Obese patients typically show poor sensitivity to the hormone insulin, a problem that can lead to diabetes. Previous reports have shown that weight loss surgery can improve this sensitivity and now, new research indicates that the type of surgery determines how fast this problem is fixed.

Restrictive weight loss surgery, such as gastric bypass, involves limiting the stomach size, ultimately causes the patient to eat less. In contrast, with malabsorptive operations, such as biliopancreatic diversion, the intestines are rearranged in such a way that the patient absorbs less of what is eaten.

The new findings indicate that with restrictive operations, the improvement in insulin sensitivity depends on the degree of weight loss. With malabsorptive operations, however, normal insulin sensitivity is restored more rapidly and is not dependent on the amount of weight loss.

The findings, are based on a study of 18 patients with severe obesity and 20 lean comparison subjects who underwent insulin sensitivity testing. The obese patients were treated with either gastric bypass or biliopancreatic diversion and then reevaluated 5 to 6 months and 16 to 24 months postoperatively.

With each procedure, weight loss averaged more than 100 pounds and occurred over roughly the same time course. However, as noted, the operations differed in their effects on insulin sensitivity.

During follow-up, a steady improvement in insulin sensitivity was seen in the gastric bypass group, but even at 16 months the sensitivity scores were still much lower than those of lean subjects.

In the biliopancreatic diversion group, by contrast, insulin sensitivity normalized by 6 months and actually exceeded that of lean subjects at 24-month follow-up, even though the subjects were still obese.

While the results suggest that biliopancreatic diversion restores insulin sensitivity more rapidly than gastric bypass, the authors note that each procedure has pros and cons and that doctors should decide which to use for a particular patient on an individual basis.