Are high levels of "good cholesterol" always good for you? A new study suggests perhaps not if you have diabetes.

HDL (high-density cholesterol) is supposed to be the "good cholesterol," or the "healthy cholesterol." It removes excess cholesterol from the blood and brings it back to the liver to be recycled.

HDL, together with estrogen, also helps stimulate the production of nitric oxide, which can relax blood vessels and lower blood pressure. HDL also has other beneficial functions as well.

It's the LDL (low-density cholesterol) that is the "bad cholesterol," or "lethal cholesterol." That's the stuff that forms plaques on your arteries that can clog them or, even worse, break off and cause heart attacks and strokes.

But now a study has shown that in "diabetics," not only does HDL fail to protect arteries by stimulating the production of nitric oxide, but it also seems to prevent arterial dilation by nitric oxide when other artery-dilating compounds are given.

The HDL from "diabetics" seems to have this effect because it associates with a 14-carbon saturated fatty acid called myristic acid. The association causes the deleterious effects.

Does this research, performed by Eric J Smart and colleagues at the University of Kentucky and published online the American Journal of Physiology Cell Physiology, mean that people with diabetes should be concerned about having high HDL levels?

The answer is that no one knows for sure yet. One problem is that the "diabetic" patients used in the study were not sufficiently described. The paper did not indicate if they were type 1 or type 2, nor how well controlled they were, as there was no mention of A1cs. It did not say how many patients were tested, nor whether the researchers tested a lot of individual samples and got the same results with all of them, or if they pooled the blood samples and tested the pooled blood.

The patients' fasting levels were 211 plus or minus 28 mg/dL, which suggests that these patients would have had high A1cs. Fasting triglyceride levels were 521 plus or minus 92 mg/dL, which is very high. Triglyceride goals are less than 150. Fasting insulin levels were three times "normal" serum levels, which suggests type 2.

The researchers used a OneTouch Ultra meter to measure blood glucose (BG) levels in the mice who were tested along with the humans. They don't say how they measured BG levels in the humans.

Hence it's difficult to know what this study really means for human diabetes patients. The researchers did a lot of controls, and the effects seem to be consistent: in both the diabetic mice and diabetic human blood samples, the HDL did not have the same beneficial effects as it did in nondiabetic blood samples.

But was this because the diabetic humans and mice had high BG levels? Or was it some metabolic defect that occurs when you have diabetes, regardless of the level of your control?