Williams Manual of Pregnancy Complications, 23 ed.

CHAPTER 72. Pregestational Overt Diabetes

Women with diabetes during pregnancy can be separated into those who were known to have diabetes before pregnancy (pregestational) and those diagnosed during pregnancy (gestational). The latter group is discussed in Chapter 71Women with diabetes antedating pregnancy have been defined using the White classification, which emphasizes that end-organ derangements, especially those involving the eyes, kidneys, and heart, have significant effects on pregnancy outcome. As shown in Table 72-1, the longer the duration of diabetes before pregnancy, the more advanced the classification. For several years now, the American College of Obstetricians and Gynecologists has no longer used this classification scheme. Instead the focus is now on when diabetes is diagnosed in relation to pregnancy and degree of maternal metabolic control.

TABLE 72-1. Classification Scheme Used from 1986 through 1994 for Diabetes Complicating Pregnancy


It is unquestioned that pregestational diabetes has a significant impact on pregnancy outcome. The embryo, as well as the fetus and the mother, can experience serious complications directly attributable to diabetes. For example, women with diabetic pregnancies are at high risk for development of preeclampsia (see Chapter 23), and this complication occurs in approximately one-half of women with diabetic nephropathy (class F).


Improved fetal surveillance, neonatal intensive care, and maternal metabolic control have reduced perinatal losses in women with overt diabetes to between 2 and 4 percent. These rates have plateaued because the two major causes of fetal death—congenital malformations and “unexplained” fetal death—remain largely unchanged by medical intervention.


Spontaneous abortion is associated with poor glycemic control during the first trimester. Only those women with initial glycohemoglobin A1 concentrations above 12 percent or persistent preprandial glucose concentrations above 120 mg/dL appear to be at increased risk for abortion.


The incidence of major malformations in infants of women with pregestational diabetes is 5 to 10 percent. These account for almost half of perinatal deaths in diabetic pregnancies. Specific types of anomalies linked to maternal diabetes and their relative incidence are summarized in Table 72-2. Diabetes is not associated with increased risk for fetal chromosomal abnormalities.

TABLE 72-2. Congenital Malformations in Infants of Women with Overt Diabetes



It is generally believed that increased severe malformations are the consequence of poorly controlled diabetes both preconceptionally as well as early in pregnancy. For example, women with lower glycosylated hemoglobin values at conception have fewer anomalous fetuses compared to women with abnormally high values. Those with good periconceptional glucose control have fewer fetal malformations when compared with women who do not. The most common single-organ system anomalies are cardiac, musculoskeletal, and central nervous system.

“Unexplained” Fetal Demise

Stillbirths without identifiable cause are a phenomenon that is peculiar to pregnancies complicated by pregestational diabetes. They are declared “unexplained” because no factors such as obvious placental insufficiency, abruption, fetal growth restriction, or oligohydramnios are apparent. These infants are typically large for age and die before labor, usually after 34 weeks’ gestation. The incidence of unexplained stillbirths is approximately 1 percent. Although these stillbirths are unexplained, some hypothesize that hyperglycemia-mediated chronic aberrations in transport of oxygen and fetal metabolites may account for these unexplained fetal deaths whereas others hypothesize that osmotically induced villous edema may lead to impaired fetal oxygen transport and death.

Explicable stillbirths due to placental insufficiency also occur with increased frequency in women with pregestational diabetes, usually in association with severe preeclampsia. Severe preeclampsia, in turn, is increased in women with advanced diabetes and vascular complications. Similarly, ketoacidosis can cause fetal death.


The incidence of macrosomia rises significantly when mean maternal blood glucose concentrations exceed 130 mg/dL. However, classifying infants as either “macrosomic” or “nonmacrosomic,” may be erroneous because virtually all infants born to a diabetic mother are growth promoted. As shown in Figure 72-1, the birth weight distribution of infants of diabetic mothers is skewed toward consistently heavier birth weights compared with normal pregnancies. This excessive growth begins early (before 24 weeks) and may be determined by early pregnancy diabetes control.


FIGURE 72-1 Distribution of birth weights—standard deviation from the normal mean for gestational age—for 280 infants of diabetic mothers and 3959 infants of nondiabetic mothers. (Redrawn, with permission, from Bradley RJ, Nocolaides KH, Brudenell JM: Are all infants of diabetic mothers “macrosomic”? BMJ 297(6663):1583–1584, 1988.)


Although diabetic pregnancies are often complicated by hydramnios, the cause is unclear. A likely, although unproven, explanation is fetal polyuria resulting from fetal hyperglycemia.

Preterm Birth

Overt diabetes antedating pregnancy is a risk factor for preterm birth. There is a twofold increase in preterm delivery in women with pregestational diabetes, and most of these preterm deliveries are indicated for superimposed preeclampsia associated with advanced diabetes.


Modern neonatal care has largely eliminated neonatal deaths due to immaturity. However, neonatal morbidity due to preterm birth continues to be a serious consequence of pregestational diabetes. Indeed, some of the morbidities in these infants of diabetic women are considered to be uniquely related to aberrations in maternal glucose metabolism.

Respiratory Distress

Conventional obstetrical teaching generally held that fetal lung maturation was delayed in diabetic pregnancies, thus placing these infants at increased risk for respiratory distress. Subsequent observations have challenged this concept of diabetes-altered fetal lung function. Gestational age, rather than overt diabetes, is likely the most significant factor governing the development of respiratory distress.


A rapid decrease in plasma glucose concentration after delivery is characteristic of the infants of diabetic mothers. Hypoglycemia in term infants is defined as a blood glucose level at or below 35 mg/dL. This state is attributed to hyperplasia of the fetal β-islet cells induced by chronic maternal hyperglycemia. Prompt recognition and treatment of the hypoglycemic infant will minimize potentially serious sequelae.


Defined as a serum calcium level less than 8 mg/dL in term infants, hypocalcemia is one of the major metabolic derangements in infants of diabetic mothers. Its cause has not been explained.


The pathogenesis of hyperbilirubinemia in infants of diabetic mothers is uncertain. Factors implicated have included preterm birth and polycythemia with hemolysis. Venous hematocrits of 65 to 70 volume percent have been observed in as many as 40 percent of infants of diabetic mothers.


Infants of diabetic mothers may have hypertrophic cardiomyopathy that occasionally progresses to congestive heart failure. These infants are typically macrosomic, and fetal hyperinsulinemia has been implicated in the pathogenesis of their heart disease.

Long-Term Cognitive Development

Maternal diabetes has a negligible impact on cognitive development of the infant.

Inheritance of Diabetes

Offspring of women with pregestational diabetes have a low risk of developing insulin-dependent diabetes, with surveys suggesting an incidence of 1 to 3 percent. The risk is 6 percent if only the father has overt diabetes. If both parents have overt diabetes, the risk is 20 percent.


Diabetes and pregnancy interact significantly such that maternal welfare can be seriously jeopardized. With the possible exception of diabetic retinopathy, however, the long-term course of diabetes is not affected by pregnancy.

Maternal deaths have become rare in women with diabetes, although mortality is increased 10-fold as a result of ketoacidosis, underlying hypertension, preeclampsia, and pyelonephritis. The rare woman with coronary artery disease (class H) is at particular risk (50 percent) of dying as a result of pregnancy.


Hypertension induced or exacerbated by pregnancy is the major complication that most often forces preterm delivery in diabetic women. Especial risk factors for preeclampsia include any vascular complications (see Table 72-1), preexisting proteinuria, and chronic hypertension. Preeclampsia does not seem to be related to glucose control. The perinatal mortality rate is increased 20-fold for preeclamptic diabetic women compared with those who are normotensive.


Although ketoacidosis affects only about 1 percent of diabetic pregnancies, it remains one of the most serious complications. A prominent factor implicated in recurrent ketoacidosis is noncompliance. Fetal loss is about 20 percent with ketoacidosis.


Approximately 80 percent of insulin-dependent diabetics develop at least one episode of infection during pregnancy compared with 25 percent in nondiabetic women. Common infections include Candidavulvovaginitis, urinary infections, puerperal pelvic infections, and respiratory tract infections. Antepartum pyelonephritis is increased fourfold in women with diabetes and can be minimized by screening for asymptomatic bacteriuria.

Diabetic Nephropathy

The incidence of class F diabetes is approximately 5 percent. Approximately half of women in class F develop preeclampsia. Chronic hypertension with diabetic nephropathy increases the risk of preeclampsia to 60 percent. Plasma creatinine values of 1.5 mg/dL or greater and protein excretion of 3 g per 24 hours or greater before 20 weeks are predictive for preeclampsia. Pregnancy does not exacerbate or modify diabetic nephropathy.

Diabetic Retinopathy

The prevalence of retinopathy is related to duration of diabetes. The first and most common visible lesions of diabetic retinopathy are small microaneurysms followed by blot hemorrhages when erythrocytes escape from the aneurysms. These areas leak serous fluid that forms hard exudates. These features are termed benign or background or nonproliferative retinopathy. These findings would place a pregnant woman into class D (see Table 72-1) regardless of the duration of diabetes. With increasingly severe retinopathy, the abnormal vessels of background eye disease become occluded, leading to retinal ischemia with infarctions that appear as cotton wool exudates. These are considered preproliferative retinopathy. In response to ischemia, there is neovascularization on the retinal surface and out into the vitreous cavity, and these vessels obscure vision when there is hemorrhage (Figure 72-2). Laser photocoagulation before these vessels hemorrhage reduces by half the rate of progression of visual loss and blindness and is indicated during pregnancy for affected women.



FIGURE 72-2 Retinal photograph from a 30-year-old diabetic woman. A. Optic nerve head showing severe proliferative retinopathy characterized by extensive networks of new vessels surrounding the optic disc. B. A portion of the acute photocoagulation full “scatter” pattern following argon laser treatment. (With permission from Elman KD, Welch RA, Frank RN, Goyert GL, Sokol RJ: Diabetic retinopathy in pregnancy: A review. Obstet Gynecol75:119, 1990.)

There is continuing debate about the effect of pregnancy on proliferative retinopathy. This complication is the one exception in which pregnancy is thought to possibly exert a detrimental effect on the long-term outcome of diabetes. Most agree that laser photocoagulation and good glycemic control during pregnancy minimizes the potential for deleterious effects of pregnancy.

Diabetic Neuropathy

Although uncommon, some pregnant women will demonstrate peripheral symmetrical sensorimotor neuropathy due to diabetes. Another form, diabetic gastropathy, is very troublesome in pregnancy because it causes nausea and vomiting, nutritional problems, and difficulty with glucose control. Treatment with metoclopramide and H2-receptor antagonists can be successful.


The goals of management are tailored somewhat uniquely for pregnant women. Management preferably should begin before pregnancy and include specific goals during each trimester.


To minimize early pregnancy loss and congenital malformations in infants of diabetic mothers, optimal medical care and patient education is recommended before conception. Unfortunately, unplanned pregnancies continue to occur in approximately 60 percent of women with diabetes, and most of these women begin pregnancy with suboptimal glucose control. The American Diabetes Association has defined optimal preconceptional glucose control using insulin to include self-monitored preprandial glucose levels of 70 to 100 mg/dL and postprandial values below 140 mg/dL and below 120 mg/dL at 1 and 2 hours, respectively. Hemoglobin A1 or A1c measurement, which expresses an average of circulating glucose for the past 4 to 8 weeks, is useful to assess early metabolic control. The most significant risk for malformations is with levels exceeding 10 percent. Folate, 400 μg/day, given periconceptionally and during early pregnancy, decreases the risk of neural-tube defects associated with diabetes.

First Trimester

Careful monitoring of glucose control is essential to management. For this reason, many obstetricians hospitalize these women during early pregnancy to institute an individualized glucose control program and to provide education concerning the ensuing months of pregnancy. Hospitalization also provides an opportunity to assess the extent of vascular complications of diabetes and to precisely establish gestational age.

Insulin Treatment

Maternal glycemic control can usually be achieved with multiple daily insulin injections and adjustment of dietary intake. Oral hypoglycemic agents are not used because they may cause fetal hyperinsulinemia and increase the risk of congenital malformations. The action profiles of commonly used insulins are shown in Table 72-3. The goals of self-monitored capillary blood glucose control recommended during pregnancy are shown in Table 72-4. Self-monitoring of capillary glucose levels using glucometers is strongly recommended as this involves the woman in her own care.

TABLE 72-3. Action Profiles of Commonly Used Insulins


TABLE 72-4. Self-Monitored Capillary Blood Glucose Goals



The Committee on Maternal Nutrition of the National Research Council has recommended a total caloric intake of 30 to 35 kcal/kg of ideal body weight, given as three meals and three snacks daily. For underweight women, this is increased to 40 kcal/kg/d. An ideal dietary composition is 55-percent carbohydrate, 20-percent protein, and 25-percent fat, with less than 10-percent saturated fat. Obese women may be managed with lower caloric intake as long as weight loss and ketonuria are avoided.


Although achieving euglycemia based on normal pregnancy blood glucose values is the goal in management of overtly diabetic women, achieving this goal is not always possible. Thus, individualized programs are often necessary to avoid both excessive hyperglycemia as well as frequent episodes of hypoglycemia. Good pregnancy outcomes can be achieved in women with mean preprandial plasma glucose values of approximately 150 mg/dL. Thus, overtly diabetic women with glucose values considerably higher than those defined as normal both during and after pregnancy can expect good outcomes.

Second Trimester

As discussed in Chapter 3, maternal serum α-fetoprotein concentration at 16 to 20 weeks is used in association with targeted ultrasound at 18 to 20 weeks in an attempt to detect neural-tube defects and other anomalies. Maternal serum α-fetoprotein values may be lower in diabetic women, and interpretation is altered accordingly.

Third Trimester

Weekly visits to monitor glucose control and to evaluate for preeclampsia are recommended. Serial ultrasonography at 3- to 4-week intervals is performed to evaluate both excessive and insufficient fetal growth as well as amnionic fluid volume. Hospitalization is recommended for women whose diabetes is poorly controlled and for those with hypertension. A program of fetal surveillance using some of the antepartum tests described in Chapter 12is usually begun between 26 and 32 weeks, depending on clinical risk factors for fetal death. According to the American College of Obstetricians and Gynecologists, antepartum testing is recommended at least weekly.


Ideally, delivery of the diabetic woman should be accomplished near term. For women whose gestational age is certain, tests to determine fetal pulmonary maturation are not done, and delivery is planned after 38 completed weeks. For others, fetal lung maturation is tested (e.g., lecithin-sphingomyelin [L/S] ratio) at about 38 weeks and, if mature, delivery is affected.

In the overtly diabetic woman within class B or C of the White classification, cesarean delivery has commonly been used to avoid traumatic birth of a large infant at or near term. In women with more advanced diabetes, especially those with vascular disease, the reduced likelihood of successfully inducing labor remote from term has also contributed appreciably to an increased cesarean rate. Labor induction may be attempted when the fetus is not excessively large and the cervix is considered favorable.

It is important to considerably reduce or cancel the dose of long-acting insulin given on the day of delivery. Regular insulin should be used to meet most or all of the insulin needs of the mother at this time, because insulin requirements typically drop markedly after delivery. We have found that constant insulin infusion by calibrated pump is most satisfactory (Table 72-5). During labor and after either cesarean or vaginal delivery, the woman should be hydrated intravenously as well as given glucose in sufficient amounts to maintain normoglycemia. Capillary or plasma glucose levels should be checked frequently, and regular insulin administered accordingly. It is not unusual for the woman to require virtually no insulin for the first 24 hours or so and then for insulin requirements to fluctuate markedly during the next few days. Infection must be promptly detected and treated.

TABLE 72-5. Insulin Management during Labor and Delivery Recommended by the American College of Obstetricians and Gynecologists (2005)



No single contraceptive method is appropriate for all women with diabetes. The use of low-dose oral contraceptives should probably be restricted to women without vasculopathy or additional risk factors such as a strong history of ischemic heart disease. The lowest dose of estrogen and progesterone should be prescribed.

Progestin-only oral or parenteral contraceptives may be used because of minimal effects on carbohydrate metabolism. Physicians have been reluctant to recommend intrauterine devices in diabetic women, primarily because of a possible increased risk of pelvic infections. For all of the reasons cited, many overtly diabetic women elect puerperal sterilization, and this should be made readily available.

For further reading in Williams Obstetrics, 23rd ed.,

see Chapter 52, “Diabetes.”