Advertisement

Excess Ticagrelor Mortality in the Food and Drug Administration Adverse Event Reporting System: Time to Recount PLATO Trial Deaths

Published:February 01, 2017DOI:https://doi.org/10.1016/j.amjmed.2016.12.037
      Despite broad utilization of oral P2Y12 platelet inhibitors such as clopidogrel, prasugrel, and ticagrelor, the comparative mortality risks of each agent are unknown. Moreover, few available randomized data are challenged with restricted populations, double-digit drug discontinuations, and incomplete follow-up.
      • Marciniak T.A.
      • Cherepanov V.
      • Golukhova E.
      • Kim M.H.
      • Serebruany V.
      Drug discontinuation and follow-up rates in oral antithrombotic trials.
      Importantly, one agent, ticagrelor, is consistently advertised as providing a mortality benefit based solely upon the controversial PLATO trial.
      • Center for Drug Evaluation and Research
      Because such a benefit has never been confirmed by any later ticagrelor trials, alternative data on comparative mortality risks linked to antiplatelet agents are crucial to inform prescribing physicians. We assessed total and 2015 annual death rates for clopidogrel, prasugrel, and ticagrelor, from the US Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) repository. Importantly, these data are public, and reporting death is mandatory (21CFR314.80) for FAERS.
      • Duggirala H.J.
      • Tonning J.M.
      • Smith E.
      • et al.
      Use of data mining at the Food and Drug Administration.
      From 7,732,656 screened FAERS cases, we excluded 7,607,053 and qualified 125,503 events, including 14,396 deaths. The distribution of deaths reported in entire FAERS, and 2015 annual fatalities for 3 antiplatelet agents are presented in the Table.
      TableTotal and 2015 Deaths Co-Reported with Oral P2Y12 Inhibitors in FAERS
      DrugEventsCases (n)Deaths (n)Deaths (%)χ2P-Value
      Compared with ticagrelor.
      PRR (95% CI)
      Compared with ticagrelor.
      ROR (95% CI)
      Compared with ticagrelor.
      ClopidogrelTotal108,08112,53811.65.59.0180.9352 (0.8851-0.988)0.9267 (0.8703-0.9866)
      2015 only13,23411568.786.331.52e−200.5961 (0.5352-0.664)0.5575 (0.4924-0.6312)
      PrasugrelTotal75626358.471.352.99e−170.6776 (0.6186-0.7422)0.648 (0.5858-0.7168)
      2015 only29271515.293.491.13e−320.4246 (0.3546-0.5084)0.3864 (0.3173-0.4707)
      TicagrelorTotal9860122212.4RR1.0001.000
      2015 only260738214.7
      CI = confidence interval; FAERS = US Food and Drug Administration Adverse Event Reporting System; PRR = proportional reporting ratio; R = reference; ROR = reporting odds ratio; χ2 = chi-squared.
      Compared with ticagrelor.
      Data from this large, uniform, US government-run international repository suggest a consistent disproportional excess of mortality associated with ticagrelor when compared with clopidogrel, and especially with prasugrel. We deliberately focused our analyses exclusively on mortality for 3 reasons. First, death is unquestionably the most important outcome measure validating the efficacy of a drug for the prevention of secondary vascular events. Second, the FAERS database analyses are always challenged by the often-uneven mixture of patients and reports, because any single event can generate multiple records. In contrast, death is a final outcome avoiding numerator confusion and repeated counts. Third, despite the heavily debated FDA decision to approve ticagrelor in 2011,
      • Center for Drug Evaluation and Research
      • Serebruany V.L.
      Paradoxical excess mortality in the PLATO trial should be independently verified.
      • DiNicolantonio J.J.
      • Tomek A.
      Misrepresentation of vital status follow-up: challenging the integrity of the PLATO trial and the claimed mortality benefit of ticagrelor versus clopidogrel.
      the drug has been expanding by consistently claiming mortality benefit following myocardial infarction, although the death benefit is lacking in the drug “Indications and Usage” portion of the insert. It is concerning that this claim of mortality's superiority over clopidogrel by gaining broad cardiovascular indications is based exclusively on the results of a single controversial PLATO trial against clopidogrel.
      • Wallentin L.
      • Becker R.C.
      • Budaj A.
      • et al.
      Ticagrelor versus clopidogrel in patients with acute coronary syndromes.
      Importantly, these death advantages after ticagrelor were never confirmed thereafter, despite massive implementation of the PARTHENON program.
      There are a few important considerations to be yielded from the current data mining. First, the sample sizes of fatalities for all 3 antiplatelet drugs are sufficient to validate such analyses and detect significant differences. The double digits of χ2 estimates suggest the sufficiency and quality of the index dataset. Second, the quantity of FAERS reports for each drug is clearly a reflection of the current clinical utilization of oral P2Y12 antiplatelet agents, where clopidogrel is still dominant with about 75% of prescriptions, while prasugrel and ticagrelor equally share the remaining use.
      • Serebruany V.L.
      • Fortmann S.D.
      Underutilization of novel antiplatelet agents - myths, generics, and economics.
      This pattern is important, suggesting a strong match between distribution of current FAERS reports with the “real-life” clinical scenarios. Third, the FAERS management should be acknowledged for their diligence in eliminating brand names from the reports, making the repository easy to explore and analyze. Fourth, ticagrelor was under the FDA scope for a potential signal of serious atrioventricular block as reported in the FAERS database. As of June 30, 2016, the FDA was evaluating the need for regulatory action.
      • US Food and Drug Administration
      This particular review resulted in the inclusion of the new section (5.5) for bradyarrhythmias in ticagrelor insert. Fifth, the numbers matter. In the PLATO trial, ticagrelor was associated with 89 fewer vascular deaths and 107 fewer all-cause deaths than clopidogrel, allegedly representing <18% mortality reduction.
      • Wallentin L.
      • Becker R.C.
      • Budaj A.
      • et al.
      Ticagrelor versus clopidogrel in patients with acute coronary syndromes.
      In FAERS, however, the latest evidence for 2015 alone indicates that the rate of ticagrelor deaths was about 40% higher when compared with clopidogrel, and almost tripled when compared with prasugrel, providing strikingly significant P-values. For example, the difference in signal between total prasugrel and ticagrelor deaths is as much as 2.99e−17. This number means that the P value for significance holds 16 zeros after the comma, and the 17th-19th digits are 299. This is especially concerning, because the latest evidence failed to confirm PLATO “mortality benefit,” because even a favorable trend was lacking in all post-PLATO outcome-driven trials (PEGASUS, PHILO, SOCRATES, and EUCLID). Importantly, the manufacturers watch the FAERS data very carefully, sometimes delaying adverse event reporting.
      • Heavey S.
      FDA warns Pfizer for not reporting side effects. Reuters, June 10, 2010.
      There was also an over-9-month recent gap (now fixed) in ticagrelor FAERS reporting, while prasugrel and clopidogrel cases are updated regularly on a quarterly basis. Finally, there are no reason(s) to believe that FAERS reporting has been biased toward ticagrelor, hiding unreported death cases for prasugrel or clopidogrel. Indeed, some cases are missing from FAERS, but differences of fatal cases among three antiplatelet agents of death distribution will remain the same. Regardless, the consistency and magnitude of excess deaths after ticagrelor suggests that the mortality signal is real. It was recounted three times by Johns Hopkins researchers, former FDA Medical Team Leader, and independent professional FAERS biostatisticians. These data severely challenge the unconfirmed ticagrelor mortality benefit observed in the PLATO trial. The FDA should urgently review their own ticagrelor FAERS dataset, focusing on disproportional excess of mortality signal beyond the acknowledged atrioventricular blockade risks. We lately see the paradigm switch from conventional trials to “real world evidence.” The 21st Century Cures Act, which President Obama signed in December 2016,
      H.R.6-114th Congress (2015-2016): 21st Century Cures Act.
      mandates that the FDA develop a plan to use “real world evidence” in regulatory decisions. Ticagrelor should become the first example of the FDA doing so. Sometimes admitting a mistake and fixing the problem saves lives and reputation.

      Acknowledgment

      HeartDrug Research LLC (Wilmington, Del) supported FAERS data mining and statistical work. Special thanks to Paul Danese, PhD, for final event count.

      References

        • Marciniak T.A.
        • Cherepanov V.
        • Golukhova E.
        • Kim M.H.
        • Serebruany V.
        Drug discontinuation and follow-up rates in oral antithrombotic trials.
        JAMA Intern Med. 2016; 176: 257-259
        • Center for Drug Evaluation and Research
        The FDA ticagrelor review of complete response.
        (Available at:) (Accessed October 4th, 2016)
        • Department of Health and Human Services
        The FDA ticagrelor secondary review.
        (Available at:) (Accessed October 5th, 2016)
        • US Food and Drug Administration
        Adverse Event Reporting System (AERS).
        (Available at:) (Accessed October 4th, 2016)
        • Duggirala H.J.
        • Tonning J.M.
        • Smith E.
        • et al.
        Use of data mining at the Food and Drug Administration.
        J Am Med Inform Assoc. 2016; 23: 428-434
        • Serebruany V.L.
        Paradoxical excess mortality in the PLATO trial should be independently verified.
        Thromb Haemost. 2011; 105: 752-759
        • DiNicolantonio J.J.
        • Tomek A.
        Misrepresentation of vital status follow-up: challenging the integrity of the PLATO trial and the claimed mortality benefit of ticagrelor versus clopidogrel.
        Int J Cardiol. 2013; 169: 145-146
        • Wallentin L.
        • Becker R.C.
        • Budaj A.
        • et al.
        Ticagrelor versus clopidogrel in patients with acute coronary syndromes.
        N Engl J Med. 2009; 61: 1045-1057
        • Serebruany V.L.
        • Fortmann S.D.
        Underutilization of novel antiplatelet agents - myths, generics, and economics.
        Thromb Haemost. 2014; 112: 4-9
        • US Food and Drug Administration
        FDA Adverse Events Reporting System (FAERS) (January-March, 2016).
        (Available at:) (Accessed October 4th, 2016)
        • Heavey S.
        FDA warns Pfizer for not reporting side effects. Reuters, June 10, 2010.
        (Available at:) (Accessed October 19, 2016)
      1. H.R.6-114th Congress (2015-2016): 21st Century Cures Act.
        (Available at:) (Accessed November 25, 2016)

      Linked Article

      • Unreliable Observations from a Confounded Analysis of a Skewed Database
        The American Journal of MedicineVol. 130Issue 8
        • Preview
          Serebruany et al1 report on all-cause mortality rates in an observational US Food and Drug Administration (FDA) database, based on voluntary reporting of adverse events, and compare fatalities reported in patients treated with clopidogrel, prasugrel, and ticagrelor. Using this database, and examining unadjusted death rates with no attempt to adjust for any of the obvious major differences between groups likely to confound the comparisons made, they claim that patients receiving ticagrelor had a death rate 40% higher than those treated with clopidogrel and triple that of patients receiving prasugrel.
        • Full-Text
        • PDF