STATISTICS HELPS TO SYSTEMATICALLY SYNTHESIZE EVIDENCE MetaAnalysis is

STATISTICS HELPS TO SYSTEMATICALLY SYNTHESIZE EVIDENCE Meta-Analysis is Informative! Joseph C. Cappelleri, Kelly H. Zou, and Martin O. Carlsson Pfizer Inc September 26, 2012 11/2/2020 Pfizer Confidential Do No Circulate 1

Outline � � What is Meta Analysis? Some History Cumulative Meta Analysis Two Examples � Example 1: Viral Load in AIDS � Example 2: Anti platelet Use Post MI � � 11/2/2020 Guidelines Tribute to Professor Ingram Olkin Conclusions References Pfizer Confidential Do No Circulate 2

What is Meta-Analysis? � � � “Meta analysis” is a statistical analysis of data from multiple studies A systematic literature review encompasses an explicit and detailed description of how a review was conducted A meta analysis is often accompanied by a systematic literature review (also referred to as a quantitative systematic review or overview) in order to identify data systematically, summarize results, and to valuate quantitatively sources of heterogeneity and bias Fig 1. Steps for Meta Analysis Cappelleri et al. (2010, Encycl. Biopharm. Stat. ) 11/2/2020 Pfizer Confidential Do No Circulate 3

Some History � The first meta analysis was performed by Simpson & Pearson (1904, BMJ). � Analyzed data comparing infection and mortality among soldiers who had volunteered for inoculation against typhoid fever in various places across the British Empire with that of other soldiers who had not volunteered � The first meta analysis of all conceptually identical experiments was found in Pratt et al. (1940/1966; Bruce Humphries). � Reviewed 145 reports on extrasensory perception experiments published from 1882 to 1939, and included an estimate of the influence of unpublished papers on the overall effect (the ‘file drawer’ problem) � Meta analysis is fundamental to evidence based medicine (EBM) [Sackett et al. (1996, BMJ; 2003, BMJ Publishing Group; 2007, Clin Orthop Relat Res) ] which integrates the best available evidence with clinical expertise and patient values and, by extension, comparative evidence research Pfizer Confidential Do No Circulate 4 (CER) which seeks to compare all available treatment options using a

Some History (Continued) � Popular tools published in the 1950 s � Mantel Haenszel (1959, J Natl Cancer Inst) test for combining 2 × 2 tables � Inverse variance weighting approach of Woolf (1955, Ann Hum Genet) � Glass (1976, Ed Researcher) first coined the term "meta analysis“ � Sacks et al. (1987, NEJM) published a seminal article on meta analysis of randomized controlled trials 11/2/2020 Pfizer Confidential Do No Circulate 5

Cumulative Meta-Analysis � Cumulative meta analysis (CMA) is a method of updating previous meta analyses with the appearance of new studies. Lau et al. (1992, NEJM); Antman et al. (1992, JAMA) � Facilitates the determination of clinical efficacy and harm and may be helpful in tracking trials, planning future trials, and making clinical recommendation for therapy � Permits the study of trends in efficacy and makes it possible to determine when a new treatment appears to be significantly effective or deleterious � When performed routinely, the earliest time when statistical significance is reached (by whatever criterion is chosen) can be identified 11/2/2020 Pfizer Confidential Do No Circulate 6

Cumulative Meta-Analysis (Continued) � For example, thrombolytic drugs did not begin to be recommended for treatment of myocardial infarction, even for a specific indication, by more than half the experts until 13 years after they could have been shown to be effective in reducing mortality. Antman et al. (1992; JAMA) � Cumulative meta analysis indicated that intravenous streptokinase (a thrombolytic drug) could have been shown to be lifesaving long before its general adoption in practice and before submission to and approval by the Food and Drug Administration Lau et al. (1992; NEJM) 11/2/2020 Pfizer Confidential Do No Circulate 7

Two Examples � Two studies have made a profound impact on patient care and medical practice �Risk of HIV transmission according to viral load �The use of anti platelet in high risk subjects 11/2/2020 Pfizer Confidential Do No Circulate 8

Example 1: Viral Load in AIDS � Attia et al. (2009, AIDS) compiled all studies to date in heterosexual couples where one partner was HIV positive but where most were not on antiretroviral therapy (ART) � They confirmed that a high viral load can significantly increase the risk of transmission � A low viral load (which is possible to achieve in a minority of people who do not receive ART) significantly reduces the risk 11/2/2020 Pfizer Confidential Do No Circulate 9

Example 1: Viral Load in AIDS (Continued) � Out of 1000 HIV positive individuals with a viral load below 400 copies/ml regularly engaging in vaginal sex with an HIV negative partner, only one transmission could be expected to occur in the course of a year � In contrast, among 1000 HIV positive individuals with a viral load above 50, 000 copies/ml, at least 90 transmissions could be expected to occur in the course of a year 11/2/2020 Pfizer Confidential Do No Circulate 10

Example 2: Anti-platelet Use in High-risk Subjects � Antithrombotic Trialists’ Collaboration (2002; BMJ) examined randomized trials of an antiplatelet regimen vs. control, or of one antiplatelet regimen versus another, in high risk patients � � Patients had acute or previous vascular disease or some other predisposing condition 287 studies involving 135 000 patients in comparisons of antiplatelet therapy versus control and 77 000 patients in comparisons of different antiplatelet regimens. 11/2/2020 Pfizer Confidential Do No Circulate 11

Example 2: Anti-platelet Use in High-Risk Subjects (Continued) � Aspirin (or another oral antiplatelet drug) is protective in most types of patients at increased risk of occlusive vascular events, including those with an acute myocardial infarction or ischaemic stroke, unstable or stable angina, previous myocardial infarction, stroke or cerebral ischaemia, peripheral arterial disease, or atrial fibrillation � Low dose aspirin (75 150 mg daily) is an effective antiplatelet regimen for long term use, but in acute settings an initial loading dose of at least 150 mg aspirin may be required � Adding a second antiplatelet drug to aspirin may produce additional benefits in some clinical circumstances, but more research into this strategy is needed 11/2/2020 Pfizer Confidential Do No Circulate 12

Guidelines � Guidelines for conducting and reporting a proper meta analysis and systematic review are available � PRISMA Liberati et al. (2009; Ann Inter Med) � MOOSE Stroup et al. (2000; JAMA) � COCHRANE Higgins and Green (2008) � INDIVIDUAL DATA Riley et al. (2010; BMJ) 11/2/2020 Pfizer Confidential Do No Circulate 13

Tribute to Professor Ingram Olkin: Ambassador of Meta-Analysis � Research on meta analysis has been greatly influenced by the work of Ingram Olkin (Becker 2007; Statistical Science) � He has made seminal contributions in particular to multivariate meta analysis More than one (dependent) outcome is measured in a study � Example: Subjects measured on several outcomes or several time points � Example: A study involves multiple treatments with comparisons to a common treatment � � His textbook (coauthored with Larry Hedges) has been cited at least 3270 times! [Hedges and Olkin (1985); Becker (2007; Statistical Science)] Pfizer Confidential Do No Circulate 11/2/2020 14

Conclusions � The advent of meta analysis , accompanied by a systematic literature review, in the medical sciences has paralleled the explosion in the number of randomized trials being conducted � Meta analysis helps to synthesize and make sense of overwhelming amounts of data � Meta analysis is integral to providing the best available research synthesis and optimal patient care in the era of EBM and CER � A properly conducted meta analysis can save untold lives and improve the quality of life of countless patients 11/2/2020 Pfizer Confidential Do No Circulate 15

References � Antithrombotic Trialists' Collaboration. Collaborative meta analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002; 324: 71 86. (Erratum in: BMJ 2002 ; 324: 141. ) � Antman EM, Lau J, Kupelnick B, Mosteller F, Chalmers TC. A comparison of results of meta analyses of randomized control trials and recommendations of clinical experts. Treatments for myocardial infarction. JAMA. 1992; 268: 240 8. � Attia S, Egger M, Müller M, Zwahlen M, Low N. Sexual transmission of HIV according to viral load antiretroviral therapy: systematic review and meta analysis. AIDS. 2009; 23: 1397 404. � Becker BJ. Multivariate meta analysis: contributions of Ingram Olkin. Statistical Science. 2007; 22: 410 6. � Cappelleri JC, Ioannidis JPA, Lau, J. Meta Analysis of Therapeutic Trials. In: Chow S C (editor), Encyclopedia of Biopharmaceutical Statistics: 3 rd Edition, Revised and Expanded. New York, NY: Informa Healthcare; 2010: 768 779. Glass GV. Primary, secondary, and meta analysis of research. Ed Researcher. 11/2/2020 Pfizer Confidential Do No Circulate 1976; 5: 3 8. � 16

References (Continued) � Hedges LV, Olkin I. Statistical Methods for Meta-Analysis. Boston, MA: Academic Press, 1985. � Higgins JPT, Green S, Eds. Cochrane Handbook for Systematic Reviews of Interventions. The Cochrane Collaboration and John Wiley: Chichester, UK, 2008. � Lau J, Antman EM, Jimenez Silva J, Kupelnick B, Mosteller F, Chalmers TC. Cumulative meta analysis of therapeutic trials for myocardial infarction. NEJM. 1992; 327: 248 54. � Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D. The PRISMA statement for reporting systematic reviews and meta analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med. 2009; 151: W 65 94. � Mantel N, Haenzel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959; 22: 719 748. 11/2/2020 Pfizer Confidential Do No Circulate 17

References (Continued) � Pratt JG, Rhine JB, Smith BM, Stuart CE, Greenwood JA. Extrasensory Perception after Sixty Years. Boston: Bruce Humphries: 1940/1966. � Riley RD, Lambert PC, Abo Zaid G. Meta analysis of individual participant data: rationale, conduct, and reporting. BMJ. 2010; 340: c 221 � Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn't. BMJ. 1996; 312: 71 72. � Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn't. 1996. Clin Orthop Relat Res. 2007 Feb; 455: 3 5. � Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB. Evidence-Based Medicine: How to Practice and Teach EBM. London: BMJ Publishing Group, 2003. � Sacks HS, Berrier J, Reitman D, Ancona Berk VA, Chalmers TC. Meta analyses of randomized controlled trials. NEJM. 1987; 19: 450 5. 11/2/2020 Pfizer Confidential Do No Circulate 18

References (Continued) � Simpson R, Pearson K. Report On Certain Enteric Fever Inoculation Statistics. BMJ. 1904; 2: 1243 1246. � Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta analysis of observational studies in epidemiology: a proposal for reporting. Meta analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000; 283: 2008 2012. Woolf B. On estimating the relation between blood group and disease. Ann Hum Genet. 1955; 19: 251 3. 11/2/2020 Pfizer Confidential Do No Circulate 19 �