A French study finding that people with Alzheimer's disease and diabetes have less memory loss than those without diabetes should be regarded with caution, American experts say.

"It's not clear from this study and others what the relationship is," said William Thies, chief medical and scientific officer of the Alzheimer's Association.

What is clear, is that having diabetes increases the risk of developing Alzheimer's disease -- a relationship acknowledged in the first sentence of a report on the French study. But there haven't been too many studies looking at the effect of diabetes on the progression of Alzheimer's, said the report by researchers at INSERM, the French National Institute for Health and Medical Research in Toulouse.

So, they followed 608 people with mild to moderate Alzheimer's disease for four years, testing their memory and thinking skills twice a year.

The 63 participants with diabetes -- 10.4 percent of the total -- started with the same average score of 20 on the test of thinking ability. There was an average overall decline of 1.24 points on that test every six months. But the decline in thinking ability scores was 0.38 points greater every six months for those without diabetes.

Previous studies of the relationship between diabetes and loss of cognitive function have had mixed results, Thies said. "Some showed a faster decline, some showed a slower decline," he said. "It is still a confused area."

One factor that has a strong influence on the results is the stage of the disease in the people being studied, Thies said. "You know that early on, the disease does not progress quickly," he said. "So, if you study people in the early stage or the late stage, that can overwhelm the results."

A close look showed that the diabetic participants in the French study had Alzheimer's disease for a shorter period of time than the non-diabetics, said Dr. Robert Friedland, chair of neurology at the University of Louisville, KY.

And the differences seen in the study "are very minimally significant, less than a point on a 30-point scale," Friedland said. "The difference in many of the mental state examination scores was very small. It was statistically significant, but clinically meaningless."

He ticked off several reasons why a difference was found. The medications taken for diabetes to help control blood sugar level could have a beneficial effect on the brain, Friedland said. "Also, patients with diabetes have more vascular disease in the heart as well as the retina, and some of their impairment might be due to that, so it was progressing more slowly," he said.There was also a possibility of misdiagnosis, since no autopsies were done in the study, Friedland said.

"The important message is that there are potentially modifiable risk factors for Alzheimer's disease, some of which are also risk factors for diabetes -- lack of physical activity, obesity," he said.