Management of type 1 and type 2 diabetes mellitus during pregnancy

Pregestational diabetes mellitus complicates approximately 0.2 percent to 0.5 percent of pregnancies. As an example, a nationwide population-based survey revealed that nearly 4 percent of pregnant women in the United States have diabetes: 88 percent had gestational diabetes mellitus, while the remaining 12 percent were women known to have diabetes. Of those with pregestational diabetes, 35 percent had type 1 and 65 percent type 2 diabetes. However, as pregnancy is deferred until the later childbearing years and population demographics change, more pregnancies can be expected to occur in women with type 2 diabetes.

Pregnancy in diabetes is associated with an increase in risk to both the fetus and the mother. There is, for example, an increased incidence of congenital anomalies and spontaneous abortions in women who are in poor glycemic control during the period of fetal organogenesis, which is nearly complete at seven weeks postconception. Thus, a woman may not even know she is pregnant at this time. It is for this reason that prepregnancy counseling and planning are essential in women of child-bearing age who have diabetes.

Women with type 2 diabetes are less likely to have preconception care and counseling, often because the diagnosis of diabetes has not yet been made. In one study, 40 percent of women with type 1 diabetes but only 14 percent of those with type 2 diabetes received preconception care.

Another problem for some women with type 2 diabetes is gaining access to adequate medical care before and after conception. Nowhere is this more evident than in Hispanic women, in whom the higher prevalence of obesity, type 2 diabetes and gestational diabetes, and higher pregnancy rates place them at higher risk for maternal and neonatal complications. Pima Indian women are also at increased risk; they have a 19-fold higher incidence of type 2 diabetes compared with the general United States population.

GLYCEMIC CONTROL AND FETAL AND MATERNAL COMPLICATIONS

Hemoglobin A1c (HbA1c) values provide the best assessment of the degree of chronic glycemic control, reflecting the average blood glucose concentration during the preceding six to eight weeks. As a result, measurement of HbA1c can, in early pregnancy, estimate the level of glycemic control during the period of fetal organogenesis. There are two important observations in this regard:

• HbA1c values early in pregnancy are correlated with the rates of spontaneous abortion and major congenital malformations (ie, diabetic embryopathy). Although most studies have been performed in women with type 1 diabetes, the same risk of hyperglycemia applies to those with type 2 diabetes

• Normalizing blood glucose concentrations before and early in pregnancy can reduce the risks of spontaneous abortion and congenital malformations nearly to that of the general population.

One report, for example, compared 110 women who were already six to 30 weeks pregnant at the time of referral with 84 women recruited before conception and then put on a daily glucose-monitoring regimen. The mean blood glucose concentration was between 60 and 140 mg/dL (3.3 and 7.8 mmol/L) in 50 percent of the latter women. The incidence of anomalies was 1.2 percent in the women recruited before conception versus 10.9 percent in those first seen during pregnancy. Very similar findings were noted in another study: 1.4 percent versus 10.4 percent incidence of congenital abnormalities. Major congenital malformations, which either require surgical correction or significantly affect the health of the child, are more common in infants of diabetic mothers (13 percent versus 2 percent in infants of nondiabetic mothers).

The increased rate of spontaneous abortion in women with poorly controlled diabetes is thought to be secondary to hyperglycemia, maternal vascular disease including uteroplacental insufficiency, and possibly immunologic factors. In addition, animal studies suggest that hyperglycemia increases the expression of an apoptosis (programmed cell death) promoting gene as early as the preimplantation blastocyst stage, resulting in increased DNA fragmentation.

Another consequence of poor glycemic control in pregnant women with diabetes is fetal macrosomia (the major manifestation of diabetic fetopathy), which leads to difficult deliveries, an increased need for cesarean delivery, and an increase in fetal morbidity. The macrosomia is caused by fetal hyperinsulinemia, the proximate causes of which are maternal and thus fetal hyperglycemia, both of which are circumvented by good glycemic control. In a study in which which 95 women with type 1 diabetes were divided retrospectively into groups based upon mean daily blood glucose concentrations during the second and third trimesters, the incidence of large-for-gestational age infants was substantially higher in those with high blood glucose concentrations (54 versus 5 percent in those with near-normal mean blood glucose values) as was overall fetal morbidity.

These findings emphasize the importance of glycemic control both before conception and throughout pregnancy.

Deleterious effects of strict glycemic control

Despite the clear benefits to the fetus of strict glycemic control, there are two potential hazards in the mothers: hypoglycemia and some worsening of diabetic retinopathy. Major complications of hypoglycemia can usually be prevented with careful monitoring and education of the mother.

Although maintenance of normoglycemia ultimately slows the development and progression of diabetic retinopathy, it may worsen after restoration of normoglycemia in both nonpregnant and pregnant women, with a particular increase in the formation of soft exudates.The degree to which this occurs during pregnancy is related to the baseline level of retinal disease and to the reduction of chronic hyperglycemia. The Diabetes in Early Pregnancy study found that, among 140 women who did not have proliferative retinopathy at the time of conception, progression of retinopathy occurred in 10 percent of those who had no retinopathy, 21 percent of those with mild background retinopathy, and 55 percent of those with severe nonproliferative retinopathy. This worsening is thought to be mediated by closure of small retinal blood vessels which were narrowed but patent.

Very strict glycemic control (mean blood glucose < or =86 mg/dL [4.8 mmol/L]) may be deleterious to the fetus and should be avoided. In one study, the overall incidence of small-for-gestational age infants was higher among women with very strict as compared with strict control, 20 percent versus 11 percent in the latter.

TREATMENT

There are several components to the treatment of diabetes in pregnant women: careful monitoring of blood glucose; administration of insulin; and diet. Attention must also be paid to obstetric surveillance and treatment during labor and delivery.

Glucose monitoring

Frequent measurements of blood glucose are mandatory in women with type 1 diabetes during pregnancy. If the first morning blood glucose value is high, testing should also be performed at bedtime and in middle of the night. Bedtime and middle-of-the-night tests are important to discover, treat, and prevent nocturnal hypoglycemia. Pregnancy is associated with an exaggerated rebound from hypoglycemia so that an elevation in morning blood glucose values may reflect hypoglycemia at night.

Urinary ketones should be measured periodically, especially when the woman is ill or when any blood glucose value is over 200 mg/dL (11.1 mmol/L). At these times ketoacidosis may occur, a complication that is associated with a high mortality rate in the fetus. In addition, ketonemia during pregnancy has been associated with decreased intelligence in offspring. In early pregnancy ketonuria sometimes occurs in women who are limiting their caloric intake because of nutritional recommendations. Women with moderate to large ketonuria should alert their physician immediately.

Target blood glucose values

We recommend that HbA1c be measured every four to six weeks and more frequently if the woman's glycemic control is poor. Both average blood glucose concentration and the HbA1c values fall by about 20 percent in normal pregnant women, and similar values should be aimed for in diabetic women to minimize fetal risk. Our blood glucose goals in a pregnant diabetic woman are:

• A fasting capillary blood glucose concentration of 55 to 65 mg/dL (3.1 to 3.6 mmol/L), about 85 percent for the venous plasma concentration.

• One-hour postprandial blood glucose concentration less than 120 mg/dL (6.7 mmol/L).

These target values are somewhat lower than those proposed by the American College of Obstetricians and Gynecologists and the American Diabetes Association :

• Fasting plasma glucose concentrations of 60 to 90-95 mg/dL (3.3 to 5.0-5.3 mmol/L)

• One-hour postprandial plasma glucose concentrations no higher than 130 to 140 mg/dL (7.2 to 7.8 mmol/L)

• Two-hour postprandial plasma glucose concentrations no higher than 120 mg/dL (6.7 mmol/L)

The effect of strict glycemic control on fetal outcome in pregnant women with diabetes has been demonstrated in several studies. One study, for example, compared the results of monitoring fasting blood glucose plus either preprandial or postprandial blood glucose monitoring in women with gestational diabetes. The postprandial monitoring was associated with the following benefits, as compared with preprandial monitoring:

• Better glycemic control (mean HbA1c value 6.5 versus 8.1 percent).

• A lower incidence of large-for-gestational age infants (12 versus 42 percent).

• A lower rate of cesarean section for cephalopelvic disproportion (12 versus 36 percent).

Diet

The optimal diet takes into account caloric intake, carbohydrate content, and distribution of meals throughout the day. The appropriate calorie intake depends upon the pregravid weight, with the following general recommendations:

• Approximately 30 kcal/kg per day if the woman is at ideal body weight.

• 24 kcal/kg per day if 20 to 50 percent above ideal body weight.

• 12 to 18 kcal/kg per day if more than 50 percent above ideal body weight.

• 36 to 40 kcal/day if more than 10 percent below ideal body weight.

The recommended distribution of calories is 40 to 50 percent carbohydrate, 20 percent protein, and 30 to 40 percent fat. Many physicians find that maintenance of glycemic control requires a diet in which carbohydrate accounts for no more than 40 percent of calories. Postprandial blood glucose values are largely dependent on the carbohydrate content of the meal, and it is the postprandial blood glucose concentration that has the most important role in macrosomia.

Most programs recommend three meals and three snacks per day. An acceptable calorie distribution would be 10 percent of calories at breakfast, 30 percent at both lunch and dinner, and 30 percent as snacks. A daily supplement of ferrous sulfate (30 mg) and folate (400 µg) is also recommended.

Insulin regimen

Most women with type 1 diabetes require at least three injections of insulin per day. A two-injection regimen can cause nocturnal hypoglycemia if the action of evening-meal dose of intermediate-acting insulin is maximal in the middle of the night.

Total daily insulin requirements usually vary considerably during gestation. After an early rise in insulin requirements between weeks 3 and 7, there often is a significant decline between weeks 7 and 15, followed by a rise during the remainder of pregnancy.

The average insulin requirement in pregnant women with type 1 diabetes is 0.7 units/kg in the first trimester, often increasing to 0.8 U/kg for weeks 18 to 26, 0.9 U/kg for weeks 27 to 36, and 1.0 U/kg for weeks 37 to term. However, the range of change in insulin requirements is broad. In one study of 237 pregnancies in women with type 1 diabetes, the mean daily insulin requirement increased by 52 units. A large maternal weight gain was associated with a greater increase in the insulin requirement. Massively obese women may need initial doses of 1.5 to 2.0 units/kg to overcome the combined insulin resistance of pregnancy and obesity.

Women with type 2 diabetes also should be treated with insulin for blood glucose control, preferably started during the preconception period. During the first trimester, insulin requirements are similar in women with type 1 and type 2 diabetes. However, as the pregnancy proceeds into the third trimester, insulin requirements increase proportionately more in women with type 2 than type 1 diabetes; in one study, for example, the respective insulin doses were 1.6 and 1.2 U/kg per day. Measures that ameliorate insulin resistance, such as avoiding excessive weight gain and moderate low-impact exercise, can be expected to improve glycemic control.

We administer a combination of regular insulin and intermediate-acting insulin (such as NPH insulin); long-acting insulin (ultralente) is avoided since it acts too slowly to respond to the ever-changing insulin needs of pregnancy. The insulin is initially distributed as follows:

• Approximately 45 percent of the total daily dose is given as NPH insulin and 22 percent as regular insulin before breakfast.

• Approximately 17 percent of the total daily dose is given as both NPH and regular insulin before dinner.

• The premeal dose of regular insulin (including lunch) is given on a sliding scale according to the blood glucose value.

Some clinicians have recommended insulin pumps to achieve optimal glycemic control during pregnancy. Most pregnant women require at least two infusion rates in a 24-hour period, in particular an increased rate in the early morning hours to counteract the increased release of the anti-insulin hormones cortisol and growth hormone. A protocol for calculating infusion rates has been described.

Oral hypoglycemic drugs should not be given to pregnant women with established diabetes because the drugs may have teratogenic effects in early pregnancy and they do not provide strict glycemic control. When metformin is used to induce ovulation in patients with the polycystic ovary syndrome and insulin resistance, it should be stopped as soon as pregnancy is diagnosed. However, small studies have not shown harmful effects from metformin in either early or late pregnancy.

Role of exercise

Gestational diabetes differs from a pregnancy in type 1 diabetes since the former is primarily a disorder of impaired glucose clearance; as a result, therapies that overcome peripheral resistance to insulin, such as exercise, are preferable to insulin. In comparison, in women with type 1 diabetes who are already taking insulin the benefits of exercise are not so clear. Exercise can contribute to the "brittleness" of diabetes, with the risk of exercise-induced hypoglycemia. Women who exercised before pregnancy can usually continue under the supervision of their obstetrician. However, exercise is not recommended in women who are deconditioned and did not exercise before pregnancy. There are also several contraindications to exercise.

OBSTETRIC SURVEILLANCE

The high perinatal mortality once associated with a diabetic pregnancy has decreased significantly, largely due to improved glycemic control. In the past, unexplained fetal death occurred in 10 to 30 percent of type 1 diabetic pregnancies. It typically occurred after the 36th week of gestation in women with poor glycemic control, associated with macrosomia, polyhydramnios, preeclampsia, and vascular disease. Fetal surveillance is therefore of utmost importance in increasing the likelihood of a good outcome for both mother and fetus, especially in the perilous third trimester.

Ultrasonography

Ultrasonography is the most useful tool for the assessment of the fetus. It can be used to:

• Estimate gestational age

• Screen for structural anomalies

• Evaluate growth

• Assess amniotic fluid volume

• Determine fetal status dynamically through Doppler and biophysical studies.

Ultrasound estimates of gestational age are most accurate if performed in early pregnancy; gestational age determined by a crown-rump length measured in the first trimester will be accurate within five days. Ultrasound estimates of gestational age are not reliable after the 28th week and cannot be used to determine the estimated date of delivery.

Macrosomia is more apparent in some fetal structures, such as the liver and abdomen. As a result, ultrasound estimates that utilize the fetal abdominal circumference to calculate fetal weight may not be as accurate as in normal fetuses in whom growth is proportional. Macrosomia is usually defined as fetal weight greater than 4.0 to 4.5 kg or birth weight above the 90th percentile for gestational age. Macrosomic fetuses are at increased risk for a prolonged second stage of labor, shoulder dystocia, operative delivery, and perinatal death.

Macrosomia occurs in approximately 88 percent of fetuses in whom the abdominal circumference and estimated fetal weight both exceed the 90th percentile. The biparietal diameter and head circumference appear to be less predictive of macrosomia.

In contrast to macrosomia, intrauterine growth restriction (usually defined as birth weight below the 10th percentile for gestational age) is uncommon in diabetic pregnancies. It is associated with uteroplacental insufficiency and primarily occurs in pregnancies complicated by diabetic vasculopathy or preeclampsia.

Congenital anomalies

Ultrasonography is essential for the evaluation of congenital anomalies. Infants of diabetic mothers are at increased risk for neural tube defects, which occur in approximately two percent of diabetic pregnancies (versus 0.1 to 0.2 percent in the general population). Other congenital anomalies that occur with higher frequency in infants of diabetic mothers include anencephaly, microcephaly, caudal regression syndrome, and genitourinary and gastrointestinal anomalies. Congenital heart disease is also common, including a specific type of hypertrophic cardiomyopathy, atrial and ventricular septic defects, transposition of the great vessels and coarctation of the aorta. Polyhydramnios can occur because of increased amniotic fluid osmolality and polyuria secondary to fetal hyperglycemia.

Risk factors for congenital anomalies include a high HbA1c value at the first prenatal visit, high serum alpha-fetoprotein concentrations, and a suggestive past medical or family history of congenital anomalies. In these women, ultrasonography should be performed to look for anomalies.

The maternal serum alpha fetoprotein value is expressed in multiples of the median (MOMs).

• A high value, corrected for maternal diabetes, should be confirmed and an ultrasound examination to rule out multiple gestation, fetal demise, or inaccurate gestational age (false-positives) should be done. Amniocentesis is performed if more definitive diagnosis is required. Amniotic fluid alpha-fetoprotein and acetylcholinesterase concentrations are high is fetuses with a neural tube defect.

• Lower values (by about 15 percent) for gestational age can occur in diabetic women and are thought to result from a growth lag in the first trimester or from poor glycemic control. Low values have also been associated with chromosomal aneuploidy (such as trisomy 21).

Maternal diabetes alone does not increase the risk for chromosomal abnormalities such as Down's syndrome. As a result, the indications for invasive testing, such as amniocentesis and chorionic villous sampling, are the same as in the general population.

Antepartum surveillance

Guidelines for antepartum surveillance vary and are usually dependent upon the clinical situation and the discretion of the physician. In women with diet-controlled gestational diabetes, fetal surveillance is usually not initiated until 40 weeks gestation, since these women are at very low risk for complications. More rigorous monitoring is recommended for women who have additional indications for closer fetal surveillance, such as hypertension.

Surveillance begins earlier in women with either gestational or pregestational diabetes treated with insulin. Most centers defer testing until the 35th week of gestation if there is excellent glycemic control, but start much earlier in women with poor control, nephropathy, or hypertension. In these women, antepartum testing is begun at 26 to 28 weeks, when fetal survival is likely if delivery were to occur. Antepartum fetal testing should be performed twice per week. Doppler umbilical artery velocimetry has shown increased placental resistance in women with vasculopathy and poor glycemic control, which increase the risk of intrauterine growth restriction and preeclampsia.

LABOR AND DELIVERY

Maternal hyperglycemia is the major cause of neonatal hypoglycemia. As a result, peripartum maintenance of maternal euglycemia is essential. The following general recommendations can be made:

• Insulin is still required before active labor, and can be given subcutaneously or by intravenous infusion with a goal of maintaining blood glucose concentrations between 70 and 90 mg/dL (3.9 and 5.0 mmol/L). One method of insulin infusion consists of intravenous administration of 15 units of regular insulin in 150 mL of normal saline at a rate of one to three units per hour.

• Normal saline may be sufficient to maintain euglycemia when labor is anticipated.

• As the mother enters active labor, insulin resistance rapidly decreases (since expulsion of the fetoplacental unit leads to cessation of production of placental growth hormone and placental lactogen, which have short half-lives) and insulin requirements fall rapidly. Thus, continuing insulin therapy is likely to lead to hypoglycemia. To prevent this, glucose should be infused at a rate of 2.5 mg/kg per min [58]. Capillary blood glucose should be measured hourly. The glucose infusion should be doubled for the next hour if the blood glucose value is less than 60 mg/dL (3.3 mmol/L). On the other hand, values of 120 mg/dL (7.8 mmol/L) or more require the administration of regular insulin subcutaneously or intravenously until the blood glucose value falls to 70 to 90 mg/dL (3.9 to 5.0 mmol/L). At this time, the insulin dose is titrated to maintain normoglycemia while glucose is infused at a rate of 2.5 mg/kg per min. Bolus doses of glucose should not be given because they can raise maternal blood glucose concentrations and increase the risk of neonatal hypoglycemia, fetal hypoxia, and fetal or neonatal acidosis.

If a cesarean section is planned, the bedtime NPH insulin dose may be given on the morning of surgery and every eight hours thereafter if surgery is delayed. Insulin requirements drop sharply after delivery, and the new mother may not require insulin for 24 to 72 hours. Insulin requirements should be recalculated at this time at approximately 0.6 units/kg per day based upon postpartum weight. Postpartum calorie requirements are approximately 25 kcal/kg per day, and somewhat higher (27 kcal/kg per day) in lactating women.

Glycemic control is somewhat more erratic in lactating diabetic women, with more frequent episodes of hypoglycemia.