During pregnancy, some women develop a form of diabetes that ends after childbirth. Little research has been done to determine whether women who become diabetic during pregnancy had a pre-existing sensitivity to sugar.

Yes. "Gestational" suggests an intimate connection to pregnancy, but the connection is not well investigated. In no published report on gestational diabetes has glucose tolerance been tested before pregnancy. And only rarely have women found to be diabetic during pregnancy been retested after delivery to see whether their glucose intolerance persisted.

We studied a random sample of 817 women, 20 to 44 years old, who weren't pregnant and who had no history of diabetes. Nearly 4 percent met criteria for gestational diabetes. Below age 35, the prevalence was 2 to 3 percent, similar to that found in studies of pregnant women.

Risk factors for non-insulin-dependent diabetes, including parental history and obesity, were more prevalent among women meeting these criteria. For pregnant women, findings have been similar. These similarities raise the hypothesis that the reasons for gestational diabetes may precede pregnancy. Finding the true cause is important for the sake of both the patients and their developing babies.

It's important for fetal morbidity and mortality because congenital anomalies are believed to arise primarily in the first trimester, when the pregnancy is often unknown.

It's also important to a better understanding of how diabetes develops, because if gestational diabetes is simply a form of impaired glucose tolerance detected in pregnancy, then studies of pregnant women should shed light on the natural history of impaired glucose tolerance and its role as a risk factor for diabetes. Without such knowledge, the term "gestational diabetes" is inappropriate. Maureen I. Harris, MD Director, National Diabetes Data Group, National Institutes of Health

No. Since we look for gestational diabetes when the pregnancy is two thirds completed and its physiological impact on glucose metabolism is most pronounced, we are of course picking up some cases of preexisting glucose intolerance. But that's not the whole story.

When women are tested serially, some who are entirely normal early in pregnancy progress to gestational diabetes before they give birth. Pregnancy's effects on insulin requirements are well known and easily demonstrated, as is the body's response to insulin. If a woman's capacity to increase her insulin secretion is limited, that will become more and more manifest as pregnancy continues. That's why serial tests don't find all cases early.

Women who've had gestation diabetes in one pregnancy may have a normal test early in a subsequent one. But they remain at risk. If patients had glucose intolerance before pregnancy, they'd be expected to test abnormally after it. Some do, but many do not. Only further study will tell how many have a preexisting condition. This would require testing a large population of women before their pregnancies and then following them with serial tests.

If Dr. Harris's view were fully supported, it would change medical practice. If a solid case could be made that most gestational diabetes represents a preexisting condition, we would have to test glucose levels early and not concentrate on the last trimester, as we do now. But present evidence indicates that our greatest testing yield comes late in pregnancy. -- Boyd Metzger, MD Center for Endocrinology, Metabolism and Nutrition, Northwestern University

1990, Physician's Weekly, a Whittle Communications Publication; reprinted with permission