Gestational diabetes mellitus (GDM) and type 2 diabetes comprise more than 95% of the diabetes complicating pregnancy. Both are characterized by a relative (not absolute) deficiency in insulin as the result of inadequate production/secretion, peripheral insulin resistance, or a combination of the two. The mainstay of medical management of type 2 diabetes for about 40 years has been oral hypoglycemic agents. It is only logical then that their use should be considered in treatment of GDM. If they could be proven to be effective and safe for both babies and mothers, the ease, convenience, patient acceptability, and relatively low cost compared to insulin would favor their widespread use and perhaps use under circumstances of ‘borderline’ glucose intolerance as a means of reducing fetal risks related to macrosomia where intervention with insulin might meet with resistance by both providers and patients alike. Although there are five classes of oral hypoglycemic agents currently in use, sulfonylureas, biguanides, meglitinides, ?-glucosidase inhibitors, and thiazolidinediones, my familiarity with the use of these agents in pregnancy is limited to only the first two and I will not venture beyond my level of experience to discuss the others.
The sulfonylureas were the first hypoglycemic drugs that reached approval in this country (for nonpregnant women) and were available even during my training. These drugs appear to exert their beneficial effects by acting directly on pancreatic beta cells to enhance the secretion of insulin from the pancreas. The use of first generation sulfonylurea compounds, such as chlorpropamide and tolbutamide, in pregnant women was discouraged in the U.S. based on limited scientific evidence. These drugs do cross the placenta and their use was associated with severe, prolonged hypoglycemia in neonates whose mothers were taking them throughout pregnancy (Zucker, et al., Pediatrics 1968;42:824-25; Kemball, et al., Arch Dis Child 1970;45:696-701). However, it is not clear from these early studies if the neonatal hypoglycemia might have simply been the result of very poor maternal blood sugar control.
Animal studies in mice and rats also suggested that these agents were teratogenic (caused birth defects) when given during critical stages of embryogenesis (Freinkel, et al., Am J Clin Nutr 1986;44:986-95; Eriksson, et al., Diabetes 1991;40(suppl 2);94-8; Styrud, et al., Teratology;1992;46:473-83). A small retrospective of only 20 women with type 2 diabetes who were taking these agents early in pregnancy reinforced these concerns (Piacquadio, et al., 1991 Lancet;338:866-69). When compared to 40 other pregnant women with type 2 diabetes who were not taking these drugs, 50% (10) of the exposed babies had congenital malformations compared to only 15% in the control group. However, again, poor blood sugar control alone in diabetics during the embryonic period is well-known to be associated with the same birth defects (including, anencephaly, cardiac defects, vertebral abnormalities, and single umbilical artery) that were reported in this small study. Frequently ignored, when the concerns noted above were raised, was a large, randomized, controlled trial published in 1971 and conducted in South Africa that compared the use of first generation sulfonylurea compounds versus insulin in the management of women with type 2 diabetes in pregnancy (Notelovitz, S Afr Med J 1971;45:226-29). In this study, no differences in the rate of birth defects were noted between the two groups.
Over the years, some degree of comfort has been achieved with the use of the second generation sulfonylurea drug, glyburide, in the pregnancy management of pregestational type 2 diabetics and GDM. In contrast to the first generation drugs, very little glyburide appears to cross the placenta to the baby. Furthermore, glyburide carries an FDA category B classification (in contrast to the first generation agents category C) suggesting minimal risk for human teratogenicity. The low risk for birth defects seems to be supported by multiple trials of sufonylureas (and the biguanide drug, metformin, to be discussed later) in early pregnancy in which the risk for major congenital malformations is correlated with poor maternal blood sugar rather than the use of any specific hypoglycemic agent (Hellmuth, et al., Diabet Med 1994;11:471-74; Towner, et al., Diabetes Care 1995;18:1446-51; Langer, et al., J Matern Fetal Med 2000;9:35-41; Hellmuth, et al., Diabet Med 2000;17:507-11; Gutzin, et al., Can J Clin Pharmacol 2003;10:179-83).
In addition to appearing very safe for use in pregnancy, glyburide is also very effective in many women with pregestational type 2 diabetes and GDM. Langer and colleagues (N Engl J Med 2000;343:1134-38) reported a randomized trial of glyburide versus insulin therapy in a randomized trial of 404 pregnant women with GDM. The mean gestational age at which therapy was started was 28 weeks. No significant differences between the groups were found for blood sugar control, fetal macrosomia, or neonatal hypoglycemia. Umbilical cord blood studies at delivery demonstrated no detectable concentrations of glyburide in women who received the drug. Remarkably, only 4% of the women on glyburide in this trial failed to achieve good glycemic control and were changed to insulin. Other smaller studies have demonstrated higher failure rates (generally ranging between 15 and 20%) and greater risk of failure in women identified to have GDM before 25 weeks and older than age 34(Conway, et al., J Matern Fetal Neonatal Med 2004;15:51-5; Kahn, et al., Obstet Gynecol 2006;107:1303-09). Furthermore, a nonrandomized, retrospective cohort trial by Jacobsen and colleagues (Am J Obstet Gynecol 2005;193:118-24), while demonstrating better efficacy of glyburide compared to insulin in achieving good blood sugar control, and similar rates of fetal macrosomia, did also note a greater risk for preeclampsia, birth injuries, and neonatal hyperbilirubinemia in the glyburide group. Confirmation of these findings will have to await future outcome studies.
When using glyburide, we usually begin at a low dose of 2.5 mg once or twice per day. Personally, I prefer the twice per day regimen. It generally takes 4 to 5 days or longer before an impact is seen on fasting and two-hour post-prandial blood sugar levels, but if the response has been inadequate within a week, it is usually safe to double the initial dose. If glycemic control is inadequate during the day, but fasting blood sugars are normal, the morning dose can be increased independently of the evening dose, just as we do with insulin. It is generally recommended that the total 24-hour dose of glyburide not exceed 20 mg and if adequate glycemic control is not achieved within 3 weeks, consider changing over to insulin. Side-effects are usually minimal but women will complain of episodic hypoglycemia and, in some cases, the use of glyburide is accompanied by excessive weight gain unless dietary intake (and exercise) is adjusted to compensate for the increased efficiency of utilization of the ingested calories! In my own experience, glyburide, coupled with diet and exercise, can achieve adequate blood sugar control in about 60-80% of pregnant women.
The other class of oral hypoglycemic agents which has received considerable attention in recent years is the biguanides group of which metformin (glucophage) is the only drug with which I have any familiarity. Metformin appears to work primarily by increasing the sensitivity (reducing the resistance) of the liver and peripheral tissues (muscle and fat) to insulin. To a lesser extent, it also appears to decrease absorption of glucose in the gastrointestinal tract. The actions of metformin have the combined effects of reducing glucose production by the liver and enhancing peripheral glucose utilization. Like glyburide, metformin is also classified by the FDA as a category B drug, however, it does have the disadvantage of being transferred across the placenta to the baby in which levels can be found that are comparable to those in maternal blood (Vanky, et al., Fertil Steril 2005;83:1575-78) and this has contributed to delayed acceptance in the management of the obstetrical patient with diabetes.
Most of the data to support the safety and efficacy of metformin in pregnancy has come from the ‘infertility’ literature, particularly that involving women with polycystic ovary syndrome (PCOS). PCOS is frequently accompanied by insulin resistance. In 2000, Glueck and colleagues (Fertil Steril 2000;74:396-7) presented a case report of a 32 year old woman with amenorrhea and infertility associated with polycystic ovary syndrome (PCOS) who had failure of 7 out of 10 IVF embryo transfers, 1 premature live birth, and two pregnancy losses at 8 and 17 weeks. She was obese, had high fasting serum insulin, androstenedione, and testosterone levels. Although not overtly ‘diabetic’, she was begun on metformin (2.55 g/day) and a weight reduction program and over the course of 4 months, her weight fell from 109 to 91.3 kg (16%), her insulin, androstenedione, and testosterone levels normalized (as did her PAI-1 activity levels).
As a follow-up to this case report, Glueck’s group (Fertil Steril 2001;75:46-52) reported preliminary results from an ongoing pilot study to determine whether metformin could reduce the rate of first trimester pregnancy loss in women with PCOS. They identified 19 women with PCOS who did not have overt diabetes and placed them on metformin (1.5-2.55 g/day) throughout pregnancy. Ten of the women had previously conceived but had miscarried 16 of their 22 pregnancies (73%). “While receiving metformin, these 10 women had 6 normal live births (60%), 1 spontaneous abortion (10%), and 3 normal ongoing pregnancies (30%)” all > 13 weeks. Up to the time of the report, among all 19 women receiving metformin, 9 (47%) had normal term live births, 2 (11%) had normal, but preterm births at 33 and 35 weeks, 6 (32%) had normal ongoing pregnancies beyond 1 weeks, 2 (10.5%) had first trimester miscarriages. For purposes of our discussion here, it is important to note that no adverse maternal side-effects, nor birth defects were attributed to metformin in this small study. Thus, metformin alone appeared to improve pregnancy outcome in a group of PCOS patients who had either had, or were at increased risk, for early pregnancy loss.
In subsequent studies, Glueck and colleagues (Fertil Steril 2002;77:520-25) demonstrated in 72 nondiabetic PCOS women that the use of metformin preconceptionally and continued throughout pregnancy not only reduced the risk of early pregnancy loss, but also reduced the risk of developing gestational diabetes (3% in the treated group vs 23% in controls). Again, though the total number of patients was small, they demonstrated no increased risk for congenital malformations in the metformin-treated patients (Glueck, et al., Hum Reprod 2002;17:2858-64). In follow-up evaluation over the first 18 months of life of 126 infants born to 109 women with PCOS who received metformin throughout pregnancy, these investigators also found no significant abnormalities of growth, motor, or social development compared to matched controls (Glueck, et al., Hum Reprod 2004;19:1323-30). A recent meta-analysis of pregnancy outcome after first trimester exposure to metformin actually has demonstrated a statistically significant protective effect for fetal malformations among metformin recipients (1.7%) compared to disease-matched controls (7.2%) (Gilbert, et al., Fertil Steril 2006;86:658-63).
Recently, preliminary results of a prospective, randomized trial comparing metformin and insulin in the management of GDM have been reported (Moore, et al., J Reprod Med 2007 52;1011-15). Although only 63 patients (32 metformin- and 31 insulin-treated) were included to date, fasting and two-hour post-prandial blood sugars were comparable between the groups and there were no differences in fetal birth weights, respiratory distress syndrome, neonatal hypoglycemia, hyperbilirubinemia, or neonatal intensive care unit admissions. None of the patients enrolled to date failed metformin therapy and required insulin and most were controlled on the initial dosing regimen of 500 mg twice per day. Maternal outcomes also did not differ between the two groups. Against this background, it should be noted that at least one older study (Helmuth, et al, S Af Med J 1984;65:635-7) reported a significantly higher prevalence of preeclampsia among 50 women treated with metformin compared to 68 treated with a sulfonylurea compound and 42 treated with insulin during pregnancy (32% vs 7% vs 10%, respectively; P < 0.001). In the same study, there was also a higher rate of perinatal mortality (11.6% vs 1.3%; P < 0.02) in the group treated with metformin in the third trimester. More recent studies do not support these earlier findings. Glueck and colleagues (Diabet Med 2004;21:829-36) found no higher risk for preeclampsia in 90 PCOS women treated throughout pregnancy compared to 252 healthy women with no history of PCOS.
When dosing metformin, we usually begin at 500 mg twice per day. Usually a beneficial effect on glycemic control will be seen within 3-5 days. Again, if the response has been inadequate within a week, the dose can be increased by 500 mg per day on a weekly basis until 2,000 (occasionally 2,500 mg) per day is reached. My first increment above the initial dosing regimen is to 500 mg three times a day, and then if necessary, 1,000 mg twice per day. If adequate glycemic control is not achieved within 3 weeks, one can either change to or add insulin to the current metformin regimen. Due to its insulin ‘sensitizing’ effects, required doses of insulin, when used in conjunction with metformin, are usually about half those necessary to achieve good blood sugar control without metformin. Side-effects of metformin include gastrointestinal intolerance and, even though food decreases its absorption, it is still suggested that metformin be given with meals during pregnancy to minimize this complication. Hypoglycemia is generally rare with metformin, but it can occur and in rare circumstances (0.03 cases per 1,000 patients) lactic acidosis has been described.
In our next post in this series, we will discuss complications related to and management of type 1 diabetes during pregnancy…