Cross‑Sectional Analysis of Pregnancy Labeling: When “Limited Data” Unjustifiably Becomes a Global Barrier to Necessary Treatment for Pregnant Women

Prescribing information (PI) serves as the primary regulatory guidance for clinicians treating pregnant women, yet often lacks comprehensive pregnancy data, leading authorities to adopt restrictive language or default contraindications. This cross-sectional study assessed consistency in pregnancy risk communication across five major regulatory authorities: FDA, EMA, TGA, Canada and MHRA. Latest PIs were obtained from official databases of the five regulatory authorities. Only English-language PIs available for at least four authorities were included. Each PI was evaluated for pregnancy contraindications, reported risks (miscarriages, major malformations), and overall tone (permissive/restrictive). McNemar and Fleiss' multi-rater Kappa methods were used to assess the authorities' agreement level. Pregnancy contraindications on EMA, TGA, Canada, and MHRA labels ranged from 18.7%, 20.1%, 23.1%, 21.8%, respectively, versus 7.4% on FDA labels (P<0.001). Restrictive phrasing often appeared despite limited clinical data. Inter-authority agreement was poor for restrictive recommendations (kappa 0.29, 95%CI:0.20-0.38). It appeared more often in EMA (64.5%), TGA (56.7%), Canada (74.5%), and MHRA (60.6%) than in FDA (34.9%) (P<0.001), though references to pregnancy safety data were similar. These inconsistencies reveal regulatory overreach that may deny patients necessary treatments, representing systemic bias against treating pregnant women. The findings underscore the need for harmonized, transparent, evidence-driven regulatory guidance distinguishing “absence of evidence” from “evidence of harm”.