to the editor: When should a pregnant woman who has diabetes receive insulin? Rightly or wrongly, the administration of insulin in pregnant women is perceived as increasing the risk to the mother and fetus enough to necessitate referral from the family practice, which often results in increasingly aggressive management. A review of bylaws at four major hospitals indicated that the administration of insulin required the involvement of an obstetrician documented in the chart before delivery. This consultation has become increasingly difficult to obtain because obstetricians often avoid involvement with high-risk cases other than their own.

Examples of aggressive management have included recommendations for insulin at two-hour postprandial glucose levels of 120 mg per dL (6.7 mmol per L). In Tennessee, recommendations vary, but few consultants have recommended the administration of insulin at these levels. An article¹ in American Family Physician on gestational diabetes cited limited studies claiming to reduce the frequency of macrosomia, which is an indirect measure of a rare event (such as shoulder dystocia or brachial plexus injury with permanent paralysis of an arm). This leads to the heuristic fallacy of managing each case with the strategy of catastrophic expectations. This advice is not compatible with community medical practice, which usually does not require insulin for the successful management of gestational diabetes mellitus.

Several patients were referred to perinatology for prescription of insulin to achieve “tight control” of blood glucose levels. Follow-up revealed that the consultants managed these cases with diet and surveillance similar to what previously had been used in the family physicians’ office. The source data² were reviewed starting with the Cochrane database.

We are concerned that the following statement from the article¹ may be misinterpreted as a standard of care: “Therefore, insulin therapy traditionally has been started when capillary blood glucose levels exceed 105 mg per dL (5.8 mmol per L) in the fasting state and 120 mg per dL…two hours after meals. These cutoff values are derived from guidelines for managing insulin in pregnant women who have type 1 diabetes.”¹ The authors¹ suggest that insulin might be started at even lower levels. While this may lower the blood glucose level, it is premature to mandate insulin on the basis of indirect measures taken from a different disease. Gestational diabetes is not the equivalent of ketosis-prone type 1 diabetes or women who were receiving insulin before becoming pregnant. The more stable condition of gestational diabetes is appropriately managed in a more conservative fashion with a latitude of practice styles.³

In our community and many others, the prescription of insulin raises the patient up into a risk category that literally prohibits continuing care by the family physician. For our uninsured patients, premature referral has been a financial disaster. In the hands of the family physician, close surveillance with ultrasonography in the third trimester appears to work just as well in avoiding the complications of macrosomia. Other maternal and fetal morbidities are extremely rare, and insulin does not prevent them. A guideline should not be elevated to a standard before its time.

in reply: I appreciate the comments from Dr. Rodney and colleagues regarding my article¹ on gestational diabetes mellitus in American Family Physician. I guarantee you that the authors of our article¹ understand the nature of the difficulties you face in obtaining obstetric consultation when working with an underserved population. The authors¹ have decades of experience providing obstetric care to underserved populations. It was certainly not our intention to refer a greater number of patients for consultation.

A central theme of the article¹ is that much of the “routine care” for patients with gestational diabetes mellitus is based on imperfect and possibly misapplied data. This article represents one groups’ view on this subject, and it is possible to find many other sources that vary in their opinions. As stated in the article,¹ this is a controversial issue. The lack of agreement among many expert bodies reflects this.

Given the wide variance in practice, there appear to be several options for family physicians:

(1) Don’t screen for gestational diabetes mellitus. The U.S. Preventive Services Task Force assigns screening an “I” rating, meaning that the evidence is insufficient to recommend for or against. The argument is that this screening test does not have data to demonstrate a reduction in adverse outcomes.

(2) Develop a community standard for the initiation of insulin in patients with gestational diabetes mellitus. There is some leeway here, and the physician may agree with the obstetricians in your community about who actually needs insulin. This may provide an option for those obstetricians who may not want to consult on your patients.

(3) Use glyburide instead of insulin. Do the hospital protocols call for referral with glyburide? The American College of Obstetricians and Gynecologists do not recommended the use of glyburide, but it is used widely in clinical practice.

As physicians who provide care to the underserved, we certainly appreciate the issues raised by Dr. Rodney and colleagues.