Objectives. We sought to describe the reproductive health of adolescent girls perinatally infected with HIV.
Methods. We estimated the incidence of first pregnancy, genital infections, and abnormal cervical cytology for 638 girls aged 13 years and older in the Pediatric AIDS Clinical Trials Group protocol 219C.
Results. Thirty-eight girls became pregnant, for a first pregnancy rate of 18.8/ 1000 person-years; 7 of these girls had additional pregnancies (95% confidence interval [CI]=13.3, 25.7). Thirty-two pregnancies resulted in live births. All girls received antiretroviral therapy during pregnancy. One infant was HIV infected, 29 were uninfected, and 2 had unknown infection status, for a rate of mother-to-child transmission of HIV in infants with known infection status of 3.3% (95% CI=0.1, 18.6). Condylomata and trichomoniasis were the most frequent genital infections. Forty-eight (47.5%) of 101 girls with Papanicolaou test examinations had abnormal cervical cytology, including atypical cells of undetermined significance (n=18), low-grade squamous intraepithelial lesions (SIL; n=27), and high-grade SIL (n=3). Many abnormalities persisted despite intervention.
Conclusions. Pregnancy rates were lower and cervical abnormalities were higher than among non–HIV-infected adolescents. These findings underscore the importance of Papanicolaou tests and promotion of safer sexual practices in this population.
Some girls infected perinatally with HIV early in the HIV/AIDS epidemic now have reached adolescence and become sexually active. The rate of pregnancy in these adolescent girls and the effectiveness of antiretroviral therapy (ART) in preventing second-generation mother-to-child transmission of HIV are not known. Two case reports have described the pregnancy outcomes of a small number of perinatally infected adolescents in the United States,1,2 and pregnancy outcomes in 28 peri-natally infected adolescents in India have been reported.3
Although reproductive health outcomes for adolescent girls with behaviorally acquired HIV infection have been studied extensively, little data exist on the reproductive health of girls with perinatal HIV infection. A high risk of genital human papillomavirus infection and its sequelae, cervical squamous intraepithelial lesions (SIL),4–6 as well as other genital infections,7 has been documented in women and adolescents infected with HIV through sexual activity and drug use. Adolescents perinatally infected with HIV are often cared for in pediatric infectious disease clinics, where reproductive health issues may not be routinely addressed.
In this study, we estimated rates of pregnancy and pregnancy outcomes, genital infections, and cervical cytological abnormalities in a cohort of perinatally HIV-infected adolescents in the United States.
The source population for this study was children enrolled in the Pediatric AIDS Clinical Trials Group (PACTG) protocol 219C. The PACTG was the pediatric counterpart of the AIDS Clinic Trial Group (ACTG), a US multi-site clinical HIV/AIDS research organization that has a major role in setting standards of care for HIV infection in the United States and the developed world. Protocol 219C has enrolled and followed HIV-infected and non–HIV-infected children at clinical centers—primarily university-affiliated pediatric infectious disease clinics—across the United States since September 2000 to study the long-term effects of in utero, neonatal, and pediatric ART exposure and the complications of pediatric HIV infection.
Protocol 219C was a revised version of PACTG protocol 219, which began in 1993; the primary change was to remove the restriction by which protocol 219 enrolled only children who were currently or previously enrolled in an ACTG clinical trial or whose mother was enrolled in an ACTG clinical trial during pregnancy. HIV-infected children enrolled in 219C if they were aged 21 years or younger, able to adhere to protocol visits, and receive medical care at a PACTG clinical site. Children were followed until loss of contact, protocol withdrawal, death, or their 25th birthday. The clinical centers involved in protocol 219C obtained approval from their respective review boards for human research, and the child or the child’s parent or guardian provided written informed consent.
The study population for our investigation was restricted to perinatally HIV-infected girls aged 13 years or older who participated in protocol 219C. A girl was considered to have perinatally acquired HIV infection if her mother was HIV infected during pregnancy, labor, or delivery according to clinical records; according to HIV DNA polymerase chain reaction or RNA assays, most girls were shown to be HIV positive during infancy or early childhood. Because of the paucity of data on rates of pregnancy and pregnancy outcomes, including mother-to-child transmission of HIV, we sought to include all pregnancies that occurred in the cohort and so used the minimum age of occurrence of pregnancy—13 years—in defining the study base.
A wide range of demographic, clinical, immunologic, and virologic data were obtained at participants’ enrollment in PACTG protocol 219C and at follow-up visits, which took place every 3 months; these data included information on living conditions, diagnostic test results and histories of clinical diagnoses, pregnancies and pregnancy outcomes, ART use, Centers for Disease Control and Prevention (CDC) disease classification,8 CD4-positive T lymphocyte number CD4+ and percentage and plasma HIV viral load copy number (obtained through blood samples), and pelvic tests and cervical Papanicolaou (Pap) tests.
The 219C protocol initially requested pelvic examinations—which included a visual inspection of the cervix and vagina and a bi-manual examination—every 3 years beginning at age 15 years or the onset of sexual activity, whichever came first. Cervical Pap tests initially were requested of sexually active girls every 3 years. Beginning in 2002, pelvic and Pap tests were administered annually. Pap tests were fixed with a spray preservative and sent to the local PACTG site’s laboratory for reading. Because the protocol 219C participants received their primary care from clinicians at the PACTG site, clinically indicated test results were recorded in addition to those specified in the 219C protocol.
All pregnancies and pregnancy outcomes recorded in the 219C database were identified and confirmed by the clinical site. Live-born infants were classified by the primary care provider as HIV infected or uninfected on the basis of HIV DNA polymerase chain reaction or RNA assays. Diagnoses of atypical squamous cells of undetermined significance (ASCUS), low-grade SIL, or high-grade SIL were made according to the Bethesda system for cervical cytological diagnosis.9 Although screening for genital infections was not performed routinely as part of protocol 219C, the physician attending the pelvic examination frequently diagnosed symptomatic cases. The diagnostic procedures for genital condylomata, trichomoniasis, chlamydial infection, gonorrhea, and syphilis were based on the PACTG site’s standard diagnostic procedures and were conducted at the institution’s local laboratory, all of which were accredited by the Colleges of American Pathology. Girls were identified as sexually active through disclosure or occurrence of pregnancy or genital infection.
We estimated the cumulative incidence of first pregnancy and the incidence rate per 1000 person-years from the girl’s 13th birthday to the date of the conception of the first pregnancy or the last cohort visit, whichever occurred first. The date of conception was determined by subtracting the gestational age of the fetus or infant from the date of the pregnancy outcome. In estimating the rate of mother-to-child transmission of HIV, pregnancies that yielded multiple births were assessed as a single transmission if any infants were HIV infected and as a single nontransmission if none were infected. The cumulative incidence of first genital condylomata, trichomoniasis, chlamydial infection, gonorrhea, and syphilis also were estimated; girls with a diagnosis of the genital infection before age 13 years were excluded from analyses for that particular outcome. Incidence rates of pregnancy and cumulative incidence of genital infections were estimated for all girls and for girls known to be sexually active.
Analysis of abnormal cervical cytology was restricted to 101 girls for whom Pap tests were given. The cumulative incidence of low- or high-grade SIL was estimated for 40 girls whose first Pap test was indicative of normal cervical cytology and who then received at least 1 subsequent Pap test (with normal or abnormal result). The treatment and persistence of cervical abnormalities were described.
Differences in CDC disease classification, CD4+ percentage, HIV RNA copy number, in utero ART exposure, and ART use according to sexual activity, pregnancy, and cervical cytology were assessed using the Kruskal–Wallis, χ2, and Fisher exact tests as appropriate.
Of 1317 perinatally HIV-infected girls enrolled in PACTG protocol 219C from September 2000 to December 2005, 638 were aged 13 years or older and were included in the study population. These 638 girls were from 75 pediatric infectious disease clinics in 24 states and districts across the United States. A total of 174 (27.3%) were known to be sexually active on the basis of disclosure, pregnancy, or genital infection. Sexually active girls were significantly older and significantly more likely to be Hispanic, to live on their own, and to have a higher HIV RNA copy number and a lower CD4+ percentage than were girls who were not known to be sexually active (Table 1). Sexually active girls were less likely to be on ART than non–sexually active girls; however, among girls on ART, there was no significant difference in the proportions on the 3 highly active antiretroviral therapies described in Table 1.
Thirty-eight girls experienced a first pregnancy between the age of 13 years and their last cohort visit (2026.2 person-years of follow-up). The overall incidence rate of first pregnancy was 18.8 per 1000 person-years (95% confidence interval [CI] = 13.3, 25.7), and the incidence among girls aged 15 to 19 years was 33.5 per 1000 person-years (95% CI = 22.8, 47.6). At the 19th birthday, the cumulative incidence of first pregnancy was 17.2% (95% CI = 11.1, 23.2) among all 638 girls in the study population and 24.2% (95% CI = 16.6, 31.8) among 174 girls known to be sexually active (Figure 1). Girls who became pregnant were significantly older than those who did not, and in general, the differences in characteristics between the 2 groups paralleled the differences between sexually active and non–sexually active girls shown in Table 1. In addition, among girls known to be sexually active, those who became pregnant were more likely to have abnormal cervical cytology during follow-up than those who did not become pregnant.
In addition to the 38 first pregnancies, 6 girls had a second pregnancy, and 1 had a third pregnancy. Twenty-eight of the first pregnancies resulted in live births (including one set of twins), 2 spontaneously aborted, and 8 were therapeutically aborted. Three of the second pregnancies were therapeutically aborted, and 3 resulted in live births; the single third pregnancy resulted in a live birth. Of the 32 pregnancies that resulted in live births, 1 infant was HIV infected, 29 were not HIV infected (with twins counted as a single event), and 2 had unknown infection status. The rate of perinatal HIV transmission among infants with known infection status was therefore 3.3% (95% CI = 0.1, 18.6). All 32 girls who delivered a live birth received ART during pregnancy; 26 were on combination therapy (i.e., at least 3 drugs) that included a protease inhibitor at some point during pregnancy, 4 received nucleoside reverse transcriptase inhibitors (NRTIs) only, and 2 received combination therapy that included a nonnucleoside reverse transcriptase inhibitor (NNRTI). One of the latter 2 was the mother of the HIV-infected infant.
The estimated cumulative incidence of genital condylomata, chlamydial infection, trichomoniasis, gonorrhea, and syphilis from age 13 to 19 years for all girls and for sexually active girls is shown in Table 2. Incidence was highest for genital condylomata and trichomoniasis. In addition to cases of condylomata in girls aged 13 years and older, 23 girls had condylomata diagnosed before the age of 13 years. The mean age at diagnosis of these 23 girls was 6.3 years (range = 2.7–11.0); 18 girls had documentation of HIV infection prior to the diagnosis of condylomata.
Of 174 girls known through disclosure, pregnancy, or genital infection to be sexually active during follow-up, 101 (58.1%) and 165 (94.8%) had a documented Pap test and pelvic examination, respectively. The mean age at the first Pap test was 16.7 years (range = 12.2–23.5). Reflecting the racial distribution of the 219C cohort, most Pap tests were given to non-Hispanic Black girls (52.5%), followed in frequency by Hispanics (35.6%), non-Hispanic Whites (9.9%), and girls of other or unknown race/ethnicity (2.0%). Thirty-six (35.6%) girls had 1 Pap test, 23 (22.8%) had 2, 17 (16.8%) had 3, and 25 (24.8%) had 4 or more.
Thirty of 101 girls with Pap tests had abnormal cervical cytology at the first examination for a prevalence of 29.7% (95% CI=20.0, 42.4), and 18 girls had abnormal cervical cytology during follow-up. The first abnormal cytological diagnosis was ASCUS for 18 girls, low-grade SIL for 27 girls, and high-grade SIL for 3 girls. The 12-month cumulative incidence of low- or high-grade SIL in 40 girls with a first normal Pap test and at least 1 subsequent Pap test was 17.5% (95% CI=4.7, 30.2; Figure 2). There were no significant differences in demographic characteristics, ART use, or HIV clinical, immunologic, or virologic status between girls with normal cervical cytology and those with abnormal cervical cytology; however, girls with abnormal cervical cytology were significantly more likely to have a genital infection during follow-up than girls with normal cervical cytology.
Of the girls with a diagnosis of abnormal cervical cytology, 7 of the 18 first diagnosed with ASCUS, 13 of the 27 first diagnosed with low-grade SIL, and 1 of the 3 girls first diagnosed with high-grade SIL underwent colposcopy, cryotherapy, excision, or a combination of these procedures. Among these 21 girls with intervention, 9 cervical abnormalities cleared, 2 regressed to less-severe SIL, 4 persisted, and 6 progressed to more-severe SIL. Among the 14 girls with cervical abnormalities and no documented intervention, 6 abnormalities cleared, 1 regressed to less-severe SIL, 6 persisted, and 1 progressed to more-severe SIL. There was no further information following the abnormal cervical diagnosis for 13 girls.
To our knowledge, this is the first study to evaluate the reproductive health of a large cohort of perinatally HIV-infected adolescent girls. Although the outcomes of pregnancy, genital infections, and abnormal cervical cytology may be considered distinct, collectively they describe the reproductive health of this population. The need for enhanced provision of reproductive health services such as contraceptive counseling and cervical cytological screening is evident: 17% of girls experienced a first pregnancy by 19 years of age and nearly half had abnormal cervical cytology. Further, only half of the girls known to be sexually active had Pap tests.
The pregnancy incidence rate of 33.5 per 1000 person-years among girls aged 15 to 19 years in our study was lower than that documented for predominantly non–HIV-infected girls of similar age in the United States in 1999 (86.7 per 1000),10 which is consistent with the findings of lower pregnancy rates in HIV-infected versus non–HIV-infected women.11 Because of the small numbers, the 219C pregnancy incidence rates were not standardized according to racial distributions in the United States; race/ethnicity has been associated with pregnancy and pregnancy outcomes.10 Very few non–HIV-infected children participated in protocol 219C until adolescence, which precluded use of reference rates in this population for the present study. In addition, some spontaneously or therapeutically aborted pregnancies may not have been documented, which could have underestimated the pregnancy rate in this population.
All adolescents who delivered a live birth were on ART—mainly a combination of 3 or more drugs, 1 of which was a protease inhibitor or NNRTI—at some point during pregnancy, and only 1 infant was HIV infected. The estimated rate of mother-to-child transmission of HIV of 3.3% (95% CI = 0.1, 18.6) is similar to rates observed in the Women and Infants Transmission Study for women receiving dual ART in which, in most cases, both drugs were NRTIs (3.8% [95% CI = 1.1, 6.5]), and for women receiving combination ART of 3 or more drugs, which included a protease inhibitor, NNRTI, or both (1.2% [95% CI = 0.0, 2.5]).12 However, there were a small number of live births in our study, and the CI is wide and also includes higher rates. In a case study of 8 perinatally HIV-infected adolescents and young adults in Puerto Rico, 5 first pregnancies resulted in live births, 4 of the 5 mothers were on ART, and none of the infants were HIV infected.1 Similarly, no mother-to-child transmission of HIV occurred among 26 infants born to perinatally HIV-infected adolescents in India.3 Further study is needed to ascertain the effectiveness of ART in preventing second-generation perinatal HIV transmission given the potential for transmission of a virus resistant to 1 or more classes of ART. Of note, none of the pregnant girls in our study were themselves exposed to ART in utero.
Rates of genital infections—including trichomoniasis, chlamydial infection, and gonorrhea—documented in our study population were lower than those found in the Reaching for Excellence in Adolescent Care and Health (REACH) cohort,13 which was expected because the REACH cohort comprised girls infected with HIV during adolescence, presumably through sexual transmission. Although screening for genital infections was not performed routinely as part of the 219C protocol, girls frequently were assessed at the time of the pelvic examination, and symptomatic cases were independently diagnosed through primary care. Thus, the reported rates represent the minimum rate of genital infections. Nevertheless, the observation that 23 of 638 girls had genital condylomata at a young age is troubling. We were unable to determine whether these young condylomata cases were related to recent exposure through sexual abuse, voluntary sexual activity, or acquisition of human papilloma virus (HPV) during delivery. However, the localization of the condylomata suggests that sexual abuse or activity was more likely to have been the source of infection.
The prevalence of abnormal cervical cytology was 29.7%, which is higher than the prevalence of 11.5% documented among 375 women aged 18 to 24 years attending a university health clinic in Montreal14 and lower than the 56.4% among 133 girls infected with HIV in adolescence in the REACH cohort.6 Further, although based on small numbers, the 36-month cumulative incidence of SIL among girls in protocol 219C (32.8%; 95% CI = 15.5, 50.0) was higher than that observed in high-risk non–HIV-infected adolescents in San Francisco (15.0%; 95% CI = 13.0, 17.0).15 The high proportion of abnormal cervical cytology in our population compared with university students also could be because of increased susceptibility to and persistence of HPV and other genital infections in these girls, as has been demonstrated in HIV-infected women.16 Although we had no virologic information on HPV infection, many cervical abnormalities persisted in affected girls, even with intervention. No associations between clinical, immunologic, or virologic profiles and cervical cytology were identified, but the analysis had limited power. In addition, the estimated cumulative incidence of SIL was based on a small number of participants, only half of whom participated in protocol 219C past age 19 years.
It is unsettling that only 58% of 174 girls known to be sexually active had documented Pap tests. The PACTG site provided the girls’ primary medical care, and testing and diagnoses from routine clinical care supplemented the data collected through the 219C protocol. The estimated rates of cervical cytological abnormalities could be misleading if girls for whom Pap tests were given had different profiles of abnormal cervical cytology than those who did not. Likewise, we may have overestimated the incidence of pregnancy and genital infection because sexual activity was identified through disclosure or the occurrence of the outcome of pregnancy or a genital infection. Nonetheless, this study provides preliminary data on a unique cohort of girls infected early in the HIV epidemic and who survived to reproductive age.
A considerable proportion of adolescents in our cohort engaged in unprotected sexual intercourse despite close and frequent contact with HIV clinics. Sexual activity in perinatally HIV-infected adolescent girls is probably more common than many pediatricians assume. Indeed, 45 pregnancies occurred, and almost 48% of the girls who had a Pap test had cytological abnormalities detected. In addition, few diagnosed cervical abnormalities cleared, even with intervention. These findings underscore the importance of obtaining sexual histories, providing counseling to prevent unintended pregnancies, screening for genital infections, administering routine Pap tests, and closely managing cervical lesions. Education on safer sexual practices is needed in this population.
Note. PACTG = Pediatric AIDS Clinical Trials Group; RNA = ribonucleic acid; CDC = Centers for Disease Control and Prevention; ART = antiretroviral therapy; HAART = highly active antiretroviral therapy (combination ART regimen of 3 or more drugs); PI = protease inhibitor; NNRTI = nonnucleoside reverse transcriptase inhibitor; NRTI = nucleoside reverse transcriptase inhibitor. aCharacteristics at or within 1 month of the last clinical visit. bGirls with missing data were not included in statistical test (Fisher exact test or χ2). cSignificantly different between the 2 groups (P < .05). dFor explanation of CDC disease classification, see reference 8. eRefers to the exposure of the adolescents themselves. Note. PACTG = Pediatric AIDS Clinical Trials Group; CI = confidence interval. aExcluded from the estimated incidence of condyloma are 23 girls in the “all girls” group and 5 girls in the “sexually active girls” group with a diagnosis of genital condyloma before age 13 years.
Characteristica Sexually Active (n = 174)b Not Sexually Active (n = 464)b Mean age (range),c y 19.0 (13.0–24.7) 15.3 (13.0–22.6) Race/ethnicity,c no. (%) Black 92 (52.9) 262 (56.5) Hispanic 62 (35.6) 120 (25.9) White 17 (9.8) 76 (16.4) Other/unknown 3 (1.8) 6 (1.3) Living conditions,c no. (%) With biological parent(s) 49 (28.2) 153 (33.0) With other adult(s) 54 (31.0) 165 (35.6) With adoptive parent(s) 22 (12.6) 113 (24.4) Foster care or residential facility 9 (5.2) 28 (6.0) On own 40 (23.0) 5 (1.1) CDC disease classification,d no. (%) A or N 50 (28.7) 159 (34.3) B 63 (36.2) 185 (39.9) C 61 (35.1) 118 (25.4) Unknown 0 (0.0) 2 (0.4) HIV RNA, copies/mL,c no. (%) <1000 41 (23.6) 212 (45.7) 1000 to <10 000 33 (19.0) 70 (15.1) ≥10 000 44 (25.3) 82 (17.7) Unknown 56 (32.2) 100 (21.6) CD4+, no. (%) <15 28 (16.1) 41 (8.8) 15–24 41 (23.6) 78 (16.8) ≥25 58 (33.3) 267 (57.5) Unknown 47 (27.0) 78 (16.8) In utero ART exposure,e no. (%) Yes 0 (0.0) 9 (1.9) No 149 (85.6) 399 (86.0) Unknown 25 (14.4) 56 (12.1) Received ART as a neonate, no. (%) Yes 0 (0.0) 2 (0.4) No 174 (100.0) 462 (99.6) Currently on ART,c no. (%) Yes 134 (77.0) 405 (87.3) No 40 (23.0) 59 (12.7) Current ART regimen (girls on ART only), no. (%) HAART with PI and NNRTI 18 (13.4) 51 (12.6) HAART with PI 82 (61.2) 226 (55.8) HAART with NNRTI 14 (10.5) 57 (14.1) ≥NRTIs 9 (6.7) 17 (4.2) Other ART 11 (8.2) 54 (13.3) Cumulative Incidence (95% CI) Infection Incident Cases by Age 19 Years, % All Girls (n = 638) Sexually Active Girls (n = 174) Condylomaa 12 4.6 (1.7, 7.4) 8.2 (3.7, 12.7) Trichomoniasis 9 4.8 (1.4, 8.3) 6.9 (2.4, 11.5) Chlamydia 9 2.7 (0.9, 4.6) 5.5 (2.0, 9.1) Gonorrhea 6 2.2 (0.3, 4.1) 3.9 (0.8, 7.0) Syphilis 2 1.1 (0.0, 2.6) 1.6 (0.0, 3.7)
Note. PACTG = Pediatric AIDS Clinical Trials Group; RNA = ribonucleic acid; CDC = Centers for Disease Control and Prevention; ART = antiretroviral therapy; HAART = highly active antiretroviral therapy (combination ART regimen of 3 or more drugs); PI = protease inhibitor; NNRTI = nonnucleoside reverse transcriptase inhibitor; NRTI = nucleoside reverse transcriptase inhibitor.
aCharacteristics at or within 1 month of the last clinical visit.
bGirls with missing data were not included in statistical test (Fisher exact test or χ2).
cSignificantly different between the 2 groups (P < .05).
dFor explanation of CDC disease classification, see reference 8.
eRefers to the exposure of the adolescents themselves.
Note. PACTG = Pediatric AIDS Clinical Trials Group; CI = confidence interval.
aExcluded from the estimated incidence of condyloma are 23 girls in the “all girls” group and 5 girls in the “sexually active girls” group with a diagnosis of genital condyloma before age 13 years.
This work was supported by the Center for Biostatistics in AIDS Research at the Harvard School of Public Health (the statistical and data analysis center of the PACTG) under National Institute of Allergy and Infectious Diseases cooperative agreement (5 U01 AI41110). The US National Institute of Allergy and Infectious Diseases and the National Institute of Child Health and Human Development funded protocol 219C (under cooperative agreement U01 AI41089).
We thank the adolescents and families for their participation in PACTG protocol 219C and the individuals and institutions involved in the conduct of 219C.
Human Participant Protection The clinical centers involved in PACTG protocol 219C obtained approval from their institutional review boards for human research, and the child or the child’s guardian provided written informed consent.