Share this article on:

Have the explosive HIV epidemics in sub-Saharan Africa been driven by higher community viral load?

Abu-Raddad, Laith J.a,b,c,*; Barnabas, Ruanne V.c,e,f,*; Janes, Hollyc,g; Weiss, Helen A.i; Kublin, James G.c; Longini, Ira M. Jrj,k; Wasserheit, Judith N.d,e,f,hthe HIV Viral Load Working Group

doi: 10.1097/QAD.0b013e32835cb927
Epidemiology and Social

Objective: The HIV epidemic has carved contrasting trajectories around the world with sub-Saharan Africa (SSA) being most affected. We hypothesized that mean HIV-1 plasma RNA viral loads are higher in SSA than other areas, and that these elevated levels may contribute to the scale of epidemics in this region.

Design and methods: To evaluate this hypothesis, we constructed a database of means of 71 668 viral load measurements from 44 cohorts in seven regions of the world. We used linear regression statistical models to estimate differences in viral load between regions. We also constructed and analyzed a mathematical model to describe the impact of the regional viral load differences on HIV epidemic trajectory.

Results: We found substantial regional viral load heterogeneity. The mean viral load in SSA was 0.58 log10 copies/ml higher than in North America (95% confidence interval 0.45–0.71); this represents about a four-fold increase. The highest mean viral loads were found in Southern and East Africa, whereas in Asia, Europe, North America, and South America, mean viral loads were comparable. Mathematical modeling indicated that conservatively 14% of HIV infections in a representative population in Kenya could be attributed to the enhanced infectiousness of patients with heightened viral load.

Conclusion: We conclude that community viral load appears to be higher in SSA than in other regions and this may be a central driver of the massive HIV epidemics in this region. The elevated viral loads in SSA may reflect, among other factors, the high burden of co-infections or the preponderance of HIV-1 subtype C infection.

Supplemental Digital Content is available in the text

aInfectious Disease Epidemiology Group, Weill Cornell Medical College – Qatar, Cornell University, Qatar Foundation – Education City, Doha, Qatar

bDepartment of Public Health, Weill Cornell Medical College, Cornell University, New York, New York

cVaccine and Infectious Disease Division

dClinical Research Division, Fred Hutchinson Cancer Research Centre, Seattle, Washington

eDepartment of Global Health

fDepartment of Medicine

gDepartment of Biostatistics

hDepartment of Epidemiology, Schools of Medicine and Public Health, University of Washington, Seattle, Washington, USA

iMRC Tropical Epidemiology Group, Faculty of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UK

jDepartment of Biostatistics

kEmerging Pathogens Institute, Colleges of Public Health and Medicine, University of Florida, Gainesville, Florida, USA.

*These authors contributed equally to this study.

Correspondence to Dr Laith J. Abu-Raddad, Infectious Disease Epidemiology Group, Weill Cornell Medical College – Qatar, Qatar Foundation – Education City, P.O. Box 24144, Doha, Qatar. Tel: +974 4492 8321; fax: +974 4492 8333; e-mail:

Received 29 July, 2012

Revised 7 November, 2012

Accepted 16 November, 2012

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (

Back to Top | Article Outline


Sub-Saharan Africa (SSA) continues to be the region hardest hit by the HIV pandemic. It is home to 70% of the world's estimated 2.7 million new HIV infections in 2010, as well as 68% of the estimated 34 million persons living with HIV [1]. Whereas in other regions HIV transmission is concentrated in populations with identifiable high-risk sexual or injecting-drug behaviors, in SSA HIV infection is prevalent in the general population with low levels of reported sexual risk behaviors [2]. Multiple behavioral, ecological, and biological host factors and viral characteristics may have contributed to the striking contrast between the HIV epidemic trajectory in SSA and other regions [2], such as differences in concurrency patterns and sexual networks [3], co-infections that drive higher HIV-1 plasma RNA viral load [4], and sub-type-specific viral dynamics [5]. Yet the interplay of factors contributing to disproportionately high HIV prevalence in the general population in SSA, and not in other regions, remains poorly understood.

HIV-1 viral load is a principal determinant of heterosexual HIV transmission [6,7]. Ample experimental and observational studies have shown that suppression of HIV viral load leads to large reduction in HIV transmission [8]. Decreases in community viral load have been associated with reductions in HIV incidence [9]. Each one log10 copies/ml increase in viral load has been found to more than double the per coital-act probability of HIV transmission [10,11].

In light of the link between HIV infectiousness and viral load level, we hypothesized that a specific biological cofactor contributed to the contrasting HIV epidemic trajectories between SSA and other regions; specifically we hypothesized that community level HIV-1 viral load in SSA is higher than in other regions, and is substantially higher to the extent that it can drive divergent epidemic trajectories between SSA and other regions. We examined this hypothesis by addressing two specific research questions: Are viral load levels higher in SSA compared to other regions? Can these higher viral loads explain, in part, the fulminant HIV transmission dynamics observed in SSA, in contrast to other regions?

Back to Top | Article Outline


Participants and settings

We obtained the mean HIV-1 plasma RNA viral load from prospective cohorts of HIV-infected, antiretroviral therapy (ART)-naive study participants across seven geographic regions (Table 1 and Supplemental Digital Content (SuppDC) Table S3.1, Overall, the studies included 15 cohorts of men, 10 cohorts of pregnant women, 14 cohorts of nonpregnant women, five cohorts of women with unknown pregnancy status, and a total of 71 668 viral load measurements (SuppDC Sections 1 and 3, Data from each cohort consisted of the mean viral load for participants in the following CD4 categories: below 200, 200–349, 350–499, and at least 500 cells/μl by sex and pregnancy status. In addition, for each study, we captured the geographic location, viral load assay used, predominant subtype in the region, and intervention description. We excluded studies in which the intervention had a potential impact on viral load, such as acyclovir, or used only baseline enrolment data.

Table 1

Table 1

Back to Top | Article Outline

Statistical methods

Mean log10 viral load was modeled using a linear regression model with the following predictors: indicators of region (North America, Europe, Asia, South America, West Africa, East Africa, Southern Africa, and South Africa), sex, CD4 category, pregnancy status, and interactions between CD4 category and sex, and between CD4 and pregnancy status (SuppDC Section 3, Pregnancy status was included as a potential confounder because lower viral loads have been observed among pregnant women compared to nonpregnant women [12]. Of the 44 cohorts, the five with unknown pregnancy status were treated as not pregnant. The effects of sub-type and assay could not be statistically adjusted due to high co-linearity with region. The model weighted each observation by the sample size. Robust standard errors are reported for the estimated model coefficients (SuppDC Table 3.2, Sensitivity analyses were conducted to assess the impact of model assumptions (SuppDC Section 3.4,

Back to Top | Article Outline

Mathematical modeling methods

We assessed the epidemiological implications of the regional differences in viral load by constructing and analyzing a deterministic compartmental mathematical model describing HIV epidemic expansion in a representative SSA population; that of Kisumu, Kenya (SuppDC Section 4, The model calculates the proportion of incident HIV infections that are directly attributable to the viral load effect, or population-attributable fraction (PAF). It stratifies the population into compartments according to HIV sero-status and stage of HIV infection, sexual-risk activity group, and exposure to a biological cofactor that heightens viral load. This cofactor modulates HIV infectiousness according to the empirical relationship between viral load and HIV per-coital transmission probability as observed initially by Quinn et al. [10] in Uganda, and affirmed recently in the Partners in Prevention study [11]. Table 2 summarizes the key assumptions in our model and a detailed description of the model and its parameterization can be found in SuppDC Section 4,

Table 2

Table 2

Recognizing that multiple uncontrolled variables, such as type of viral load assay and changes in sexual activity due to co-infection-associated morbidities, may affect viral load or its effect on HIV epidemic trajectory, we reported conservative predictions for the impact of the viral load effect on HIV epidemic expansion and conducted multiple sensitivity and uncertainty analyses on the model predictions (SuppDC Sections 4.4.3-5 and 4.5, We assumed a 20% smaller increase in mean viral load in SSA relative to North America than emerged from our database analyses (0.46 log10 rather than 0.58 log10; the results assuming the full 0.58 log10 increase can be found in SuppDC Section 4.4.2, We also assumed a 20% sexual-activity reduction with the heightened viral load to account for potential illness-associated abstinence during co-infections that increase HIV-1 viral load. The latter is a conservative assumption because malaria morbidity, a frequent co-infection in SSA, generally lasts for less than 10% of the period of heightened viral load due to malaria [4,13–16]. Moreover, only a very small fraction of the prevalent herpes simplex virus type 2 (HSV-2) seropositive persons suffer from clinically apparent ulcers, and even among those, HSV-2 reactivations are predominately asymptomatic [17,18]. Of note, the assumed 20% reduction in sexual activity reflects a reduction throughout HIV infection natural history and not merely a reduction in sexual activity during a transient episode of a specific co-infection, which biases our model results towards a lower impact of the viral load effect. Finally, although co-infection-induced viral load does not appear to accelerate HIV disease progression to AIDS or death in SSA [19,20], we accounted for this potential effect in the model to assess whether such mechanism could influence our predictions (SuppDC Sections 4.4.3 and 4.5,

Back to Top | Article Outline


There was striking heterogeneity in mean viral load (Table 1). Compared with North America, viral load levels were significantly higher in SSA (Fig. 1). The estimated mean log10 viral load was 0.29 higher [95% confidence interval (CI) 0.11–0.47] in West Africa, 0.71 higher (95% CI 0.48–0.93) in East Africa, and 0.74 higher (95% CI 0.55–0.92) in Southern Africa, excluding the South African study sites (SuppDC Table S3.2, The estimated mean log10 viral load was modestly, but significantly, higher in Asia than in North America [0.14 higher (95% CI 0.03–0.26)], but no difference was seen in mean log10 viral load between Europe or South America and North America (P = 0.66 and 0.49, respectively).

Fig. 1

Fig. 1

Compared to North America, the mean difference in viral load across the three resource-poor sub-regions of SSA, East, West, and Southern Africa, with each sub-region weighted by its HIV-infected population size, is 0.58 log10 (95% CI 0.45–0.71). South Africa, with its substantially lower burden of tropical co-infections that drive higher HIV-1 viral load (see Discussion section), particularly malaria [21], and better access to care than much of SSA, was considered separately. Indeed, mean viral load in South Africa was lower than that in North America. However, this anomaly may be largely due to the use of the bDNA viral load assay for mainly subtype C infections which, in this laboratory, has been associated with consistently lower viral load results (SuppDC Figure S3.2, [22]. When data from this laboratory are excluded, mean viral load borders on being significantly higher in South Africa than in North America [0.19 log10 (95% CI −0.02 to 0.39) versus −0.33 log10 (95% CI −0.49 to −0.18) with inclusion of the bDNA assay results]. SuppDC Section 3.5, includes a discussion and further analysis of the data from South Africa.

Applying our mathematical model to assess the epidemiological implications of such differences in viral load in a setting representative of a hyper-endemic HIV epidemic in SSA (Kisumu, Kenya), the viral load effect substantially augmented HIV epidemic expansion (Fig. 2a). The viral load effect contributed an excess HIV prevalence (prevalence with no viral load effect subtracted from prevalence with the viral load effect) of 2.9% (95% CI 1.4–3.6%) at the epidemic peak, and 2.7% (95% CI 1.5–3.2%) at the endemic equilibrium (CI estimated using the CI of the mean viral load in resource-poor SSA). The viral load effect fuelled an excess HIV incidence rate (incidence rate with no viral load effect subtracted from incidence rate with the viral load effect) of 0.69 (95% CI 0.39–1.04) per 100 person-years at the epidemic peak, and 0.48 (95% CI 0.27–0.56) per 100 person-years at endemic equilibrium. The proportion of incident HIV infections that is directly attributable to the viral load effect (PAF) increased with time, reaching 14.4% (95% CI 7.7–20.5%) at the epidemic peak in the mid to late 1990s (Fig. 2b). The proportion of cumulative incident HIV infections that is directly attributable to the viral load effect (PAFcum) is 13.9% by 2010.

Fig. 2

Fig. 2

Our calculations show that the relative impact of the viral load effect is largest in the majority of the population exhibiting the lowest sexual risk behavior, the general population. HIV prevalence in the low-risk group increased by 22.5%, whereas that in the higher risk groups increased by 4.4–11.1% (SuppDC Table S4.3, We estimate that in Kisumu, with an adult population of about 200 000, the viral load effect has contributed, directly or indirectly through onward transmission, over 30 000 excess HIV infections from 1980 to 2010 out of a total of approximately 135 000 HIV infections during this period. Additional results of the epidemiological impact of the viral load effect can be found in SuppDC Section 4,

The substantial impact of the viral load effect was robust to multiple sensitivity and uncertainty analyses (SuppDC Sections 3.4, 4.4.3–5, and 4.5, We performed a broad range of sensitivity analyses to the magnitude of the heightened viral load, reduction in sexual activity associated with the heightened viral load (the heightened viral load could be due to co-infections that lead to co-morbidities as discussed in Methods and Discussion sections), and enhancement of the rate of disease progression associated with the heightened viral load. We found that a reduction in coital frequency with the elevated viral load greater than 34% is needed to balance the increase in infectiousness due to the viral load effect, whereas any increase in viral load greater than 0.25 log10 would result in considerable population-level impact for this effect. The strong effect was robust to variations in the model parameters, including both the behavioral and biological inputs used to parameterize the model. As expected, our model predictions are primarily sensitive to the exact magnitudes of both the heightened viral load and the assumed level of sexual activity reduction associated with it. The lower the reduction in sexual activity, the higher is the impact of the viral load effect.

Back to Top | Article Outline


We found substantial regional differences in viral load with SSA having about four times the viral load levels found elsewhere. The highest viral load levels were seen in Southern and East Africa. Modeling the impact of the elevated viral load in a representative population, Kisumu, Kenya, we demonstrated that the elevated viral load in SSA could be a central driver of the massive HIV epidemics that erupted in this part of the world. The viral load effect is particularly important driver of HIV transmission among the general population and may explain, in part, the general population HIV epidemics that uniquely characterize this region. The robustness of our results to multiple sensitivity and uncertainty analyses points to a potentially substantial role for this biological cofactor in shaping HIV epidemic trajectories in SSA and warrants further investigation.

The higher regional viral load in SSA poses a question about the causes of such elevated viral load levels. We suspect that the high burden of infectious diseases beyond HIV in SSA, particularly malaria, tuberculosis, HSV-2, helminthes, and other tropical diseases [20], may have caused to a large extent the higher viral load in SSA. Mounting epidemiological and laboratory evidence suggests that these co-infections induce transient, but substantial increases in HIV-1 viral load, due to enhanced HIV replication associated with the immune response (see SuppDC Section 2, for description of evidence). A recent systematic review and meta-analysis found that acute malaria increases HIV-1 viral load by 0·67 log10 (95% CI 0.15, 1.19), active tuberculosis by 0.40 log10 (95% CI 0.13–0.67), and HSV-2 infection by 0·18 log10 (95% CI 0.01, 0.34) [23]. Furthermore, high levels of serum immune activation markers have been found in African populations compared to those in industrialized countries [24–26], possibly reflecting extended exposure to a range of pathogens. Together, the comparatively high burden of infectious diseases in SSA, and the links among co-infections, immune activation and increased HIV replication, suggest that viral load may be higher in this region than elsewhere, and that immune activation due to co-infections may be an important causal mechanism.

The first work to explore the impact on HIV epidemic trajectory of a co-infection that increases HIV-1 viral load, for the case of malaria, suggested a small, but tangible effect in fuelling HIV spread [4]. But the impact of a single co-infection on HIV transmission may be relatively limited, whereas the combined effects of multiple recurrent or persistent co-infections could potentially result in considerable population-level impact. The transient effect of one co-infection increasing viral load may not be discernible at the population level, but the cumulative effect of multiple potentially overlapping co-infections on raising the viral load, throughout the course of HIV infection, may be substantial as schematized in Fig. 3 [27]. Such effect would alter the natural history of HIV infection for the individual and consequently the epidemic trajectory for the community.

Fig. 3

Fig. 3

Another reason, in addition to co-infections, that may contribute to explaining the elevated mean viral loads in SSA is virus subtype. Emerging data suggest that persons infected with HIV-1 subtype C infection, the dominant type in SSA particularly in Southern Africa, may maintain high viral loads even after acute infection [28].

Our approach assumes a log-linear relationship between HIV-1 viral load and HIV per coital-act transmission probability, based primarily on empirical evidence from cohort studies of sero-discordant couples [10,11]. These studies may under or over-estimate the relationship between viral load and risk of transmission due to selection bias such as the selection of more ‘resistant’ couples in the recruitment of discordant partnerships with high viral load-infected individuals. A linear relationship between viral load and risk of transmission, assuming implicitly, for example, a simplistic concept for the establishment of the infection in terms of clonal expansion of a single infecting virion, would imply a stronger effect for the higher viral load on transmission, and more differential epidemiologic impact as a consequence of the differences in community viral load.

To our knowledge, the database we assembled in this study of summary measures of more than 70 000 viral load measurements from cohorts representing every major region of the globe, is the most extensive ever analyzed to assess the regional viral load differences and their impact on HIV transmission. However, there are limitations to these exploratory analyses. We used heterogeneous datasets from studies not designed to investigate this effect. Our hypothesis of higher-community viral load in SSA was examined using regional viral load data measured on participants recruited for different reasons (Table S3.1). The recruitment strategies could, in principle, be a source of selection bias where the viral load data were measured on individuals who may not necessarily represent the community of HIV-infected persons, and this may confound the assessment of the regional viral load differences. Of note, however, this vast amount of viral load data was collected from participants who were recruited using diverse recruitment strategies. The diversity of the modes of recruitment suggests a minimal bias in recruitment of persons who are more likely to have higher (or lower) viral load in one region as opposed to another.

The widest difference in viral load between SSA and developed settings was observed in the CD4 at least 500 category (Table 1), a category that potentially could encompass samples including persons within acute infection. The observed differences in viral load could accordingly reflect differences in the distribution of persons across HIV natural history stages. However, all except one of the viral load studies in this analysis have targeted recruitment of HIV sero-prevalent participants, suggesting minimal contribution of acute infection. Beyond acute infection, we stratified our analysis in four CD4 categories and this should implicitly, at least in part, correct for potential differences between the regions in the distribution of persons across HIV natural history stages.

Our ecological analysis did not control for viral load assay type which could potentially affect the observed regional viral load differences. Although we were not able to control for viral load assay, PCR-based assays were used in both North America and SSA. Viral load assays would have to provide consistently lower absolute viral load readings in North America and Europe, and higher readings in SSA, for this difference in viral load to be seen. Historically, the opposite trend has been observed; commercially available RNA tests are generally optimized to detect subtype B, which predominates in North America and Europe, and are frequently suboptimal in detecting HIV-1 subtypes found in other parts of the world [29]. Furthermore, despite the heterogeneity of assays used within North America and Europe, mean viral loads were similar. Another limitation of our database is the small number of viral load cohorts from Asia and South America, where only a single country is represented from each region. Finally, pregnancy data were incomplete in five cohorts.

Nevertheless, our study provides a tantalizing ‘smoking gun’, suggesting both substantially higher viral loads in SSA compared to other regions, and a potentially profound viral load effect on HIV infectiousness and epidemic trajectory. This is despite conservative assumptions in the mathematical model for the magnitude of the heightened viral load and the potential impact of co-infection-associated morbidities on sexual activity.

Several lines of evidence support our findings. The viral load differences we found (∼ 0.3 to 0.7 log10) are similar to those observed earlier (0.7 log10) in comparing a cohort of 49 Malawians to a cohort of 61 US and Swiss HIV-positive patients matched by CD4 cell count [30]. Yet, existing evidence suggests that African descent among those residing in North America and Europe is associated with lower, not higher, viral load [31,32]. Therefore, it is likely that ecological factors, such as frequently recurrent or persistent co-infections and HIV-1 sub-type, rather than host biology factors, explain these regional viral load differences. Recent data further indicate that individuals living in areas of high malaria prevalence have more than twice the odds of being HIV-positive than individuals living in areas with low malaria prevalence [odds ratio (OR) of 2.24, 95% CI 1.62–3.12 for men and 2.44, 95% CI 1.85–3.21 for women] [33].

Additional lines of evidence support our thesis of higher HIV infectiousness per coital act in SSA compared to other regions. A recent systematic review and meta-analysis found HIV transmission probability per coital act in low-income countries (predominantly SSA countries) to be six times that in high-income settings [34]. The viral load regional differences identified here of 0.58 log10 implies that the per-coital transmission probability in SSA is only 1.68 times that in North America and Europe (SuppDC Table S4.2,; smaller than that observed in this meta-analysis [34]. This suggests that other factors may also contribute to the higher HIV infectiousness in SSA. The recent landmark HPTN 52 clinical trial found four-fold higher HIV sero-conversion rate among the sero-discordant partnerships in the African sites compared to the non-African sites [8], further supporting higher HIV infectiousness in SSA.

One of the truly remarkable research advances of the past 2 years was the demonstration of the very high efficacy of ART in reducing HIV transmission [8,35]. This outcome attests to the critical role of viral load in driving HIV epidemics. Observational studies and intervention trials designed specifically to elucidate the role of co-infections in HIV infectiousness and epidemic trajectory – such as the on-going trial evaluating the impact of helminthes treatment on HIV-1 viral load in Kenya [36] – are essential. They will determine whether aggressive prevention and treatment of selected, persistent or frequently recurrent, co-infections should be included in randomized controlled trials of combination HIV prevention packages, and in HIV prevention programs and policy recommendations. As we are learning from trials evaluating HSV-2 treatment for HIV prevention [37,38], this may require development of new regimens that target the biological mechanisms underpinning these interactions.

A potential role for co-infections in the exceptionally fulminant spread of HIV in Africa is of great interest because these infections offer feasible intervention points with co-benefits that lie at the intersection of HIV prevention and other major health programs. Until we can assure immediate and sustained access to ART for all people diagnosed with HIV infection, complementary strategies will be essential. Addressing co-infections potentially may offer such a complementary strategy for the control of HIV in SSA.

Back to Top | Article Outline


The authors are grateful to the investigators who generously shared their data with us or assisted in obtaining different datasets:

Burkina Fasos: Nicolas Nagot, Philippe Mayaud and Philippe Van de Perre for the ANRS 1285 Study Group; Denmark: Nicolai Lohse and Niels Obel for the Danish HIV Cohort Study; India: Amita Gupta, Ramesh S. Paranjape, Madhuri Thakar, Nikhil Gupte and Robert Bollinger; Kenya: Jared Baeten, Julie Overbaugh, Kishorchandra Mandaliya, Grace John Stewart, Phelgona Otieno, Julie Overbaugh and Dorothy Mbori-Ngacha; Malawi: James Kublin, Padmaja Patnaik, Irving Hoffman, Bill Miller, Charles Jere, Richard Pendame, Terrie Taylor and Malcolm Molyneux; Malawi; Tanzania; Zambia: Taha Taha, Lynda Emel, Tom Fleming, Elizabeth Brown, Anthony Mwatha, Lei Wang, Moses Sinkala, George Kafulafula, Gernard Msamanga, Megan Valentine and Robert Goldenberg for HPTN 024 Team; Peru: Aldo Lucchetti, Rosario Zuñiga, Jorge Sanchez, Connie Celum, Jared Baeten, Richard Zuckerman, Wil Whittington, Jesus Peinado and Juan Guanira; South Africa: Gavin Churchyard, Katherine Fielding, Sinead Delany, Nkuli Mlaba, Godspower Akpomiemie, Tim Clayton, Wendy Stevens, Helen Rees, Philippe Mayaud, Gabriela Paz Bailey and David Lewis; Spain: Jordi Casabona for the PISCIS Study Group; Tanzania: Wafaie Fawzi, Donna Spiegelman, Ellen Hertzmark, Ferdinand Mugusi, and Gernard Msamanga, for the Tanzania Vitamin and HIV Infection Trial Team; The Netherlands: Frank de Wolf and Colette Smit for the Dutch HIV Monitoring Foundation; Uganda: Neil French; Canada: D. William Cameron and Curtis L. Cooper; USA: John T. Brooks and Rose Baker for HOPS/HIV Insight, Alvaro Munoz and Lisa Jacobson for the MACS cohort and Ronald Bosch for the ACTG trials group; various datasets: Christl Donnelly, Steve Self, Thomas Skillman, and Daniel Meade

Authors and contributors: The concept of this study was conceived by L.J.A. All authors contributed to the study design. R.V.B. and L.J.A. collected the HIV-1 plasma RNA viral load data. R.V.B. constructed and managed the database and was responsible for collating the data. H.J. conducted the statistical analyses. L.J.A. conducted the mathematical modeling analyses and wrote the first draft of the paper. All authors contributed to the analysis, discussion of the results, and writing of the manuscript.

Disclose funding received for this work: Primarily, the Qatar National Research Fund (QNRF), a Qatar Foundation funded program (NPRP 08-068-3-024). RVB acknowledges also NCRR/NIH (5 KL2 RR025015) and CFAR/NIH (P30 AI027757) funding. Additional support was provided by the Fred Hutchinson Cancer Research Center and the HIV Vaccine Trials Network. Some of the data provided to this study were supported in part by the AIDS Clinical Trials Group funded by the National Institute of Allergy and Infectious Diseases (AI 68636, AI 38858, AI 68634, and AI 38855).

Back to Top | Article Outline

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline


1. UNAIDS. UNAIDS 2011 World AIDS Day Report (available at [Accessed 3 January 2012]. 2011.
2. Pettifor AE, Levandowski BA, Macphail C, Miller WC, Tabor J, Ford C, et al. A tale of two countries: rethinking sexual risk for HIV among young people in South Africa and the United States. J Adolesc Health 2011; 49:237e231–243e231.
3. Morris M. Sexual networks and HIV. AIDS 1997; 11:S209–S216.
4. Abu-Raddad LJ, Patnaik P, Kublin JG. Dual infection with HIV and malaria fuels the spread of both diseases in sub-Saharan Africa. Science 2006; 314:1603–1606.
5. Kaul R, Cohen CR, Chege D, Yi TJ, Tharao W, McKinnon LR, et al. Biological factors that may contribute to regional and racial disparities in HIV prevalence. Am J Reprod Immunol 2011; 65:317–324.
6. Fideli US, Allen SA, Musonda R, Trask S, Hahn BH, Weiss H, et al. Virologic and immunologic determinants of heterosexual transmission of human immunodeficiency virus type 1 in Africa. AIDS Res Hum Retroviruses 2001; 17:901–910.
7. Gray RH, Wawer MJ, Brookmeyer R, Sewankambo NK, Serwadda D, Wabwire-Mangen F, 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.
8. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med 2011; 365:493–505.
9. Das M, Chu PL, Santos GM, Scheer S, Vittinghoff E, McFarland W, et al. Decreases in community viral load are accompanied by reductions in new HIV infections in San Francisco. PLoS One 2010; 5:e11068.
10. Quinn TC, Wawer MJ, Sewankambo N, Serwadda D, Li CJ, Wabwire-Mangen F, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med 2000; 342:921–929.
11. Lingappa JR, Hughes JP, Wang RS, Baeten JM, Celum C, Gray GE, et al. Estimating the impact of plasma HIV-1 RNA reductions on heterosexual HIV-1 transmission risk. PLoS One 2010; 5:e12598.
12. Nguyen RHN, Gange SJ, Serwadda D, Wabwire-Mangen F, Sewankambo KN, Wawer M, et al. Reduced odds of livebirth associated with HIV-RNA viral load: Rakai, Uganda In: XIV International AIDS Conference: Barcelona: International AIDS Society, Stockholm; 2002: 395.
13. Bruce-Chwatt LJ. A Longitudinal Survey of Natural Malaria Infection in a Group of West African Adults. West Afr Med J 1963; 12:199–217.
14. Rogier C, Ly AB, Tall A, Cisse B, Trape JF. Plasmodium falciparum clinical malaria in Dielmo, a holoendemic area in Senegal: no influence of acquired immunity on initial symptomatology and severity of malaria attacks. Am J Trop Med Hyg 1999; 60:410–420.
15. Snow RW, Craig MH, Newton CR, Steketee RW. The public health burden of Plasmodium falciparum malaria in Africa: Deriving the numbers, Working Paper No. 11, Disease Control Priorities Project, Bethesda, Maryland: Fogarty International Center, National Institutes of Health, In; 2003.
16. Kublin JG, Patnaik P, Jere CS, Miller WC, Hoffman IF, Chimbiya N, et al. Effect of Plasmodium falciparum malaria on concentration of HIV-1-RNA in the blood of adults in rural Malawi: a prospective cohort study. Lancet 2005; 365:233–240.
17. Mark KE, Wald A, Magaret AS, Selke S, Olin L, Huang ML, et al. Rapidly cleared episodes of herpes simplex virus reactivation in immunocompetent adults. J Infect Dis 2008; 198:1141–1149.
18. Abu-Raddad LJ, Magaret AS, Celum C, Wald A, Longini IM Jr, Self SG, et al. Genital herpes has played a more important role than any other sexually transmitted infection in driving HIV prevalence in Africa. PLoS ONE 2008; 3:e2230.
19. Jaffar S, Grant AD, Whitworth J, Smith PG, Whittle H. The natural history of HIV-1 and HIV-2 infections in adults in Africa: a literature review. Bull World Health Organ 2004; 82:462–469.
20. Morgan D, Mahe C, Mayanja B, Whitworth JA. Progression to symptomatic disease in people infected with HIV-1 in rural Uganda: prospective cohort study. BMJ 2002; 324:193–196.
21. Snow RW, Guerra CA, Noor AM, Myint HY, Hay SI. The global distribution of clinical episodes of Plasmodium falciparum malaria. Nature 2005; 434:214–217.
22. Casteling A, Sim J, Williamson C, Loubser S, van Brummelen R, Napier G, Martin D. Comparison of viral load technologies for samples from South African patients. In: AIDS 2000. Durban; 2000.
23. Barnabas RV, Webb EL, Weiss HA, Wasserheit JN. The role of coinfections in HIV epidemic trajectory and positive prevention: a systematic review and meta-analysis. AIDS 2011; 25:1559–1573.
24. Rizzardini G, Piconi S, Ruzzante S, Fusi ML, Lukwiya M, Declich S, et al. Immunological activation markers in the serum of African and European HIV-seropositive and seronegative individuals. AIDS 1996; 10:1535–1542.
25. Kalinkovich A, Weisman Z, Leng Q, Borkow G, Stein M, Greenberg Z, et al. Increased CCR5 expression with decreased beta chemokine secretion in Ethiopians: relevance to AIDS in Africa. J Hum Virol 1999; 2:283–289.
26. Cohen CR, Moscicki AB, Scott ME, Ma Y, Shiboski S, Bukusi E, et al. Increased levels of immune activation in the genital tract of healthy young women from sub-Saharan Africa. AIDS 2010; 24:2069–2074.
27. Cohen MS, Pilcher CD. Amplified HIV transmission and new approaches to HIV prevention. J Infect Dis 2005; 191:1391–1393.
28. Novitsky V, Ndung’u T, Wang R, Bussmann H, Chonco F, Makhema J, et al. Extended high viremics: a substantial fraction of individuals maintain high plasma viral RNA levels after acute HIV-1 subtype C infection. AIDS 2011; 25:1515–1522.
29. Emery S, Bodrug S, Richardson BA, Giachetti C, Bott MA, Panteleeff D, et al. Evaluation of performance of the Gen-Probe human immunodeficiency virus type 1 viral load assay using primary subtype A, C, and D isolates from Kenya. J Clin Microbiol 2000; 38:2688–2695.
30. Dyer JR, Kazembe P, Vernazza PL, Gilliam BL, Maida M, Zimba D, et al. High levels of human immunodeficiency virus type 1 in blood and semen of seropositive men in sub-Saharan Africa. J Infect Dis 1998; 177:1742–1746.
31. Gras L, Jurriaans S, Bakker M, van Sighem A, Bezemer D, Fraser C, et al. Viral load levels measured at set-point have risen over the last decade of the HIV epidemic in the Netherlands. PLoS One 2009; 4:e7365.
32. Muller V, von Wyl V, Yerly S, Boni J, Klimkait T, Burgisser P, et al. African descent is associated with slower CD4 cell count decline in treatment-naive patients of the Swiss HIV Cohort Study. AIDS 2009; 23:1269–1276.
33. Cuadros DF, Branscum AJ, Crowley PH. HIV-malaria co-infection: effects of malaria on the prevalence of HIV in East sub-Saharan Africa. Int J Epidemiol 2011; 40:931–939.
34. Boily MC, Baggaley RF, Wang L, Masse B, White RG, Hayes RJ, et al. Heterosexual risk of HIV-1 infection per sexual act: systematic review and meta-analysis of observational studies. Lancet Infect Dis 2009; 9:118–129.
35. Donnell D, Baeten JM, Kiarie J, Thomas KK, Stevens W, Cohen CR, et al. Heterosexual HIV-1 transmission after initiation of antiretroviral therapy: a prospective cohort analysis. Lancet 2010; 375:2092–2098.
36. Walson JL. Empiric Therapy of Helminth Co-infection to Reduce HIV-1 Disease Progression (THE or PHE), 2008. See [Accessed 10 January 2013].
37. Hayes R, Watson-Jones D, Celum C, van de Wijgert J, Wasserheit J. Treatment of sexually transmitted infections for HIV prevention: end of the road or new beginning?. AIDS 2010; 24 (Suppl 4):S15–S26.
38. Tanton C, Abu-Raddad LJ, Weiss HA. Time to refocus on HSV interventions for HIV prevention?. J Infect Dis 2011; 204:1822–1826.

co-infection; epidemic; HIV; mathematical model; sub-Saharan Africa; viral load

Supplemental Digital Content

Back to Top | Article Outline
© 2013 Lippincott Williams & Wilkins, Inc.