Menu
Menu
Menu
Menu
Objective: Adolescents living with HIV are sexually active and engaged in risky sexual behaviors. Knowledge on how and to what extent adolescents in HIV care are affected by pregnancy is needed so as to adopt better preventive services. We estimated 4-year pregnancy incidence and correlates among HIV-infected female adolescents in HIV care in urban Côte d’Ivoire. Design: We conducted retrospective analysis of a pediatric prospective cohort of the International epidemiological Databases to Evaluate AIDS (IeDEA) West Africa Collaboration. Female patients with confirmed HIV infection aged 10–19 years, having at least one clinical visit in 2009 to health facilities participating in the pediatric IeDEA West African cohort in Abidjan, Côte d’Ivoire, were included. Data on incident pregnancies were obtained through medical records and interviews with health professionals. Pregnancy incidence rate was estimated per 100 person-years (PY). Poisson regression models were used to identify factors associated with the first pregnancy and provided incidence rate ratios (IRR) with 95% confidence intervals (CI). Results: In 2009, 266 female adolescents were included, with a median age of 12.8 years (interquartile range, IQR: 10.0–15.0), CD4 cell counts of 506 cells/mm3 (IQR: 302–737), and 80% on antiretroviral treatment. At the 48th month, 17 new pregnancies were reported after 938 PY of follow-up: 13 girls had one pregnancy while 2 had two pregnancies. Overall incidence rate of pregnancy was 1.8/100 PY (95% CI: 1.1–2.9). High incidence was observed among those aged 15–19 years: 3.6/100 PY (95% CI: 2.2–5.9). Role of maternal death in the risk of pregnancy was at the limit of statistical significance (adjusted IRR: 3.1, 95% CI: 0.9–11.0; ref. non-maternal orphans). Conclusions: Incidence of pregnancy among HIV-infected adolescents in care aged 15–19 years reached a level observed in adult cohorts in Sub-Saharan Africa. Health personnel in pediatric care have to intensify their efforts to provide more realistic and age-adapted reproductive health services to meet the needs of adolescent patients already confronting issues of sexuality. Vulnerability of maternal orphans merits further investigation.
The International epidemiological Databases to Evaluate AIDS (IeDEA) initiative (www.iedea-hiv.org), launched in 2006, is a consortium of leading clinicians and epidemiologists. The present analysis was conducted in four health facilities participating in the pediatric IeDEA West African HIV cohort (pWADA) in Abidjan, Côte d’Ivoire, namely CIRBA, CePReF, Yopougon, and Cocody University Hospitals. HIV-infected children aged <10 years at the time of their HIV diagnosis were seen at least every 3 months according to national guidelines. All female patients with a confirmed HIV infection, who had at least one contact during the calendar year 2009 with one of the four pediatric pWADA clinics mentioned above, and aged 10–19 years at their first visit in 2009 were included in our study. Patients without any single follow-up visit during the study period were excluded from the analysis. Age, vital status of mother and father, weight, height, CD4 cell count, hemoglobin at the time of inclusion in the study, and the date of anti-retroviral therapy (ART) initiation (if on ART) were extracted from the pWADA database. When biological or clinical data were not recorded exactly on the date of the first visit in 2009 (i.e. baseline of our study), we used the most updated data within a range of 3 months. Episodes of pregnancies were sought through multiple sources. In most cases, they were self-declared and recorded in patient's files. In addition to these written records, health care workers were asked to list all pregnancies which could have occurred between 2009 and 2013 and to provide detailed information about each case. The information provided by a given personnel was cross-checked by other staffs to ensure its validity. When the information on gestational age, date of delivery, pregnancy outcome, or a child's HIV status could not be found in medical records, health care workers phoned the patients or relevant services (adult or obstetric) to acquire necessary information. Pregnancy intention for the first pregnancy was assessed by health personnel over the phone with the London Measure of Unplanned Pregnancy (LMUP) (www.lmup.com/) for those still in follow-up in 2016. The LMUP is a psychometrically validated measure of pregnancy and planning/intention (11). In the present study, we used a French version of this tool which we had adapted to the local context with midwives in the study sites through a translation back translation method (12). Each adolescent contributed to the denominator from the time they entered into the study until either 31 December 2013, date of the 20th birthday, date of death or loss-to-follow-up, date of transfer out, or date of first pregnancy, whichever occurred first. Loss to follow-up was confirmed when the patient did not report for any follow-up for at least 6 months, and for whom vital status could not be confirmed. Pregnant adolescents were right-censored at the estimated conception date and subsequently uncensored and re-included in analysis after a 15-month period in the case of a term live-birth. The 15 months of censoring was chosen based on 9 months of pregnancy and 6 months of post-partum abstinence, which is a common practice among women in Côte d'Ivoire (13). When a pregnancy was terminated by a spontaneous miscarriage or induced abortion, the adolescent was right-censored on the estimated date of conception and uncensored at the date of pregnancy interruption. Re-inclusion of adolescents with pregnancy episode allowed analysis of recurrent pregnancies. Delayed entry was allowed for those who were pregnant at baseline (n=2). Baseline characteristics were described by median values with interquartile range (IQR) for continuous variables and frequencies for categorical variables, stratified according to the occurrence of pregnancy over the study period from 2009 to 2013. Pregnancy incidence was calculated per 100 person-years (PY) of follow-up with their 95% confidence intervals (95% CI). Poisson regression method was used to estimate the incidence rate of pregnancy according to baseline patient characteristics. Factors associated with the occurrence of first pregnancy were analyzed using Poisson regression. Immunodeficiency was defined as <350 cells/mm3 as per the 2010 WHO recommendations for ART initiation (14). Age group was computed as a time-dependent variable, summarizing for each girl the number of months contributing to the period of a given age category. Univariate analyses were run with all the covariables of interest. The variables included in multivariate analyses were selected at the threshold of p=0.25 through a stepwise descending method. In the final model, statistical significance was considered two-sided when p<0.05. Unadjusted incidence rate ratios (IRRs) and adjusted IRR (aIRR) are reported with their 95% CI. Data were entered into Microsoft Access 2003, and all the statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC).
N/A
