[Restrict the use of macrolides in pregnant women: The advantages and disadvantages of research with data from health care databases].

In this Commentary on a British study of congenital malformations after use of macrolide antibiotics by pregnant women with data from general practitioners, the importance of high-quality healthcare databases for pharmacovigilance is emphasized. The authors of the British study compared three groups during the study period January 1990 through June 2016: 104,605 children from mothers treated during pregnancy with a macrolide or penicillin, a group of 53,735 siblings, and 82,314 children whose mothers were treated with a macrolide before conception. A significantly increased risk of 55% of congenital malformations was seen for use of macrolides during the first semester in comparison to use of penicillins, especially cardiovascular malformations. Although this well-performed study is the largest up till now, some potential limitations are discussed such as potential exposure and outcome misclassification, as well as confounding by indication.