Effect of PrEP for MSM in UK: considerations for scale-up
Download the PDF here
Download the PDF here
Effect of pre-exposure prophylaxis and combination HIV prevention for men who have sex with men in the UK: a mathematical modeling study [Jan 2016]....see below following commentary
"Our analysis confirmed the importance of implementation of a combination of interventions for effective HIV control in MSM.13, 17 The provision of PrEP as part of a combination strategy, even to a quarter of highly sexually active individuals, could prevent more than 7000 new HIV infections in the UK before the end of this decade. The relatively small coverage the programme requires is feasible because around half of MSM in the UK have shown interest in participating in PrEP18 and treatment-as-prevention programmes.19 Combined programmatic launch would be aided by the fact that the most at-risk men are more likely to use PrEP than are other high-risk MSM.......In a modelling study in UK MSM, Phillips and colleagues27 reported that immediate ART would prevent 32% of new infections during 2006-10 compared with our 100% coverage estimates of 24% during 2014-20......Both our and Phillips and colleagues'27 studies suggested the same 62% incidence reduction at the maximum coverage if more frequent testing and early ART were implemented simultaneously......In conclusion, our analysis suggests that a combination of PrEP, expansion of HIV testing, and test-and-treat programmes implemented in small groups of MSM in the UK could prevent thousands from HIV within a few years of implementation, particularly if risk-compensation behaviour is successfully avoided. Integration of this enhanced version of conventional interventions with novel biomedical technologies is likely to deliver the optimum approach to HIV prevention that could determine the future course of the HIV epidemic in the UK."
HIV transmission in MSM: considerations for PrEP scale-up - Comment
The Lancet HIV Published Online: 13 January 2016
Emily A Arnold, Wayne T Steward
The modelling exercises described in Narat Punyacharoensin and colleagues' study1 in The Lancet HIV show the value of incorporation of pre-exposure prophylaxis (PrEP) into combination HIV prevention in men who have sex with men (MSM).2 Through a well conceived set of simulations, the investigators examined the number of infections that could be averted in MSM in the UK by use of PrEP, HIV testing, treatment of HIV-infected individuals, and sexual risk reduction.3, 4 Punyacharoensin and colleagues estimated that the greatest number of infections would be prevented in a practical combination prevention programme that included PrEP alongside annual HIV testing for HIV-negative men and immediate treatment for HIV-positive men. In addition to promising findings from the PROUD study,5 in which PrEP reduced the risk of HIV infection in MSM by 86%, the evidence is mounting for national policies in the UK and elsewhere that incorporate PrEP as a viable and accessible prevention strategy for MSM.
Punyacharoensin and colleagues' models also emphasise the importance of effective scale-up. The investigators estimated that the numbers of infections averted would be more heavily affected by variations in the proportion of MSM receiving PrEP than by variations in its efficacy.1 This finding warrants consideration because PrEP scale-up requires a unique combination of expertise and engagement. Traditionally, HIV intervention efforts have differed by recipient serostatus and location of delivery. For HIV-negative individuals, the main goals are prevention and screening, which involve discrete services, such as HIV testing, that are often delivered in community-based organisations or sexual health clinics. These services are segregated from other parts of the health-care system, partly to minimise the effect of stigma.6 Indeed, young and middle-aged MSM might avoid primary care facilities for their sexual health needs because they fear discrimination.7 For HIV-positive individuals, the main goal is treatment, which requires ongoing interaction with medical facilities. Delivery of PrEP does not fit neatly into this existing health-care delivery system: it is an intervention for HIV-negative individuals that requires medical expertise and ongoing interaction with the health-care system.
To enable the scale-up of PrEP to effective coverage levels, the public health system must strengthen ties among community-based organisations, sexual health clinics, and medical providers. These efforts could parallel those for linking of newly diagnosed individuals from testing to care sites,8 but they could also take different forms, such as medical providers overseeing PrEP programmes at community-based organisations. The diversified expertise arising from collaboration among these different partners would better position the health-care system to address the barriers to uptake and use of PrEP. To maximise success, PrEP implementation needs to include the ability to engage with MSM communities and reach high-risk individuals, and should provide a point of access that is convenient and does not elicit fears of stigma, adherence counselling and monitoring for drug side-effects to ensure that individuals use PrEP effectively, and support for the uptake of complementary behavioural prevention strategies (eg, reduced number of partners and condom use). Additionally, medical providers need guidance about how to manage patients receiving PrEP, and how drug and ancillary care costs should be covered.9, 10 Finally, gay communities should be mobilised and educated about PrEP use and its use not only in prevention of HIV transmission but also in promotion of sexual health and wellbeing.11, 12
Collaboration among community-based organisations, sexual health clinics, and medical providers would aid the rollout of these complementary components of a PrEP programme and promote cross-training and capacity building. Community-based organisations could assist with community outreach and education, while medical providers could increase capacity and cultural competency in provision of PrEP to MSM. Medical providers might benefit from improved skills during discussions of sexual health, which would help to ensure that at-risk individuals can comfortably ask primary care providers about PrEP, and community-based organisations might benefit from enhanced in-house expertise about biomedical forms of intervention, which would better position them to respond to new prevention approaches and improve upon services to complement medical care for HIV-infected individuals.
Punyacharoensin and colleagues' models emphasise the potential effect that PrEP could have on the HIV epidemic. Work should now be done at the health systems and policy levels to help realise that effect.
Effect of pre-exposure prophylaxis and combination HIV prevention for men who have sex with men in the UK: a mathematical modeling study
Lancet HIV Jan 13 2016 Narat Punyacharoensin, William John Edmunds, Daniela De Angelis, Valerie Delpech, Graham Hart, Jonathan Elford, Alison Brown, 0 Noel Gill,
Richard Guy White
HIV transmission in men who have sex with men (MSM) in the UK has shown no sign of decreasing in the past decade. Additional prevention measures are needed. We aimed to estimate the effect of various potential interventions implemented individually and in combination on prevention of HIV infection.
We extended a deterministic partnership-based mathematical model for HIV transmission, informed by detailed behavioural and surveillance data, to assess the effect of seven different HIV interventions implemented in MSM (aged 15-64 years) in the UK during 2014-20, including increasing rates of HIV testing, test-and-treat programmes, pre-exposure prophylaxis (PrEP), and sexual behavioural changes. We did sensitivity analyses on risk compensation.
We predicted a baseline of 16 955 new infections (IQR 13 156-21 669) in MSM in the UK during 2014-20. At a coverage of ≤50%, testing twice a year outperformed all other interventions. Of all intervention combinations, only the combined effect of test and treat and annual HIV testing (61·8%, IQR 47·2-81·8, of total incidence) was greater than the sum of effects of the two interventions individually (32·6%, 23·7-46·0, and 23·9%, 16·5-33·3, respectively). Simultaneous PrEP, expansion of HIV testing, and initiation of test-and-treat programme in 25% of high-activity MSM could save 7399 (IQR 5587-9813) UK MSM from HIV infection (43·6%, IQR 32·9-57·9, of total incidence). An increase in unsafe sex or sexual partners to 50% or more could substantially reduce the effect of interventions, but is unlikely to negate the prevention benefit completely.
PrEP could prevent a large number of new HIV infections if other key strategies including HIV testing and treatment are simultaneously expanded and improved. Without PrEP, HIV incidence in MSM in the UK is unlikely to decrease substantially by the end of this decade.
In the UK, an estimated 103 700 people were living with HIV in 2014, around 43% of whom were men who have sex with men (MSM), which was equivalent to 4·9% prevalence in MSM aged 15-44 years.1 On the basis of estimates from a CD4 back-calculation model,2 about 2600 MSM in England and Wales have been infected with HIV every year in the past 10 years, with a slight increase to 2800 in 2014, and without alternative prevention measures incidence is expected to remain at this level throughout the decade.3
Attempts to prevent HIV transmission in the UK have focused on promotion of proper and consistent condom use and increasing coverage and frequency of HIV testing.1 But because these interventions alone have been insufficient to reduce HIV incidence in UK MSM over time, more attention has been paid to alternative strategies of HIV infection control, including use of pre-exposure prophylaxis (PrEP). Two PrEP trials—the PROUD study in England4 and the IPERGAY study in France and Canada5—had to discontinue the placebo group and provide PrEP to all eligible participants after their interim results showed the high protective effect of PrEP.
Because the combination of conventional and alternative prevention measures is now being considered, an increased understanding is needed of their potential effect on HIV transmission in MSM in the UK. We did this mathematical modelling study to assess the potential effect of various HIV prevention interventions, both individually and in combination, on HIV transmission during 2014-20. We also investigated the sensitivity of our results to variation in intervention coverage, PrEP effectiveness, and the effects of potential risk compensation.
Our analysis confirmed the importance of implementation of a combination of interventions for effective HIV control in MSM.13, 17 The provision of PrEP as part of a combination strategy, even to a quarter of highly sexually active individuals, could prevent more than 7000 new HIV infections in the UK before the end of this decade. The relatively small coverage the programme requires is feasible because around half of MSM in the UK have shown interest in participating in PrEP18 and treatment-as-prevention programmes.19 Combined programmatic launch would be aided by the fact that the most at-risk men are more likely to use PrEP than are other high-risk MSM.20
One of the main concerns about PrEP is that risk compensation could reduce the effectiveness of the intervention.21 Our risk-compensation analyses suggested that a substantial increase in unprotected anal intercourse would be needed to negate the population-level benefits of PrEP. However, in reality, the negative effect of risk compensation could be multiplied, for example, in PrEP users with poor drug adherence in whom there is an increase in unsafe sex and number of partners simultaneously. Thus, approximation of increased risk-taking in real-world situations, with the behavioural insights gathered from clinical trials, is highly important but might require more appropriate study designs other than a conventional longitudinal analysis.22 With sufficient knowledge and understanding, the complementary risk-reduction programmes could be tailor-made to help adjust perceptions of risks and benefits underlying the mechanism of risk compensation in PrEP users.
Another issue regarding PrEP is its protective effectiveness against HIV. In the present study we assumed a baseline effectiveness of 44% for the daily tenofovir and emtricitabine, which was suggested by a clinical trial in MSM in six non-European countries.16 The HIV protective effects of PrEP relied greatly on drug adherence of programme participants,14, 16 which could differ greatly between the trial and actual implementation in the UK. Findings from the PROUD open-label trial suggested a relative incidence reduction of 86% in participants who received PrEP immediately compared with those who received PrEP after a 12 month deferral period.4 Moreover, the PrEP programme is not a stand-alone intervention, but rather a combination of several components including medicine use, HIV testing, motivational interviewing, and risk reduction. The effectiveness of PrEP will thus rely greatly on the quality of these complementary programmes and might exceed our estimates if all are successfully adopted.
By comparison with PrEP, our findings showed that the test-and-treat programme has a smaller effect on HIV prevention at the population level, which is consistent with modelling studies in other settings.23, 24 This finding might be explained by the already small contribution of MSM with diagnosed HIV to total infections,3 because after diagnosis there are reductions in high-risk behaviours, and many of these men in whom infection is diagnosed begin ART promptly, especially as the CD4 count threshold for treatment eligibility has changed in the UK.25 Moreover, now that the individual benefit of immediate ART has been established for people with HIV infection, the high proportion of MSM with HIV infection beginning immediate or very early ART is likely to increase still further.
The key strengths of test and treat lie in various aspects. First, the robustness of the programme in relation to risk-compensation behaviour. This robustness could be explained by the benefits of ART in reducing the probability of HIV transmission between serodiscordant couples,26 which exceed the adverse effects of plausible risk compensation.
Second, the capability of the programme to be effective at low coverage makes it suitable for use in various practical situations in which full coverage is almost impossible to reach. Third, because early ART is only offered to diagnosed MSM, the number of participants needed to reach the targeted incidence reduction is much smaller than that needed by other programmes focusing on a pool of susceptibles and undiagnosed men.
Finally, test and treat was the only intervention investigated here that, when implemented along with increased HIV testing frequency and coverage, provided additional benefits over the sum of the two independent effects. This result is because of the completely non-overlapping target groups (undiagnosed vs diagnosed MSM) between the two interventions and the increased number of diagnosed individuals to be treated early with ART.
In a modelling study in UK MSM, Phillips and colleagues27 reported that immediate ART would prevent 32% of new infections during 2006-10 compared with our 100% coverage estimates of 24% during 2014-20. The slight difference in the estimates is probably due to different timescales and the underlying assumptions of the interventions. We assumed that men diagnosed at a CD4 cell count of 350 cells per μL or more, except in those with primary HIV infection, had the same ART initiation rate as those diagnosed at counts of less than 350 cells per μL, whereas in Phillips and colleagues' study,27 ART was initiated immediately in all MSM diagnosed with HIV after 2000. That study also reported a 25% decline in incidence if HIV test rate has been increasing since 2000 until 2010, when 68% of all MSM were tested each year,27 whereas in our study testing once a year with 50-75% coverage would lower the incidence by 25-30%. Both our and Phillips and colleagues'27 studies suggested the same 62% incidence reduction at the maximum coverage if more frequent testing and early ART were implemented simultaneously.
The three ambitious 90-90-90 targets proposed by UNAIDS are intended to reduce HIV incidence by 90% by 2030, compared with that in 2010.28 The first target called for 90% of people living with HIV in 2020 to be aware of their infection.28 An estimated 84% of MSM living with HIV in the UK in 2013 were aware of their infection, so the 90% target should be achievable in the near future. The other two targets that expected 90% of diagnosed individuals to be treated with ART, and 90% of those receiving ART to achieve an undetectable viral load,28 have already been reached in the UK.1 Despite these achievements, the reduction in incidence by 2030 is unlikely without additional interventions. Our baseline scenario suggested a constant level of 2000-3000 new infections in UK MSM every year up to 2020, with no sign of a decrease. Through using linear regression on our 2014-20 estimates to extrapolate the future incidence, we concluded that at least 50% of MSM would need to be consistently tested annually or to use PrEP to attain a 90% incidence reduction by 2030. Both testing and PrEP strategies each raised the proportion of MSM living with HIV and aware of their infection to around 95% while maintaining the proportion of MSM living with HIV who were on treatment at around 90%. Thus, achievement of the three 90-90-90 targets might not lead to the 2030 goal29 and, in the UK MSM setting, achievement of and sustaining a 95-90-90 target by 2020 would be required.
Our study has several limitations. First, our analyses are, by necessity, a simplification of the real world, and might overestimate actual programme effectiveness. For illustrative purpose we show results at 100% coverage, but this is clearly unrealistic. However, the overall conclusions remained similar when the coverage was reduced and we also applied a much smaller level of coverage in the practical scenarios to reflect more conservative outcomes. Second, we did not explicitly include the effects of ART resistance30 and co-infection with other sexually transmitted diseases31 on HIV transmission. Unless the prevalence of co-infection and drug-resistant strains changes substantially during the intervention period, their effects on the effect of interventions are not likely to be influential. Third, we did not account for the UK regional difference. Although this limitation might not have a substantial effect on the effect of interventions at the national level, the regional difference might play an important part in some specific areas. Fourth, changes of risk within individuals was also not accounted for. High-risk behaviour may be relatively transient and take place occasionally over time. However, this should have no substantial effect on our findings because we saw no evidence of a major fluctuation in the overall proportions of high-risk MSM in the UK over the past decade. Fifth, even with our practical scenarios the scale of the modelled interventions could be practically too large, and a more targeted programme aimed at groups with highest risks could be more feasible. Finally, our analysis did not include cost-effectiveness, which is also crucial to inform public health decision making. That any highly effective programmes against HIV transmission suggested in this study might not be cost effective in the real UK setting is a possibility. Although a simple cost approximation of implementation of the interventions can be done with the estimated numbers of programme participants provided in the tables, further in-depth economic analyses are needed to inform timely and accurate decision making. Future work should include investigation of the effect of changes in other behavioural factors (eg, partnership duration and mixing preferences) on the effect of interventions.
In conclusion, our analysis suggests that a combination of PrEP, expansion of HIV testing, and test-and-treat programmes implemented in small groups of MSM in the UK could prevent thousands from HIV within a few years of implementation, particularly if risk-compensation behaviour is successfully avoided. Integration of this enhanced version of conventional interventions with novel biomedical technologies is likely to deliver the optimum approach to HIV prevention that could determine the future course of the HIV epidemic in the UK.
The model estimates of the number of new HIV infections in UK MSM in the past decade showed good consistency with the national estimates (figure 1). The status-quo scenario estimated a total of 16 955 HIV infections in MSM in the UK, with around 2400 new infections annually during 2014-20 (table 1), which is equivalent to about 6·6 new HIV infections every day. Without additional interventions, 52 268 MSM are expected to be living with HIV in the UK by 2020, with an annual incidence of 0·34% (table 1).
With 100% programme coverage, PrEP prevented the greatest number of HIV infections (table 1). Even when targeted only at high-activity men, PrEP was more effective than all other individual interventions that targeted the entire UK MSM population (table 1, appendix p 32). A large number of MSM could also be protected from HIV infection by decreasing the number of repeat sexual partners by 0·5 times (table 1). Decreasing the proportion of men who had unprotected anal intercourse with repeat sexual partners by 0·5 times reduced incidence by almost a third during the intervention period (table 1).
An estimated 5522 new infections (32·6% of total incidence) would be prevented by testing annually at 100% coverage (table 1). Twice-yearly testing in all UK MSM increased the number of prevented infections to 7089, which is equivalent to a 41·8% reduction in incidence (table 1). Early provision of ART (test and treat) reduced total incidence by 23·9%; one test-and-treat participant could prevent 0·322 new infections during 2014-20, which is around 17 times higher than PrEP (unit effect 0·019; table 1).
Alteration of the coverage of interventions had a large effect on prevention of new infections (figure 2). Reducing repeat sexual partnerships was the most affected intervention, with a decrease of about 75% in effect as coverage reduced from 100% to 25%, whereas the same reduction in coverage resulted in a 55% decrease in effect for testing once a year, and of 37% for testing twice a year (figure 2). The robustness to variation in programme coverage suggests that expansion of HIV testing and treatment might still be effective even when adopted by a smaller proportion of MSM. At 25% coverage of all interventions, testing twice a year produced the greatest effect; the effect of test-and-treat programmes was greater than for most other interventions and was on par with PrEP (figure 2). The other individual interventions tended to show a roughly linear association between coverage and effect (figure 2).
With an assumed effectiveness of 44%, if coverage was 50% rather than 100%, PrEP would fail to prevent 4144 new infections (figure 2). Increasing the effectiveness raised the number of infections prevented, irrespective of the coverage (figure 2). The PrEP intervention with 100% coverage and 20% effectiveness was able to prevent 1275 more cases than 25% coverage with 100% effectiveness (6893 cases vs 5618 cases). However, at low to moderate effectiveness (≤60%) but high coverage (≥75%), increasing PrEP coverage seemed to have a slightly greater effect than did increasing effectiveness (figure 2). On average, a 1% increase in PrEP coverage led to an increase of 1·90% in effect, whereas a 1% increase in PrEP effectiveness led to an increase of 1·25% (figure 2).
We recorded diminishing returns for most combined interventions. For example, programmes to increase frequency of HIV testing and decrease unprotected anal intercourse with repeat sexual partners were estimated to each prevent about a third of new infections when implemented separately at 100% coverage, whereas combining both programmes achieved only 49·2% incidence reduction (table 1). Only the combination of annual HIV testing with the test-and-treat programme was able to provide an incidence reduction greater than the sum of the effects of the two individually (appendix, p 36).
We investigated 12 practical scenarios (table 2). The first scenario (3.1) assumed that PrEP was provided to 25% of the high-activity HIV-negative MSM in the UK and, for high-activity men who did not use PrEP, 25% had an annual HIV test. This strategy prevented 26·4% of HIV infections, an effect that increased to 37·2% when PrEP coverage was increased to 50% (scenario 3.2). Addition of test and treat to scenario 3.1 prevented 7399 cases (43·6% of total incidence), the highest number among all the practical scenarios assessed (scenario 3.3). Scenario 3.3 also provided the largest unit effect (0·053). Replacement of test-and-treat strategies with programmes to encourage fewer sexual partners (scenario 3.4) and less unprotected anal intercourse (scenario 3.5) had a lower effect than the test-and-treat scenarios (appendix, p 40).
In the analysis of risk compensation in PrEP users, scenario 3.6 assumed that all PrEP men in scenario 3.1 completely stopped using condoms with repeat sexual partners. This assumption resulted in HIV incidence effect falling to 11·5% (table 2). Surprisingly, scenario 3.7, which assumed that all PrEP men in scenario 3.1 acquired twice as many repeat sexual partners, increased incidence effect to 28·0% compared with scenario 3.1 (26·4%; table 2). This result was because more sexual partnerships in PrEP men would require a larger number of non-PrEP partners, including infected men, which would reduce the probability that non-PrEP susceptibles had serodiscordant relationships, particularly with undiagnosed MSM. Combining the two risk compensations from scenarios 3.6 and 3.7 resulted in the incidence effect of 22·3% (scenario 3.8; table 2).
In scenario 3.9, we assumed that 25% of the entire MSM population reduced repeat sexual partnerships by half, and 25% of high-activity MSM had an annual HIV test. The effect was 22·2% in this scenario (table 2). When the test-and-treat programme with 25% coverage was added to scenario 3.9, the effect increased to 40·6% (scenario 3.10; table 2). However, the benefit of early ART was completely negated if high-activity men tested for HIV half as frequently (scenario 3.11; table 2). If even more risk-compensation behaviour took place, with the assumption that all ART-treated MSM completely stopped using condoms with repeat sexual partners on the basis of the benefit of immediate ART, the effect would fall to 15·8% (scenario 3.12; table 2).
The number of new infections prevented was sensitive to sexual risk compensation, particularly an increase in the number of repeat sexual partners. Most interventions could tolerate up to 75% or more increase in both risk compensations before the benefits were completely negated. However, the sensitivity of the model outcomes to decreasing frequency of HIV testing was relatively less pronounced compared with the sexual risk compensation (appendix, p 19).