Dispensing of Potentially Harmful Prescription Drugs in 1.8 Million Pregnant Women in France: A Nationwide Study Based on Two Risk Classification Systems.

INTRODUCTION

Nationwide prevalence of potentially harmful drug prescribing during pregnancy is unknown in France, and several risk classification systems (RCS) exist to guide prescribers.

OBJECTIVE

The aim of this study was to estimate the nationwide prevalence of potentially harmful drug prescribing during pregnancy in France and to describe maternal characteristics associated with this prescription.

METHODS

This drug utilisation study, conducted on the French health databases (67 million beneficiaries), included all pregnancies beginning in 2016-2017, regardless of pregnancy outcome. Potentially harmful drug prescribing was defined as at least one reimbursement during pregnancy of Swedish RCS category D drugs, Australian RCS category D/X drugs, or contraindicated drugs in France for drugs not listed in these two RCSs. Maternal characteristics associated with potentially harmful drug prescribing were described using a univariate logistic regression analysis.

RESULTS

Among the 1,844,447 pregnant women identified, the prevalence of potentially harmful drug prescribing was higher according to the Australian RCS (3.9%) than according to the Swedish RCS (2.2%), with good agreement between the two RCSs (Kappa = 0.81 [0.74-0.87]). This prevalence increased to 9.2% and 6.9%, respectively, when considering contraindications in France. Prescribing of teratogenic drugs, including retinoids and valproate, was highest during the first trimester, whereas prescribing of foetotoxic drugs decreased after the first trimester but remained high for nonsteroidal anti-inflammatory drugs (N = 10,021). In women with no chronic diseases, polymedication (five or more drugs) was the strongest maternal characteristic associated with potentially harmful drug prescribing in both RCSs.

CONCLUSIONS

Potentially harmful drug prescribing during pregnancy is not uncommon in France. This study supports the comparative analysis of RCS to assess potentially harmful drug prescribing in claims databases.