Metformin in the first trimester and risks for specific birth defects in the National Birth Defects Prevention Study

• Published in: Birth defects research Vol. 110; no. 7; p. 579
• Main Authors: Dukhovny, Stephanie, Van Bennekom, Carla M, Gagnon, David R, Hernandez Diaz, Sonia, Parker, Samantha E, Anderka, Marlene, Werler, Martha M, Mitchell, Allen A
• Format: Journal Article
• Language: English
• Published: United States 17.04.2018
• Subjects: Abnormalities, Drug-Induced - prevention & control; Adult; Female; Humans; Hypoglycemic Agents - adverse effects; Logistic Models; Metformin - adverse effects; Pregnancy; Pregnancy Trimester, First; Risk; metformin; National Birth Defects Prevention Study; birth defects; diabetes; oral hypoglycemic agents; polycystic ovarian syndrome
• DOI: 10.1002/bdr2.1199

We assessed associations between first-trimester metformin use for pregestational diabetes and specific major birth defects. We compared risks associated with first-trimester metformin use by diabetic women to nondiabetic women on no diabetes medication; we calculated crude odds ratios by exact logistic regression and adjusted by inverse probability weighting. Confounding by diabetes was assessed by comparing risks for metformin-exposed diabetic women to those for insulin-exposed diabetics and nondiabetics treated with metformin for subfertililty. Among 9,279 nonmalformed controls and 24,375 malformed cases, diabetics who used metformin (with or without insulin) had increased adjusted odds ratios (aORs) for several birth defects associated with diabetes. However, women treated with metformin for subfertility had aORs similar to or lower than those for diabetic metformin users, and many approximated the null. For atrial septal defect secundum, anorectal defects, and limb reduction defects, the estimates for metformin when used for subfertility were 2-3-fold. While metformin use for diabetes was associated with an increased risk of many birth defects, when metformin was used for subfertility most defects had aORs that approximated the null, while only three defects had modestly increased aORs, two of which had lower confidence bounds that included the null. Our study does not suggest that metformin poses an appreciable risk for major birth defects, but further studies are necessary.