Can Highly Active Antiretroviral Therapy Reduce the Spread of HIV?

EPIDEMIOLOGY AND SOCIAL SCIENCE Can Highly Active Antiretroviral Therapy Reduce the Spread of HIV? A Study in a Township of South Africa Bertran Auvert, MD, PhD,*†‡§ Sylvia Males,† Adrian Puren, MD, PhD,㛳 Dirk Taljaard,¶ Michel Caraël, PhD,# and Brian Williams, PhD** Background: Calls have been made for the large-scale delivery of highly active antiretroviral therapy (HAART) to people infected with HIV in developing countries. If this is to be done, estimates of the number of people who currently require HAART in high HIV prevalence areas of sub-Saharan Africa are needed, and the impact of the widespread use of HAART on the transmission and, hence, spread of HIV must be assessed. Objectives: To estimate the proportion of people eligible for combination antiretroviral therapy and to evaluate the potential impact of providing HAART on the spread of HIV-1 under World Health Organization (WHO) guidelines in a South African township with a high prevalence of HIV-1. Design: A community-based cross-sectional study in a township near Johannesburg, South Africa, of a random sample of approximately 1000 men and women aged 15 to 49 years. Methods: Background characteristics and sexual behavior were recorded by questionnaire. Participants were tested for HIV-1, and their CD4+ cell counts and plasma HIV-1 RNA loads were measured. The proportion of people whose CD4+ cell count was less than 200 cells/mm3 and who would be eligible to receive HAART under WHO guidelines was estimated. The potential impact of antiretroviral drugs on the spread of HIV-1 in this setting was determined by estimating among the partnerships engaged in by HIV-1–positive individuals the Received for publication July 1, 2003; accepted November 26, 2003. From *INSERM U88–IFR 69, Saint-Maurice, France; †AP-HP, Hôpital Ambroise Paré, Boulogne-Billancourt, France; ‡Université de Versailles, Saint-Quentin-en-Yvelines, France; §UFR médicale Paris, Ile-de-France– Ouest, Garches, France; 㛳National Institute for Communicable Diseases, Johannesburg, South Africa; ¶Progressus CC, Johannesburg, South Africa; #UNAIDS, Geneva, Switzerland; and **16 rue de la canonnière, Geneva, Switzerland. Supported by the Agence Nationale de Recherche contre le SIDA (ANRS2002-1265), Paris, France; National Institute for Communicable Diseases, Johannesburg, South Africa; and Institut National de la Santé et de la Recherche Médicale, Paris, France. Reprints: Bertran Auvert, INSERM U88, 14, rue du Val d’Osne, 94415 SaintMaurice Cedex, France (e-mail: bertran.auvert@paris-ouest. univ-paris5.fr). Copyright © 2004 by Lippincott Williams & Wilkins proportion of spousal and nonspousal partnerships eligible to receive HAART and then by calculating the potential impact of HAART on the annual risk of HIV-1 transmission due to sexual contacts with HIV-1–infected persons. The results were compared with those obtained using United States Department of Health and Human Services (USDHHS) guidelines. Results: The overall prevalence of HIV-1 infection was 21.8% (19.2%–24.6%), and of these people, 9.5% (6.1%–14.9%) or 2.1% (1.3%–3.3%) of all those aged 15 to 49 years would be eligible for HAART (ranges are 95% confidence limits). In each of the next 3 years 6.3% (4.6%–8.4%) of those currently infected with HIV-1 need to start HAART. Among the partnerships in which individuals were HIV-1–positive, only a small proportion of spousal partnerships (7.6% [3.4%–15.6%]) and nonspousal partnerships (5.7%, [3.0%– 10.2%]) involved a partner with a CD4+ cell count below 200 cells/mm3 and would have benefited from the reduction of transmission due to the decrease in plasma HIV-1 RNA load under HAART. Estimates of the impact of HAART on the annual risk of HIV-1 transmission show that this risk would be reduced by 11.9% (7.1%– 17.0%). When using USDHHS guidelines, the proportion of HIV-1– positive individuals eligible for HAART reached 56.3% (49.1%– 63.2%) and the impact of HAART on the annual risk of HIV-1 transmission reached 71.8% (64.5%–77.5%). Conclusions: The population impact of HAART on reducing sexual transmission of HIV-1 is likely to be small under WHO guidelines, and reducing the spread of HIV-1 will depend on further strengthening of conventional prevention efforts. A much higher impact of HAART is to be expected if USDHHS guidelines are used. Key Words: HAART, HIV, South Africa, CD4+, viral load, transmission (J Acquir Immune Defic Syndr 2004;36:613–621) S ub-Saharan Africa remains the region most severely affected by HIV/AIDS. Approximately 3.4 million new infections occurred in 2001, bringing the total number of people living with HIV/AIDS in this region to 28.1 million. In South Africa, the prevalence of HIV-1 infection is among the highest in the world, with almost 1 in 9 South Africans living with HIV/AIDS.1 J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 613 J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 Auvert et al In developed countries, striking improvements have been reported in the health status and life expectancy of HIVinfected patients as a result of the widespread use of antiretroviral therapy (ART).2–10 Ninety percent of people infected with HIV live in the developing world, however, where only 4% of those who need ART currently have access to the drugs they require.11 Several studies have evaluated the feasibility of delivering ART to patients in resource-limited settings,12,13 but despite international pressure to implement highly active antiretroviral therapy (HAART) in such countries, treatment requirements have yet to be precisely characterized. In developing countries with high HIV prevalence, such as South Africa, the fraction of the population that would be eligible for HAART under World Health Organization (WHO) guidelines11 is not precisely known. It is also not known how this fraction will change if the United States Department of Health and Human Services (USDHHS) guidelines14 are used. Such information is needed to calculate the cost of scaling up ART and to prepare health systems to deliver these treatments. In addition, the potential impact of the widespread use of antiretroviral drugs on the spread of HIV remains unclear and requires evaluation. There are biologic and epidemiologic reasons to believe that ART will reduce sexual transmission of HIV. Biologic studies have shown that antiretroviral drugs decrease HIV in seminal fluid15 and in cervicovaginal secretions.16 An epidemiologic study of discordant couples has shown that the use of zidovudine by infected men was associated with a 50% reduction in the risk of transmission of HIV to their female sexual partner,17 suggesting that the wide use of HAART would slow down the spread of HIV in the countries where the route of transmission is mainly heterosexual. The objectives of this study are to estimate the proportion of the population needing HAART in a township in South Africa under WHO guidelines, to estimate the short-term impact of providing ART on the spread of HIV, and to assess the impact of using USDHHS guidelines on these estimations. MATERIALS AND METHODS Survey In April 2002, a population-based cross-sectional study was conducted in a township 40 km south of Johannesburg, South Africa. Households were selected by a 2-stage random sampling technique. Index houses were randomly selected from a map obtained from the local municipal offices. Using each index house as a starting point, a cluster of households was identified by starting to the right of the index house and counting households around the street block and adjacent street blocks until 50 households had been reached. A selfweighting random sample of 20 households was then chosen from each cluster. All men and women aged 15 to 49 years who slept in the selected households the night before the study 614 team’s visit were eligible for inclusion in the study. The consent form was presented in the language of the respondent, who was invited to take part in the study, and those who agreed were asked to sign the consent form. The response rate was 68%. Eligible participants were transported to a local facility for the interviews and the collection of blood and urine samples. If eligible participants were not at home, the study team made up to 3 repeat visits on different days at different times. The field work was done in the late afternoon, when residents returned from work, so as to reach as many residents as possible. Field work was conducted on Monday to Thursday and again on Saturday to ensure that people who work during the week could be reached. When it was not possible for participants to go directly to the interviewing points or for household members who were not home, appointments were made at times that suited the participants, and these appointments were followed up. The questionnaire used in this study was based on a UNAIDS questionnaire.18 The interviewers completed the questionnaire during a private interview in the preferred language of the interviewee. Data were collected on background and behavioral characteristics. Sexual partners were divided into spousal and nonspousal partners. The spousal partners were partners to whom the respondents were married or lived with as married. The nonspousal partners were all the other partners. The questionnaire allowed for a detailed description of all the nonspousal partners during the last 12 months, including those with whom the respondent had only 1 sexual contact. In addition, specific questions were asked about the use of condoms during the last months with nonspousal partners. During the survey, participants with symptoms of sexually transmitted infections (STIs) were encouraged to go to the local STI clinic for treatment. Participants who wished to know their HIV status were offered a separate free enzymelinked immunosorbent assay (ELISA) test with pre- and posttest counseling to be arranged through the normal clinical channels. Blood samples were tested for syphilis, HIV-1, CD4+ count, and plasma HIV-1 RNA load. Urine samples were tested for chlamydial infection. When results were available, a trained nurse delivered the syphilis and Chlamydia infection test results directly to the participants. Participants with positive STI results were encouraged to seek treatment at the local STI clinic. Individuals with fewer than 200 CD4+ cells/mm3 were included in a specific program that involved voluntary counseling and testing, prevention of opportunistic infections, and access to ART. Pregnant women were informed of the possibility of reducing the mother-to-child transmission (MTCT) of HIV-1 during pregnancy and childbearing. The cost of transportation to health facilities where the MTCT program was available free of charge was borne by the study. © 2004 Lippincott Williams & Wilkins J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 Laboratory Procedures Following the interview, trained nurses collected whole blood and urine (first flow) samples. The urine samples were stored at 4°C and then transported the next day to the laboratory, where they were stored at −70°C before being analyzed. Two EDTA blood tubes of 20 mL of venous blood were taken and transported at room temperature to the laboratory the following morning. One tube was centrifuged at 400g for 10 minutes, and 5 aliquots of plasma were then taken and frozen at −70°C. The second tube of blood was used to determine the CD4+ count. An ELISA screen (Genscreen HIV1/2; version 2, BioRad, France and Wellcozyme HIV recombinant; Abbott Murex, Dartford, UK) and ELISA confirmation (Vironostika HIV Uni-Form II plus O; BioMerieux, Boxtel, Netherlands) were carried out on plasma to test for HIV-1 infection. Plasma HIV-1 RNA load was determined by reversetranscription polymerase chain reaction (PCR) using an assay designed to detect all M-group subtypes (Amplicor HIV-1 Monitor Test, version 1.5; Roche Diagnostic Systems, Branchburg, NJ).19 CD4+ cell counts were determined by BD FACSOUNT analysis (BD Biosciences, Belgium). Syphilis testing was performed using a rapid plasma reagin (RPR) screen (Macro-Vue RPR Card Tests; Becton Dickinson Microbiology Systems, Cockeysville, MD), followed by a fluorescent treponemal antibody absorption (FTA-ABS) confirmatory test (FTA-ABS Test Sorbent; BioMérieux, France). A positive RPR (at any titer) and FTA-ABS were taken as evidence of recently acquired and/or untreated syphilis. Urine samples were tested for Chlamydia infection using a qualitative DNA amplification method (Amplicor CT/NG Test; Roche Diagnostic Systems). Ethics Ethical clearance was obtained from the University of Witwaterstrand Committee for Research on Human Subjects on February 8, 2002 (protocol study no. M020103). Data Management Laboratory results and data generated from questionnaires were entered twice into a database (Microsoft Access, Redmond, WA) by different people. The 2 entries were compared, and discrepancies were corrected. The data were checked for inconsistencies. The files were then imported into the Statistical Package for Social Sciences (SPSS 8.0 for Windows, Chicago, IL) and prepared for statistical analysis. Statistical Methods Estimation and Statistical Tests The Clopper-Pearson method, which is known to produce slightly conservative 2-sided confidence intervals,20 was © 2004 Lippincott Williams & Wilkins HAART and the Spread of HIV used to estimate confidence limits of proportions. (Unless otherwise stated, ranges give 95% confidence limits.) Medians of quantitative data were calculated with their interquartile range (IQR). Quantitative data were compared between subgroups using the Kruskal-Wallis test. The correlation between plasma HIV-1 RNA load and CD4+ count was analyzed using the nonparametric Spearman correlation coefficient and by regression of log10 of the plasma HIV-1 RNA load against the CD4+ count. Estimation of the Proportion of the Population Needing HAART Under WHO Guidelines Current WHO recommendations are that all patients with CD4+ counts below 200 cells/mm3 should be offered ART,11 and we used this criterion to estimate the proportion of the 15- to 49-year-olds in that population currently requiring HAART. To estimate conservatively the proportion needing to start HAART in the next 3 years, we assumed that the CD4+ cell counts decline by an average 50 cells/mm3/y in untreated infected subjects.21–24 As a result, the proportion of the 15- to 49-year-old population needing to start HAART in the next 3 years under WHO guidelines was determined by the proportion of the sample with CD4+ counts in the range of 201 to 350 cells/mm3. Estimation of the Impact of HAART on the Short-Term Spread of HIV-1 Under WHO Guidelines Assuming that all people with CD4+ cell counts less than 200/mm3 receive HAART, we used 2 approaches to estimate the short-term impact of providing HAART on the spread of HIV-1. The first approach assumes that any HIV-1–positive person receiving HAART will become less infectious. We thus calculated among partnerships engaged in by HIV-1–positive individuals the proportion of spousal and nonspousal partnerships eligible to receive HAART and, as a consequence, the transmission of HIV-1 that could be reduced by HAART. In the second approach, we used data on the annual risk of HIV-1 transmission as a function of plasma HIV-1 RNA load from a study in Uganda25 to estimate the potential number of new HIV-1 infections per year per HIV-1–infected person. From this, we estimated the relative decrease in the annual risk of HIV-1 transmission assuming that HAART reduces plasma HIV-1 RNA load to fewer than 400 copies/mL. This approach allows for the contribution to the spread of HIV-1 by individuals with low CD4+ counts, who are more likely to have high plasma HIV-1 RNA loads. Impact of Using USDHHS Guidelines To assess the impact of using USDHHS guidelines on the proportion of the population requiring HAART, we recalculated this proportion using USDHHS guidelines recommending initiation of HAART with CD4+ counts below 350 cells/mm 3 or a plasma HIV-1 RNA load above 55,000 copies/mL.14 The impact of HAART on the short-term spread 615 J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 Auvert et al of HIV-1 under USDHHS guidelines was estimated as described previously. RESULTS Prevalence of HIV-1, Syphilis, and Chlamydial Infections Background Characteristics and Sexual Behavior Most households have electricity (89.0% [85.8%– 91.6%]) and piped water (88.2% [84.9%–90.8%]), but flush toilets are less common (31.2%, [27.2%–35.6%]). The median (IQR) of the combined monthly income per household is 884 (500–1500) South African Rands, corresponding to about 88 (50–150) Euros. The median (IQR) number of persons per household is 4 (3–5), and the median (IQR) number of persons per room is 1.4 (1–2). A total of 930 people agreed to participate in the survey. The male-to-female ratio was 1:1.12. Background characteristics of the sample are given in Table 1. At the time of the interview, 90.9% (88.8%–92.6%) of all participants reported having ever had sex. Among those who had experienced sexual intercourse, 34.7% (31.5%–38.0%) said that they had never used a condom. Among men and women who had experienced sexual intercourse, 68.7% (63.8%–73.2%) and 48.6% (43.9%–53.3%), respectively, reported having had at least 1 nonspousal sexual partner in the last 12 months. Among men and women who were married or living as married, the corresponding figures were 32.6% (24.9%–41.3%) and 11.7% TABLE 1. Background Characteristics of Men and Women included in the Survey Median age (IQR) years Ethnic group (% in each category) Sotho Tswana Xhosa Zulu Other Primary school completed (%) Occupation (% in each category) Employed Student Unemployed Other Marital status (%) Married or living as married Single 616 (7.8%–17.0%), respectively. Among those who had had sex in the last month with nonspousal partners, 39.2% (33.5–45.3%) reported that they always used condoms. Men (n = 438) Women (n = 492) Total (n = 930) 25 (19–33) 28 (20–37) 26 (20–35) 33.8 7.5 9.6 36.8 12.3 36.6 6.5 13.0 35.4 8.5 35.3 7.0 11.4 36.0 10.3 84.5 83.3 83.9 37.4 29.9 27.6 5.0 25.0 22.2 50.2 2.6 30.9 25.8 39.6 3.8 30.8 69.2 41.9 58.1 36.7 63.3 The overall prevalence of HIV-1 infection was 21.8% (19.2%–24.6%): 17.4% (14.1%–21.4%) among men and 25.7% (range: 21.9%–30.0%) among women. The highest prevalence of HIV-1 by age was 34.4% (19.2%–53.2%) among men aged 35 to 39 years and 46.4% (34.4%–58.7%) among women aged 25 to 29 years. The median age of HIV1–infected people was 31 (IQR: 26–37) years for men and 23 (IQR: 19–32) years for women. Among those having a spousal partnership and those having a nonspousal partnership, the prevalence of HIV-1 was 25.6% (21.1%–30.7%) and 23.8% (20.1%–27.9%), respectively. The prevalence of syphilis was 3.2% (1.9%–5.4%) for men and 9.6% (7.4%–12.8%) for women. The prevalence of Chlamydia infection was 6.2% (4.3%–8.9%) for men and 6.9% (5.0%–9.6%) for women. Distribution of Plasma HIV-1 RNA Load The median (IQR) plasma HIV-1 RNA load was 55,750 (10,750–172,000) copies/mL (4.7 [4.0–5.2] copies/mL − log10), and the difference between men and women was not significant (P = 0.59 by Kruskal-Wallis test). The median (IQR) plasma HIV-1 RNA load in participants with CD4+ counts less than 200 cells/mm3 was 160,000 (72,900–410,000) copies/mL (5.2 [4.9 – 5.6] copies/mL − log10), and in participants with CD4+ counts higher than 200 cells/mm3, it was significantly lower at 46,800 (9407–149,500) copies/mL (4.7 [4.0–5.2]) copies/mL − log10) (P = 0.000 by Kruskal-Wallis test). Distribution of CD4+ Counts The median (IQR) CD4+ cell count in the HIV-1– negative and –positive participants was 1128 (911–1371) cells/mm3 and 475 (321–735) cells/mm3, respectively, and the difference was statistically significant (P = 0.000 by KruskalWallis test). Among HIV-1–negative men, the median (IQR) CD4+ cell count was 1057 (850–1316) cells/mm3, and among women, it was slightly higher at 1180 (963–1436) cells/mm3 (P = 0.000 by Kruskal-Wallis test). Among HIV-1–positive participants, the median CD4+ count was 488 (321–740) cells/mm3 and not statistically different between men and women (P = 0.17 by Kruskal-Wallis test). The distribution of CD4+ cell counts in HIV-1–infected individuals is given in Figure 1. Characteristics of HIV-1–Infected Subjects Eligible for Antiretroviral Therapy Under WHO Guidelines Taking a CD4+ cell count of 200 cells/mm3 as the critical level for the initiation of HAART, 9.5% (6.1%–14.9%) of © 2004 Lippincott Williams & Wilkins J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 FIGURE 1. Distribution of CD4+ cell count among HIV-1– infected persons (with the upper limit of the 95% confidence interval). HIV-1–infected people, or 2.1% (1.3%–3.3%) of 15- to 49year-olds, should be provided with HAART. The median age of these people was 33 (IQR: 26–36) years, with a male-tofemale ratio of 1:2.2. Of these persons, 36.8% (20.4%–73.9%) were married and 47.4% (25.2%–70.5%) reported at least 1 nonspousal sexual partner in the last 12 months. At the time of the study, 18.9% (13.8%–25.2%) of HIV1–infected people had CD4+ cell counts between 200 and 349 cells/mm3 such that 6.3% (4.6%–8.4%) of those currently infected with HIV-1, or 1.4% (0.97%–1.9%) of 15- to 49-yearolds, should start HAART in each of the next 2 years. On average, plasma HIV-1 RNA load in copies/mL − log10 falls by a factor of 1.43 (1.32–1.56) for each 100 cells/mm3 decline in CD4+ count, as shown in Figure 2. There is substantial dispersion in the data, however, and a high proportion of individuals have a high plasma HIV-1 RNA load even though their CD4+ cell count is above 200 cells/mm3. HAART and the Spread of HIV 200 cells/mm3. Of the nonspousal partnerships in the last 12 months involving HIV-1–positive individuals, 5.7% (3.0%– 10.2%) had a CD4+ cell count below 200 cells/mm3. Of the spousal and nonspousal partnerships in the last 12 months engaged in by HIV-1–positive individuals, a total of 6.3% (3.8%–9.6%) had a CD4+ cell count below 200 cells/mm3 The plasma HIV-1 RNA load stratified by CD4+ cell count is given in Table 2. Combining the survey data from the present study with data on the annual risk of infection from a study in Uganda,26 we are able to estimate the probability that 1 current infection will give rise to a secondary infection in 1 year (annual risk of HIV-1 transmission). By assuming that those who receive HAART cease to be infectious, we also show in Table 2 that the provision of HAART to all infected subjects with a CD4+ cell count below 200 cells/mm3 will reduce the annual risk of HIV-1 transmission by 11.9% (7.1%– 17.0%). Impact of Using USDHHS Guidelines If HAART were given to individuals with a CD4+ cell count below 350 cells/mm3 or a plasma HIV-1 RNA load greater than 55,000 copies/mL, more people would need HAART and the impact on the short-term spread of HIV-1 would be greater. If this were done, we estimate (data not shown) that 56.3% (49.1%–63.2%) of people infected with HIV-1 would have required HAART at the time of the study. As a result, 50.0% (39.5%–60.5%) of spousal and 59.3% (52.0%–66.2%) of nonspousal partnerships would potentially benefit from the resulting reduction of transmission, and the annual risk of HIV-1 transmission would be reduced by 71.8% (64.5%–77.5%). DISCUSSION Impact of HAART on the Short-Term Spread of HIV-1 Under WHO Guidelines Among spousal partnerships involving HIV-1–positive individuals, 7.6% (3.4%–15.6%) had a CD4+ cell count below FIGURE 2. Plasma HIV-1 RNA load by CD4+ count among HIV-1–positive individuals. The regression line is plasma HIV-1 RNA load (copies/mL ⳮ log10) = 5.40 ⳮ 1.57 10ⳮ3 CD4+ count (Spearman ␳ = ⳮ0.53, P = 0.000). © 2004 Lippincott Williams & Wilkins Number of People Who Need HAART Marked improvements have been reported in the health status and life expectancy of HIV-1–infected individuals and coincide with the widespread use of antiretroviral drugs.2–10 It is estimated that only 230,000 HIV-infected people in poor and middle-income countries are currently being treated with ART and that half of these people live in Brazil. In sub-Saharan Africa, where 70% of the HIV-1–infected people live, almost no or limited triple-combination HIV-1 treatment is used.24 Several studies have investigated the feasibility and efficacy of ART in resource-poor settings and confirm that antiretroviral drugs can be successfully provided in developing countries.12,13 Treatment requirements have not been well characterized in sub-Saharan countries, however. The current community-based study was conducted in a South African township of the Gauteng province with a very high prevalence of HIV-1. In this community, the prevalence of HIV and syphilis among women that we found (9.6% and 25.7%, respectively) was comparable to the prevalence (31.6% and 6.0%, respec- 617 J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 Auvert et al TABLE 2. Estimates of the Potential Impact of HAART on the Annual Risk of HIV-1 Transmission Total (6) Plasma HIV-1 RNA load (copies/mL) Annual risk of HIV-1 transmission (person/y) Percentage of HIV-1–positive population (%) Weighted annual risk of HIV-1 transmission without HAART (person/y) Percentage of HIV-1–positive population with CD4+ counts >200 cells/mm3 Weighted annual risk of HIV-1 transmission with HAART (person/year) <399 400–3499 3500–9999 10,000–49,999 >49,999 NA 0.04 0.12 0.14 0.23 NA 3.1 (1.1–6.5) 0 8.2 (4.7–12.9) 0.00327 (0.00188–0.00516) 12.2 (8.0–17.7) 0.0147 (0.00960–0.0212) 25.5 (19.6–32.2) 0.0357 (0.0274–0.0451) 51.1 (43.8–58.2) 0.118 (0.101–0.134) 0.171 (0.1591–0.183) 3.1 (1.1–6.5) 8.2 (4.7–12.9) 12.2 (8.0–17.7) 23.5 (17.5–30.0) 43.4 (36.3–50.6) 90.3 (85.3–94.1) 0.00327 (0.00188–0.00516) 0.0147 (0.00960–0.0212) 0.0329 (0.0245–0.0420) 0.0997 (0.0835–0.116) 0.151 (0.136–0.165) 0 0 100 The table gives estimates of the annual risk of HIV-1 transmission as a function of plasma HIV-1 RNA load,27 the proportion of the present population falling into each plasma HIV-1 RNA load band, the weighted annual risk of HIV-1 transmission, the proportion of the population that will not receive HAART under present guidelines, and the weighted annual risk of HIV-1 transmission with the provision of HAART. The decrease in the annual risk of HIV-1 transmission from 0.171/person/y without HAART to 0.151/person/y with HAART corresponds to a reduction of 11.9% (7.1%–17.0%). NA, not applicable. tively) reported for the same province in the antenatal survey conducted in 2002 by the South African National Department of Health.27 We estimate that 9.5% of all adults infected with HIV-1, or 2.1% of those aged 15 to 49 years, would be eligible for antiretroviral drugs under WHO guidelines that recommend initiating HAART at CD4 + counts less than 200 cells/mm3. In addition, a further 6.3% of all adults infected with HIV-1, or 1.4% of those aged 15 to 49 years, should start HAART each year. Because HAART will reduce mortality, the number of individuals who need HAART will increase. As data on HAART coverage and survival of patients receiving HAART in Africa become available, it will be possible to estimate more precisely the demand for HAART and the cost of providing it.28,29 Although the empiric data provided by this study represent an important first step, detailed public health and economic studies of the feasibility of testing and treating these individuals will allow for a precise estimate of the feasibility and cost of scaling up ART in settings with a high HIV prevalence. Such studies will make it possible to refine recent studies that have attempted to estimate the real cost of an effective response to the global AIDS epidemic.30,31 Estimates of the fraction of the population needing HAART in sub-Saharan Africa depend on the natural history of the epidemic and, in particular, on its magnitude and maturity. Because the epidemic in South Africa has developed recently but very rapidly,1,32 this study should be repeated in 618 other sub-Saharan African settings, especially in countries where the prevalence is lower and the epidemic is more mature and even declining. Nevertheless, it is unlikely that the proportion of HIV-positive people needing HAART will be dramatically different in other places in sub-Saharan Africa, where the epidemic is driven primarily by heterosexual contact. HIV-1 has the same impact on the immune system, and the epidemic has probably reached an endemic state even in more recently affected countries. When using USDHHS guidelines, we found that a high proportion (more than 50%) of people infected with HIV-1 would have needed HAART at the time of the study. The drastic difference found when using WHO and USDHHS guidelines is due to the strong difference in our HIV-positive population between the fraction having CD4+ counts below 200 cells/mm3 (about 10%) and the fraction having a viral load above 55,000 copies/mL (about 50%). This indicates that the proportion of people needing HAART critically depends on the choice of guidelines applied. A recent study conducted in the Rakai population in Uganda found that 20% of infected persons had viral loads greater than 55,000 copies/mL and thus needed to receive HAART under USDHHS guidelines.33 The reasons for such a difference when compared with our study are unclear. It could be due to the difference in the sample collection. Viral loads were determined from recently sampled plasma in our study and from archived serum in the Rakai study, and such variation in the collection and storage can at © 2004 Lippincott Williams & Wilkins J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 least partly explain the difference between the 2 studies.34 The difference could also be due to the population sample. We have designed a specific cross-sectional approach for our study compared with the cohort used in the Rakai study. Potential Impact of HAART on the Spread of HIV The impact that HAART might have on the spread of HIV by reducing the infectivity of treated subjects and preventing subsequent sexual transmission remains unclear.35 A mathematic model suggested that the widespread use of antiretroviral drugs could curb the HIV epidemic in a gay community in San Francisco.36 A recently published stochastic simulation showed that in Uganda, using USDHHS guidelines, HAART will have a limited impact on the spread of HIV.33 Here, we have used 2 indicators to evaluate the short-term impact of the widespread use of HAART on HIV-1 transmission in a South African township where the prevalence of HIV-1 is high. The first indicator revealed that only a small proportion of spousal partnerships (7.6%) and nonspousal partnerships (5.7%) formed by HIV-1–infected individuals will potentially benefit from the likely reduction of transmission due to the decrease in plasma HIV-1 RNA load induced by HAART. The second indicator suggests that the annual risk of HIV-1 transmission would be reduced by 11.9%. The low numeric values of the 2 indicators show that the impact of HAART on the spread of HIV, by reducing the infectivity of treated subjects and preventing subsequent sexual transmission, will be small under WHO guidelines. The first indicator takes into account the sexual activity of HIV-1–positive people in estimating the impact of HAART on the spread of HIV-1. Nevertheless, this indicator does not take into account that in a proportion of the partnerships of HIV-1–positive individuals, both partners could be HIV-1– positive. Therefore, this indicator is likely to be an overestimation of the proportion of sexual relationships that would benefit from the reduction of transmission due to the decrease in plasma HIV-1 RNA load by HAART. The second indicator was based on a study conducted in Uganda because of the lack of South African data on the relationship between HIV-1 viral load and sexual transmission of HIV-1. Therefore we cannot exclude that the result could have been slightly different if such data were available. Because the average plasma HIV-1 RNA load only increases slowly as CD4+ count falls, the infectiousness of individuals who are eligible for HAART is close to the infectiousness of those who are not. As a result, we can estimate the impact of HAART on the spread of HIV-1, as a first approximation, by the proportion of HIV-1–positive people who would receive HAART. This is confirmed by this study, where 9.5% of HIV-1–infected people need HAART and the reduction in transmission was estimated to be 11.9%. © 2004 Lippincott Williams & Wilkins HAART and the Spread of HIV Many people are not eligible for HAART because their CD4+ counts are above 200 cells/mm3 even though they have a high plasma HIV-1 RNA load, and they will continue to contribute substantially to the spread of HIV after the introduction of HAART. In addition, individuals who are in the early asymptomatic period and have a negative HIV test, that is, those who are newly infected or primary HIV-1–infected individuals are not taken into account in the current study, and it is believed that they contribute substantially to the spread of HIV because of their high plasma HIV-1 RNA load.37 This group is likely to be small, however, because the window period of the ELISA HIV test used here is short. Our estimation of the impact of HAART on the spread of HIV has been calculated assuming full coverage and that HAART reduces plasma HIV-1 RNA load to <400 copies/mL. Because we used a relationship between plasma HIV-1 RNA load and HIV transmission25 in which no transmission occurs with such a low value of viral load, our hypothesis is equivalent to assuming that HAART completely suppresses infectiousness. It thus seems likely that the estimation given in this study is an overestimation for at least 3 reasons. First, it is unlikely for cultural and practical reasons that all eligible people infected with HIV will receive HAART. Second, studies conducted in Africa have highlighted the problems of nonadherence and drug resistance among treated patients.38–44 Third, it is unlikely that HIV-infected individuals receiving HAART will be completely noninfectious. Indeed, in HIV-1– infected men who are successfully receiving HAART, the virus may still be present in seminal cells and transmitted sexually,45 and it has been shown that treating HIV-1–infected men with zidovudine reduces but does not eliminate heterosexual transmission of HIV.17 Finally, despite effective ART, high seminal plasma HIV-1 RNA loads occur during gonococcal urethritis46 and patients may still be infected as evidenced by continued shedding of cells harboring the HIV provirus.47 When using USDHHS guidelines, we found that the annual risk of HIV-1 transmission would be substantially reduced by more than two thirds. The drastic difference found when using WHO and USDHHS guidelines is due to the marked difference in the populations eligible for HAART under the 2 sets of guidelines. This indicates that the choice of guidelines to be used in a developing country is of critical importance on the potential impact of ART on the heterosexual spread of HIV. The aim of these guidelines is to recommend the point during the course of HIV infection at which ART should be initiated, but this point remains uncertain.48 A consequence of our study is that the design of guidelines for developing countries has serious public health implications that should be taken into consideration. The relatively small impact of HAART when using USDHHS guidelines that was found in Uganda is in contrast with what was found in our study. The difference is likely the result of the lower proportion of HIVpositive persons eligible for HAART in the Rakai population 619 Auvert et al J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 (about 20%) in comparison to the proportion in the South African population of this study (about 50%). Consequences for Prevention Under WHO guidelines, the limited effect of even the widespread use of HAART on the spread of HIV in subSaharan Africa found in this study suggests that HAART will not substantially reduce the heterosexual spread of HIV. Scaling up HAART should not lead to any relaxation in prevention efforts to reduce the spread of HIV. In particular, we advocate, as others have,49 that the budget allocated for prevention should not be in competition with the budget allocated to treatment and should not be reduced. In this context, the prevention of HIV infection should be based on established cost-effective prevention strategies such as condom distribution, blood and injection safety measures, treatment of STIs, and changes of sexual behavior.50 In the population under study here, the prevalence of both syphilis and Chlamydia is high, sexual risk behavior is common, and condom use is not optimal in nonspousal partnerships, although it is higher than in some other cities of subSaharan Africa.51–53 This situation, characterized by low condom use, high rates of curable STIs, and sexual risk behavior, is common in sub-Saharan Africa, and prevention efforts to reduce the spread of HIV need to be substantially strengthened. The development of HAART represents a major advance in the fight against HIV/AIDS, and even though HAART has to be administered by trained health workers in health centers satisfying the minimum requirements for the delivery of such drugs,11 there is no doubt that HAART, which is already available in some countries in sub-Saharan Africa such as Senegal,42 will soon become available in other countries. The consequences of the availability of HAART on prevention are difficult to predict. HAART could facilitate HIV prevention. The targeted availability of an effective therapy could lead to an increase in demand for HIV testing and counseling, which has been shown to be effective in reducing risky sexual behaviors in heterosexual couples,35,54 and to a lessening of the stigma associated with AIDS.13 HAART could also have a negative impact on prevention, however, because recent studies in San Francisco and Amsterdam have provided evidence of an increase in unprotected sex among men who have sex with men, possibly due to the availability of HAART.11,55,56 Detailed studies conducted in sub-Saharan Africa on the impact on prevention of the widespread use of HAART will be needed to judge adequately the complementary approaches of prevention and treatment of the HIV/AIDS epidemic in subSaharan communities highly infected by HIV. ACKNOWLEDGMENT The authors thank all the participants who agreed to take part in the survey, to answer the questions we put to them, and 620 to provide blood samples. Special thanks go to Reathe Taljaard and Gaph Pathedi, who helped to make the survey possible. Ewalde Cutler, Precious Magooa, Melody Nzama, Moses Mashiloane, and Japh Sibeko from the National Institute for Communicable Diseases, Johannesburg, South Africa, provided excellent technical assistance in regard to the laboratory testing. The authors thank Emmanuel Lagarde, INSERM U88, France, for his invaluable assistance and support. REFERENCES 1. UNAIDS. AIDS Epidemic Update. WHO, Geneva, Switzerland, 2001. 2. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver’s West End. Can J Public Health. 2000;91:125–128. 3. Vittinghoff E, Scheer S, O’Malley P, et al. Combination antiretroviral therapy and recent declines in AIDS incidence and mortality. J Infect Dis. 1999;179:717–720. 4. Moore RD, Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS. 1999;13:1933–1942. 5. Mocroft A, Vella S, Benfield TL, et al. Changing patterns of mortality across Europe in patients infected with HIV-1. EuroSIDA Study Group. Lancet. 1998;352:1725–1730. 6. Palella FJ, Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med. 1998;338: 853–860. 7. Dore GJ, Li Y, McDonald A, et al. Impact of highly active antiretroviral therapy on individual AIDS-defining illness incidence and survival in Australia. J Acquir Immune Defic Syndr. 2002;29:388–395. 8. Egger M, Hirschel B, Francioli P, et al. Impact of new antiretroviral combination therapies in HIV infected patients in Switzerland: prospective multicentre study. Swiss HIV Cohort Study. BMJ. 1997;315:1194–1199. 9. Hogg RS, Yip B, Kully C, et al. Improved survival among HIV-infected patients after initiation of triple-drug antiretroviral regimens. CMAJ. 1999;160:659–665. 10. Gange SJ, Barron Y, Greenblatt RM, et al. Effectiveness of highly active antiretroviral therapy among HIV-1 infected women. J Epidemiol Community Health. 2002;56:153–159. 11. World Health Organization. Scaling Up Antiretroviral Therapy in Resource Limited Settings: Guidelines for a Public Health Approach. WHO, 2002, Geneva, Switzerland. 12. Laurent C, Diakhate N, Gueye NF, et al. The Senegalese government’s highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS. 2002;16:1363–1370. 13. Farmer P, Leandre F, Mukherjee JS, et al. Community-based approaches to HIV treatment in resource-poor settings. Lancet. 2001;358:404–409. 14. US Department of Health and Human Services. Guidelines for the Use of ARV Agents in HIV-Infected Adults and Adolescents. Washington, DC. US Department of Health and Human Services; 2002. 15. Pereira AS, Kashuba AD, Fiscus SA, et al. Nucleoside analogues achieve high concentrations in seminal plasma: relationship between drug concentration and virus burden. J Infect Dis. 1999;180:2039–2043. 16. Hart CE, Lennox JL, Pratt-Palmore M, et al. Correlation of human immunodeficiency virus type 1 RNA levels in blood and the female genital tract. J Infect Dis. 1999;179:871–882. 17. Musicco M, Lazzarin A, Nicolosi A, et al. Antiretroviral treatment of men infected with human immunodeficiency virus type 1 reduces the incidence of heterosexual transmission. Italian Study Group on HIV Heterosexual Transmission. Arch Intern Med. 1994;154:1971–1976. 18. UNAIDS. Looking deeper into the HIV epidemic: a questionnaire for tracing sexual networks. In: Best Practice Collection, Key Material 98/27. Geneva: UNAIDS; 1998:1–24. 19. Triques K, Coste J, Perret JL, et al. Efficiencies of four versions of the AMPLICOR HIV-1 MONITOR test for quantification of different subtypes of human immunodeficiency virus type 1. J Clin Microbiol. 1999; 37:110–116. © 2004 Lippincott Williams & Wilkins J Acquir Immune Defic Syndr • Volume 36, Number 1, May 1 2004 20. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med. 1998;17:857–872. 21. World Health Organization. Basic Science in HIV/AIDS: An Update. Geneva: WHO; 1999:64. 22. Kaleebu P, Ross A, Morgan D, et al. Relationship between HIV-1 Env subtypes A and D and disease progression in a rural Ugandan cohort. AIDS. 2001;15:293–299. 23. Fauci AS, Pantaleo G, Stanley S, et al. Immunopathogenic mechanisms of HIV infection. Ann Intern Med. 1996;124:654–663. 24. Anastos K, Gange SJ, Lau B, et al. Association of race and gender with HIV-1 RNA levels and immunologic progression. J Acquir Immune Defic Syndr. 2000;24:218–226. 25. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med. 2000;342:921–929. 26. Gray RH, Wawer MJ, Brookmeyer R, et al. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet. 2001;357:1149–1153. 27. Department of Health. National HIV and syphilis antenatal seroprevalence survey in South Africa. Department of Health, Pretoria, South Africa, 2002. 28. Weidle PJ, Malamba S, Mwebaze R, et al. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients’ response, survival, and drug resistance. Lancet. 2002;360:34–40. 29. Frater AJ, Dunn DT, Beardall AJ, et al. Comparative response of African HIV-1-infected individuals to highly active antiretroviral therapy. AIDS. 2002;16:1139–1146. 30. Attaran A, Sachs J. Defining and refining international donor support for combating the AIDS pandemic. Lancet. 2001;357:57–61. 31. Schwartlander B, Stover J, Walker N, et al. AIDS. Resource needs for HIV/AIDS. Science. 2001;292:2434–2436. 32. Williams BG, Gouws E. The epidemiology of human immunodeficiency virus in South Africa. Philos Trans R Soc Lond B Biol Sci. 2001;356: 1077–1086. 33. Gray RH, Li X, Wawer MJ, et al. Stochastic simulation of the impact of antiretroviral therapy and HIV vaccines on HIV transmission; Rakai, Uganda. AIDS. 2003;17:1941–1951. 34. Ginocchio CC, Wang XP, Kaplan MH, et al. Effects of specimen collection, processing, and storage conditions on stability of human immunodeficiency virus type 1 RNA levels in plasma. J Clin Microbiol. 1997;35: 2886–2893. 35. Wood E, Braitstein P, Montaner JS, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa. Lancet. 2000;355:2095–2100. 36. Velasco-Hernandez JX, Gershengorn HB, Blower SM. Could widespread use of combination antiretroviral therapy eradicate HIV epidemics? Lancet Infect Dis. 2002;2:487–493. 37. Leynaert B, Downs AM, de Vincenzi I. Heterosexual transmission of human immunodeficiency virus: variability of infectivity throughout the course of infection. European Study Group on Heterosexual Transmission of HIV. Am J Epidemiol. 1998;148:88–96. 38. Weidle PJ, Malamba S, Mwebaze R, et al. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients’ response, survival, and drug resistance. Lancet. 2002;360:34–40. © 2004 Lippincott Williams & Wilkins HAART and the Spread of HIV 39. Vergne L, Malonga-Mouellet G, Mistoul I, et al. Resistance to antiretroviral treatment in Gabon: need for implementation of guidelines on antiretroviral therapy use and HIV-1 drug resistance monitoring in developing countries. J Acquir Immune Defic Syndr. 2002;29:165–168. 40. Adje C, Cheingsong R, Roels TH, et al. High prevalence of genotypic and phenotypic HIV-1 drug-resistant strains among patients receiving antiretroviral therapy in Abidjan, Cote d’Ivoire. J Acquir Immune Defic Syndr. 2001;26(5):501–506. 41. Nyazema NZ, Khoza S, Landman I, et al. Antiretrovial (ARV) drug utilisation in Harare. Cent Afr J Med. 2000;46:89–93. 42. Laurent C, Diakhate N, Gueye NF, et al. The Senegalese government’s highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS. 2002;16:1363–1370. 43. Frater AJ, Dunn DT, Beardall AJ, et al. Comparative response of African HIV-1-infected individuals to highly active antiretroviral therapy. AIDS. 2002;16:1139–1146. 44. Harries AD, Nyangulu DS, Hargreaves NJ, et al. Preventing antiretroviral anarchy in sub-Saharan Africa. Lancet. 2001;358:410–414. 45. Zhang H, Dornadula G, Beumont M, et al. Human immunodeficiency virus type 1 in the semen of men receiving highly active antiretroviral therapy. N Engl J Med. 1998;329:1803–1809. 46. Sadiq ST, Taylor S, Kaye S, et al. The effects of antiretroviral therapy on HIV-1 RNA loads in seminal plasma in HIV-positive patients with and without urethritis. AIDS. 2002;16:219–225. 47. Vernazza PL, Troiani L, Flepp MJ, et al. Potent antiretroviral treatment of HIV-infection results in suppression of the seminal shedding of HIV. The Swiss HIV Cohort Study. AIDS. 2000;14:117–121. 48. Phillips AN, Lepri AC, Lampe F, et al. When should antiretroviral therapy be started for HIV infection? Interpreting the evidence from observational studies. AIDS. 2003;17:1863–1869. 49. Marseille E, Hofmann PB, Kahn JG. HIV prevention before HAART in sub-Saharan Africa. Lancet. 2002;359:1851–1856. 50. Creese A, Floyd K, Alban A, et al. Cost-effectiveness of HIV/AIDS interventions in Africa: a systematic review of the evidence. Lancet. 2002; 359:1635–1643. 51. Carael M, Cleland J, Deheneffe JC, et al. Sexual behaviour in developing countries: implications for HIV control. AIDS. 1995;9:1171–1175. 52. Auvert B, Buve A, Ferry B, et al. Ecological and individual level analysis of risk factors for HIV infection in four urban populations in sub-Saharan Africa with different levels of HIV infection. AIDS. 2001;15(Suppl 4): S15–S30. 53. Lagarde E, Auvert B, Chege J, et al. Condom use and its association with HIV/sexually transmitted diseases in four urban communities of subSaharan Africa. AIDS. 2001;15(Suppl 4):S71–S78. 54. Roth DL, Stewart KE, Clay OJ, et al. Sexual practices of HIV discordant and concordant couples in Rwanda: effects of a testing and counselling programme for men. Int J STD AIDS. 2001;12:181–188. 55. Dukers NH, Goudsmit J, de Wit JB, et al. Sexual risk behaviour relates to the virological and immunological improvements during highly active antiretroviral therapy in HIV-1 infection. AIDS. 2001;15:369–378. 56. Katz MH, Schwarcz SK, Kellogg TA, et al. Impact of highly active antiretroviral treatment on HIV seroincidence among men who have sex with men: San Francisco. Am J Public Health. 2002;92:388–394. 621