ADOLESCENT FEMALES have the highest rate of Chlamydia trachomatis infection. Adolescent behavioral risks for sexually transmitted diseases include multiple sex partners, inconsistent condom use, or reduced access to health care.1 Biologic factors that may increase susceptibility include cervical changes at puberty, including cervical ectopy, and lack of acquired immunity.
Cervical ectopy is a common physiologic process in adolescence (Figure 1), and also occurs during pregnancy and in response to hormonal contraceptive use.2 The mechanisms of cervical ectopy have been described elsewhere.2–4
Ectopy is thought to increase chlamydial risk by exposing a larger surface area of the target columnar epithelium to a potential infectious inoculum. Although the association between cervical ectopy and chlamydia has been extensively studied, nearly all previous work assessed ectopy by unaided visual assessment.5–9 This approach is unstandardized, nonquantitative, and subject to substantial observer bias because distinguishing between ectopy and immature metaplasia is difficult without acetic acid application or magnification.
Few previous studies have evaluated the relation between the extent of ectopy and chlamydia risk.7 Studies that have evaluated the impact of oral contraceptives (OC) and chlamydial infection have found inconsistent associations after stratifying for ectopy.5,6 Assessing these physiologic and behavioral issues is complex. Adolescent and adult females10 may have different biologic and behavioral responses to contraceptives, and the use and effect of hormonal contraceptives may vary by age and reproductive history. In addition, oral contraceptive formulations have changed over time. The specific effect of estrogen or progesterone on chlamydial infection has not been widely assessed.5
In this cross-sectional study, we hypothesized that in adolescents ectopy would be directly associated with chlamydial infection. In contrast to most previous studies, we quantified ectopy from cervical photographs using cervicography and computerized planimetry, a new standardized reproducible technique.11,12 Using these measurements, we compared prevalence of chlamydia in adolescents using combined estrogen-progestin OCs with those using an injectable progestin-only contraceptive, depot-medroxyprogesterone acetate (DMPA).
Eligibility and Recruitment
Subjects were recruited sequentially from January 1996 to July 1997 at the Baltimore City Sexually Transmitted Disease Clinics and the Adolescent Clinic at the University of Maryland at Baltimore. We screened all sexually active females between the ages of 11 to 20 years who were undergoing a pelvic examination. Patients were excluded if they had used antibiotics for the treatment of chlamydia within the last month, were currently more than 6 weeks pregnant, or had a past cervical biopsy for a cervical abnormality. Ninety-seven patients were included in this cross-sectional analysis. The study was approved by the Institutional Review Boards of Johns Hopkins University, the University of Maryland at Baltimore, and the Baltimore City Health Department.
Study Protocol and Questionnaire
After obtaining informed consent from participants, a brief questionnaire was administered by one interviewer (DJ). The questions included self-reported history of menarche and douching, number and type of past sexually transmitted diseases, pregnancy, sexual behavior, and hormonal contraceptive use. Sexual history variables included age at first intercourse, number of lifetime and current (last 6 months) partners, frequency of sex (daily, times per week/ month) over the last 6 months with one-time, occasional, and steady partners, and frequency of barrier contraceptive use. Contraception history included ages at and duration of lifetime and current usage of low-dose combination OCs, DMPA, and levonorgestrel subdermal implants (Norplant, Wyeth-Ayerst, Paoli, PA).
Endocervical specimens were obtained for C trachomatis testing by polymerase chain reaction (PCR).13,14 After specimen collection, mucus was removed from the ectocervix and a solution of 5% acetic acid was applied to the ectocervix. The instrument was aligned and the cervix was brought into adequate view and focused. Acetic acid was reapplied and two photographs were taken 15 seconds after reapplication.
Cervical photographs were taken with a Cerviscope (Cerviscope; National Testing Laboratories, Fenton, MO) using a standardized protocol.11 The standard focal length and preadjusted settings allowed for comparison of the absolute measurements of the cervix across patients.
Cervical Measurements Using Computerized Planimetry
Total area measurements of the ectocervix (cervical portio). Cervical measurements were calculated from the photographs using computerized planimetry11 in a method that has been described previously.12 Briefly, the images were scanned into a Power Macintosh computer (Apple Computer Corp., Cupertino, CA). Four areas were estimated and outlined: (1) the outermost border of visible ectocervix (cervical portio), (2) the external cervical os, (3) cervical ectopy, and (4) the transformation zone (the area between the original and the neosquamocolumnar junction, including immature metaplasia). Cervical ectopy was defined as dark-red areas on the portio adjacent to the external cervical os, and the transformation zone (T zone) was defined as lighter areas between the dark-red ectopy and the pink stratified epithelium of the portio. Using Adobe Photoshop 3.0 alias (Adobe Systems, San Jose, CA), the images were enhanced and enlarged. Using an image program (National Institutes of Health Image 1.43, National Institutes of Health, Bethesda, MD), the areas within the outlines were computed.11 The areas (mm2) were calculated based on a conversion scale of 112 pixels to 5 mm.
Area of ectopy/T zone versus percent ectopy/T zone. We directly quantified the total areas of ectopy, T zone and the ectocervix. Percent measurements are also mentioned throughout the text so that our results could be compared with previous studies and adjusted for the total size of the ectocervix (cervical portio).
The calculations for percent ectopy/T zone were as follows: The specific area (ectopy or T zone) divided by the area of the ectocervix minus the area of the os, or EQUATION 1 and EQUATION 2
We analyzed whether chlamydial infection was more prevalent among females with larger amounts of ectopy or T zone, and among females using OCs or DMPA. Analyses were adjusted for potential confounders such as age, sexual partners, sexual frequency, condom use, douching, and reproductive or gynecologic history (menarche, parity, past sexually transmitted diseases).
Univariate analysis. We analyzed the frequency and percent of chlamydial infection by strata of the risk factor or potential confounder. Differences among levels were evaluated with the Fisher exact test. Unadjusted odds ratios and 95% CIs of infection between strata were obtained from simple logistic regression. All analyses were performed with SAS version 6.11 (SAS Institute, Cary, NC).
Multivariate analysis. We used multivariate logistic regression to calculate the adjusted odds of chlamydial infection by the amount of cervical ectopy (area or percent), the amount of T zone (area or percent), and by hormonal contraceptive use (OC, DMPA). Potential confounders were examined in multivariate models if they were associated with the risk factor and chlamydial infection on univariate analysis, or if they were predictors of chlamydia in the documented literature. The final models each contain only three variables because of small numbers. Variables that were not significantly associated with chlamydia are indicated at the bottom of the multivariate tables.
Categorization of Variables
The areas and percent of ectopy and T zone were divided into approximate tertiles. Ectopy was not normally distributed (25% of participants had no ectopy). All categories were established before data analysis based on biology and previous literature, when available.15 The categories were (1) area of ectopy: none (0 mm2), mild (<22 mm2), and moderate (>22 mm2); (2) percent ectopy: none (0%), mild (0.01-8%), and moderate (>8%); (3) T zone area: small (<129 mm2), moderate (129-198 mm2), and extensive (>199 mm2); and (4) percent T zone: mild (<31%), moderate (31-46%), and extensive (>47%). Age and sexual and reproductive history were categorized by the median, tertile, or average ages of reproductive or sexual events obtained from the literature.1
Cervical specimens were tested for chlamydia by PCR using the Amplicor system (Roche Molecular Diagnostics, Branchburg, NJ) according to the manufacturer's instructions.13,14
Description of Population
Although cervical photographs were taken in 97 patients, nine photographs were unreadable because the cervix was not adequately visualized. These nine participants who were subsequently excluded from the analysis of ectopy and T zone did not differ by demographic, reproductive, or gynecologic histories.
The median age of this predominantly black (91%) sample was 17 years (Table 1), and sexual debut occurred at a median of 14.0 years. Half of the participants had been sexually active for > 3 years and 41% had been pregnant. At study enrollment, 12 females (12%) were using OCs, 15 (15%) were using DMPA, 2 (2%) were using Norplant, and 64 (66%) were not using any form of hormonal contraception.
Cervical measures and cervical maturation. Sixty of 88 participants (74%) had ectopy. C trachomatis was detected by polymerase chain reaction in 15 of 97 females (15%) (Table 1). Chlamydia prevalence was not significantly different among the three ectopy strata (Table 2). Adolescents with moderate ectopy (27.3 percent) tended to have a greater rate of infection than adolescents with mild ectopy (11.1%, P = 0.08) on univariate analysis only. Neither T zone area nor T zone percent was significantly associated with chlamydia.
Hormonal contraception. Chlamydial infection was associated with current DMPA use. Forty percent of current DMPA users had chlamydia, compared with 8% of OC users and 12% of nonhormone users (P = 0.03) (Table 3). Use of DMPA for more than 3 months conferred the greatest risk of infection compared with use for 1 to 3 months or with nonhormone use (P < 0.001). Duration of OC use was not associated with chlamydial infection.
Sexual history and sexually transmitted diseases. In contrast to previous studies, chlamydial infection was more common in women with fewer lifetime partners (32% versus 10%, P < 0.001) (Table 3). The prevalence of chlamydia did not differ significantly by type of partner, sexual frequency, or condom use. There was a nonsignificant trend for higher prevalence of chlamydia in persons without a history of gonorrhea or trichomoniasis.
Association between hormonal contraception and other covariates. DMPA users were generally similar to OC users and nonhormonal contraceptive users in terms of demographics and behavioral and gynecologic history; however, DMPA users were more likely to be monogamous and to have more frequent sex than OC users. However, these variables are unlikely to be confounders because they were associated with hormonal contraceptive use (the risk factor), but not with chlamydial infection (the outcome).
We evaluated the independent contribution of cervical ectopy and T zone and hormonal contraceptives to chlamydial infection adjusted for covariates (Table 4). Chlamydia was not independently associated with any amount of cervical ectopy even after adjustment (model 1).
Chlamydial infection was more likely in DMPA users compared with nonhormone users (adjusted OR, 5.44; 95% CI, 1.25-23.6) even when adjusted for age and lifetime number of partners (model 2). The strength of the association between DMPA and chlamydia diminished when adjusted for age and T zone size (adjusted OR, 3.6; 95% CI, 0.72-17.68) (model 3). Oral contraceptive users did not have an increased likelihood of chlamydial infection in this study (adjusted OR, 0.92; 95% CI, 0.10-8.78), adjusted for age and number of partners (model 2), or for age and T zone size (model 3).
Paradoxically, chlamydia was somewhat less likely with more lifetime partners (adjusted OR, 0.31; 95% CI, 0.08-1.17) when adjusted for age and current contraceptive use (models 1 and 2). No other variables were associated with chlamydial infection, including age, number of partners during the past 6 months, condom use, frequency of sex during the past 6 months, reproductive history, past hormonal contraceptive use, or history of gonorrhea or trichomoniasis.
Using a standardized, reproducible measurement technique for cervical ectopy, we examined two chlamydia risk factors that are especially relevant to young women: cervical ectopy and hormonal contraception. We found that carefully measured ectopy was not associated with prevalence of chlamydia.
After adjusting for age and number of sexual partners, DMPA users were more likely to be infected with chlamydia compared with nonhormone users. Although this relation was no longer significant after adjustment for T zone size, we think that this decrease is explained by the strong association between DMPA and T zone size rather than by confounding by T zone. No independent relation between T zone and chlamydia was found. OC users did not have an elevated prevalence of chlamydial infection compared with DMPA or nonhormone users.
Our results contrast with most previous cross-sectional studies in adolescents that found a relation between ectopy and chlamydial infection.9,16,17 Our results coincide with one previous longitudinal study that did not find an elevated risk of infection with greater amounts of ectopy.7 We think that previous studies may have used less-sensitive measurements with the potential misclassification of T zone and inflammation as ectopy.7 It easier to distinguish between the areas of ectopy and T zone using cervical photographs after acetic acid application because the T zone may appear red with unaided visual assessment. Computerized planimetry allowed for the reliable quantification of the different cervical zones and the detection of small areas of ectopy not seen by unaided visual assessment. These techniques also improved measurement reliability.11
To assay for chlamydia, we used PCR techniques that increases sensitivity by 20-30% compared with culture.18 With culture techniques, ectopy may improve detection of the organism because the columnar cells are exposed to the vaginal environment. Use of PCR may reduce this detection bias because fewer organisms are needed to detect an infection.
The relation between OCs and chlamydia has been debated.7–9,19–21 A second controversy relates to the contribution of estrogen or progesterone components. Our data suggest that progesterone (e.g., DMPA) is more likely than estrogen to increase the risk of chlamydial infection. Few other studies have evaluated DMPA as a risk factor for chlamydial infection.22
The inconsistent results among previous studies on the relation between OC use and chlamydia risk may be due to multiple factors, including varying duration of hormone exposure8 or behavior changes in women using different contraceptive methods. Adolescents frequently change or discontinue contraception, have variable compliance, and change the frequency of intercourse.1 For example, one study found a higher prevalence of chlamydia among adolescents using OCs for more than 6 months.8 Similarly, in our study no person currently using OCs or DMPA for less than 4 months was infected. Finally, although dose of estrogen or progesterone in OCs has not been associated with an increased rate of chlamydia,5,23 our study may differ because current OC preparations have less estradiol and progestin, are less androgenic, and are are more similar to each other than past OC preparations.
The mechanism by which progesterone might elevate the risk of chlamydial infection is not well known. Progesterone might alter attachment or growth of chlamydia24 and stimulate hyperplasia of columnar cells, thereby increasing the number of target cells24 or maintain the number of columnar cells by inhibiting metaplasia. DMPA may directly affect the T zone because progesterone alone inhibits complete maturation of squamous epithelium.25 Although OCs, which contain estrogen and progesterone, are thought to elevate chlamydial risk primarily by producing ectopy, it is unlikely that progesterone alone promotes infection by ectopy. In another study, we found no association between DMPA or OCs and ectopy.26
Because of our small sample size and cross-sectional design, we were limited in our ability to detect small differences in infection rates, to adjust for multiple confounders, to test for interaction between ectopy and hormone use, or to derive more precise odds estimates.
Cervicography was well tolerated and less subject to observer bias than previous techniques. Applying these methods in a large prospective study will improve our understanding of the complex relation between DMPA, OC use, cervical development, and the risks of infection with chlamydia. It will be important to obtain baseline measurements of ectopy and T zone before hormonal contraceptive use. Because DMPA use is increasing worldwide, it is important to gain a better understanding of the potential risk of DMPA use, which may increase reproductive sequelae among young women.
1. Zabin LS, Hayward SC, eds. Adolescent Sexual Behavior and Child-bearing. Newbury Park: Sage Publications, 1993.
2. Jordan J, Singer S, eds. The Cervix. London: WB Saunders, 1976.
3. Jordan J, Singer A. Effect of oral contraceptive steroids upon epithelium and mucus. In: Jordan J, Singer A, eds. The Cervix. London: WB Saunders, 1976:192–209.
4. Ferenczy A, Wright TC. Anatomy and histology of the cervix. In: Kurman R, ed. Blaustein's Pathology of the Female Genital Tract. New York: Springer-Verlag, 1994:185–201.
5. Louv WC, Austin H, Perlman J, Alexander WJ. Oral contraceptive use and the risk of chlamydia and gonococcal infections. Am J Obstet Gynecol 1989; 160:396–402.
6. Harrison HR, Costin M, Meder JB, et al. Cervical Chlamydia trachomatis
infection in university women: relationship to history, contraception, ectopy, and cervicitis. Am J Obstet Gynecol 1985; 153:244–251.
7. Rahm VA, Odlind V, Pettersson R. Chlamydia trachomatis
in sexually active teenage girls. Factors related to genital chlamydial infection: a prospective study. Genitourin Med 1991; 67:317–321.
8. Oh MK, Feinstein RA, Soileau EJ, Cloud GA, Pass RF. Chlamydia trachomatis
cervical infection and oral contraceptive use among adolescent girls. J Adolesc Health Care 1989; 10:376–381.
9. Blythe MJ, Katz BP, Orr DP, Caine VA, Jones RB. Historical and clinical factors associated with Chlamydia trachomatis
genitourinary infection in female adolescents. J Pediatr 1988; 112:1000–1114.
10. Critchlow CW, Wolner-Hanssen P, Eschenbach DA, et al. Determinants of cervical ectopia and of cervicitis: age, oral contraception, specific infection, smoking, and douching. Am J Obstet Gynecol 1995; 173:534–543.
11. Gilmour E, Ellerbrock TV, Koulos JP, et al. Measuring cervical ectopy: direct visual assessment versus computerized planimetry. Am J Obstet Gynecol 1997; 176:108–111.
12. Jacobson DL, Peralta L, Farmer M, Graham NMH, Wright TC, Zenilman J. Cervical ectopy and the transformation zone measured by computerized planimetry in adolescents. Int J Gynecol Obstet 1999; 66:7–17.
13. Jaschek G, Gaydos CA, Welsh LE, Quinn TC. Direct detection of Chlamydia trachomatis
in urine specimens from symptomatic and asymptomatic men by using a rapid polymerase chain reaction assay. J Clin Microbiol 1993; 31:1209–1212.
14. Bauer HM, Greer CE, Manos MM. Determinants of genital human papillomavirus infection by consensus PCR amplification. In: Herrington CS, McGee JO, eds. Diagnostic Molecular Pathology: A Practical Approach. Vol. II-Cell and Tissue Genotyping. Oxford: Oxford University Press, 1992:131–152.
15. Guijon F, Paraskevas M, Brunham R. The association of sexually transmitted diseases with cervical intraepithelial neoplasia: a case-control study. Am J Obstet Gynecol 1985; 151:185–190.
16. Chacko MR, Lovchik JC. Chlamydia trachomatis
infection in sexually active adolescents: prevalence and risk factors. Pediatrics 1984; 73:836–840.
17. Rahm VA, Odlind V, Gnarpe H. Chlamydia trachomatis
among sexually active teenage girls: influence of sampling location and clinical signs on the detection rate. Genitourin Med 1990; 66:66–69.
18. Bauwens J, Clark A, Loeffelholz M, Herman S, Stamm W. Diagnosis of Chlamydia trachomatis
urethritis in men by polymerase chain reaction assay of first-catch urine. J Clin Microbiol 1993; 31:3013–3016.
19. Shafer M, Beck A, Blain B, et al. Chlamydia trachomatis:
important relationships to race, contraception, lower genital tract infection, and Papanicolaou smear. J Pediatr 1984; 104:141–146.
20. Fraser JJ, Rettig PJ, Kaplan DW. Prevalence of cervical Chlamydia trachomatis
and Neisseria gonorrhoeae
in female adolescents. J Pediatr 1983; 71:333–336.
21. Jaffe LR, Siqueira LM, Diamond SB, Diaz A, Spielsinger NA. Chlamydia trachomatis
detection in adolescents: a comparison of direct specimen and tissue culture methods. J Adolesc Health Care 1986; 7:401–404.
22. Clark RA, Kissinger P, Williams T. Contraceptive and sexually transmitted diseases protection among adult and adolescent women infected with human immunodeficiency virus. Int J STD AIDS 1996; 7:439–442.
23. Avonts D, Sercu M, Heyerick P, Vandermeeren I, Meheus A, Piot P. Incidence of uncomplicated genital infections in women using oral contraception or an intrauterine device: a prospective study. Sex Transm Dis 1990; 17:23–29.
24. Tuffrey M, Taylor-Robinson D. Progesterone as a key factor in the development of a mouse model for genital-tract infection with Chlamydia trachomatis.
FEMS Microbiol Lett 1981; 12:111–115.
25. Ferenczy A, Winkler B. Benign diseases of the cervix. In: Kurman R, ed. Blaustein's Pathology of the Female Genital Tract. New York: Springer-Verlag, 1994:203–227.
26. Jacobson DL, Peralta L, Farmer M, Graham NM, Wright TC, Zenilman J. Cervical ectopy and the transformation zone measured by computerized planimetry in adolescents. Int J Gynaecol Obstet 1999; 66:7–17.