Second-generation antipsychotics and pregnancy complications: the impact of confounding by indication
- PDF / 137,162 Bytes
- 2 Pages / 595.276 x 790.866 pts Page_size
- 17 Downloads / 151 Views
LETTER TO THE EDITOR
Second-generation antipsychotics and pregnancy complications: the impact of confounding by indication Olorunfemi A. Oshagbemi 1
&
Abraham Daniel 2 & Ismaeel Yunusa 3
Received: 12 August 2020 / Accepted: 14 October 2020 # Springer-Verlag GmbH Germany, part of Springer Nature 2020
We read with great interest the publication by Ellfolk et al. (2020) [1]. The authors conducted a retrospective cohort study using a population-based database in Finland from 1996 to 2016. The study aimed to evaluate exposure to secondgeneration antipsychotics (S-GAs) and pregnancy complications. In this study, the authors compared S-GAs with firstgeneration antipsychotics (F-GAs) and patients unexposed to any medication. The authors found that patients exposed to SGA are associated with an increased risk of gestational diabetes (GD) compared with those unexposed to any drug. They discovered that S-GAs were not associated with GD in comparison with F-GAs. The authors concluded that S-GAs use during pregnancy was associated with an increased risk of pregnancy complications, including GD. Although via acting directly on glucose homeostasis, it is biologically plausible that antipsychotics can be associated with gestational diabetes. To our knowledge, no observational study confirmed this hypothesis. We believe that the study was impacted by confounding by indication, which needs to be discussed. One of the methodological standards advocated for the evaluation of risks emanating from drug exposures in pharmacoepidemiological studies by the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance is the use of an active comparator group, which has been found to mitigate confounding by indication/ frailty and reduce selection bias [2]. In the study, the authors focused their conclusion on the increased risk of GD observed * Ismaeel Yunusa [email protected] 1
Quantitative Safety and Epidemiology, Novartis Pharma AG, Basel, Switzerland
2
Division of Pharmacy, University of Tasmania, Hobart, Tasmania, Australia
3
College of Pharmacy, University of South Carolina, Columbia, SC, USA
among pregnant women exposed to S-GAs compared with unexposed patients. However, in the secondary analysis using FGAs (active comparators), the increased risk of GD initially observed disappeared (adjusted OR 0.97; 95% CI (0.78– 1.20)). This scenario depicts a typical bias resulting from the use of inappropriate comparators [3]. Table 1 of their study’s maternal characteristics showed that patients in the S-GAs group had a greater proportion of subjects with psychotic and other severe mental disorders for which antipsychotics will most likely be prescribed compared with nonusers. Notably, there were higher proportion of women with pregestational diabetes in the S-GAs group (thus at a higher risk of the GD) than nonusers. Given the obvious systematic error in the contributing cohort exposure estimates and its influence as the dominant source of study error, the authors could have conducted a quantitative bias assessmen
Data Loading...
