Diabetes in Pregnancy

Epidemiology of Diabetes in Women in the United States

Studies indicate that the prevalence of preexisting diabetes and gestational diabetes among women of childbearing age is increasing in the United States. ^, Based on the 2011–16 National Health and Nutrition Examination Survey (NHANES), diabetes is estimated to affect 4.5% of US women of childbearing age (20–44 years), with 30% of cases undiagnosed. The increasing prevalence of diabetes will continue to have a profound impact on obstetrics and pediatrics in the upcoming decades. When offspring of women with diabetes are compared with those of weight-matched (nondiabetic) controls, the risk of serious birth injury is doubled, the likelihood of cesarean delivery is tripled, and the incidence of newborn intensive care unit admission is quadrupled.

Before the 20th century, pregnancy in a woman with diabetes requiring insulin portended death of the mother, the child, or both. Today, centers providing meticulous metabolic and obstetric surveillance report perinatal loss rates approaching but still higher than those seen in the general population. ^, Despite steadily falling perinatal mortality rates ( Fig. 59.2 ), fetal and neonatal mortality remain threefold or fourfold higher for mothers with type 1 or type 2 diabetes than for the normoglycemic population. Congenital fetal anomalies, many of them life-threatening and debilitating, remain three to four times more common in pregnancies of women with diabetes as compared to those without. Macrosomia and birth injury also occur several times more frequently in neonates of women with diabetes. Studies indicate that the magnitude of such risks is proportional to the degree of maternal hyperglycemia. To a great extent, the excessive fetal and neonatal morbidity of diabetes in pregnancy is preventable or at least reducible by meticulous prenatal and intrapartum care, including excellent control of hyperglycemia. This chapter reviews the pathophysiology of this complex group of hyperglycemic disorders and identifies the obstetric interventions that can improve outcome.

Type 1 Diabetes

T1DM is a chronic autoimmune disease in which destruction of, or damage to, beta cells in the islets of Langerhans results in insulin deficiency and hyperglycemia. T1DM accounts for 5% to 10% of all patients diagnosed with diabetes. However, T1DM may represent a relatively greater proportion of women in the reproductive age group with nongestational diabetes because of the relatively earlier age at onset of T1DM compared with T2DM.

Markers of the destructive immune response include islet cell autoantibodies, autoantibodies to insulin, autoantibodies to glutamic acid decarboxylase 2 (formerly designated GAD65), and autoantibodies to the tyrosine phosphatase–related islet antigen 2. It is now clear that T1DM occurs in genetically susceptible individuals, in concert with an environmental trigger. More than 60 genetic loci associated with T1DM have been discovered in family, candidate gene, and genome-wide association studies; many of these loci implicate the immune system in the disease. Epidemiologic and other studies suggest a triggering role for enteroviruses; many other environmental factors are under investigation.

T1DM usually is characterized by an abrupt clinical onset after a period of immune destruction of beta cells that might have been in progress for some time; however, the course of the disease can be quite variable. Although diabetic ketoacidosis (DKA) is considered a cardinal feature of the disease, only 30% of children present with DKA. Beta cell destruction continues after the clinical onset of diabetes, usually leading to absolute insulinopenia with resultant lifelong requirements for insulin replacement. T1DM was previously referred to as juvenile-onset diabetes, but it is now well understood that it can occur at any age. Later-onset T1DM, sometimes referred to as latent autoimmune diabetes of adults (LADA), typically has a more indolent course with longer preservation of beta cell function. Thus there is some debate as to whether LADA is a distinct entity from T1DM.

Therapy for T1DM always includes insulin, delivered via multiple daily injections or continuous subcutaneous insulin infusion (CSII, commonly referred to as insulin pump therapy).

Type 2 Diabetes

Unlike T1DM, which is primarily a disease of severe insulin deficiency, both insulin resistance and beta cell dysfunction are considered fundamental to the underlying pathophysiology of T2DM. The latter involves a loss of balance between insulin sensitivity (opposite of insulin resistance) and insulin (i.e., beta cell) response. The relationship between these two factors can be expressed as the disposition index (DI = insulin sensitivity × insulin secretory response; i.e., the normal inverse relationship between the two factors can be expressed as a constant). A decline in the disposition index is associated with the development of T2DM.

In T2DM, decreased insulin sensitivity (increased insulin resistance) and inadequate insulin secretory response lead to hyperglycemia. Decreased insulin sensitivity (insulin resistance) in individuals with T2DM results in the inability of insulin to suppress lipolysis in adipose tissue and endogenous/hepatic glucose production. Although insulin resistance plays a large role in T2DM, there is substantial heterogeneity in the degree of insulin resistance or insulin deficiency among individuals with T2DM, and the pathophysiology underlying this heterogeneity is an area of active investigation. ^, Many predisposing factors for T2DM are related to decreased insulin sensitivity, including obesity, diet, a sedentary lifestyle, family history, epigenetics, puberty, advancing age, and, of particular concern to the obstetrician, the maternal hormonal environment of pregnancy. Although it was formerly believed that T2DM was primarily a disorder of older individuals, there has been a significant increase in the prevalence of T2DM among children and adolescents since 1990. ^,

A variety of noninsulin agents are available for treatment outside of pregnancy. Typical first-line pharmacologic therapy for T2DM is metformin. Other frequently used agents include sulfonylureas, glucagon-like peptide-1 agonists (GLP1a), sodium-glucose transporter 2 inhibitors (SGLT2i), and dipeptidyl peptidase 4 inhibitors (DPP4i). Alpha-glucosidase inhibitors and thiazolidinediones (TZD) are used less commonly.

Prediabetes

Because the onset of T2DM is usually insidious, hyperglycemia insufficient to satisfy the diagnostic criteria for diabetes is often categorized as impaired fasting glucose (100 to 125 mg/dL) or impaired glucose tolerance (2-hour glucose level of 140 to 199 mg/dL 2 hours after administration of a 75 g oral glucose load). Prediabetes can also be identified based on a hemoglobin A _1c (HbA _1c ) value between 5.7% and 6.5%. This moderate hyperglycemia state is similar in severity to that seen in most women with gestational diabetes. In a prospective study, the cumulative risk of developing overt diabetes over 8 years was nearly three times higher among those with prediabetes at baseline compared to those without prediabetes (23.2% versus 8.5%; P < .001; annual risk, 3.9% versus 1.5%). Though prediabetes is not specific to any cause of hyperglycemia, prediabetes most commonly represents developing T2DM.

Gestational Diabetes

Gestational diabetes (GDM) is defined as glucose intolerance diagnosed during pregnancy that is not clearly preexisting diabetes. The underlying pathophysiology of GDM in most instances is similar to that observed for T2DM: an inability to maintain an adequate insulin response because of the significant decreases in insulin sensitivity with advancing gestation. However, a small subset of women with gestational diabetes actually have pre–type 1 diabetes. A recent study from Finland found that almost 6% of women with GDM developed T1DM within 7 years of pregnancy.

It is estimated that 5% to 9% of pregnant women will be diagnosed with GDM using the two-step procedure (50-g glucose challenge test, which if elevated is followed by 100-g, 3-hour oral glucose tolerance test [OGTT]) that is standard in the United States. Diagnosing GDM is important because therapy can reduce pregnancy complications and can lead to downstream measures to prevent or delay future maternal T2DM. The use of the 3-hour, 100-g OGTT for diagnosing GDM was based on John O’Sullivan’s effort in the 1960s to establish criteria that could predict risk for developing T2DM after pregnancy. The glucose limits (thresholds) for the 3-hour 100-g OGTT were defined to identify women in the 98th percentile of glucose response. However, problems with the 3-hour OGTT—including the propensity for patient emesis after ingesting such a large glucose load, which invalidates the test—led to the development of the two-step testing regimen currently used widely in the United States. This involves an initial, nonfasting 50-g “glucose challenge test” followed, if the glucose value 1 hour later was above 140 mg/dL (7.8 mmol/L), by the 3-hour 100-g OGGT. Thus the latter is obtained only if the challenge result is abnormal. However, the two-step diagnostic system inherently imposes delay in diagnosing GDM, with a false-negative rate of 4%. Furthermore, the typical follow-up rate for performing a 100-g OGTT after a positive GCT is only 85%.

In 2010, the International Association of Diabetes and Pregnancy Study Groups (IADPSG) recommended new criteria for the diagnosis of GDM based on the results of a one-step 2-hour 75-g OGTT, summarized in Table 59.1 . These criteria are based on the results of a large (∼25,000 participants) multinational, observational trial known as the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study, which meticulously tracked adverse pregnancy and neonatal outcomes. The IADPSG 2-hour testing criteria were based on the risks of excessive adverse perinatal outcomes (odds ratios [ORs] >1.75) for higher newborn birth weight, body fat, and cord C-peptide level, adjusted for confounding variables. The IADPSG criteria for GDM diagnosis have been adopted widely globally but not by the American College of Obstetricians and Gynecologists (ACOG), largely because of concern for the substantial increase in women diagnosed with GDM—up to 18% using the new criteria, compared to 5% to 9% with the two-step 3-hour OGTT system—without clear evidence of benefit. Accordingly, in the United States the current recommendations by the ADA for diagnosing GDM include either the one-step or the two-step regimen, whereas ACOG continues to recommend the continued use of the two-step diagnostic protocol GDM.

The current ADA protocol involves:

• Early pregnancy screening of women at high risk for pregestational diabetes ( Box 59.2 ) using criteria for nonpregnant persons.

  Box 59.2

  Diagnostic Criteria for Prediabetes ^a

  Data from American Diabetes Association. Standards of medical care in diabetes—2021. Diabetes Care. 2021; 44(Suppl. 1):S200–S210.

  FPG, Fasting plasma glucose; HbA1c, glycosylated hemoglobin; OGTT, oral glucose tolerance test.

  • HbA _1c of 5.7%–6.4%

  • FPG of 100–125 mg/dL

  • A 75-g OGTT 2-h blood glucose of 140–199 mg/dL

• At 24 to 28 weeks’ gestation, routine and uniform screening for GDM via either a one-step 75-g, 2-hour OGTT or a two-step regimen consisting of:

  • A 50-g, 1-hour glucose challenge test, which may be administered in the fasting or nonfasting state. A threshold value of 135 mg/dL or greater or 140 mg/dL or greater can be used at the discretion of the provider.

  • For glucose challenge test results exceeding the selected threshold, a 100-g, 3-hour OGTT is performed. Two abnormal values meeting or exceeding the values shown in Table 59.2 are required for the diagnosis.

    TABLE 59.2

    Three-Hour 100-g Oral Glucose Tolerance Test for Gestational Diabetes

    Data from American College of Obstetricians and Gynecologists Practice Bulletin No. 190. Gestational diabetes mellitus. Obstet Gynecol. 2018;131:e49–e64.

    Assessment for GDM	Plasma Glucose Level After a 100-g Glucose Load, mg/dL (mmol/L)
    Fasting	95 (5.3)
    1 h	180 (10.0)
    2 h	155 (8.6)
    3 h	140 (7.8)
    Test Prerequisites
    1-h, 50-g glucose challenge result ≥135 or 140 mg/dL
    Overnight fast of 8–14 h
    Seated, not smoking during the test
    Two or more values must be met or exceeded for a diagnosis of GDM, though the 2018 ACOG guidelines suggest that women with 1 abnormal value may be treated.

    GDM, Gestational diabetes mellitus.

Genetic and Other Causes of Diabetes

The ADA’s fourth classification of diabetes, comprising specific types of diabetes attributed to “other causes,” includes single gene disorders (monogenic diabetes), diseases of the exocrine pancreas (e.g., cystic fibrosis), and drug- or chemical-induced diabetes, such as occurs in the treatment of human immunodeficiency virus infection or after organ transplantation. Cystic fibrosis–related diabetes (CFRD) is becoming more common in pregnancy as treatments for cystic fibrosis improve and allow for increased survival and childbearing. One well-characterized genetic disease that is often included under the heading of monogenic diabetes is the glucokinase ( GCK ) form of MODY described by Hattersley and colleagues. Because the glucokinase enzyme phosphorylates glucose to glucose 6-phosphate, the first enzymatic step in glucose utilization in the pancreas and liver, a heterozygous inactivating GCK mutation results in mild lifelong hyperglycemia, usually with a fasting glucose level that is elevated disproportionately relative to postprandial levels.

It has been estimated that between 0.5% and 6% of pregnant women with GDM have GCK MODY. If the heterozygous mutation is present in the fetus, then the altered glucose sensing by the fetal pancreas will result in a decrease in insulin secretion. Insulin is a primary stimulus for growth in the fetus, and decrease in fetal insulin secretion results in delayed fetal growth and possible intrauterine growth restriction (IUGR). Depending on whether the mother or fetus, or both, has a defect in the glucokinase gene, the phenotype of the infant can vary from IUGR through normal fetal growth or macrosomia. Features of GDM suggesting a diagnosis of GCK MODY include elevated prepregnancy hemoglobin A _1c (5.7%–7.6%), fasting hyperglycemia in the setting of normal body mass index (BMI), family history of prediabetes/diabetes suggestive of autosomal dominant inheritance, and persistence of hyperglycemia postpartum. Genetic testing is commercially available and is indicated (after genetic counseling) in individuals with a suspicion for GCK MODY.

Normal Glucose-Tolerant Pregnancy

Glucose homeostasis is a balance between insulin resistance and insulin secretion. Diabetes of all types involves a loss of balance between insulin sensitivity (opposite of insulin resistance) and insulin secretory (i.e., beta cell) response. As noted above, the relationship between these two factors can be expressed as the DI. Women who maintain normal glucose tolerance during gestation are those able to maintain adequate insulin secretory response in the face of pregnancy-related changes in insulin sensitivity.

Glycemia changes dynamically during gestation, even in women without diabetes. A decrease in maternal fasting plasma glucose (FPG) concentrations in early pregnancy results from increasing plasma volumes in early gestation, and it can recur due to increased fetoplacental utilization in late pregnancy. Changes in insulin secretory response, discussed below, may also contribute. In contrast to fasting glucose levels, postprandial glucose levels increase with advancing gestation. The physiologic changes in late pregnancy that lead to these changes are well characterized, whereas early pregnancy physiology is less well characterized.

Late pregnancy is characterized by substantial insulin resistance (diminished insulin sensitivity). Alterations in insulin resistance affect both peripheral glucose metabolism, which takes place primarily in skeletal muscle, and endogenous (primarily hepatic) glucose production. Peripheral insulin resistance is defined as the decreased ability of insulin to stimulate glucose uptake primarily in skeletal muscle and, to a lesser degree, in adipose tissue. ^, Various methods are used to assess insulin sensitivity in vivo, including mathematical models of fasting glucose and insulin modeling (e.g., homeostasis model assessment, Matsuda OGTT), ^, the intravenous glucose tolerance test, and the gold standard, the hyperinsulinemic-euglycemic clamp. Most of these measures have identified a significant 50% to 60% decrease in insulin sensitivity in late gestation.

In a lean woman with normal glucose tolerance, there is a significant 30% increase in basal hepatic glucose production by the third trimester of pregnancy ( Fig. 59.3 ), suggestive of the development of hepatic insulin resistance. This is associated with a significant increase in fasting insulin concentrations ( Fig. 59.4 ). In the postprandial state outside of pregnancy and in lean pregnant women, increasing insulin concentrations enhance glucose uptake into skeletal muscle and adipose tissue and almost completely suppress hepatic glucose production, even in this insulin-resistant state. Women with obesity, even those with normal glucose tolerance, have a decreased ability for insulin to completely suppress hepatic glucose production in late pregnancy. These data support the concept that decreased hepatic insulin sensitivity in late gestation is more severe in women with obesity compared to those without. Lean women often have greater pregravid insulin sensitivity compared with overweight or obese women. These differences manifest before pregnancy and persist in late pregnancy because of the 50% to 60% decrease in insulin sensitivity that occurs in all pregnancies late in gestation. (see Fig. 59.4 ).

In contrast to late pregnancy, several lines of evidence suggest that insulin sensitivity is enhanced in early pregnancy. First, women with T1DM typically experience an increased risk of hypoglycemia during the first trimester. Though this has been ascribed to attempts to achieve tight glycemic control, a study of women with well-controlled diabetes demonstrated reduced insulin requirements between 9 and 16 weeks’ gestation, suggestive of enhanced insulin sensitivity. Similarly, studies using euglycemic-hyperinsulinemic clamps show a small but significant augmentation of insulin sensitivity at 12–14 weeks’ gestation compared to the pregravid state.

In addition to the changes in insulin sensitivity discussed above, there are profound changes in the pancreatic beta cell insulin secretory response during gestation that begin in early pregnancy. The changes in insulin secretory response are often described as compensatory for the profound insulin resistance that develops by late pregnancy. Indeed, the decreases in insulin sensitivity in late pregnancy are accompanied by an increase in insulin response. The insulin response to a glucose load increases approximately threefold compared with pregravid measures ( Fig. 59.5 ). Yet, insulin secretory response increases by 12–14 weeks’ gestation, prior to the development of insulin resistance and in the setting of a mild increase in insulin sensitivity. This results in a markedly elevated DI in early pregnancy, suggestive of a profound pregnancy effect on beta cell function.

Pregnancy Complicated by Diabetes

Alterations in glucose metabolism in women with diabetes have been most extensively examined in women with GDM. Across the body weight spectrum, in women with GDM, there is an increase in basal endogenous glucose production, similar to that observed in subjects with normal glucose tolerance, although fasting insulin concentrations, particularly in late gestation, are greater on average than in normal glucose-tolerant women. Insulin infusion assays indicate that the ability of insulin to suppress endogenous glucose production is decreased in women with GDM compared with a matched control group (approximately 80% versus 95%).

When insulin sensitivity is estimated antepartum or postpartum, there is a significant decrease noted among women who developed GDM compared with normal glucose-tolerant women. However, during pregnancy, the average percentage decrease in total insulin sensitivity (endogenous plus peripheral) is approximately the same in women with GDM as in matched controls (i.e., 50% to 60%) ( Fig. 59.6 ). Thus the decreased insulin sensitivity observed during pregnancy in the woman who develops GDM is significantly affected by her pregravid metabolic status.

As described above, the relationship between insulin sensitivity and insulin response has been characterized as a hyperbolic curve or, when multiplied, as the DI. A curve that is “shifted down and to the left” can be plotted for individuals with GDM, indicating reduced beta cell function for their level of insulin sensitivity, manifesting in a lower DI. The cause of beta cell dysfunction in GDM may be heterogeneous and includes those with chronic insulin resistance.

Scant data are available on changes in glucose metabolism in pregnant women with T1DM and T2DM. Schmitz and coworkers evaluated the longitudinal changes in insulin sensitivity in women with T1DM in early and late pregnancy and after delivery and showed a 50% decrease in insulin sensitivity in late gestation. There was no significant difference in insulin sensitivity in these women in early pregnancy or within 1 week of delivery compared with nonpregnant women with T1DM. Longitudinal changes in insulin dosage requirements during pregnancy in women with well-controlled T1DM before conception were reported by Garcia-Patterson and colleagues. There were variable changes in insulin requirements in early gestation, with decreases between 9 and 16 weeks’ gestation. The greatest increase was observed between 16 and 37 weeks, with a plateau or small decline thereafter (see Fig. 59.6 ). These changes in insulin dosage requirements in women with well-controlled T1DM support the earlier observations of an increase in insulin sensitivity in early gestation and a significant decrease in insulin sensitivity in late gestation. Thus, based on available data, women with T1DM have similar alterations in insulin sensitivity as are observed in women with normal glucose tolerance.

Mechanisms of Gestational Insulin Resistance

As discussed above, insulin resistance is defined as a condition in which physiologic insulin concentrations elicit decreased biologic response in target tissues. The metabolic action of insulin in target cells requires the orchestrated activation of complex molecular steps that regulate signal transduction. Autophosphorylation of insulin receptor tyrosine residues is the initial mandatory step allowing recruitment and activation of downstream effectors such as insulin receptor substrate-1 (IRS-1). The mechanisms responsible for pregnancy-induced insulin resistance have been characterized at the molecular level. In normal pregnancy, the ability of the insulin receptor to transduce an intracellular signal is diminished. In late pregnancy, skeletal muscle IRS-1 content is lower than in nonpregnant women. Downregulation of IRS-1 closely parallels the decreased ability of insulin to stimulate 2-deoxyglucose uptake in skeletal muscle in vitro. In GDM, the insulin receptor also displays a decreased ability to undergo tyrosine phosphorylation. This decreased receptor phosphorylation results in 25% less glucose transport activity and is not observed in pregnant women with normal glucose tolerance.

The insulin resistance of pregnancy is reversed shortly after delivery, consistent with the marked decrease in insulin requirements clinically in women managed with insulin. Waters and colleagues reported that there is a 120% increase in insulin sensitivity in women with GDM within 1–5 days of delivery with a concomitant decrease in insulin response. Further, only 9% of the variance in insulin sensitivity can be ascribed to variations in gestational weight gain. The placenta has long been suspected of producing hormonal factors related to these adaptations in maternal glucose metabolism. The most studied candidates include placenta-derived hormones such as human placental lactogen, progesterone, and estrogens. More recent epigenetic work has demonstrated that placental DNA methylation at specific sites is associated with insulin sensitivity in late pregnancy. Mendelian randomization suggested at least five of these loci where DNA methylation causally affects insulin sensitivity, though the downstream effects of these methylation changes have not yet been worked out.

Despite decades of intensive research, placental factors and mechanisms by which they modify insulin action in maternal tissues are poorly understood. The role of tumor necrosis factor-α (TNF-α) in insulin resistance was first revealed in studies investigating how TNF-α impairs insulin action. Since then, a wide variety of factors, including nutrients such as fatty acids and amino acids, have been found to induce insulin resistance through IRS-1 serine phosphorylation. TNF-α and other cytokines produced by the placenta can be released locally as well as in the maternal systemic circulation. In pregnant women, circulating TNF-α concentrations were inversely correlated with insulin sensitivity, as estimated from insulin clamp studies. In the third trimester, increased circulating free fatty acids may also contribute to the insulin resistance as described in nonpregnant individuals.

The cooperation of inflammatory and metabolic factors in blunting the effects of insulin has recently gained attention. The term metabolic inflammation is applied to situations of low-grade chronic inflammation observed in several metabolic disorders such as obesity and diabetes. Adipose tissue plays an essential role in initiation of the inflammatory response. Not only is it a storage depot for excess calories but it also actively releases fatty acids and secretes a variety of adipocytokines that impact endocrine and metabolic function. Pregnancy in a woman with obesity and GDM is in a state of metabolic inflammation. Macrophages originating from the maternal systemic circulation invade maternal adipose tissue and the placenta, increasing local and systemic release of proinflammatory cytokines. Altogether, these observations suggest that in pregnancy, the placenta, in addition to the expanding adipose tissue mass, is a primary cause of systemic inflammation and that the resulting altered homeostasis and insulin resistance propagate to skeletal muscle and the liver.

Maternal Morbidity

Women with pregestational diabetes are at risk for obstetric and medical complications. ^, Though those with preexisting diabetes account about 2% of deliveries, their pregnancies are complicated by significant increases in hypertension (OR = 14.2), preeclampsia (OR = 3.4), cesarean delivery (OR = 11.3), and preterm birth (OR = 4.4). The relative risk (RR) of these problems is proportional to the duration and severity of disease. Evers and colleagues reported on maternal morbidity in a cohort of 323 pregnant women with T1DM. Glycemic control was excellent, but rates of preeclampsia (12.7%), preterm birth (32%), cesarean delivery (44%), and maternal mortality (60 deaths per 100,000 pregnancies) were considerably higher than in the nondiabetic population.

Retinopathy

Diabetic retinopathy is the leading cause of blindness between the ages of 24 and 64 years. Some form of retinopathy is present in virtually 100% of women who have had T1DM for 25 years or more; approximately 20% of these women are legally blind. The topic of diabetic retinopathy in pregnancy has been recently reviewed. The pattern of progression of diabetic retinopathy is relatively predictable, proceeding from mild nonproliferative abnormalities associated with increased vascular permeability to more severe nonproliferative changes characterized by vascular closure and then finally to proliferative vasculopathy marked by the growth of new blood vessels in the retina and posterior surface of the vitreous. Vision loss from diabetic retinopathy results from several mechanisms: impairment of central vision by macular edema or capillary nonperfusion, new proliferative retinal blood vessels and accompanying fibrous tissue distorting the retina leading to tractional retinal detachment, and, finally, new blood vessels may rupture and causing preretinal and/or vitreous hemorrhage.

Although older studies suggested that rapid induction of glycemic control in early pregnancy stimulates retinal vascular proliferation, recent investigations have shown that the severity and duration of diabetes before pregnancy have a greater effect. Bourri and colleagues documented retinopathy progression in 375 T1DM patients in whom prepregnancy retinopathy was present in 30.3%. The best predictors of retinopathy progression were elevated prepregnancy HbA _1c and duration of diabetes ≥10 years. Notably, the use of CSII significantly decreased the risk of retinopathy progression. Postpartum, the rate of progression was only 4.1%, whereas the rate of regression was 9.3%. Other studies have reported similar results. Temple and colleagues studied 179 women with pregestational T1DM, performing dilated fundal examination at the first prenatal visit and at 24 and 34 weeks’ gestation. Progression to proliferative diabetic retinopathy occurred in only 2.2% of their subjects, and moderate progression occurred in 2.8%. However, progression was significantly greater in women who had had diabetes for more than 10 years (10% versus 0%).

Trials failed to show any acceleration in microvascular complications when pregnant and nonpregnant diabetic subjects were closely followed and compared. In the study by Arun and Taylor, women who had T1DM with a mean duration of 14.4 ± 8.2 years, a mean age at pregnancy of 29.8 ± 5.5 years, and a mean HbA _1c value of 8.2 ± 2.0% before pregnancy with tighter control during pregnancy were followed for 5 years. Four women required laser therapy during pregnancy, but none required it subsequently. Baseline retinopathy status was the only independent risk factor that predicted retinopathy progression.

Screening for retinopathy by a qualified ophthalmologist is recommended before pregnancy and again during the first trimester for patients with pregestational diabetes because of the demonstrated effectiveness of laser photocoagulation therapy in arresting progression. Patients with minimal disease should be reexamined yearly after delivery. Those with significant retinal pathology may require monthly follow-up.

In patients with macular edema, off-label use of intravitreal anti–vascular endothelial growth factor (anti-VEGF) agents in nonpregnant subjects has demonstrated superior improvement in visual acuity compared with laser photocoagulation. However, use of this agent in pregnant women has been limited to small case series. ^,

Nephropathy

Diabetes is the most common cause of end-stage renal disease (ESRD) in the United States and Europe. In the United States, diabetes is the leading cause of kidney failure, accounting for 44% of all new cases of kidney failure. The pathophysiology of diabetic renal disease is influenced by several factors, including genetic susceptibility, glycemic control, and the duration and severity of coexisting hypertension. Additional renal tissue insults, such as repeated urinary tract infections, excessive glycogen deposition, glycosylation, and papillary necrosis, promote deterioration of renal function. Though the kidney is typically normal at the onset of diabetes, within a few years glomerular basement membrane thickening can be identified. By 5 years, there is expansion of the glomerular mesangium resulting in diffuse diabetic glomerulosclerosis. All patients with marked mesangial expansion exhibit proteinuria exceeding 400 mg in 24 hours. The peak incidence of nephropathy occurs after about 16 years of diabetes.

Effect of Pregnancy on Progression of Nephropathy

Although some clinicians discourage pregnancy in women with diabetic renal disease because of concerns about permanent renal deterioration accruing pregnancy, recent data consistently indicate that pregnancy does not measurably alter the natural history of diabetic renal disease.

Progression of diabetic nephropathy is closely related to the degree of glycemic control, with a threshold value of HbA _1c of 6.5% in several studies. To the extent that many women have better glycemic control during pregnancy, delay or slowing of renal function deterioration can be expected. A study of renal function for 4 years before and 4 years after pregnancy in 11 patients with diabetic nephropathy showed that the gradual rise in serum creatinine over that period was unaffected by the intervening pregnancy. Imbasciati and coworkers performed a longitudinal study of 58 women with chronic renal disease whose mean serum creatinine level was 6 mg/dL at the start and 6 mg/dL at the end of pregnancy. Although women with glomerular filtration rates below 40 mL/min and proteinuria greater than 1 g/d had increased risk of delivering a low-birth-weight infant, even those with worse values had modest changes in renal function when post- and prepregnancy indices were compared.

Young and associates monitored 32 diabetic women without nephropathy (group I) and 11 diabetic women with nephropathy (group II) through pregnancy and for 1 year after delivery. In both groups, there was an increase in urinary albumin excretion during pregnancy (592 versus 119 mg/24 h, P = .0001) but no difference in postpartum albuminuria, creatinine, or creatinine clearance compared to antepartum values. Group II had a higher prevalence of chronic hypertension (72.7% versus 21.9%, P = .004) and preeclampsia (63.6% versus 6.3%, P = .0003) and lower gestational age at delivery (36 versus 38 weeks, P = .003), but pregnancy was not associated with development or progression of diabetic nephropathy.

Rossing and colleagues evaluated the effect of pregnancy on deterioration of renal function in 93 women older than 20 years of age. They compared groups of never-pregnant and pregnant women who received similar medical therapy and who had similar degrees of renal function at the start of the study. The results are shown in Fig. 59.7 . Based on this excellent prospective study, it is evident that pregnancy neither alters the time course of progression in renal disease nor increases the likelihood of transition to ESRD.

Cardiovascular Complications

Cardiovascular complications experienced by pregnant women with diabetes include chronic hypertension, pregnancy-induced hypertension, and, rarely, atherosclerotic heart disease. In composite studies of all types of diabetic pregnancies, the incidence of hypertensive disorders during pregnancy varied from 15% to 30%, with the rate of hypertension increased fourfold over that for the nondiabetic population.

Diabetic Ketoacidosis

Diabetic ketoacidosis (DKA) during pregnancy is a medical emergency for the mother and the fetus. Pregnant women with T1DM are at increased risk for DKA, although the incidence in pregnancy and morbidity of this complication have decreased from 20% or more in the older literature to less than 1% in recent reports. The rate of intrauterine fetal death, formerly as high as 35% with DKA during pregnancy, has dropped to 5% or less.

Precipitating factors for DKA include pulmonary, urinary, or soft tissue infections; poor compliance; and unrecognized new onset of diabetes. Because severe DKA threatens the life of the mother and fetus, prompt treatment is essential. Fetal well-being is in jeopardy until maternal metabolic homeostasis is reestablished. High levels of plasma glucose and ketones are readily transported to the fetus, which may be unable to secrete enough insulin to prevent DKA and in utero hypoxia. Early in the illness, hyperglycemia and ketosis are moderate. If hyperglycemia is not corrected, diuresis, dehydration, and hyperosmolality follow. Pregnant women in the early stages of DKA respond quickly to appropriate treatment of the initiating cause (e.g., broad-spectrum antibiotics), additional doses of intravenous insulin, and volume replacement.

Patients with advanced DKA usually present with typical findings, including hyperventilation, normal or obtunded mental state (depending on the severity of the acidosis), dehydration, hypotension, and a fruity odor to the breath. Abdominal pain and vomiting may be prominent symptoms. The diagnosis of DKA is confirmed by the presence of hyperglycemia (glucose >200 to 300 mg/dL) and base deficit of −4 mEq/L or greater. As many as one-third of patients in the early or very late stages of DKA have initial blood glucose levels lower than 200 mg/dL. A pregnant diabetic patient with a history of poor food intake or vomiting for longer than 12 to 16 hours should have a thorough workup for DKA, including a complete blood cell count and electrolyte determinations. A serum bicarbonate level lower than 18 mg/dL or an anion gap exceeding 10 to 15 mEq/L should prompt performance of an arterial blood gas analysis. In all cases of DKA, the diagnosis is confirmed by arterial blood gases demonstrating a metabolic acidemia with base excess exceeding −4 mEq/L.

Table 59.4 contains a protocol for treatment of DKA. The important steps in management include the following:

• Search for and treat the precipitating cause. Typical initiators include pyelonephritis and pulmonary or gastrointestinal viral infections.

• Perform volume resuscitation that is both vigorous (3 to 4 L of physiologic intravenous fluid over the first 2 hours) and sustained (a total of 6 to 8 L is frequently required over the first 24 hours). The patient will continue to generate vascular volume deficits until her glucose levels and acidosis are largely resolved. A physiologic fluid such as 0.9% NaCl or lactated Ringer solution should be used and continued until the acidosis is substantially corrected. Potassium chloride should be added to the infusate when the plasma potassium level nears the lower limit of normal.

• Use insulin to correct hyperglycemia. Although intermittent injections may be used, a continuous infusion of regular or short-acting insulin (i.e., lispro or aspart) allows frequent adjustments. When given as a continuous infusion, insulin 1 to 2 U/h gradually corrects the patient’s glucose abnormality over 4 to 8 hours. Attempts to normalize plasma glucose levels rapidly (i.e., in less than 2 to 3 hours) may result in hypoglycemia and physiologic counter-regulatory responses.

• Monitor serum bicarbonate levels and arterial blood gas base deficits every 1 to 3 hours to guide management. Even after the plasma glucose level is normalized, acidemia may persist, as evidenced by continuing abnormalities in the patient’s electrolyte concentrations. Unless volume therapy is continued until the patient’s electrolyte stores and plasma concentrations have substantially returned to normal, DKA may reappear, and the cycle of metabolic derangement will be renewed.

If DKA occurs after 24 weeks’ gestation, the status of the fetus should be continuously monitored by fetal heart rate monitoring or a biophysical profile or both. However, even if the fetal status is questionable during the phase of therapeutic volume and plasma glucose correction, emergency cesarean delivery should be avoided. Usually, correction of the maternal metabolic disorder is effective in normalizing fetal status. Nevertheless, if a reasonable effort has been expended in correcting the maternal metabolic disorder and the fetal status remains a concern, delivery should not be delayed if the maternal condition is stable.

Fetal Morbidity and Mortality Associated With Diabetes

The rate of perinatal mortality in diabetic pregnancy has decreased 30-fold since the discovery of insulin in 1922 and the institution of intensive obstetric and infant care in the 1970s. Improved techniques of maintaining maternal euglycemia have led to later timing of delivery and reduced iatrogenic respiratory distress syndrome (RDS). Nevertheless, the perinatal mortality rates reported for diabetic women remain approximately three times those observed in the nondiabetic population, with congenital malformations, RDS, and extreme prematurity accounting for most deaths. Recognizing these risks, many experts have recommended early delivery (at 37 to 38 weeks’ gestation) to optimize fetal outcome.

However, in 2011, SMFM and the Eunice Kennedy Shriver National Institutes of Child Health and Human Development held a conference addressing the issues of late preterm and early term birth. If GDM is well controlled on diet or pharmacologic therapy and the estimated fetal weight is less than the 90th percentile, one need not consider early term delivery. If GDM or preexisting diabetes is complicated by conditions that increase the risk for the mother (e.g., severe preeclampsia) or the fetus (nonreassuring fetal status), the timing of delivery should be based on the underling medical or obstetric condition. If a woman with GDM or preexisting diabetes has poor glycemic control despite reasonable efforts, consideration should be given to late preterm or early term delivery after evaluation of the risks and benefits of continuing the pregnancy. Even among women with excellent glycemic control, delivery at term increases the risk of fetal demise, fetal macrosomia, and birth injury. Additional prospective data are needed to better understand the short- and long-term risks and benefits of this obstetric dilemma.