Colonization of the genitourinary tract by group B streptococcus (GBS) is common among pregnant women1–4 and is a major cause of perinatal morbidity and mortality in the United States.5,6 Group B streptococcus colonization is strongly associated with both early- and late-onset GBS sepsis, pneumonia, and meningitis in neonates7 and also is associated with several maternal complications, including chorioamnionitis and postpartum endometritis.8 Despite its importance, few epidemiologic studies have examined general risk factors for GBS colonization in pregnant women. Prior studies found associations between colonization and age, parity, race, tobacco use, education, and diabetes,9–13 but results were inconsistent and no investigation was population-based. Further, these studies do not examine occupational risk factors for colonization, such as working in health care. A greater understanding of the factors associated with colonization could aid efforts to provide antibiotic prophylaxis to women colonized with GBS or vaccines to high-risk women, reducing the overall incidence of infectious perinatal complications.6
Based on evidence for increased carriage rates of other bacteria, specifically methicillin-resistant Staphylococcus aureus, in health care workers14,15 we hypothesized that women in health care occupations may also have an increased prevalence of genital tract GBS colonization. To investigate this and other risk factors for GBS colonization, including race/ethnicity, body mass index (BMI), smoking, diabetes, and alcohol consumption, we conducted a population-based, case-control study among pregnant women using birth certificate and hospital discharge data.
PATIENTS AND METHODS
Subjects were identified from Washington State birth certificate data linked to hospital discharge data obtained from the Washington State Comprehensive Hospital Abstract Reporting System (CHARS) for births between 1997 and 2002. Only women giving birth to live singleton infants were included. Cases of GBS colonization were defined as either 1) the mother’s birth hospitalization discharge record containing the International Classification of Disease, Ninth Revision (ICD-9) codes 041.02 (GBS) or V02.51 (carrier or suspected carrier of GBS) or 2) the neonate’s birth discharge record containing the ICD-9 code 041.02 (GBS) and either 771.81 (septicemia of the newborn) or 771.83 (bacteremia of the newborn). Combining these 2 methods, 44,998 cases of GBS colonization were identified; only 434 of these were identified from neonatal birth discharge records. Because we could not differentiate between early and late neonatal GBS disease and these cases represented less that 1% of our sample, we elected to exclude these cases. After these exclusions, 44,564 cases of GBS colonization remained. Controls were then randomly selected in a 2:1 ratio from among noncases and were frequency-matched to cases by year of delivery. To avoid counting multiple observations for a single individual, only the earliest birth was used in the analysis for mothers with more than one birth during the study interval. The final sample included 40,459 cases and 84,268 controls. Patients with missing data were excluded from the multivariable analysis so as not to introduce bias from missing data imputation.
All variables, except where noted, were obtained from data on the Washington State Birth Certificate. The primary risk factors of interest were maternal occupation, race/ethnicity, tobacco and alcohol use during pregnancy, diabetes, and body mass index. Occupation was categorized a priori as non–health care worker, higher-risk health care worker, and lower-risk health care worker, a classification based upon degree and frequency of contact with patients. Higher-risk occupations included registered nurse (1.5% of our sample), licensed practical nurse (0.2%), physician (0.4%), nursing student (none in our sample), midwife (0.01%), osteopath (none), and naturopath (none). Lower-risk occupations were podiatrist (< 0.01%), chiropractor (0.02%), dentist (0.08%), dietician (0.08%), optometrist (0.02%), pharmacist (0.4%), psychiatrist/psychologist (0.2%), medical technician (1.3%), ultrasound/X-ray technician (0.1%), and physical/speech therapist (0.6%). Additional variables were categorized as follows: race/ethnicity as white, black, Hispanic, Native American, or other; tobacco use as nonsmokers, 1–10 cigarettes/day, or 10 cigarettes/day or more; alcohol use as none or any use during pregnancy; and diabetes as none, pregestational, or gestational. Body mass index (BMI, kg/m2) was calculated from the mother’s height obtained by linkage to the Washington State Department of Licensing and reported prepregnancy weight on the Washington State Birth Certificate. Body mass index was then categorized according to National Institutes of Health guidelines (underweight < 18.5, normal 18.5–24.9, overweight 25.0–29.9, obese 30.0–39.9, and morbidly obese ≥ 40).16
The mother’s age, parity, marital status, education level, access to prenatal health care, concurrent herpes or hepatitis, and household income were evaluated as potential confounders of the relationship between GBS colonization and risk factors. Education was reported as highest completed grade and categorized as no high school, some high school, and high school graduate. Income was calculated from the median annual income level of the mother’s census tract of residence and categorized into quintiles (first quintile being the lowest income). Access to health care was classified as adequate, intermediate, or inadequate using the Kotelchuck Adequacy of Prenatal Care Utilization Index.17
Multiple logistic regression was used to identify independent predictors of GBS colonization, with GBS status as the dependent variable. Initially, a base model was fitted, using all variables considered a priori to be either potentially associated with GBS colonization or potential confounders of other risk estimates. Variables not found to be significantly associated with GBS colonization and not found to confound other risk estimates (using a greater than 10% change in estimates approach) were removed from the model. Collinearity between variables was assessed with variance inflation factors, and none was greater than 1.5. Interaction terms that we suspected a priori to be significant (BMI and race/ethnicity, BMI and smoking, and smoking and race/ethnicity) were tested. Sensitivity analyses were performed by restriction to observations in the last 2 years of the study period, stratifying by year, and by restriction to only the 041.02 ICD-9 code. All tests were 2-tailed, and P < .05 was considered statistically significant. This investigation was conducted under the auspices of a data use agreement approved by the Internal Review Board of the University of Washington and the Washington State Department of Health.
A total of 44,564 cases of GBS colonization were identified over the 6-year period, from 413,728 live singleton births, for an overall prevalence of 10.8% (95% confidence interval [CI] 10.7–10.9%). Eighty-five percent of cases were identified by the V02.51 ICD-9 code on the maternal birth hospitalization discharge record. The prevalence of GBS colonization as defined in this study increased from the beginning to the end of the study period (Fig. 1). We excluded 4,105 cases from our analyses because they were not the earliest birth in our sample for an individual mother.
In univariate analyses, cases and controls were similar with regard to age, median annual income, parity, marital status, prenatal care use, and presence of diabetes (Table 1). Cases tended to be more highly educated, with 55.1% of cases having a high school diploma compared with 50.9% of controls. Cases and controls also differed with regard to their distribution of race/ethnicity, with a higher percentage of cases being white or black and more controls being Hispanic.
After fitting our base model, marital status, concurrent hepatitis, and concurrent herpes were not associated with GBS colonization and did not confound other associations; therefore, these variables were removed from the model. We included age and parity in our final model because of their relevance to obstetric epidemiology. The year of delivery was included in our model because subjects were matched on this variable. No significant effect modification was found between BMI and race/ethnicity, BMI and smoking, and smoking and race/ethnicity.
After adjustment for occupation, race/ethnicity, BMI, income, education, smoking, prenatal care use, alcohol use, diabetes, parity, age, and year of delivery, the following characteristics were independently associated with GBS colonization (Table 2): higher-risk health care occupation, black race, race categorized as “other,” overweight, obese, severely obese, median income in second quintile, median income in third quintile, median income in fourth quintile, median income in fifth quintile, some high school education, high school graduate, and adequate access to prenatal care. Hispanic ethnicity and smoking during pregnancy were associated with a significantly decreased odds of colonization. Diabetes (either pregestational or gestational), alcohol use, age, and parity were not associated with GBS colonization.
Analyses performed by restricting observations to the last 2 years of the study period and by stratification by year had no significant effect on odds ratios (ORs) for colonization (data not shown). When only cases identified by the 041.02 ICD-9 code were analyzed, the OR for colonization in the higher-risk health care worker group was 1.59 (95% CI 1.21–2.10).
Our study used a large population-based database to determine risk factors for GBS colonization among pregnant women. Our results demonstrate a modest association between GBS colonization and health care workers with a high frequency of direct patient contact (OR 1.22). This association was not found among female health care workers with minimal direct patient contact, suggesting that direct exposure to patients may be a mode of inoculation.
Nonsmoker status, black race, high BMI, higher income, greater educational attainment, and access to prenatal care were also associated with increased odds of GBS colonization in our study. These results are generally consistent with previous reports on the epidemiology of GBS colonization and neonatal GBS disease,18 although there are some important differences. The most comprehensive of these prior investigations used data from the Vaginal Infections and Prematurity Study.9 Our study confirms many of those findings in a larger population-based cohort but suggests that higher socioeconomic status, at least with regard to education and income, is associated with a greater odds of colonization. We believe this may be a true association and not just a reflection of increased screening for GBS among women of higher socioeconomic status because the finding persisted after controlling for adequacy of prenatal care. Other smaller studies examined risk factors for GBS colonization at single institutions.10–13 These studies generally found older age, greater parity, and diabetes to be associated with GBS colonization, although results were not consistent. Importantly, our study found no association between diabetes (either gestational or pregestational) and GBS colonization, suggesting that this may not be a significant risk factor as previously believed.
Interestingly, our results also show an increase in the prevalence of GBS colonization by study year, a finding most likely due to an increase in GBS screening during pregnancy. Since 1996, consensus guidelines from the American Academy of Pediatrics,19 the American College of Obstetricians and Gynecologists (ACOG),20,21 and the Centers for Disease Control and Prevention22,23 have advocated universal screening during pregnancy. As routine prenatal screening for GBS became more common, more cases were identified.
The underlying etiology of the association between GBS colonization and the risk factors that we identified is not clear. Currently, little is known about the biologic mechanism of colonization, which consequently limits the interpretation of its epidemiology. The natural reservoir for GBS is the gastrointestinal tract. Genitourinary colonization, therefore, may occur as a function of hygiene,24,25 sexual practice,26,27 or underlying immune system polymorphisms that reduce innate ability to eliminate the organism.28,29 Further, development of female GBS colonization, either nosocomially or in other nonhospital patient care settings, is suggested by the higher colonization rates in health care workers with frequent direct patient contact. Although serotypes of GBS that cause early neonatal disease are found in the maternal genital tract, serotypes that cause late onset GBS disease are less commonly represented in the genital tract of the mothers of these infants.30 Thus, it may be that infants with late onset GBS disease acquire the infection in the nursery or from other family members.31,32 If nosocomial transmission of GBS occurs, vaccination of health care workers might be important, both to protect the workers themselves and to limit the spread of GBS to susceptible patients. However, it is important to acknowledge that these findings are preliminary. Future research with longitudinal follow-up that examines women who are originally negative and then become positive is needed and may improve our understanding of the risk factors for colonization.
One important nonbiologic reason for our findings could be residual confounding by access to care or quality of care. It is possible that nonsmokers, black women, and women who are overweight, more highly educated, and in the upper quintile of income are more likely to receive prenatal care in a setting that more often provides GBS screening. Thus, they would be more likely to be found to be colonized. Similarly, Hispanic women may be screened less often and therefore less likely to be found to be colonized. Although we adjusted for the Kotelchuck Index as a measure of prenatal care use, it is possible that this index incompletely captured all of the effect of prenatal care.
In addition to residual confounding, other limitations of this study are noteworthy. First, because we used birth certificate and hospital discharge data for this study, both disease and exposure misclassification are possible. Identification of cases of GBS colonization was performed with ICD-9 codes rather than microbiological data. Almost certainly, there is inadequate case ascertainment by these methods in Washington State because the overall prevalence during the years of study was 10.8% and the estimated population prevalence is 15–20%.12 We have no way to differentiate “not colonized” from “not screened.” Additionally, data for our exposures and covariates of interest also were obtained from birth certificate data that are largely self-reported. However, although both disease and exposure misclassification may have occurred, it was likely nondifferential (ie, there is no reason to suspect differences in birth certificate or discharge data based upon presence or absence of risk factors) and would therefore be expected to attenuate the risk estimates. Thus, the associations that we did identify are likely underestimates of the actual risks. Second, these findings are associations and do not necessarily imply causality. The pathophysiologic mechanisms responsible for the observed associations are unknown.
Despite these limitations, this study provides new insight into the epidemiology of GBS colonization among pregnant women using a large population-based cohort. Although there are no specific data about the accuracy of reporting of GBS in hospital discharge records, the data sources for this study are generally considered reliable and provide information for a diverse population that includes all Washington State live births.33–36
Further research into the epidemiology of maternal GBS colonization is needed. Although universal screening and intrapartum antibiotic prophylaxis effectively reduce the incidence of GBS disease,6 screening can be incomplete and intrapartum antibiotics represent a significant burden on health care resources. In the future, new therapies, such as vaccines,37,38 may curtail the need for intrapartum antibiotics and further reduce morbidity and mortality related to GBS. Efforts to identify those at highest risk of colonization, and therefore most likely to benefit from these preventive therapies when culture results are not available, will be improved with greater understanding of the risk factors for GBS colonization.
1. Hoogkamp-Korstanje JA, Gerards LJ, Cats BP. Maternal carriage and neonatal acquisition of group B streptococci. J Infect Dis 1982;145:800–3.
2. Dillon HC Jr, Gray E, Pass MA, Gray BM. Anorectal and vaginal carriage of group B streptococci during pregnancy. J Infect Dis 1982;145:794–9.
3. Jones DE, Kanarek KS, Lim DV. Group B streptococcal colonization patterns in mothers and their infants. J Clin Microbiol 1984;20:438–40.
4. Dillon HC Jr, Khare S, Gray BM. Group B streptococcal carriage and disease: a 6-year prospective study. J Pediatr 1987;110:31–6.
5. Gibbs RS, Schrag S, Schuchat A. Perinatal infections due to group B streptococci. Obstet Gynecol 2004;104:1062–76.
6. Schrag SJ, Zywicki S, Farley MM, Reingold AL, Harrison LH, Lefkowitz LB, et al. Group B streptococcal disease in the era of intrapartum antibiotic prophylaxis. N Engl J Med 2000;342:15–20.
7. Benitz WE, Gould JB, Druzin ML. Risk factors for early-onset group B streptococcal sepsis: estimation of odds ratios by critical literature review. Pediatrics 1999;103:e77.
8. Krohn MA, Hillier SL, Baker CJ. Maternal peripartum complications associated with vaginal group B streptococci colonization. J Infect Dis 1999;179:1410-5.
9. Regan JA, Klebanoff MA, Nugent RP. The epidemiology of group B streptococcal colonization in pregnancy. Vaginal Infections and Prematurity Study Group. Obstet Gynecol 1991;77:604–10.
10. Baker CJ, Barrett FF, Yow MD. The influence of advancing gestation on group B streptococcal colonization in pregnant women. Am J Obstet Gynecol 1975;122:820–3.
11. Schauf V, Hlaing V. Group B streptococcal colonization in pregnancy. Obstet Gynecol 1976;47:719–21.
12. Anthony BF, Okada DM, Hobel CJ. Epidemiology of group B Streptococcus: longitudinal observations during pregnancy. J Infect Dis 1978;137:524–30.
13. Ramos E, Gaudier FL, Hearing LR, Del Valle GO, Jenkins S, Briones D. Group B streptococcus colonization in pregnant diabetic women. Obstet Gynecol 1997;89:257–60.
14. Cesur S, Cokca F. Nasal carriage of methicillin-resistant Staphylococcus aureus among hospital staff and outpatients. Infect Control Hosp Epidemiol 2004;25:169–71.
15. Eveillard M, Martin Y, Hidri N, Boussougant Y, Joly-Guillou ML. Carriage of methicillin-resistant Staphylococcus aureus among hospital employees: prevalence, duration, and transmission to households. Infect Control Hosp Epidemiol 2004;25:114–20.
16. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Obesity Education Initiative. National Institutes of Health. National Heart, Lung, and Blood Institute. Obes Res 1998;6 suppl 2:51S–209S.
17. Kotelchuck M. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health 1994;84:1414–20.
18. Schuchat A, Oxtoby M, Cochi S, Sikes RK, Hightower A, Plikaytis B, et al. Population-based risk factors for neonatal group B streptococcal disease: results of a cohort study in metropolitan Atlanta. J Infect Dis 1990;162:672–7.
19. Revised guidelines for prevention of early-onset group B streptococcal (GBS) infection. American Academy of Pediatrics Committee on Infectious Diseases and Committee on Fetus and Newborn. Pediatrics 1997;99:489–96.
20. ACOG committee opinion. Prevention of early-onset group B streptococcal disease in newborns. Number 173–June 1996. Committee on Obstetric Practice. American College of Obstetrics and Gynecologists. Int J Gynaecol Obstet 1996;54:197–205.
21. Prevention of early-onset group B streptococcal disease in newborns. ACOG Committee Opinion No. 279. Obstet Gynecol 2002;100:1405–12.
22. Prevention of perinatal group B streptococcal disease: a public health perspective [published erratum in MMWR Morb Mortal Wkly Rep 1996;45:679]. MMWR Recomm Rep 1996;45(RR-7):1–24.
23. Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of perinatal group B streptococcal disease. Revised guidelines from CDC. MMWR Recomm Rep 2002;51(RR-11):1–22.
24. Martino JL, Vermund SH. Vaginal douching: evidence for risks or benefits to women’s health. Epidemiol Rev 2002;24:109–24.
25. Ness RB, Hillier SL, Richter HE, Soper DE, Stamm C, McGregor J, et al. Douching in relation to bacterial vaginosis, lactobacilli, and facultative bacteria in the vagina. Obstet Gynecol 2002;100:765–772.
26. Manning SD, Tallman P, Baker CJ, Gillespie B, Marrs CF, Foxman B. Determinants of co-colonization with group B streptococcus among heterosexual college couples. Epidemiology 2002;13:533–9.
27. Meyn LA, Moore DM, Hillier SL, Krohn MA. Association of sexual activity with colonization and vaginal acquisition of group B Streptococcus in nonpregnant women. Am J Epidemiol 2002;155:949–57.
28. Abreu MT, Arditi M. Innate immunity and toll-like receptors: clinical implications of basic science research. J Pediatr 2004;144:421–9.
29. Zhang D, Zhang G, Hayden MS, Greenblatt MB, Bussey C, Flavell RA, et al. A toll-like receptor that prevents infection by uropathogenic bacteria. Science 2004;303:1522–6.
30. Davies HD, Adair C, McGeer A, et al. Antibodies to capsular polysaccharides of group B Streptococcus in pregnant Canadian women: relationship to colonization status and infection in the neonate. J Infect Dis 2001;184:285–91.
31. Takayanagi T, Tanaka H, Yoshinaga M, Aoki M, Teshima H, Fukuda M, et al. An outbreak of group B streptococcus infection in a neonatal nursery and subsequent trial for prophylaxis of nosocomial transmission. Acta Paediatr Jpn 1994;36:88–90.
32. Haase R, Nagel F, Hirsch W, Sitka U. Severe late-onset group B streptococcal infection: A case report [in German]. Z Geburtshilfe Neonatol 2003;207:186–9.
33. Lydon-Rochelle MT, Holt VL, Cardenas V, Nelson JC, Easterling TR, Gardella C, et al. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. Am J Obstet Gynecol 2005;193:125–34.
34. Parrish KM, Holt VL, Connell FA, Williams B, LoGerfo JP. Variations in the accuracy of obstetric procedures and diagnoses on birth records in Washington State, 1989. Am J Epidemiol 1993;138:119–27.
35. Ossiander EM, Emanuel I, O’Brien W, Malone K. Driver’s licenses as a source of data on height and weight. Econ Hum Biol 2004;2:219–27.
36. Dobie SA, Baldwin LM, Rosenblatt RA, Fordyce MA, Andrilla CH, Hart LG. How well do birth certificates describe the pregnancies they report? The Washington State experience with low-risk pregnancies. Matern Child Health J 1998;2:145–54.
37. Robbins JB, Schneerson R, Vann WF, Bryla DA, Fattom A. Prevention of systemic infections caused by group B streptococcus and Staphylococcus aureus by multivalent polysaccharide-protein conjugate vaccines. Ann N Y Acad Sci 1995;754:68–82.
38. Baker CJ, Paoletti LC, Wessels MR, Guttormsen HK, Rench MA, Hickman ME, et al. Safety and immunogenicity of capsular polysaccharide-tetanus toxoid conjugate vaccines for group B streptococcal types Ia and Ib. J Infect Dis 1999;179:142–50.