Gestational diabetes mellitus: a syndrome with phenotypic and genotypic heterogeneity

Genetic Risk Factors and Gene-Lifestyle Interactions in Gestational Diabetes

Paralleling the increasing trends of maternal obesity, gestational diabetes (GDM) has become a global health challenge with significant public health repercussions. In addition to short-term adverse outcomes, such as hypertensive pregnancy disorders and fetal macrosomia, in the long term, GDM results in excess cardiometabolic morbidity in both the mother and child. Recent data suggest that women with GDM are characterized by notable phenotypic and genotypic heterogeneity and that frequencies of adverse obstetric and perinatal outcomes are different between physiologic GDM subtypes. However, as of yet, GDM treatment protocols do not differentiate between these subtypes. Mapping the genetic architecture of GDM, as well as accurate phenotypic and genotypic definitions of GDM, could potentially help in the individualization of GDM treatment and assessment of long-term prognoses. In this narrative review, we outline recent studies exploring genetic risk factors of GDM and later type 2 diabetes (T2D) in women with prior GDM. Further, we discuss the current evidence on gene-lifestyle interactions in the development of these diseases. In addition, we point out specific research gaps that still need to be addressed to better understand the complex genetic and metabolic crosstalk within the mother-placenta-fetus triad that contributes to hyperglycemia in pregnancy.

The Need for Personalized Risk-Stratified Approaches to Treatment for Gestational Diabetes: A Narrative Review

Gestational diabetes mellitus (GDM) is common and is associated with an increased risk of adverse pregnancy outcomes. However, the prevailing one-size-fits-all approach that treats all women with GDM as having equivalent risk needs revision, given the clinical heterogeneity of GDM, the limitations of a population-based approach to risk, and the need to move beyond a glucocentric focus to address other intersecting risk factors. To address these challenges, we propose using a clinical prediction model for adverse pregnancy outcomes to guide risk-stratified approaches to treatment tailored to the individual needs of women with GDM. This will allow preventative and therapeutic interventions to be delivered to those who will maximally benefit, sparing expense, and harm for those at a lower risk.

The possible role of epigenetics in gestational diabetes: cause, consequence, or both

Gestational diabetes mellitus (GDM) is defined as the glucose intolerance that is not present or recognized prior to pregnancy. Several risk factors of GDM depend on environmental factors that are thought to regulate the genome through epigenetic mechanisms. Thus, epigenetic regulation could be involved in the development of GDM. In addition, the adverse intrauterine environment in patients with GDM could also have a negative impact on the establishment of the epigenomes of the offspring.

Impaired circulating glucagon-like peptide-1 response to oral glucose in women with previous gestational diabetes

BACKGROUND AND OBJECTIVE

Women with previous gestational diabetes (pGDM) are at risk of developing Type 2 diabetes. Glucagon-like peptide-1 (GLP-1) potentiates the insulin response to oral glucose, and its secretion is diminished in Type 2 diabetes. The aim of the study was to see if decreased GLP-1 secretion might be an early abnormality in the progression to Type 2 diabetes and would therefore be diminished in women with pGDM.

PATIENTS AND METHODS

Eleven women with pGDM and previously documented normal glucose tolerance and 11 control women underwent a 75 g oral glucose tolerance test (OGTT). Circulating plasma glucose, insulin, nonesterified fatty acids (NEFA) and GLP-1 concentrations were sampled.

RESULTS

One of the women with pGDM had impaired glucose tolerance and was excluded from the study. All other women had normal glucose tolerance. The women with pGDM had higher fasting glucose concentrations than controls (5.1; 4.9-5.3 vs. 4.8; 4.4-5.1 mmol/l, median; interquartile range, P = 0.04) and greater circulating glucose area under the curve (AUC) following the oral glucose load (930; 818-1015 vs. 668; 584-737 min x mmol/l, P = 0.0007). Fasting insulin concentrations and total insulin AUC were similar. The initial (0-30 min) insulin response was decreased in the pGDM women (AUC 3981; 2783-4795 vs. 6167; 5009-8145 min x pmol/l, P = 0.05). The initial (0-30 min) GLP-1 response was reduced in the pGDM women (AUC 816; 663-984 vs. 1163; 872-2024 min x pmol/l, P = 0.02).

CONCLUSION

A reduced initial GLP-1 response to oral glucose may therefore be an early abnormality in the progression to Type 2 diabetes.

Altered protein kinase C activation associated with rat embryonic dysmorphogenesis

It has been suggested that protein kinase C (PKC) is involved in the etiology of diabetic complications. The aim of the present study was to investigate the putative involvement of different PKC isoforms (alpha, beta1, beta 2, gamma, delta, epsilon, and zeta) in the embryopathy of diabetic rat pregnancy. Embryos were collected from normal and diabetic rats and assayed for PKC activity, PKC mRNA levels, and PKC protein distribution on gestational d 10 and 11. Embryos of diabetic rats showed markers of increased activity of PKC-alpha, PKC-beta1, PKC-gamma, PKC-delta, and PKC-zeta compared with embryos of normal rats on d 10. In addition, the malformed embryos had further increased PKC-gamma, and PKC-delta activity markers compared with nonmalformed embryos of diabetic rats on gestational d 10. In contrast, maternal diabetes caused only two alterations in PKC activity markers on gestational d 11, i.e. both PKC-alpha and PKC-zeta were decreased in embryos of diabetic rats. We found increased mRNA levels of PKC-beta 1 and PKC-zeta on d 10 in embryos of diabetic rats and decreased mRNA levels of PKC-gamma on d 11 in embryos of diabetic rats. Malformed embryos from diabetic rats showed increased distribution of PKC-beta 1 and PKC-beta 2 protein in the tissue compared with nonmalformed embryos from diabetic rats and embryos from normal rats. We conclude that diabetic rat embryopathy may be associated with increased activity and enhanced tissue distribution of several PKC isoforms in early organogenesis.

Clinical characteristics and outcome of pregnancy in women with gestational hyperglycaemia with and without antibodies to beta-cell antigens

AIMS

To evaluate the prevalence of beta-cell autoantibodies in women with gestational diabetes and impaired glucose tolerance, and identify clinical characteristics differentiating hyperglycaemic patients with and without autoantibodies.

METHODS

One hundred and twenty-three pregnant patients with gestational diabetes, 84 with impaired glucose tolerance and 290 with normoglycaemia were evaluated for anti-islet cell antibodies, glutamic acid decarboxylase (GAD) autoantibodies, and the components of the metabolic syndrome.

RESULTS

Autoantibody positivity was 8.9%, 17.9% and 0.3% in patients with diabetes, impaired tolerance and normoglycaemia, respectively. Hyperglycaemic patients with autoantibodies had lower body mass index, waist, weight gain at the time of the screening test and a lower percentage of previous pregnancies than those without autoantibodies. In addition, their fasting insulin values were significantly lower and inversely related to the presence of autoantibodies (odds ratio (OR) = 0.64; 95% confidence interval (CI) 0.42-0.96), the lowest values being found in anti-GAD+ patients. Autoantibody-positive women with diabetes were more frequently treated with insulin than negative patients (OR = 7.21; 95% CI 1.85-28.08).

CONCLUSIONS

Autoantibody-positive women with gestational hyperglycaemia displayed fewer features of insulin resistance and required more frequent insulin therapy than negative women and presumably had presymptomatic Type 1 diabetes. If this conclusion is corroborated by the follow-up of larger series, clinical and immunological distinction of types of gestational hyperglycaemia would be useful.

Chromosomal anomalies among the offspring of women with gestational diabetes

A limited body of data over the past 35 years has suggested that autoimmunity may be responsible for some cases of aneuploidy. The role of diabetes mellitus in the etiology of chromosomal anomalies has been infrequently studied. This study was designed to compare the prevalence of chromosome abnormalities among the offspring of women with gestational diabetes and the offspring of women without it. The authors used data from 7,332 women who underwent amniocentesis in a prospective study of pregnancy outcome (1984-1988) and examined the prevalence of autosomal and sex chromosome defects associated with gestational diabetes. Among the offspring of 231 women with gestational diabetes, the crude prevalence of chromosomal defects was twice as high as that seen in the offspring of 7,101 women without gestational diabetes. These anomalies were predominantly numeric sex chromosome defects. After adjusting for potential confounding by maternal age, body mass index, education, and first-trimester exposures in multiple logistic regression analysis, the authors found that women with gestational diabetes were 7.7 times as likely (95% confidence interval: 2.8, 21.1) to have an infant with a numeric sex chromosome defect as those without gestational diabetes. These results support the theory that some women who develop gestational diabetes may have underlying biochemical changes that induce nondisjunction and the development of chromosomal defects.

Autoimmune gestational diabetes mellitus: a distinct clinical entity?

This review gives an update of the present knowledge on what is defined here as autoimmune gestational diabetes mellitus (GDM). Autoimmune phenomena associated with type 1 diabetes mellitus (DM) can be detected in a subgroup of women with GDM. Islet autoantibodies are present in sera from women with GDM with variable frequency. Distinct phenotypic and genotypic features may be recognised in this subset of women with GDM, which are representative of a distinct clinical entity. Furthermore, these women are at increased risk of developing type 1 DM after pregnancy. However, the eventual progression of the autoimmune destruction of beta-cells in these subjects may follow different time-course patterns thus leading to variable forms of presentation of autoimmune DM. As a high-risk group for type 1 diabetes, women with previous autoimmune GDM may be candidates for potential immune intervention strategies.

Epidemiological analysis of outcomes of pregnancy in gestational diabetic mothers

The association between maternal diabetes mellitus and congenital defects has been well documented. However, few data exist on the potential teratogenic effect of gestational diabetes (GD). We analyzed 19,577 consecutive infants with malformations of unknown cause and compared the offspring of mothers with GD with those of nondiabetic mothers. The children with each of 20 types of selected anomalies among the two groups were used to calculate the odds ratio (OR). Because we used as a reference group for each congenital defect the rest of malformed infants, the value of the OR gives us the specificity between the association of GD and each congenital defect. Our analysis strongly supports the suggestion that GD is a significant risk for holoprosencephaly, upper/lower spine/rib, and renal and urinary system anomalies. GD is a heterogeneous disorder, which includes previously unrecognized and newly diagnosed nongestational diabetes mellitus (DM). Thus, it is possible that the teratogenic effect is related to latent DM. However, because it is not possible at this time to differentiate between these situations, pregnancies complicated by GD should be considered at risk for congenital anomalies. Prenatal ultrasound examination should be aimed particularly at the detection of abnormalities of the central nervous system, the renal and urinary system, and the spine/rib developmental field.

Lactic acidosis and other mitochondrial disorders

The ability of mitochondria to oxidize substrates and generate energy is integral to normal homeostasis and to the ability of cells to survive in the face of impending energy failure. Lactic acidosis is a common and readily apparent biochemical marker for mitochondrial dysfunction. However, lactic acidosis represents only the most obvious example in which acquired or congenital abnormalities of mitochondrial oxidative phosphorylating capacity contribute to the pathobiology and phenotypic expression of a broad spectrum of clinical disorders. Consequently, interventions that improve mitochondrial function or prevent mitochondrial energy failure may have widespread therapeutic implications.

Measurement of glucose metabolism and insulin secretion during normal pregnancy and pregnancy complicated by gestational diabetes

Gestational diabetes affects 2-3% of pregnant women and is associated with foetal complications including macrosomia and an increased likelihood of developing diabetes in later life. We have therefore studied seven women with gestational diabetes and five control women both during the third trimester of pregnancy and again 2-3 months post-partum, using the minimal model analysis of the frequently sampled labelled ([6,6-2H2]-glucose) intravenous glucose tolerance test. Glucose tolerance (glucose Kd) was significantly reduced in the women with gestational diabetes compared with the normal pregnant women both in pregnancy (1.16 +/- 0.11 vs 1.78 +/- 0.23%/min; p < 0.05) and post-partum (1.47 +/- 0.22 vs 2.59 +/- 0.43%/min; p < 0.05) and increased significantly in the control women after delivery (p < 0.05). Glucose effectiveness was not significantly different between the women with gestational diabetes and the control group either during or after pregnancy. Insulin sensitivity was significantly lower during pregnancy than after delivery in the women with gestational diabetes (p < 0.05). There was no significant difference in basal insulin secretion in the two groups during pregnancy or post-partum. However, during pregnancy the control subjects significantly increased (p < 0.001) their insulin secretion over a period of 20 min in response to an intravenous glucose tolerance test (96.2 +/- 42.7 pmol/kg) compared with post-partum values (58.3 +/- 25.2 pmol/kg) while in the women with gestational diabetes insulin secretion was similar in pregnancy (65.5 +/- 9.3 pmol/kg) and after delivery (57.7 +/- 15.7 pmol/kg). These data suggest that the glucose intolerance in gestational diabetes compared to normal pregnancy is due to reduced insulin sensitivity and an impaired ability in gestational diabetes to increase insulin secretion in response to glucose.

Indication and results of insulin therapy for gestational diabetes mellitus

The aim of this study was to determine whether amniotic fluid insulin concentration (AFI) is a better parameter than mean maternal blood glucose values (MBG) for deciding about insulin therapy in patients with gestational diabetes. MBG's were calculated on the base of 9 blood glucose levels during a 24 hour period after one week of diet therapy. In a prospective trial between 1987 and 1989 in Karlsburg, 123 gestational diabetic patients were randomized into two groups. Treatment was either based on the concentration of AFI or MBG levels. In a second series in Berlin, 103 patients were offered amniocentesis. 81 patients agreed and 22 refused. Treatment was then analogous to that in Karlsburg. In both groups of the randomized population, strict metabolic control was achieved. There was no difference regarding pregnancy complications. Earlier labor induction and higher cesarean section rates were seen in the non-invasive group (p < 0.05). The incidence of diabetic fetopathy and neonatal hypoglycemia was significantly lower in the invasive group (p < 0.01), even though the metabolic control parameters did not differ between the two groups. The results in Berlin correspond to these findings. In conclusion, AFI enables the recognition of any hyperinsulinism reaction to the maternal metabolic situation. We recommend the additional measurement of the AFI concentration between 28 and 36 weeks as the direct fetal parameter for deciding about insulin treatment.

Mitochondrial genetics: a paradigm for aging and degenerative diseases?

Studies of diseases caused by mitochondrial DNA mutations suggest that a variety of degenerative processes may be associated with defects in oxidative phosphorylation (OXPHOS). Application of this hypothesis has provided new insights into such diverse clinical problems as ischemic heart disease, late-onset diabetes, Parkinson's disease, Alzheimer's disease, and aging.

Maternally transmitted diabetes and deafness associated with a 10.4 kb mitochondrial DNA deletion

Diabetes mellitus (DM) is one of the most common chronic disorders of children and adults. Several reports have suggested an increased incidence of maternal transmission in some forms of DM. Therefore, we tested a pedigree with maternally transmitted DM and deafness for mitochondrial DNA mutations and discovered a 10.4 kilobase (kb) mtDNA deletion. This deletion is unique because it is maternally inherited, removes the light strand origin (OL) of mtDNA replication, inhibits mitochondrial protein synthesis, and is not associated with the hallmarks of mtDNA deletion syndromes. This discovery demonstrates that DM can be caused by mtDNA mutations and suggests that some of the heterogeneity of this disease results from the novel features of mtDNA genetics.

Screening recommendations for gestational diabetes mellitus

As a result of extensive experiences in multiple centers and a review of the current literature, we conclude that a plasma glucose level obtained 1 hour after a 50 gm oral glucose challenge is the "best" gestational diabetes mellitus screening test. This universal screening is performed at least once during pregnancy. The screening threshold should be no higher than 140 mg/dl, or an unacceptable loss in sensitivity occurs. Universal screening for gestational diabetes mellitus is justified by morbidity reduction, cost, and protocol simplicity and ease.

Gestational diabetes mellitus among African-American women

Gestational diabetes mellitus (GDM) is associated with increased risk of poor outcomes for the pregnancy. It is a strong risk factor for subsequent diabetes. The epidemiology of GDM in African-American women is not well known. It has not been demonstrated that their risk factors are similar in character and weight to those among White women. There is considerable multicollinearity among GDM risk factors such as age, parity, obesity, hypertension, and family history of diabetes, and this needs to be sorted out. This review is based on the results of a nested case-control study to evaluate the frequency of, and the relationships of the known risk factors with, the onset of GDM among African-American women. All cases of GDM within a cohort of women seen at any of the county health department clinics in Jefferson County, Alabama from 1981 to 1987 were identified. The cohort represents approximately 63% of all African-American pregnancies in the county during the period. With few exceptions (5.1% based on fasting plasma glucose greater than or equal to 120 mg/dl), potential GDM cases (7.1%) were selected on the basis of a 2 h post 100 g carbohydrate meal screening plasma glucose measure at their second prenatal visit and again at 28-32 weeks greater than or equal to 115 mg/dl and diagnosed on the basis of the results of an oral glucose tolerance test (OGTT) using the criteria of O'Sullivan and Mahan. Women with any prior history of diabetes (even in pregnancy), 1.6%, were excluded. The frequency of the new diagnosis of GDM among African-American women in this pregnancy in the cohort was 2.5% of pregnancies and 3.4% of women, which is similar to the values reported in the other studies. Controls were selected from women with negative screening tests who delivered after a GDM subject. The results reported in this paper reflect 358 cases (86% of all eligible GDM cases identified) and 273 controls. Cases were significantly older (28.3 vs. 21.7 years), of higher gravidity (2.7 vs. 1.9), more obese (76.7 vs. 61.7 kg), gained weight more rapidly (0.34 vs. 0.28 kg/week), had more hypertension in this pregnancy (28.2 vs. 2.6%), and there was a higher proportion with a family history of diabetes (41.3 vs. 16.5%) (p less than 0.001 for all comparisons). Because there were significant correlations among the risk factors in both cases and controls, multivariable logistic regression analyses were performed.(ABSTRACT TRUNCATED AT 400 WORDS)

Aetiology of gestational diabetes

Glucose tolerance deteriorates in human pregnancy, but about 97-98% of all pregnant women retain a normal glucose tolerance and only 2-3% develop gestational diabetes. The data reviewed show that the diabetogenicity of pregnancy is not due to diminished secretion of insulin or disproportional secretion of proinsulin or glucagon, nor is an increased insulin degradation involved. Only quantitative differences in insulin secretion have been observed between normal pregnant women and women with gestational diabetes. The insulin responses to an oral glucose load or a test meal are thus lower in gestational diabetic women than in normal pregnant women, despite significantly higher plasma glucose concentrations in the gestational diabetics. Also the insulin responses to intravenous glucose injections or infusions are abnormal in gestational diabetics when compared with normal pregnant women, a difference which is still detectable for some time after the completion of pregnancy in at least a fraction of gestational diabetic women. There is thus ample evidence that the diabetogenicity of pregnancy is related to a pronounced peripheral resistance to insulin. The resistance is of a similar magnitude in normal pregnant women and women with gestational diabetes, and it does not seem to be caused by significant alterations in insulin receptor binding to target tissues. The insulin resistance of the whole body is increased to about three times that seen in the non-pregnant state. The increased resistance is caused by post-insulin receptor events and is probably brought about by the cellular effects of the increased plasma levels of one or more of the pregnancy-associated hormones and free cortisol. There is evidence that the resistance is predominantly located to the muscle tissue, where significant reductions in certain key enzymes in glucose and lipid metabolism have been demonstrated. Published evidence points to a similar degree of insulin resistance in normal pregnant women and normal weight women with gestational diabetes. Most normal pregnant women are able to counteract the peripheral resistance by a significant augmentation of their basal and nutrient-stimulated insulin secretion. However, a few (2-3%) of the women do not appear to have the capability to produce a sufficiently large increase in insulin secretion and hence cannot overcome the peripheral resistance. These are the women who become glucose intolerant to such an extent that the diagnostic criteria for gestational diabetes are fulfilled.

Amniotic fluid prolactin in the third trimester of pregnancies complicated by gestational or pregestational diabetes mellitus

Amniotic fluid concentrations of immunoreactive prolactin were measured during the third trimester in 184 diabetic gravidas and correlated with concurrent levels of prolactin in maternal plasma. Prolactin measurements concorded with previously published estimates in normal gravid women and averaged 825 +/- 32 ng/mL (mean +/- SEM) in amniotic fluid and 168 +/- 6.5 ng/mL in simultaneously sampled plasma. Cross-sectional and longitudinal analyses indicated that the prolactin levels in amniotic fluid of pregnant diabetics declined significantly between weeks 32 and 40 of gestation, whereas plasma levels did not change consistently during the same interval. Mean values for amniotic fluid prolactin did not correlate with simultaneous prolactin concentrations in plasma, nor with maternal age, clinical estimates of polyhydramnios, amniotic fluid creatinine content, or lecithin/sphingomyelin (L/S) ratios or subsequent birth weight of the offspring. Clear-cut correlations with overall maternal glucose regulation could not be demonstrated. However, subtle effects may be operative since amniotic fluid prolactin displayed weak but significant correlations with concurrent levels of maternal plasma glucose, and mean values for hemoglobin A1c (HbA1c) but not with mean values for fasting plasma glucose (FPG). Amniotic fluid prolactin concentrations were significantly greater in patients with pregestational diabetes (White classes C, D, and F) than in women with gestational diabetes mellitus (GDM) (our classes A1, A2, and B1). The differences could not be accounted for by differences in metabolic regulation, maternal age, or weights of these two populations.(ABSTRACT TRUNCATED AT 250 WORDS)

Insulin sensitivity and B-cell responsiveness to glucose during late pregnancy in lean and moderately obese women with normal glucose tolerance or mild gestational diabetes

We used the minimal model technique to obtain concurrent measurements of whole-body insulin sensitivity and pancreatic B-cell responsiveness to glucose during the third trimester of pregnancy. Insulin sensitivity in normal pregnant women (n = 8) was reduced to only one third that of a group of nonpregnant women (n = 7) of similar age and relative weight. This marked insulin resistance was compensated by reciprocal enhancement of the first and second-phase insulin responses to intravenous glucose, which were increased threefold as compared with the nonpregnant women. Women with gestational diabetes mellitus (n = 16) had mean insulin sensitivity that was similar to that of the normal pregnant group, which indicates that insulin action was appropriate for the late phase of pregnancy in the gestational diabetic group. By contrast, the mean first-phase insulin response was significantly reduced in women with gestational diabetes mellitus, as compared with that of normal pregnant women (p less than 0.001). However, approximately one fifth of the group with gestational diabetes mellitus had first-phase responses that did not fall below the 95% confidence interval for the mean in normal pregnant women. The mean second-phase response was also lower in the group with gestational diabetes, although the difference was of borderline statistical significance (p less than 0.09). Our findings reveal the quantitative nature of the reciprocal changes in insulin sensitivity and B-cell function that normally accompany late pregnancy. They further indicate that during the third trimester, mild gestational diabetes is characterized by an impairment of pancreatic B-cell function rather than an exaggeration of the normal insulin resistance of late pregnancy.

Increased risk for gestational diabetes mellitus associated with insulin receptor and insulin-like growth factor II restriction fragment length polymorphisms

Gestational diabetes mellitus (GDM) is defined as glucose intolerance with onset or first recognition during pregnancy. We have examined restriction fragment length polymorphisms (RFLPs) near "candidate diabetogenic genes" as one approach to identify molecular markers for GDM genes. Genotypes for insulin hypervariable region (HVR), insulin-like growth factor II (IGF2), insulin receptor (INSR), and glucose transporter (GLUT1) RFLPs were studied in 96 GDM and 164 control subjects, matched to GDM for race, age, and gravidity. Logistic regression analysis was used to explore the relationship between genotypes at these candidate gene loci and GDM, while adjusting for the effects of potential confounding variables. Among black subjects, the INSR allele 1 (P = 0.001) and interactions between INSR allele 1 with body mass index (BMI) (P = 0.002) and history of DM in subject's mother (P = 0.004) contributed significantly to GDM risk. Among Caucasian subjects, a similar relationship between the INSR allele 1 (P = 0.007) and INSR allele 1-BMI interactions (P = 0.011) on GDM risk were observed. In Caucasians, an additional significant risk factor was determined by an INSR allele 1-IGF2 allele 2 interaction (P = 0.018). No risk factors were identified in Hispanic subjects. These data continue to support the hypothesis that GDM is a heterogeneous disorder with respect to phenotypic and genotypic features. Furthermore, our data suggest that risk for GDM in black and Caucasian subjects is not due to obesity perse but to interactions between obesity and INSR alleles. In Caucasian women, INSR and IGF2 alleles interact to confer additional risk for GDM. Thus genes underlying susceptibility to GDM in some women may be similar to genes conferring risk to NIDDM, while in others novel genes may contribute to GDM risk.

Restriction fragment length polymorphisms of the insulin gene hypervariable region in gestational onset diabetes mellitus

Restriction fragment length polymorphisms in the insulin gene hypervariable region are compared among 93 women with gestational onset diabetes mellitus and 146 women with normal glucose tolerance during pregnancy. No significant differences in gene or genotype frequencies were observed in the overall sample (p greater than 0.50). However, an increased frequency of one allele (class 1) was observed among nonoverweight patients with gestational onset diabetes mellitus with elevated fasting plasma glucose levels compared with age-, race-, and parity-matched control subjects (p = 0.061). These data suggest that gestational onset diabetes mellitus is a heterogeneous disorder with respect to both genotypic and phenotypic characteristics, and that restriction fragment length polymorphisms near the insulin gene may serve as a molecular marker for susceptibility to gestational onset diabetes mellitus only in some women.