1. Mortality and Antithrombotics: Focus on FAERS Repository
Victor Serebruany, MD, PhD
Owner, HeartDrug™ Research LLC;
Johns Hopkins University;
Busan, December 10, 2016

2. Limitations of Clinical Trials
Small sample size;
Heavily selective population;
Lack of elderly, children, pregnancy, co-morbidities, impaired renal function;
Narrow indications;
Short duration, no chronic use;
Potential industry bias

3. FDA Postmarketing Surveillance
Drug Utilization data
External databases:
Passive Surveillance (FAERS)
Active
Surveillance
Background Incidence Rates

4. FAERS
FDA Adverse Event Reporting System (FAERS)
Voluntary, “spontaneous” reporting system
Sponsors required to report (21CFR314.80)
Computerized database
Origin 1969; > 7.5 million reports
Contains human drug and “therapeutic” biologic reports
Exception - vaccines (VAERS)

5. FAERS Strengths Open to public Includes all U.S. marketed products
Simple, inexpensive reporting system
Detection of events not seen in clinical trials (“signal generation”)
Especially good for events with rare background rate, short latency
Case series evaluation: identification of trends, drug indication, population, and other clinically significant emerging safety concerns
Open to public

6. FAERS Limitations Duplicate reporting; Extensive underreporting;
Quality of report is variable;
Reporting biases;
Difficult to attribute events with a high background rate, confounders, long latency outcomes;
Hard to mine effectively

7. FDA should monitor FAERS
Daily “in-box” review of reports
All serious unlabeled reports;
Serious directreports;
Periodic and “enhanced pharmacovigilance” reports
Periodic safety reports
Main mission: identify and monitor “Safety Signals”
Work with epidemiologists, and medical officers

8. Objective: Assess mortality with antithrombotics in FAERS
Clopidogrel vs. Prasugrel vs. Ticagrelor
NOACs vs. Warfarin
Dabigatran vs. Rivaroxiban vs. Apixaban vs. Edoxaban vs. Warfarin

9. Total fatalities co-reported with oral P2Y12 inhibitors in FAERS
Drug Cases (n) Deaths (n/%) Chi-square p-value PRR (95%-CI) ROR (95% - CI) _________________________________________________________________________________________ Clopidogrel 108,081 12,538; 11.6% ( ) ( ) Prasugrel 7, ; 8.4% < ( ) ( ) Ticagrelor 9, ; 12.4%

10. Annual 2015 deaths co-reported with oral P2Y12 inhibitors in FAERS
Drug Cases (n) Deaths (n/%) Chi-square p-value PRR (95%-CI) ROR (95% - CI) _________________________________________________________________________________________ Clopidogrel 13,234 1,156; 8.7% < ( ) ( ) Prasugrel 2, ; 5.2% < ( ) ( ) Ticagrelor 2, ; 14.7%

13. Death events co-reported with NOACs and warfarin in FAERS.
Drug Total Cases (n) Deaths (n;%) chi-square p-value PRR (95% CI)
____________________________________________________________________________
Dabigatran , ,989 (15.11%) e ( )
Rivaroxaban , ,318 (9.79%) e ( )
Apixaban , ,693 (9.52%) e ( )
Edoxaban (7.02%) e ( )
All NOACs , ,917 (11.63%) e ( )
Warfarin , ,493 (12.67%)

15. Conclusions:
Total and 2015 annual fatalities in FAERS are higher after ticagrelor than with clopidogrel, challenging PLATO trial results
Prasugrel is linked to remarkably less deaths than ticagrelor
NOACs appear to be associated with a mild mortality benefit over warfarin
Among NOACs FAERS reveals favorable trends for factor-Xa inhibitors (apixaban, rivaroxaban, and edoxaban), but unfavorable signal for the direct thrombin inhibitor (dabigatran)