FOR SEVERAL REASONS, it is difficult to determine the prevalence and incidence of pelvic inflammatory disease (PID) among adolescents. Undercounting of cases occurs because infected youths have mild symptoms and do not seek care. Experts estimate that approximately two thirds of all women who have PID have no or mild symptoms and never seek care, so the diagnosis is never made. 1 In prospective studies, loss to follow-up is common; in studies that rely on patient recall of a previous diagnosis of PID, adolescents may not remember, understand, or choose to mention the diagnosis. 2 An unavoidable problem results from diagnostic inaccuracy: even if appropriate clinical criteria are used for diagnosis, approximately one third of clinically apparent cases are not actually PID by the “gold standard” diagnostic test, laparoscopy, which is not practical for general use. Underdiagnosis occurs as well. 1
Only a few studies of PID have evaluated incidence and/or prevalence in a reasonably well-defined closed population of adolescents, and none of these were performed recently. 1,3 In the 1988 U.S. National Survey of Family Growth (cycle IV), a retrospective population-based study depending on self-report, 2.9% of participants aged 15 to 19 years disclosed a history of PID. 2 In Lund, Sweden, researchers who provided all clinical services for patients with sexually transmitted infections (STIs) studied PID in 1977 to 1980. The mean annual incidence of PID was 2.5% among adolescents 15 to 19 years old. During this period, the prevalence of chlamydial infection was 22.6% among symptomatic adolescents attending a clinic for STIs and 4.7% among asymptomatic adolescents in screening programs. The prevalence of gonorrheal infection was 11.5% in symptomatic and 0.8% among asymptomatic youths. 4
Although many cases of PID have a polymicrobial cause, both Neisseria gonorrhoeae and Chlamydia trachomatis are important causes. The incidence of PID increases as the incidence of these infections increases. 1,3
At the Harris County Juvenile Detention Center (HCJDC), youths are often incarcerated for several weeks pending adjudication. All youths undergo complete medical evaluations soon after admission and are seen for follow-up examination if symptoms suggestive of PID develop. A previous study has shown that approximately 25% of young women detained at the HCJDC have chlamydial infections. 5 The purpose of this study was to measure the prevalence of PID at the time of incarceration and the incidence of PID in this well-defined, stable cohort of high-risk youths.
The HCJDC serves Harris County, Texas, which has a population of approximately 3.4 million people. An average of 620 youths are incarcerated each month (range, 503–750). About 150 (25%) of these youths are females, of whom 65% are released within 48 hours. The remaining 50 young women admitted each month are incarcerated for a period averaging 21 days. About 5% are arrested because they are sex workers; the others are runaways, are parole violators, or have committed a variety of crimes such as aggravated assault, burglary, and arson. These youths have no opportunity for heterosexual sexual activity after incarceration.
We conducted our study between April 18, 2000, and February 23, 2001. If a female complained of lower abdominal pain or other symptoms suggestive of an STI at her screening health assessment on the day of admission, at her complete medical evaluation several days later, or at any other time, she underwent a pelvic examination by one of two experienced adolescent medicine specialists. No one refused this examination. One of the clinicians also visited the females’ living quarters weekly to ask if any had developed lower abdominal pain or other STI symptoms and to encourage them to seek care if such symptoms did appear.
All youths, whether symptomatic or not, provided a first-catch urine sample that was tested for chlamydial infection with the Gen-Probe Amplified Chlamydia trachomatis Assay, a transcription-mediated amplification procedure that qualitatively detects chlamydial ribonucleic acid. This test had a sensitivity/specificity of 94%/100% in one typical study. 6
For the diagnosis of PID, we used the minimal criteria recommended by the CDC: the presence of lower abdominal tenderness; pain on cervical motion; and adnexal tenderness. 7 All who met these criteria also underwent a complete physical examination, and during their pelvic examination, we analyzed saline and KOH wet preparations and the pH of vaginal secretions and performed a cervical test for gonorrhea and chlamydia. We used the Gen-Probe Pace 2C Assay, a combination DNA hybridization test that qualitatively detects chlamydial and gonorrheal ribonucleic acid. The sensitivity/specificity of this test is approximately 96%/99%. 8 If this test was positive, Gen-Probe Pace 2 Assays for each organism were done; these also are DNA hybridization tests that qualitatively detect chlamydial and gonorrheal ribonucleic acid. Investigators have reported sensitivities/specificities for these tests of 96%/98% for chlamydia 9 and 100%/99.7% for gonorrhea. 10 The urine and endocervical tests were those in use at the Houston Department of Health and Human Services, whose laboratory did the testing for free.
Girls who met the criteria for PID also underwent laboratory tests for determination of complete blood cell count, with white blood cell (WBC) differential, and erythrocyte sedimentation rate (ESR); a pregnancy test; a urinalysis and urine culture; an HIV and RPR test; and other tests if indicated by the findings of the pelvic examination. Those patients who underwent pelvic examinations but did not meet the criteria for the diagnosis of PID underwent only the cervical chlamydia and gonorrhea tests, wet-preparation assessment, and an HIV and RPR test, unless other tests were clinically indicated.
All young women who were not pregnant were included in the study. Pregnant adolescents were excluded because we often did not have accurate information on the duration of pregnancy. We therefore were unable to determine if these youths had reached the stage of pregnancy at which they were no longer at risk of PID. Prevalent cases were defined as those in which the pain began before the first medical evaluation on the day of admission, and incident cases were those in which the patients were asymptomatic at the first evaluation and subsequently developed pain. All nonpregnant youths were included in our calculation of the prevalence of PID at admission; nonpregnant youths without prevalent PID were included in the calculations of the incidence of PID during follow-up. The prevalence and incidence of PID were also determined in the subset of youths who admitted to being sexually active and were either heterosexual or bisexual and in the subset of sexually active youths who tested positive for chlamydial infection and/or gonorrhea.
Incidence density was calculated on the basis of person-months of follow-up. 11 Incidence density measures the rate at which cases occur and is appropriate for prospective studies with varying lengths of follow-up. Detainees were followed from their admission to their release, to their treatment for chlamydial and gonorrheal infection, to the development of PID, to the outcome of interest, or to the end of the study.
Urine tests for chlamydia were usually done several days after admission, and several more days passed before we were informed of positive results and treated infected youths, so these at-risk adolescents contributed person-time to the study. Follow-up ended on February 23, 2001, the last day of the study, for subjects still in the detention center.
Because the incidence density indicates how fast cases are occurring but does not measure an individual's risk of infection, we also determined the cumulative incidence, which does measure individual risk. 11 We calculated the cumulative incidence for the first 31 days, using the Kaplan–Meier method. 11 This method is appropriate in studies with varying lengths of follow-up.
We treated youths who met the criteria for PID with ceftriaxone (250 mg intramuscularly once) and doxycycline (100 mg twice daily for 14 days). They were reevaluated at 2–3, 7, 14, and 21 days. If they had not improved by day 2 or 3, metronidazole (500 mg orally twice a day for 10 days) was added to the antibiotic regimen; we also reviewed their clinical and laboratory findings as we considered alternative diagnoses.
Other STIs were diagnosed by the following methods: culture of lesions for herpes simplex infection; saline wet preparations for trichomoniasis; and visual inspection for venereal warts. Diagnosis of bacterial vaginosis depended on the occurrence of three of the following four findings (Amsel's criteria) 12: thin, gray, homogeneous vaginal discharge; vaginal pH >4.5; >20% clue cells in the saline wet preparation; and a positive whiff test.
The exact method was used to calculate confidence intervals for incidence and prevalence values; differences in proportions were evaluated by Fisher exact test. 11
This study was approved by the Committee for the Protection of Human Subjects of the University of Texas–Houston Health Sciences Center and by the administration of the Harris County Juvenile Probation Department. The subjects gave informed consent.
Of the 356 nonpregnant subjects, 319 said that they were sexually active. Of these, 300 reported being exclusively heterosexual, 13 bisexual, and 6 lesbian. Of the 313 sexually active heterosexual or bisexual youths, 78 (24.9%) were positive for chlamydial and/or gonorrheal infections: 71 for chlamydia, 2 for gonorrhea, and 5 for both. Of the 98 subjects who had cervical testing for gonorrhea and chlamydia during a pelvic examination (some of whom had PID and others who were evaluated for possible cervical infection), 20 (20%) were positive for chlamydia and 7 (7%) were positive for gonorrhea, so the ratio of chlamydial to gonorrheal infections in youths who were tested for both organisms was 2.9:1.
The demographic and STI risk factors of the 319 subjects who were sexually active and heterosexual or bisexual are described in Table 1. There were similar numbers of blacks, Hispanics, and whites. The various measures of sexual activity identified this as a high-risk population for PID:9 for example, 16% had a history of previous gonorrheal or chlamydial infection; 5.6% had a history of previous PID; and 8% had been sex workers. Of the approximately 65% of youths who provided information on the age and race/ethnicity of their sexual partners, 23% had had partners 3 to 4 years older and 33% had had partners >5 years older. Approximately 20% of blacks and Hispanics had had partners of other races/ethnicities, compared with 56% of whites.
The mean duration of follow-up was 21.1 days for all subjects. Some subjects had 0 days of follow-up because they were treated for PID or presumptively for chlamydial or gonorrheal infection on the day of their admission to the HCJDC. Four adolescents had more than 100 days of follow-up; no case of PID occurred after day 10 of follow-up. All gonorrheal and chlamydial infections identified in subjects who did not have PID were treated by day 31. To avoid bias by outliers, we calculated the incidence values for the first 31 days of follow-up only.
Table 2 summarizes the data on the prevalence of PID. Among all nonpregnant youths, the prevalence of PID at admission was 14/356 (3.9%). Among the 313 sexually active heterosexual or bisexual youths, 4.5% had prevalent PID, and among the 78 sexually active youths who were found to have chlamydial or gonorrheal infections, 10% had prevalent PID.
Table 3 presents the cumulative incidence (the risk) and the incidence density (the rate) of PID among those with no evidence of PID at initial examination. Among all nonpregnant detainees, the risk of developing PID was 1.9%, and the rate of disease development was 2.6 cases per 100 person-months; for sexually active subjects, the values were 2.4% and 3.3 cases per 100 person-months; and for those with confirmed chlamydial and/or gonorrheal infections, the values were 10% and 8.2 cases per 100 person-months.
The 14 persons with prevalent PID had had pain for a mean of 34 days and a range of 7 to 56 days before admission; 3 had had PID diagnosed before admission but had failed to complete the antibiotic treatment regimen; and 2 had been told in the previous 2 months that they had chlamydial or gonorrheal infections but had not received therapy. Four others had a history of successfully treated episodes of chlamydial or gonorrheal infection, and 1 had had a previous episode of PID. Sexual histories were similar to those of other youths, except that a higher proportion were sex workers: 4/14 (29%) versus 23/294 (8%) (difference significant:P = 0.047). Five patients were black, 8 Hispanic, and 1 white.
Ten of the 14 persons with prevalent PID met one or more of the CDC's additional criteria that are designed to increase the specificity of this diagnosis:5 5 had positive tests for gonorrhea and/or chlamydial infection as well as abnormal vaginal discharge (excess WBCs); 3 had positive tests only; and 2 had abnormal discharge only. Of those with positive tests, 7 had isolated chlamydial infections and 1 had an isolated gonorrheal infection. Ten had bacterial vaginosis, including 2/4 of those who had both normal vaginal discharge and negative tests for chlamydial and gonorrheal infection. None had systemic symptoms or elevated WBC counts or ESRs. After treatment was begun, 9 were followed for 14 or more days and had resolution of symptoms and signs of PID; 1 had improved at 7 days and was then discharged; and 4 were discharged before any follow-up (Figure 1).
The 5 persons with incident PID had pain for a mean of 5 days and a range of 1 to 10 days before diagnosis. All developed symptoms within 10 days of incarceration. One had recently had an untreated chlamydial infection and 1 had a history of PID. One was a sex worker. One case was black, one was Hispanic, and three were white.
Two of the 5 incident cases met additional diagnostic criteria for PID: 2 had chlamydial infections, 1 of whom also had an abnormal vaginal discharge. For two the data were incomplete; they had refused a speculum examination because of vaginal bleeding. One was vomiting; none had fever; and none had elevated WBC counts or ESRs. Four had onset of symptoms during their menses and complained of more severe and persistent pain than usual. After we began therapy, we observed 2 patients for 4 or more days, and their symptoms and signs of PID resolved; 2 others had improved at 2 to 4 days and were then discharged; 1 had improved at 2 days and was discharged on day 7.
Of the other 86 youths who underwent a pelvic examination, 17 had one of the following symptomatic STIs: 12, trichomoniasis; 2, venereal warts; 2, culture-positive herpes simplex virus infection; and 1, bartholinitis. Thirty-two had only bacterial vaginosis, and 8 had vaginosis combined with an STI. Of the 23 who had chlamydial and/or gonorrheal cervical infections, 10 had abnormal vaginal discharge that was not explained by another infection.
Although there is inaccuracy in the clinical diagnosis of PID, we had two experienced clinicians applying consistent criteria, and our patients’ findings were typical of those described for women with PID. 1 The cases were of mild disease, with few systemic symptoms and normal WBC counts and ESRs. Most prevalent cases (9/14; 64%) involved current or recent untreated chlamydial or gonorrheal infection or both, as did 3/5 (60%) of cases of incident disease; other studies have reported a prevalence of these infections of 25% to 75% of inpatients who have PID. 1 In most of the incident cases (4/5), symptoms developed during menses. For those subjects who remained in detention long enough for adequate follow-up, symptoms and signs resolved with therapy. Of interest was the fact that a significantly higher proportion of subjects with prevalent PID than of all sexually active females were sex workers (4/14 [29%] versus 23/294 [8%];P = 0.047).
The advantage of this study was our ability to examine a stable population of high-risk adolescents who could be carefully evaluated at baseline. Those who lacked evidence of PID could then be followed carefully for the appearance of symptoms of this disease. Thus, we were able to identify mild as well as more severe disease and to achieve complete ascertainment of all cases that met our clinical criteria. One characteristic of this group was the lack of continuing sexual activity. Continuing intercourse can increase risk of disease by introducing chlamydial or gonorrheal infections in previously uninfected women. In addition, the mechanical and physiological events of intercourse may increase risk in ways that are currently poorly understood but that may include infection of fallopian tubes by bacteria adhering to spermatozoa. 13 The impact of this lack of continuing sexual activity on the risk of PID in our subjects is unknown.
Because of our screening program for chlamydia, asymptomatic infected youths were identified and treated within 31 days, and usually sooner, so that their risk of PID was probably decreased after this. In addition, our symptomatic subjects with gynecologic complaints but without PID were promptly evaluated, so those who had gonorrheal or chlamydial infection also were treated within a month. Some subjects probably continued to be at risk, because other organisms sometimes cause PID and because some subjects undoubtedly had unrecognized gonorrheal or chlamydial infection, given that asymptomatic subjects were not tested for gonorrhea and that the sensitivity of our tests for chlamydial infection was not 100%. The decrease in risk is supported by the fact that all of our incident cases occurred within the first 10 days following incarceration. We therefore believed that it was appropriate to calculate the incidence values for only the first 31 days of follow-up.
Our cumulative incidence of 2.2%/month is much higher than the rate of 2.5%/year that Westrom et al. found in Lund, Sweden. 4 Part of the difference is probably attributable to the apparent higher rate of chlamydial infection among our subjects: in Sweden, the prevalence was similarly high in adolescents with symptoms of an STI (23%) but not in asymptomatic youths (5%). In addition, we may have seen adolescents with mild symptoms who would not have sought even readily available medical care, and the Swedish researchers confirmed 90% of their cases with laparoscopy, so they may have had fewer false-positive results. Our incidence is also much higher than the incidence of 2.9% reported by adolescents 15 to 19 years old in the 1988 National Survey of Family Growth. 2 Part of this difference may be a result of underreporting in that study; in addition, the rate of chlamydial infection has apparently increased significantly since the 1980s, although there are problems in the accuracy and comparability of surveillance data over this period of time. 14
How does this high incidence and prevalence of PID in our population relate to risk factors for PID? Aral and Wasserheit 15 have discussed two such factors at the level of the individual and of the population. One factor is recurrent or persistent infection with Chlamydia. Individual risks for such infection—for example, inconsistent condom use and high numbers of partners—appear to be less important than such population-related factors as age, race/ethnicity, neighborhood, and geographic area. The age of our subjects (adolescents) puts them in a high-risk group, as does their geographic location in the South; the majority of our subjects were black, and blacks are known to be at high risk for chlamydial infection. 1,14 In our population, however, Hispanics and whites have rates of chlamydial infection similar to that in blacks. 5 As an additional risk factor, incarcerated youths have been found to have the highest prevalence of chlamydial infection of any subgroup of adolescents. 5,16
Aral and Wasserheit 15 identify critical delays in detection and treatment of chlamydial and gonorrheal cervical infections as the second most important set of risk factors for upper-tract disease. At the individual level, women with symptomatic disease are known to often delay seeking care and to fail to comply with therapy. Even though the mean duration of their symptoms was 34 days, only 3 of the 14 young women in our study who had prevalent PID had sought medical care before incarceration, and these 3 did not comply with their treatment regimens. At the population level, potential problems include lack of knowledge of where to seek care and inaccessibility or unacceptability of services. We did not study these factors in our population, but we know from experience that few incarcerated youths know where to go for medical care, and therefore rely on equally uninformed parents for this information, and that the available services for symptomatic indigent youths are few and often user-unfriendly, as are screening and treatment programs for asymptomatic sexually active adolescents.
In conclusion, this is one of the few recent studies to assess the prevalence and incidence of symptomatic PID in a reasonably well-defined population of adolescents. Among the sexually active subjects, we found a high prevalence (4.5%), incidence density (3.3 cases/100 person-months), and cumulative incidence (2.2%/month) of PID. These findings underscore the need for continued efforts to improve the prevention, identification, and treatment of chlamydial and gonorrheal infection in high-risk populations of adolescents.
1. Westrom L, Eschenbach D. Pelvic inflammatory disease. In: Holmes KK, Sparling PF, Mardh P-A, et al., eds. Sexually Transmitted Diseases. 3rd ed. New York: McGraw-Hill Company, 1999: 783–810.
2. Aral SO, Mosher WD, Cates W Jr. Self-reported pelvic inflammatory disease in the United States, 1988. JAMA 1991; 266: 2570–2573.
3. Simms I, Stephenson JM. Pelvic inflammatory diseases epidemiology: what do we know and what do we need to know? Sex Transm Infect 2000; 6: 80–87.
4. Westrom L, Svennson L, Wolner-Hansen P, Mardh PA. Chlamydial and gonococcal infections in a defined population of women. Scan J Infect Dis Suppl 1982; 32: 77–82.
5. Risser JMH, Risser WL, Gefter LA, Brandstetter DM, Cromwell PF. Implementation of a screening program for chlamydia infection in incarcerated adolescents. Sex Transm Dis 2001; 28: 43–46.
6. Crotchfelt KA, Pare B, Gaydos C, Quinn TC. Detection of Chlamydia trachomatis by the Gen-Probe AMPLIFIED Chlamydia Trachomatis Assay (AMP CT) in urine specimens from men and women and endocervical specimens from women. J Clin Microbial 1998; 36: 361–364.
7. Centers for Disease Control and Prevention. 1988 Guidelines for treatment of sexually transmitted disease. MMWR Morb Mortal Wkly Rep 1998; 47 (RR-1):79–80.
8. Iwen PC, Walker RA, Warren KL, Kelly DM, Hinrichs SH, Linder J. Evaluation of a nucleic acid-based test (PACE 2C) for simultaneous detection of Chlamydia trachomatis and Neisseria gonorrhoeae in endocervical specimens. J Clin Microbiol 1995; 33: 2587–2591.
9. Warren R, Dwyer B, Plackett M, Pettit K, Rizv N, Baker AM. Comparative evaluation of detection assays for Chlamydia trachomatis. J Clin Microbiol 1993; 31: 1663–1666.
10. Chapin-Robertson K, Reece EA, Edberg SC. Evaluation of the Gen-Probe PACE II assay for the direct detection of Neisseria gonorrhoeae in endocervical specimens. Diagn Microbiol Infect Dis 1992; 8: 645–649.
11. Szklo M, Nieto FJ. Epidemiology: Beyond the Basics. Gaithersburg, MD: Aspen Publishers, 2000: 58–71.
12. Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, Holmes KK. Nonspecific vaginitis: diagnostic criteria and microbial and epidemiologic associations. Am J Med 1983; 74: 14–22.
13. Friberg J, Confino E, Suarez M, Gleicher N. Chlamydia trachomatis attached to spermatozoa recovered from the peritoneal cavity of patients with salpingitis. J Reprod Med 1987; 32: 120–122.
14. Division of STD Prevention. Sexually Transmitted Disease Surveillance, 1999. Department of Health and Human Services, Atlanta: Centers for Disease Control and Prevention, 2000: 7–24.
15. Aral SO, Wasserheit JN. Social and behavioral correlates of pelvic inflammatory disease. Sex Transm Dis 1998; 28: 378–385.
16. Oh MK, Smith KR, O'Cain M, Kilmer D, Johnson J, Hook EW III. Urine-based screening of adolescents in detention to guide treatment for gonococcal and chlamydial infections. Arch Pediatr Adolesc Med 1998; 152: 52–56.