Differences Between HIV-Infected Men and Women in Antiretroviral Therapy Outcomes — Six African Countries, 2004–2012
|MMWR Weekly |
Volume 62, No. 47
November 29, 2013
PDF of this issue
Differences Between HIV-Infected Men and Women in Antiretroviral Therapy Outcomes — Six African Countries, 2004–2012
November 29, 2013 / 62(47);946-952
Evaluation of differences between human immunodeficiency virus (HIV)-infected men and women in antiretroviral therapy (ART) enrollment characteristics and outcomes might identify opportunities to improve ART program patient outcomes and prevention impact. During September 2008–February 2012, retrospective cohort studies to estimate attrition of enrollees (i.e., from death, stopping ART, or loss to follow-up) at 6-month intervals after ART initiation were completed among samples of adult men and women (defined as aged ≥ 15 years or aged ≥ 18 years) who initiated ART during 2004–2010 in six African countries: Côte d'Ivoire in western Africa; Swaziland, Mozambique, and Zambia in southern Africa; and Uganda and Tanzania in eastern Africa. Records for 13,175 ART enrollees were analyzed; sample sizes among the six countries ranged from 1,457 to 3,682. In each country, women comprised 61%–67% of ART enrollees. Median CD4 count range was 119–141 cells/µL for men and 137–161 cells/µL for women. Compared with women, a greater percentage of men initiated ART who had World Health Organization (WHO) HIV stage IV disease. In cohorts from western Africa and southern Africa, the risk for attrition was 15%–26% lower among women compared with men in multivariable analysis. However, in eastern Africa, differences between men and women in risk for attrition were not statistically significant. Research to identify country-specific causes for increased attrition and delayed initiation of care among men could identify strategies to improve ART program outcomes among men, which might contribute to prevention of new HIV infections in female partners.
In each of the six countries, a representative sample of ART facilities was selected. To keep the studies feasible, small facilities were excluded from the sample frames of countries with > 100 ART facilities at the time of sampling. Therefore, in Côte d'Ivoire and Mozambique, facilities that had enrolled < 50 adults on ART were excluded, whereas in Zambia, Uganda, and Tanzania, facilities that had enrolled < 300 adults on ART were excluded (Table 1).
From the eligible number of facilities in Côte d'Ivoire (78), Swaziland (31), and Mozambique (94), totals of 34, 16, and 30 study facilities, respectively, were randomly selected using probability-proportional-to-size sampling (Table 1). From the eligible number of facilities in Zambia (129), Uganda (114), and Tanzania (85), six study facilities were purposefully (nonrandomly) selected in each country to represent different types of ART facilities.
At each selected facility, a sample frame of study-eligible ART patients was created, and simple random sampling used to select the desired sample size of patient medical records. Eligibility criteria included having started ART during 2004–2010 and ≥ 6 months before data abstraction. Data were abstracted by trained study personnel from ART medical records onto standardized abstraction forms.
Attrition was the primary outcome of interest. A patient was considered lost to attrition if the record showed 1) the patient had died, 2) the patient had stopped ART because of a personal or clinician decision, or 3) the patient had not attended the facility in the 90 days preceding data abstraction for either medication refill or a clinician visit, in which case the patient was considered lost to follow-up.
Variables routinely collected on Ministry of Health medical records at ART initiation including sex, age, CD4 count, WHO HIV disease stage, and ART regimen, were entered into standardized abstraction forms. Data were analyzed using statistical software, and study design was controlled for during analysis. Data for Côte d'Ivoire, Swaziland, and Mozambique were weighted to account for the probability-proportional-to-size sampling.
To estimate the effect of sex on attrition, Cox proportional hazards regression models were used to estimate unadjusted and adjusted hazard ratios, 95% confidence intervals, and p-values. Multivariate models included only cases with complete data for sex, age, CD4 count, WHO HIV disease stage, and ART regimen. The proportional hazards assumption was assessed using visual methods and the Grambsch and Therneu test (1). For all countries, a shared frailty model was used to account for intrafacility correlation. Chi-square tests were used to compare distributions of categorical variables, and t-tests were used to compare distributions of continuous variables between men and women. Attrition at 6-month intervals after ART initiation was analyzed using the Kaplan-Meier product-limit estimate of the survivor function.
The sample sizes in the six countries ranged from 1,457 to 3,682, with a total of 13,175 records analyzed (Table 2). In each country, 61%–67% of ART enrollees were women. In each country, men were significantly older than women at ART initiation. Median age range for men was 37–40 years, and for women was 32–35 years.
Compared with women, median CD4 count at ART initiation was significantly lower among men in Côte d'Ivoire (119/µL compared with 155/µL), Swaziland (121/µL compared with 161/µL), and Uganda (128/µL compared with 144/µL). Median CD4 count was lower for men compared with women in Mozambique, Zambia, and Tanzania, although these differences were not statistically significant (p> 0.05).
In all countries, men initiated ART at a more advanced WHO disease stage (Table 2). For example, a higher proportion of men than women initiated ART at WHO stage IV disease in Côte d'Ivoire (26% compared with 21%), Swaziland (18% compared with 10%), Mozambique (20% compared with 13%), Zambia (11% compared with 9%), Uganda (16% compared with 10%), and Tanzania (30% compared with 27%).
For both men and women across all six countries, nevirapine-containing first-line regimens were more common than efavirenz-containing or protease inhibitor–containing regimens, or triple nucleoside reverse transcriptase inhibitor regimens (Table 2). The distribution of first-line ART regimen choices for men differed from choices for women in all countries, with women more likely to be prescribed nevirapine-containing regimens than men (Table 2).
In all countries, point estimates for attrition during the first 4.5 years of ART were lower for women than men (Table 3). This difference was statistically significant in countries in western and southern Africa, where women had 21%–27% lower rates of attrition in unadjusted analysis and 15%–26% lower rates of attrition in adjusted analysis (Table 3). In Uganda and Tanzania, women had 7%–18% lower rates of attrition in unadjusted analysis and 11%–12% lower rates of attrition in adjusted analysis, although the associations between sex and attrition rates were not statistically significant in these two countries.
Virginie Ettiègne-Traoré, MD, Moise Zanga Tuho, MD, Ministry of Health; Fayama Mohamed, PhD, Directorate General of Budget and Finance, Côte d'Ivoire. Charles Azih, MD, Ministry of Health, Swaziland. Francisco Mbofana, MD, National Institute of Health, Mozambique. Modest Mulenga, MD, Tropical Diseases Research Center, Zambia. Fred Wabwire-Mangen, PhD, The Infectious Diseases Institute, Makerere Univ College of Health Sciences, Uganda. Gideon Kwesigabo, PhD, Muhimbili Univ of Health and Allied Sciences, Tanzania. Joseph Essombo, MD, Elizabeth Glaser Pediatric AIDS Foundation, Côte d'Ivoire. Harrison Kamiru, DrPH, Harriet Nuwagaba-Biribonwoha, MD, International Center for AIDS Care and Treatment Programs-Columbia Univ, Swaziland. Kwasi Torpey, PhD, FHI 360, Zambia. Eric Van Praag, MD, FHI 360, Tanzania. Ya Diul Mukadi, MD, FHI 360, Haiti. Olivier Koole, MD, Joris Menten, MSc, Robert Colebunders, PhD, Institute of Tropical Medicine, Dept of Clinical Sciences, Belgium. Lisa J. Nelson, MD, Dept of HIV/AIDS, World Health Organization, Switzerland. Georgette Adjorlolo-Johnson, PhD, Elizabeth Glaser Pediatric AIDS Foundation, California. Julie Denison, PhD, Sharon Tsui, MPH, Social and Behavioral Health Sciences, FHI 360, Washington, DC. Carol Dukes Hamilton, MD, Timothy Mastro, MD, Global Health, Population & Nutrition, FHI 360, Durham, North Carolina. David Bangsberg, MD, Massachusetts General Hospital. Kunomboa A. Ekra, MD, Joseph S. Kouakou, MD, Center for Global Health, CDC-Côte d'Ivoire. Peter Ehrenkranz, MD, Trong Ao, ScD, Center for Global Health, CDC-Swaziland. Charity Alfredo, MD, Kebba Jobarteh, MD, Center for Global Health, CDC-Mozambique. Seymour Williams, MD, Center for Global Health, CDC-South Africa. Ray W. Shiraishi, PhD, Andrew Baughman, PhD, Simon Agolory, MD, George Bicego, PhD, Thomas Spira, MD, Aaron Zee, MPH, Jonathan Kaplan, MD, Tedd V. Ellerbrock, MD, Andrew F. Auld, MBChB, Div of Global HIV/AIDS, Center for Global Health, CDC. Corresponding contributor: Andrew F. Auld, CDC, 404-639-8997, firstname.lastname@example.org.
Equitable access to ART for both men and women is a principle endorsed by most African governments and international donors, including the U.S. President's Emergency plan for AIDS Relief and the Global Fund to Fight HIV/AIDS, Tuberculosis, and Malaria (2,3). However, in Africa, proportionally more HIV-infected women are accessing ART services than men (4). Evaluating differences in ART enrollment characteristics and treatment outcomes between men and women might help program managers understand these differences in ART enrollment and identify opportunities for ART program improvement.
This report has two main findings: 1) among representative samples of adult ART enrollees in six African countries, men were more likely than women to initiate ART with advanced HIV disease, and 2) men had higher attrition risk than women after ART initiation. However, differences between men and women in rate of attrition varied by country, being larger and statistically significant in western and southern African cohorts, but smaller and not statistically significant in cohorts from eastern Africa (Uganda and Tanzania).
As in other reports from African ART programs (2,5–7), men initiated ART at more advanced HIV disease stages than women. Late initiation of ART among men has commonly been attributed to sex differences in health-seeking behavior, with men considered more likely to delay access to health care for various reasons, including stigma, male norms that discourage admitting ill health, and employment responsibilities (2). However, some have proposed that the prioritization of maternal and child health services by global and national public health organizations in Africa has resulted in inequitable access to health services, including ART (2,3). Recent reports suggest that whereas nearly all African countries include initiatives focused on women in their national AIDS strategies, only 10% of countries are effectively engaging men and boys in the national AIDS response (8). To address delayed enrollment in ART among men, increased attention from national governments and international donors to identify and implement evidence-based strategies that achieve earlier HIV testing and ART among men might be needed (2,3).
As in other studies from western and southern Africa (2,5,6,9), adjusting for possible baseline predictors of ART outcomes, including CD4 count, WHO HIV disease stage, age, and ART regimen, did not fully account for the increased rate of attrition among men in Côte d'Ivoire, Swaziland, Mozambique, and Zambia. This suggests unmeasured factors are contributing to either increased rates of death or loss to follow-up among men (2). Some reports have suggested sex differences in health-seeking behavior, biologic differences in response to ART, increased male risk for opportunistic infections, worse adherence to ART pill-taking among men, or background differences in mortality rates by sex in the general population are responsible for higher attrition rates among men taking ART (2).
In this analysis, although rates of attrition were marginally higher among men than women in Uganda and Tanzania, the effect of sex on attrition risk was smaller than that observed in cohorts from western and southern Africa. This might indicate variations in the effect of sex on attrition risk by country or region. One possible explanation is that internal and cross-border migration patterns vary by country and region. Historically, cross-border migration for work, which is more common among men than women, varies by region in Africa, being most common in western (10) and southern Africa, where South Africa is a hub for migrant labor from surrounding countries (10). Further research into variations in the effect of sex on attrition risk by country or region might inform interventions to reduce rates of male attrition.
The findings in this report are subject to at least two limitations. First, missing data for certain covariates of interest at ART initiation might have introduced some measurement error, and might have affected estimates of hazard ratios. Second, because of differences in cohort size, there was greater power to detect differences in outcomes between men and women in Swaziland, Mozambique, and Côte d'Ivoire than in Zambia, Uganda, and Tanzania, which limits ability to make conclusions about country variations in the effect of sex on attrition risk.
Across six countries in Africa, men initiated ART with more advanced disease and had statistically significant higher attrition in western and southern African cohorts. Reasons for differences between men and women in ART enrollment are not fully understood but might include lack of emphasis by donors and national governments on the importance of engaging men early in ART (3). Higher attrition rates among men are not fully explained by traditional predictors of poor outcomes (e.g., low CD4 count and advanced WHO HIV disease stage). Identifying and implementing evidence-based interventions to improve male enrollment and retention in ART programs is important to reduce male AIDS-related mortality and might contribute to prevention of new HIV infections in female partners (3).
- Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81:515–26.
- Cornell M, Schomaker M, Garone DB, et al. Gender differences in survival among adult patients starting antiretroviral therapy in South Africa: a multicentre cohort study. PLoS Med 2012;9(9)e1001304.
- Cornell M, McIntyre J, Myer L. Men and antiretroviral therapy in Africa: our blind spot. Trop Med Int Health 2011;16:828–9.
- Muula AS, Ngulube TJ, Siziya S, et al. Gender distribution of adult patients on highly active antiretroviral therapy (HAART) in Southern Africa: a systematic review. BMC Public Health 2007;7:63.
- Auld AF, Mbofana F, Shiraishi RW, et al. Four-year treatment outcomes of adult patients enrolled in Mozambique's rapidly expanding antiretroviral therapy program. PLoS One 2011;6(4)e18453.
- Stringer J, Zulu I, Levy J, et al. Rapid scale-up of antiretroviral therapy at primary care sites in Zambia: feasibility and early outcomes. JAMA 2006;296:782–93.
- Mosha F, Muchunguzi V, Matee M, et al. Gender differences in HIV disease progression and treatment outcomes among HIV patients one year after starting antiretroviral treatment (ART) in Dar es Salaam, Tanzania. BMC Public Health 2013;13:38.
- Joint United Nations Programme on HIV/AIDS. UNAIDS report on the global AIDS epidemic. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS; 2012. Available at http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2012/gr2012/20121120_unaids_global_report_2012_with_annexes_en.pdf .
- Toure S, Kouadio B, Seyler C, et al. Rapid scaling-up of antiretroviral therapy in 10,000 adults in Côte d'Ivoire: 2-year outcomes and determinants. AIDS 2008;22:873–82.
- African Development Bank. Migration patterns, trends and policy issues in Africa. Tunis, Tunisia: African Development Bank; 2010. Available at http://www.afdb.org/fileadmin/uploads/afdb/Documents/Project-related-Procurement/WORKING%20119%20word%20document%20AA.pdf .