No field better illustrates the triumphs and the problems of evidence-based medicine than heart failure. In the early 1980s, investigators sought to reduce the high mortality of systolic heart failure by using a new drug class, the angiotensin converting enzyme (ACE) inhibitors. They recruited mostly male patients with a reduced systolic ejection fraction measured by angiography or radionuclide ventriculography. Sure enough, ACE inhibitors reduced mortality in this group, and many trials followed, comparing different ACE inhibitors and different dosages. Then the angiotensin II receptor blockers (ARBs) arrived, leading to a clutch of similar studies in a similar male-dominated population of younger patients with systolic heart failure due to infarction and ischaemia. A systematic review (Int J Clin Pract 2008;62:1397–402) looks at the trials of ARBs in heart failure, either in head-on comparisons with ACE inhibitors or in combination with them. The most striking thing is how small the effect sizes are and how much the confidence intervals overlap.

Turning to the real-life population of patients we see with clinical heart failure, the situation …