Epidemiology and Social
The new role of antiretrovirals in combination HIV prevention: a mathematical modelling analysis
Cremin, Idea; Alsallaq, Ramzib; Dybul, Markc,d; Piot, Petere; Garnett, Geoffreyf; Hallett, Timothy B.a
aDepartment of Infectious Disease Epidemiology, Imperial College London, UK
bCollege of Nursing Global, New York University, New York, New York
cO’Neill Institute for National and Global Health Law, Georgetown University, Washington, District of Columbia
dGeorge W. Bush Institute, Dallas, Texas, USA
eLondon School of Hygiene and Tropical Medicine, UK
fBill and Melinda Gates Foundation, Seattle, Washington, USA.
Correspondence to Ide Cremin, Department of Infectious Disease Epidemiology, Imperial College London, St. Mary's Campus, Norfolk Place, London W2 1PG, UK. Tel: +44 20 7594 3631; e-mail: email@example.com
Received 3 September, 2012
Revised 29 October, 2012
Accepted 14 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 (http://www.AIDSonline.com).
Background and objectives: Antiretroviral drugs can reduce HIV acquisition among uninfected individuals (as pre-exposure prophylaxis: PrEP) and reduce onward transmission among infected individuals (as antiretroviral treatment: ART). We estimate the potential impact and cost-effectiveness of antiretroviral-based HIV prevention strategies.
Design and methods: We developed and analysed a mathematical model of a hyperendemic setting with relatively low levels of condom use. We estimated the prevention impact and cost of various PrEP interventions, assuming a fixed amount of spending on PrEP; investigated the optimal role of PrEP and earlier ART in terms of epidemiological impact and cost; and systematically explored the impact of earlier ART and PrEP, in combination with medical male circumcision services; on HIV transmission.
Results: A PrEP intervention is unlikely to generate a large reduction in HIV incidence, unless the cost is substantially reduced. In terms of infections averted and quality adjusted life years gained, at a population-level maximal cost-effectiveness is achieved by providing ART to more infected individuals earlier rather than providing PrEP to uninfected individuals. However, early ART alone cannot reduce HIV incidence to very low levels and PrEP can be used cost-effectively in addition to earlier ART to reduce incidence further. If implemented in combination and at ambitious coverage levels, medical male circumcision, earlier ART and PrEP could produce dramatic declines in HIV incidence, but not stop transmission completely.
Conclusion: A combination prevention approach based on proven-efficacy interventions provides the best opportunity for achieving the much hoped for prevention advance and curbing the spread of HIV.
In 2011, over eight million people were receiving antiretroviral treatment (ART) in low-income and middle-income countries . Yet without substantial reductions in new infections, numbers in need of HIV treatment will continue to increase and the targets of universal access will become harder to achieve . In recent years, there has been an unprecedented pace of scientific development in HIV prevention. The HPTN052 trial demonstrated that providing early ART to infected individuals dramatically reduces heterosexual transmission to an uninfected partner . This randomized controlled trial confirmed a wealth of observational data [4–7] and contributed to intense discussions about the potential for expanded access to treatment to generate large declines in HIV incidence [8–10].
Concomitantly, there were encouraging results from trials testing the efficacy of pre-exposure prophylaxis (PrEP), whereby uninfected individuals at risk of infection take antiretrovirals, either orally or topically, to reduce the risk of becoming infected with HIV [11–14]. First, the CAPRISA 004 trial in South Africa found that a topical gel containing 1% tenofovir reduced the acquisition of infection among heterosexual women by 39% [Incidence rate ratio: 0.61; 95% confidence interval (CI) 0.40–0.94] . Following this, the iPrEx trial found that use of a daily oral pill of tenofovir and emtricitabine reduced the risk of HIV infection among MSM and transgender women by 44% (hazard ratio: 0.56; 95% CI 0.37–0.85) . Subsequently, the Partners PrEP trial, carried out in Kenya and Uganda among heterosexual serodiscordant couples, reported 67% (95% CI: 44–81) fewer infections among those receiving tenofovir alone and 75% (95% CI 55–87) fewer infections among those receiving tenofovir in combination with emtricitabine . Most recently, a randomized controlled trial among young adult heterosexual men and women in Botswana found a 62.6% (95% CI 21.5–83.4) reduction in the risk of infection among those receiving tenofovir with emtricitabine .
However, the FEM-PrEP trial (testing the effect of daily oral tenofovir-emtricitabine)  and the tenofovir only arms (oral and 1% gel) of the Vaginal and Oral Interventions to Control the Epidemic trial  have been closed due to futility. The reasons for these findings are currently being investigated but low adherence likely played a role .
Taken as a whole, these studies provide evidence for the efficacy of PrEP in different populations, for different routes of transmission (anal and vaginal sex), by different route of administration (oral and topical) and different formulations (tenofovir alone as a gel and tenofovir with emtricitabine). They also point to the importance of adherence (as suggested by efficacy correlated with drug concentrations in mucosal tissues), and the vulnerability of daily dosing regimens and, possibly, the frequency of sex acts in certain studies [18–20].
These new findings have raised important questions for HIV prevention in hyperendemic settings. Could PrEP reduce HIV incidence in these hyperendemic settings? How should relatively expensive antiretroviral drugs (used in earlier treatment or as PrEP) be distributed for optimal benefit and cost-effectiveness? More broadly, what is the best use of three interventions – medical male circumcision, early ART and PrEP – and by how much could HIV incidence be reduced by all acting in combination? We developed and analysed a mathematical model to address these questions. We focused on KwaZulu-Natal, South Africa, which has one of the largest HIV epidemics in the world.
Model structure and parameterization
A deterministic compartmental mathematical model was developed to represent heterosexual transmission of HIV in a hyperendemic setting (see Supplemental Digital Content for model definition and parameter values, http://links.lww.com/QAD/A280). The model is stratified by age, sex, male circumcision status, behavioural risk, ART use and PrEP use. Heterogeneity in sexual risk behaviour is represented as groups with different rates of partner change and condom use. Levels of condom use depend on behavioural risk group and change over time to reflect the reported increases in condom use in South Africa (see Supplemental Digital Content for details, http://links.lww.com/QAD/A280).
The natural course of HIV infection is represented as six stages of disease progression defined by their infectiousness and duration [21,22]. Infected individuals can be initiated on ART once their CD4 cell count reaches 200 μl or less (as was common in national guidelines), 350 μl or less (current WHO guidelines) or immediately following HIV testing (which is a proposed intervention). ART is assumed to increase survival [23,24] and reduce infectiousness .
Annual population size and age structure for KwaZulu-Natal were obtained from Statistics South Africa . Age-specific mortality rates were obtained from the earliest available (1990) WHO life tables for South Africa . Prevalence data from three cross-sectional population-based household surveys in KwaZulu-Natal, South Africa , and an incidence estimate from a longitudinal population-based HIV survey in the Umkhanyakude district of KwaZulu-Natal , were used to calibrate the model (see Supplemental Digital Content, http://links.lww.com/QAD/A280). Calibration of age-specific incidence curves was guided by data from a longitudinal population-based HIV survey in the Umkhanyakude district of KwaZulu-Natal .
Modelling antiretroviral-based interventions
Individuals that are on ART and retained in programs are assumed to have a risk of transmission per sex act to uninfected partners that is 96% less than the risk of transmission from those in chronic stages of infection not on ART . ART coverage parameters reflect the proportion of newly eligible individuals that initiate ART. The model is parameterized such that individuals receiving early ART do so an average of 1 year following infection. A drop-out rate of 7 per 100 PY is assumed (based on a KwaZulu-Natal Department of Health report of 7 per 100 PY rate of ART interruption) .
The efficacy of PrEP (defined in the model as the reduction in acquisition of infection per PrEP protected sex act when used that day) was assumed to be 75%. The adherence to PrEP (defined as the proportion of a PrEP user's sex acts which benefit from the effects of PrEP) was assumed to be high, at 95%. Together these parameters give an overall level of protection to PrEP users of 70% (see Supplemental Digital Content for details, http://links.lww.com/QAD/A280), within the range observed in the Partners PrEP trial and the TDF2 trial in Botswana . In terms of PrEP implementation, in all scenarios scale-up begins in 2013 (unless otherwise stated), coverage is reached in 5 years and maintained thereafter. PrEP users are assumed to spend an average of 5 years using PrEP.
A total delivery cost of US$ 600 per person per year was assumed for ART, irrespective of when treatment is initiated for the patient. This average cost is within a range of empirical cost estimates [32–35] and has been assumed in previous modelling studies . A total delivery cost of US$ 250 per person per year was assumed for PrEP . This estimate is rounded from US$ 252 and includes testing (US$ 20), human resources (US$ 91), facilities (US$ 23) and the antiretrovirals (US$ 118). All cost calculations include a 3% annual discount rate.
A baseline ‘status quo’ scenario simulated the trends in the epidemic (Figure S2, http://links.lww.com/QAD/A280) and the scale-up of ART which has occurred in KwaZulu-Natal (Table S10, http://links.lww.com/QAD/A280). The impact of additional interventions was compared to this baseline, in terms of the number of infections averted and the additional costs.
Could pre-exposure prophylaxis reduce population-level incidence and be cost saving over the long term?
To investigate what impact could be achieved with a PrEP intervention added to the current level of intervention in this setting, we examined two scenarios in which US$ 50 million, on average, is spent on PrEP annually over a 10-year period, 2013–2023. This amount is equivalent to 2% of total spending on HIV programs in South Africa in the financial year 2010/2011 , or approximately 6% of spending in KwaZulu-Natal, assuming spending is proportional to burden of infection in each province. Assuming PrEP costs US$ 250 per person-year , this amount of spending is sufficient to provide PrEP to 7.3% of uninfected 15–24 year olds, or to 4.4% of uninfected 15–54 year olds. This produces a modest decrease in incidence when PrEP is distributed only to 15–24 year olds (Fig. 1a). A very similar reduction in incidence is produced when PrEP is distributed to 15–54 year olds. Overall, 3.2 and 3.6% of infections are averted from 2013 to 2023 when 15–24 year olds and 15–54 year olds receive PrEP, respectively. Costs per infection averted, when calculated annually, decrease steadily with increasing time after the introduction of PrEP. The average cost per infection averted over the first 10 years of the intervention (from 2013 to 2023) is calculated to be US$ 10 540 and, US$ 9390 when 15–24 and 15–54 year olds receive PrEP, respectively. Total spending (on ART and PrEP) increases as PrEP is scaled-up before stabilizing at a higher level when PrEP is distributed to 15–24 year olds (Fig. 1b), and when PrEP is distributed to 15–54 year olds.
From a programs perspective, the preventive benefit of a PrEP intervention does not translate into a reduction in annual cost. However, from a cohort perspective, investment in PrEP could defray later ART costs for that cohort. Thus, to determine if a PrEP intervention could be cost saving over the long term, a cohort analysis is used (whereby a cohort of individuals aged 15–19 in 2013 are tracked for 35 years, until they had aged to be 50–54 years). Providing PrEP to 7.3% of uninfected young people (15–24 years) only, results in a slight reduction in ART costs over time for a 5-year cohort (Fig. 1c). Providing PrEP to 4.4% of uninfected 15–54 year olds produces a very similar reduction. The eventual reduced ART costs are insufficient to offset spending on PrEP over time and the cumulative intervention cost is always higher than the cumulative baseline cost when PrEP is distributed to 15–24 year olds (Fig. 1d) and when PrEP is distributed to 15–54 year olds. Assuming the same fixed financial cap of US$ 50 million annually over 10 years, with PrEP provided to 15–24 year olds only, PrEP would need to cost less than US$ 43 per person-year in order to eventually be cost saving for a cohort of individuals, in this setting.
What is the optimal role for pre-exposure prophylaxis in combination with antiretroviral treatment?
Several countries have already indicated intentions to scale-up ART for clinical need and medical male circumcision (wherein prevalence of male circumcision is low) to higher levels than at present. We therefore started by assuming that 80% of those that reached a CD4 cell count of 200/μl are initiated on ART (after a waiting period) and that coverage of male circumcision is scaled-up to 80%, from 2013 onwards. This counterfactual scenario, against which the impact of additional antiretroval-based interventions are measured, results in incidence decreasing to 1.2% by 2023, as compared to 2.4% in the ‘status quo’ scenario.
We then considered a set of potential additional ART interventions and a set of PrEP interventions (Table 1). Figure 2a charts the percentage of infections averted over 10 years and the total additional cost of each permutation of each ART and PrEP intervention. The black line (the ‘cost frontier’) connects those interventions that provide the greatest impact with respect to total additional cost. Following the ‘cost frontier’ shows how any increased spending on antiretroviral-based prevention could be optimally allocated.
On the basis of these assumptions, additional spending should first be directed to providing high coverage of treatment, first to all those with a CD4 cell count of 350/μl or less, and then to those testing positive for HIV irrespective of CD4 cell count. The maximum impact to be gained by such a strategy is 35% of infections averted over 10 years (a 61.5% reduction in incidence by 2023), at a total additional cost by 2023 of US$ 1.8 billion (an average of US$ 180 million annually, 22% of HIV spending for KwaZulu-Natal in 2010/2011, assuming spending proportional to burden of infection in each province) (Fig. 2a). In addition, this strategy produces the lowest cost per infection averted of all sixteen scenarios, at $10 530. In this model, the impact of earlier ART is limited by the substantial proportion of new HIV infections that are generated by those in early HIV infection (28% in 2013), often before treatment can be initiated.
With additional spending, incidence can be reduced further by introducing a PrEP intervention following the earlier ART intervention. Providing PrEP first to young people (aged 15–24 years) in addition to early ART averts up to 41% of infections (with 40% coverage), at a total additional cost of US$ 4.1 billion (Fig. 2a). The marginal cost per infection averted (relative to the early ART intervention) associated with this use of ART and PrEP, at US$ 39 900 is much higher than for expanding earlier ART. The highest impact – and most expensive – intervention would be to provide PrEP to 80% of the entire uninfected population (aged 15–54 years) in addition to an earlier ART intervention. This would avert 59% of infections over 10 years (and reduce HIV incidence by 88% by 2023, to 0.15%), but at a total additional cost over 10 years of US$ 9.5 billion, and a total cost per infection averted of US$ 20 500, relative to the counterfactual scenario. Thus, increasing the impact from 35% of infections averted (by means of early ART) to 59% of infections averted (by means of early ART and PrEP), would require US$ 7.7 billion to be invested in PrEP, in this setting.
In all scenarios incidence continues to decrease after 2023. The levels of impact generated are highly sensitive to the level of PrEP adherence assumed (Figure S8, http://links.lww.com/QAD/A280). Very similar results are produced when the counterfactual scenario and intervention scenarios do not assume scale-up of medical male circumcision (Figure S7, http://links.lww.com/QAD/A280). This is because intervention impact is calculated as the reduction in incidence relative to incidence in the counterfactual scenario.
The optimal cost and impact boundary delineated by various scenarios of ART and PrEP is sensitive to the cost of a person year of PrEP and of ART. For PrEP provided to young people to be the best way to spend additional resources (to reduce HIV incidence) after providing full access to treatment to those with CD4 cell count 350/μl or less and not earlier ART, the annual per person cost for PrEP would need to be less than US$ 50, an 80% reduction from US$ 250. Similarly, the extra per-person cost to initiate individuals on ART early, due to the increased efforts in HIV testing and linkage to treatment, would need to be over US$ 6500, for it to be more cost-effective to provide PrEP to young people (in conjunction with universal access to treatment for those with CD4 cell count of 350 cells/μl or less) before aiming to provide earlier treatment.
In order to also include some of the important direct therapeutic benefits of ART, the analysis was repeated for an outcome of quality adjusted life years (QALYs) gained (Fig. 2b, Table S11, http://links.lww.com/QAD/A280). The overall finding – that scaling up earlier ART should be prioritized initially – remains. However, QALYs gained per dollar spent are maximized when providing high coverage of ART at CD4 cell count 350/μl or less. This is because initiating ART early (an average of 1 year following infection) is assumed not to provide increased survival as compared to initiating ART at CD4 cell count 350/μl or less and initiating ART at higher CD4 cell counts is not associated with substantial clinical benefits (as compared to the clinical benefits when initiating ART at lower CD4 cell counts). This assumption is currently being tested by the START trial . However, observational cohort studies have suggested a clinical benefit of ART at CD4 cell counts more than 350/μl, the scale of which will determine whether providing ART earlier than 350 could maximise QALYs gained per dollar .
The potential impact of antiretroviral-based combination prevention
We took a systematic approach to examine impact with respect to different coverages of early ART, PrEP among 15–24 year olds and medical male circumcision in combination (Table 2). Across 2000 different sets of interventions, the reduction in incidence at 2023, as compared to a counterfactual of 95% ART coverage at CD4 cell count 350/μl or less from 2013 onwards, ranged from 8.8 to 70.3%, with a mean reduction of 50.4% (reducing incidence to 0.5%). The distribution of impact is left-skewed with the majority of combinations producing an impact within a high but narrow range of 45–65% (Figure S9, http://links.lww.com/QAD/A280).
A distinct pattern is observed whereby the reduction in incidence increases as coverage of each of the three interventions increases, producing a series of relatively symmetrical ‘contours’ of impact (Fig. 3). In order to achieve an incidence reduction of greater than 65%, male circumcision coverage must be at least 60–75%. With this level of coverage of male circumcision, the highest levels of impact (>65% incidence reduction) are only observed when early ART coverage is at least 64% and PrEP coverage among uninfected 15–24 year olds is at least 69%. Given a high level of male circumcision (75–90% coverage), the highest levels of impact are observed when early ART coverage is at least 40% and PrEP coverage among uninfected 15–24 year olds is at least 47%. If coverage of male circumcision and early ART do not exceed 60%, then impact generally does not exceed a 60% reduction in incidence.
Finally, the impact of a combination of these interventions on the annual number of new infections is illustrated in Fig. 4 (see Table 3 for corresponding assumptions). Scaling-up ART such that coverage for those with CD4 cell count 200/μl or less reaches 100% results in the annual number of new infections stabilizing at approximately 120 000 from 2018 onwards. Scaling-up ART such that coverage for those with CD4 cell count 350/μl or less reaches 95% results in the annual number of new infections decreasing steadily to 71 500 by 2023. Scaling-up medical male circumcision from 2015 results in the annual number of new infections decreasing to 53 100 by 2023. Introduction of early ART in 2016 substantially reduces the annual number of new infections, to 38 100 by 2023. Finally, a PrEP intervention, introduced in 2017, reduces the annual number of new infections slightly further, to 29 100 by 2023.
Starting to scale-up medical male circumcision before scaling-up ART further increases the relative impact to be gained by medical male circumcision. However, the overall impact at 2023 of this set of interventions is very similar to the scenario in which medical male circumcision is scaled-up after ART at a CD4 cell count of 350/μl or less (Figure S10, http://links.lww.com/QAD/A280). Altogether, both these combinations of interventions of proven efficacy would have reduced HIV incidence by 84% (0.4% in 2023) compared with the baseline ‘status quo’ scenario.
We estimated the potential impact of various PrEP interventions: alone; in combination with earlier ART; and in combination with earlier ART and medical male circumcision. First, assuming coverage levels that seem feasible based on current resources, a PrEP intervention in this setting would have modest impact, unless the cost of PrEP can be reduced considerably. Second, assuming current estimates for costs of PrEP and ART, scaling up earlier ART should be initially prioritized over scaling up PrEP. However, in our predictions early ART alone is insufficient to reduce HIV incidence to very low levels. PrEP could play a role in addition to earlier ART in driving greater reductions in incidence. The optimal combination of PrEP and ART is highly dependent on how inexpensive PrEP could become and how expensive providing ART to all infected individuals would be. Finally, high coverages of all three interventions are required to generate the greatest impact, reducing incidence to relatively low levels by 2023.
The idea to use antiretrovirals to curb the spread of HIV is well established [40,41]. Less certain is an understanding of where the optimal balance lies between providing antiretrovirals for infected versus uninfected individuals, if they are to be used for prevention purposes. Each future transmission in a population will involve an infected person, of whom there are many fewer than uninfected persons. The ‘optimal balance’ for the use of antiretrovirals, therefore, depends on the prevalence of HIV and patterns of partnership formation in the population; in addition to the efficacy and cost. These analyses demonstrate that, unless the cost of PrEP is reduced substantially (compared with current estimates), earlier ART for infected individuals is preferable to PrEP for uninfected individuals, in terms of cost, impact on incidence and potentially QALYs gained. For the situation of stable serodiscordant couples the optimal balance for the use of antiretrovirals is different because one partner is infected and the other is uninfected, and therefore PrEP could be more cost-effective for stable serodiscordant couples .
Three mathematical modelling studies have previously investigated the impact of PrEP in South Africa. Firstly, Pretorius et al. estimated that, if PrEP is prioritized for 15–35-year-old women, 5 to 12% of infections would be averted in South Africa over the period 2014 to 2025 if coverage among 15–35-year-old women is 30 to 60%, respectively, assuming PrEP efficacy of 90%. The corresponding costs per infection averted are over US$ 20 000. On the basis of the same efficacy and intervention assumptions, our impact estimates are higher, at 16 and 30% of infections averted with 30 and 60% coverage among 15–35 year old women, respectively, whereas our costs per infection averted are substantially lower at US$ 6000 and US$ 6300, respectively. However, our model is parameterized for KwaZulu-Natal, where incidence is higher than national levels. Second, Williams et al. predicted female use of topical PrEP to be highly cost-effective, even with low rates of gel use (i.e. in 25% of sex acts) with an estimated cost per infection averted within a range of US$ 562–4222, and this is because a very low cost per application (US$ 0.50) is assumed. Third, Walensky et al. also modelled female use of topical PrEP and estimated that PrEP could be highly cost-effective, at US$ 2700 per life year saved. Elsewhere, modelling analyses have found PrEP could provide a substantial reduction in transmission in other African settings [45,46], in India , and could be cost-effective among MSM in the United States [48,49] and in Peru .
Estimates of costs per infection averted by means of PrEP are highly sensitive to the number of individuals receiving PrEP relative to their risk of acquiring HIV, as uninfected individuals are not all at equal risk of acquiring HIV. If PrEP were allocated regardless of the level of transmission in an area and the behavioural profile of PrEP users, its cost-effectiveness and impact are likely to be low. If, however, PrEP were provided in hyperendemic areas or used among very high-risk groups (assuming they can be identified at a reasonable cost), cost-effectiveness and impact could be improved. In this regard, a more focused approach to deliver PrEP may allow a more focused, intensive approach to ensuring higher rates of adherence, and therefore effectiveness and cost-effectiveness.
There has been much interest and enthusiasm regarding combination HIV prevention [51,52], but limited quantitative analyses of what the impact could be and what the best combinations of interventions are. For a given epidemiological context, the overall effectiveness of combination prevention, in terms of impact on transmission, will depend on the efficacy and coverage of each component intervention. The efficacies of each intervention are known with different levels of certainty. The efficacies of medical male circumcision and of successful early ART are well established, whereas that of PrEP is less certain. Given that the efficacies of the interventions analysed here are known (notwithstanding uncertainty regarding PrEP), we, therefore, varied the levels of coverage of each intervention. In summary, these analyses indicate considerable reductions in incidence can be achieved by implementing several interventions at relatively low levels, but given the constraints of our assumptions regarding a maximum coverage of 90%, even with high coverages very few of the simulations reduced HIV incidence more than 70% (as compared to 95% coverage of ART at CD4 cell count of 350 cells/μl). Furthermore, this model underscores the difference between individual-level efficacy of each intervention and the population-level impact which could be achieved by scaling-up these interventions.
Our focus on medical male circumcision and antiretroviral-based prevention in these analyses does not imply that we are advocating for resources to be diverted from condom promotion or other behavioural interventions. Male condoms when used consistently are highly effective . However, barriers to their use remain, particularly within long-term stable partnerships . In South Africa, reported levels of condom use have been increasing since 1998 . Modelling analyses suggest that increased levels of condom use have had a strong contribution to the recent decline in HIV incidence in South Africa, accounting for 23–37% of the reduction in incidence depending on the model used . It is, therefore, essential that condom promotion and distribution efforts be intensified. In addition, delay in sexual debut and reduction in numbers of sexual partners have been correlated with declines in HIV prevalence and estimated incidence in some settings in sub-Saharan Africa .
Medical male circumcision is a highly cost effective, one-off procedure, conferring a lifelong reduction in risk of acquisition for uninfected men , assuming there is no risk compensation to overcome the effect. Provision of male circumcision is strongly recommended for prevention in generalized epidemic settings with low prevalence of male circumcision [58,59]. Antiretroviral-based prevention, however, is more expensive and complex, raising issues concerning HIV testing and the emergence of drug resistance. Furthermore, unlike medical male circumcision, the efficacy of early ART and PrEP strongly depend on user adherence. Strategic allocation of limited resources is essential . However, considerations apart from cost-effectiveness from a prevention standpoint will, and should, influence allocation of these resources. Modelling analyses have shown that women would benefit indirectly from scale-up of medical male circumcision, but these indirect benefits would take several years to become apparent .
The importance of using realistic assumptions when modelling HIV control strategies has recently been discussed . Here, when estimating the impact of a PrEP intervention in addition to the current situation in this setting, we used an intervention budget based on current availability of resources and then estimated what impact would be achieved, given this budget. The scenario providing early ART and PrEP to 80% of a population could rightly be criticized as being currently unobtainable, however, the aim of comparing a range of early ART and PrEP interventions was to objectively gain strategic insight regarding the relative impact and cost of allocating antiretrovirals to infected individuals to reduce onward transmission versus to uninfected individuals to reduce acquisition. Equally, the purpose of analysing three interventions at high coverages was to gain understanding of the trade-offs between them. Despite enormous progress in scaling-up treatment, the optimistic coverage assumptions, used to gain theoretical insight as to what the optimal strategies could be, are in contrast to the current reality where estimated ART coverage for South Africa was 55% in 2010 , and refusal rates among those eligible for ART are as high as 20% in some settings in South Africa .
The objectives of this analysis were to investigate the potential impact of a PrEP intervention, explore antiretroviral-based prevention and antiretroviral-based prevention in combination with medical male circumcision. Thus, a detailed representation of an ART programme was not the focus of this analysis. Development of drug resistance on ART and the differentials in mortality between first and subsequent years following ART initiation are not included. Importantly, clinical benefits (in addition to survival) of treatment above CD4 cell count 200/μl are not modelled. However, QALYs have been calculated to provide a broader measure rather than just HIV prevention. The total life years lived under different scenarios, and hence estimates of QALYs, are very sensitive to assumptions regarding survival and drop-out on early and late ART. Importantly, QALYs do not capture any of the community or societal benefits of providing ART, such as reduced levels of orphanhood or increased levels of productivity and household income. The assumption of a 96% reduction in risk of transmission among individuals on ART is optimistic; trials are ongoing to quantify the population-level effectiveness of early ART. Other limitations include those regarding PrEP. Foremost, that the emergence of drug resistance following inadvertent use of PrEP while infected is not included in this model. The assumption of 70% PrEP effectiveness is highly optimistic, yet it is within the range observed in the Partners PrEP trial  and the TDF2 trial in Botswana .
The developments in HIV prevention science over the past 2 years have raised exciting and important discussion regarding the possibility of ending the HIV epidemic [64–66]. These analyses demonstrate that large reductions in incidence can be achieved without stringently optimistic assumptions regarding a single intervention. Less idealistic assumptions (albeit still highly optimistic), as compared to previous modelling work, are sufficient to achieve such large declines in incidence, provided that resources can be allocated optimally among interventions. The best role for antiretrovirals is to contribute to the existing collection of interventions, which when applied together can reduce the spread of HIV through a focus on the local drivers of the epidemic, the groups most at risk, locations with high rates of transmission, and the most cost-effective ways to implement a combination prevention strategy.
All authors contributed to this study from its conception to finalization of the manuscript:
T.B.H., G.G., P.P., I.C. and M.D conceptualized the study, I.C., R.A. and T.B.H. developed the mathematical model; I.C. carried out the model analyses; all authors contributed to interpretation of the results and writing the manuscript.
We thank Jeff Eaton, James Truscott and Peter White for useful discussions.
Funding source: This work was funded by the Bill and Melinda Gates Foundation.
Conflicts of interest
There are no conflicts of interest.
2. Over M. Achieving an AIDS transition. Preventing infections to sustain treatment. Washington, DC: Center for Global Development; 2011.
3. 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.
4. Bunnell R, Ekwaru JP, Solberg P, Wamai N, Bikaako-Kajura W, Were W, et al. Changes in sexual behavior and risk of HIV transmission after antiretroviral therapy and prevention interventions in rural Uganda. AIDS 2006; 20:85–92.
5. Attia S, Egger M, Muller M, Zwahlen M, Low N. Sexual transmission of HIV according to viral load and antiretroviral therapy: systematic review and meta-analysis. AIDS 2009; 23:1397–1404.
6. 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.
7. Reynolds SJ, Makumbi F, Nakigozi G, Kagaayi J, Gray RH, Wawer M, et al. HIV-1 transmission among HIV-1 discordant couples before and after the introduction of antiretroviral therapy. AIDS 2011; 25:473–477.
8. Montaner JS, Hogg R, Wood E, Kerr T, Tyndall M, Levy AR, et al. The case for expanding access to highly active antiretroviral therapy to curb the growth of the HIV epidemic. Lancet 2006; 368:531–536.
9. Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2009; 373:48–57.
10. The Economist. The end of AIDS? The 30 years war. http://http://www.economist.com
/node/18772276; 2011. [Accessed 2 June 2011].
11. Abdool Karim Q, Abdool Karim SS, Frohlich JA, Grobler AC, Baxter C, Mansoor LE, et al. Effectiveness and safety of tenofovir gel, an antiretroviral microbicide, for the prevention of HIV infection in women. Science 2010; 329:1168–1174.
12. Grant RM, Lama JR, Anderson PL, McMahan V, Liu AY, Vargas L, et al.Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med 2010; 363:2587–2599.
13. Baeten JM, Donnell D, Ndase P, Mugo NR, Campbell JD, Wangisi J, et al. Antiretroviral prophylaxis for HIV prevention in heterosexual men and women. N Engl J Med 2012; 367:399–410.
14. Thigpen MC, Kebaabetswe PM, Paxton LA, Smith DK, Rose CE, Segolodi TM, et al. Antiretroviral preexposure prophylaxis for heterosexual HIV transmission in Botswana. N Engl J Med 2012; 367:423–434.
15. Van Damme L, Corneli A, Ahmed K, Agot K, Lombaard J, Kapiga S, et al. Preexposure prophylaxis for HIV infection among African women. N Engl J Med 2012; 367:411–422.
16. Microbicide Trials Network. MTN Statement on Decision to Discontinue Use of Oral Tenofovir Tablets in VOICE, a Major HIV Prevention Study in Women. http://www.mtnstopshiv.org/node/3619
[Accessed 5 December 2012].
17. Van Damme L, Corneli A, Ahmed K, Agot K, Lombaard J, Kapiga S, et al.The FEM-PrEP trial of emtricitabine/tenofovir disoproxil fumarate (Truvada) among African women
. In: 2012 Conference on Retroviruses and Opportunistic Infections
; 2012 http://www.retroconference.org/2012b/Abstracts/45406.htm
18. Karim SS, Karim QA. Antiretroviral prophylaxis: a defining moment in HIV control. Lancet 2011; 378:e23–e25.
19. Kashuba AD, Patterson KB, Dumond JB, Cohen MS. Preexposure prophylaxis for HIV prevention: how to predict success. Lancet 2012; 379:2409–2411.
20. Celum C, Baeten JM. Tenofovir-based preexposure prophylaxis for HIV prevention: evolving evidence. Curr Opin Infect Dis 2012; 25:51–57.
21. Wandel S, Egger M, Rangsin R, Nelson KE, Costello C, Lewden C, et al. Duration from seroconversion to eligibility for antiretroviral therapy and from ART eligibility to death in adult HIV-infected patients from low and middle-income countries: collaborative analysis of prospective studies. Sex Transm Infect 2008; 84 (Suppl 1):i31–i36.
22. Hollingsworth TD, Anderson RM, Fraser C. HIV-1 transmission, by stage of infection. J Infect Dis 2008; 198:687–693.
23. Mahy M, Lewden C, Brinkhof MW, Dabis F, Tassie JM, Souteyrand Y, et al. Derivation of parameters used in Spectrum for eligibility for antiretroviral therapy and survival on antiretroviral therapy. Sex Transm Infect 2010; 86 (Suppl 2):ii28–34.
24. Kitahata MM, Gange SJ, Abraham AG, Merriman B, Saag MS, Justice AC, et al. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med 2009; 360:1815–1826.
27. Shisana O, Rehle T, Simbayi LC, Zuma K, Jooste S, Pillay-van-Wyk V, et al. South African National HIV Prevalence, Incidence, Behaviour and Communication Survey, 2008: a turning tide among teenagers? Cape Town: HSRC Press; 2009.
28. Barnighausen T, Wallrauch C, Welte A, McWalter TA, Mbizana N, Viljoen J, et al. HIV incidence in rural South Africa: comparison of estimates from longitudinal surveillance and cross-sectional cBED assay testing. PLoS One 2008; 3:e3640.
29. Barnighausen T, Tanser F, Gqwede Z, Mbizana C, Herbst K, Newell ML. High HIV incidence in a community with high HIV prevalence in rural South Africa: findings from a prospective population-based study. AIDS 2008; 22:139–144.
30. The Epidemiology Unit of KwaZulu-Natal Department of Health. KwaZulu-Natal Epidemiology Bulletin. Weekly Monitoring System of the Antiretroviral Therapy for HIV/AIDS in KZN. http://www.kznhealth.gov.za/epibulletin10.pdf
. [Accessed 5 December 2012].
32. Long L, Brennan A, Fox MP, Ndibongo B, Jaffray I, Sanne I, et al. Treatment outcomes and cost-effectiveness of shifting management of stable ART patients to nurses in South Africa: an observational cohort. PLoS Med 2011; 8:e1001055.
33. Menzies NA, Berruti AA, Berzon R, Filler S, Ferris R, Ellerbrock TV, et al. The cost of providing comprehensive HIV treatment in PEPFAR-supported programs. AIDS 2011; 25:1753–1760.
34. Vella V, Govender T, Dlamini SS, Moodley I, David V, Taylor M, et al. Cost-effectiveness of staff and workload profiles in retaining patients on antiretroviral therapy in KwaZulu-Natal, South Africa. AIDS Care 2011; 23:1146–1153.
36. Pretorius C, Stover J, Bollinger L, Bacaer N, Williams B. Evaluating the cost-effectiveness of preexposure prophylaxis (PrEP) and its impact on HIV-1 transmission in South Africa. PLoS One 2010; 5:e13646.
40. Anderson RM, Gupta S, May RM. Potential of community-wide chemotherapy or immunotherapy to control the spread of HIV-1. Nature 1991; 350:356–359.
41. Garnett GP, Bartley L, Grassly NC, Anderson RM. Antiretroviral therapy to treat and prevent HIV/AIDS in resource-poor settings. Nat Med 2002; 8:651–654.
42. Hallett TB, Baeten JM, Heffron R, Barnabas R, de Bruyn G, Cremin I, et al. Optimal uses of antiretrovirals for prevention in HIV-1 serodiscordant heterosexual couples in South Africa: a modelling study. PLoS Med 2011; 8:e1001123.
43. Williams BG, Abdool Karim SS, Karim QA, Gouws E. Epidemiological impact of tenofovir gel on the HIV epidemic in South Africa. J Acquir Immune Defic Syndr 2011; 58:207–210.
44. Walensky RP, Park JE, Wood R, Freedberg KA, Scott CA, Bekker LG, et al. The cost-effectiveness of preexposure prophylaxis for HIV infection in South African women. Clin Infect Dis 2012; 54:1504–1513.
45. Abbas UL, Anderson RM, Mellors JW. Potential impact of antiretroviral chemoprophylaxis on HIV-1 transmission in resource-limited settings. PLoS One 2007; 2:e875.
46. van de Vijver DA, Derdelinckx I, Boucher CA. Circulating HIV type 1 drug resistance will have limited impact on the effectiveness of preexposure prophylaxis among young women in Zimbabwe. J Infect Dis 2009; 199:1310–1317.
47. Vissers DC, Voeten HA, Nagelkerke NJ, Habbema JD, de Vlas SJ. The impact of preexposure prophylaxis (PrEP) on HIV epidemics in Africa and India: a simulation study. PLoS One 2008; 3:e2077.
48. Desai K, Sansom SL, Ackers ML, Stewart SR, Hall HI, Hu DJ, et al. Modeling the impact of HIV chemoprophylaxis strategies among men who have sex with men in the United States: HIV infections prevented and cost-effectiveness. AIDS 2008; 22:1829–1839.
49. Paltiel AD, Freedberg KA, Scott CA, Schackman BR, Losina E, Wang B, et al. HIV preexposure prophylaxis in the United States: impact on lifetime infection risk, clinical outcomes, and cost-effectiveness. Clin Infect Dis 2009; 48:806–815.
50. Gomez GB, Borquez A, Caceres CF, Segura ER, Grant RM, Garnett GP, et al. The potential impact of pre-exposure prophylaxis for HIV prevention among men who have sex with men and transwomen in Lima, Peru: a mathematical modelling study. PLoS Med 2012; 9:e1001323.
51. Schwartlander B, Stover J, Hallett T, Atun R, Avila C, Gouws E, et al. Towards an improved investment approach for an effective response to HIV/AIDS. Lancet 2011; 377:2031–2041.
52. Kurth AE, Celum C, Baeten JM, Vermund SH, Wasserheit JN. Combination HIV prevention: significance, challenges, and opportunities. Curr HIV/AIDS Rep 2011; 8:62–72.
53. Weller S, Davis K. Condom effectiveness in reducing heterosexual HIV transmission. Cochrane Database Syst Rev 2002:CD003255.
54. Chimbiri AM. The condom is an ’intruder’ in marriage: evidence from rural Malawi. Soc Sci Med 2007; 64:1102–1115.
55. Johnson LF, Hallett TB, Rehle TM, Dorrington RE. The effect of changes in condom usage and antiretroviral treatment coverage on human immunodeficiency virus incidence in South Africa: a model-based analysis. J R Soc Interface 2012; 9:1544–1554.
56. UNAIDS. Report on the Global AIDS epidemic 2010. Geneva: UNAIDS; 2010.
57. Njeuhmeli E, Forsythe S, Reed J, Opuni M, Bollinger L, Heard N, et al. Voluntary medical male circumcision: modeling the impact and cost of expanding male circumcision for HIV prevention in eastern and southern Africa. PLoS Med 2011; 8:e1001132.
58. WHO, UNAIDS. Joint Strategic Action Framework to Accelerate the Scale-Up of Voluntary Medical Male Circumcision for HIV Prevention in Eastern and Southern Africa 2012–2016. Available at: http://whqlibdoc.who.int/unaids/2011/JC2251E_eng.pdf
. [Accessed 5 December 2012].
60. UNAIDS/WHO/SACEMA Expert Group on Modelling the Impact and Cost of Male Circumcision for HIV Prevention. Male circumcision for HIV prevention in high HIV prevalence settings: what can mathematical modelling contribute to informed decision making?PLoS Med 2009; 6:e1000109.
61. Vermund SH. Modeling interventions to assess HIV epidemic impact in Africa. J Acquir Immune Defic Syndr 2011; 58:121–124.
62. WHO, UNAIDS, UNICEF. Global HIV/AIDS Response. Epidemic update and health sector progress towards universal access. Progress Report. Geneva: WHO; 2011.
63. Katz IT, Essien T, Marinda ET, Gray GE, Bangsberg DR, Martinson NA, et al. Antiretroviral therapy refusal among newly diagnosed HIV-infected adults. AIDS 2011; 25:2177–2181.
64. Shattock RJ, Warren M, McCormack S, Hankins CA. AIDS. Turning the tide against HIV. Science 2011; 333:42–43.
65. Padian NS, McCoy SI, Karim SS, Hasen N, Kim J, Bartos M, et al. HIV prevention transformed: the new prevention research agenda. Lancet 2011; 378:269–278.
66. Cohen J. HIV prevention. Halting HIV/AIDS epidemics. Science 2011; 334:1338–1340.
This article has been cited 1 time(s).
antiretroviral therapy; combination prevention; HIV; mathematical models; pre-exposure prophylaxis
Supplemental Digital Content
© 2013 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.