The health benefits of moderate alcohol use may have been overestimated, according to new research.

Numerous previous studies have suggested that light to moderate drinking may help reduce the likelihood of cardiovascular diseases and all-cause mortality, but a recent study published in the British Medical Journal described their findings as "contentious" and possibly the result of a combination of selection bias, poor controls and a tendency to overlook older participants.

Researchers including Craig Knott of University College, London, used data from the Health Survey for England (1998-2008) to assess 18,368 adults aged 50-64 years and 34,523 adults over the age of 65.

The only significant mortality benefits detected were among men aged 50-64 who consumed 15-20 units of alcohol a week, or up to 1.5 units on the heaviest day, and women aged 65 and over who drank less than 10 units a week, or up to 4.5 units on the heaviest day.

The researchers say their results "may have better isolated the true effect of alcohol consumption on mortality".