ever used these other methods]). More ob/gyns than FPs reported using wet-mount (74% versus 53%; RR = 1.4 [95% CI = 0.9–2.1]). Physicians who believed that BV causes preterm delivery were more likely than those who did not to use wet-mount to diagnose BV in symptomatic pregnant patients (74% versus 42%). Few physicians screened for BV in asymptomatic pregnant patients: only 8% always performed wet-mount and another 19% sometimes performed wet-mount. In comparison, 53% of responders always screened asymptomatic pregnant patients for group B streptococcal (GBS) colonization, 62% always screened for chlamydia, and 60% always screened for gonorrhea.
Nearly all physicians (95%) prescribed either oral or intravaginal metronidazole to treat BV in nonpregnant patients. Of the 102 physicians providing obstetric care, 67 (61%) reported prescribing either oral metronidazole (56%) or oral clindamycin (18%) to treat BV in pregnancy. However, 45% of those physicians also reported prescribing intravaginal metronidazole for BV in pregnancy (Table 2). Only 12% of physicians prescribed metronidazole or clindamycin (in any formulation) for BV during the first trimester, including for symptomatic patients.
Physicians who believed that BV causes preterm delivery were more likely to prescribe oral metronidazole for treating BV in pregnant patients (68% versus 23%). They were also more likely to use intravaginal clindamycin for treating pregnant patients (47% versus 19%). No substantial difference was noted in the use of intravaginal metronidazole to treat BV in pregnant patients.
Although most physicians reported asking at least some or all patients about vaginal douching (75%; 156/208), only 15% (32/208) asked all patients about douching. Similarly, although 82% of physicians (170/208) counseled at least some patients about douching, only 12% (25/208) counseled all women on douching practices. Of physicians counseling patients about douching, 61% (104/170) advised them not to douche at all.
The vaginal-infection management score indicated that, overall, only 27% of physicians caring for pregnant patients (27/101) had a high vaginal-infection management score (of 4 or 5), and physicians who believed that BV causes PTB were much more likely to have a high score (43% versus 7%; RR = 6.4 [95% CI = 2.1–20.0]). Ob/gyns caring for pregnant patients were more likely to have a high score than were family physicians providing such care (31% versus 18%; RR = 1.7 [95% CI = 0.6–4.4]).
The survey results indicated that a substantial minority of physicians were not using recommended methods for diagnosis and treatment of BV. Although two thirds of physicians used wet-mount to diagnose BV in symptomatic pregnant patients, a substantial proportion (20%) used vaginal culture, which is not recommended because of its lack of sensitivity and specificity for Gardnerella vaginalis, the bacteria most commonly associated with BV. 8 Similarly, whereas the majority of physicians were treating BV with oral metronidazole (the preferred therapy), nearly one half were also prescribing a regimen that is not recommended: intravaginal clindamycin for treatment of BV in pregnant patients. 8 This finding is similar to a recent study of prenatal care providers in Georgia, where 38% of physicians reported use of intravaginal preparations to treat BV in pregnancy. 7
Only oral (versus intravaginal) regimens have been demonstrated to decrease PTB, presumably by treating both lower and upper reproductive tract infection. In fact, intravaginal clindamycin is associated with a slight increase in PTB. 16,17 However, at the time the data for our study were collected (1999), intravaginal clindamycin was still being marketed for treatment of BV in pregnancy, an indication for which it had FDA approval. Although we did not ask why physicians avoided metronidazole (as well as clindamycin) in the first trimester of pregnancy, the likely reasoning is a fear of teratogenicity and subsequent birth defects.
A report on the use of metronidazole described this fear as a “traditional” concern of many physicians but noted that other studies supported the safe use of metronidazole in pregnancy and advocated its use at the initial prenatal visit if it is needed. 18 The STD Treatment Guidelines of the CDC discuss concerns regarding the use of oral metronidazole in pregnancy but conclude that it is the preferred therapy for BV in pregnancy and recommend its use. Specifically, a long course of therapy (i.e., 250 mg three times daily for 7 days) is recommended, rather than a single-dose regimen.
Adoption of clinical guidelines by physicians is often slow, as was seen similarly with GBS disease. Guidelines for the prevention of neonatal GBS disease were issued in 1992. 19 A study of Canadian obstetric care providers indicated that 10% of providers adhered to the guidelines in 1994, and this percentage increased to 29% by 1997. 20 Although questions regarding screening for GBS in our survey were asked for the sake of comparison and were not a primary focus of the study, we found that only 53% of respondents reported screening all pregnant patients. Nearly half of respondents did not screen for GBS, indicating either poor compliance overall with this preventive intervention or reliance on prophylactic antibiotics (a choice provided by the CDC in the 1996 revision of the GBS prevention guidelines). 21 Adoption of guidelines by physicians for managing BV would be expected to follow a similar pattern, although educational interventions can speed the process. A crossover study of physicians’ practices for diagnosing and treating BV both before and after being provided clinical guidelines indicated that an increase (from 55% to 77%) in use of diagnostic procedures, as well as subsequent diagnosis of BV, occurred in the short-term (i.e., 4 weeks; longer-term data on physician behavior were not collected). 22
In our survey, good management of BV was more common among physicians who graduated from medical school after 1979. Recent medical school graduates were also more likely to believe that evidence exists that BV causes PTB. This finding might reflect the focus of the recent medical literature, which has highlighted the role of BV in adverse pregnancy outcomes. Professional education might help to improve knowledge of BV among physicians in practice for >20 years, resulting in improved diagnosis and management of BV.
Physicians who were aware of the association of BV with PTB were also more likely to treat BV with the most effective regimens (i.e., with systemic metronidazole). This treatment choice, the only one shown to be effective in the prevention of PTB in high-risk women when given as a prolonged regimen, might be attributed to an increased familiarity with BV as a disease entity in general or a desire by physicians to treat BV more aggressively if they perceive it as being a cause of more serious problems, such as PTB.
Prepregnancy treatment of BV is important because many women do not seek pregnancy care immediately, and in some cases, BV infection might become established in the fetal membranes in early pregnancy. This infection can lead to chorioamnionitis, a potentially dangerous infection that is not curable during pregnancy. 23 Prepregnancy treatment of BV is a concept that is consistent with other public health interventions aimed at promoting healthy pregnancies by targeting all women of reproductive age. Other examples include dietary recommendations for folate and for smoking cessation. 24 Identifying and treating BV in women before they become pregnant also precludes the issue of using metronidazole in the first trimester. In addition, because of the strong association of douching with BV and possibly PTB, physicians should ask patients about douching and provide information that discourages this activity. 25
In our study, the response rate of 41% limits the interpretation of our findings. Respondents possibly were more interested in the topic than those who did not respond; if so, this bias would tend to overestimate physicians’ knowledge of BV and other vaginal infections. Similarly, we made additional attempts to collect responses from physicians practicing in high-risk areas; however, the findings from those physicians were similar to those from the entire study group. Other characteristics of respondents’ and nonrespondents’ practices, such as patient volume (and thus time available to complete the survey), could have introduced bias, although these characteristics are difficult to estimate.
Reduction in the rates of BV in pregnant women and in women who might become pregnant could have a substantial public health effect. One observational study determined the population-attributable risk for PTB related to BV to be 6%. 3 The attributable risk for PTB related to BV might be as high as 40% for PTBs at <32 weeks of gestation among black women. 26 Another study estimated that annually in the United States, BV in pregnant women might lead to 4000 excess perinatal deaths and 4000 children with long-term neurologic sequelae (such as cerebral palsy and similar conditions) attributable to PTB. 27 BV might also play a more insidious role in morbidity by enhancing the transmission of HIV. A prospective study of pregnant women in Malawi, a country in which the HIV prevalence is high, demonstrated that BV is associated with acquisition of HIV infection. Greater disturbances of vaginal flora progressively increased HIV seroconversion both during pregnancy and postnatally. 28
Our study results suggest the need for a community-wide program both to increase physicians’ knowledge of BV and to encourage women to learn about BV. Women should be encouraged not only to seek diagnosis and treatment of symptoms that might indicate vaginal infections but also to avoid behaviors that might contribute to development of BV and other vaginal infections. Some of these behaviors (e.g., douching) might be culturally based, and the practice is passed on from mother to daughter. 29 Unsafe sexual behaviors are already a focus of HIV and STD prevention programs, but education discouraging unsafe sexual behaviors could also be integrated into a community health education campaign aimed at all vaginal infections. Changing such behaviors will require a sensitive approach and will need to involve trusted community members. Improving prevention and treatment of BV and other vaginal infections not only might affect the rate of PTBs but also could decrease HIV transmission.
However, changing physicians’ behavior as it relates to disease prevention is a notoriously difficult task. As noted in the Guide to Clinical Preventive Services of the U.S. Preventive Services Task Force, there are numerous barriers to implementing practice guidelines, particularly as they relate to preventive care. 30 Nonetheless, our survey suggests that physicians who were aware of the consequences of BV chose a more effective approach for diagnosing and treating vaginal infections in their pregnant patients. This finding indicates that educating physicians might be a good starting point for prevention of morbidity attributable to BV.
1. Peters KD, Lochanek KD, Murphy SL. Deaths: Final Data for 1996. National Vital Statistics Reports, Vol. 47, no. 9. Hyattsville, Maryland: National Center for Health Statistics, 1998.
2. Centers for Disease Control and Prevention. Preterm singleton births: United States, 1989–1996. MMWR Morb Mortal Wkly Rep 1999; 48: 185–189.
3. McGregor JA, French JI, Richter R, et al. Antenatal microbiologic and maternal risk factors associated with prematurity. Am J Obstet Gynecol 1990; 163: 1465–477.
4. American Academy of Pediatrics, American College of Obstetrics and Gynecology. Guidelines for Perinatal Care. 4th ed. Washington, DC: ACOG, 1997.
5. Hillier SL, Nugent RP, Eschenback DA, et al. Association between bacterial vaginosis and preterm delivery of low-birth-weight infants. N Engl J Med 1995; 333: 737–742.
6. Chandran R. Cochrane for clinicians: screening for and treating asymptomatic bacterial vaginosis in pregnancy. Am Fam Phys 2002; 66: 780.
7. Weisbord JS, Koumans EH, Toomey KE, Grayson C, Markowitz LE. Sexually transmitted diseases during pregnancy: screening, diagnosis and treatment practices among prenatal care providers in Georgia. South Med J 2001; 94: 47–53.
8. McGregor JA, French JI, Parker R, et al. Prevention of premature birth by screening and treatment for common genital tract infections: results of a prospective controlled evaluation. Am J Obstet Gynecol 1995; 173: 157–167.
9. Gibbs RS, Eschenback DA. Use of antibiotics to prevent preterm birth. Am J Obstet Gynecol 1997; 177: 375–380.
10. Centers for Disease Control and Prevention. 2002 Guidelines for treatment of sexually transmitted diseases. MMWR Morb Mortal Wkly Rep 2002; 51(RR-6): 42–44.
11. Schwebke JR. Diagnostic methods for bacterial vaginosis. Int J Gynecol 1999; 67: S21–S23.
12. McDonald HM, O'Laughlin JA, Fracog RV, Jolley PT. Bacterial vaginosis in pregnancy and efficacy of short-course oral metronidazole treatment: a randomized controlled trial. Obstet Gynecol 1994; 84: 343–348.
13. Caro-Paton T, Carvajal A, de Diego IM, et al. Is metronidazole teratogenic? A meta-analysis. Br J Clin Pharmacol 1997; 44: 179–182.
14. Kekki M, Kurki T, Pelkonen J, Kurkinen-Räty M, Cacciatore B, Paavonen J. Vaginal clindamycin in preventing preterm birth and peripartal infections in asymptomatic women with bacterial vaginosis: a randomized, controlled trial. Obstet Gynecol 2001; 97( 5 pt 1): 643–648.
15. Kurkinen-Räty M, Vuopala S, Koskela M, et al. A randomised controlled trial of vaginal clindamycin for early pregnancy bacterial vaginosis. BJOG 2000; 107: 1427–1432.
16. Joesoef MR, Hillier SL, Wiknjosastro G, et al. Intravaginal clindamycin treatment for bacterial vaginosis: effects on preterm delivery and low birth weight. Am J Obstet Gynecol 1995; 173: 1527–1531.
17. McGregor JA, French JI, Jones W, et al. Bacterial vaginosis is associated with prematurity and vaginal fluid mucinase and sialidase: results of a controlled trial of topical clindamycin cream. Am J Obstet Gynecol 1994; 170: 1048–1060.
18. McGregor JA, Roberts AM. First antenatal visits and metronidazole. Am J Obstet Gynecol 1997; 176: 887–888.
19. ACOG Technical Bulletin Number 170. Washington, DC: ACOG, 1992.
20. Davies HD, Adair CE, Schuchat A, Low DE, Sauve RS, McGeer A. Alberta Neonatal Group B Streptococcal Network and the Toronto Invasive Bacterial Diseases Network. Physicians’ prevention practices and incidence of neonatal group B streptococcal disease in 2 Canadian regions. CMAJ 2001; 164: 479–485.
21. Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease: a public health perspective. MMWR Morb Mortal Wkly Rep 1996; 45(RR-7): 1–24.
22. Langsford MJ, Dobbs FF, Morrison GM, Dance DA. The effect of introduction of a guideline on the management of vaginal discharge and in particular bacterial vaginosis in primary care. Fam Pract 2001; 3: 253–257.
23. Newton ER, Piper J, Peairs W. Bacterial vaginosis and intraamniotic infection. Am J Obstet Gynecol 1997; 176: 672–677.
24. Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong LC. Impact of folic acid fortification of the US food supply on the occurrence of neural tube defects. JAMA 2001; 285: 2981–2986.
25. Bruce FC, Fiscella K, Kendrick JS. Vaginal douching and preterm birth: an intriguing hypothesis. Med Hypoth 2000; 54: 448–452.
26. Goldenberg R, Iams J, Mercer B, et al. The preterm prediction study: the value of new vs standard risk factors in predicting early and all spontaneous preterm births. Am J Public Health 1998; 88: 233–238.
27. Goldenberg RL, Andrews MW, Yuan AC, MacKay HT, St. Louis ME. Sexually transmitted diseases and adverse outcomes of pregnancy. Clin Perinat 1997; 24: 23–41.
28. Taha TE, Hoover DR, Dallabetta GA, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 1998; 12: 1699–1706.
29. Sasikala R, Low N, Jones SB, Pozniak A. Bacterial vaginosis, ethnicity, and the use of genital cleaning agents: a case control study. Sex Transm Dis 1999; 26: 404–409.
30. U.S. Preventive Services Task Force. Guide to clinical preventive services. 2nd ed. Alexandria, Virginia: International Medical Publishing, 1996.