As my hypertension story of the year for PracticeUpdate Cardiology 2020, I felt I had to address a prior issue by discussing the HYGIA trial again. HYGIA was my story of the year for 2019, as I felt then it had potential to be a practice-changing study.1
To review, the report in the European Heart Journal (EHJ) stated that HYGIA was a trial randomizing 19,084 hypertensive patients to ingest their entire daily dose of all their hypertension medication at bedtime (n = 9552) or all of them upon awakening (n = 9532).2 As background, there are compelling data to indicate that nocturnal hypertension is more of a risk factor for CVD than daytime hypertension; so, it stood to reason that dosing of blood pressure meds at bedtime was worth studying. The results were strongly in favor of nighttime dosing of blood pressure medication. As I stated last year, “Simply put, the results of this trial are flabbergasting at first glance. The results reported almost seem too good to be true; those randomized to bedtime BP dosing had a 45% relative reduction in the primary CVD outcome.”
Having first approached the report with skepticism (asking myself why this wasn’t published in NEJM or Lancet or JAMA?), after reading the full paper, I nonetheless felt that the findings should be taken in good faith both because they were published in what was, at the time, the world’s leading cardiology journal (EHJ) and because I felt both the authors and editorial process deserved to be trusted.
However, in the wake of HYGIA—as well two other prominent blood pressure studies retracted from NEJM this year3,4—I now understand that all published reports must be scrutinized by the reader closely and one cannot assume that peer reviewers or editors have comprehensively screened for and culled potentially fraudulent findings. Perhaps this is already your own position; but, personally, up to now, I tended to trust papers in high-impact journals. No more.
To be clear, HYGIA remains under investigation and so we cannot yet know for sure that there was fraud. However, all signals point to inconsistencies in the data, and the EHJ editors themselves have published a notice of concern in which they state they, “recommend to interpret the major results and conclusions with caution until further notice”. All of the concerns are too numerous for me to go into much detail here and, for interested readers, are outlined in the referenced articles.5,6,7 The main concerns are: 1) HYGIA may not have even been a randomized trial at all; 2) the protocol on ClinicalTrials.gov contains numerous problems (including 15 primary endpoints!); 3) there was no adjudication of endpoints and no reporting on how the study was monitored; 4) apparently only 607 of over 19,000 participants dropped out over 6 years due to missing ABPM data (which is very implausible); and 5) there was a 40% reduction in non-CVD death (also highly implausible).
In conclusion, HYGIA reminds us to be ever vigilant for indicators of the dark side of clinical research. In 2019, my stance was that investigators should always be considered innocent until proven guilty. In legal terms, this is the presumption of innocence approach. Although I do not think by any means that we need to yet universally apply a presumption of guilt approach to clinical research, after 2020, I now realize that we have to scrutinize everything ourselves and cannot assume the editors or peer reviewers will pick up major problems in the data.