For many years, diabetes experts assumed that the closer they could get a diabetic patient's blood sugar level to what is considered normal, the better.

But recent studies now suggest that a more tailored treatment approach, taking into account age and other health problems, may be a better way to go to control heart disease.

"How sick a patient is matters and how long they've had diabetes matters," said Elbert Huang, a clinical and health care policy researcher at the University of Chicago. "Those are new ideas that no one really ever suspected," Mr Huang said in a telephone interview.

Aggressive treatment strategies that use drugs, diet and exercise to dramatically reduce blood sugar levels have been shown to reduce the risks of diabetic complications such as blindness and kidney disease.

But recent studies looking at whether intensive treatment helps prevent the most deadly complications of diabetes - heart attacks and strokes - have been mixed.

A federally sponsored trial called Accord was stopped in February 2008 because there were 20 per cent more deaths among diabetics with heart problems who got intensive treatment compared to those who were treated more conservatively.

An international trial of aggressive measures to control blood sugar, however, found no signs that such treatment increases the risk of death.

Since then, several teams have been working to make sense of the findings.

Last spring, researchers said a trial of older veterans with diabetes found that people who started intensive treatment within the first 15 years of diagnosis had fewer heart risks.

But in people who started intensive treatment more than 20 years after their diabetes diagnosis, the heart risks doubled.

This week, a study of diabetics in Italy published in the Annals of Internal Medicine offers some clues about why. It found that strictly controlling blood glucose levels in patients with other serious health problems did not reduce their risk of a heart attack or stroke.

But they did find that firm glucose control - defined as keeping hemoglobin A1c, a measure of long-term blood sugar control, levels below 7 percent - did cut heart risks in diabetics with few other health concerns.

"It turns out the reason some studies have been positive and some have been negative is that there are at least two subgroups in here," said Sheldon Greenfield of the University of California Irvine, one of the study's authors.

"It looks like patients who have a lot of other medical conditions just do not benefit very much, if at all, from tight control, whereas patients who are younger and have little of what we call comorbidity, appear to benefit," Dr Greenfield said in a telephone interview.

He said the findings suggest that in some cases, doctors should aim for just moderate blood glucose levels - a hemoglobin A1c of between seven and eight - and treat the coexisting conditions more aggressively.

The findings "raise serious questions about guidelines advocating a single approach for all diabetics," said Sherrie Kaplan of the University of California Irvine's School of Medicine, who worked on the study.