Metformin in Pregnancy and Breastfeeding
Fetal Risk Summary
Metformin is an oral, biguanide, antihyperglycemic agent that is chemically and pharmacologically unrelated to the sulfonylureas. Its mechanism of action is thought to include decreased hepatic glucose production, decreased intestinal absorption of glucose, and increased peripheral uptake and utilization of glucose (1,2). The latter two mechanisms result in improved insulin sensitivity (i.e., decreased insulin requirements) (1,2).
Reproduction studies found no evidence of impaired fertility in male and female rats and no evidence of teratogenicity in rats and rabbits at doses up to 600 mg/kg/day, approximately 2 times the maximum recommended human dose on a mg/m2 basis (1). A partial placental barrier to metformin was observed, however, based on fetal concentrations (1).
Shepard (3) and Schardein (4) cited a study in rats that observed teratogenesis (neural tube closure defects and edema) in rat fetuses exposed to metformin. The drug did not appear to be a major teratogen because less than 0.5% of the rat fetuses in mothers fed 5001000 mg/kg developed anophthalmia and anencephaly (3). Higher doses in this study were embryotoxic (5).
A 1994 abstract described the teratogenic effects of high metformin concentrations on early somite mouse embryos exposed in vitro (6). At levels much higher than those obtained clinically, metformin produced neural tube defects and malformations of the heart and eye. In contrast, a study that also appeared in 1994 observed no major malformations in mouse embryos exposed in culture to similar concentrations of metformin (7). About 10% of the embryos, however, demonstrated a transient delay in closure of cranial neuropores.
In an experiment using the human single-cotyledon model with placentas obtained from diabetic patients and normal controls, researchers measured the effect of metformin on the uptake and transport of glucose in both the maternal-to-fetal and fetal-to-maternal directions (8). A second publication described only the results of the experiments studying the effect of metformin on the maternal-to-fetal direction of glucose (9). Compared with controls, metformin had no effect on the movement of glucose in either direction (8,9).
Among 26 women undergoing treatment for polycystic ovary syndrome with metformin, 1.5 g/day for 8 weeks, 3 became pregnant during treatment (10). The women were involved in a study to determine whether metformin was effective in normalizing the condition that is characterized by insulin resistance and hyperandrogenism. One of the pregnancies aborted after 2 months, and the outcomes of the other two were not mentioned.
A number of References have described the use of metformin during all stages of gestation for the control of maternal diabetes (11,12,13,14,15,16,17,18 and 19). A 1979 Reference described the therapy of gestational diabetes that included the use of metformin alone (N=15), glyburide alone (N=9), or metformin plus glyburide (N=6) in women who were not controlled on diet alone and did not require insulin (13). None of the newborns developed symptomatic hypoglycemia, and only one infant had a congenital defect (ventricular septal defect). Although the treatment group was not specified, ventricular septal defects are commonly associated with poorly controlled diabetes occurring early in gestation (19).
A 1979 Reference described the pregnancy outcomes of 60 obese women who received metformin in the 2nd and the 3rd trimester for diabetes (preexisting [N=39] or gestational [N=21]) that was not controlled by diet alone (14). The drug was not effective in 21 (54%) and 6 (29%) of the patients, respectively. The pregnancy outcomes, in terms of hyperbilirubinemia, polycythemia, necrotizing enterocolitis, and major congenital abnormalities, were not better than those observed in another group of insulin-dependent and noninsulin-dependent diabetic women treated by the investigators, except for perinatal mortality. None of the newborns had symptoms of hypoglycemia. The three infants with congenital malformations had defects (two heart defects and one sacral agenesis) that were most likely caused by poorly controlled diabetes occurring early in gestation.
In another report, metformin in combination with diet (N=22) and glyburide (N=45) was used during pregnancy for the treatment of preexisting diabetes (15). No cases of lactic acidosis or neonatal hypoglycemia were observed with metformin and diet alone, but the latter complication did occur when glyburide was added. Neonatal hypoglycemia is a well-known complication of sulfonylurea agents if they are ingested too close to delivery (see Glyburide). In an earlier Reference, a total of 56 patients had received metformin up to 24 hours of delivery without adverse effects in the newborns (12).
Metformin was used during the 1st trimester in 21 pregnancies described in a 1984 report (16). A minor abnormality (polydactyly) was observed in a newborn whose mother had taken metformin and glyburide. Based on this and previous experience, these investigators proposed a treatment regimen for the management of women with noninsulin-dependent diabetes mellitus (NIDDM) (i.e., type II diabetes mellitus) who become pregnant consisting of diet and treatment with metformin and glyburide as necessary (17). If this regimen did not provide control of the blood glucose, a change to insulin therapy was recommended (17). These investigators concluded that the drug therapy was not teratogenic and did not cause ketosis and that neonatal hypoglycemia was preventable if the therapy was changed to insulin before delivery.
A 1990 Reference addressed the problem of treating diabetes in tropical countries where the availability of medical support and facilities is poor (18). The author recommended that gestational diabetes, not responding to diet alone, be treated with sulfonylurea agents or metformin or both because the initiation of insulin therapy in developing countries was difficult.
The fetal effects of oral hypoglycemic agents on the fetuses of women attending a diabetes and pregnancy clinic were reported in 1991 (19). All of the women (N=21) had NIDDM and were treated during organogenesis with oral agents (1 with metformin, 2 with phenformin, 17 with sulfonylureas, and 1 with an unknown agent), with a duration of exposure of 328 weeks. A control group (N=40) of similar women with NIDDM who were also attending the clinic, matched for age, race, parity, and glycemic control, was used for comparison. Both groups of patients were changed to insulin therapy at the first prenatal visit. From the study group, 11 infants (52%) had congenital malformations, compared with 6 (15%) from the controls (p<0.002). Moreover, six of the newborns from the study group (none in the control group) had ear defects, a malformation that is observed, albeit uncommonly, in diabetic embryopathy (19). No defects were seen in the one infant exposed to metformin. Sixteen live births occurred in the exposed group, compared with 36 in controls. The groups did not differ in the incidence of hypoglycemia at birth (53% vs. 53%), but three of the exposed newborns had severe hypoglycemia lasting 2, 4, and 7 days, respectively, even though the mothers had not used the oral agents close to delivery. In one of these cases, the mother had been taking metformin 1500 mg daily for 28 weeks. Hyperbilirubinemia was noted in 10 (67%) of 15 exposed liveborn infants, compared with 13 (36%) of the controls (p<0.04), and polycythemia and hyperviscosity requiring partial exchange transfusions were observed in 4 (27%) of 15 exposed vs. 1 (3.0%) control (p<0.03) (one exposed infant was not included in these data because the child presented after completion of the study).
In summary, although the use of metformin may be beneficial for decreasing the incidence of fetal and/or newborn morbidity and mortality in developing countries where the proper use of insulin is problematic, insulin is still the treatment of choice for this disease. Moreover, insulin, unlike metformin, does not cross the placenta and, thus, eliminates the additional concern that the drug therapy itself is adversely affecting the fetus. Carefully prescribed insulin therapy will provide better control of the mother's blood glucose, thereby preventing the fetal and neonatal complications that occur with this disease. High maternal glucose levels, as may occur in diabetes mellitus, are closely associated with a number of maternal and fetal adverse effects, including fetal structural anomalies if the hyperglycemia occurs early in gestation. To prevent this toxicity, most experts, including the American College of Obstetricians and Gynecologists, recommend that insulin be used for types I and II diabetes occurring during pregnancy and, if diet therapy alone is not successful, for gestational diabetes (20,21).
Breast Feeding Summary
No reports describing the use of metformin during human lactation or measuring the amount excreted in milk have been located. Metformin is excreted in the milk of lactating rats, obtaining levels comparable to those in the plasma (1). Because of its low molecular weight (about 166), the passage of metformin into human milk should be anticipated. The effect on the nursing infant from exposure to this agent via the milk is unknown.
- Product information. Glucophage. Bristol-Myers Squibb, 1997.
- Klepser TB, Kelly MW. Metformin hydrochloride: an antihyperglycemic agent. Am J Health-Syst Pharm 1997;54:893903.
- Shepard TH. Catalog of Teratogenic Agents. 8th ed. Baltimore, MD: Johns Hopkins University Press, 1995:270.
- Schardein JL. Chemically Induced Birth Defects. 2nd ed. New York: Marcel Dekker, 1993:4178.
- Onnis A, Grella P. The Biochemical Effects of Drugs in Pregnancy. Volume 2. West Sussex, England: Ellis Horwood Limited, 1984:179.
- Miao J, Smoak IW. In vitro effects of the biguanide, metformin, on early-somite mouse embryos (abstract). Teratology 1994;49:389.
- Denno KM, Sadler TW. Effects of the biguanide class of oral hypoglycemic agents on mouse embryogenesis. Teratology 1994;49:2606.
- Elliott B, Schuessling F, Langer O. The oral antihyperglycemic agent metformin does not affect glucose uptake and transport in the human diabetic placenta (abstract). Am J Obstet Gynecol 1997;176:S182.
- Elliott B, Langer O, Schuessling F. Human placental glucose uptake and transport are not altered by the oral antihyperglycemic agent metformin. Am J Obstet Gynecol 1997;176:52730.
- Velazquez EM, Mendoza S, Hamer T, Sosa F, Glueck CJ. Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. Metabolism 1994;43:64754.
- Brearley BF. The management of pregnancy in diabetes mellitus. Practitioner 1975;215:64452.
- Jackson WPU, Coetzee EJ. Side-effects of metformin. S Afr Med J 1979;56:11134.
- Coetzee EJ, Jackson WPU. Diabetes newly diagnosed during pregnancy. A 4-year study at Groote Schuur Hospital. S Afr Med J 1979;56:46775.
- Coetzee RJ, Jackson WPU. Metformin in management of pregnant insulin-independent diabetics. Diabetologia 1979;16:24145.
- Coetzee EJ, Jackson WPU. Pregnancy in established non-insulin-dependent diabetics. A five-and-a-half year study at Groote Schuur Hospital. S Afr Med J 1980;58:795802.
- Coetzee EJ, Jackson WPU. Oral hypoglycaemics in the first trimester and fetal outcome. S Afr Med J 1984;65:6357.
- Coetzee EJ, Jackson WPU. The management of non-insulin-dependent diabetes during pregnancy. Diabetes Res Clin Pract 1986;5:2817.
- Gill G. Practical management of diabetes in the tropics. Trop Doct 1990;20:410.
- Piacquadio K, Hollingsworth DR, Murphy H. Effects of in-utero exposure to oral hypoglycaemic drugs. Lancet 1991;338:8669.
- American College of Obstetricians and Gynecologists. Diabetes and pregnancy. Technical Bulletin No. 200, December 1994.
Coustan DR. Management of gestational diabetes. Clin Obstet Gynecol 1991;34:55864.