Pathophysiology/Complications N A L
A R T I C L E
Gestational Diabetes: Should It Be Added to the Syndrome of Insulin Resistance? CHARLES M. CLARK, JR., MD CHUNFU QIU, PHD BARBARA AMERMAN, MS, RD BEVERLY PORTER, MSN
NAOMI FINEBERG, PHD SALEH ALDASOUQI, MD ALAN GOLICHOWSKI, MD
OBJECTIVE — The significance of gestational diabetes mellitus (GDM) results from its short-term detrimental effects on the fetus and its long-term prediction of N1DDM in the mother. We compared several variables associated with insulin resistance between GDM and non-GDM pregnant women to show the similarities between GDM and NIDDM (and thus insulin resistance). RESEARCH DESIGN A N D M E T H O D S — On the basis of a 3-h oral glucose tolerance test (OGTT), 52 GDM patients and 127 non-GDM patients were recruited from pregnant, nondiabetic women who had a nonfasting 1-h 50-g glucose screening test ^7.2 mmol/1 (130 mg/dl) performed between 16 and 33 weeks of gestation (a total of 518 of 3,041 women drawn from six community health care prenatal clinics were screened positive). During the OGTT, several potential markers of insulin resistance were measured at fasting and 2-h time points, in addition to the standard glucose measurements. The relationship of these variables with the diagnosis of GDM was studied. RESULTS — GDM patients, compared with non-GDM patients, had 1) higher prepregnancy weight (P = 0.011), prepregnancy BM1 (P = 0.006), C-peptide at fasting (P = 0.002) and at 2 h (P < 0.001), insulin at fasting (P = 0.001) and at 2 h (P < 0.001), triglycerides at fasting (P = 0.005) and at 2 h (P = 0.003), free fatty acids at fasting (P = 0.017), fi-hydroxybutyrate at fasting (P = 0.007); and 2) lower HDL cholesterol at fasting (P = 0.029). These variables were all predictive of GDM (P < 0.036) individually. Using stepwise logistic regression with all of these variables available, fasting (P = 0.019) and 2-h (P < 0.001) insulin levels, fasting free fatty acids (P = 0.031), and fasting 3-hydroxybutyrate (P = 0.036) were statistically significant as jointly predictive of GDM. Comparisons between GDM patients and non-GDM patients matched by BMI confirmed that the metabolic abnormalities persisted when difference in BMI was taken into account. Concomitant blood pressure measurements in women with GDM did not differ significantly from those without GDM. CONCLUSIONS — Our results show that many of the known metabolic components of the syndrome of insulin resistance (syndrome X) are predictive of GDM. These results are in keeping with the argument that GDM is one phase of the syndrome of insulin resistance. We suggest that GDM be looked upon as a component of the syndrome of insulin resistance that provides an excellent model for the study and prevention of NIDDM in a relatively young agegroup.
From the Department of Medicine (C.M.C., B.P, N.E, S.A.), the Department of Obstetrics and Gynecology (A.G.), the Regenstrief Institute for Health Care (C.M.C., C.Q.), the Wishard Memorial Hospital (B.A.), and the Richard Roudebush VA Medical Center (C.M.C.), Indiana University School of Medicine, Indianapolis, Indiana. Address correspondence and reprint requests to Charles M. Clark, Jr., MD, Regenstrief Health Center, Indiana University Medical Center, 1001 West Tenth St., Indianapolis, IN 46202-2859. E-mail:
[email protected]. Received for publication 12 September 1996 and accepted in revised form 20 November 1996. ANCOVA, analysis of covariance; ANOVA, analysis of variance; BOHB, (3-hydroxybutyrate; GDM, gestational diabetes mellitus; OGTT, oral glucose tolerance test.
DIABETES CARE, VOLUME 20, NUMBER 5, MAY
1997
G
estational diabetes mellitus (GDM), defined as glucose intolerance that has its onset or first recognition during pregnancy (1-8), occurs in 2 - 5 % of pregnancies (5). Since its first description by O'Sullivan and Mahan (1) in the early 1960s, GDM has been one of the most controversial syndromes in the field of diabetes. This applies to the lack of a consensus about its definition, diagnosis, clinical importance, and treatment (2). The importance of the recognition of GDM stems from fetal morbidity (mainly macrosomia) and from its importance as an independent predictor of NIDDM (2,4). The latter is best exemplified by a very high and cumulative "conversion rate" of patients previously diagnosed with GDM to NIDDM. Bevier et al. (4) listed several studies reporting conversion rates from 6.8 to 92% for combined impaired glucose tolerance and overt diabetes and 2.9 to 50% for diabetes alone; and Buchanan and Catalano (5) reported a conversion rate of 6 to 62% (or a cumulative risk of 5-10% per year) for the development of diabetes. These wide ranges result from different follow-up durations and different diagnostic criteria in GDM diagnosis. Geographic, socioeconomic, and ethnic variations play important roles as risk factors in determining the prevalence of development of NIDDM in women with a history of GDM. For example, Hispanic women demonstrated a higher conversion rate (54%) within 5 years (5). More recently, besides the aforementioned longitudinal observations of a high conversion rate of GDM to NIDDM, the close relationship between GDM and NIDDM has been emphasized. Pendergrass and colleagues (3) discussed the temporal similarities between GDM and NIDDM in several aspects, including metabolic abnormalities (hyperinsulinemia, insulin resistance, and P-cell hypofunction) and risk factors. In Western countries, NIDDM is a part, or a manifestation of, the syndrome of insulin resistance (9-16). Reaven (12) referred to the constellation of hyperinsulinemia, insulin resistance, obesity, dyslipidemia, hypertension, and NIDDM or
867
GDM and the insulin resistance syndrome
impaired glucose tolerance as syndrome X. The association of these disorders with atherosclerosis and coronary artery disease is at the center of contemporary research in metabolism. In addition to the above-mentioned components of the syndrome, more have been suggested in subsequent publications; examples include hyperuricemia and polycystic ovary syndrome (9). However, GDM has not yet been routinely listed as part of the syndrome, despite a bulk of evidence to suggest and support such a listing. Because it may take years or decades for overt NIDDM to become manifest, it is becoming crucial to identify individuals at risk of developing NIDDM in order to initiate efforts designed to prevent its development. Therefore, efforts to develop methods to increase the efficiency of detecting GDM appear to be warranted. We conducted this study to test the hypothesis that GDM is part of the syndrome of insulin resistance, an early phase of the syndrome present in a young group of the general population. We thus hypothesized that hyperinsulinemia, hypertriglyceridemia, and low HDL cholesterol would be present in women diagnosed with GDM at the time of diagnosis (at screening). We also hypothesized that such parameters, if statistically proven to correlate with the diagnosis of GDM, could be helpful to improve the predictability of GDM at an early stage of pregnancy.
RESEARCH DESIGN AND METHODS Patient population We recruited 52 GDM patients and 127 non-GDM patients from pregnant women who were positive in 1-h glucose screens in prenatal clinics in a major teaching hospital and five outreach clinics in Indianapolis between 23 May 1994 and 26 October 1995. In the clinics, a 1-h glucose screen was standard care for all pregnant women who were at 16-33 weeks' gestation. A positive screen was defined as a glucose value of 7.2 mmol/1 (130 mg/dl) or higher following a nonfasting 50-g glucose load. There were 3,041 women screened in the period, and 518 of them were positive. Within about 1 week, each woman with a positive screen was given a fasting 3-h 100-g-load oral glucose tolerance test (OGTT). A positive OGTT was defined as two or more values that were abnormal: fasting 5.3 mmol/1 (96 mg/dl) or above, 1 h 9.5 mmol/1 (172 mg/dl) or above, 2 h 8.3 mmol/1 (150 868
mg/dl) or above, 3 h 7.4 mmol/1 (134 mg/dl) or above. Before the initiation of the first OGTT, patients were informed of the study, and consents were obtained at this time. Patients seen at the hospital were excluded from our study if J) patients were 17 years old or younger and without parental consent; 2) patients were at 33 weeks' gestation; 3) patients' consents were unavailable; 4) patients moved out of Indiana or were lost to followup; or 5) patients were prisoners. In the outreach clinics, the exclusion criteria above applied plus patients were excluded if BMI was
