Pregnancy in Preexisting Diabetes Thomas A
Total Page:16
File Type:pdf, Size:1020Kb
Chapter 36 Pregnancy in Preexisting Diabetes Thomas A. Buchanan, M.D. SUMMARY ata from birth certificates in the United macrosomia (an excessively large infant). Stillbirths States indicate that maternal diabetes com- are now uncommon in diabetic pregnancies; congeni- plicates 2%-3% of all pregnancies, but tal malformations and complications of maternal hy- these data may underestimate the true pertensive disorders account for most of the 1.5- to Dprevalence of maternal diabetes in pregnancy. Two 2-fold increase in perinatal mortality compared with major forms of maternal diabetes may occur during nondiabetic pregnancies. Macrosomia appears to be pregnancy: preexisting or "pregestational" diabetes, the most frequent fetal complication, affecting 10%- and gestational-onset or gestational diabetes mellitus 33% of infants, depending on the definition used for (GDM). Only the former is known prior to pregnancy, macrosomia. Macrosomia increases the risk of birth and this form constitutes ~10% of cases of maternal trauma and has been associated with a long-term risk diabetes. Thus, prevalence rates for pregestational of obesity in offspring. diabetes appear to be in the range of 0.1%-0.3% of all pregnancies. These pregnancies are at risk for both Maternal risks in diabetic pregnancies are greatest in maternal and fetal complications. the presence of preexisting microvascular disease (ret- inopathy and nephropathy). Diabetic retinopathy is Fetal complications of maternal diabetes can be di- present in 15%-66% of women early in pregnancy, and vided into two major categories. Complications that the retinopathy frequently worsens during gestation, arise from the effects of maternal diabetes on early especially when severe background or proliferative fetal development (i.e., in the first trimester) include changes are present early on. Laser photocoagulation spontaneous abortions and major congenital malfor- therapy prior to pregnancy can reduce the risk that mations. In the absence of special preconceptional proliferative retinopathy will worsen during gesta- diabetes management, spontaneous abortions occur tion. Overt diabetic nephropathy is present before in 7%-17% of diabetic pregnancies and major malfor- pregnancy in 5%-10% of patients; of these, two-thirds mations occur in 7%-13%. Rates of both complica- manifest hypertensive disorders during gestation. The tions are highest in women with the most marked hypertensive disorders precede pregnancy in approxi- hyperglycemia during the first trimester, and the rates mately half of the cases and develop during pregnancy of malformations appear to be decreasing in countries in the other half. Overt diabetic nephropathy in moth- and medical centers where standards of diabetes care ers increases the prevalence of intrauterine growth result in improved maternal blood glucose control retardation and prematurity in infants; fetal morbidity prior to and during early pregnancy. The most promi- and mortality increase as well. The long-term impact nent fetal complications that can arise during the of pregnancy on diabetic retinopathy and neph- second and third trimesters are stillbirth and ropathy in mothers is not known. • • • • • • • 719 INTRODUCTION tion of reasonably good national data on pregnancies complicated by maternal diabetes. There is no national surveillance program for diabetes during pregnancy in the United States. As a result, it is not possible to determine true national prevalence PREVALENCE OF DIABETES IN WOMEN OF CHILDBEARING AGE rates for diabetes during pregnancy or for the various maternal and fetal complications that can occur when diabetes and pregnancy coexist. Data to help estimate Data from the 1991-92 National Health Interview Sur- prevalence rates for diabetes and its complications vey (NHIS)3,4 on the prevalence of known diabetes in during pregnancy come from several sources. Since white and black women age 18-44 years are shown in 1989, birth certificates in most states and the District Figure 36.1. These data are based on self-reporting of of Columbia have included information on a variety of physician-diagnosed diabetes and they indicate that maternal and infant risk factors, including diabetes1. 1.2% of white women and 2.2% of black women in the The birth certificate data provide the first national age group (525,000 and 140,000 women, respec- estimates of the prevalence of diabetes during preg- tively) have been diagnosed by a physician as having nancy. However, the certificates do not distinguish diabetes. The responses did not distinguish between between the focus of this chapter, diabetes that existed IDDM and NIDDM, which have different age distribu- prior to pregnancy—pregestational diabetes, includ- tions in the population (see Chapter 2). Data from the ing insulin-dependent diabetes mellitus (IDDM) and National Health and Nutrition Examination Surveys non-insulin-dependent diabetes mellitus (NIDDM)— (NHANES), in which medical history and oral glucose and diabetes that is first detected during pregnancy tolerance testing were used to ascertain diabetes, indi- (GDM, discussed in Chapter 35). Birth certificate data cate that an additional 0.7%-1.3% of women age 20-44 may also suffer from inaccurate reporting of maternal years have undiagnosed diabetes (Figure 36.2). Of the and fetal complications (e.g., only 65% of maternal women without diabetes, many have impaired glucose diabetes was recorded on birth certificates surveyed in tolerance (IGT) (Figure 36.2), a condition in which Tennessee in 1989)2. blood glucose concentrations are above normal but not in the diabetic range (Chapter 2). When diabetes Other data sources include regional or statewide data and IGT estimates are combined, 10%-18% of non- derived from a combination of birth certificate and pregnant women age 20-44 years have some type of hospital record information and published reports abnormal glucose tolerance that would be associated from individual medical centers. The former source with fetal or maternal risks if those women became may be the most complete for a specific region, al- pregnant. though the magnitude of inaccurate reporting on birth certificates and hospital discharge summaries is difficult to assess. The latter source may suffer from at Figure 36.1 least two forms of bias related to patterns of patient Prevalence of Diagnosed Diabetes in Women Age referral and care. First, the medical centers that have 18-44 Years, U.S., 1991-92 published their patient data were predominantly spe- cialized referral centers. It is likely that these centers 2.5 White Black 6006 managed the most complicated cases of maternal dia- 5005 betes, so that prevalence rates of various maternal 2 complications may be overestimated compared with the entire population. Second, physicians in most of 4004 1.5 these centers have extensive experience in the man- agement of diabetes during pregnancy, so that mater- 3003 nal and fetal outcomes might be better for a given 1 severity of diabetes than would be true for less special- 2002 ized medical centers. 0.5 1001 Because of the limitations imposed by the lack of 0 0 national data for many aspects of pregestational diabe- Percent Number tes in pregnancy in the United States, some informa- tion from other countries has been included in this Data are based on self-reported information on physician-diagnosed diabetes chapter, particularly when the structure of the health in the 1991-92 National Health Interview Surveys. care system in those countries has allowed the collec- Source: References 3 and 4 720 Figure 36.2 Figure 36.4 Prevalence of NIDDM and IGT in Women Age Birth Rates by Maternal Age for All Black and 20-44 Years, U.S., 1976-82 White Women, U.S., 1992 20 18 IGT White Black Undiagnosed diabetes 16 Diagnosed diabetes 14 15 12 10 10 8 6 5 4 2 0 0 Non-Hispanic Black Mexican Puerto Rican Cuban 15-19 20-24 25-29 30-34 35-39 40-44 White American(New York, NY)American Maternal Age (Years) Diagnosed diabetes is based on self-reported information on physician-diag- Rates were determined from data on all birth certificates filed in the U.S. in nosed diabetes; undiagnosed NIDDM and impaired glucose tolerance (IGT) 1992. were determined by a 75-g, 2-hour oral glucose tolerance test interpreted acccording to World Health Organization criteria. Source: Reference 5 Source: 1976-80 Second National Health and Nutrition Examination Survey and 1982-84 Hispanic Health and Nutrition Examination Survey women age 25-29 years, and for black women age 20-24 years and birth rates declined almost linearly at Precise details on the age distribution of diabetes in older ages (Figure 36.4). This decline contrasts with women of childbearing age are not available. How- the rising prevalence rates of diabetes with increasing ever, data have been collected on the maternal age age among women who are pregnant, as presented distribution for all live births and on birth rates ac- below. cording to maternal age in the entire U.S. population5. In 1992, 39% of births occurred to women age <25 years and only 10% occurred to women age ≥35 years PREVALENCE OF PREGESTATIONAL (Figure 36.3). The birth rate was highest for white DIABETES IN PREGNANT WOMEN Figure 36.3 Distribution of All Live Births by Maternal Age for Data from birth certificates indicate that 2%-3% of Black and White Women, U.S., 1992 pregnancies in the United States are complicated by some form of maternal diabetes1,6-8. These data do not 35 White Black distinguish between pregestational diabetes and GDM. However, since prevalence rates for the latter 30 condition are in the range of 2%-4% when routine 25 blood glucose screening is employed during preg- nancy9,10 (see Chapter 35), it is likely that: 1) the birth 20 certificate data underestimate the overall prevalence 2 15 of maternal diabetes during pregnancy ; and 2) a mi- nority of diabetic pregnancies occur in women with 10 pregestational diabetes. 5 Age-specific prevalence rates for all types of diabetes 0 in white and black pregnant women, based on U.S.