In 2011, an estimated 2.5 million people became infected with HIV-1 . Alongside behavior change, male circumcision and condom use, there is an urgent need for novel HIV prevention strategies. Daily oral preexposure prophylaxis (PrEP) with tenofovir and emtricitabine can prevent 44–75% of new HIV infections [2–4]. Two studies [5,6] have found no protective effect of PrEP on prevention of new infections, but this was probably due to limited adherence.
The use of PrEP can result in the emergence and spread of drug resistance  if individuals on PrEP are infected with HIV before, or while taking PrEP. Only a single point mutation in the viral genome is required for resistance to tenofovir (K65R), and another single point mutation is required for resistance to emtricitabine (M184V) . Drug resistance can therefore quickly emerge in HIV-infected individuals who use PrEP. Indeed, resistance was shown to develop in most patients who started PrEP in the trials whilst also having an unrecognized acute infection [2–5]. However, in those individuals who became infected following assignment to PrEP, resistance developed in only a few, perhaps due to suboptimal adherence [2–4].
There is a concern that the preventive benefits of PrEP could be offset in the long term by an increase in drug resistance to commonly prescribed antiretroviral drug regimens . The WHO recommends the use of tenofovir in first-line regimens. In addition, any first-line regimen is recommended to include lamivudine or emtricitabine , which have comparable resistance profiles. The concern over resistance is highlighted by the US Food and Drug Administration (FDA) that approved PrEP under the condition that drug resistance is evaluated in viral isolates from individuals who become infected while using PrEP .
Determining the impact of PrEP on the development of drug resistance requires prospective epidemiological studies. These studies would have to be unfeasibly large, expensive and time-consuming. Mathematical modelling has therefore been used to predict whether PrEP can increase drug resistance in infected populations [11–14]. However, these mathematical models can make diverse predictions. This heterogeneity can be the result of differences in assumptions used in reconstructing HIV transmission and drug resistance, differences in the risk behavior structure, differences in the setting being modelled or simply differences in the way the question is posed and the results articulated.
We compared standardized outcomes from three independent mathematical models that determined the impact of PrEP on HIV transmission and drug resistance in areas in sub-Saharan Africa where antiretroviral therapy (ART) is available. The outcomes of the models were standardized so that differences in results could not be due to differences in the way the results were articulated.
Materials and methods
We reviewed PubMed for mathematical models that studied the impact of PrEP in the presence of ART on HIV drug resistance in sub-Saharan Africa (Keywords: PrEP, resistance, model). We also reviewed the proceedings of the main HIV conferences for similar models (the conferences considered were the Conference on Retroviruses and Opportunistic Infections, CROI; the meeting of the International AIDS Society, IAS; and the AIDS meeting). Three groups agreed to participate in the model comparison exercise. The models included are the Synthesis Transmission Model [15,16], the South African Transmission Model  and the Macha Transmission Model  (Table 1) [2,4,17,18]. The Synthesis Transmission model reports that PrEP will not increase the number of people living with a drug-resistant virus [15,16]. The other models used a different metric and found that drug resistance can increase amongst infected individuals after PrEP implementation [13,14].
The Synthesis Transmission Model is an individual-based stochastic model that simulates the HIV epidemic in sub-Saharan Africa starting in the 1980s and incorporates age (range 15–65 years), sex, condom-less sex, CD4+ cell count, specific antiretroviral drugs and resistance. For this model comparison, the model was calibrated to the HIV epidemic in South Africa. The overall adult HIV prevalence was 15.4% in 2013, when PrEP is introduced. Availability of ART starts in 2003 with initiating therapy in those with WHO stage 4 or CD4+ cell count less than 200 cells/μl. After 2010, the model assumes that ART is initiated at a CD4+ cell count less than 350 cells/μl . PrEP introduction in the model was implemented in the form of a programme targeting sero-discordant couples currently having condom-less sex.
The South African Transmission Model is a deterministic mathematical model that simulates the HIV epidemic in the adult population (15–49 year olds) of South Africa. The model assumes an HIV prevalence of 17% at the end of 2003 when roll-out of ART was started. The model assumes a treatment eligibility threshold of CD4+ cell count of less than 200 cells/μl until the end of 2009 when the threshold changes to CD4+ cell count less than 350 cells/μl. The model is stratified according to sex, sexual activity level, stage of HIV infection, drug resistance and use of ART or PrEP. In this comparison, we use the base-case scenario .
The Macha Transmission Model is a deterministic mathematical model that focuses on Macha, a rural area in southern Zambia. The model assumes an HIV prevalence of 7.7% from 2002 until 2009. Treatment is started in patients with a CD4+ cell count of less than 350 cells/μl. The model is stratified according to sexual activity level, stage of HIV infection, drug resistance due to PrEP and use of ART or PrEP . For this model comparison, the model was extended to simulate acquired resistance due to ART on the basis of previously reported data .
Assumptions of the models regarding preexposure prophylaxis
The models assume that PrEP becomes available in areas where antiretroviral drugs have already been used for HIV treatment for 8–10 years. All models assume that people receiving PrEP will be tested for HIV at intervals between 3 and 6 months (Table 1).
The models use different assumptions about the uptake of a future PrEP intervention. The Synthesis Transmission Model assumes that eventually 5% of the entire uninfected population will use PrEP. The South African Transmission Model and the Macha Transmission Model assume that 30 and 15% of the uninfected population, respectively, will receive PrEP once it is rolled out.
The models assume that PrEP has a high efficacy in preventing infection with HIV and that the effectiveness of PrEP in daily practice depends on adherence [2–4]. The Synthesis Transmission Model assumes that PrEP prevents all infections with HIV when a patient is fully adherent. When patients are partially adherent, the model assumes that percentage reduction in effectiveness of PrEP is equivalent to the percentage of PrEP doses taken . The South African Transmission Model also assumes that the effectiveness of PrEP depends on adherence, but used the efficacy of the Partners PrEP study . The South African Transmission Model and the Macha Transmission Model assume an average PrEP effectiveness of 75  and 44% , respectively.
Assumptions of the models regarding drug resistance
The outcomes of the models are standardized according to three important events that contribute to drug resistance: acquired resistance due to treatment with antiretroviral drugs, transmission of drug-resistant HIV at the time of infection and acquired drug resistance due to the use of PrEP whilst infected. In the following paragraphs, we discuss the assumptions (summarized in Table 1) made in the different models regarding these events.
The Synthesis Transmission Model assumes that drug resistance due to PrEP is characterized by the M184V and/or K65R mutations [15,16]. The other models do not specifically represent different drug-resistant mutations but assume that resistance due to PrEP results from the M184V mutation, as was previously reported [2–4].
Epidemiological studies report wide variations in the risk of acquired drug resistance during treatment [19–21]. The proportion of people on ART in whom acquired resistance developed by the end of the first year was 7% in the Synthesis Transmission Model, 16% in the South African Transmission Model [22,23] and 7% in the Macha Transmission Model . All models assume that the risk of acquired resistance gradually decreases after 1 year of ART.
The models all assume that transmission of drug-resistant HIV depends on the plasma HIV RNA viral load. The Synthesis Transmission Model assumes that the risk of transmitting wild-type or drug-resistant HIV is the same for a given plasma HIV RNA viral load. However, because individuals with drug-resistant virus are more likely to be on antiretroviral drugs, they are less likely to transmit because they have a lower plasma HIV RNA viral load . Given that a person with a virus containing the M184V mutation is the source of a new infection, the probability that this mutation is transmitted is 20%; the corresponding figure for K65R is 70% [25,26]. The South African Transmission Model assumes that drug-resistant virus acquired during treatment has a reduced fitness on average  and is therefore 25% less transmissible  than wild-type virus. Conversely, a virus with transmitted drug resistance is assumed to be equally transmissible as wild-type virus. The Macha Transmission Model assumes that resistance to PrEP involves the M184V mutation, which is associated with a 50% lower plasma HIV RNA level than a wild-type virus  and is assumed to be 50% less transmissible . In addition, all of the models assume residual virological efficacy of antiretrovirals against drug-resistant viruses , resulting in partial effectiveness of PrEP in preventing infection with a resistant virus. In particular, the Synthesis Transmission Model assumes a partial effectiveness of PrEP against a drug-resistant virus (in the presence of K65R or M184V) that is 50% lower than the effectiveness of PrEP against a wild-type virus. PrEP is assumed not effective against a virus containing both the K65R and the M184V mutations. Similarly, the South Africa Transmission Model and the Macha Transmission Model assume that drug resistance reduces the effectiveness by 75 and 50%, respectively .
The assumptions relating to the emergence of drug resistance amongst persons who continue or start using PrEP after becoming infected differ for the three models. Emergence of drug resistance in the Synthesis Transmission Model depends on number of active drugs, viral load and adherence. It is assumed that continued use of PrEP after infection results in the emergence of drug resistance in 25–44% of patients after 3 months with the risk continuing at the same rate thereafter. The South African Transmission Model assumes that drug resistance will emerge in 33% of persons on average after 1 month of inappropriate use of PrEP after becoming infected and with the same risk thereafter. The Macha Transmission Model uses a worst-case scenario, meaning that resistance will always develop after PrEP failure.
The Synthesis Transmission Model and the South African Transmission Model assume that once PrEP is stopped, the virus can revert to a majority variant that is not resistant to antiretroviral drugs while off ART. There is evidence that the M184V mutation, which is the most frequently observed in PrEP failure, can revert within weeks to a wild-type virus that is susceptible to antiretroviral treatment . The Synthesis Transmission Model assumes that M184V reverts to a wild-type virus at a rate of 80% per 3 months. This rate is comparable to the South Africa Transmission Model that assumes that reversion takes places after an average of 1.5 months. The Macha Transmission Model follows a worst-case scenario assuming that reversion will not occur. The other models assume that after reversion, drug resistance remains present in a minority of viruses  resulting in an increased likelihood of developing HIV drug resistance after start of ART.
To enable comparison between the models, each model simulated two strategies over 20 years: In the first strategy, ART was provided, and in the second strategy, both PrEP and ART were provided. The key outputs of the models for comparison were the prevalence of HIV in the general population, the prevalence of HIV drug resistance in the general population, the proportion of infected individuals with resistant infections and the source of drug-resistant infection.
Impact of antiretroviral therapy and preexposure prophylaxis on the prevalence of HIV
All models assume that access to ART will increase in the next 20 years and that ART can prevent new HIV infections . Baseline model projections suggest that if PrEP is not implemented, HIV will decrease by a modest amount without PrEP: by almost 2% in the South-African Transmission Model and by less than 1% in the two other models (Fig. 1).
All models find that availability of PrEP will result in greater decreases in HIV prevalence than a situation in which PrEP is not available. The greatest decrease is predicted in the South African Transmission Model (from 18.2% prevalence at implementation of PrEP to 13.8% after 20 years) (Fig. 1b). Smaller reductions in HIV prevalence are reported by the Synthesis Transmission Model, which reports a decrease from 15.5 to 12.7% (Fig. 1a), and the Macha Transmission Model, which finds a reduction from 8.0 to 6.8% (Fig. 1c). Results from previous models have suggested that implementing PrEP in addition to ART can result in a reduction in the prevalence of HIV compared with the use of ART alone [34,35]. The larger decrease in HIV prevalence reported by the South African transmission model therefore seems to be ascribed to the greater proportion of uninfected individuals (30%) who are assumed to receive PrEP in this model as compared with the other models (PrEP coverage of 5% in the Synthesis Transmission Model and 15% in the Macha Transmission Model) (Table 1).
Impact of antiretroviral therapy and preexposure prophylaxis on HIV drug resistance
All of the models find that HIV drug resistance in the general population will increase in the next 20 years, and to fairly similar levels regardless of whether PrEP is used or not (Fig. 2). In the Synthesis Transmission Model, the population prevalence of drug resistance (measured as the proportion of drug resistance in the total population) will increase from less than 1% to approximately 4% in the next 20 years (Fig. 2a). In the South African Transmission Model, the population prevalence will increase from about 1% to just over 2% (Fig. 2b), and in the Macha Transmission Model, the population prevalence of resistance will rise from about 1% but will remain at less than 2% (Fig. 2c). The comparatively large increase observed in the Synthesis Transmission Model might be due to the higher long-term probability of acquiring resistance during treatment assumed in this model, which reaches a prevalence of 29.8% after 20 years (Table 1). Notably, all models suggest that PrEP will only have a modest impact on the population prevalence of HIV drug resistance with an increase of at most 0.34% (or an additional 34 individuals infected with a drug-resistant virus out of 10 000 individuals) as found in the Macha Transmission Model (Fig. 2c). Similarly, the South African Transmission Model predicts an increase of 0.18%. The Synthesis Transmission Model predicts that PrEP will result in a 0.29% reduction in the population prevalence of HIV drug resistance, which is ascribed to prevention of new HIV infections due to PrEP.
In addition, the models predict that the prevalence of HIV drug resistance amongst those infected will increase over the next 20 years with and without the use of PrEP (Fig. 2d–f). The largest increase is observed in the Synthesis Transmission Model, which predicts that, without PrEP, the prevalence of infected patients carrying a drug-resistant virus will increase over the next 20 years from 4 to 29%, when PrEP is not available, or to 32% when PrEP is used (Fig. 2d). Drug resistance in the South African Transmission Model is predicted to rise from 5 to 14% when PrEP is not available, and from 5 to 17% if PrEP is used (Fig. 2e). In the Macha Transmission Model, resistance is predicted to increase from 13 to 18% if PrEP is not implemented, but to about 25% if PrEP is used (Fig. 2f). The relatively large discrepancy between the level of drug resistance in infected individuals with and without PrEP in the Macha Transmission Model is due to the assumption that all individuals who become infected whilst using PrEP will develop drug resistance. Conversely, the other models assumed that acquiring resistance due to PrEP depends on adherence and that not all individuals will acquire a resistant virus (Table 1).
Factors contributing to HIV drug resistance
Figure 3 shows the proportional contribution of the three different factors that contribute to HIV drug resistance after 20 years. Of all persons living with a drug-resistant HIV infection after 20 years, the majority (50–63% across models) is due to drug resistance arising from combination ART for treatment of HIV. Transmission of resistance is the cause of drug resistance in 40–50% of individuals across models. The cause of resistance in the remainder of persons living with a drug-resistant infection (less than 4%) can be directly attributed to the acquisition of resistance whilst infected and using PrEP.
We compared three independently developed mathematical models that predicted the impact of PrEP implementation on HIV drug resistance in sub-Saharan Africa. To facilitate the comparison, the models were reanalyzed to report common outcomes. The models represented different generalized HIV epidemic settings and different PrEP intervention strategies. Despite these differences, all models predict that the prevalence of drug resistance will increase in the coming 20 years due to increased acquired and transmitted resistance. But, importantly, PrEP is predicted to have a limited impact on future levels of drug resistance and just a small proportion (less than 4%) of resistant infections are predicted to be directly attributable to PrEP. This result was consistently found in models with high and low PrEP coverage.
The models made different assumptions regarding the acquisition, loss and transmission of drug resistance and the effectiveness of PrEP, which reflects the existing uncertainty about these processes. Nevertheless, the relative consistency between the models is reassuring and shows that PrEP is not likely to have a major impact on future levels of drug resistance.
The models included in this comparison have several limitations. First, the risk of acquired resistance due to antiretroviral treatment in sub-Saharan Africa used in the models was based on available data from the literature [17,21–23]. This risk for resistance has generally been collected in settings where laboratory monitoring techniques (estimation of plasma HIV RNA load and genotypic resistance tests) were not routinely available. Previous studies have shown that availability of such laboratory monitoring techniques is associated with a reduction in the incidence of drug resistance [21,23]. This reduction can be due to a reduced accumulation of drug resistance associated mutations, as virological failure is identified in a timely manner. In addition, patients experiencing virological failure can be advised to improve adherence, which may then reduce the risk of resistance. If laboratory techniques become widely available in the coming years, then the risk of acquired drug resistance is expected to decrease and therefore the proportion of drug resistance due to PrEP could increase. The absolute number of drug-resistant infections that can be ascribed to PrEP would, however, be expected to remain limited, and more frequent monitoring of persons on PrEP for breakthrough infection would be expected to further limit resistance.
Our standardized model comparison used a simple classification of drug resistance. As such, a distinction between particular drug resistance associated mutations or resistance to particular classes of antiretroviral drugs was not considered.
The mathematical models assume that individuals with an undiagnosed acute infection can start using PrEP. However, the models did not assume that resistance will develop faster if PrEP is used during the acute stage compared with if PrEP is used during chronic infection. Randomized clinical trials have found that resistance due to PrEP is predominantly found among patients who start PrEP with an unrecognized acute infection [2–5], suggesting that drug resistance mutations when using PrEP are potentiated by high viral replication. The resultant underestimation of the contribution of PrEP to drug resistance in the models compared here is likely to be small, as the acute stage has a brief duration of 10–16 weeks , meaning that few people will start PrEP in this phase of infection.
The purpose of this model comparison is to highlight the potential contribution of PrEP to resistance, given this has been a major issue in the FDA approval of PrEP and in public health arguments concerning PrEP [8,10,37]. Therefore, the simulated interventions were simplified to enable comparison between models, and the results should not be taken as our prediction or recommendation for how to scale-up PrEP. The comparison did not standardize and simulate the roll-out of antiretroviral drugs, adherence to antiretroviral treatment, particular antiretroviral drug treatment or the availability of viral load monitoring.
In conclusion, drug resistance will always be a risk with the use of antiretroviral drugs. Drug resistance due to ART and transmission of drug resistance will, however, far exceed drug resistance due to PrEP. Expanding access to antiretroviral drugs will require careful planning so that most infections can be averted at the lowest cost. However, with good monitoring of persons initiating and remaining on PrEP, drug resistance should not be a reason to withhold PrEP.
D.vdV., U.A., C.B., V.C., J.E., K.L., J.M. and T.H. conceived the study. V.C. and A.P. contributed data from the Synthesis Transmission Model. U.A., R.G. and J.M. contributed data from the South Africa Transmission Model. D.vdV., B.N., C.B. and K.S. contributed data from the Macha Transmission Model. DD.vdV. and B.N. analysed the combined data from all models. D.vdV. and B.N. wrote the first draft of the manuscript and U.A., C.B., V.C., J.E., R.G., K.L., J.M., A.P., K.S. and T.H. contributed to data interpretation and development of the manuscript. All authors contributed to subsequent drafts and reviewed and approved the final manuscript.
V.C. and A.P. acknowledge support from UCL Research Computing (Legion Cluster).
This work was supported by the HIV Modeling Consortium, which is supported by a grant from the Bill and Melinda Gates Foundation to Imperial College London. D.vdV. B.N. and C.B. acknowledge support from the Aids Fonds, Amsterdam, the Netherlands (grant 2010035) and the European Union (Chain, DynaNets). K.L. acknowledges support from the Wellcome Trust. J.E. and T.H. thank the Bill and Melinda Gates Foundation. U.L.A., R.G. and J.W.M. acknowledge grant support from the Bill and Melinda Gates Foundation (OPP1005974).
Conflicts of interest
There are no conflicts of interest.
This work was presented in part at the 19th International AIDS Conference (AIDS, Washington DC, July 2012; abstract FRLBX04)
1. UNAIDSGlobal report. UNAIDS report on the global AIDS epidemic 2012
. Geneva:Joint United Nations Programme on HIV/AIDS (UNAIDS); 2012.
2. 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
3. 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
4. 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
5. 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
6. Marrazzo J, Ramjee G, Nair G, Palanee T, Mkhize B, Nakabiito C, et al. Preexposure prophylaxis for HIV in women: daily oral tenofovir, oral tenofovir/emtricitabine, or vaginal tenofovir gel in the VOICE study (MTN 003)
. Program and abstracts of the 20th Conference on Retroviruses and Opportunistic Infections 2013; Atlanta: March 3–6. Abstract no. 26LB.
7. van de Vijver DA, Boucher CA. The risk of HIV drug resistance following implementation of preexposure prophylaxis
. Curr Opin Infect Dis
8. Hurt CB, Eron JJ Jr, Cohen MS. Preexposure prophylaxis and antiretroviral resistance: HIV prevention at a cost?
. Clin Infect Dis
9. WHORapid advice: antiretroviral therapy for HIV infection in adults and adolescents
. 2009; Geneva:World Health Organisation, http://www.who.int/hiv/pub/arv/rapid_advice_art.pdf
10. FDAPress announcement: FDA approves first drug for reducing the risk of sexually acquired HIV infection. Evidence-based approach enhances existing prevention strategies
. Silver Spring, MD:US Food and Drug Administration; 2012.
11. Abbas UL, Anderson RM, Mellors JW. Potential impact of antiretroviral chemoprophylaxis on HIV-1 transmission in resource-limited settings
. PLoS One
12. Abbas UL, Hood G, Wetzel AW, Mellors JW. Factors influencing the emergence and spread of HIV drug resistance arising from rollout of antiretroviral preexposure prophylaxis (PrEP)
. PLoS One
13. Abbas UL, Glaubius R, Mubayi A, Hood G, Mellors JW. Antiretroviral therapy and preexposure prophylaxis: combined impact on HIV-1 transmission and drug resistance in South Africa
. J Infect Dis
14. Nichols BE, Boucher CA, van Dijk JH, Thuma PE, Nouwen JL, Baltussen R, et al. Cost-effectiveness of pre-exposure prophylaxis (PrEP) in preventing HIV-1 infections in rural Zambia: a modeling study
. PLoS One
15. Phillips AN, Pillay D, Garnett G, Bennett D, Vitoria M, Cambiano V, et al. Effect on transmission of HIV-1 resistance of timing of implementation of viral load monitoring to determine switches from first to second-line antiretroviral regimens in resource-limited settings
16. Cambiano V, Pillay D, Lundgren J, Phillips A. Preexposure prophylaxis: impact on resistance of targeting sero-discordant couples
. Journal of the International AIDS Society. 19th International AIDS Conference Washington, DC, July 22–27 2012Abstract no. LBPE26.
17. Hamers RL, Sigaloff KC, Wensing AM, Wallis CL, Kityo C, Siwale M, et al. Patterns of HIV-1 drug resistance after first-line antiretroviral therapy (ART) failure in 6 sub-Saharan African countries: implications for second-line ART strategies
. Clin Infect Dis
18. Hollingsworth TD, Anderson RM, Fraser C. HIV-1 transmission, by stage of infection
. J Infect Dis
19. Phillips AN, Dunn D, Sabin C, Pozniak A, Matthias R, Geretti AM, et al. Long term probability of detection of HIV-1 drug resistance after starting antiretroviral therapy in routine clinical practice
20. UK-CHICLong-term probability of detecting drug-resistant HIV in treatment-naive patients initiating combination antiretroviral therapy
. Clin Infect Dis
21. Nichols BE, Boucher CA, van de Vijver DA. HIV testing and antiretroviral treatment strategies for prevention of HIV infection: impact on antiretroviral drug resistance
. J Intern Med
22. Barth RE, van der Loeff MF, Schuurman R, Hoepelman AI, Wensing AM. Virological follow-up of adult patients in antiretroviral treatment programmes in sub-Saharan Africa: a systematic review
. Lancet Infect Dis
23. Gupta RK, Hill A, Sawyer AW, Cozzi-Lepri A, von Wyl V, Yerly S, et al. Virological monitoring and resistance to first-line highly active antiretroviral therapy in adults infected with HIV-1 treated under WHO guidelines: a systematic review and meta-analysis
. Lancet Infect Dis
24. Quinn TC, Wawer MJ, Sewankambo N, Serwadda D, Li C, Wabwire-Mangen F, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group
. N Engl J Med
25. Corvasce S, Violin M, Romano L, Razzolini F, Vicenti I, Galli A, et al. Evidence of differential selection of HIV-1 variants carrying drug-resistant mutations in seroconverters
. Antivir Ther
26. Turner D, Brenner B, Routy JP, Moisi D, Rosberger Z, Roger M, et al. Diminished representation of HIV-1 variants containing select drug resistance-conferring mutations in primary HIV-1 infection
. J Acquir Immune Defic Syndr
27. Nijhuis M, Deeks S, Boucher C. Implications of antiretroviral resistance on viral fitness
. Curr Opin Infect Dis
28. Schuurman R, Nijhuis M, van Leeuwen R, Schipper P, de Jong D, Collis P, et al. Rapid changes in human immunodeficiency virus type 1 RNA load and appearance of drug-resistant virus populations in persons treated with lamivudine (3TC)
. J Infect Dis
29. 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
30. Castagna A, Danise A, Menzo S, Galli L, Gianotti N, Carini E, et al. Lamivudine monotherapy in HIV-1-infected patients harbouring a lamivudine-resistant virus: a randomized pilot study (E-184V study)
31. Jain V, Sucupira MC, Bacchetti P, Hartogensis W, Diaz RS, Kallas EG, et al. Differential persistence of transmitted HIV-1 drug resistance mutation classes
. J Infect Dis
32. Li JZ, Paredes R, Ribaudo HJ, Svarovskaia ES, Metzner KJ, Kozal MJ, et al. Low-frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: a systematic review and pooled analysis
33. 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
34. 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
35. 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
36. Pilcher CD, Joaki G, Hoffman IF, Martinson FE, Mapanje C, Stewart PW, et al. Amplified transmission of HIV-1: comparison of HIV-1 concentrations in semen and blood during acute and chronic infection
37. Michael NL. Oral preexposure prophylaxis for HIV – another arrow in the quiver?
. N Engl J Med