Contents
Contributors
Search

Pregnancy is a complex metabolic state that involves dramatic alterations in the hormonal milieu (increases in estrogen, progesterone, prolactin, cortisol, human chorionic gonadotropin, placental growth hormone and human placental lactogen) as well as an increasing burden of fuel utilization by the conceptus. Metabolically, the first trimester is characterized by increased insulin sensitivity and accelerated starvation in some women, with an increased turnover of maternal metabolic fuels and an earlier transition from carbohydrate to fat utilization in the fasting state (1; 2). The second and third trimesters, in contrast, are characterized by insulin resistance with a nearly 50% decrease in insulin mediated glucose disposal (assessed by the hyperinsulinemic-euglycemic clamp technique) and a 200-300% increase in the insulin response to glucose (3; 4; 5; 6). This serves to meet the metabolic demands of the fetus, which requires 80% of its' energy as glucose, while maintaining euglycemia in the mother. The placental and fetal demands for glucose are considerable and approach the equivalent of ~150 grams per day of glucose in the third trimester (7; 8). In addition, the maternal metabolic rate increases by ~300 kcal/day in the third trimester. These increased nutritional needs place the mother at risk for ketosis which occurs much earlier than usual when without adequate oral or intravenous nutrients, frequently referred to as "accelerated starvation of pregnancy" (9; 2; 10). Glucose transport to the fetus occurs in direct proportion to maternal glucose levels, and is augmented by a five-fold increase in a placental glucose transporter, (GLUT-1) which increases transplacental glucose flux even in the absence of maternal hyperglycemia (11). At the same time it has been demonstrated that in the third trimester of normal pregnancy there is decreased expression of the GLUT-4 glucose transporter protein in maternal adipose tissue (12, 13) and decreased translocation of GLUT-4 to the plasma membrane in skeletal muscle, both of which contribute to the insulin resistance of pregnancy.

Human placental growth hormone (hPGH) has been recently characterized as a metabolically active hormone capable of causing severe insulin resistance in transgenic animals which express this hormone at levels comparable to those measured in the third trimester of pregnancy (14). This key hormone may mediate insulin resistance as does excess pituitary growth hormone (pit GH) when it is administered or expressed chronically. Human placental growth hormone differs from pit GH by 13 amino acids. It almost completely replaces pit GH in the maternal circulation by 20 weeks, and it is unregulated by growth hormone releasing hormone (15; 16) The placenta is responsible for the production of hormones which reprogram maternal physiology to become insulin resistant in the 2nd and 3rd trimester of pregnancy to ensure an adequate supply of nutrients to the growing fetus (17). Most probably, this is due to an increase in placental growth hormone (18; 14) in combination with human placental lactogen, progesterone, and possibly circulating TNFa, which correlates with maternal insulin resistance measured by hyperinsulinemic-clamps (19). Human placental lactogen may play a key role in stimulating insulin production in human islets (20; 21; 17) in order for the mother to increase her insulin secretion 2-3 fold.

Normal pregnancy is associated with lower fasting plasma glucose and fasting hyperinsulinemia presumably due to an increase in glucose uptake by the fetoplacental unit as well as a decrease in glycogenolysis and hepatic glucose production (1). Postprandial glucoses are modestly elevated associated with maternal postprandial hyperinsulinemia, especially in the second and third trimester. Glucose is not the only fuel altered in normal pregnancy. Amino acids are decreased, whereas triglycerides, cholesterol, and free fatty acids are increased; the latter may serve to further increase the insulin resistance of pregnancy (6).

Effect of Metabolic Changes on Diabetes Management Throughout Gestation

Diabetes should optimally be under tight control before conception. During the first trimester before the placenta increases the production of hormones, nausea, accelerated starvation, and increased insulin sensitivity (possibly by estriol) may place the mother at risk for hypoglycemia. Women must be counseled that their insulin requirements in the first trimester are likely to decrease by 10-20% (22). This is especially true at night when prolonged fasting and continuous fetal glucose utilization places the woman at even a higher risk for hypoglycemia. Women with Type 1 diabetes mellitus must have a bedtime snack and usually need to have their evening dose of NPH insulin lowered and moved from suppertime to bedtime to avoid early morning hypoglycemia (23). The effect of short-term maternal hypoglycemia on the fetus is not well understood. During the first trimester, glycemic control just above the normal range (hemoglobin A1C < 7.0%) may thus be safer than "normal" and may decrease the risk of fetal hypoglycemia.

After 20 weeks of gestation, peripheral insulin resistance increases insulin requirements so that it is not unusual for a pregnant woman to require twice as much insulin as she did prior to pregnancy. If the woman becomes ill, her insulin requirements are likely to be even higher due to the addition of high counterregulatory hormones in the face of pre-existing insulin resistance of pregnancy. Diabetic ketoacidosis carries the highest risk of fetal mortality in the third trimester thought in part due to the extreme insulin resistance in these patients and insulin requirments to treat DKA that are nearly twice as high as in the second trimester (24). It has been demonstrated that postprandial hyperglycemia is the strongest risk factor for fetal macrosomia (25). Therefore, tight glucose control in women with Type 1 diabetes usually requires insulin administration with each meal with a short acting insulin preparation such as Lispro (26). Frequent monitoring allows appropriate insulin dosage adjustments. The maintenance of normal glucose control is the key to prevention of complications such as fetal malformations in the first trimester, macrosomia in the second and third trimesters, and neonatal metabolic abnormalities.

PRECONCEPTION COUNSELING AND PREPARING THE WOMAN WITH PRE-EXISTING DIABETES FOR PREGNANCY

Glycemic Control Objectives

Most important in preconception counseling is the message of optimal glucose control prior to conception (27; 28). In a retrospective study, < 40% of women attempted to achieve optimal glycemic control before becoming pregnant (29). Four times as many fetal and neonatal deaths and congenital abnormalities occurred in a group of women who did not receive prenatal counseling in comparison to those who did (30). Hyperglycemia is a known teratogen (31). The incidence of congenital abnormalities in offspring of diabetic mothers in the early era of insulin use was 33%. Over the past decade, with the advent of home blood glucose monitoring and more rigid objectives, this percentage has fallen to <10% of offspring (32; 33; 34). In one of the largest studies in 462 pregnancies in women with Type 1 DM from 10 centers over a 5 year period in England, there were 76% live births, 17% pregnancy losses, and 2% still births (five times the rate of the normal population). The congenital malformation rate was nearly 10% (94/1000 versus 9.7/1000) and directly related to the Hemoglobin A1C. (35). Epidemiologic and prospective studies have shown that the level of HgbA1C in the 6 months before conception and during the first trimester correlates with the incidence of major malformations such as neural tube and cardiac defects as well as spontaneous abortions. Hyperglycemia in the mouse will modulate the expression of an apoptosis regulatory gene as early as the preimplantation blastocyst stage, and this can be prevented by treating with insulin (36). These experimental results may help to explain the high spontaneous abortion rate in women with poorly controlled diabetes. The neural tube is completely formed by 4 weeks and the heart by 6 weeks after conception; many women do not even know they are pregnant at these times. It has been demonstrated that women with a normal HgbA1C at conception and during the first trimester have no increased risk while women with a HgbA1C of 11-12% or a fasting blood glucose >260 have up to a 25% risk of major malformations (28;37). The randomized, prospective Diabetes and Complications Trial has shown that timely institution of intensive therapy for blood glucose control is associated with rates of spontaneous abortion and congenital malformations that are similar to those in the nondiabetic population (38). In 270 pregnancies studied, the hemoglobin A1C at conception was 7.4% in the intensively treated mothers and 8.1% in the conventional group but similar during later gestation (6.6%) in both groups. Nine congenital malformations were identified (4.7%), eight which occurred in women originally assigned to conventional therapy.

Oral Hypoglycemic Agents

Oral hypoglycemic agents such as sulfonylureas and metformin do not appear to be teratogenic, although there is a concern about the risk of fetal and neonatal hypoglycemia or lactic acidosis if pregnant women take these medications (39). Small trials have recently been conducted using Metformin pre-conception and throughout the first trimester in women with polycystic ovary disease to improve fertility and prevent early miscarriage (40; 41). In one trial, Metformin was continued through pregnancy in 65 women (41). In the metformin group, 62 pregnancies resulted in live births. Of these, 53 were term deliveries and 8 were preterm. All babies were normal neonates with appropriate size for gestational age. One baby demonstrated achondrodysplasia. However, metformin, unlike glyburide, crosses the placenta and until more definitive experience in pregnancy is obtained, it should be discontinued and replaced with insulin if treatment is necessary in the first trimester (42). In a cohort of 50 pregnant women treated with metformin, 68 women treated with a sulphonylurea and 42 women treated with insulin, there was an increase in preeclampsia in the metformin group and an increase in perinatal mortality (39). The only possible exception to its continued use in the first trimester may be in women with polycystic ovarian disease with recurrent fetal losses where it has been suggested that continuing its use during the first trimester may decrease early pregnancy losses (41; 40). However, the woman should be counseled that such treatment is off label and not approved for use in pregnancy.

Glyburide has recently been shown in a randomized controlled trial of 400 women not to cross the placenta or significantly effect fetal insulin levels. However exposure was limited to after 24 weeks gestation so the effect on embryogenesis was not studied (43). It is also not approved for use in pregnancy, however its use is increasing in women with gestational diabetes. There are no sufficient data available on any of the new thiazolidinediones and exposure during the first trimester. Accordingly, it is recommended that oral hypoglycemic agents be avoided during pregnancy, with the possible exception of glyburide, which should be limited to the late second and third trimester of pregnancy if used. However if a woman conceives while taking these agents, they should simply be replaced by insulin. Women with Type 2 diabetes who are actively trying to become pregnant should be switched to insulin during the preconception period because it may take some time to determine the ideal insulin dose prior to the critical time of embryogenesis.

Risks of Pregnancy Complications

Women who are taking ACE-inhibitors should be counseled that these agents are contraindicated in the second and third trimesters of pregnancy because of the risk of fetal anuria, which may be irreversible (44). Although first trimester exposure alone has not been shown to cause problems, women who are actively trying to conceive and who have no history of infertility should probably be switched to a safer agent before pregnancy (calcium channel blocker, methyldopa, hydralazine (45). A woman who is being treated with an ACE-inhibitor for significant diabetic nephropathy and who is not actively trying to conceive should be told to stop her ACE-inhibitor as soon as she misses a period and to obtain a pregnancy test (46). At that time she can be switched safely to an alternative agent. Although mild renal disease does not seem to be accelerated by pregnancy, women with more severe renal disease are at a very high risk of pregnancy complications and progression of their renal disease (47; 48). Therefore, women with diabetic nephropathy should be counseled to have their children when their diabetes is optimally controlled and preferably early in the course of their nephropathy. Preeclampsia complicates ~20% of pregnancies in women with Type 1 diabetes (49). The rate of preeclampsia ranges from 9%-66%, the highest rates in women with increased severity of diabetes by White's classification and especially in women with diabetic nephropathy (proteinuria >300 mg) or who are hypertensive (49). Unfortunately, there is no known effective therapy to prevent preeclampsia. Although antihypertensive therapy should be instituted to maintain BP at <130-140/80-90 in order to protect the maternal kidney, it does not prevent preeclampsia, nor does baby aspirin (49). Women with longstanding Type 1 DM, especially those with nephropathy or hypertension, should be told that there is a substantial risk for preeclampsia complicating the pregnancy and this can lead to growth restriction, preterm delivery, and fetal lung immaturity in the infant (46).

Proliferative retinopathy may also progress during pregnancy; it is imperative, therefore, that this condition be optimally treated with laser therapy prior to pregnancy. The rapid institution of tight control in the face of increased volume expansion, anemia, placental angiogenic growth factors, and the hypercoagulable state of pregnancy may accellerate disease in women with moderate to severe retinopathy (50). Optimal glycemic control and treatment of retinopathy prior to conception are of paramount importance.

Women with Type I diabetes have a 5-10% risk of developing autoimmune thyroid disease first diagnosed in pregnancy (usually Hashimoto's thyroiditis). TSH should be checked prior to pregnancy since the fetus is completely dependent on maternal thyroid hormone in the first trimester (27). Because of the high morbidity and mortality of coronary artery disease in pregnancy, women with multiple cardiac risk factors such as hyperlipidemia, hypertension, smoking, advanced maternal age (>35) or a strong family history should have their cardiac status assessed with functional testing prior to conception (33) Lipid lowering agents should be discontinued before conception since there is inadequate data about their safety during pregnancy. However, if a woman has severe hypertriglyceridemia, which places her at high risk for pancreatitis, it may be necessary to continue fibrate therapy if a low fat diet, fish oils, or niacin therapy is not effective or tolerated. All women should also be taking folic acid supplements (1 mg per day) before conception.

Offspring of women with Type 1 DM have a risk of developing Type 1 DM of about 1-3% (32). The risk is higher to the offspring if the father has Type 1 DM rather than the mother (~6%) and if both parents have Type 1 diabetes, the risk if ~20%. The infant can be tested with HLA typing and a battery of antibodies associated with Type 1 diabetes to better predict the risk of developing Type 1 diabetes later in childhood (51).

Smoking continues to be the leading cause of low birth weight infants in patients with and without diabetes and places the infant at increased risk for respiratory infections, reactive airway disease, and sudden infant death syndrome. Smoking cessation efforts need to be intensified before conception since agents such as the nicotine patch and Wellbutrin are not approved for use during pregnancy (46).

DIABETIC CLASSIFICATION: THE WHITE CRITERIA FOR SEVERITY OF DIABETES DURING GESTATION

Although this classification is foreign to internists, it is used by obstetricians who take care of pregnant women with diabetes and is advocated by the American College of Obstetricians and Gynecologists in order to stratify both maternal and fetal risk in pregnant women with diabetes. Priscilla White, working with Elliot Joslin at the Joslin Clinic observed that a patient's age at onset of diabetes, its duration, and the severity of complications including vascular disease, nephropathy, and retinopathy influenced maternal and perinatal outcome adversely (Table 1). She developed a classification scheme in 1949 that is still widely used in the obstetric community due to its predictive value in identifying patients who are at greatest risk for obstetric complications during pregnancy. The updated classification scheme allows physicians to focus on intensified management and fetal surveillance on those patients who pose the highest risk for poor maternal and obstetric outcome during pregnancy. Pre-gestational (women with pre-existing diabetes) diabetic women are designated by the letters B,C,D,F,R,T, and H according to their duration of diabetes and complications. There is not a separate classification scheme for Type 1 and Type 2 diabetes but the initial scheme was developed for women with Type 1 DM.

PRE-EXISTING DIABETES AND PREGNANCY

Management and Treatment of Pre-exising Diabetes and Pregnancy

The goals of blood glucose control during pregnancy are rigorous. Optimally, the pre-meal whole blood glucose should be less than <95 mg/dl, the 1 hour postprandial glucose <140mg/dl and the 2 hour glucose <120mg/dl (52). Since fetal macrosomia (overgrowth) is more strongly related to the postprandial glucose excursions, pregnant diabetic women need to monitor their pre-meal and postprandial glucose values regularly (25). Type 1 diabetic patients usually require 3-4 injections per day or an insulin pump to achieve adequate control during pregnancy. Lispro (Humalog) insulin may be especially helpful in women with hyperemesis or gastroparesis because it can be dosed after a successful meal and still be effective (53). Increasingly, pregnant women with Type I diabetes are being managed with a flexible intensive self management program in which they learn to dose their short acting insulin according to a pre-meal correction factor and carbohydrate to insulin ratio (23). There is now adequate experience with short acting insulin Lispro demonstrating that it does not cross the placenta and is not associated with any increased risk in retinopathy (54). It is extremely useful in minimizing postprandial hyperglycemia and avoiding hypoglycemia in patients with Type 1 DM during pregnancy (53). Two doses of NPH, Lente, or Ultralente are typically used to try and achieve a basal insulin since Glargine is not approved in pregnancy due to concerns of its potential mitogenic effect and higher affinity to the IGF-1 receptor. However, in animal studies there is no evidence of reproductive toxicity or embryotoxicity with insulin glargine (55).The evening NPH or Lente dose usually needs to be moved to before bedtime to avoid nocturnal hypoglycemia and prevent fasting hyperglycemia.

The insulin pump is gaining favor in the treatment of Type 1 DM in pregnancy. In a non-randomized trial in which 24 women began insulin pump therapy during pregnancy and were compared to 12 women using the pump before pregnancy and 24 women treated with multiple insulin injections. There was no deterioration of glycemic control and maternal and perinatal outcomes were similar (56). However, 2 of the 24 women who began using the pump in pregnancy developed ketoacidosis due to pump failure compared to no cases of ketoacidosis in the two other groups. Therefore, it may be optimal to begin pump therapy before pregnancy due to the steep learning curve involved with its use. Women with Type 2 diabetes may achieve adequate glycemic control with twice daily injections because they tend to be more insulin resistant and experience less hypoglycemia. However, if postprandial lunch excursions are too high, three injections a day of a rapid acting insulin (Lispro) may be necessary. Perinatal outcomes were better with four times daily compared to twice daily regimens in both women with Type 2 DM and GDM in a randomized study (57).

Maternal hypoglycemia is common and often severe in pregnancy in women with Type 1 DM. In a series of 84 pregnant women with Type 1 DM, hypoglycemia requiring assistance from another person occurred in 71% of patients with a peak incidence at 10-15 weeks gestation (58). One third of subjects had a least one severe episode resulting in seizures, loss of consciousness, or injury. In pregnancy there appears to be an increase in hypoglycemic unawareness with the institution of intensive therapy, and this is worsened by the nausea and vomiting of pregnancy. There is also data to suggest that the counter-regulatory hormonal responses to hypoglycemia are diminished in pregnancy (59; 60). One of the highest risk periods for severe hypoglycemia is between midnight and 8:00 a.m, but diabetic women who have gastroparesis or hyperemesis gravidarum are at the greatest risk for daytime hypoglycemia (61). Occasional monitoring in the middle of the night is recommended in women with Type I diabetes because of the increased risk of nocturnal hypoglycemia, especially if they have hypoglycemia unawareness. The physician must have a low threshold for bringing the expectant mother into the hospital to optimize education and glycemic control. The risk of hypoglycemia to the fetus is difficult to study but animal studies indicate that hypoglycemia is potentially teratogenic during organogenesis which would translate into a gestational age between 3-10 weeks in the human (26). Exposure to hypoglycemia in utero may have long-term effects on the offspring including neuropsychological defects (26) so intensive efforts must made to avoid it. The patient should have a glucagon kit and carry easily absorbed carbohydrate with her at all times.

Failure to achieve optimal control in early pregnancy may have teratogenic effects in the first 3-10 weeks of gestation or lead to early fetal loss. Poor control later in pregnancy increases the risk of intrauterine fetal demise, macrosomia, cardiac septal enlargement in the fetus, and metabolic complications in the newborn (62). An early dating ultrasound is necessary to accurately determine the gestational age of the fetus and a formal anatomy scan at 18-20 weeks should be done to evaluate for fetal anomalies (63). A fetal echocardiogram should be offered at 20-22 weeks, especially if the HbgA1C was elevated during the first trimester. Women with Type 1 diabetes can be at risk for macrosomic infants (due to excess delivery of nutrients to the fetus from poor glycemic control) or intrauterine growth restriction (IUGR) due to the common finding of poor placental perfusion in women with longstanding diabetes and microvascular disease (64). Most recently, it is being recognized that although the mother may have glucoses in the target range, the fetus may still demonstrate abnormal growth (large for gestational age i.e. LGA) due to excessive nutrients being shunted to the fetus. This abnormal growth is usually in a characteristic pattern of head to body disproportion. The fetus exhibits advanced growth in the abdominal circumference measurement due to excessive subcutaneous fat, compared to the head measurement (65). This places the mother at an increased risk for cesarean section due to difficult delivery of the baby's enlarged abdomen. This abnormal growth pattern can be seen between 29-32 weeks. Increasingly, fetal criteria and growth patterns are dictating the aggressiveness of maternal glycemic treatment rather than simply using mother's glucoses as the goal for therapy (65).

Due to the increased risk of uteroplacental insufficiency and an intrauterine fetal demise in patients with longstanding Type 1 diabetes in those women with microvascular disease, diabetic nephropathy, hypertension, or evidence of poor intrauterine growth, fetal surveillance is recommended at ~28-30 weeks with twice weekly non-stress testing (NST) to ensure fetal well being (63). Serial ultrasounds are used to monitor growth and if the estimated fetal weight is less than the 10th percentile (small for gestational age or SGA), umbilical artery and uterine artery dopplers may be useful to estimate the degree of uteroplacental insufficiency and predict poor obstetric outcome (66). Delivery management is made according to maternal well being, growth of the fetus, evidence of uteroplacental insufficiency, and the results of fetal surveillance. An amniocentesis may be useful to determine fetal lung maturity if delivery before 38-39 weeks is being considered (63). A cesarean section (C-section) may be recommended for obstetric indications such as severe preeclampsia with an unfavorable cervix, estimated fetal weight >4000 grams with head to body disproportion, history of a C-section, or fetal distress. If there are no obstetric indications for a C-section, a vaginal delivery is optimal in the woman with Type 1 DM due to the higher risk of infectious complications, thromboembolism, and delayed recovery with C-section.

Diabetic Ketoacidosis in Pregnancy

Pregnancy predisposes the mother to accelerated starvation, which can result in ketonuria after an overnight fast (67). DKA may therefore occur at lower glucose levels, often referred to as "euglycemic DKA" of pregnancy, and may develop more rapidly than it does in non-pregnant individuals (10; 68; 9). Women also have a lower buffering capacity due to the progesterone-induced respiratory alkalosis resulting in a compensatory metabolic acidosis. Furthermore, euglycemic DKA is not uncommon in pregnancy due to earlier ketosis in pregnant women (2) and glomerular hyperfiltration in pregnancy which causes glucosuria at lower serum glucoses. Any pregnant woman with Type 1 diabetes unable to keep down food or fluids should check urine ketones at home and if positive, a chemistry panel should be ordered to rule out an anion gap even if the maternal glucose is < 200 mg/dl.

In a study of 20 consecutive cases of DKA, only 65% of fetuses were alive on admission to the hospital (24). Once the patient was hospitalized and treated, the risk of fetal loss declined dramatically. Risk factors for fetal loss included DKA presenting later in pregnancy (mean gestational age 31 weeks versus 24 weeks); glucose > 800 mg/dl; BUN > 20 mg/dl; osmolality > 300 mmol/L; high insulin requirements; and longer duration until resolution of DKA. The fetal heart rate must be monitored continuously until the acidosis has resolved. There was no maternal mortality in this small series. Causes of DKA in pregnancy are often different with infection less common as a precipitant (69). Of the infectious causes, pyelonephritis was the most common. However, there is often no precipitant other than emesis in the pregnant woman who can develop starvation ketosis very quickly. In a series of 37 pregnant women with DKA, emesis alone accounted for 42% of the cases (60% of these women had gastroparesis), and 17% were non-compliant. Beta agonist therapy, pump failure, infection, undiagnosed pregnancy, and new onset diabetes each accounted for 8% of the cases (69). Prolonged fasting is a common precipitant for DKA and it has been shown that even women with GDM can become severely ketotic if they are given terbutaline (to prevent contractions) or betamethasone (to accelerate fetal lung maturity) in the face of prolonged fasting (2). It is imperative to remember that the pregnant woman unable to take glucose orally requires an additional 100-150 grams of intravenous glucose to meet the metabolic demands of the pregnancy (7). Without adequate carbohydrate (often a D10 glucose solution is needed), fat will be burned for fuel and the patient in DKA will remain ketotic (1).

Diabetic Nephropathy and Hypertensive Disorders in Pregnancy

Women with pre-existing proteinuria often have a significant progressive increase in protein excretion, frequently into the nephrotic range, in part due to the 30-50% increase in glomerular filtration rate (GFR) that occurs during pregnancy (70). In most cases the proteinuria returns to the pre-pregnancy baseline after delivery. In some patients, however, the proteinuria can become massive and result in significant edema, hypoalbuminemia, and a hypercoagulable state. Factors which could accelerate nephropathy in pregnancy include the hyperfiltration of pregnancy, increase in protein intake, hypertension, and withdrawal of Ace Inhibitors (33). Fortunately, in an observational study of 26 women with Type 1 DM with preserved renal function, the decline in creatine clearance over time appeared to be no different in women who became pregnant versus those who did not (47).

While women with mild renal insufficiency are not at an appreciable risk for irreversible progression of their nephropathy (71; 72), those with more severe renal insufficiency (creatinine >1.5 mg/dl) have a 30-50% risk of a permanent pregnancy-related decline in GFR (48). In a series of 36 women with Type 1 DM and nephropathy, maternal and obstetric outcomes were strongly dependent on the degree of maternal renal function (48). In women with a creatinine clearance of >80 cc/min, the prematurity rate was 19% and the mean birth weight was 2670 grams in comparison to women with a creatinine clearance of 30-80 cc/min in whom 60% of the infants were premature and the mean birth weight was only 1640 gms. Overall, ~50% of the patients developed nephrotic range proteinuria, 97% of the patients required antihypertensive treatment, and 20% of the children had psychomotor retardation. Approximately 30% of mothers developed end stage renal disease and required dialysis at follow-up in 10 years and 4 women had died.

Long term outcome of infants born to mothers with nephropathy was studied in a 3 year follow-up of infants of ten diabetic mothers with stage IV diabetic nephropathy compared to 30 diabetic women without nephropathy (73). The mean birthweight was 1000 grams less in those women with nephropathy, births were premature in 60% of the women with nephropathy but in none without nephropathy, and 30% of the infants born to mother with nephropathy showed respiratory distress syndrome compare to 6% without nephropathy. Nephrotic range proteinuria developed in 70% of women with nephropathy in contrast to none without nephropathy. Three years postpartum, 60% of the children of nephropathic mothers had a body weight <50th percentile in comparison to none of the children of the women without nephropathy. In addition, the children of mothers with nephropathy started to speak significantly later (15 months versus 12 months) and had infectious disease more commonly (60% versus 6%).

Women with diabetic nephropathy are also at extremely high risk of developing preeclampsia which often leads to prematurity and intrauterine growth restriction. Even women with microalbuminuria are at a higher risk of preeclampsia than women without microalbuminuria. In a study of 240 women categorized according to their urinary albumin excretion (normal, <30 mg/24hr; microalbuninuria, 30-300 mg/24 hr; and diabetic nephropathy, >300 mg/24 hr), the incidence of preterm delivery, small for gestational age, and preeclampsia were highly significantly different (74). Of all deliveries in women with normal urinary albumin excretion, microalbuminuria, and diabetic nephropathy, 35, 62, and 91% were preterm, 2, 4, and 45% were small for gestational age, and preeclampsia developed in 6, 42, and 64% of the women, respectively. Blood pressure control is imperative to try to minimize the deterioration of renal function. The goal for blood pressure control is not as low in pregnancy (<130/80) due to the concerns about decreasing uteroplacental blood flow in the face of high vascular resistance in women at high risk of preeclampsia (75, 33).

Hypertension should be treated in the pregnant woman with pre-existing diabetes at a BP level of ~140/90, especially if the patient has underlying diabetic nephropathy. Preeclampsia must be ruled out since the risk is so high in women with pre-exisiting diabetes. Treating mother's blood pressure has not been shown to prevent preeclampsia. Agents such as Methydopa, Hydralazine, Calcium channel blockers, Clonidine, or Labetalol can all be used. Clearly there is most clinical experience using Methyldopa, Hydralazine, Nifedipine, or Labetalol in pregnancy but there may be as yet an unproven benefit to using Amilodipine or Diltiazem to prevent further increases in proteinura. Ace-Inhibitors are contraindicated in the second and third trimester of pregnancy due to the risks of fetal anuria, oligohydramnios, and pulmonary hypoplasia.

Women with severe renal insufficiency should be counseled that their chances for a favorable obstetric outcome may be higher with a successful renal transplant. Women with good function of their renal allografts who have only mild hypertension, do not require high doses of immunosuppressive agents, and are 1-2 years out from their renal translplant have a much better prognosis than women with severe renal insufficiency. Successful pregnancy outcomes have been reported in 89% of these patients (76).

Diabetic Retinopathy in Pregnancy

Progression of retinopathy during pregnancy is well documented. This phenomenon is most prevalent in women with high-risk diabetic eye disease such as severe preproliferative or proliferative retinopathy. Recent data suggest that rapid institution of tight control may be associated with subsequent progression of retinopathy. However it is unclear whether the tight control often achieved during pregnancy or the changes of pregnancy per se, including increased cardiac output by 20-40%, the production of placental angiogenic factors, anemia, the hypercoagulable state of pregnancy or pregnancy-induced hypertension account for this deterioration (33). In the Diabetes Control and Compliations trial, the effect of pregnancy was studied in women randomized to the intensive group before pregnancy versus those assigned to the conventional treatment who were then intensively controlled as soon as pregnancy was documented (77). Compared to the non-pregnant group, pregnant women had a 1.6 fold risk of worsening retinopathy from before to during pregnancy in the intensive group compared to 2.5 fold greater in the control group. However, by the end of the DCCT, the degree of retinopathy and albuminuria in subjects who became pregnant were not different than those women who did not.

In 179 prenancies in women with Type 1 diabetes who were followed prospectively, progression of retinopathy occurred in 5% of women. Risk factors for progression were duration of diabetes >10 years (10% versus 0%), moderate to severe background retinopathy (30% vesus 3.7%), and a trend for those women who had the greates fall in hemoglobin A1C (50). Fortunately, retinal changes appear to regress in the postpartum period in many patients (78). Given this uncertainty, it is best to intensify glycemic control and to stabilize retinopathy before conception (79). However, laser surgery is as effective in preventing blindness during pregnancy as it is outside of pregnancy and can be done safely. Women with low-risk eye disease should be followed by an ophthalmologist during pregnancy, but significant vision-threatening progression of retinopathy is rare in these individuals (80).

Coronary Artery Disease in Pregnancy

Pregnancy causes ~25% increase in cardiac output, a significant decrease in system vascular resistance which can shunt blood away from the coronary arteries, and an increase in oxygen consumption, all of which contribute to a decrease in the ability of maternal coronary blood flow to meet the demands of the myocardium. At labor and delivery there is a 60-80% increase in cardiac output caused by the release of venocaval obstruction, autotransfusion of uteroplacental blood, and rapid mobilization of extravascular fluid resulting in a marked increase in venous return and stroke volume (33). This in combination with acute blood loss and an activation of catecholamines at labor and delivery can result in acute mycardial ischemia and decompensation in a woman who has preexisting coronary artery disease. The mortality rate of a myocardial infarction in pregnancy prior to 1980 was ~60-70% versus 0/10 in reported cases since that time (33), thought secondary to improved care or counseling in diabetic women with coronary artery disease. However, maternal morbidity is extremely high. Women with any complaints suggestive of ischemic heart disease should be fully evaluated in pregnancy. Carefully monitored treadmill testing or stress echocardiography are not contraindicated nor is a revascularization procedure if indicated. It is recommended that women with preexisting coronary artery disease not become pregnant in spite of good left ventricular function, even though the risk of mortality is less if the myocardial infarction occurred before pregnancy than during pregnancy. Women with longstanding diabetes and especially those with other risk factors for coronary artery disease (hyperlipidemia or hypertension) should be evaluated for asymptomatic coronary artery disease before becoming pregnant (46). HMG Co-A reductase inhibitors are contraindicated in pregnancy, but if necessary, triglyceride lowering agents such as Gemfibrozil or Niacin can be used (42). There is no data yet available on Ezetimibe in pregnancy.

GESTATIONAL DIABETES

Diagnosis of Gestational Diabetes

Gestational diabetes mellitus (GDM) is defined as glucose intolerance of variable severity with onset or first recognition during pregnancy (52; 81). The incidence of GDM ranges from 2-14%% of pregnancies throughout the world and is highest in ethnic groups that have a higher incidence of Type 2 diabetes (Hispanic Americans, African Americans, Native Americans, and Pacific Islanders). A tremendous degree of controversy continues regarding the benefit of screening and treating for gestational diabetes due to the absence of high-quality randomized controlled trials (82; 81). The impact of hyperglycemia on adverse maternal and neonatal health outcomes is probably continuous. Although insulin therapy decreases the incidence of fetal macrosomia for those women with more severe degrees of hyperglycemia, the magnitude of any effect on maternal and other neonatal health outcomes is unclear (82). The ADA and ACOG recommends screening for pregnant women for gestational diabetes (52; 81) but the U.S. Preventive Medicine Task Force concludes that the evidence is insufficient to recommend for or against routine screening for gestational diabetes due to the absence of adequate randomized controlled trials (82).

The criteria for diagnosis in the United States has recently changed and the Carpenter and Coustan criteria have been adopted by the ADA and the Fourth International Workshop-Conference on Gestational Diabetes (52; 83; 84). Screening recommendations have been stratified according to low risk status, average risk status, and high risk status of GDM. Most obstetricians employ universal screening of all women at 24-28 weeks which is a reasonable approach, especially in a population that contains ethnic groups with a higher prevalence of GDM. However in populations with a lower incidence of GDM, selective screening has been shown to be effective (85) (86).

Women with a fasting blood glucose >125 mg/dl or a random or postprandial glucose of > 200 mg/dl meet the criteria for GDM and this precludes the need for any glucose challenge. All other high risk status women should be given a 50 gm glucose challenge (Glucola test) or proceed directly to a 100 gm oral glucose tolerance test as soon as they establish prenatal care. If initial testing is normal, repeat testing should be done at 24-28 weeks gestation.

Average Risk Status: These are women who do not fall in the low risk or high risk status. They should receive a 50 gm glucose challenge at 24-28 weeks and if positive, undergo diagnostic testing with a 100 gm 3 hour oral glucose tolerance test (3 hr OGTT).

50 Gram Glucola: The 50 gram glucose challenge is the accepted screen for the presence of GDM but, if positive, must be followed by a diagnostic 100 gm 3 hour oral glucose tolerance test (3 hr OGTT). A positive screen is in the range of 130-140 mg/dl. The sensitivity and specificity of the test will depend on what threshold value is chosen and the cutoff may be selected according to the prevalence of GDM in the population being screened. The test does not have to be done fasting, but a serum sample must be drawn exactly one hour after administering the oral glucose.

100 Gram 3 hour OGTT: The 100 gm 3 hour test must be done after 3 days of an unrestricted carbohydrate diet and while the patient is fasting. A positive test requires that 2 values be met or exceeded. One abnormal value should be followed with a repeated 3 hour test one month later because a single elevated value increases the risk of macrosomia and one-third of patients will ultimately meet the diagnostic criteria for GDM.

Pathophysiology of Gestational Diabetes

GDM is caused by abnormalities in at least 3 aspects of fuel metabolism: insulin resistance, impaired insulin secretion, and increased hepatic glucose production (6). The insulin resistance of normal pregnancy is thought to be due primarily to the effects of increased production of human placental lactogen and possibly placental growth hormone, both which peak in the third trimester when insulin resistance is greatest (14; 17; 87). An increase in circulating TNF-a has also been reported recently in human pregnancy (19), and may be an important additional contributing factor to insulin resistance. Insulin resistance during pregnancy is usually compensated for by a considerable increase in insulin secretion. However, in 3-8% of women, insulin resistance is more profound and this challenge, combined with decreased pancreatic beta-cell reserve, triggers GDM. Women diagnosed with GDM appear to have abnormalities in insulin secretion that contributes to the development of GDM (88). Investigators have shown more pronounced insulin resistance during pregnancy in GDM patients compared to women with normal glucose tolerance, that may contribute to hyperglycemia in addition to defects in insulin secretion (3). Although diabetes usually remits after pregnancy, 30-50% of women diagnosed with GDM go on to develop type II diabetes mellitus (T2DM) later in life, particularly if obesity is present. GDM shares many of the characteristics of T2DM. Both are aggravated by increasing obesity and age, suggesting that the components of insulin resistance and decreased insulin secretion, which lead to GDM, may be common to NIDDM. Thus, GDM may represent an unmasking of the genetic predisposition of NIDDM induced by the hormonal changes of pregnancy (5).

Although insulin resistance is a universal finding in pregnancy and GDM, the cellular mechanisms for this type of insulin resistance are multi-factorial and just beginning to be understood. Insulin binding to its receptor is unchanged in pregnant and GDM subjects, and in skeletal muscle, the major glucose consuming tissue, the levels of glucose transporter protein GLUT4 are unchanged in pregnancy and GDM. Pregnancy reduces the capacity for insulin-stimulated glucose transport independent of obesity, due in part to a tissue-specific decrease in insulin receptor phosphorylation and decreased expression of Insulin Receptor Substrate-1 (IRS-1), a major docking protein in skeletal muscle. In addition to these mechanisms, in muscles from GDM subjects, there is a decreased ability of insulin to fully stimulate tyrosine phosphorylation of the insulin receptor, not found in normal pregnancy (4; 89). Decreased tyrosine phosphorylation likely contributes to further decreases in insulin receptor activity, less phosphorylation of IRS-1, and other down-stream signaling events important for GLUT4 trafficking to the plasma membrane, resulting in lower glucose transport compared to normal pregnancy. GDM subjects also tend to have higher circulating FFA and reduced PPARg expression in adipose tissue, a target for thiazolidinediones. There is also evidence for a decrease in the number of glucose transporters (GLUT-4) in adipocytes in GDM subjects and an abnormal translocation of these transporters that results in reduced ability of insulin to recruit them to the cell surface, which contributes to the overall insulin resistance of GDM (12).

Risks to the Mother with Gestational Diabetes

The immediate risks to the mother with GDM are an increased incidence of cesarean section (~30%), preeclampsia (~20-30%), and polyhydramnios (~20%) which can result in preterm labor (63). The long-term risks to the mother are related to recurrent GDM pregnancies and the substantial risk of developing Type 2 DM. Women with GDM represent a group of patients with an extremely high risk (~50%) of developing Type 2 diabetes in the subsequent 5-10 years. Women with fasting hyperglycemia, GDM diagnosed prior to 24 weeks (preexisting glucose intolerance), obesity, those belonging to an ethnic group with a high prevalence of Type 2 DM, or who demonstrate impaired glucose tolerance at 6 weeks postpartum, have the highest risk (90; 91; 92; 93) of developing Type 2 GDM. Women with impaired glucose tolerance postpartum have up to an 80% risk of developing Type 2 DM within five years and should be targeted for primary prevention (94). Counseling with regard to diet, weight loss, and exercise is essential and is likely to improve insulin sensitivity. Such dietary modifications should be adopted by the family since the infant is also at increased risk of developing impaired glucose tolerance (95; 96). Whether or not medications that may improve insulin sensitivity (metformin and thiazolidinediones) could be used in this high risk group of patients to prevent the development of Type 2 diabetes is under current investigation.

Risks to the Infant from Gestational Diabetes

Macrosomia is the major risk to the fetus in women with GDM. Many theories have been generated over the years to explain the macrosomia associated with diabetes in pregnancy (96; 6; 5; 31; 97). Overall, the theory of excessive fetal insulin due to increased transport of maternal fuel to the conceptus holds the most credence and has the most supportive data (Freinkel hypothesis). Diabetes in pregnancy is associated with increased delivery of glucose and amino acids to the fetus via the maternal circulation (98). These fuels can stimulate increased production of fetal insulin which promotes somatic growth. Other maternal substrates (e.g., free fatty acids, triglycerides) add to the burgeoning supply of fetal substrate and further support excessive growth (8?). It is, therefore, the goal of management of pregnancies complicated by diabetes to normalize the above parameters with good metabolic control. Maternal obesity appears to be an independent risk factor since some mothers who appear to have optimal metabolic control still give birth to macrosomic infants (99; 97). Furthermore, macrosomia is not limited to the diabetic population; in fact, approximately 25% of macrosomic infants are born to mothers without GDM. It has recently been shown that women may have glucoses within target range yet there is excess shunting of glucose to the fetus as demonstrated by increased amniotic fluid insulin levels reflecting fetal hyperinsulinemia (100). In some countries, amniotic fluid insulin levels are thought to be best predictor of macrosomia and decisions about treatment therapy in the mother are based on this evidence of fetal hyperinsulinemia (100).

Even with the advent of screening and aggressive management of GDM, the incidence of neonatal complications ranges from 12-28% (63; 96). Macrosomia places the mother at increased risk of requiring a cesearean section and the infant at risk for shoulder dystocia. In a randomized trial it was demonstrated that if insulin therapy is started in women with GDM whose maternal glucoses are at target levels on diet alone but whose fetuses are demonstrate excessive growth, the rate of fetal macrosomia can be decreased (65). Shoulder dystocia can result in Erb's palsy, clavicular fractures, fetal distress, low APGAR scores, and even birth asphyxia when unrecognized (101). Shoulder dystocia occurs nearly 50% of the time when a 4500 gram infant is delivered vaginally (102). Preterm labor can result due to polyhydramnios from the fetus ultrafiltrating glucose through the kidneys. In mothers who have poor glycemic control, respiratory distress syndrome may occur in up to 31% of infants while cardiac septal hypertrophy may be seen in 35-40% (31; 63). With extremely poor glucose control, there is also an increased risk of fetal mortality due to fetal acidemia and hypoxia. Common metabolic abnormalities in the infant of a GDM mother include neonatal hypoglycemia, hypocalcemia, hyperbilirubinemia and polycythemia. Neonatal hypoglycemia is common in women in suboptimal glycemic control because the infant may continue to produce excessive insulin for up to 24 hours after birth before the normal feedback loop is operating.

Women with GDM who require insulin or glyburide or those who are not taking insulin but have suboptimal glycemic control should undergo fetal surveillance at ~32 weeks gestation and an earlier delivery should be considered after fetal lung maturity is confirmed by amniocentesis (63). An ultrasound for growth to look for head to body disproportion and evidence of the fetus being large for gestational age should be obtained at ~28-32 weeks (65), especially if it would influence treatment. Ultrasonography can often predict the risk of fetal macrosomia by measuring the abdominal circumference of the fetus at 29-33 weeks. An estimated fetal weight of > 4500 grams carries such a high risk of shoulder dystocia that an elective cesearan section is usually recommended (63; 102). Women with good dating criteria, a favorable cervix, and an estimated fetal weight <4000 grams are often electively induced at 38-39 weeks in an attempt to decrease macrosomic births (103).

The long-term sequelae of GDM for offspring are much more controversial. Reports of an increased risk of adolescent obesity and of Type 2 diabetes are compelling and it appears that fetal islet hyperplasia occurs in-utero with maternal hyperglycemia resulting in an increased risk of developing Type 2 diabetes in teenage years or as a young adult (104). In Pima Indians, the incidence of childhood Type 2 DM at 10-14 years in the offspring of GDM mothers was 20 times higher compared to the offspring of non-diabetic mothers and 5-fold higher than that of pre-diabetic mothers who develop Type 2 DM after pregnancy (105). Elevated amniotic fluid insulin levels (due to fetal hyperinsulinemia as a result of maternal hyperglycemia) predicted teenage obesity in one study, independent of fetal weight, and one-third of these offspring had impaired glucose tolerance by 17 years of age (106). This scenario creates enormous potential on a public health level for the incidence of Type 2DM to escalate as these children with impaired glucose tolerance become mothers themselves, perpetuating the cycle.

Treatment Strategies for Gestational Diabetes

Women with GDM should be taught home glucose monitoring to ensure that their glycemic goals are being met throughout the duration of pregnancy. The best therapy for GDM depends entirely on the severity of the glucose intolerance and on the mother's response in addition to the effect on fetal growth. In at least half of the cases, diet alone will maintain the fasting and postprandial blood glucose values within the target range. Since postprandial glucose levels have been most strongly associated with the risk of macrosomia (25) modest carbohydrate restriction to ~45% of total calories may be helpful to blunt the postprandial glucose excursions. Women who weigh more than 130% of ideal body weight should be restricted to a caloric intake of ~20-25 kcal/kg body weight and advised to limit their weight gain to no more than 15 lbs. For obese women (BMI >30, a 30% caloric restriction (an intake of ~1800 calories per day) has been shown to reduce hyperglycemia and plasma triglycerides with no increase in ketonuria (27). None of the oral diabetes medications (sulfonylureas, metformin, acarbose, or the thiazolidinediones) are currently approved for use in pregnancy. However, a recent multicenter trial including 400 women with GDM were randomized to either insulin or glyburide after 24 weeks gestation and maternal glycemic control, macrosomia, neonatal hypoglycemia, and neonatal outcome were no different between the groups. Most importantly, the cord-serum insulin concentrations were similar in the two groups and glyburide was not detected in the cord serum of any infant (43) . Although Glyburide is not FDA approved for use in pregnancy, an increasing number of diabetes centers are gaining experience with its use and there have not been any significant number of adverse reports associated with its use in pregnancy.

Women who have fasting blood glucose levels > 95 mg/dl, 1 hour postprandial glucose levels >140 mg/dl or 2 hour postprandial glucose levels > 120 mg/dl should be started on insulin therapy (107). Glyburide may be an alternative if women strongly prefer the option of an oral hypoglycemic agent and understand that glyburide is not FDA approved for use in pregnancy. No other oral hypoglycemic should be substituted because only glyburide has been shown not to cross the placenta. Those with large for gestational fetuses by ultrasound or who have head to body disproportion are also candidates for insulin or glyburide therapy even though maternal glucoses are within target range on diet alone (65). GDM can usually be treated with twice daily injections of NPH and Regular insulin but occasionally postprandial glycemic excursions are so excessive that three times daily mealtime injections of Lispro (Humalog) are necessary. Alternative treatment with glyburide remains to be defined but starting doses of 2.5 q.d. to b.i.d. were used in the study and titrated up to a maximal dose of 10 mg b.i.d. Up to 15% of patients will fail maximum dose Glyburide therapy and need to be switched to insulin, especially if dietary restriction is not carefully followed (43). Hypoglycemia tends to be an infrequent occurrence in these patients because of their underlying insulin resistance.

It has been demonstrated that moderate exercise is well tolerated in pregnancy (108). Fetal safety has been established if the maternal heart rate is maintained < 150 beats per minute at durations of less than 1 hour and if the mother is well hydrated and does not get overheated. Two out of three trials in pregnancy have shown that a combination of aerobic and weight bearing exercise three times per week can achieve glycemic control and infant birth weights that are similar to those seen in women who are treated with insulin (108). Establishing a regular routine of modest exercise during pregnancy may also have long lasting benefits for the GDM patient who clearly has an appreciable risk of developing Type 2 diabetes in the future. Home glucose monitoring must be continued throughout pregnancy to determine whether or not insulin therapy will be necessary. Women at risk for preterm labor or conditions predisposing to growth restriction are not candidates for a controlled exercise program.

POSTPARTUM CARE OF THE WOMAN WITH DIABETES

A number of critical issues including maintenance of glycemic control, diet, exercise, weight loss, blood pressure management, breastfeeding, contraception and postpartum thyroiditis need to be addressed in the postpartum period. It has been demonstrated that the majority of women with pre-existing diabetes, even those who have been extremely compliant and who have had optimal glycemic control during pregnancy, have a dramatic worsening of their glucose control after the birth of their infant (109). Furthermore, many quit seeking medical care for their diabetes. The postpartum period is relatively neglected, therefore, as both the new mother and her physician relax their vigilance. However, this period offers a unique opportunity to institute health habits that could have highly beneficial effects on the quality of life of both the mother and her infant.

Postpartum Issues in Women with Type 1 Diabetes

Home glucose monitoring should be continued vigilantly in the postpartum period because insulin requirements drop almost immediately and often dramatically at this time, increasing the risk of hypoglycemia (62). Women with Type 1 DM often need to decrease their insulin by ~50%, immediately after delivery and may have a "honeymoon" period for several days in which their insulin requirements are minimal (61). Also, women who are candidates for an ACE-inhibitor can be started on one of these agents at this time as they have not been shown to appear in breast milk. (42). Breastfeeding is encouraged in women with Type 1 DM and may decrease the risk of Type 1 DM in the offspring but requires an additional 400-500 calories or oral intake in order to prevent hypoglycemia. (62) Women with Type 1 diabetes have been reported to have a 30% incidence of postpartum thyroiditis (110). Hyperthyroidism can occur in the 2-4 month postpartum period and hypothyroidism may present in the 4-8 month period. Given the significance of this disorder, a TSH measurement should be offered at 6 months postpartum and before this time if a patient has symptoms (111).

Postpartum Issues in Women with GDM

Women with a history of GDM should have their glycemic status reassessed at 6 weeks postpartum. A weight loss program consisting of diet and exercise should be instituted for women with GDM in order to improve their insulin sensitivity and hopefully to prevent the development of Type 2 diabetes (112). Hyperglycemia generally resolves in the majority of patients during this interval but up to 10% of patients will fulfill criteria for Type 2 DM (113; 90; 91; 95). At the minimum, a fasting blood glucose should be done to determine if the woman has persistent diabetes (glucose >125 mg/dl) or impaired fasting glucose tolerance (glucose of at least 110 mg/dl). A 75 gm 2 hour glucose tolerance test is recommended by the ADA since most women with impaired glucose intolerance will be missed if only a FBG is checked (52). A 2 hour value of at least 200 mg/dl establishes a diagnosis of diabetes and a 2 hour value of at least 140mg/dl but less than 200 mg/dl makes the diagnosis of impaired glucose tolerance. The importance of diagnosing impaired glucose intolerance lies in its value in predicting the future development of Type 2 diabetes. In one series, a diagnosis of impaired glucose tolerance was the most potent predictor of the development of Type 2 diabetes in women with a history of GDM; 80% of such women developed diabetes in the subsequent 5-7 years (94). Intensified efforts promoting diet, exercise and weight loss should be instituted in these patients. Nondiabetic women with a history of GDM should have annual measurements of their fasting glucose levels and lipid profiles. The TRIPOD study demonstrated that the use of a thiozolidinedione postpartum in women with a history of GDM and persistent impaired glucose intolerance decreased the development of Type 2 diabetes. The rate of Type 2 DM in the 133 women randomized to Troglitazone was 5.4% versus 12.1% in the 133 women randomized to placebo at a median follow-up of 30 months (114). The protection from diabetes was closely related to the degree of reduction of insulin secretion three months after randomization and persisted 8 months after the medication was stopped. Neither Metformin or the thiozolidinediones are approved at this time for the prevention of Type 2 diabetes. However these interventions are being intensely studied for primary prevention in this extremely high risk group.

Women should be encouraged to breastfeed unless difficulties in glycemic control arise. Women who breastfeed appear to have a lower incidence of developing Type 2 diabetes and it also appears to decrease the risk of developing infant obesity and impaired glucose tolerance (115). None of the oral agents are approved for use while breastfeeding as it appears that the sulfonylureas and metformin cross into breast milk (42). It is recommended that insulin be continued in diabetic mothers who choose to breastfeed. Calcium intake should be at least ~1500 mg per day since exclusive breast-feeding for an extended period of time can cause a modest decrease in bone density (116).

Contraception in Women with Pre-existing or Gestational Diabetes

It should be documented at every visit that women are using or have been offered an effective birth control method. The vast majority of contraceptive methods are relatively safe in women with diabetes who do not have poorly controlled hypertension or hypertriglyceridemia and who are not at increased risk for thromboembolic disease (95). Triglycerides should be measured after the initiation of oral contraceptives in all women with diabetes or a history of GDM because of the significant incidence of hypertriglyceridemia and the associated risk of pancreatitis with oral estrogen use in these women (117). Whether the estrogen containing contraceptive transdermal patch causes less triglyceride elevation in women with diabetes has yet to be studied. Low dose combined oral contraceptives have been shown to be effective and to have minimal metabolic effects in women with diabetes (118). In a retrospective cohort of 904 women with GDM, combined oral contraceptives did not influence the development of Type 2 diabetes (119). Progestational agents such as Norplant, Depo-Provera, and Norethindrone are also alternatives. There is no increase in pelvic inflammatory disease with the use of intrauterine devices in women with well controlled Type 1 or Type 2 diabetes after the post-insertion period (95). Therefore, this may be an attractive choice in older women who do not desire future pregnancies.

CONCLUSION

The obstetric outlook for pregnancy in women with pre-existing diabetes continues to improve enormously as rapid advances in diabetes management, fetal surveillance, and neonatal care emerge. Perhaps, the greatest challenge to face is the growing number of women developing gestational diabetes as the obesity epidemic increases. The development of Type 2 DM in the mother and glucose intolerance in the offspring set the stage for a perpetuating cycle that must be aggressively addressed with effective primary prevention strategies.