Thyroid function during pregnancy and the postpartum period: iodine metabolism and disease states.

OBJECTIVE: To review various aspects of thyroid function during and early after pregnancy. METHODS: We discuss biochemical and potential pathologic changes in the thyroid associated with the gestational and postpartum periods. RESULTS: Urinary iodine excretion during the last trimester of gestation in healthy euthyroid women shows that, in areas with mild iodine intake, iodine supplementation is necessary during pregnancy and the postpartum period. This measure should be considered in iodine-sufficient areas as well. Thyroglobulin is the main biochemical marker of persistent thyroidal stimulation. Alterations in thyroid volume during pregnancy can persist after delivery, especially in breast-feeding mothers. In most patients, the goitrogenic stimulus of pregnancy can be suppressed with iodine supplementation. Autoimmune thyroid disease during pregnancy and the postpartum period is reflected by monitoring of thyroid peroxidase antibodies (TPO-Ab). Women with positive test results for TPO-Ab early in gestation showed a highly significant decrease in free thyroxine and increased thyroid-stimulating hormone levels late in gestation. The main marker of Graves' disease during pregnancy is thyroid-stimulating antibodies. Nonautoimmune gestational hyperthyroidism differs from Graves' disease in that thyroid-stimulating antibodies are not detectable. CONCLUSION: Clinicians should be alert to the fact that pregnancy can induce thyroidal pathologic conditions.

Comparison of serum placental protein hormone levels in diabetic and normal pregnancy.

Conflicting data exist concerning maternal serum concentrations of placental hormones during pregnancy in women with diabetes mellitus. To resolve some of these discrepancies, women participating in the NICHD-Diabetes in Early Pregnancy Study were studied. In this collaborative study, pregnancy was identified within 21 days of conception by serum hCG measurements. We prospectively collected 185 blood samples from 35 insulin-dependent diabetic women and 166 blood samples from 31 control women, all between 5 and 37 weeks gestation. Serum concentrations of hCG, pregnancy-specific beta-1-glycoprotein, placental lactogen, and hCG alpha were measured serially. The relationship between serum hormone, fasting blood glucose, 1-h postprandial blood glucose, and glycosylated hemoglobin concentrations was compared. Serum hCG alpha levels were significantly lower in the diabetic women than in control women at multiple time points during the first and second trimesters, while no consistent differences in the serum concentrations of hCG or pregnancy-specific beta-1-glycoprotein were found between pregnant diabetic and control women. Serum placental lactogen levels were significantly lower in diabetic women at 9-10 weeks and 20 weeks gestation. There were no correlations between fasting blood glucose, 1-h postprandial blood glucose, or glycosylated hemoglobin and any of the placental protein levels in the diabetic women. These data are consistent with a defect in synthesis and/or secretion of hCG alpha by the cytotrophoblast during the first two trimesters of pregnancy in insulin-requiring diabetic women.

Relation of hemoglobin A1 and blood glucose to cardiac function in diabetes mellitus.

To examine the relation of short- and long-term changes in glucose metabolism to cardiac function, radionuclide cineangiography and echocardiography were performed in 10 young insulin-dependent diabetic patients without clinical evidence of heart disease. Cardiac assessments were performed before and after both acute variations in blood glucose, and induction of chronic "tight glucose control" involving normalization of hemoglobin A1 concentrations. In diabetic patients, left ventricular (LV) ejection fraction (EF) at normal blood glucose concentration was indistinguishable from values in 11 normal subjects. However, during hyperglycemia (about 300 mg/dl), the average EF at rest was 61%, significantly higher than that during normoglycemia (56%, p less than 0.001). No significant change in LV diastolic dimension was noted in association with shifts between high and normal blood glucose concentrations. Normalization of hemoglobin A1 was achieved within 6 to 25 weeks. This alteration had no significant effect on LVEF, mitral valve E-F slope, or the response of systolic function to blood glucose levels. In addition, no correlation was found between LVEF and hemoglobin A1 concentrations in 4 of 5 evaluation periods. Thus, in young insulin-dependent diabetic patients without overt heart disease, variation in blood glucose concentration is associated with small but significant variation in EF at rest; normalization of hemoglobin A1 has no significant effect on LVEF or the response of systolic function to blood glucose levels.

Anti-insulin antibody titers do not influence control or insulin requirements in early pregnancy.

IgG antibodies (AB) to insulin have been reported to influence insulin requirements and control in patients who have taken insulin for prolonged periods of time. Nineteen pregnant type I diabetic patients (C-peptide less than 0.03 pmol/ml) were studied in their fifth week of gestation after the establishment of normoglycemia. Mean age was 27.5 yr and duration of diabetes, 14.2 yr (range: 1-23 yr). IgG AB to beef and pork insulin were measured. IgG AB to insulin were encountered in all diabetic patients (range: 103-6736 microU/ml). None of the nondiabetic pregnant controls in their fifth week of gestation (N = 17) had detectable (greater than 50 microU/ml) AB levels. The antibody titer did not affect the insulin requirement (P greater than 0.2, NS) or ability to achieve normoglycemia. AB levels were correlated with years of treatment with conventional insulin preparations (r = 0.73; P less than 0.001). At 5 wk postmenstruation the mean AB level in the patients with less than 10 yr duration of diabetes (N = 7) was 727 microU/ml and mean insulin requirement was 0.7 U/kg/24 h. In the group of patients with greater than 10 yr duration of diabetes (N = 12) the mean antibody titer was 3716 microU/ml and the insulin requirement was also 0.7 U/kg/24 h. IgG AB to insulin increase with increasing duration of treatment with beef/pork insulin. IgG AB do not affect the insulin requirement or the ability to achieve normoglycemia during early pregnancy.

Malformations in infants of diabetic mothers: problems in study design.

Infants of diabetic mothers have a two- to threefold increased risk of having congenital malformations. Primary prevention of these malformations may be possible if the teratogenic mechanisms in diabetes can be identified. The Diabetes in Early Pregnancy Study (DIEP) addresses the questions of high malformation rates and fetal losses using an innovative study design. Women are recruited prior to pregnancy, monitored to ensure prompt diagnosis of pregnancy, and followed carefully for diabetic problems or fetal losses. To date, 450 diabetic and 550 normal control prepregnant women have been recruited. A basal body temperature monitoring system has been successful in identifying pregnant women within 21 days of conception. Before initiating such a study, problems of feasibility, study design, and possible bias had to be solved. This report summarizes our experience in designing and conducting a prospective pregnancy study. The DIEP demonstrates that a truly prospective pregnancy study is feasible.