Toward universal criteria for gestational diabetes: the 75-gram glucose tolerance test in pregnancy.

OBJECTIVES: The purpose of this study was to determine the distribution of values for the 75 gm glucose tolerance test in pregnancy and to define glucose intolerance by the relationship between maternal glucose values and neonatal macrosomia. STUDY DESIGN: A total 3505 unselected pregnant women were given a 75 gm, 2-hour glucose tolerance test. Diet or insulin therapy was offered only to patients with a fasting plasma glucose level > or = 105 mg/dl or a 2-hour post-glucose-load value > or = 200 mg/dl. Birth weights of live-born singletons delivered from 36 to 42 weeks whose mothers had a fasting plasma glucose level < 105 mg/dl and 2-hour post-glucose-load value < 200 mg/dl were used to calculate relationships between glucose levels and birth weights. RESULTS: At 24 to 28 weeks' gestation the mean and SD plasma glucose values were fasting 83.6 (8.9) mg/dl, 1 hour 128.4 (32.9) mg/dl, and 2 hour 108.4 (24.8) mg/dl. In a multiple logistic regression model the factors found to be statistically significantly associated with macrosomia were maternal race, parity, prepregnancy body mass index, weight gain, gestational age at testing, fasting plasma glucose level, and 2-hour post-glucose-load value. A positive association was found between maternal glucose values and birth weight percentiles. No clinically meaningful glucose threshold values relative to birth weight or macrosomia were found. CONCLUSION: In the absence of a meaningful threshold relationship between glucose tolerance test values and clinical outcome, criteria defining gestational diabetes will probably be established by consensus.

Do the current standards for glucose tolerance testing in pregnancy represent a valid conversion of O'Sullivan's original criteria?

The venous plasma values for the 100 gm glucose tolerance test endorsed by the Second International Workshop-Conference on Gestational Diabetes represent an arbitrary conversion from O'Sullivan's original values. The latter were calculated from whole blood by means of the Somogyi-Nelson method. The factor used to convert the whole blood values to plasma glucose values was empirically derived in a population in whom pregnancy status was not stated. A conversion factor derived from a nonpregnant population may systematically overestimate plasma glucose concentration when applied to a pregnant population. Paired specimens obtained from 995 consecutive pregnant patients were analyzed by the Somogyi-Nelson method on whole blood and the glucose oxidase method on plasma. A conversion formula was derived and was used to estimate plasma glucose values and 95% confidence limits equivalent to whole blood values. Application of this formula yields plasma glucose oxidase glucose tolerance test values of fasting, 96 mg/dl; 1 hour, 172 mg/dl; 2 hours, 152 mg/dl, and 3 hours, 131 mg/dl. These values are all lower than those currently endorsed.

How reliable is the fifty-gram, one-hour glucose screening test?

The Second International Workshop-Conference on Gestational Diabetes recommended screening all pregnant women with a 1-hour 50 gm oral glucose screening test, given without regard to time of day or time elapsed from the last meal. This study was designed to evaluate the reproducibility of that test, given under those clinical conditions. Thirty women with gestational diabetes and 80 control volunteers between 24 and 28 weeks' gestation were tested at the same time of day on two successive days. Of the 30 women with diabetes, three (10%) had glucose screening test results below the 135 mg/dl threshold on both days and 10 (33%) had results that straddled the threshold on successive days. Test results of 11 control women also straddled the threshold. Among these 21 women with results that straddled the threshold, there was no difference in mean glucose screening test results in diabetic women compared with nondiabetic ones. We conclude that, as currently recommended, the 1-hour glucose screening test is moderately reproducible. Reliance should not be placed on a single normal test result, particularly among patients with risk factors.

Screening for gestational diabetes with the one-hour 50-g glucose test.

A one-year experience of screening for gestational diabetes is reported. Patients with any of seven risk factors were screened at the time of prenatal registration. Those without risk factors, and those not found to be diabetic by 24 weeks' gestation, were tested later in pregnancy. Of 4116 patients, 77% had at least one risk factor. The prevalence of diabetes in patients with risk factors was significantly greater than among those with no risk factors (P less than .001). Of 936 patients who had no risk factors, four were found to have diabetes. Multiple logistic regression analysis suggested that family history, obesity, and age over 25 years contributed significantly to the prediction of gestational diabetes. More than 10% of gestational diabetics had screening values between 135-139 mg/dL. Among patients whose early screening values were elevated and whose initial glucose tolerance tests were normal, the odds of being classified ultimately as a gestational diabetic were 7.3 times that of patients whose initial screening tests were normal. Selective screening based on risk factors including maternal age may enhance detection of diabetes early in gestation.

The effect of dopamine infusion on insulin and glucagon secretion in man.

The effect of dopamine on the release of insulin and glucagon was investigated in eight adult subjects. Dopamine infused iv at a rate of 4 mug/kg/min X 2 hr, unaccompanied by cardiovascular changes, induced a prompt and significant increase in plasma insulin and glucagon and in plasma glucose levels. These findings suggest that dopamine may stimulate both alpha and beta cells.

Effects of dopamine infusion on pituitary hormone secretion in humans.

Dopamine, infused at a rate of 4 mug/kg/min for 3-4 h unaccompanied by any significant changes in cardiovascular dynamics, induced a prompt and sustained suppression of circulating prolactin (PRL) levels in normal men and women as well as in patients with hyperprolactinemia. At the discontinuation of the infusion, there was a marked rebound in PRL levels in normal subjects and a rapid return to basal levels in hyperprolactinemic patients. Dopamine infusion also induced a significant fall in LH levels in the normal subjects with a marked rebound in LH levels following the infusion. No significant changes in GH, TSH, and FSH levels were observed. These data indicate that in man a dopaminergic mechanism ixists in the regulation of PRL secretion and that dopamine also exerts an inhibitory effect of LH release.

Serum prolactin patterns in early human gestation.

The relation between levels of human prolactin (HPRL), other protein hromones, and estradiol in serum immediately prior to and for the first weeks after conception was determined. HPRL was measured by radioimmunoassay in serum samples obtained daily from these women during the menstrual cycle in which conception took place and for several weeks therafter. It was found that the concentration of HPRL initally increased above nonpregnant levels 32 to 36 days after the luteinizing hormone (LH) peak. The patterns of estradiol and HPRL were similar in early gestation, while there was no similarity between the patterns of HPRL and human placental lactogen. These results are in aggreement with other studies showing that high levels of estrogen influence HPRL secretion in the human subject.