Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 Anti-HIV Drugs for Prevention Bernard Hirschel Division of HIV/AIDS Geneva University Hospital Geneva, Switzerland.

Similar presentations


Presentation on theme: "1 Anti-HIV Drugs for Prevention Bernard Hirschel Division of HIV/AIDS Geneva University Hospital Geneva, Switzerland."— Presentation transcript:

1 1 Anti-HIV Drugs for Prevention Bernard Hirschel Division of HIV/AIDS Geneva University Hospital Geneva, Switzerland

2 2 Prevention at an Impasse Sexual behaviour Condoms Circumcision Microbicides Vaccines

3 3 The theory is just fine… Estimated rate of HIV after 10 years’ cohabitation in a heterosexual couple always using condoms Adapted from W. Cates, FHI

4 4 Practice rather less so… Estimated rate of HIV after 10 years’ cohabitation in a heterosexual couple at typical rates of condom use Adapted from W. Cates, FHI

5 5 ART: potentially more efficacious (than any of the previously evaluated prevention methods)

6 6 « Rakai » Study: Transmission risk as a function of plasma viral load Quinn et al. N Engl J Med 2000 No transmission if VL « undetectable »

7 7 Mother to Child Transmission AZT HAART % of infected children Adapted from Coovadia and Lallemant, NEJM 2004

8 8 Effect of HAART on heterosexual transmission 476 heterosexual couples in Madrid The “index” patient was HIV +, and consulted between Among their sexual partners, the only risk factor for HIV was exposure to the “index” patient All partners tested to establish prevalence of HIV among partners Del Romero J, BMJ 2010

9 9 Prevalence of HIV infection in the partners, at entry 44/4760/149 Percent of partenairs infected Del Romero J, BMJ 2010

10 10 Incidence of HIV infection in the partners, during follow-up TreatmentCouple-yearsNew infections No HAART8635 HAART4170 Del Romero J, BMJ 2010

11 11 Condoms plus HAART in comparison to condoms alone: 3 African studies What they had in common: Sero-discordant heterosexual couples Condom promotion How they differed: Study A*, without HAART Study B**, with HAART Study C***, a period without HAART followed by a period with HAART * Wawer M et al. Lancet 2009 **Bunnell R et al. Abstract 29, 15th CROI, Boston 2008; AIDS 2006 ***Donnell D, Lancet 2010

12 12 Results Study A*Study B**Study C*** « before »« after » Condoms HAART- - Infections %/year * Wawer M et al. Lancet 2009; 374: **Bunnell R et al. Abstract 29, 15th CROI, 2009 *** Donnell D, Lancet 2010

13 13 In conclusion Circumstantial evidence indicates that: HAART lowers MTCT HAART lowers heterosexual transmission HAART appears more efficacious than condoms (or has a marked additional effect when used in combination with condoms), in sero-discordant heterosexual couples

14 14 What has been the effect of ART on the AIDS epidemic ? Introduction of ART, Expansion of ART,

15 15 Bull. OFSP 2001 Newly Diagnosed HIV Infections in Switzerland AZT HAART

16 16 Montaner J, CROI 2010 Newly Diagnosed HIV Infections, and N of pts on HAART in B.C., Canada

17 17 Amsterdam, MSMs New infections per seropositive individual If > 1, the epidemic grows, if < 1, it shrinks. 1.5 Theoretical R(t) in 2002, without HAART, but with the 2002 sexual behavior Bezemer D et al., CROI 2001, AIDS 2008 R(t)

18 18 Conclusions Introduction of ARVs in the 1990ies coincided with a decrease in new HIV infections Other things being equal, without HAART, new infections might have been 50 to a 100 percent more frequent by 2000

19 19 Montaner J, CROI 2010 Newly Diagnosed HIV Infections, and N of pts on HAART in B.C., Canada IDUs only

20 20 Swiss HIV Cohort: More patients with stably suppressed viral load Adapted from Ledergerber et al. CROI 2010

21 21 More patients are treated, and less are infectious (VL > 500, pink) > < 500

22 22 Swiss Cohort: N of pts with viremia > 500

23 23 Theory: What you would expect The number of potentially infectious persons starts to fall....and after a lag of a few years newly discovered infections decline in parallel

24 24 Switzerland: Newly Diagnosed HIV Infections, and N of pts with viremia > 500 in the SHCS* * SHCS = Swiss HIV Cohort Study

25 25 Conclusions (2000 to 2010) 1)Expansion of treatment, and better efficacy, diminished the pool of potentially infectious persons 2)The number of newly discovered infections, after years of stability or even increase, may be declining again

26 26 What would happen to the epidemic if even more infected persons were treated?

27 Montaner, Hogg et al. Unpublished, 2006 Treat all Treat 30% HIV infections per 1000 population Projections for British Columbia

28 Costs of treatment Treat all Treat 30% $ Hogg et al. Unpublished, 2006

29 Costs of HAART Billions de $ Small investment Great savings Lima VD et al. JID 2008 Hogg et al. Unpublished, 2006

30 30 If Lima and Montaner are right, it is enough to treat > 75% of those with CD4 counts < 350, and after a while, there will be close to zero new infections: HIV will disappear (Lima et al. JID, 2008) Treatment coverage 50 % 75 % 90 % 100 %

31 ARV for CD4 < 350 Granich et al But if Granich and Williams are right, HIV will never be defeated by treating only those with CD4 counts below 350. One needs to test the whole population frequently, and treat all those found to be infected (green curve)

32 32 All models are wrong, but some are useful… Box, G.E.P., Robustness in the strategy of scientific model building, in Robustness in Statistics, R.L. Launer and G.N. Wilkinson, Editors. 1979, Academic Press: New York.

33 33 Blower et al., Science 2000 Percent of Cases Treated 50%60%70%80%90% Percent of New Infections Prevented

34 34 50%60%70%80%90% Percent of New Infections Prevented Depending on assumptions regarding risk behaviour and selection of drug resistance, at 70% of patients treated, the effect on new infections could range from 40% prevented to doubled

35 35 Kahn JG et al. 17th CROI, 2010, abstract 965 Compared to treating only patients with less than 350 CD4 cells, universal ART would reduce costs by 4.9 billion US$ … « I therefore believe that you have substantially underestimated, and deliberately misrepresented, the costs of your proposed elimination strategy »

36 36 Let’s declare a model armistice, and instead get some data ! Dabis, Newell, Hirschel, Lancet 2010

37 37 HPTN* 052 Myron Cohen et al.: Randomised study to evaluate HAART in preventing sexual transmission in sero-discordant couples * HPTN: Health Prevention and Treatment Network

38 38 Inclusion Criteria 1750 couples (3500 individuals), fully recruited CD4 = cells/mm3 One partner HIV+, the other HIV- Endogamous: 93% say that they had only one sex partner during the last six months.

39 39 Randomisation In the intervention group, ART starts right away, i.e. at a CD4 level between 350 and 550 In the control group, according to local indication (CD4 = 200 to 250, rising more recently to 350)

40 40 Endpoints 1.Transmissions 2.Sustainability, with a planned follow-up of 5+ years Results expected in 2015

41 41 Limitations of HPTN052 Stable serodiscordant heterosexual couples are only part of the problem In the general population individual randomisation is not feasible, because it would necessitate tracking and testing of all sex partners

42 42 Measure the incidence of HIV before and after introduction of ART, or before and after expansion of ART Examples: British Columbia (J. Montaner et al.), San Francisco, Switzerland General population: The before-and-after approach

43 43 Switzerland: Newly Diagnosed HIV Infections, and N of pts with viremia > 500 in the SHCS* * SHCS = Swiss HIV Cohort Study

44 44 Newly Diagnosed HIV Infections, and N of MSMs with viremia > 500 in the SHCS* MSMs * SHCS = Swiss HIV Cohort Study

45 45 February to June 2008: A campaign to eliminate acute HIV infections in MSMs in Switzerland

46 46 Problems with the before-and- after approach 1.Not all evidence goes into the same direction 2.B after A  B because of A (If HIV incidence falls after expansion of ART, it is not certain that the expansion caused the fall)

47 47 « Children whose mothers listenend to Mozart during pregnancy are more intelligent » « The H1N1 vaccine has caused my multiple sclerosis because 5 days afterwards I started having double vision » (Tribune de Genève, 15 mars 2010) The « post hoc ergo propter hoc » - fallacy is not new…

48 48 How To Do Better Compare two regions, at the same time: – Region 1: Test-and-Treat – Region 2: Continue as before Measure, in both regions, the number of new HIV infections Better but not perfect: The two regions will differ in many respects: – Number and type of sex partners – Use of condoms – Prevalence of circumcision – Age

49 49 … but if one compared not two, but thirty regions, 15 with expanded ART, 15 without, and each time, HIV incidence falls in the « Test-and-Treat » regions but remains the same in the control regions …

50 50 This is the essential idea behind the methodology called « cluster-randomized trial », where the unit of randomization is not the individual, but a community of individuals, for instance a village … but if one compared not two, but thirty regions, 15 with expanded ART, 15 without, and each time, HIV incidence falls in the « Test-and-Treat » regions but remains the same in the control regions …

51 51 ANRS Working group (AC12, B. Hirschel, F. Dabis): Priorities 2010 – 2015: Treatment as prevention (TasP), or « Test-and-Treat (TnT) » Together with a partner in Africa

52 52 The Africa Center for Health and Population Studies Durban Johannisburg Marie-Louise Newell

53 53 Ukuphila kwami, ukuphila kwethu* Test-and-Treat A cluster-randomized trial in Hlabisa, KwaZulu-Natal, South Africa Francois Dabis and Marie-Louise Newell (PIs), Africa Centre and Hlabisa Department of Health collaborators Pilot Phase 2010 – 2013, Trial Phase Approximate budget of 3’000’000 € for the Pilot Phase only, with important downstream commitments ANRS is « interested », and looking for additional financing from other sources * My Health for Your Health

54 54 Basic Plan Screen « everybody » 2 arms : –Intervention Clusters: Treat all who screen HIV+ –Control clusters: HIV+ with treatment indications according to local guidelines, but using the type of HAART prescribed in the Intervention Clusters.

55 55 Endpoints Primary –Incident HIV infections, as measured by repetitive 6- monthly screening Secondary –Acceptability and results of widespread testing –Behavioral modifications –Costs and cost-effectiveness –Morbidity and mortality in HIV+ HIV related « Non-HIV-related » HAART-related Test-and-Treat is also a «when-to-start » study

56 56

57 57 Hurdles 1.Attrition 2.Harm (to the individual) versus benefit (to the community) 3.Cost and sustainability

58 58 Attrition… Intervention Group Not all will be tested Of those who are tested, some will not receive their results Of those who receive results and are HIV+, not all will be treated Of those who are treated, not all will have effective treatment Of those with effective treatment, not all will continue

59 59 Attrition… Intervention Group Not all will be tested Of those who are tested, some will not receive their results Of those who receive results and are HIV+, not all will be treated Of those who are treated, not all will have effective treatment Of those with effective treatment, not all will continue Control Group Some already on ART Proportion on ART expected to increase: –Expansion of access –Revision of indications (limit of 350 CD4 cells instead of 200) Of those who remain off treatment, many will use other prevention methods: –Condoms –Microbicides –Circumcision

60 Attrition: Example from MTCT Source: Stringer EM et al. AIDS 2003; 17:3077 (Lusaka, Zambia)

61 61 Harm versus Benefit Consider risks and benefits both to the individual, and the community Health benefits –Persons in intervention clusters will probably have less HIV-related diseases. This will be a secondary endpoint in the trial –Effect of tuberculosis could be particularly beneficial: Occurs at higher CD4 counts « Spill-over » into the HIV-negative population

62 62 Risks and beneftis (continued) Medical risks –Asymptomatic individuals with intact immune systems may derive little or no benefit, and probable side effects, from HAART Non-medical risks –Without widespread acceptance, test-and-treat programs cannot succeed –Test-and treat-programs must avoid undue pressure on individuals to get tested and begin treatment –On the plus side, the perception that treated patients are not longer infectious, may decrease stigma and discrimination

63 63 Resources and Sustainability If treatment has a preventive effect, it will increase pool of people potentially eligible for HAART –Will increase pressure for availability of ARVs –Will increase costs in the short run –Long-term sustainability and resistance is certainly an issue. –Operational research would provide years of follow-up and surveillance of infection, without randomisation

64 64

65 65

66 66

67 67 Acknowledgements Africa Center: M.-L. Newell S. Timberlake, J. Amon ANRS: Brigitte Bazin, J.-F. Delfraissy, F. Dabis J. Montaner R. Granich M. Cohen Th. Wägli A. Calmy, E. Boffi, J. Ambrosioni, C. Delhumeau F. Schöni, N. Low, B. Althaus, M. Egger, O. Keiser B. Vidondo, M. Gebhart

68 68 Spare Slides

69 69 N on HAART And Newly Diagnosed Infections in Switzerland, N on HAART Newly discovered HIV Why this marked decrease from 2008 to 2009 ? New drugs with less resistance? New infections N treated

70 70 N on HAART And Newly Diagnosed Infections in Switzerland, N treated MSM N on HAART Newly discovered HIV Total new

71 71 N on HAART And Newly Diagnosed Infections in Switzerland, N treated MSM Hetero N on HAART Newly discovered HIV Total new

72 72 Attrition: Example from MTCT Pregnant women Received counseling % Tested % Received test results % Positive % Received nevirapine –Mothers % –Children % Partners tested % ObservedIdealObs/Ideal Source: Stringer EM et al. AIDS 2003; 17:3077 (Lusaka, Zambia)

73 73 Newell et al. 32 districts 16 with, 16 without intervention Approx persons

74 74 Risk Compensation Increased sexual risk taking in intervention clusters –May, or may not matter while patients are on HAART –Has to be considered in the local context (where infection rates at times have been shockingly high even in trial settings with condom promotion) –Some reassurance provided by the circumcision trials, where there was no increase in sexual risk taking in intervention groups

75 75 Follow-up Test-and-Treat 1: Years 1 and 2: Evaluation Test-and-Treat 2: Years 3 to 5 (to 10?)

76 76 TnT1 compared to TnT2 TnT1TnT2 Focus onefficacysustainability Randomisation unitClusterNo randomisation Indication for HAART in intervention groupAll HIV+All HIV+ Indication for HAART in control groupusual careNo control group Follow-up2 years3-8 years

77 77 N on HAART N on HAART increases in Switzerland

78 78 N on HAART N on HAART And Newly Diagnosed HIV Infections in Switzerland Newly diagnosed HIV N treated

79 79 N on HAART N on HAART And Newly Diagnosed HIV Infections in Switzerland Newly diagnosed HIV N treated

80 80 Newly Diagnosed HIV Infections, and N of pts with > 500 in Switzerland IDUs

81 81 Newly Diagnosed HIV Infections, and N of pts with viremia > 500 in Switzerland Heterosexuals


Download ppt "1 Anti-HIV Drugs for Prevention Bernard Hirschel Division of HIV/AIDS Geneva University Hospital Geneva, Switzerland."

Similar presentations


Ads by Google