Gestational Diabetes: Risk or Myth?

Thomas A. Buchanan and Siri L. Kjos

University of Southern California School of Medicine, Los Angeles, California 90033

Address all correspondence and requests for reprints to: Dr. Buchanan, Room 6602, General Hospital, 1200 North State Street, Los Angeles, California 90033. E-mail: buchanan{at}hsc.usc.edu

There is a wide spectrum of views regarding the clinical significance of the entity known as gestational diabetes mellitus (GDM). At one end of the spectrum are advocates of strict evidence-based medical care who argue that there is insufficient evidence for preventable morbidity resulting from asymptomatic maternal hyperglycemia during pregnancy to warrant routine detection and treatment of GDM. At the other end are advocates of universally strict glycemic control during pregnancy. They argue that even minimal maternal hyperglycemia is potentially dangerous and worthy of detection and aggressive treatment. As is often the case when such wide ranging views abound, truth and prudence seem to lie somewhere between the extremes. There is evidence that asymptomatic hyperglycemia during pregnancy is attended by important morbidity in the mother and offspring; at least some of the morbidity can be reduced by detection and treatment of GDM. However, over the range of glycemia generally encountered in GDM, only a subset of women and their infants are at risk for morbidity related to maternal hyperglycemia. Ignoring all asymptomatic hyperglycemia during pregnancy will result in inattention to important morbidities in the population. Aggressive detection and treatment of even minimal hyperglycemia during pregnancy will lead to unnecessary overtreatment of many women. The logical solution to this dilemma is the development of methods to stratify patients with GDM, regarding perinatal and long-term health risks, so that treatments can be tested and applied in the highest risk cases. The emphasis of this commentary will be on measures that can be taken in that regard.

Hyperglycemia and Perinatal Morbidity

There is quite firm evidence that increasing maternal glycemia during pregnancy is associated with an increasing risk of clinically important morbidity in infants. The most obvious evidence comes from women with established diabetes, whose infants suffer an increased risk of congenital anomalies, in proportion to maternal glycemia, during the first trimester and an increased risk of several perinatal complications, in proportion to maternal glycemia, during later pregnancy. These problems are not limited to women with diagnosed diabetes. At least one cross-sectional study of women with hyperglycemia first detected during pregnancy revealed that presentation with a fasting serum glucose more than 120 mg/dL was associated with a doubling of the risk of congenital anomalies in newborns (1). Blinded (2) and unblinded (3) evaluations of maternal glucose levels during the latter half of pregnancy have revealed a fairly slow and continuous increase in complications, like macrosomia and cesarean delivery, as maternal glucose levels rise in the subdiabetic range. These findings indicate that rising maternal glycemia should be viewed as a risk factor for adverse perinatal outcomes. There may be a glucose threshold above which the risk of congenital anomalies increases sharply in early pregnancy, but there does not seem to be any such threshold for complications related to excess fetal nutrition and overgrowth in later pregnancy. Thus, maternal glycemia cannot be used to discriminate clearly between low and high risks for late-pregnancy complications, which are the major concern in GDM.

The continuous relation between maternal glycemia and macrosomia-related perinatal risks accounts for much of the controversy in the diagnosis and management of GDM. Without a biological threshold for risk, there can be no proof of a superior set of diagnostic criteria or therapeutic guidelines based on maternal glucose alone. One of three general approaches can be taken: 1) set a low and very sensitive glucose threshold for identifying at-risk pregnancies and treat all patients who exceed that threshold; 2) set a very high and insensitive threshold and treat only patients who exceed it; or 3) set a low glucose threshold and, among women who exceed it, use additional information to identify patients at high or low risk when making decisions about monitoring and treatment. The first approach has been taken by advocates of universally strict glycemic control during pregnancy. They have shown that it is possible to eliminate the 10–20% excess risk of complications such as fetal macrosomia in women with diagnosed GDM, by having all patients perform capillary glucose monitoring and by administering insulin to the 50–70% of patients whose glucose exceeds very strict postprandial glucose thresholds. The fraction of women treated with this approach is generally quite high, compared with the fraction of pregnancies in which morbidity is prevented. Nonetheless, this approach has been reported to be cost effective in one retrospective analysis (4). The cost effectiveness of this approach may decrease as glucose thresholds for diagnosing GDM are lowered (5), resulting in more low-risk women being given the diagnosis.

The second approach, of setting a high threshold for abnormality, is essentially what advocates of strict evidence-based care advocate: detect and treat only overt diabetes. There is very little hard evidence to reveal what such an approach would do to perinatal outcomes, but it seems likely that important morbidities, and even mortality, will be overlooked. Indeed, O’Sullivan (6) observed a 4-fold increase in perinatal mortality in women with GDM, in an observational study conducted before there was a standard policy for treatment of the condition. Similarly, the Toronto Tri-Hospital GDM Project (2) reported increased rates of macrosomia and cesarean delivery when women with mild gestational diabetes were allowed to go untreated. Thus, setting a high glucose threshold for diagnosing or treating GDM does not seem to offer any advantage over aggressively treating all women with GDM.

The third approach, using measures other than maternal glucose to enhance the assessment of risk in individual pregnancies, holds the potential to reduce overtreatment of low-risk pregnancies while minimizing the chances of missing important morbidity or mortality. Our group has advocated this approach. Once GDM is diagnosed [preferably by a sensitive set of criteria such as those recommended by the 4th Workshop Conference on GDM (5)], an easily obtainable measure of maternal glycemia can be used (as a so-called first cut) to identify low-risk and very high-risk individuals. Fasting serum or plasma glucose measurements in clinic can be useful in this regard. Data from Sacks et al. (3) indicate that women with GDM and fasting serum glucose concentrations of <=85 mg/dL are at very low risk for perinatal complications; they probably do not need intensive monitoring or treatment other than dietary management. We have found a high rate of macrosomia in women with fasting glucose concentrations of >=105 mg/dL and suggest that they be managed with exogenous insulin treatment to lower that risk. The majority (50–60%) of patients have fasting serum glucose concentrations between 85 and 105 mg/dL and have an intermediate risk of perinatal complications. In this group of pregnancies, we (7) found that a fetal abdominal circumference below the 70th percentile for gestational age between 29–33 weeks gestation identified approximately two thirds of pregnancies with a low perinatal risk, if managed with nutritional therapy and without maternal self-glucose monitoring. The other approximately one third of pregnancies with a fetal AC >=70th percentile had a high risk of macrosomia that could be eliminated by aggressive insulin treatment for the last 6–10 weeks of pregnancy. Weiss et al. (8) have advocated an analogous approach using the serum fructosamine concentration, as a first cut, to identify a large number of very-low-risk pregnancies that do well with nutritional therapy. Amniotic fluid insulin levels are then used in the smaller, at-risk group to identify hyperinsulinemic infants in need of maternal insulin treatment. Their approach may be too invasive for widespread application in the United States, but it demonstrates the utility of applying simple first-cut measures of glycemia to detect no-risk individuals, followed my more complicated and nonglucose measurements to fine-tune the risk assessment among the at-risk pregnancies. The time is right for comparison of this general concept for antepartum management of GDM to standard glycemia-only approaches, to define the relative efficacy and cost-effectiveness of each approach.

Hyperglycemia and Long-Term Morbidity

The mother

The major maternal complication of GDM is a lifetime risk of diabetes that seems to be at least 50%, much higher than the prevalence of diabetes in the general population. A minority of patients have circulating immune markers during pregnancy that suggest a risk of type 1 diabetes; the majority seem to be at risk for type 2 diabetes. Virtually no physiological data are available from subsets of women with GDM and evidence for pancreatic autoimmunity. Physiological studies in immune-negative patients have generally indicated that they have two abnormalities common in other individuals at risk for type 2 diabetes: chronic insulin resistance and inappropriately low insulin secretion.

Virtually all pregnant women develop some degree of insulin resistance. Immune-negative women with GDM are slightly more insulin resistant during late pregnancy than are normal pregnant women (9). Nonpregnant women with a history of GDM have also been reported to be insulin resistant, compared with nonpregnant women who maintained normal glucose tolerance during pregnancy. These observations suggest that women with GDM encounter two distinct and additive forms of insulin resistance: short-term physiological insulin resistance during pregnancy, and more chronic insulin resistance that may be analogous to insulin resistance in other groups at risk for type 2 diabetes.

Normal women compensate for insulin resistance by increasing insulin secretion to maintain normal glucose tolerance. Women with GDM have limited ability to increase their insulin secretion. As a result, they have very blunted insulin responses to nutrients during pregnancy, compared with the augmented insulin responses of normal pregnant women (9). After pregnancy, women with GDM also have been reported to have quantitative defects in insulin secretion, relative to their degree of insulin resistance. Thus, screening for glucose intolerance during pregnancy seems to identify women with limited pancreatic B-cell reserve. The defect may be the result of autoimmunity in a minority of patients. The basis for the B-cell defect in immune-negative patients has not been determined. As discussed below, the defect could be related to their chronic insulin resistance.

The presence of chronic insulin resistance in women with a history of GDM is somewhat surprising. Such resistance would not be expected if the normal, acquired insulin resistance of pregnancy simply served as a stress test for B-cells, allowing the clinical detection of women with limited B-cell reserve caused, for example, by inherited factors. The persistence of insulin resistance after pregnancies complicated by GDM suggests that chronic insulin resistance is linked somehow to the B-cell dysfunction that is a prominent feature of the condition. Such linkage could occur if either defect caused or contributed to the other or if they resulted from some common biochemical abnormality. Epidemiological studies, demonstrating that one pregnancy increases the risk of type 2 diabetes in women with a history of GDM (10), whereas multiple pregnancies have very little effect on diabetes risk in the general population (11), have led us to postulate that insulin resistance actually contributes to B-cell dysfunction in women with GDM. Thus, their B-cell defect may be characterized as maladaptation to insulin resistance, rather than as an isolated defect in B-cell function. To the extent that this hypothesis is true, amelioration of insulin resistance should help to preserve B-cell function and prevent or delay diabetes in women with a history of GDM. We are currently testing that prediction in a cohort of high-risk Latino women (12).

The fraction of women at risk for diabetes, after GDM, is high. Thus, it could be argued that all women who have had GDM deserve care aimed at reducing their risk of diabetes. Although no measures have been proven to be effective in that regard, the pathophysiological considerations above suggest that avoidance of insulin resistance may be important in minimizing the risk of type 2 diabetes after GDM. No data are available regarding minimization of the risk of type 1 diabetes in patients with evidence for islet or B-cell autoimmunity. Both groups should have regular follow-up and education about the symptoms of diabetes to facilitate early detection and early treatment to prevent long-term diabetic complications. Women who seem to be at risk for type 2 diabetes should avoid obesity, inactivity, and medications that worsen insulin resistance (e.g. supraphysiological doses of glucocorticoids or nicotinic acid). When resources are limited, some measures can be used to identify women with the highest short-term diabetes risk, who are most in need of behavioral interventions and careful follow-up. Important among those measures are elevated fasting glucose levels and poor insulin secretion during pregnancy, glucose intolerance and poor insulin secretion after pregnancy, and obesity. Persistence of glucose intolerance after delivery seems to be the best clinically available predictor of future diabetes (13). That observation provides a rationale for postpartum glucose tolerance testing, not only to identify the approximate 10% of women who have diabetes after delivery, but also to identify a very high-risk group with impaired glucose tolerance who need intensive follow-up and behavioral intervention. Finally, women with a history of GDM should participate in effective family planning to assure that they have the opportunity to enter any subsequent pregnancies with plasma glucose concentrations in the low-risk range for fetal malformations. In that regard, it is reassuring that low-dose combination oral contractive preparations do not seem to increase the risk of type 2 diabetes in women with recent GDM (14).

The offspring

Studies in rodent models of chemically-induced diabetes have revealed important defects in insulin action and B-cell function in adult animals that have been exposed to a diabetic intrauterine environment. Epidemiological studies in humans (15, 16) have revealed higher-than-expected rates of obesity, diabetes, and impaired glucose tolerance in children and young adults who were exposed to either established or gestational diabetes in utero. Though the relative contributions of genetics and the intrauterine environment to these problems in humans are still being studied, at least some environmental component seems to be operative. For example, Pettitt and colleagues (personal communication) have observed higher rates of obesity and higher blood glucose levels in offspring exposed to diabetes in utero, compared with their siblings who were born before the mother had developed diabetes. Thus, the potential for health problems in infants of mothers with GDM is very likely not limited to the prenatal and perinatal periods. Limited information is currently available regarding identification of the highest-risk offspring. Birth weights and weight-to-length ratios do not seem to be very useful, because they are not closely related to the risk of obesity or hyperglycemia in childhood. Maternal antepartum glucose levels and fetal and neonatal insulin levels seem to be more predictive, but in a continuous fashion without clear cut-points for offspring at highest risk. Given this information, it seems prudent to recommend careful follow-up of growth patterns and glucose levels for all offspring of women with GDM. More work is needed to determine whether long-term risks to offspring can be reduced by careful antepartum diabetes management, as well as to develop optimal management strategies for the children.

Summary

As currently defined (5), gestational diabetes is associated with important perinatal and long-term health risks. Many of the risks increase, in relation to the severity of maternal hyperglycemia. For perinatal risks to infants, the relationship seems to be continuous. Maternal fasting glucose levels can be used to identify subsets of patients with very low and very high risks. The majority of pregnancies lie between these two extremes; and nonglucose measures, such as fetal ultrasound, can be used to enhance risk assessment, thereby minimizing over- and undertreatment of patients. The major long-term maternal risk is development of type 1 or type 2 diabetes, predominantly the latter. The risk increases continuously, in relation to maternal glycemia during and, especially, after pregnancy. Patients seem to have a B-cell defect that is characterized by maladaptation to insulin resistance. The B-cell defect is predictive of future diabetes, supporting the testing and clinical application of interventions that minimize insulin resistance to delay or prevent diabetes. Women with impaired glucose tolerance in the first few months postpartum are at particularly high risk for diabetes and should receive the most intensive education, intervention, and follow-up. Offspring of women with GDM are at increased risk for obesity and have an unexpectedly high prevalence of elevated glucose levels during childhood and adolescence. Both genetic and intrauterine environmental influences are likely to contribute to these abnormalities. Optimal strategies to detect and prevent the long-term risks to offspring remain to be established.

Received January 15, 1999.

Revised February 16, 1999.

Accepted February 24, 1999.

References

  1. Schafer UM, Songster G, Xiang A, Berkowitz K, Buchanan TA, Kjos SL. 1997 Congenital malformations in offspring of women with hyperglycemia first detected during pregnancy. Am J Obstet Gynecol. 177:1165–1171.[Medline]
  2. Sermer M, Naylor CD, Gare DJ, et al. 1995 Impact of increasing carbohydrate intolerance on maternal fetal outcomes in 3637 women without gestational diabetes. Am J Obstet Gynecol. 173:146–156.[Medline]
  3. Sacks DA, Greenspoon JG, Abu-Fadil S, Herny HM, Wolde-Tsadik G, Yao JFF. 1995 Toward universal criteria gestational diabetes: the 75 gram glucose tolerance test in pregnancy. Am J Obstet Gynecol. 172:607–614.[Medline]
  4. Kitzmiller JL, Elixhauser A, Carr S, et al. 1998 Assessment of costs and benefits of management of gestational diabetes mellitus. Diabetes Care. [Suppl 2] 21:B123–B130.
  5. Metzger BE, Coustan DR, the Organizing Committee. 1998 Summary and recommendations of the 4th International Workshop Conference on Gestational Diabetes Mellitus. Diabetes Care. [Suppl 2] 21:B161–B167.
  6. O’Sullivan JB, Charles D, Mahan CM, Dandrow RV. 1973 Gestational diabetes and perinatal mortality rate. Am J Obstet Gynecol. 116:901–904.[Medline]
  7. Buchanan TA, Kjos SL, Montoro MN, et al. 1994 Use of fetal ultrasound to select metabolic therapy for pregnancies complicated by mild gestational diabetes. Diabetes Care. 17:275–283.[Abstract]
  8. Hofmann HMH, Weiss PAM, Purstner P, et al. 1990 Serum fructosamine and amniotic fluid insulin levels in patients with gestational diabetes and healthy control subjects. Am J Obstet Gynecol. 162:1174–1177.[Medline]
  9. Xiang AH, Peters RK, Trigo E, Kjos SL, Lee WP, Buchanan TA Multiple metabolic defects during late pregnancy in women at high risk for type 2 diabetes mellitus. Diabetes. In press.
  10. Peters RK, Kjos SL, Xiang A, Buchanan TA. 1996 Long-term diabetogenic effect of a single pregnancy in women with prior gestational diabetes mellitus. Lancet. 347:227–230.[Medline]
  11. Manson JE, Rimm EB, Colditz GA, et al. 1992 Parity and incidence of non-insulin-dependent diabetes mellitus. Am J Med. 93:13–18.[Medline]
  12. Azen SP, Berkowitz K, Kjos S, Peters R, Xiang A, Buchanan TA. 1998 TRIPOD: a randomized placebo-controlled trial of troglitazone in women with prior gestational diabetes mellitus. Controlled Clin Trials. 19:217–231.[CrossRef][Medline]
  13. Kjos SL, Peters RK, Xiang A, Henry OA, Montoro MN, Buchanan TA. 1995 Predicting future diabetes in Latino women with gestational diabetes: utility of early postpartum glucose tolerance testing. Diabetes. 44:586–591.[Abstract]
  14. Kjos SL, Peters RK, Xiang A, Thomas D, Schafer U, Buchanan TA. 1998 Oral contraception and the risk of type 2 diabetes in Latino women with prior gestational diabetes. JAMA 280::533–538.
  15. Pettitt DJ, Nelson RG, Saad MF, Bennett PH, Knowler WC. 1993 Diabetes and obesity in the offspring of Pima Indian women with diabetes during pregnancy. Diabetes Care. [Suppl 1]16 :310–314.
  16. Silverman BL, Rizzo T, Green OC, et al. 1991 Long-term prospective evaluation of offspring of diabetic mothers. Diabetes. [Suppl 2] 40:121–125.