The American College of Obstetricians and Gynecologists (ACOG) defines GDM to be carbohydrate intolerance with onset or recognition during pregnancy [21]. Risk factors for GDM include obesity, family history of DM, age > 25 years old, ethnicity (African-American, Hispanic, Native American, Asian, and Pacific Islander), previous delivery of a baby weighing over 9 pounds, polycystic ovarian syndrome, and personal history of glucose intolerance. Given the prevalence of risk factors in most populations in the United States, universal screening has been generally adopted. ACOG recommends a 2-step approach for the diagnosis of GDM. If the diabetic screen (a 50-g, 1-hour glucose challenge test) performed at 24–28 weeks’ gestation results in a value of ≥140 mg/dl, a 100-g, 3-h diagnostic oral glucose tolerance test is recommended (see Table 4.4). An early diabetic screen at 12–14 weeks is recommended in women who are at increased risk for pregestational diabetes such as those with obesity, strong family history of diabetes, or a previous history of GDM. If its result is normal, screening should be repeated at 24–28 weeks’ gestation.

Other professional organizations offer other diagnostic criteria for diabetes in pregnancy [22]. The International Association of Diabetes and Pregnancy Study Group (IADPSG) and American Diabetes Association (ADA) accept the following criteria for overt diabetes in pregnancy: fasting plasma glucose ≥ 126 mg/dL, A1c ≥ 6.5 %, or random plasma glucose ≥ 200 mg/dL that is subsequently confirmed by elevated fasting glucose or A1c [23, 24]. Both organizations also recommended diabetes screening by employing a single 2-h, 75-g glucose tolerance test, with diagnostic thresholds noted in Table 4.4. This approach has been calculated to increase the number of gravidas diagnosed with gestational diabetes from 5–6 % to 15–20 % of pregnant women [5]. Given this, an NIH consensus development conference and ACOG have recommended continued use of the 2-step approach described above because of concern that the adoption of the new criteria “would increase the prevalence of GDM, and the corresponding costs and interventions, without clear demonstration of improvements in the most clinically important health and patient-centered outcomes” [25].

Maintaining euglycemia is paramount to mitigating the risks associated with hyperglycemia during pregnancy. In particular, treatment reduces risk for macrosomia, shoulder dystocia, and preeclampsia [26]. Among 25,000 gravidas in a worldwide prospective cohort study of gravidas without diagnosed diabetes, midpregnancy maternal glucose levels had a linear positive association with fetal birth weight, without evident threshold effect. In the United States, the ADA and ACOG recommend the goals of fasting and 2-h postprandial glucose concentrations of <95 mg/dL and <120 mg/dL [21, 27].

Among patients with newly diagnosed diabetes in pregnancy, we begin a regimen of glucose measurements in the fasting state and 2 h after each main meal. Concurrently, patients are provided dietary counsel to establish reduced carbohydrate intake (40 % of calories from carbohydrate, 20 %–30 % from protein, and 30–40 % from fat in 3 meals with planned snacks) in a stable diet as to timing, amount, and type of food. After 2–7 days of observation, those patients found to maintain normal fasting glucose values are encouraged to continue to pursue dietary change to lower postprandial values to target range. In the absence of obstetric contraindications, moderate physical activity is safe during pregnancy and may improve glucose control by enhancing insulin sensitivity.

For most women with GDM, these interventions will be sufficient without pharmacologic treatment. However, women with preexisting diabetes and some with GDM will require medication to achieve adequate glucose control. Insulin does not cross the placenta and its use offers two other advantages over oral agents during pregnancy. First, insulin treatment can be tailored to the timing of meals to minimize risk of hypoglycemia while achieving a euglycemic profile. Second, insulin treatment can generally achieve an acceptable euglycemic profile within two weeks even in an initially poorly controlled diabetic patient, a much shorter interval than is generally required by treatment with glyburide. A typical regimen includes either long- or intermediate-acting insulin in combination with short acting. NPH insulin and a short-acting insulin (regular (crystalline zinc) insulin or insulin lispro or insulin aspart) are prescribed prior to breakfast and the evening meal. Occasionally the evening dose of NPH insulin is moved to bedtime to more effectively treat fasting hyperglycemia. Limited data suggests that insulin detemir and insulin glargine may be safely used during pregnancy, but randomized trials comparing these agents to NPH are not available. However, if a woman has well-controlled pregestational diabetes on one of these agents and becomes pregnant, it is reasonable to continue it during pregnancy. Table 4.2 outlines the pharmacokinetics of different insulins.

The long-term effects of fetal exposure to sulfonylureas or metformin among offspring are unknown. Whether glyburide crosses the placenta in significant amounts is controversial, but it has not been shown to be associated with adverse fetal outcomes when compared to insulin. However, practical and theoretical limitations to its use are relevant to diabetes management. In one study, approximately 20 % of pregnant women with fasting glucoses over 110 mg/dL treated with glyburide failed to reach targeted glycemic control and ultimately required treatment switched to insulin [28]. Further, despite the attractive anticipation that frequent insulin adjustments can be avoided with glyburide use, it can cause hypoglycemia because of the inevitable mismatch between meal schedules and its pharmacodynamic effects. Metformin does cross the placenta but has not been shown to be associated with adverse perinatal outcomes. Up to 35 % of gravidas treated with metformin require insulin treatment to achieve adequate control, however [29]. We find that as compared with oral agents, insulin dosing can more nimbly respond to changing requirements due to increasing insulin resistance as pregnancy progresses and changes in diet due to early nausea, gastroparesis, or the patient’s adoption of recommended diet. While this approach is more laborsome than using oral agents, we employ a diabetes nurse-educator in reviewing glucose data with patients at frequent intervals and find that most GDM and pregestational diabetic patients achieve euglycemic targets within 2 weeks of initiating treatment.

The interface between internist and obstetric offices is based on the expressed needs of obstetrical prenatal care providers, who must remain the primary site of care for pregnant women. This often requires the internist to actively define their services in a dialog with the obstetrical staff and physicians. The delegation of internist duties to the patient should be explicitly documented in initial correspondence from either party to provide the framework for subsequent patient education, treatment, and reciprocal reporting by both parties during and after pregnancy. Often, a third party physician, one who has been the primary prepregnancy caregiver, should participate in establishing roles as well. In our own practice, after initial conversations, we document our own understanding of each provider’s role, while requesting any modification advised by the other party. We find that initial education and treatment of the diabetic gravida means more frequent visits with the obstetric internist. This usually shifts to a more frequent interaction with the obstetrician as the pregnancy progresses. We provide the referring obstetrician with documentation of each visit with the internal medicine office.

Office visits are opportunities to assess glucose control face-to-face with patients. As observed by others [30], we frequently find the reported glucose values and those recorded in the memory of glucose meters to be discordant. Review of glucose data from the patient’s meter also may provide insight into erratic timing of meals and lapses in glucose testing that allows for a more reality-based discussion of impediments to better glucose control. In order to develop a durable trust relationship with patients, we inform patients at the start of treatment that we will be reviewing glucose meter data at each office visit. We also emphasize that we view our therapeutic role as supportive and recognize the many physical, logistical, and psychological impediments to achieving glycemic control. Coexisting conditions such as hypertension, retinopathy, and nephropathy are addressed at these visits. After 20 weeks’ gestation, evidence for possible superimposed preeclampsia should be sought. Tables 4.3 and 4.4 outline management steps from prior to and during pregnancy.

Unless requested specifically by the referring obstetrical care office, we do not address issues of pregnancy viability and dating, screening for fetal disorders, and surveillance of fetal health and growth during pregnancy, as these issues are the general purview of obstetrical care. As pregnancy progresses, antenatal testing for fetal well-being and appropriate growth will be instituted. Types of tests of fetal well-being and the scheduling of fetal growth assessments will depend on individual pregnancy concerns. Fetal growth assessment at 28 gestational weeks’ finding fetal abdominal circumference values in the upper quartile for gestational age increases the risk of fetal macrosomia at birth. More aggressive treatment of maternal hyperglycemia in this group has been shown to reduce the incidence of birth macrosomia [31].

The risk for fetal acidosis and neonatal hypoglycemia has been associated with maternal hyperglycemia during labor and delivery. Management of maternal glucose concentrations among gravidas treated with insulin involves several considerations. Active labor is estimated to produce up to twice resting energy expenditure. Thus, labor is associated with insulin-independent glucose utilization that may appreciably reduce maternal insulin requirements in the fasting patient. ACOG recommends maintaining glucoses between 70 and 110 mg/dL during labor, though some authors suggest higher levels may be acceptable [32, 33]. Accomplishing this requires frequent glucose monitoring and, for some patients, insulin infusion. Type 1 diabetic patients require constant insulin infusion during labor, but often at less than one unit each hour during active labor. In such patients, we recommend that an insulin solution and a separate glucose solution be hung as ancillary to the main crystalloid line containing either lactated Ringers or normal saline solutions. We do not recommend subcutaneous insulin treatment during labor because of the unpredictable insulin requirements during this time. A suggested protocol for the management of insulin during labor and delivery is illustrated in Table 4.5.