A follow-up program for women with previous gestational diabetes mellitus.

OBJECTIVE: To analyse the patterns of attendance in a gestational diabetes mellitus (GDM) follow-up program for detection of impaired glucose tolerance and diabetes mellitus. DESIGN: Retrospective cohort study using computerised data from the GDM follow-up program. PARTICIPANTS AND SETTING: All women with GDM who delivered at the Mercy Hospital for Women in Victoria between 1 January 1981 and 31 December 1995. OUTCOME MEASURES: Enrollment and maintenance in the follow-up program. Predictors of attendance analysed were attendance for the postnatal oral glucose tolerance test (OGTT), severity of GDM, insulin requirement in pregnancy, age at index pregnancy, country of birth, patient booking status and year of index pregnancy. RESULTS: There were 3524 women with GDM delivered during the study period. Attendance for postnatal OGTT was 71% and increased from 43.7% to 69.5% to 84.4% during the three five-year periods of the study (P < 0.00001). Entry into the follow-up program was 58% (1743 of 2986 eligible). A further 538 women (15.3%) were awaiting the postnatal OGTT or first follow-up OGTT. By December 1995, 45% of women who had entered the program had been lost to follow-up. Enrollment in the follow-up program was significantly predicted by insulin requirement in pregnancy (odds ratio [OR], 2.22; 95% confidence interval [95% CI], 1.57-3.13), attendance for postnatal OGTT (OR, 1.94; 95% CI, 1.64-2.29), private patient status (OR, 1.31; 95% CI, 1.12-1.54), severity of GDM (OR, 1.50; 95% CI, 1.24-1.82) and age 30 years or more (OR, 1.37; 95% CI, 1.17-1.60). Maintenance in the follow-up program was significantly associated with attendance for postnatal OGTT (OR, 2.67; 95% CI, 2.19-3.24), insulin requirement in pregnancy (OR, 2.56; 95% CI, 1.87-3.50), age 30 years or more (OR, 1.59; 95% CI, 1.34-1.88) and severity of GDM (OR, 1.55; 95% CI, 1.28-1.89). CONCLUSIONS: There are major difficulties with both recruiting women with GDM into a follow-up program and ensuring their continued attendance. However, a postnatal OGTT and consultation is the most important remediable factor for continuation in a follow-up program. The dedication of the follow-up team administrators rather than the clinical variables of the patients was probably the main determinant of compliance with the follow-up program.

Studies of postnatal diabetes mellitus in women who had gestational diabetes. Part 1. Estimation of the prevalence of unrecognized prepregnancy diabetes mellitus.

This study investigated the prevalence of undiagnosed diabetes in women in the reproductive age group in a Victorian population by analysis of the results of glucose tolerance testing in 57,563 pregnancies. Gestational diabetes (GD) was diagnosed in 4,243 pregnancies and in 2,957 (69.7%) of these, postnatal glucose tolerance testing was performed. Diabetes mellitus was diagnosed within 26 weeks of delivery in 59 women, 55 of whom were diagnosed by the postnatal glucose tolerance test (GTT). There were 4 women with GD who developed diabetic ketosis during pregnancy (3) or within 12 weeks of delivery (1). By consideration of the results of the antenatal and postnatal GTTs, it was deduced that 53% (31 of 59) of the women with diabetes diagnosed after delivery may have had unrecognized prepregnancy diabetes. Consideration of the entire glucose-tolerance tested population led to the conclusion that approximately 1 in 1,031 women in the reproductive age group in our community have unrecognized prepregnancy diabetes mellitus.

Studies of postnatal diabetes mellitus in women who had gestational diabetes. Part 2. Prevalence and predictors of diabetes mellitus after delivery.

An important part of the management of women with gestational diabetes (GD) is their subsequent follow-up after delivery. At this postnatal visit a glucose tolerance test (GTT) is essential. We have analysed the results of the postnatal GTT's in 2,957 women whose pregnancies were complicated by GD. Diabetes mellitus was diagnosed in 59 women (2.0%) in the first 6 months after delivery. As stated in Part 1 of this paper, 31 of these 59 women may have had unrecognized prepregnancy diabetes mellitus. The significant independent predictors of postnatal diabetes mellitus on logistic regression analysis in these women were severity of GD, Asian origin and the 1-hour plasma glucose level during the antenatal GTT.

Identification and treatment of women with hyperglycaemia diagnosed during pregnancy can significantly reduce perinatal mortality rates.

We wished to determine whether gestational diabetes was associated with an increased perinatal mortality rate, and to investigate the cause for the observed increase in the incidence of gestational diabetes. We therefore reviewed the results of glucose tolerance tests and pregnancy outcome in 116,303 pregnancies, 1971-1994, at the Mercy Hospital for Women. The main outcome measurements were the presence or absence of gestational diabetes, and perinatal mortality. Over the entire period of the study, gestational diabetes was associated with an increased risk of perinatal mortality (Mantel-Haenszel adjusted odds ratio 1.53, 95% CI 1.13-2.06, p = 0.0069). Women with gestational diabetes that was only diagnosed retrospectively had a higher perinatal mortality rate than their contemporaries with normal glucose tolerance (OR 2.31, 95% CI 1.37-3.91, p = 0.0025). Women in whom a glucose tolerance test was not performed continued to have a higher perinatal mortality rate than women who were tested (adjusted OR 2.21, 95% CI 1.56-3.12, p < 0.00001). There has been an increase in the prevalence of gestational diabetes from 2.9% to 8.8%. Some of this is due to changes in population characteristics (increases in maternal age, obesity and proportion from South-East Asia), but there was still an independent increase over time. We conclude that identification and treatment of women with gestational diabetes can reduce perinatal mortality rates. Similarly to diabetes mellitus in the total population, the prevalence of gestational diabetes has increased over time.

Effect of follow-up of women with gestational diabetes on the ratio of IDDM to NIDDM in pregnancy.

OBJECTIVE: We wished to test the hypothesis that the diagnosis of diabetes in women with previous gestational diabetes in our follow-up program had altered the ratio of IDDM to NIDDM in our pregnant population. RESEARCH DESIGN AND METHODS: We identified all pregnancies managed at the Mercy Hospital for Women in Melbourne, Australia, from 1971 to 1994 that were complicated by prepregnancy diabetes. In these 374 pregnancies, we identified those women who had previously been diagnosed with gestational diabetes mellitus (GDM). The changing prevalences over time of prepregnancy IDDM and NIDDM, as well as the contribution to both of these conditions made by women who had previously had GDM, were calculated. RESULTS: Over the period of the study, there was an increase in the prevalence of IDDM from 0.15 to 0.44% (chi 2 for trend, P < 0.00001) and NIDDM from 0.03 to 0.11% (chi 2 for trend, P = 0.0001). The proportion of all women with diabetes with NIDDM did not change significantly (16.7-20%). There was a progressive increase in the proportion of women with NIDDM who had had GDM (from 8.3 to 39.1%), but the trend was not statistically significant (P = 0.059). Women with NIDDM were more likely (20 of 64, 31.3%) to have had gestational diabetes in the past than women with IDDM (12 of 310, 3.9%, odds ratio 11.3, 95% CI 5.16-24.7, P < 0.0001). CONCLUSIONS: Despite finding relatively young women to have NIDDM, our GDM follow-up clinic has not yet altered significantly the ratio of IDDM to NIDDM in pregnancy.

Maternal serum triglyceride, glucose tolerance, and neonatal birth weight ratio in pregnancy.

OBJECTIVE: To determine the value of measuring serum triglyceride (TG) levels early in pregnancy for predicting late-gestation glucose tolerance and neonatal birth weight ratio (BWR) (birth weight corrected for gestational age). RESEARCH DESIGN AND METHODS: The relationships between morning nonfasting TG measured early in pregnancy (gestational age 12 +/- 6 weeks [mean +/- SD]) and glucose tolerance measured by a 3-h 50-g oral glucose tolerance test (OGTT) late in pregnancy (gestational age 30 +/- 3 weeks) and BWR were investigated in 388 women attending routine antenatal care. The data were analyzed for all women in addition to subgroups of Australian/Western European-born (n = 246) and Asian-born (n = 97) women. RESULTS: Morning nonfasting TG positively correlated with the OGTT glucose area under the curve (OGTT-GAUC) (r = 0.23, P < 0.0001) in all subjects. This correlation was stronger in the subset of subjects who had TG measured between 9 and 12 weeks of gestation (r = 0.35, P = 0.0001) and was particularly strong in Asian-born women who had TG measured within this period (r = 0.71, P < 0.0001). Mean TG and the 2- and 3-h OGTT values were higher in Asian-born subjects compared with Australian/Western European-born subjects (P = 0.004, P < 0.0001, and P = 0.02, respectively). TG correlated positively with BWR in all subjects (r = 0.12, P = 0.02), in Asian-born subjects (r = 0.23, P = 0.02), and in subjects with gestational diabetes mellitus (GDM) (r = 0.60, P = < 0.001). CONCLUSIONS: TG, if measured between 9 and 12 weeks of gestation, has moderate predictive value for subsequent glucose tolerance in pregnancy. TG is also predictive of BWR in GDM subjects. Further studies are warranted to investigate the role of early TG measurement in the screening and management of GDM. Metabolic heterogeneity exists between Asian-born and Australian/Western European-born women, the significance of which is still unclear and warrants further study.

Prevalence of antibodies to glutamic acid decarboxylase in women who have had gestational diabetes.

OBJECTIVES: Our purpose was to determine the prevalence of autoantibodies to glutamic acid decarboxylase in women who had had gestational diabetes, including those in whom insulin-requiring or non-insulin-requiring diabetes mellitus has since developed. STUDY DESIGN: The study group comprised 734 women with previous gestational diabetes who were consecutive attendees to a follow-up clinic. These women were tested for autoantibodies to glutamic acid decarboxylase with a radioimmunoprecipitation assay. We similarly tested 104 women in whom permanent diabetes mellitus developed after gestational diabetes, of whom 20 were using insulin and 84 were not. Those using insulin also had fasting C-peptide levels measured. RESULTS: Thirteen of the 734 (1.8%, 95% confidence interval 0.9% to 3.0%) women with previous gestational diabetes were positive for autoantibodies to glutamic acid decarboxylase. Of the 20 women with diabetes treated with insulin, 12 had insulin deficiency confirmed by low levels of C peptide; all 12 were positive for autoantibodies to glutamic acid decarboxylase. Of the 84 women with diabetes not requiring insulin, 6 (7.1%, 95% confidence interval 2.7% to 14.9%) were positive for autoantibodies to glutamic acid decarboxylase. CONCLUSIONS: The prevalence of autoantibodies to glutamic acid decarboxylase in women with previous gestational diabetes was 1.8%. Our data also showed that insulin-dependent diabetes mellitus will develop in 1.7% of women with gestational diabetes. A positive test for autoantibodies to glutamic acid decarboxylase may help in the early identification of insulin-dependent diabetes mellitus. Adult-onset insulin-dependent diabetes mellitus developed in only 5.2% (12/230) of women with previous gestational diabetes who later had diabetes mellitus.

Factors predictive of recurrent gestational diabetes diagnosed before 24 weeks' gestation.

The purpose of this study was to determine which patient and pregnancy characteristics in the first pregnancy complicated by gestational diabetes mellitus (GDM) were associated with the diagnosis of GDM before 24 weeks' gestation in a subsequent pregnancy--early recurrent GDM. The case notes of 180 women who previously had GDM diagnosed and who had glucose tolerance tests performed before 24 weeks' gestation in their next ongoing pregnancy were reviewed. Factors examined included severity of GDM, insulin requirement, racial origin, macrosomia, obesity, age, family history of diabetes, preeclampsia, and parity. Multivariate analysis showed that women with early recurrent GDM were more likely, in their first pregnancy with GDM, to have needed insulin (odds ratio [OR] 11.26; 95% confidence interval [CI] 2.02 to 62.65), to be more often of non-Northern European origin (OR, 5.53; 95% CI, 2.46 to 12.44), to have had a macrosomic infant (OR, 4.01; 95% CI, 1.40 to 11.49) or severe GDM (OR, 3.52; 95% CI, 1.60 to 7.76), and were more often 30 years or more of age (OR, 2.27; 95% CI, 1.05 to 4.90). Obesity, family history, fasting plasma glucose levels, and parity were not significant risk factors. However, even without any of the significant risk factors, logistic regression modeling suggested that a woman who has had GDM in a previous pregnancy has a 5.1% (95% CI, 2.2 to 11.6%) chance of having early recurrent GDM. We therefore continue to recommend that all women who have had GDM diagnosed previously should have glucose tolerance testing performed early (before 24 weeks' gestation) in any future pregnancies.

Value of cardiotocography in women with antepartum haemorrhage--is it too late for caesarean section when the cardiotocograph shows ominous features?

Caesarean section is thought to be indicated by an ominous antepartum cardiotocograph (CTG). However, the fear remains that infants delivered for this indication in the presence of antepartum haemorrhage, especially when premature, are destined to have severe hypoxic neurological damage. We therefore reviewed our experience of cardiotocography in women with antepartum haemorrhage (APH) from 1989 to 1992. There were 472 women with APH who had a CTG performed. Of them, 68 had abruptio placentae and 317 had an APH of undetermined cause. For the group with abruptio placentae, the perinatal mortality rate (PMR) was 230.7 per 1,000 when the CTG was abnormal, but only 18.2 per 1,000 if the CTG was normal (odds ratio 16.2, 95% confidence interval [CI] 1.53-171.9, p = 0.02). For APH of undetermined cause, the corresponding rates were 90.9 per 1,000 and 9.8 per 1,000 (odds ratio 10.1, 95% CI 0.96-105.8, p = 0.13). There were no perinatal losses in women with APH due to placenta praevia (87 cases). There were 6 cases of critical fetal reserve identified on a CTG in women with abruptio or APH of undetermined cause. All were delivered by Caesarean section, with 4 surviving infants, 3 with normal neurological outcome and 1 lost to follow-up. There were 3 cases of APH resulting in an infant with cerebral palsy, all of whom had had a normal antepartum CTG. Our data suggest that cardiotocography allows pregnancy to be safely prolonged in pregnancies complicated by abruptio placentae or APH of undetermined cause, and that Caesarean section is an appropriate form of delivery when the CTG becomes abnormal in these cases.

Maternal glucose tolerance and obstetric complications in pregnancies in which the offspring developed type I diabetes.

OBJECTIVE: To identify possible in utero risk factors in children who develop type I diabetes and to determine the risk of development of type I diabetes in the children of women with gestational diabetes. RESEARCH DESIGN AND METHODS: All known children with type I diabetes born at the Mercy Hospital for Women whose mothers had glucose tolerance tests (GTTs) performed during pregnancy were identified. The results of the mothers' GTTs were compared with those of the hospital population, as were their obstetric complications. RESULTS: We identified 38 children with type I diabetes born at this hospital whose mothers had GTTs performed during pregnancy. Only one of these mothers had gestational diabetes, compared with 5.6% in the overall hospital population (adjusted odds ratio 0.69, 95% confidence interval 0.12-3.84, P = 0.99). There were no differences in the blood glucose levels between the mothers of the children who developed diabetes and the general hospital population. The birth weights of the children destined to develop diabetes also showed no deviation from the expected distribution, and there were no outstanding features of the mothers' obstetric histories. CONCLUSIONS: Maternal blood glucose level is not an important determinant of the child's risk of developing type I diabetes.

Value of early glucose tolerance testing in women who had gestational diabetes in their previous pregnancy.

The patterns of glucose tolerance, clinical characteristics, and follow-up results of 1,027 pregnant women who had gestational diabetes diagnosed in their previous pregnancy, were analyzed. Glucose tolerance testing was performed before 24 weeks (mean 16.6 +/- 4.5 weeks) in 180 women (group 1); when the result was normal the test was repeated at 26-30 weeks' gestation. In this group the incidence of recurrent gestational diabetes was 49.4%, and early testing diagnosed 61.8% of cases. Group 2 consisted of 685 women in whom glucose tolerance was tested only at 26-30 weeks' gestation. In this group the incidence of recurrent gestational diabetes was 34.0%. Group 3 consisted of 162 women in whom glucose tolerance was not tested in the subsequent pregnancy. Perinatal mortality rates in Groups 1 to 3 were 2.2%, 0.6% and 3.1% respectively in the pregnancy subsequent to that in which gestational diabetes was first diagnosed. The risk of emerging diabetes mellitus on follow-up was greater in women in whom gestational diabetes was diagnosed early than in those diagnosed at the usual time (33.3% versus 12.5%, p < 0.05). Analysis of the 13 perinatal deaths in Groups 1-3 revealed 5 that were potentially avoidable; none of the 3 women with recurrent gestational diabetes in the early tested group who had perinatal deaths received insulin in their subsequent pregnancies. Although perinatal deaths are unavoidable in some high-risk pregnancies associated with gestational diabetes, this study suggests that early diagnosis of gestational diabetes may allow further reduction of perinatal mortality.(ABSTRACT TRUNCATED AT 250 WORDS)

Amniotic fluid insulin values in women with gestational diabetes as a predictor of emerging diabetes mellitus.

Amniotic fluid insulin levels were estimated in 30 women with insulin-dependent diabetes, 216 with gestational diabetes and 27 with normal glucose tolerance. Results were correlated with birth-weight, incidences of fetal macrosomia and neonatal hypoglycaemia, and the risk of the mothers with gestational diabetes developing diabetes mellitus on follow-up. The women with prepregnancy diabetes had significantly higher amniotic fluid insulin values and showed a significant correlation between raised liquor insulin values (> 97th percentile) and hypoglycaemia in the infant (p = 0.039). In the gestational diabetic pregnancies there were highly significant associations between elevated liquor insulin values and macrosomia (p < 0.0045) and birth-weight (p < 0.00004), and a weak correlation with neonatal blood glucose levels (p = 0.042). Women with gestational diabetes who later developed permanent diabetes mellitus had higher mean amniotic fluid insulin levels than those whose glucose tolerance remained normal on follow-up (p < or = 0.0072) and more of them had a level greater than the 97th percentile than those whose glucose tolerance remained normal (odds ratio 6.48, 95% confidence interval 1.51-27.8, p = 0.0094). However a high amniotic fluid insulin level was of less clinical value for detection of women destined to develop diabetes (7 of 25, 28%) than was the need for insulin therapy during pregnancy (18 of 39, 46%).

Gestational diabetes and follow-up among immigrant Vietnam-born women.

Gestational diabetes is associated with an increased risk of fetal macrosomia and perinatal death. Immigrant mothers from Vietnam who delivered in the Mercy Hospital for Women between January 1, 1979 and December 31, 1990 were investigated to assess their risk of gestational diabetes, the factors that were associated with gestational diabetes, and the prevalence of diabetes mellitus on follow-up. These mothers were compared with Australian-born mothers attending the same hospital and who delivered in the same period. Using a logistic regression model, gestational diabetes was found to be more common in Vietnam-born mothers who were older, who were primigravidas, or were underweight and the risk of gestational diabetes increased over the time period of the study. The adjusted relative risk of gestational diabetes for Vietnam-born women was 1.43 (95% confidence limits 1.10, 1.86) compared with Australian-born women. The incidence of gestational diabetes was 7.8% (144 of 1,839) in Vietnam-born mothers and 4.3% (1,173 of 27,086) in Australian-born mothers. Vietnam-born mothers also had a greater risk of diabetes mellitus on follow-up; 25% (17 of 68) of those with follow-up testing had developed diabetes mellitus within 9 years of diagnosis of gestational diabetes, in comparison with an incidence of 9% (52 of 581) of Australian-born mothers with follow-up testing. Vietnam-born mothers should have glucose tolerance testing performed during pregnancy to detect gestational diabetes and those diagnosed should have long-term follow-up to detect the development of diabetes mellitus.

Excessive birth weight and maternal glucose tolerance--a 19-year review.

The incidence of birth-weight of 4,540 g (10 lb) or more rose from 0.87% in the years 1971 to 1977 to 1.16% in the 12 years from 1978 to 1989 with a concomitant increase in hyperglycaemia in our antenatal population. The relationship between excessive birth-weight and maternal glucose tolerance was investigated in the light of these observations. The results from glucose tolerance tests performed routinely during the pregnancies of 510 women who delivered infants with a birth-weight of 4,540 g or more were compared with those from a control series of 5,003 women with consecutively tested pregnancies. Glucose tolerance in subsequent pregnancies was also compared with the control series, and in 1991 the study group women were investigated for emergence of permanent diabetes mellitus. Excessive birth-weight was associated with maternal hyperglycaemia (p < 0.05) but not with gestational diabetes; 79% of infants with birth-weight > or = 4,540 g were born to mothers who were not hyperglycaemic. There was no increase in glucose intolerance in subsequent pregnancies in the study group and only 2 of 49 women with follow-up testing had diabetes mellitus. Birth-weight > or = 4,540 g occurred in 1.1% of the total population and 1.1% of women with gestational diabetes, and was related to maternal hyperglycaemia in about 1 in 5 cases. The increased incidence of excessive birth-weight infants was not related to the increased incidence of gestational diabetes in our pregnant population. Birth-weight > or = 4,540 g had a poor association with later development of diabetes.

Incidence and severity of gestational diabetes mellitus according to country of birth in women living in Australia.

Gestational diabetes mellitus (GDM) was diagnosed in 1928 of 35,253 (5.5%) tested pregnancies at the Mercy Maternity Hospital in Melbourne between 1979 and the end of 1988. Compared with women born in Australia and New Zealand, the incidence of GDM was significantly greater in women born on the Indian subcontinent (15%); in women born in Africa (9.4%), Vietnam (7.3%), Mediterranean countries (7.3%), and Egypt and Arabic countries (7.2%); and in Chinese (13.9%) and other Asian (10.9%) women. There was no significant difference for women born in the United Kingdom and northern Europe (5.2%), Oceania (5.7%), North America (4.0%), or South America (2.2%). With the World Health Organization criteria as a guide to the severity of hyperglycemia, compared with mothers born in Australia and New Zealand, there were significant increases in the incidences of the more severe grades of GDM in parturients born in the Mediterranean region, Asia, the Indian subcontinent, Egypt, and Arabic countries. The incidence of GDM increased significantly in all racial groups, rising from 3.3% during 1979-1983 to 7.5% during 1984-1988.

When is a maternal death a maternal death? A review of maternal deaths at the Mercy Maternity Hospital, Melbourne.

The limiting of the reporting of maternal deaths to those that are included in the criteria of the World Health Organization excludes deaths which yield useful information for further improvements in clinical performance. In this series of 22 maternal deaths, six deaths would have been excluded from reporting: one "direct" obstetric death of pre-eclampsia; one "indirect" death as a result of renal and cardiac failure; two deaths as a result of postnatal depression which led to suicide three and four months postpartum, respectively; and two deaths of cancers, where diagnostic delay may have been a result of the coexistent pregnancy. The importance of primary pulmonary hypertension, cardiomyopathy and psychiatric illness is emphasized. We endorse the recent recommendation of the International Federation of Gynaecology and Obstetrics (FIGO) that all maternal deaths that occur more than 42 days after the end of a pregnancy should be assessed for possible relationships with childbirth, and suggest that a time limit of one year would include all deaths that are worthy of scrutiny.