Elsevier

Contraception

Volume 77, Issue 5, May 2008, Pages 371-376
Contraception

Original research article
Hormonal contraception and HIV prevalence in four African countries

https://doi.org/10.1016/j.contraception.2008.01.012Get rights and content

Abstract

Background

The HIV seroprevalence among women aged 15–24 years was compared according to their pattern of contraceptive use in four African countries: Kenya, Lesotho, Malawi and Zimbabwe.

Study Design

Data were derived from Demographic and Health Surveys (DHS) conducted between 2003 and 2006 on representative samples, totaling 4549 women.

Results

It is indicated that users of depo-medroxyprogesterone acetate (DMPA) have a significantly higher seroprevalence than nonusers [odds ratio (OR)=1.82, 95% CI=1.63–2.03] and higher than users of oral contraceptives and users of traditional methods. The results were confirmed in a multivariate analysis including as controls, age, duration since first intercourse, urban residence, education, number of sexual partners in the last 12 months and marital status. A somewhat smaller net effect (OR=1.34, 95% CI=1.10–1.63) was found. In contrast, oral contraceptives and traditional methods did not show any risk for HIV (OR=0.96 and 0.92, respectively).

Conclusion

The increased risk of DMPA was present in three of the four countries investigated, and significant in Zimbabwe and Lesotho, the countries with the highest HIV seroprevalence. The HIV risk attributable to DMPA remained small altogether and was estimated as 6% in the four countries combined.

Introduction

The relationship between hormonal contraception and susceptibility to the heterosexual transmission of HIV in developing countries remains controversial and has been the source of several reviews [1], [2], [3], [4], [5]. Part of the controversy stems from the variety of the epidemiological studies: some studies distinguish between oral contraceptives based on estrogens, and injectables or implants based on progestins, but others merged them together; some studies focused on young women, others on older women or women of any age; some studies focused on high risk or heavily exposed women, for example, commercial sex workers (CSWs) or sexually transmitted diseases (STDs) clinic patients, while others deal with general populations or subgroups: attendants of antenatal clinics or users of family planning clinics (FPC). Many studies had small sample size and not enough statistical power to confirm the effect. In addition, controls used for multivariate analysis were not always the same, and some are directly linked to the risk of HIV/AIDS, in particular patients with STDs.

First observations made in the early 1990s suggested an increased risk associated with hormonal contraception. In Thailand, a prospective study conducted among CSWs found a high risk ratio associated with the use of depot medroxyprogesterone acetate (DMPA) [6]. In Kenya, a similar prospective study among CSWs found a high risk ratio associated with oral contraceptives [7]. However, another study conducted later in Kenya found no effect of oral contraception but a significant effect of DMPA [8]. In Rwanda, a prospective study of pregnant women recruited in antenatal clinics found a significant effect of oral contraception [9]. A meta-analysis conducted in 28 countries suggested a 20% increased risk of HIV infection associated with oral contraception, and somewhat higher risks were found in Africa and when studies were conducted more rigorously [1]. Another review of 15 prospective studies concluded that hormonal contraception, and in particular DMPA, could increase the risk of HIV-1 acquisition among high-risk women but not in the general population [5]. Of course, various biases might be involved in these studies, since numerous interactions could exist between sexual behavior and the use of hormonal contraception [10].

Biological studies have also shown the complexity of the mechanisms involved in the differential susceptibility to HIV infection associated with hormonal contraception and the possible differential effects of estrogen-containing (pill) vs. progestin-only contraceptives (DMPA, implants). Simian models showed that simian immunodeficiency virus (SIV) transmission is possible through the vaginal epithelium [11] but that an intact lining is a formidable barrier to the virus [12]. Therefore, one could hypothesize that breaks in, or thinning of, the vaginal epithelium may be responsible for increased risk of vaginal transmission [13]. In simian models, progesterone leads to reversible vaginal thinning, whereas estrogens seem to be associated with thickening of the vaginal epithelium [14]. Sexual hormones can modify susceptibility to SIV among nonhuman primates, and progesterone may facilitate its vaginal transmission [15], whereas estrogens may be protective [16], [17]. In humans, the mechanisms might be somewhat different, and the physiological effects of progestins may be smaller than in monkeys [18], [19], [20]. Precise mechanisms by which hormones might modulate HIV susceptibility remain unclear [4]. Furthermore, age might be an important cofactor since the vaginal epithelium of adolescent and very young women is thinner than at older reproductive ages. In animal models, age was found as a factor of thickness and integrity of the vaginal epithelium [21]. In epidemiological studies, younger women appear most susceptible to HIV-1 infection, and a large proportion of HIV infections occurs prior to age 25 [9], [22].

The aim of this article is to help resolve these controversies from a demographic perspective. Some countries in sub-Saharan Africa have among the highest rates of HIV seroprevalence in the world and also use large amounts of hormonal contraception. In particular, young women of Southern and Eastern Africa often use DMPA and have high rates of HIV seroprevalence. Although use of DMPA among 15- to 24-year-old women is not higher than in many Latin American countries, where HIV seroprevalence is much lower, it might still be a risk factor. The Demographic and Health Surveys (DHS) provide an opportunity to investigate the relationship between HIV seroprevalence and hormonal contraception in the general population, since they provide both information for large samples of women and numerous potential cofactors. They permit the comparison of HIV seroprevalence among young women according to their pattern of use of hormonal contraception.

Section snippets

Study population

Demographic and Health Surveys are conducted on nationally representative samples of women in their reproductive ages, with sizes ranging from 3000 to 20,000 (see DHS Web site for more details [23]). They provide data on numerous aspects of demographic and health characteristics, in particular the use of contraceptive methods, and since 2001, some of the recent surveys include data on HIV seroprevalence. In Africa, 12 surveys with information on HIV were available at the time of this study,

Results

The samples in the four countries investigated (Kenya, Malawi, Lesotho and Zimbabwe) cumulated 4549 women aged 15–24 years, sexually active and with known HIV serologic status. The HIV prevalence differs between the countries and ranges from 8.2% in Kenya, 11.8% in Malawi, 18.9% in Zimbabwe, to 20.2% in Lesotho (Table 1). Contraceptive prevalence is rather high for African countries and varies from country to country. The proportion of women who ever used any method ranges from 47% in Kenya to

Discussion

Representative samples of the populations in the four countries investigated show that ever use of DMPA increased somewhat the risk of HIV among young women but that ever using the pill had no effect. The DMPA effect does not seem to be due to a selection bias associated with using contraception, since both traditional methods and the pill had no associated risk, since the effect was consistent in the four countries investigated and since the results were robust to various controls added in the

Acknowledgment

This study was not funded from an outside sort; all the authors are supported by their institutions and the study had no external cost.

References (29)

  • H.L. Martin et al.

    Hormonal contraception, sexually transmitted diseases, and risk of heterosexual transmission of human immunodeficiency virus type 1

    J Infect Dis

    (1998)
  • M. Bulterys et al.

    Incident HIV-1 infection in a cohort of young women in Butare, Rwanda

    AIDS

    (1994)
  • H. Sangi-Haghpeykar et al.

    Consistency of condom use among low-income hormonal contraceptive users

    Perspect Sex Reprod Health

    (2005)
  • C.J. Miller et al.

    Mechanism of genital transmission of SIV: a hypothesis based on transmission studies and the location of SIV in the genital tract of chronically infected female rhesus macaques

    J Med Primatol

    (1992)
  • Cited by (21)

    • Contraceptive method and self-reported HIV status among women in Malawi

      2017, Contraception
      Citation Excerpt :

      The link between DMPA and increased risk of HIV acquisition among women remains unclear [10]. Contrary to the current findings, an earlier study of DHS data from 2003 to 2006 in four African countries, including Malawi, found that progestin-only injectable users had higher, although nonsignificant, HIV seroprevalence than nonusers [15]. This may be because that study used an analysis population of women aged 15–24, whereas the present analysis included a wider range of reproductive-aged women.

    • Increased prevalence of sexually transmitted viral infections in women: The role of female sex hormones in regulating susceptibility and immune responses

      2011, Journal of Reproductive Immunology
      Citation Excerpt :

      In a 10-year cohort study, women sex workers from Kenya who used either the injectable contraceptive DMPA or oral contraceptive pills were at increased risk of acquiring HIV-1 compared to women who did not use contraceptives (Lavreys et al., 2004). A recent study has also shown that of 4549 young, African women, DMPA users had higher seroprevalence for HIV-1 and that approximately 6% of HIV risk was attributable to DMPA use (Leclerc et al., 2008). In addition, women who use progesterone-based contraceptives displayed accelerated HIV-1 disease progression and mortality, compared to women who did not (Stringer et al., 2007).

    • Letter to the Editor

      2008, Contraception
    View all citing articles on Scopus
    View full text