American Journal of Epidemiology Copyright © 1992 by The Johns Hopkins University School of Hygiene and Public Health
Vol 135, No 1 Printed in U S.A
AD rights reserved
Chlamydia trachomatis Infection in Women: A Need for Universal Screening in High Prevalence Populations?
Hillard S. Weinstock,1 Gail A. Bolan,2 Robert Kohn,2 Carlos Balladares,2 Arthur Back,2 and Geraldine Oliva2
Chlamydia trachomatis is the most prevalent sexually transmitted bacterial pathogen. Nevertheless, selective, rather than universal, screening for chlamydia has been recommended, largely because testing is expensive and requires considerable technical expertise. A total of 1,348 women in four family planning clinics in San Francisco, California, were screened from March 1987 to January 1988 to identify criteria for selective screening. Of these, 9.2% had a positive chlamydia test using direct fluorescence. Logistic regression analysis identified five factors associated with infection: age less than 25 years, cervical friability, single marital status, a new sexual partner within the past 3 months, and lack of barrier contraceptive use. No single risk factor or combination of risk factors had both a high sensitivity and a high positive predictive value for infection. While screening all women who were unmarried would detect 93% of those with chlamydia, the positive predictive value of 10.7% was not much higher than the overall prevalence. Conversely, screening all women with cervical friability, which had a positive predictive value of 23.2%, would only detect 1 1 % of those with chlamydia. On the basis of the authors' findings, selective screening should not be used in high prevalence populations in which all women are at risk and should be screened for chlamydia. Am J Epidemiol 1992;135:41-7. chlamydia; risk factors; sensitivity and specificity (epidemiology); sexually transmitted diseases
Chlamydia trachomatis is the most prevalent sexually transmitted bacterial patho-
gen in the United States today. An estimated three to four million men, women, and infants have chlamydial infections each year (1). Women, however, suffer most from the consequences of infection: pelvic inflammatory disease, tubal damage, and adverse outcomes of pregnancy (2). The direct and indirect costs of chlamydial infections have been estimated to be approximately 1.5 billion dollars annually (3). Most chlamydial infections among women are asymptomatic, and, for this reason, screening of sexually active women has been recommended (1,4). Universal screening has not been implemented, however, because cultures for chlamydia are expensive and require considerable technical expertise. Antigen detection methods for C. trachomatis have made screening more practical (5, 6). While these methods are usually less
Recerved for publication April 1, 1991, and in final form September 12, 1991. 1 Division of STD/HIV Prevention, Centers for Disease Control, Atlanta, GA. 2 San Francisco Department of Public Health, San Francisco, CA. Reprint requests to Technical Information Services, National Center for Prevention Services, Centers for Disease Control, 1600 Clrfton Road, N.E., Mailstop E06, Atlanta, GA 30333. This study was supported in part by Centers for Disease Control grant 09H000037-18-0. The authors thank the staffs of the San Francisco District Health Centers for their assistance and Drs. E R. Alexander, Stuart Berman, WDIard Cates, Jr., Joel Greenspan, Arthur Reingold, and Robert Rotfs for their critical review of the manuscript and helpful suggestions. Presented in part at the Twenty-ninth Intersaence Conference on Antimicrobial Agents and Chemotherapy, Houston, TX, September 17-20, 1989, and at the Association of Reproductive Health Professionals Meeting, San Diego, CA, October 19-21, 1989.
41
42
Weinstock et a).
expensive and require less technical skill than do cultures, their costs still discourage widespread chlamydia screening, particularly in public health facilities. Consequently, selective, rather than universal, screening of sexually active women has been advocated as an essential component of chlamydia control programs (1). Several authors have determined criteria that would permit such selective screening (4, 7-12). In this study, we hoped to identify criteria for selective screening and to test the performance of criteria proposed by others in an ethnically heterogeneous population. MATERIALS AND METHODS Study population
Women between the ages of 13 and 50 years who were attending one of four family planning clinics between March 1987 and January 1988 for contraceptive counseling, routine annual examination, or the diagnosis and treatment of urogenital symptoms were consecutively screened. Patients who had taken antibiotics within the past 3 weeks or who had already been screened for chlamydia within the past year were excluded from the study. Those whose tests were inadequate (i.e., tests in which there were fewer than 20 epithelial cells obtained from the cervical specimen) were also excluded from the analysis. The four family planning clinics, located in racially diverse neighborhoods throughout the city of San Francisco, California, are administered by the San Francisco Department of Public Health and serve predominantly lower- and middle-income women. Patient evaluation and laboratory methods
In order to standardize data collection techniques, the patients were interviewed and examined by nurse practitioners or physicians who had undergone training. At all four family planning clinics, a standardized questionnaire was completed by each clini-
cian regarding each patient's sexual and medical history, symptoms, and findings on cervical examination. A pelvic examination was performed on each patient, and endocervical specimens were obtained with Dacron-tipped plastic swabs (Du Pont, Wilmington, Delaware). Mucopurulent cervical discharge was defined by the presence of pus on the endocervical swab. Cervical friability was defined by bleeding induced by insertion of the endocervical swab. The presence of Trichomonas vaginalis was diagnosed by the observation of motile organisms in a wet mount preparation. The direct smear fluorescent antibody test (Syva Micro-trak, Palo Alto, California) was used to detect the presence of chlamydia antigen. The specimens were processed and interpreted according to standard procedures at the San Francisco Department of Public Health Laboratory. The detection of 10 or more elementary bodies was considered a positive result. Specimens were regarded as inadequate if fewer than 20 epithelial cells were seen on the slide. Statistical methods
Categorical variables were examined using Pearson's chi-square. The two-tailed Fisher exact test was used when appropriate. Multivariate logistic regression analysis was performed using the SAS computer program (13). On the basis of the strength of the univariate association and results from previous studies (4, 7-12, 14), nine risk factors were chosen for inclusion in the logistic regression analysis: age less than 25 years, marital status, black race, cervical friability, mucopurulent cervical discharge, a new sexual partner within the past 3 months, Trichomonas on wet mount preparation, oral contraceptive use, and barrier contraceptive use. Information on all of these characteristics was available for 1,048 women (78 percent of the total study population). The chlamydia positivity rate in this subpopulation was 9.4 percent. A stepwise backward elimination process was used; only variables with p values of 0.05 or less were included in the
Chlamydia in High Prevalence Populations
final model. The sensitivity of a risk factor to detect chlamydial infection was defined as the probability of the risk factor being present among all patients with a positive chlamydia test. The positive predictive value of a risk factor for infection was defined as the proportion of positive tests among all women with the risk factor. RESULTS
The study population consisted of 1,348 of the 1,598 women (84.4 percent) seen in the family planning clinics during the study period. Of these 1,598 women, 115 patients had taken antibiotics within the past 3 weeks; 34 had been screened for chlamydia within the past year, and 101 had inadequate specimens collected for testing. Table 1 summarizes the demographic and clinical characteristics of the study population. Chlamydia was detected in 124 women, 9.2 percent of the total study population. Positivity rates within each clinic were as follows: 7.2 percent of the women in clinic A, 14.7 percent in clinic B, 9.6 percent in clinic C, and 8.8 percent in clinic D. Rates within the clinics remained constant during the 11-month period of the study. Age was inversely correlated with positivity rates for chlamydial infection: 27 percent for women aged 13-14 years (n = 15), 16 percent for women aged 15-19 (n = 319), 9 percent for women aged 20-24 (« = 500), 7 percent for women aged 25-29 (n = 275), and 3 percent for women aged 30 and over (« = 239). Univariate analysis identified the following factors as those statistically associated with chlamydial infection (table 2): age less than 25 years, black race, single marital status, attendance at family planning clinic B, cervical friability, mucopurulent cervical discharge, and having a new sexual partner within the past 3 months. When an analysis was performed of women who were asymptomatic, that is, without abdominal pain, vaginal discharge, or dysuria (63.3 percent of the total study population), the relative risks for these seven factors were similar.
43
TABLE 1. Characteristics of 1,348 women (mean age: 24.1 years) screened for Chlamydia in four family planning clinics, San Francisco, California, 1987-1988 Characteristic Clinic A B C D Race White Black Hispanic Asian Unmarried Genitourinary symptoms Cervical friability Mucopurulent cervical discharge Current contraceptive method Oral Barrier Other None New sexual partner in past 3 months
% 31.8 9.6 40.9 17.7 33.0 31.0 23.6 12.4 78.9 36.7 4.4 4.8 39.5 27.3 9.2 24.6 20.3
When patients were stratified by the clinics they attended, these factors remained associated with chlamydia. Logistic regression analysis determined four characteristics that were independently associated with increased risk of chlamydial cervical infection (table 3): age less than 25 years, cervical friability, single marital status, and a new sexual partner within the preceding 3 months; barrier contraceptive use was protective against chlamydial infection. Oral contraception was not associated with the detection of chlamydia in this population. No statistically significant interactions were present between age and contraceptive use or between age and race. The sensitivities and positive predictive values of the five independent risk factors for cervical chlamydial infection were determined (table 4). A program that selectively screens women in this population who have any one of these risk factors would have a sensitivity of 100 percent and a positive predictive value of 9.4 percent. This compares with the overall prevalence of chla-
44
Weinstock et al.
TABLE 2. Untvariate analysis of characteristics associated with chlamydlal infection in women from four family planning clinics, San Francisco, California, 1987-1988 Variable
Age
Vol 135, No 1 Printed in U S.A
AD rights reserved
Chlamydia trachomatis Infection in Women: A Need for Universal Screening in High Prevalence Populations?
Hillard S. Weinstock,1 Gail A. Bolan,2 Robert Kohn,2 Carlos Balladares,2 Arthur Back,2 and Geraldine Oliva2
Chlamydia trachomatis is the most prevalent sexually transmitted bacterial pathogen. Nevertheless, selective, rather than universal, screening for chlamydia has been recommended, largely because testing is expensive and requires considerable technical expertise. A total of 1,348 women in four family planning clinics in San Francisco, California, were screened from March 1987 to January 1988 to identify criteria for selective screening. Of these, 9.2% had a positive chlamydia test using direct fluorescence. Logistic regression analysis identified five factors associated with infection: age less than 25 years, cervical friability, single marital status, a new sexual partner within the past 3 months, and lack of barrier contraceptive use. No single risk factor or combination of risk factors had both a high sensitivity and a high positive predictive value for infection. While screening all women who were unmarried would detect 93% of those with chlamydia, the positive predictive value of 10.7% was not much higher than the overall prevalence. Conversely, screening all women with cervical friability, which had a positive predictive value of 23.2%, would only detect 1 1 % of those with chlamydia. On the basis of the authors' findings, selective screening should not be used in high prevalence populations in which all women are at risk and should be screened for chlamydia. Am J Epidemiol 1992;135:41-7. chlamydia; risk factors; sensitivity and specificity (epidemiology); sexually transmitted diseases
Chlamydia trachomatis is the most prevalent sexually transmitted bacterial patho-
gen in the United States today. An estimated three to four million men, women, and infants have chlamydial infections each year (1). Women, however, suffer most from the consequences of infection: pelvic inflammatory disease, tubal damage, and adverse outcomes of pregnancy (2). The direct and indirect costs of chlamydial infections have been estimated to be approximately 1.5 billion dollars annually (3). Most chlamydial infections among women are asymptomatic, and, for this reason, screening of sexually active women has been recommended (1,4). Universal screening has not been implemented, however, because cultures for chlamydia are expensive and require considerable technical expertise. Antigen detection methods for C. trachomatis have made screening more practical (5, 6). While these methods are usually less
Recerved for publication April 1, 1991, and in final form September 12, 1991. 1 Division of STD/HIV Prevention, Centers for Disease Control, Atlanta, GA. 2 San Francisco Department of Public Health, San Francisco, CA. Reprint requests to Technical Information Services, National Center for Prevention Services, Centers for Disease Control, 1600 Clrfton Road, N.E., Mailstop E06, Atlanta, GA 30333. This study was supported in part by Centers for Disease Control grant 09H000037-18-0. The authors thank the staffs of the San Francisco District Health Centers for their assistance and Drs. E R. Alexander, Stuart Berman, WDIard Cates, Jr., Joel Greenspan, Arthur Reingold, and Robert Rotfs for their critical review of the manuscript and helpful suggestions. Presented in part at the Twenty-ninth Intersaence Conference on Antimicrobial Agents and Chemotherapy, Houston, TX, September 17-20, 1989, and at the Association of Reproductive Health Professionals Meeting, San Diego, CA, October 19-21, 1989.
41
42
Weinstock et a).
expensive and require less technical skill than do cultures, their costs still discourage widespread chlamydia screening, particularly in public health facilities. Consequently, selective, rather than universal, screening of sexually active women has been advocated as an essential component of chlamydia control programs (1). Several authors have determined criteria that would permit such selective screening (4, 7-12). In this study, we hoped to identify criteria for selective screening and to test the performance of criteria proposed by others in an ethnically heterogeneous population. MATERIALS AND METHODS Study population
Women between the ages of 13 and 50 years who were attending one of four family planning clinics between March 1987 and January 1988 for contraceptive counseling, routine annual examination, or the diagnosis and treatment of urogenital symptoms were consecutively screened. Patients who had taken antibiotics within the past 3 weeks or who had already been screened for chlamydia within the past year were excluded from the study. Those whose tests were inadequate (i.e., tests in which there were fewer than 20 epithelial cells obtained from the cervical specimen) were also excluded from the analysis. The four family planning clinics, located in racially diverse neighborhoods throughout the city of San Francisco, California, are administered by the San Francisco Department of Public Health and serve predominantly lower- and middle-income women. Patient evaluation and laboratory methods
In order to standardize data collection techniques, the patients were interviewed and examined by nurse practitioners or physicians who had undergone training. At all four family planning clinics, a standardized questionnaire was completed by each clini-
cian regarding each patient's sexual and medical history, symptoms, and findings on cervical examination. A pelvic examination was performed on each patient, and endocervical specimens were obtained with Dacron-tipped plastic swabs (Du Pont, Wilmington, Delaware). Mucopurulent cervical discharge was defined by the presence of pus on the endocervical swab. Cervical friability was defined by bleeding induced by insertion of the endocervical swab. The presence of Trichomonas vaginalis was diagnosed by the observation of motile organisms in a wet mount preparation. The direct smear fluorescent antibody test (Syva Micro-trak, Palo Alto, California) was used to detect the presence of chlamydia antigen. The specimens were processed and interpreted according to standard procedures at the San Francisco Department of Public Health Laboratory. The detection of 10 or more elementary bodies was considered a positive result. Specimens were regarded as inadequate if fewer than 20 epithelial cells were seen on the slide. Statistical methods
Categorical variables were examined using Pearson's chi-square. The two-tailed Fisher exact test was used when appropriate. Multivariate logistic regression analysis was performed using the SAS computer program (13). On the basis of the strength of the univariate association and results from previous studies (4, 7-12, 14), nine risk factors were chosen for inclusion in the logistic regression analysis: age less than 25 years, marital status, black race, cervical friability, mucopurulent cervical discharge, a new sexual partner within the past 3 months, Trichomonas on wet mount preparation, oral contraceptive use, and barrier contraceptive use. Information on all of these characteristics was available for 1,048 women (78 percent of the total study population). The chlamydia positivity rate in this subpopulation was 9.4 percent. A stepwise backward elimination process was used; only variables with p values of 0.05 or less were included in the
Chlamydia in High Prevalence Populations
final model. The sensitivity of a risk factor to detect chlamydial infection was defined as the probability of the risk factor being present among all patients with a positive chlamydia test. The positive predictive value of a risk factor for infection was defined as the proportion of positive tests among all women with the risk factor. RESULTS
The study population consisted of 1,348 of the 1,598 women (84.4 percent) seen in the family planning clinics during the study period. Of these 1,598 women, 115 patients had taken antibiotics within the past 3 weeks; 34 had been screened for chlamydia within the past year, and 101 had inadequate specimens collected for testing. Table 1 summarizes the demographic and clinical characteristics of the study population. Chlamydia was detected in 124 women, 9.2 percent of the total study population. Positivity rates within each clinic were as follows: 7.2 percent of the women in clinic A, 14.7 percent in clinic B, 9.6 percent in clinic C, and 8.8 percent in clinic D. Rates within the clinics remained constant during the 11-month period of the study. Age was inversely correlated with positivity rates for chlamydial infection: 27 percent for women aged 13-14 years (n = 15), 16 percent for women aged 15-19 (n = 319), 9 percent for women aged 20-24 (« = 500), 7 percent for women aged 25-29 (n = 275), and 3 percent for women aged 30 and over (« = 239). Univariate analysis identified the following factors as those statistically associated with chlamydial infection (table 2): age less than 25 years, black race, single marital status, attendance at family planning clinic B, cervical friability, mucopurulent cervical discharge, and having a new sexual partner within the past 3 months. When an analysis was performed of women who were asymptomatic, that is, without abdominal pain, vaginal discharge, or dysuria (63.3 percent of the total study population), the relative risks for these seven factors were similar.
43
TABLE 1. Characteristics of 1,348 women (mean age: 24.1 years) screened for Chlamydia in four family planning clinics, San Francisco, California, 1987-1988 Characteristic Clinic A B C D Race White Black Hispanic Asian Unmarried Genitourinary symptoms Cervical friability Mucopurulent cervical discharge Current contraceptive method Oral Barrier Other None New sexual partner in past 3 months
% 31.8 9.6 40.9 17.7 33.0 31.0 23.6 12.4 78.9 36.7 4.4 4.8 39.5 27.3 9.2 24.6 20.3
When patients were stratified by the clinics they attended, these factors remained associated with chlamydia. Logistic regression analysis determined four characteristics that were independently associated with increased risk of chlamydial cervical infection (table 3): age less than 25 years, cervical friability, single marital status, and a new sexual partner within the preceding 3 months; barrier contraceptive use was protective against chlamydial infection. Oral contraception was not associated with the detection of chlamydia in this population. No statistically significant interactions were present between age and contraceptive use or between age and race. The sensitivities and positive predictive values of the five independent risk factors for cervical chlamydial infection were determined (table 4). A program that selectively screens women in this population who have any one of these risk factors would have a sensitivity of 100 percent and a positive predictive value of 9.4 percent. This compares with the overall prevalence of chla-
44
Weinstock et al.
TABLE 2. Untvariate analysis of characteristics associated with chlamydlal infection in women from four family planning clinics, San Francisco, California, 1987-1988 Variable
Age
