OBJECTIVES : To describe the distribution of pregnancy exposure reports submitted for insulin therapies in the US

METHODS : The FDA Adverse Event Reporting System (FAERS) was used to identify maternal exposure reports spontaneously submitted between Q2/2000 and Q2/2017 for insulin products that have pregnancy exposure events. Insulin products identified by generic names, and MedDRA preferred terms (PT) corresponding to fetal and maternal exposures were used to create a maternal exposure (ME) custom term. Reporting Rates (EBGM) and 95%CI were used to assess insulin-related ME during pregnancy, and compared to those in other anti-diabetes medications (ADM). Rates with 95%CI lower limit ≥2.0 are considered significant insulin-ME associations.

RESULTS : During the analysis period, a total of 3,912 and 1,253 ME reports were submitted for all insulins and other ADM, respectively. Compared to other ADM, insulins showed significant reporting rates of ME (EBGM=0.38; 95%CI=0.37-0.40 and EBGM=2.23; 95%CI=2.14-2.32, respectively). Number of ME reports and reporting rates for individual insulin products were: aspart (n=522; EBGM=5.06; 95%CI=4.72-5.42), NPH (n=29; EBGM=4.55; 95%CI=3.32-6.11), detemir (n=360; EBGM=4.54; 95%CI=4.16-4.95), human (n=128; EBGM=3.57; 95%CI=3.08-4.12), glulisine (n=80; EBGM=2.38; 95%CI=1.97-2.84), degludec (n=37; 2.25; 95%CI=1.71-2.93), lispro (n=1153; EBGM=2.05; 95%CI=1.95-2.15), and glargine (n=562; EBGM=1.40; 95%CI=1.30-1.50). Numbers and reporting rates of specific PT for insulins were: ME timing unspecified (n=21; EBGM=1.91), ME during pregnancy (n=1769; EBGM=2.60), ME before pregnancy (n=30; EBGM=1.23), maternal drugs affecting fetus (n=371; EBGM=1.30), fetal exposure during pregnancy (n=835; EBGM=2.07), and fetal exposure during breast feeding (n=217; EBGM=2.72). Corresponding PT-specific EBGM for other ADM were: 0.23; 0.46; 0.17; 0.45; 0.30; and 0.14.

CONCLUSIONS : Reporting of ME for insulin therapy appears to reflect the guidelines of diabetes management during pregnancy. These events do not reflect safety outcomes, rather a proxy for insulin treatment during pregnancy. Signal detection of congenital anomalies or fetal exposures in diabetes therapies using FAERS should be interpreted with caution due to confounding by indication.