1. SMART

2. Limitations of Current Antiretroviral Therapy Inability to eradicate HIV Limited number of agents/classes Development of HIV resistance Occurrence of serious side effects Difficulty with long-term adherence

3. SMART Short-term trials Observational studies Expert opinion Basis for Antiretroviral Guidelines Only partially evidence-based

4. SMART 2001 DHHS Guidelines Plasma HIV Clinical Category CD4+ Count RNA Recommendation Symptomatic Treat!! Asymptomatic > 200/mm 3 Any Treat. But… and < 350/mm 3 Asymptomatic < 200/mm 3 AnyTreat! Asymptomatic > 350 Treat? > 55,000 Asymptomatic > 350 Treat?? < 55,000

5. SMART A Large, Randomized Trial Comparing Two Strategies for Management of Anti- Retroviral Therapy

6. SMART The SMART Study is a: Multicenter Study of the Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA) Funded by: NIH: The National Institute of Allergy and Infectious Diseases Division of AIDS

7. SMART Countries Participating in SMART CPCRA RCC Brazil Canada Peru South Africa United States Sydney RCC Argentina Australia Chile Israel Japan New Zealand Thailand Uruguay Copenhagen RCC Austria Belgium Denmark Estonia Finland Germany Lithuania Luxembourg Norway Poland Portugal Russia Spain London RCC France Greece Ireland Italy Morocco Switzerland United Kingdom

8. SMART SMART Study Design Drug Conservation (DC) Strategy [Stop or defer ART until CD4+ 350] Virologic Suppression (VS) Strategy [Use ART to maintain viral load as low as possible throughout follow-up] Participants with CD4+ > 350 n = 3000 Expected follow-up period: 7-8 years

9. SMART The SMART Study Questions What is the optimal way to use ART to achieve longest disease-free survival? Which treatment strategy is associated with –Better adherence –Fewer drug side effects –Fewer metabolic complications –Better quality of life –Less drug resistance? Which strategy is more cost effective?

10. SMART Study Treatment Comparison Viral Suppression Strategy Continuous Antiretroviral Therapy VS Drug Conservation Strategy Episodic Antiretroviral Therapy

11. SMART Primary Objective To compare the VS and DC strategies in prolonging survival without progression of disease.

12. SMART Other Major Clinical Outcomes Death Major cardiovascular and metabolic complications Serious HIV progression events Combined endpoint of death, HIV disease progression, and major cardiovascular and metabolic events Grade 4 adverse events

13. SMART SMART Target Population HIV-infected patients CD4+ cell counts > 350 cells/mm 3 Older than 13 years of age Willing to initiate, modify, or stop antiretroviral therapy, in accordance with the randomized assignment.

14. SMART Viral Suppression (VS) Strategy Goal Maintain maximum possible suppression of viral load Treatment Strategy Utilize any available HIV treatments irrespective of CD4 cell count

15. SMART Potential RISKS More cumulative side effects Adherence more difficult Resistance to more agents with virologic failures Fewer active anti-HIV agents available when risk of disease is higher Potential BENEFITS Maximal suppression of viral load more likely Lower risk of resistance due to viral suppression Lower risk of lasting CD4 depletion Potential lower risk of HIV transmission Viral Suppression Strategy

16. SMART Drug Conservation (DC) Strategy Goal Conserve anti-HIV drug options while the risk of disease progression is low Treatment Strategy Defer use of anti-HIV agents until CD4 cell count is 350 using any available HIV treatments

17. SMART Potential RISKS Stopping and restarting anti-HIV agents may increase the risk of drug resistance Risk of lasting damage to immune system Potential increased risk of HIV transmission Potential BENEFITS Potent anti-HIV agents available when risk of disease increases Fewer side effects Less anti-HIV drug use may reduce the risk of drug resistance Adherence may be better with episodic therapy Drug Conservation Strategy

18. SMART Enrollment from 8 January 2002 to 11 January 2006 –33 countries –318 sites –5,472 patients Date randomization screens closed: 11 January 2006 Mean follow-up: 15 months SMART Participants and Timetable

19. SMART Canadian Participation 13 sites across Canada First Canadian site opened in December 2004 10 of 13 sites were opened for enrolment

20. SMART Canadian Enrolment (to 11 January 2006) Site CHUL (Ste-Foy, Quebec) Toronto Hospital (Ontario) McMaster (Hamilton, Ontario) Halifax (Nova Scotia) CHUS (Fleurimont, Quebec) Montreal Chest (Quebec) Windsor (Ontario) St. Joseph’s (London, Ontario) L’actuel (Montreal, Quebec) Medical Halles (Ste-Foy, Quebec) Total = Enrollment 14 5 8 9 13 4 7 4 29 9 102/150

21. SMART Canadian Sites Three sites not opened for enrolment: Quartier Latin (Montreal, Quebec) Sudbury (Ontario) DIDC (Vancouver, British Columbia)

22. SMART SMART Enrollment by Quarter Total Number Enrolled 240 243 186 153 117 127 170 250 292 240 293 351 503 623 547 942 195

23. SMART Enrollment by Geographic Region 57% 26% 10% 3% 1%

24. SMART SMART Study: Enrollment Through 11 January 2006 DCVSTotal Number Enrolled Study Main272027525472

25. SMART Baseline Characteristics Patients with data available as of 10 Dec 2005 Number244324644907 Age (years; mean)464646 Female (%)252726 Race: Black (%)303231 White (%)575556 Other (%)131313 Likely mode of infection: Sexual contact, same sex (%)535052 Sexual contact, opposite sex (%)424443 Injection drug use (%)101010 Other/ unknown (%)898 DC GroupVS Group Total

26. SMART Baseline Characteristics Patients with data available as of 10 Dec 2005 Number244324644907 CD4+ (cells/mm 3 ; median)599598598(466, 792) CD4+ nadir (cells/mm 3 ; median)253253253(153, 364) HIV RNA ≤ 400 copies/mL (%)69.869.469.6 HIV RNA < 1000 copies/mL (%)74.373.974.1 Highest log HIV RNA (log copies/mL; median)4.84.84.8 DC GroupVS Group Total IQR

27. SMART Baseline Characteristics Patients with data available as of 10 Dec 2005 Number244324644907 Prior AIDS-related illnesses (%)24.423.423.9 Hepatitis B co-infection (%)2.52.32.4 Hepatitis C co-infection (%)16.114.915.5 DC GroupVS Group Total

28. SMART Baseline Characteristics Patients with data available as of 10 Dec 2005 Number244324644907 ART History ART naïve (%)4.75.14.9 PI experienced (%)68.966.567.7 NNRTI experienced (%)64.664.364.5 On ART at baseline (%)83.382.582.9 Time since first prescribed ART (years; median)666(3, 8) DC GroupVS Group Total IQR

29. SMART Mean: 661 Median, [ 25 th ; 75 th ]: 598[ 466 ; 792 ] Distribution of Baseline CD4+ Cell Count No. of Patients

30. SMART Mean: 270 Median, [ 25 th ; 75 th ]: 253 [ 153 ; 364 ] Distribution of CD4+ Nadir Prior to Enrollment No. of Patients

31. SMART Intent-to-treat DC versus VS comparisons –Kaplan-Meier survival curves and Cox’s proportional hazard models used to compare treatment groups for: Progression to AIDS or death Survival Major cardiovascular and metabolic events Serious disease progression events Grade 4 events –Data cutoff date used in clinical events analyses: 10 December 2005; at that time 5,007 patients were randomized. SMART Statistical Methods

32. SMART Follow-up by Treatment Group Number of follow-up 761 (4.6%) 1,061 (6.4%) visits missed (%) Number lost to 48 (1.9%) 55 (2.2%) follow-up (%) Median months of 10 (4,23) 10 (4,23) follow-up (IQR) Average follow-up (months)14.7 14.7 Total person years 3062 3077 through 10 Dec 2005 DC Group VS Group

33. SMART Hypothetical CD4+ Cell Count Patterns Over Follow-up

34. SMART Confirmed Clinical Events + - 1 through 10 December 2005 DC GroupVS Group RR (DC/VS) P-valueNRate [95% CI] Progression of disease or death (primary endpoint)933.1441.42.15 [1.50, 3.08]<0.0001 Death471.5290.91.63 [1.02, 2.58]0.04 Group I 1 events592.0371.21.62 [1.07, 2.44]0.02 Major CVD or metabolic complications 2 832.9732.51.14 [0.83, 1.56]0.41 + Reviewed and adjudicated by Endpoint Review Committee 1 Fatal or non-fatal MI (reported as supplemental event or diagnosed by ECG), stroke, CAD requiring surgery, kidney failure, cirrhosis. 2 Fatal or non-fatal MI (reported as supplemental event or diagnosed by ECG), CAD requiring drug treatment or an invasive procedure, stroke, myocarditis, pericarditis, diabetes requiring drug treatment, pancreatitis, lactic acidosis, osteonecrosis.

35. SMART Time to Disease Progression or Death Confirmed events through 10 December 2005

36. SMART Time to Death Events through 10 December 2005

37. SMART SMART Primary Composite Endpoint (Disease Progression or Death) and Components; Confirmed Events Through 10 Dec Favors VS ► ► Favors DC No. of Patients with Events Endpoints Relative Risk (95% CI) Progression of Disease or Death137 Death76 Serious Progression Event19 Other Progression Event46 Serious Progression of disease or death91 > 2.2 1.6 5.8 2.9 1.9

38. SMART Time to Group I Event 1 Confirmed events through 10 December 2005 1 Fatal or non-fatal MI (reported as supplemental event or diagnosed by ECG), stroke, CAD requiring surgery, kidney failure, cirrhosis.

39. SMART SMART Group I Events and Components: Confirmed Events Through 10 December No. of Patients with Events Subgroups Total96 Cardiovascular78 Liver15 Renal9 Favors VS ► ► Favors DC Relative Risk (95% CI) 1.6 1.5 2.0

40. SMART Confirmed Clinical Events + - 2 through 10 December 2005 DC GroupVS Group RR (DC/VS) P-valueNRate [95% CI] Serious progression of160.530.15.82 [1.68, 20.2]0.01 disease 1 Serious progression of disease 1 or death591.9321.01.88 [1.22, 2.90]0.003 Grade 4 events1575.41334.51.19 [0.94, 1.49]0.15 Total follow-up time (person-years)30623077 + Reviewed and adjudicated by Endpoint Review Committee. 1 Progressive multifocal leukoencephalopathy, lymphoma, visceral Kaposi’s sarcoma, AIDS dementia complex, toxoplasmosis, histoplasmosis, cryptococcosis, MAC, wasting syndrome, cytomegalovirus disease.

41. SMART Rate of AIDS-related Illnesses through 10 December 2005 DC GroupVS Group Events 1 Rate 2 Aspergillosis, invasive0010.0 Candidiasis, esophageal170.650.2 Candidiasis of bronchi, trachea, or lungs 20.100 CMV disease 10.000 Cryptococcosis, extrapulmonary 10.000 Encephalopathy, HIV-related, stage 2 or higher 10.000 Herpes simplex30.120.1 Herpes zoster, disseminated30.110.0 Kaposi’s sarcoma60.210.0 Lymphoma40.110.0 TB, pulmonary or extrapulmonary20.120.1 MAC, extrapulmonary10.010.0 PCP60.220.1 Pneumonia, bacterial 3 70.220.1 Wasting syndrome due to HIV30.100 Patients with any AIDS-related event541.8170.6 1 One patient may have several illnesses. Each occurrence is counted, but recurrent illnesses of the same type are counted only once. 2 Per 100 person-years. 3 Recurred within 1 year. Events 1 Rate 2 AIDS – related event

42. SMART Causes of Death Deaths through 10 December 2005 DC GroupVS Group Deaths 1 % AIDS-related36.413.4 Hepatic complications36.4413.8 Pancreatic/ GI complications12.100 Cancer, excluding AIDS-related714.9517.2 Cardiovascular complications1123.41034.5 Violent/ Accident919.1931.0 Other, none of the above 2 612.826.9 Unknown919.126.9 Total number of deaths47100.029100.0 1Patients may have multiple causes of death. 2Other includes the following causes: DC (cerebral empyema, renal failure – 2 patients, ARDS/septic shock, bacterial meningitis, COPD/pneumonia) and VS (intra-abdominal sepsis/multi-organ failure, pneumonia). Deaths 1 % Death Classification

43. SMART Progression of Disease or Death by Baseline CD4+ Confirmed Events Through 10 December 2005 0.07 350 - 449 21.7213.36172.401.4 0.30 [0.7,2.7] 450 - 54920.4233.8060.964.0 <0.005 [1.6,9.9] 550 - 64915.5143.0361.292.4 0.08 [0.9,6.2] ≥ 65042.5352.67151.192.3 0.01 [1.2,4.1] CD4+ (cells per mm 3 ) % of patients in sub- group EventsRate 1 EventsRate 1 RR 2 (DC/VS) [95%CI] P-Value 3 Inter- action P-Value 2 for RR 1 Per 100 person-years, time to first event. 2 Cox proportional hazards model within subgroup 3 Subgroup by treatment group interaction, Cox proportional hazards model DC GroupVS Group

44. SMART Progression of Disease or Death for Baseline CD4+ Subgroups No. of Patients with Events Subgroups Total137 CD4 (cells/mm 3 ) 350 - 44938 450 - 54929 550 - 64920 ≥ 65050 Favors VS ► ► Favors DC Relative Risk (95% CI) 2.2 1.4 4.0 2.4 2.3

45. SMART Progression of Disease or Death by Nadir CD4+ Confirmed Events Through 10 December 2005 0.42 < 10016.4183.6461.133.2 0.01 [1.3,8.2] 100-19918.4162.79101.981.4 0.38 [0.6,3.1] 200-29925.621 3.09111.402.2 0.03 [1.1,4.6] 300-39920.0254.09111.842.1 0.04 [1.1,4.4] ≥ 40019.5132.0160.932.1 0.13 [0.8,5.6] Nadir CD4+ (cells per mm 3 ) % of patients in sub- group EventsRate 1 EventsRate 1 RR 2 (DC/VS) [95%CI] P-Value 3 Inter- action P-Value 2 for RR 1 Per 100 person-years, time to first event. 2 Cox proportional hazards model within subgroup 3 Subgroup by treatment group interaction, Cox proportional hazards model DC GroupVS Group

46. SMART Progression of Disease or Death for Nadir CD4+ Subgroups No. of Patients with Events Subgroups Total137 Nadir CD4 (cells/mm 3 ) 100 – 19926 200 – 29932 300 – 39936 ≥ 40019 < 10024 Favors VS ► ► Favors DC Relative Risk (95% CI) 2.2 3.2 1.4 2.2 2.1

47. SMART Progression of Disease or Death by Baseline Viral Load for those on ART Confirmed Events Through 10 December 2005 <0.01 ≤ 40081.7553.02170.943.2 <0.005 [1.9,5.6] > 40018.3193.04182.851.1 0.87 [0.6,2.0] HIV RNA (copies/mL) % of patients in sub- group EventsRate 1 EventsRate 1 RR 2 (DC/VS) [95%CI] P-Value 3 Inter- action P-Value 2 for RR DC GroupVS Group 1 Per 100 person-years, time to first event. 2 Cox proportional hazards model within subgroup 3 Subgroup by treatment group interaction, Cox proportional hazards model

48. SMART Progression of Disease or Death for Baseline HIV RNA Subgroups among Those Taking ART No. of Patients with Events Subgroups Total137 HIV RNA (copies/ mL) (only patients on ART at baseline) ≤ 40072 > 40037 Favors VS ► ► Favors DC Relative Risk (95% CI) 2.2 3.2 1.1

49. SMART Percent of Patients on ART at each Month of Follow-up and Treatment Group

50. SMART CD4+ Change from Baseline, Mean + 2SE (in cells/mm 3 ) by Month of Follow-up and Treatment Group

51. SMART Hypothetical CD4+ Cell Count Patterns Over Follow-up

52. SMART Percent with HIV RNA ≤ 400 copies/mL by Month of Follow-up and Treatment Group

53. SMART Summary - 1 There is an increased risk in the DC group compared to the VS group of: –Progression to AIDS, including death –Death –Serious disease progression events –Major cardiovascular, renal and liver events

54. SMART Summary - 2 The increased risk of the DC group compared to the VS group did not differ according to subgroups defined by baseline and nadir CD4+ cell count. For the subgroup of patients who entered with a viral load < 400 copies/mL on treatment, risk of progression was over 3- fold higher in the DC group compared to the VS group. For other subgroups examined, risk was always greater in the DC group than the VS group.

55. SMART Conclusion of the SMART executive committee Episodic use of ART based on CD4+ cell count levels as per the SMART study design is inferior to continuous ART for the management of treatment experienced patients and thus should not be routinely recommended.

56. SMART Do you agree with the conclusion of the executive committee? Questions:

57. SMART Do you agree with the DSMB's recommendation to: (a) stop enrolment? (b) consider changing the therapy for patients in the DC arm? Questions:

58. SMART Given that the relative risks between the 2 arms change over time, how should patients who are 1 year into the study be managed? At 2 years? 3 years? Questions:

59. SMART Can the original question (long term comparison) still be answered by, say, re- designing the study? If so, how? Any suggestions? Questions:

60. SMART SMART Baseline CVD Risk factors (to Nov 2005) Current smoker 41.8 % Diabetes mellitus 7.6 % Prior MI 1.8 % Prior stroke 1.6 % Peripheral vascular disease 1.7 %

61. SMART SMART Baseline CVD Risk (cont.) Major ECG abnormalities7.9 % Coronary artery disease 2.8 % Congestive heart failure 0.9 % Antihypertensive drugs 20.3 % Cholesterol >= 240 or lipid lowering drugs 23.3 %

62. SMART NIAID Stops Intermittent HAART Trial Reuters Health Information 2006. © 2006 Reuters Ltd. NEW YORK (Reuters Health) Jan 18 - The National Institute of Allergy and Infectious Diseases has halted enrollment in a large international trial comparing continuous highly active antiretroviral therapy (HAART) with intermittent therapy, guided by CD4+ cell counts. The trial, known as Strategies for Management of Antiretroviral therapy (SMART), quickly showed that patients do better on continuous HAART, according to a statement released by the NIAID on Wednesday. Patients in the intermittent arm had twice the risk of dying or progressing to AIDS. "Furthermore, there was an increase in major complications such as cardiovascular, kidney and liver diseases in the participants on the drug conservation arm," the statement continued. "These complications have been associated with (HAART), and it was hoped that they would be seen less frequently in those patients receiving less drug," it added.

63. SMART NIAID Stops Intermittent HAART Trial (Continued) "We were surprised to learn that in the short-term, episodic antiretroviral therapy carries such an increased risk without evidence of sparing patients the known side effects associated with ART," said co- investigator Dr. Wafaa El-Sadr of the Harlem Hospital Center and Columbia University in New York. The patients in the drug-sparing arm stopped treatment when CD4+ counts reached 350 cells per microliter and resumed treatment when CD4+ counts dropped below 250 cells per microliter. When the trial was stopped, it had enrolled 5472 of a target of 6000 patients at 318 clinical sites in 33 countries. The trial was halted January 11, 2006 after an average follow-up of 15 months.