medwireNews: The risk for a birth defect may be increased if the father is taking metformin around the time of spermatogenesis, say researchers.
“The observed effect size is similar to that of maternal age greater than 45 years, a recognized risk factor, with 4.8% birth defects among liveborn singletons in our data,” write Maarten Wensink (University of Southern Denmark, Odense) and study co-authors in the Annals of Internal Medicine.
The team used Danish nationwide registry data from 1997 to 2016 to identify 1,116,779 singleton births to mothers without diabetes or hypertension. Of these babies, 7029 had fathers who were taking diabetes medications at the time of spermatogenesis, defined as during the 3 months prior to conception.
The most common exposure was to insulin (n=5298). Among babies with this exposure, the prevalence of major birth defects according to the European Concerted Action on Congenital Anomalies and Twins guidelines was 3.3%, which was identical to the rate in the 1,109,750 men who were not taking diabetes drugs.
And the sex ratio among insulin-exposed babies was also virtually identical to that in the unexposed group, with 51.3% and 51.4%, respectively, being male.
By contrast, the birth defect rate was higher, at 5.2% and 5.1% in the 1451 babies exposed to metformin and the 647 exposed to sulfonylureas (with 44.2% also exposed to metformin), respectively. In addition, these babies had an altered sex ratio, with 49.4% and 49.3%, respectively, being male.
There were only 276 exposures to other diabetes medications, so the researchers did not investigate these further.
Unlike the fathers taking insulin, those taking metformin or sulfonylureas were older than men not taking diabetes medications, had a lower income level, and were less well educated.
But after adjusting for these variables, as well as birth year, and maternal age, smoking status, and education, the odds ratio for birth defects associated with metformin use was a significant 1.40. That associated with sulfonylureas was a nonsignificant 1.34, whereas there was no sign of an association with insulin use.
In an analysis to assess the potential impact of unmeasured confounders, the researchers found that a single confounder would have to be associated with both metformin and birth defects with an odds ratio of at least 2.17 in order to fully account for the relationship seen in this study.
“Because this is a strong association, it seems difficult to attribute the observed association entirely to unobserved confounding,” they say, although they note the possibility that it could weaken the strength of the association or render it nonsignificant.
The team also found that type 2 diabetes itself did not account for the link between metformin and birth defects, the rate of which was 3.1% in the absence of metformin but 4.6% when exposure occurred. Moreover, the association between metformin exposure and birth defects was observed only if the father was taking it at the time of spermatogenesis – not if he took it before or after this period.
In a linked editorial, Germaine Buck Louis (George Mason University, Fairfax, Virginia, USA) says the study authors “deserve commendation for addressing a critical data gap regarding paternal exposures and birth defects.”
She calls for more research, noting that although links between medications and birth defects is nothing new, metformin’s “attributable risk could be large given its prevalent use.”
Buck Louis also highlights the need for clinical guidance for pregnancy planning where the prospective father has type 2 diabetes.
“Important in this guidance will be communicating that the adverse relationship was specific to metformin during the period of spermatogenesis,” she says.
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2022 Springer Healthcare Ltd, part of the Springer Nature Group
Ann Intern Med 2022; doi:10.7326/M21-4389
Ann Intern Med 2022; doi:10.7326/M22-0770