Title:Paroxetine Use During Pregnancy and Perinatal Outcomes Including Types of Cardiac Malformations in Quebec and France: A Short Communication
Volume: 7
Issue: 3
Author(s): Anick Berard, Odile Sheehy, Christine Damase-Michel and Sophie Crespin
Affiliation:
Keywords:
Paroxetine, pregnancy, cardiac malformations, spontaneous abortions, Quebec Pregnancy Registry, EFEMERIS
France, detection bias
Abstract: Background: Antidepressants, more specifically SSRIs, represent one example of a widely prescribed class of
medications in pregnant women for which less than adequate pregnancy data have been available since the first drug in
this class was marketed 20 years ago. Moreover, findings from studies performed after 2005, when health governmental
authorities issued warnings regarding first trimester exposure to paroxetine and the risk of cardiac malformations, may be
the result of detection bias if physicians were investigating more their pregnant patients that used paroxetine than the
others.
Objectives: To estimate the prevalence of 1) paroxetine use during pregnancy, and 2) diagnosed cardiac malformations in
the Quebec and France populations.
Methods: Two distinct pregnancy databases were used for this ecologic study: the Quebec Pregnancy Registry and the
French EFEMERIS database.
Results: In Quebec, among the 109,344 eligible pregnancies, 1,612 (1.5%) were exposed to paroxetine. Prevalence of
paroxetine use during pregnancy increased from 0.7% to 1.2% between 1998 and 2003, simultaneously to the increase of
the prevalence of cardiac malformation diagnoses. In France, among 40,317 eligible pregnancies, 173 (0.4%) were
exposed to paroxetine; between 2004 and 2008 the number of paroxetine fillings and cardiac malformation diagnoses
remained constant.
Conclusions: Despite differences in the Quebec and French healthcare systems, increase in paroxetine prevalence use
during pregnancy was already consistent with an increase in the prevalence of cardiac malformations, even before the
warning on the risk of cardiac malformations in newborns in 2005-2006, limiting the possibility of detection bias.
