Patients using Avandia, the brand name given by GlaxoSmithKline to rosiglitazone, ran a 45% higher risk of a heart attack than those using a placebo or alternative treatment, according to the results of the study published in a leading medical journal

The US regulator the food and drug administration (FDA) responded by issuing a formal alert to doctors. And congressional leaders in Washington announced that the House oversight committee would hold hearings on GSK's handling of the drug next month.

GSK, the UK's leading pharmaceutical company, disputed the study's findings, arguing it was based on "incomplete evidence and a methodology that the author admits has significant limitations".

Avandia has been taken by more than 6 million people worldwide since it was licensed eight years ago. The drug is one of a group that lower blood glucose levels for people with type 2 diabetes - the late onset form of the disease which is linked to obesity. The FDA warned doctors to be aware of the study published online by the New England Journal of Medicine linking it to heart disease.

The original trials, which led to the drug being approved, did not look at its effect on the heart, say the study's authors, Steven Nissen and Kathy Wolski from the Cleveland clinic in Ohio. Dr Nissen told the Guardian: "I would have preferred it if, right after the drug was launched they'd done a large-scale cardiovascular outcome study. Eight years on, unfortunately we still haven't had a complete one."

In an attempt to establish whether Avandia could cause heart problems, the authors pulled together results of 42 studies into the drug lasting more than 24 weeks.

GSK said it "strongly disagrees" with the findings. The review was not the best way to establish whether a drug has side-effects, it said, because it pulled together information from many trials designed in different ways. The best way to be certain is to run a large, long-term clinical trial looking at risks and benefits in patients. Several of these are in progress and some have been published. One found a small increased heart risk; others found none.