Avoidable harm to patients
Recommended treatments for heart attacks that had appeared in textbooks published over a period of 30 years were compared with evidence that could have been taken into account had the authors systematically reviewed the results of fair tests of treatment reported during that time. 
Science is cumulative
but scientists don’t accumulate evidence scientifically
The impact of this was devastating. In some cases, patients with heart attacks were being deprived of life-saving therapies (for example, clotbusting drugs). In other cases, doctors continued to recommend treatments long after fair tests had shown they were lethal (for example, the use of drugs that reduce heart rhythm abnormalities in patients having heart attacks).
The failure to combine the results of studies in systematic reviews as new evidence becomes available continues to harm patients.
Blood substitutes that need no refrigeration or crossmatching are an obviously attractive alternative to real blood for the treatment of haemorrhage. Unfortunately these products increase the risk of heart attacks and death. Furthermore, a systematic review of the randomized trials reported since the late 1990s reveals that their dangers could and should have been recognized several years earlier than they were.