To the Editor:
Little is known about the frequency of hormonal contraceptive use among HIV-infected women. To evaluate the trends and correlates of medroxy-progesterone acetate (MDPA), levonorgestrel implants, and oral contraceptive pills (OCPs) in this population, a retrospective study was conducted in a large urban clinic in New Orleans. The Medical Center of Louisiana at New Orleans HIV Outpatient Program (HOP) is a Centers for Disease Control (CDC)–supported Adult Spectrum of Disease (ASD) Study site. The ASD database contains selected demographic, clinical, and laboratory information collected through retrospective chart reviews every 6 months. HIV RNA, or viral load, information was systematically collected only after January 1, 1997. The ASD database does not systematically collect information on condom use, hysterectomies, and tubal ligations.
Trends in selected characteristics, including demographic information, clinical information, pregnancy frequencies, and use of specific hormonal contraceptive methods, were determined among 3 cohorts of women enrolled into HOP aged 13 to 42 years using data from the ASD database. The 3 cohorts included women enrolled during 3 different periods: January 1, 1994 through December 31, 1996; January 1, 1997 through December 31, 1999; and January 1, 2000 through December 31, 2002. Factors associated with DMPA and OCP use were determined for the entire cohort.
Two primary statistical analyses were performed on the entire sample of women. A Cochran-Armitage test for trend was used to compare trends in demographic characteristics, clinical factors, contraceptive method choices, and pregnancy frequencies over the 3 different periods. To evaluate factors associated with DMPA and OCP use over time, logistic regression was performed with generalized estimating equations (GEEs). Time was included in each logistic model to account for any secular trends. Factors identified as significant in the bivariate or crude GEE models were included in a multivariate model. Because of the availability of viral load and highly active antiretroviral therapy (HAART) information only after 1996, 2 multivariate models were analyzed using data only from cohorts 2 and 3 (1997–2002). The first model excluded viral load and antiretroviral (ARV) information, and the second model included these variables. All statistical analyses were conducted using SAS version 8.0 (SAS Institute, Cary, NC), and statistical significance was determined using a cutoff of 0.05.
The demographic features, hormonal contraceptive choices, and proportion of women having at least 1 pregnancy for each of the 3 cohorts are shown in Table 1. Over time, the population became significantly older and contained a larger proportion of African Americans and a smaller proportion of women with a history of injection drug use. Although the proportion of women having a CD4 cell count less than 200 cells/mm3 remained the same, a significantly larger number of women with nondetectable viral load levels were in the most recent cohort compared with cohort 2. Trends in ARV use comparing the most recent cohort with cohort 2 showed a decrease in protease inhibitor (PI) use but an increase in nonnucleoside reverse transcriptase inhibitor (NNRTI) use.
The proportion of women experiencing pregnancy did not significantly change over time, but the highest frequency of hormonal contraceptive use occurred in the most recent cohort. Although the proportions of women choosing OCPs were similar for the 3 cohorts, trends for increasing MDPA use and decreasing levonorgestrel implants were statistically significant.
Factors associated with MDPA or OCP use in bivariate logistic regression (GEE) analyses are shown in Table 2, and multivariate analyses evaluating factors associated with MDPA and OCP use are shown in Table 3. Among all cohorts, younger women (14–24 years old) were significantly more likely to be taking MDPA than those aged 25 to 42 years. When predictors of MDPA use in the post-HAART era were considered, age, NNRTI therapy, and CD4 cell count remained significantly associated with the use of MDPA.
Younger women were more than twice as likely to be taking OCPs than those aged 25 to 42 years in all cohorts. African-American women, those with a history of injection drug use, and those with a CD4 cell count less than 200 cells/mm3 were significantly less likely to be taking OCPs. When only women in cohorts 2 and 3 were considered, age, race, injection drug use history, and CD4 cell count remained significantly associated with OCP use. Furthermore, women on PI therapy in cohorts 2 and 3 were less likely to be taking OCPs compared with those not on PI therapy.
This is the first study describing patterns and trends of hormonal contraceptive use spanning the pre- and post-HAART eras among a large urban cohort of HIV-infected women. Approximately 1 in 5 women aged 13 to 42 years in the study population used hormonal contraception. This proportion would be even higher if only nonsterilized women were included. Although information on tubal ligations and hysterectomies was not available from the ASD for the period of this study, prior published studies 1,2 showed proportions of the HOP clinic female population who had undergone a tubal ligation to be 30% for the period November 1990 through November 1993 and 27% for the period June 1994 through November 1995. Information on intrauterine devices and new forms of hormonal contraception was also not collected, but these methods were rarely used by the HOP clinic population before January 2003.
MDPA was the most common hormonal contraceptive method used, and its use consistently increased over time. The trends showing the increasing use of MDPA and decreasing use of levonorgestrel implants are even more dramatic if data from November 1990 through November 1993 are included. A study reviewing contraceptive methods used by the HOP female population at this time showed that only 5% used MDPA but that 15% had levonorgestrel implants. 1 The growing proportion of women on MDPA underlines the importance of investigating potential pharmacokinetic interactions between MDPA and selected ARV therapies. An ongoing AIDS Clinical Trial Group study should yield data on this area in the near future.
Although younger age, higher CD4 cell counts, and lack of injection drug use history were all associated with both MDPA and OCPs, African-American women preferred MDPA. MDPA was also used significantly more by women prescribed an NNRTI. This is not surprising given the concerns regarding early fetal exposure to efavirenz.
The results of this study cannot be generalized to other clinic populations. Geographic variations in hormonal contraception use among HIV-infected women are unknown, but information on presumed HIV-negative adolescents is available from Youth Risk Behavior Surveys conducted on a sample of 9th through 12th graders representative of all students in the United States. The Youth Risk Behavior study showed that more female adolescents used OCPs compared with MDPA (21.1% vs. 5.7% in 2001). 3 The relatively high proportion of HOP clinic women opting to use MDPA may be a regional phenomenon. Larger studies to evaluate national use and trends of hormonal contraceptive methods in the HIV-infected population would be of interest, particularly to assist in prioritizing pharmacokinetic studies between future ARV therapies and selected hormonal contraceptive methods.
Rebecca A. Clark, MD, PhD*
Katherine P. Theall, MPH†
*HIV Outpatient Program, Louisiana State University Health Science Center, New Orleans, Louisiana
†Department of Epidemiology, Tulane School of Public Health and Tropical Medicine, New Orleans, Louisiana
1. Clark RA, Kissinger P, Williams T. Contraception among HIV-infected women and adolescents. Int J STD AIDS
2. Clark R, Kissinger P, Bedimo A, et al. Determination of factors associated with condom use among women infected with human immunodeficiency virus. Int J STD AIDS
. 1997;8: 229–233.
© 2004 Lippincott Williams & Wilkins, Inc.
3. Anderson JE, Santelli J, Gilbert BC. Adolescent dual use of condoms and hormonal contraception. Sex Trans Dis