Advertisement
Research Article

Potential Impact of Antiretroviral Chemoprophylaxis on HIV-1 Transmission in Resource-Limited Settings

  • Ume L. Abbas mail,

    To whom correspondence should be addressed. E-mail: abbasu@dom.pitt.edu

    Affiliation: Division of Infectious Diseases, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America

    X
  • Roy M. Anderson,

    Affiliation: Department of Infectious Disease Epidemiology, Imperial College, Faculty of Medicine, University of London, London, United Kingdom

    X
  • John W. Mellors

    Affiliation: Division of Infectious Diseases, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America

    X
  • Published: September 19, 2007
  • DOI: 10.1371/journal.pone.0000875

Abstract

Background

The potential impact of pre-exposure chemoprophylaxis (PrEP) on heterosexual transmission of HIV-1 infection in resource-limited settings is uncertain.

Methodology/Principle Findings

A deterministic mathematical model was used to simulate the effects of antiretroviral PrEP on an HIV-1 epidemic in sub-Saharan Africa under different scenarios (optimistic, neutral and pessimistic) both with and without sexual disinhibition. Sensitivity analyses were used to evaluate the effect of uncertainty in input parameters on model output and included calculation of partial rank correlations and standardized rank regressions. In the scenario without sexual disinhibition after PrEP initiation, key parameters influencing infections prevented were effectiveness of PrEP (partial rank correlation coefficient (PRCC) = 0.94), PrEP discontinuation rate (PRCC = −0.94), level of coverage (PRCC = 0.92), and time to achieve target coverage (PRCC = −0.82). In the scenario with sexual disinhibition, PrEP effectiveness and the extent of sexual disinhibition had the greatest impact on prevention. An optimistic scenario of PrEP with 90% effectiveness and 75% coverage of the general population predicted a 74% decline in cumulative HIV-1 infections after 10 years, and a 28.8% decline with PrEP targeted to the highest risk groups (16% of the population). Even with a 100% increase in at-risk behavior from sexual disinhibition, a beneficial effect (23.4%–62.7% decrease in infections) was seen with 90% effective PrEP across a broad range of coverage (25%–75%). Similar disinhibition led to a rise in infections with lower effectiveness of PrEP (≤50%).

Conclusions/Significance

Mathematical modeling supports the potential public health benefit of PrEP. Approximately 2.7 to 3.2 million new HIV-1 infections could be averted in southern sub-Saharan Africa over 10 years by targeting PrEP (having 90% effectiveness) to those at highest behavioral risk and by preventing sexual disinhibition. This benefit could be lost, however, by sexual disinhibition and by high PrEP discontinuation, especially with lower PrEP effectiveness (≤50%).

Introduction

While the search is ongoing for a safe and effective HIV-1 vaccine, encouraging data from animal studies [1][5] have ignited interest in pre-exposure chemoprophylaxis (PrEP) with antiretrovirals as a strategy to prevent HIV-1 infection [6]. The potential impact of targeted or widespread PrEP on HIV-1 epidemics is uncertain and major determinants of its utility have not been defined. Several clinical trials to address the efficacy of PrEP are underway, but these will take considerable time to complete and will not specifically address the potential public health benefit [7], [8]. We therefore developed a mathematical model of a heterosexual HIV-1 epidemic and analyzed the potential for HIV-1 prevention from PrEP under different scenarios of effectiveness, duration of use, population coverage, emergence and spread of drug resistance and increased sexual risk behavior.

Methods

Model Structure

We refined our previously described deterministic mathematical model of HIV-1 disease progression and heterosexual transmission by incorporating demographic and sexual behavioral details, and by the introduction of PrEP [9]. Briefly, the model population was stratified according to gender, age, sexual activity level, disease state, PrEP status, and HIV-1 drug resistance. Model input parameters were chosen to simulate a mature epidemic in southern sub-Saharan Africa. Parameter assignments were made from recent literature on HIV-1 disease progression [10], [11], infectivity [12], and sexual behavior [13][15]. The model consists of coupled nonlinear differential equations describing the population and epidemiological stratifications outlined in Figure 1. Model parameters are shown in Tables 1 and 2, and model equations and details are provided in the Appendix S1.

thumbnail

Figure 1. Simplified Flow Diagram of Model with PrEP Implementation.

doi:10.1371/journal.pone.0000875.g001
thumbnail

Table 1. Model Parameters for the Simulated HIV-1 Epidemic

doi:10.1371/journal.pone.0000875.t001
thumbnail

Table 2. Model Parameters for Pre-exposure Chemoprophylaxis (PrEP) Implementation

doi:10.1371/journal.pone.0000875.t002

Model Output and Introduction of PrEP

The model's dynamical behavior was investigated using numerical methods. The key model outputs were: HIV-1 prevalence; HIV-1 incidence; cumulative new HIV-1 infections; and cumulative deaths from AIDS. PrEP was introduced (as once daily oral antiretroviral dosing) at endemic equilibrium when HIV-1 prevalence in sexually active adults (15–49 year-olds) was approximately 20%. The implementation of PrEP was simulated both in the absence and presence of sexual disinhibition of the individuals on PrEP, where disinhibition is defined as increased rate of sex partner change. We summed over 20 years of PrEP implementation the number of new infections and the total number of persons on PrEP, to make comparisons between the epidemics with and without PrEP at each simulation time-step. The key output variables employed in these comparisons were: the % change in the cumulative new HIV-1 infections; the ratio of HIV-1 infections prevented to person-years of PrEP; the ratio of HIV-1 infections prevented to persons enrolled in PrEP; and the ratio of the cost of PrEP to the number of infections prevented.

Effectiveness of PrEP

Our model represents the transmission of HIV-1 as a Poisson process [16][18]. The probability of transmission per heterosexual partnership, β, between an individual (on PrEP) of gender g, activity level k, and age i, with an (infected) individual of opposite gender g′, activity level l and age j is given by:
where Ψ is the number of sex acts within the partnership; γ is the probability of HIV-1 transmission per sex-act (infectivity) based on the disease stage, Ω, and drug resistance status, Θ, of the infected partner; and ξθ is the effectiveness of PrEP. Effectiveness is defined as the probability of preventing HIV-1 transmission per sex-act through PrEP and is given by the product of the average efficacy of PrEP, ξ (the degree of protection provided, from HIV-1 transmission per sex-act) and the average level of adherence to PrEP, θ (assuming once daily dosing and that doses are missed at random). In a partnership, where the infected partner harbors major drug-resistant variants (discussed below), the probability of transmission of resistant virus is υβ, while that of wild-type virus is (1−υ)β, and the effectiveness of PrEP against resistant virus is ιξθ. The parameters ξ, θ, υ and ι assume values between 0 and 1 (Table 2).

Modeling Drug Resistance

We sub-classified the HIV-1 infected individuals based on their PrEP status (naïve, on or off), type of drug resistance (primary or secondary), and simplified population dynamics of drug-resistant HIV-1 (persistence or reversion), to represent the individuals' drug resistance status (Figure 1 and Table 2). Our model assumptions for HIV-1 drug resistance are as follows. In an infected individual, the HIV-1 population is comprised of a set of related variants, termed as viral quasispecies [19]. Before the introduction of PrEP, all HIV-1 infected individuals harbor a dominant population of wild-type (drug-sensitive) virus [20]. Drug-resistant mutants are selected by drug pressure in a fraction (termed selection, π, having a value between 50% to 100%) of those individuals who become infected by wild-type virus while on PrEP (e.g. emergence of mutants with: K65R with tenofovir [21][23]; M184V with emtricitabine [24]; and M184V+K65R with tenofovir+emtricitabine [25], [26]). This type of resistance is termed secondary resistance [27]. A proportion of susceptible individuals, both on and not on PrEP, become infected by drug-resistant mutants through transmission from their sexual partner. This type of resistance is termed primary resistance [27]. Upon removal of selection pressure, either by discontinuation of PrEP [23] or transmission to a drug naïve individual [28], the drug-resistant mutants decline after a period of persistence, due to outgrowth by wild-type virus (reversion) [29][31]. Prior to reversion, the drug-resistant mutants are the dominant (major) viral variants [20], [32], whereas following reversion these become minor variants [32], [33]. Compared to individuals with wild-type virus, individuals with drug-resistant variants can have: i) decreased probability of transmission per sex act (infectivity, γΩΘ, having a relative value of 50% to 100%) due to lower level viremia from PrEP use [23], [24] or from diminished viral replicative fitness [34][36]; and ii) decreased viral transmission fitness (probability per partnership that a resistant rather than wild-type virus will be transmitted, υ, with a value between 20% to 100%) [37][39]; but the same rate of disease progression due to temporary predominance of drug-resistant mutants. Individuals with minor drug-resistant variants behave as individuals with wild-type virus in terms of infectivity and disease progression and likewise do not transmit drug resistant mutants. The re-emergence of major variants due to subsequent drug challenge (e.g. antiretroviral therapy) [40], [41] was not modeled.

Sensitivity Analyses

We performed sensitivity analyses [42] to determine the relative influence of PrEP-related model input parameters (Table 2) on the predicted decrease in new infections. For multivariate time-dependent sensitivity analyses, we rank transformed input and output data obtained using Latin hypercube sampling [43], [44] and 1000 simulation runs for epidemic scenarios with and without sexual disinhibition, and from this derived partial rank correlation coefficients (PRCCs) and standardized rank regression coefficients (SRRCs) [45], [46].

PrEP Scenarios

The impact of PrEP was determined by simulating three different scenarios: namely, optimistic, neutral and pessimistic (Table 2). For each of these scenarios, we simulated: i) the implementation of PrEP in the sexually active population in general (non-targeted strategy); ii) PrEP targeted to the two highest sexual activity levels (targeted-by-activity strategy); and iii) PrEP initiation targeted to the group 15–20 years of age (targeted-by-age strategy). We performed univariate sensitivity analysis for each scenario, in which we measured the change in infections arising from variation of each PrEP-related input parameter over its specified range (Table 2). We also analyzed the interplay between key PrEP-related input parameters and the number of infections prevented. Finally, we estimated the potential reduction in the number of new adult infections in all of southern sub-Saharan Africa, by extrapolating our results to current epidemiological data from that region [47].

Software: Model construction, simulations and sensitivity runs were implemented concurrently in Berkeley Madonna (version 8.3.8; Robert I. Macey and George F. Oster) and Vensim DSS (version 5.5d; Ventana Systems, Inc.). Data were analyzed using Microsoft Excel (version 11.8; Microsoft Corporation) and Stata SE (version 9.2; StataCorp LP).

Results

Our mathematical model stratifies the population based on gender, age, sexual activity level, disease state, PrEP status, and HIV-1 drug resistance (Figure 1), and its dynamical behavior is analyzed numerically. We introduced PrEP at endemic equilibrium and simulated optimistic, neutral and pessimistic scenarios (Table 2). For each scenario we simulated three strategies of PrEP implementation: i) in the sexually active population in general (non-targeted strategy); ii) in the two highest sexual activity levels (targeted-by-activity strategy); and iii) in the group15–20 years of age (targeted-by-age strategy). Each strategy was simulated both with and without sexual disinhibition of the individuals on PrEP. To determine the epidemiological impact of PrEP, we compared the epidemics with and without PrEP up to 20 years and determined the % change in the cumulative new HIV-1 infections; the ratio of HIV-1 infections averted to person-years of PrEP; the ratio of HIV-1 infections averted to persons enrolled in PrEP; and the ratio of the cost of person-years of PrEP to the number of infections averted.

In the simulated epidemic, adult HIV-1 prevalence was 20% at endemic equilibrium with the ratio of female to male prevalence of 1.66 [48]. Simulated trends in female prevalence are shown in Figure 2. They mimic observed patterns among urban antenatal clinic attendees in Zambia [49].

thumbnail

Figure 2. Trends in HIV-1 Prevalence among Urban Antenatal Clinic Attendees in Zambia from 1994 to 2004 [49] and the Simulated Adult Female Population.

doi:10.1371/journal.pone.0000875.g002

The results for both univariate and multivariate sensitivity analyses of the predicted impact of PrEP were similar, thus only multivariate results are presented. Table 3 shows sensitivity analyses of the predicted decline in cumulative new HIV-1 infections for 20 years after PrEP implementation. In general, coefficient (PRCC and SRCC) values near 1 indicate a strong positive influence of the input variable on prevention of infections, whereas values near -1 indicate a strong negative influence. Values near 0 indicate little, if any, influence [50]. In the scenario without sexual disinhibition occurring, the rate of PrEP discontinuation (inverse of the average duration of PrEP use) was the strongest determinant overall of reduction in infections and its effect persisted over time (PRCC ranged from −0.94 at year 5 to −0.97 at year 20). The next most important determinants were the effectiveness of PrEP (composite of efficacy and adherence) and the fraction of individuals covered (coverage) by PrEP (PRCCs of 0.92 and 0.88 at year 10, respectively), with both parameters having a positive influence. The time to achieve target coverage had a strong negative influence on infections prevented at year 5 (PRCC of −0.82), which declined over long durations of time. A weak negative influence (PRCC of −0.13) was seen for secondary drug resistance (resistance developing on PrEP), though this persisted over time.

thumbnail

Table 3. Results of Sensitivity Analyses for Model Parameters Affecting the Decrease in Cumulative New HIV-1 Infections (%)

doi:10.1371/journal.pone.0000875.t003

In the scenario in which sexual disinhibition occurred among individuals on PrEP, the effectiveness of PrEP emerged as the strongest positive determinant of infections prevented with a PRCC ranging from 0.87 to 0.91. The increase in at-risk behavior was the strongest negative determinant of infections prevented having a PRCC of −0.75 at year 10. Though PrEP discontinuation rate and coverage remained significant with disinhibition scenario, their effect was attenuated on average by about 33% and 50%, respectively, compared to the scenario without disinhibition.

SRRC values confirmed the above findings. In the absence of sexual disinhibition, the influence on infections prevented was strongest for the rate of PrEP discontinuation (SRRC range: −0.56 to −0.81) followed by effectiveness (SRRC range: 0.44 to 0.56) and coverage (SRRC range: 0.34 to 0.49). With sexual disinhibition (SRCC of −0.42), the effectiveness of PrEP became the predominant influence on the infections prevented (SRRC range: 0.71 to 0.76).

Table 4 compares the outcomes in the optimistic, neutral and pessimistic scenarios 10 years after the introduction of PrEP. These scenarios respectively assume optimistic, neutral and pessimistic sets of PrEP-related input parameters in Table 2. The potential impact of PrEP was impressive for the optimistic scenario, but was negligible for the pessimistic scenario, illustrating the importance of key chemoprophylaxis parameters on outcome. For each scenario, the greatest decline in infections was achieved with the non-targeted strategy, whereas the lowest cost of PrEP per infection averted was obtained with the targeted-by-activity strategy. Specifically, a 74% reduction in infections occurred for the optimistic scenario, 24.9% for the neutral scenario and 3.3% for the pessimistic scenario with the non-targeted strategy. These figures declined to 28.8%, 6.8% and 0.8%, respectively, with the targeted-by-activity strategy. However, the cost of person-years of PrEP per infection averted over the 10 year intervention time span fell substantially with the targeted strategy; from $6,812 to $638 for the optimistic scenario, from $9,629 to $ 1,160 for the neutral scenario, and from $20,164 to $2,949 for the pessimistic scenario. The targeted-by age strategy yielded intermediate declines in infections (45.5%, 14.5% and 2.0%), although the cost of person-years of PrEP per infection averted remained high ($5,723, $8,968 and $20,202). Overall, the numbers of infections averted per person-year of PrEP and per person enrolled in PrEP were highest for the optimistic targeted-by-activity strategy (0.33 and 1.74). Similar results were seen after 20 years of PrEP (data not shown).

thumbnail

Table 4. Outcomes for Optimistic, Neutral and Pessimistic Scenarios after Ten Years of PrEP Implementation

doi:10.1371/journal.pone.0000875.t004

Sexual disinhibition of individuals on PrEP progressively eroded the declines in infections for all scenarios (Table 4), although this effect was modest for the optimistic scenario. Specifically, at year 10 in the optimistic scenario with a 100% increase in at-risk behavior, the decline in infections was 62.7% for the non-targeted strategy and 17.7% for the targeted-by activity strategy (reduced from 74% and 28.8%, respectively, without disinhibition). The infections increased by 1.9% for the neutral scenario with targeted-by activity strategy, and increased by 7%, 2.5% and 4.4% for the three pessimistic scenario strategies (non-targeted, targeted by sexual activity and targeted by age group). Such increases were also seen with the optimistic scenario when lower levels of effectiveness were assumed. For example, at 50% effectiveness, infections increased by 8% for the non-targeted strategy as the result of a 100% increase in at-risk behavior (Figure 3). With sexual disinhibition, the decline in infections was also influenced negatively by the infectivity of individuals with secondary resistance and the probability of transmission of resistant virus from individuals with secondary resistance; however, these effects were weak with PRCCs of −0.16 and −0.08, respectively, at year 10 (Table 3). Other input parameters related to drug-resistance (Table 2) were not significantly associated with outcome.

thumbnail

Figure 3. Contour Graph for Decline in Cumulative Infections (%) as a Function of Effectiveness of PrEP and Increase in Risk Behavior Assuming Optimistic Scenario.

Negative numbers reflect increase in infections.

doi:10.1371/journal.pone.0000875.g003

Table 5 quantifies the predicted impact of PrEP for selected countries in southern Sub-Saharan Africa, as well as for the overall region, where the median country-prevalence of HIV-1 is about 20% [47]. In South Africa, up to 1.5 million new HIV-1 infections could be averted over 10 years by PrEP coverage of 75% of high sexual activity groups [51]. The corresponding estimate of the infections averted for Zambia is 0.36 million, for Botswana 0.13 million and for Lesotho 0.09 million. For southern sub-Saharan Africa as a whole up to 3.2 million new HIV-1 infections could be averted over 10 years at a cost of roughly $2.0 billion for PrEP. All these estimates assume high levels of efficacy and adherence to PrEP.

thumbnail

Table 5. Potential Impact of PrEP Introduced in 2007 on HIV-1 Infections in Southern Sub-Saharan Africa

doi:10.1371/journal.pone.0000875.t005

Discussion

Data from animal studies show that systemic antiretrovirals can prevent infection of macaques by simian immunodeficiency virus [1], [3][5]. The safety and efficacy of once daily oral antiretroviral PrEP in humans are under clinical trials in the Unites Sates, Latin America, Africa and Asia [7], [8]. However, these studies are not designed to address the population-level impact of PrEP on rates of HIV-1 transmission over many years. Using a carefully stratified and well-parameterized deterministic model of HIV-1 transmission, our analyses suggest that PrEP could have a profound impact on the HIV-1 epidemic, if effectiveness is high (high levels of efficacy and adherence) and usage persists over a decade or more; that is over much of the typical duration of sexual activity of an individual. Though the maximum effect of PrEP was observed at the highest level of coverage (75% of susceptible sexually active individuals) with good continuous adherence, such coverage and adherence are not realistic. Furthermore, the cost of untargeted PrEP per infection prevented is relatively high at $6,812. PrEP targeted to persons with greatest sexual activity produced significant declines in infections and had the lowest cost of person-years of PrEP per infection averted at $638. This result is noteworthy for two reasons: the highest activity groups comprised only 13.3% and 2.8% of the male and female model populations; and PrEP was introduced at endemic equilibrium when these high activity groups become saturated and play a lesser role in the spread of HIV-1 compared to earlier stage epidemics [52]. Models of HIV-1 vaccine implementation have also suggested that targeting by sexual activity could have a significant epidemiological impact [53]. In contrast to these vaccine models, we found that targeting by age group had less of an impact than a non-targeted approach and very similar estimates of cost of person-years of PrEP per infection averted. This is because, unlike the assumption of a one time (± booster) vaccination, PrEP requires continual use in the presence of ongoing risk of sexual transmission.

Sensitivity analyses showed that the effectiveness of PrEP was the most important determinant of the magnitude of decline in infections. This is especially the case in the scenario with sexual disinhibition of the individuals on PrEP. The high effectiveness assumed in the optimistic scenario was also the foremost reason why this scenario yielded the best outcomes overall, including cost of person-years of PrEP per infection averted. When effectiveness was lower, infections increased for all scenarios with increased risk taking behaviors of those on PrEP. The decline in infections was also very sensitive to the PrEP discontinuation rate (inverse of the average duration of PrEP use) and the level of coverage. We modeled the PrEP discontinuation rate as distinct from adherence, which was represented within our composite parameter of effectiveness. Our data suggest that continual access to PrEP would be of great importance and permanent discontinuation would undermine the epidemiological gains if PrEP use was short-lived in relation to individuals' typical duration of sexual activity.

Our representation of the evolution and transmission of drug resistance is crude. Further model development is required in this area and will be the subject of additional study. Nevertheless, the parameters directly related to drug resistance did not emerge as key determinants of the outcome of PrEP. These results may be explained by a greater contribution of other parameters impacting HIV-1 transmission, such as PrEP effectiveness.

The feasibility of PrEP as an HIV-1 prevention strategy would not only depend on its safety and efficacy, but also on its incremental cost-effectiveness compared to other intervention strategies in resource-poor settings. Our simple comparison between implementation of PrEP to a “do nothing” strategy revealed that the cost of person-years of PrEP per infection averted for the optimistic scenario with targeted-by-activity strategy of $638–$2147, compared favorably with the projected cost of $3900 per infection averted over the period 2005–2015 with the UNAIDS comprehensive prevention package [54]. This same study projected a savings of $4700 in forgone treatment and care costs. Using mathematical models, other investigators have reported that PrEP is a cost-effective strategy among high-risk men who have sex with men in New York City [55], and among populations in low-income settings [56].

Sub-Saharan Africa has about 63% of the HIV-infected population of the world totaling 22.4 million adults [47], [57]. Though epidemics of various characteristics are affecting this region [58], the vast majority of the population are not yet infected with HIV-1 and thus effective prevention strategies are urgently needed. Our analyses indicate that approximately 2.7 to 3.2 million new HIV-1 infections could be averted in southern sub-Saharan Africa over the next 10 years by targeting PrEP to population groups with the highest sexual activity concomitant with preventing increased risk behaviors in those on PrEP. It is not always easy, however, to identify those with high sexual activity patterns, except if they are involved in commercial sexual activities [14], [59].

The estimate of the number of infections averted for South Africa alone is 1.5 million infections. This high individual and public health benefit will require sustained access to PrEP. In addition, the integration of PrEP programs with voluntary counseling and testing services and other prevention programs (e.g. promotion of condom use, circumcision, and identification and treatment of sexually transmitted diseases (STDs) will be key in controlling spread of HIV-1) [60].

There are some important limitations of our current model structure and the assumptions embedded within it. The precise quantitative detail of our predictions will be affected by variations in the structure and sexual activity patterns of different populations, for which data are very limited, especially on sexual mixing patterns. However, we employed a well-established template of sexual behavior [61][65], with robust epidemiological and demographic parameterization, broadly applicable to southern sub-Saharan Africa. The actual impact of PrEP will depend on the PrEP agent or agents used as well as the physiological, behavioral and viral characteristics of the HIV-1 infected target population. Primate studies of PrEP suggest superiority of tenofovir plus emtricitabine over tenofovir alone [1], [2]. Natural polymorphisms in HIV-1 subtypes have been postulated to play an important role in drug resistance pathways [66], including the propensity of HIV-1 subtype C virus that is predominant in Sub-Saharan Africa [67], for more frequent and rapid development of the K65R tenofovir-resistance mutation, noted by some investigators [68], [69] though not by others [70]. Although there is substantial uncertainty regarding PrEP-related parameters, we employed wide ranges (within plausible bounds) for our input parameters and performed extensive sensitivity analyses. There are significant differences between the demographic and HIV/AIDS epidemiological trends estimates predicted by different agencies, largely as a result of the methods employed in analysis and prediction [15], [47], [71], [72]. In addition, except for South Africa [73], estimates of HIV-1 incidence have not been measured directly at the population level in most African countries, and reliable country-specific estimates are rarely available excepting from a few well-defined study sites with long term surveillance [74]. We elected to employ the demographic and HIV/AIDS epidemiological estimates from UNAIDS where applicable [47], [71], [75]. Our optimistic analyses assume a high level of effectiveness for PrEP, which may not be the case because of more limited drug activity and/or medication adherence. However, data on both efficacy of the potential PrEP agents [1], and adherence in Africa [76] justify some degree of optimism. In a macaque study [1] in which animals received weekly rectal simian human immunodeficiency virus challenges, 83% (5/6) of the controls became infected after 14 challenges, whereas 100% (6/6) of the macaques that received subcutaneously a combination of 22 mg tenofovir and 20 mg emtricitabine per kg once daily remained uninfected . A meta-analysis [76] of 31 North American studies (17,573 patients total) indicated a pooled estimate of 55% of the populations achieving adequate levels of adherence, whereas analysis of 27 African studies (12,116 patients total) indicated a pooled estimate of 77%. About 71% of the former and 66% of the latter studies used patient self-report to assess adherence and similar thresholds for adherence monitoring (>80% to 100%). The authors concluded that although adherence remained a concern in North America, favorable levels of adherence could be achieved in sub-Saharan Africa.

We excluded from our analyses the impact of antiretroviral therapy for infected persons [9], various other influences on transmission (e.g. STDs [77], [78], circumcision [79] and condom use [16]), as well as a formal cost-effectiveness analysis [80]. These and other refinements will be addressed in future work. Nevertheless, the key conclusion of this study is that PrEP can be a cost effective intervention given high efficacy, good adherence and long-term use, especially if sexual disinhibition is prevented.

Supporting Information

Appendix S1.

doi:10.1371/journal.pone.0000875.s001

(0.16 MB DOC)

Acknowledgments

ULA thanks Robert Eberlein, Ronald Iman, Jonathan Dushoff, Thomas Buettner, Elias Halvas and Thomas Rehle for discussions and the Ministry of Health Zambia for data [49].

Author Contributions

Contributed to the development of the mathematical model: JM RA UA. Interpretation of findings: JM RA UA. Preparation of the manuscript: JM RA UA. Collated estimates of model variables: UA. Constructed and analyzed the model HIV-1 epidemics: UA. Wrote the manuscript: UA. Saw and approved the final version of the manuscript: JM RA UA.

References

  1. 1. Garcia-Lerma J, Otten R, Qari S, Jackson E, Luo W, et al. (2006) Prevention of rectal SHIV transmission in macaques by tenofovir/FTC combination [Abstract 32LB]. 13th Conference on Retroviruses and Opportunistic Infections. Denver, CO.
  2. 2. Subbarao S, Otten RA, Ramos A, Kim C, Jackson E, et al. (2006) Chemoprophylaxis with tenofovir disoproxil fumarate provided partial protection against infection with simian human immunodeficiency virus in macaques given multiple virus challenges. J Infect Dis 194: 904–911.
  3. 3. Tsai CC, Follis KE, Sabo A, Beck TW, Grant RF, et al. (1995) Prevention of SIV infection in macaques by (R)-9-(2-phosphonylmethoxypropyl)adenine​. Science 270: 1197–1199.
  4. 4. Van Rompay KK, McChesney MB, Aguirre NL, Schmidt KA, Bischofberger N, et al. (2001) Two low doses of tenofovir protect newborn macaques against oral simian immunodeficiency virus infection. J Infect Dis 184: 429–438.
  5. 5. Van Rompay KK, Miller MD, Marthas ML, Margot NA, Dailey PJ, et al. (2000) Prophylactic and therapeutic benefits of short-term 9-[2-(R)-(phosphonomethoxy)propyl]adenin​e(PMPA) administration to newborn macaques following oral inoculation with simian immunodeficiency virus with reduced susceptibility to PMPA. J Virol 74: 1767–1774.
  6. 6. Grant RM, Buchbinder S, Cates W Jr, Clarke E, Coates T, et al. (2005) Promote HIV chemoprophylaxis research, don't prevent it. Science 309: 2170–2171.
  7. 7. AIDS Vaccine Advocacy Coalition PrEP watch. Available at: http://www.prepwatch.org/ (accessed: August 22, 2007).
  8. 8. Centers for Disease Control and Prevention CDC trials of pre-exposure prophylaxis for HIV prevention: clinical trials in Botswana, Thailand, and the United States. Available at: http://www.cdc.gov/hiv/resources/factshe​ets/prep.htm (accessed: August 22, 2007).
  9. 9. Abbas UL, Anderson RM, Mellors JW (2006) Potential impact of antiretroviral therapy on HIV-1 transmission and AIDS mortality in resource-limited settings. J Acquir Immune Defic Syndr 41: 632–641.
  10. 10. Morgan D, Mahe C, Mayanja B, Okongo JM, Lubega R, et al. (2002) HIV-1 infection in rural Africa: is there a difference in median time to AIDS and survival compared with that in industrialized countries? AIDS 16: 597–603.
  11. 11. Morgan D, Malamba SS, Orem J, Mayanja B, Okongo M, et al. (2000) Survival by AIDS defining condition in rural Uganda. Sex Transm Infect 76: 193–197.
  12. 12. Wawer MJ, Gray RH, Sewankambo NK, Serwadda D, Li X, et al. (2005) Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 191: 1403–1409.
  13. 13. Ferry B, Carael M, Buve A, Auvert B, Laourou M, et al. (2001) Comparison of key parameters of sexual behaviour in four African urban populations with different levels of HIV infection. AIDS 15: Suppl441–50.
  14. 14. Morison L, Weiss HA, Buve A, Carael M, Abega SC, et al. (2001) Commercial sex and the spread of HIV in four cities in sub-Saharan Africa. AIDS 15: Suppl461–69.
  15. 15. United States Agency for International Development & Macro International Inc. Demographic and Health Surveys. Available at: http://www.measuredhs.com/hivdata/ (accessed: August 22, 2007).
  16. 16. Foss AM, Vickerman PT, Heise L, Watts CH (2003) Shifts in condom use following microbicide introduction: should we be concerned? AIDS 17: 1227–1237.
  17. 17. May RM, Anderson RM (1988) The transmission dynamics of human immunodeficiency virus (HIV). Philos Trans R Soc Lond B Biol Sci 321: 565–607.
  18. 18. Weinstein MC, Graham J, Siegel JE, Fineberg HV (1989) Cost-effectiveness analysis of AIDS prevention programs: concepts, complications and illustrations. In: Turner C, Miller H, Moses L, editors. Confronting AIDS: Sexual behavior and intravenous drug use. Washington DC: National Academy Press. pp. 471–499.
  19. 19. Eigen M, Schuster P (1977) The hypercycle. A principle of natural self-organization. Part A: Emergence of the hypercycle. Naturwissenschaften 64: 541–565.
  20. 20. Coffin JM (1995) HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 267: 483–489.
  21. 21. (2004) Investigator's Brochure: tenofovir disoproxil fumarate, 10th edn. Foster City, CA: Gilead Sciences.
  22. 22. Van Rompay KK, Johnson JA, Blackwood EJ, Singh RP, Lipscomb J, et al. (2007) Sequential emergence and clinical implications of viral mutants with K70E and K65R mutation in reverse transcriptase during prolonged tenofovir monotherapy in rhesus macaques with chronic RT-SHIV infection. Retrovirology 4: 25.
  23. 23. Van Rompay KK, Singh RP, Heneine W, Johnson JA, Montefiori DC, et al. (2006) Structured treatment interruptions with tenofovir monotherapy for simian immunodeficiency virus-infected newborn macaques. J Virol 80: 6399–6410.
  24. 24. García-Lerma JG, Qari S, Otten R, Johnson J, Kim C, et al. (2006) Blunted viraemia and slow drug resistance emergence in rhesus macaques failing chemoprophylaxis with emtricitabine. Antivir Ther 11: Suppl 152.
  25. 25. (2005) Investigator's Brochure: emtricitabine/tenofovir disoproxil fumarate tablets, 2nd edn. Foster City, CA: Gilead Sciences.
  26. 26. Margot NA, Waters JM, Miller MD (2006) In vitro human immunodeficiency virus type 1 resistance selections with combinations of tenofovir and emtricitabine or abacavir and lamivudine. Antimicrob Agents Chemother 50: 4087–4095.
  27. 27. Clavel F, Hance AJ (2004) HIV drug resistance. N Engl J Med 350: 1023–1035.
  28. 28. Atchison R, Liegler T, Javier J, Fiss E, Hecht F, et al. (2005) Early and isolated reversion of transmitted RT K65R in a multi-drug resistant infection detected using a novel quantitative allele-specific PCR. Antivir Ther 10: Suppl 145.
  29. 29. Deeks SG, Wrin T, Liegler T, Hoh R, Hayden M, et al. (2001) Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viremia. N Engl J Med 344: 472–480.
  30. 30. Gandhi RT, Wurcel A, Rosenberg ES, Johnston MN, Hellmann N, et al. (2003) Progressive reversion of human immunodeficiency virus type 1 resistance mutations in vivo after transmission of a multiply drug-resistant virus. Clin Infect Dis 37: 1693–1698.
  31. 31. Hance AJ, Lemiale V, Izopet J, Lecossier D, Joly V, et al. (2001) Changes in human immunodeficiency virus type 1 populations after treatment interruption in patients failing antiretroviral therapy. J Virol 75: 6410–6417.
  32. 32. Charpentier C, Dwyer DE, Mammano F, Lecossier D, Clavel F, et al. (2004) Role of minority populations of human immunodeficiency virus type 1 in the evolution of viral resistance to protease inhibitors. J Virol 78: 4234–4247.
  33. 33. Palmer S, Boltz V, Maldarelli F, Kearney M, Halvas EK, et al. (2006) Selection and persistence of non-nucleoside reverse transcriptase inhibitor-resistant HIV-1 in patients starting and stopping non-nucleoside therapy. AIDS 20: 701–710.
  34. 34. Deeks SG, Hoh R, Neilands TB, Liegler T, Aweeka F, et al. (2005) Interruption of treatment with individual therapeutic drug classes in adults with multidrug-resistant HIV-1 infection. J Infect Dis 192: 1537–1544.
  35. 35. Weber J, Chakraborty B, Weberova J, Miller MD, Quinones-Mateu ME (2005) Diminished replicative fitness of primary human immunodeficiency virus type 1 isolates harboring the K65R mutation. J Clin Microbiol 43: 1395–1400.
  36. 36. White KL, Margot NA, Wrin T, Petropoulos CJ, Miller MD, et al. (2002) Molecular mechanisms of resistance to human immunodeficiency virus type 1 with reverse transcriptase mutations K65R and K65R+M184V and their effects on enzyme function and viral replication capacity. Antimicrob Agents Chemother 46: 3437–3446.
  37. 37. Leigh Brown AJ, Frost SD, Mathews WC, Dawson K, Hellmann NS, et al. (2003) Transmission fitness of drug-resistant human immunodeficiency virus and the prevalence of resistance in the antiretroviral-treated population. J Infect Dis 187: 683–686.
  38. 38. Turner D, Brenner B, Routy JP, Moisi D, Rosberger Z, et al. (2004) Diminished representation of HIV-1 variants containing select drug resistance-conferring mutations in primary HIV-1 infection. J Acquir Immune Defic Syndr 37: 1627–1631.
  39. 39. Yerly S, Jost S, Telenti A, Flepp M, Kaiser L, et al. (2004) Infrequent transmission of HIV-1 drug-resistant variants. Antivir Ther 9: 375–384.
  40. 40. Delaugerre C, Valantin MA, Mouroux M, Bonmarchand M, Carcelain G, et al. (2001) Re-occurrence of HIV-1 drug mutations after treatment re-initiation following interruption in patients with multiple treatment failure. AIDS 15: 2189–2191.
  41. 41. Izopet J, Souyris C, Hance A, Sandres-Saune K, Alvarez M, et al. (2002) Evolution of human immunodeficiency virus type 1 populations after resumption of therapy following treatment interruption and shift in resistance genotype. J Infect Dis 185: 1506–1510.
  42. 42. Saltelli A, Chan K, Scott EM, editors. (2000) Sensitivity analysis. West Sussex: John Wiley & Sons Ltd.
  43. 43. Blower SM, Dowlatabadi H (1994) Sensitivity and Uncertainty Analysis of Complex-Models of Disease Transmission - an HIV Model, as an Example. Int Stat Rev 62: 229–243.
  44. 44. McKay MD, Beckman RJ, Conover WJ (1979) A comparison of three methods for selecting values of input variables in the analysis of output from a computer code. Technometrics 21: 239–245.
  45. 45. Iman RL, Helton JC, Campbell JE (1981) An approach to sensitivity analysis of computer models: Part I-introduction, input variable selection and preliminary variable assessment. J Qual Technol 13: 174–183.
  46. 46. Iman RL, Helton JC, Campbell JE (1981) An approach to sensitivity analysis of computer models: Part II- ranking of input variables, response surface validation, distribution effect and technique synopsis. J Qual Technol 13: 232–240.
  47. 47. UNAIDS (2006) Report on the global AIDS epidemic. Available at: http://www.unaids.org/en/HIV_data/2006Gl​obalReport/default.asp (accessed: August 22, 2007).
  48. 48. Garcia-Calleja JM, Gouws E, Ghys PD (2006) National population based HIV prevalence surveys in sub-Saharan Africa: results and implications for HIV and AIDS estimates. Sex Transm Infect 82: Suppl 364–70.
  49. 49. Ministry of Health, Central Board of Health, Government of Republic of Zambia (2005) Zambia 2004 Antenatal Clinic Sentinel Surveillance Survey.
  50. 50. Iman RL, Johnson ME, Schroeder TA (2002) Assessing hurricane effects. Part 1. Sensitivity analysis. Reliab Eng Syst Safe 78: 131–145.
  51. 51. Aral SO (2000) Behavioral aspects of sexually transmitted diseases: core groups and bridge populations. Sex Transm Dis 27: 327–328.
  52. 52. Anderson RM, May RM (1991) Infectious diseases of humans: dynamics and control. Oxford: Oxford University Press.
  53. 53. Stover J, Garnett GP, Seitz S, Forsythe S (2002) The Epidemiological Impact of an HIV/AIDS Vaccine in Developing Countries. The World Bank. 2811 2811. pp. 1–30.
  54. 54. Stover J, Bertozzi S, Gutierrez JP, Walker N, Stanecki KA, et al. (2006) The global impact of scaling up HIV/AIDS prevention programs in low- and middle-income countries. Science 311: 1474–1476.
  55. 55. Desai K, McGreevey WP, Ackers ML, Hall HI, Hu DJ, et al. (2006) Modeling the potential impact of HIV chemoprophylaxis strategies among men who have sex with men in the United States: HIV infections prevented and cost-effectiveness [Abstract THAD0101]. XVI International AIDS Conference. Toronto, Canada.
  56. 56. Grant R, Lama J, Goicochea P, Levy V, Porco T (2006) Cost-effectiveness analysis of HIV chemoprophylaxis [Abstract THLB0102]. XVI International AIDS Conference. Toronto, Canada.
  57. 57. UNAIDS/WHO Epidemic Update: December 2006 Available at: http://data.unaids.org/pub/EpiReport/200​6/2006_EpiUpdate_en.pdf (accessed: August 22, 2007).
  58. 58. Asamoah-Odei E, Garcia Calleja JM, Boerma JT (2004) HIV prevalence and trends in sub-Saharan Africa: no decline and large subregional differences. Lancet 364: 35–40.
  59. 59. UNAIDS/WHO Working Group on HIV/AIDS/STI (2003) Estimating the size of populations at risk for HIV: issues and methods. Geneva: World Health Organization.
  60. 60. Wegbreit J, Bertozzi S, DeMaria LM, Padian NS (2006) Effectiveness of HIV prevention strategies in resource-poor countries: tailoring the intervention to the context. AIDS 20: 1217–1235.
  61. 61. Anderson RM, May RM, Ng TW, Rowley JT (1992) Age-dependent choice of sexual partners and the transmission dynamics of HIV in Sub-Saharan Africa. Philos Trans R Soc Lond B Biol Sci 336: 135–155.
  62. 62. Anderson RM, Ng TW, Boily MC, May RM (1989) The influence of different sexual-contact patterns between age classes on the predicted demographic impact of AIDS in developing countries. Ann N Y Acad Sci 569: 240–274.
  63. 63. Garnett GP, Anderson RM (1993) Factors controlling the spread of HIV in heterosexual communities in developing countries: patterns of mixing between different age and sexual activity classes. Philos Trans R Soc Lond B Biol Sci 342: 137–159.
  64. 64. Garnett GP, Anderson RM (1994) Balancing sexual partnerships in an age and activity stratified model of HIV transmission in heterosexual populations. IMA J Math Appl Med Biol 11: 161–192.
  65. 65. Garnett GP, Anderson RM (1995) Strategies for limiting the spread of HIV in developing countries: conclusions based on studies of the transmission dynamics of the virus. J Acquir Immune Defic Syndr Hum Retrovirol 9: 500–513.
  66. 66. Kantor R, Katzenstein DA, Efron B, Carvalho AP, Wynhoven B, et al. (2005) Impact of HIV-1 subtype and antiretroviral therapy on protease and reverse transcriptase genotype: results of a global collaboration. PLoS Med 2: e112.
  67. 67. Hemelaar J, Gouws E, Ghys PD, Osmanov S (2006) Global and regional distribution of HIV-1 genetic subtypes and recombinants in 2004. AIDS 20: W13–23.
  68. 68. Brenner BG, Oliveira M, Doualla-Bell F, Moisi DD, Ntemgwa M, et al. (2006) HIV-1 subtype C viruses rapidly develop K65R resistance to tenofovir in cell culture. AIDS 20: F9–13.
  69. 69. Doualla-Bell F, Avalos A, Brenner B, Gaolathe T, Mine M, et al. (2006) High prevalence of the K65R mutation in human immunodeficiency virus type 1 subtype C isolates from infected patients in Botswana treated with didanosine-based regimens. Antimicrob Agents Chemother 50: 4182–4185.
  70. 70. Miller MD, Margot N, McColl D, Cheng AK (2007) K65R development among subtype C HIV-1-infected patients in tenofovir DF clinical trials. AIDS 21: 265–266.
  71. 71. Department of Economic and Social Affairs, Population Division (2005) World population prospects: the 2004 revision. New York: United Nations.
  72. 72. U.S. Census Bureau International Data Base Available at: http://www.census.gov/ipc/www/idb/ (accessed: August 22, 2007).
  73. 73. Shisana O, Rehle T, Simbayi L, Parker W, Zuma K, et al. (2005) South African national HIV prevalence, HIV incidence, behavior and communication survey, 2005. Capetown: HSRC Press.
  74. 74. Shelton JD, Halperin DT, Wilson D (2006) Has global HIV incidence peaked? Lancet 367: 1120–1122.
  75. 75. UNAIDS/WHO StatementApr 12. 2006 Estimating the status of the AIDS epidemic in countries.
  76. 76. Mills EJ, Nachega JB, Buchan I, Orbinski J, Attaran A, et al. (2006) Adherence to antiretroviral therapy in sub-Saharan Africa and North America: a meta-analysis. JAMA 296: 679–690.
  77. 77. Blower S, Ma L (2004) Calculating the contribution of herpes simplex virus type 2 epidemics to increasing HIV incidence: treatment implications. Clin Infect Dis 39: Suppl 5240–247.
  78. 78. Grassly NC, Lowndes CM, Rhodes T, Judd A, Renton A, et al. (2003) Modelling emerging HIV epidemics: the role of injecting drug use and sexual transmission in the Russian Federation, China and India. Int J Drug Policy 14: 25–43.
  79. 79. Williams BG, Lloyd-Smith JO, Gouws E, Hankins C, Getz WM, et al. (2006) The potential impact of male circumcision on HIV in sub-Saharan Africa. PLoS Med 3: e262.
  80. 80. Drummond MF, Sculpher MJ, Torrance GW, O'Brien BJ, Stoddart GL (2005) Methods for the economic evaluation of health care programmes. New York: Oxford University Press.
  81. 81. Deschamps M-M, Pape JW, Hafner A, Johnson WDJ (1996) Heterosexual transmission of HIV in Haiti. Ann Intern Med 125: 324–330.
  82. 82. Moatti JP, Prudhomme J, Traore DC, Juillet-Amari A, Akribi HA, et al. (2003) Access to antiretroviral treatment and sexual behaviours of HIV-infected patients aware of their serostatus in Cote d'Ivoire. AIDS 17: Suppl 369–77.
  83. 83. Cong ME, Heneine W, Garcia-Lerma JG (2007) The fitness cost of mutations associated with human immunodeficiency virus type 1 drug resistance is modulated by mutational interactions. J Virol 81: 3037–3041.
  84. 84. Brenner BG, Routy JP, Petrella M, Moisi D, Oliveira M, et al. (2002) Persistence and fitness of multidrug-resistant human immunodeficiency virus type 1 acquired in primary infection. J Virol 76: 1753–1761.
  85. 85. Lifson JD, Rossio JL, Arnaout R, Li L, Parks TL, et al. (2000) Containment of simian immunodeficiency virus infection: cellular immune responses and protection from rechallenge following transient postinoculation antiretroviral treatment. J Virol 74: 2584–2593.
  86. 86. Rosenwirth B, ten Haaft P, Bogers WM, Nieuwenhuis IG, Niphuis H, et al. (2000) Antiretroviral therapy during primary immunodeficiency virus infection can induce persistent suppression of virus load and protection from heterologous challenge in rhesus macaques. J Virol 74: 1704–1711.
  87. 87. Gentleman A, Kumar H (2006) AIDS drug provokes patent battle in India. The International Herald Tribune.
  88. 88. Gilead Sciences Inc. (2005) Gilead reduces prices for Viread® and Truvada® in the developing world. Available at: http://www.gilead.com/wt/sec/pr_750303 (accessed: August 22, 2007).
  89. 89. Surveillance and Survey and Laboratory Working Groups (2006) Interim recommendations for the use of the BED capture enzyme immunoassay for incidence estimation and surveillance. Atlanta: Centers for Disease Control and Prevention.