Researchers who compared two diabetes trials said on Friday they are getting some insight into why patients in one were more likely to die after aggressive treatment, while patients in another were not.

Weight gain, the use of multiple drug combinations and perhaps even getting blood sugar too low, too fast may all have been factors in causing early deaths, experts said.

The trials, one in the United States and Canada and the other in Australia, were both designed to measure the benefits of using drugs to dramatically lower blood sugar in diabetes.

But the American trial, called Accord, was halted in February after researchers found patients were more likely to die with more treatment. Results from the Australian trial, called Advance, showed the treatment reduced the risk of kidney disease by one-fifth without increasing the risk of death.

"These findings reinforce that blood glucose lowering in diabetes is safe and has an important role to play in the prevention of serious complications," said Dr. Anushka Patel of the University of Sydney, who presented findings from the study at the American Diabetes Association meeting in San Francisco.

Details of both trials were also released online in the New England Journal of Medicine on Friday.

More Deaths

Five percent of patients given intensive drug treatment in Accord died of any cause during the 3.5 years before the trial was stopped, compared to 4 percent of patients given standard treatment, said Dr. Hertzel Gerstein of McMaster University in Ontario and colleagues.

And cardiovascular deaths were higher among those who got more treatment -- 2.6 percent versus 1.8 percent. This translated to 41 excess deaths among the heavily treated group. There is no clear explanation, the Accord team said.

Before the Accord study, doctors had assumed that lowering blood glucose to the normal range would protect patients from the complications associated with diabetes, such as heart attacks, kidney disease, nerve damage and blindness.

For now, several experts agreed, the goal should be to get average blood sugar levels as measured by hemoglobin A1c to 7 and no lower.

The Advance study involved 11,140 patients with hemoglobin A1c of 7.5 percent at the start of the study.

One group got aggressive treatment to reduce their hemoglobin A1c to 6.5 percent or below, far below the current target in the United States of under 7 percent. The other group got standard care.

After an average of 5 years, the group that got aggressive blood sugar control had a 10 percent overall reduced risk of serious diabetes complications and a 21 percent reduced risk of kidney disease, a common complication of diabetes.

Too Low?

The Accord trial involved 10,000 patients whose hemoglobin A1c was 8 on average, and the goal was to get it down to 6.

In Accord, patients were more likely to have severe low blood sugar crises that required medical help, noted Dr. Robert Dluhy and Dr. Graham McMahon, both editors at the New England Journal of Medicine.

There were some other key differences between the Advance and Accord studies.

Patients in the intensive arm of Advance were treated primarily with moderate-release gliclazide, a drug in an older class known as sulfonylureas.

Dluhy and McMahon noted that few patients in Advance got drugs in a new class called thiazolidinediones or glitazones. But rosiglitazone -- GlaxoSmithKline Plc's Avandia -- was given to 90 percent of the intensive therapy group in Accord.

Within 4 months, A1c levels fell from 8.1 percent to 6.7 percent in Accord.

Patients who got their levels lower took a greater variety of drugs from several classes. They also had their prescriptions changed more often -- four times a year on average in the intense treatment group compared to twice a year for the standard therapy group.

The people treated more intensively in Accord gained more weight -- 3.5 kg (7.7 pounds) compared to 0.4 kg (0.9 pound), noted Dr. William Cefalu of Pennington Biomedical Research Center at Louisiana State University in Baton Rouge.

"The weight gain is not a trivial point," he wrote in a commentary -- noting that in neither study were patients encouraged to eat a healthier diet or exercise, both proven to help lower the risks and symptoms of both diabetes and heart disease.