"We saw that those patients who were on a beta blocker had a higher incidence of death and myocardial infarction (heart attack)," Itani said in a telephone interview.

Over the month after surgery, 2.94 per cent of the beta-blocker group had a heart attack compared with 0.74 per cent who did not get the drugs. And 2.52 per cent of people in the beta blocker group died, compared with 0.25 per cent not taking the drugs.

And even though the patients who died were on beta-blockers, they had fast pulses that suggested the drugs were not having the intended effect.

He said his study suggests beta blockers should not be given routinely before surgery. But for patients with heart problems who need the drugs, he said, surgeons need to carefully monitor their heart rate before surgery to ensure it is well controlled.

Dr. Jeffrey Peters, chairman of the surgery department at the University of Rochester in New York who was not part of the study, said the evidence is enough to give him pause.

"We need to take a step or two back," he said.