Sum and substance in the Jikei Heart Study

1The addition of the angiotensin-II type-1 receptor antagonist (ARB), valsartan, to conventional therapy in patients with hypertension, coronary heart disease, or heart failure reduced the primary composite endpoint by 39% (95% CI 21–53%). We know how diffi cult it is to do a randomised trial in Japan. The Jikei investigators are therefore to be congratulated for this trial in more than 3000 patients. Nevertheless, one should not accept at face value the main conclusions of the Jikei report. The investigators suggest that addition of valsartan to conventional therapy prevented more cardiovascular complications than added non-ARB therapy and that, in line with previous publications of one of the co-authors, 2 the benefi t of the ARB could not entirely be explained by diff erences in blood pressure. The major weakness of the Jikei Heart Study stems from the compromises the investigators had to accept to make their trial feasible. The sample size and incidence of “hard” events were lower than in many other trials. 3–5 In terms of standardised daily doses, the added non-ARB therapy (the comparator) did not catch up with the addition of valsartan until after the second year of follow-up. Furthermore, the Jikei Heart Study had a prospective randomised open-blinded-label endpoint (PROBE) 6 design. Although an independent and blinded endpoint committee adjudicated endpoints, PROBE does not protect against possible bias of the investigators in reporting events or admitting patients to the hospital. New admissions for angina pectoris (relative risk reduction 65%, 95% CI 42–80%) and heart failure (47%, 6–69%) were the main drivers of the benefi t in the primary composite endpoint in the Jikei study. We fi nd it diffi cult to understand how in an open trial, in which investigators were fully aware of the treatment being administered, admission could be part of the primary outcome measure. The 77 cerebrovascular events also included nine cases of transient ischaemic attacks, a “weak” event not usually