(319) The PregVal Project. A Population-based Cohort to Assess Drug Use and Safety During Pregnancy.

Background: Many existing pregnancy cohorts used to assess in-utero drug safety are based on data on livebirths only, have varying mother-infant linkage rates; representativeness of the base population also varies, and certain key variables are lacking.

Objectives: To build a pregnancy cohort inclusive of the entire population within a universal health care system, including several features previously unavailable, to present key information available and describe the main characteristics.

Methods:

Design: Retrospective population-based cohort, comprised by all pregnant women and their offspring (Jan 1, 2009 -Dec 31, 2021), which will be followed from 1y pre-conception until end of study or death. Data source: Valencia Health System Integrated database (VID), Spain, covering a  5 million population; Statistical analyses: 1) Descriptive analysis of the characteristics of the cohort (sociodemographics, chronic comorbidity and health services use at conception, and pregnancy-related diseases); 2) Rates of pregnancy outcomes (congenital anomalies, abortion, elective termination, stillbirth (SB), and among live births (LB): low birth weight, preterm birth); 3) Frequency of drug use by therapeutic group.

Results: We identified 620107 pregnancies during the study period. Mean age of the cohort was 31.4 y, 26.2% were foreigners, 17.8% were at risk of social exclusion. Regarding clinical variables, 2.1% had chronic hypertension,1.7 % diabetes, 7.1% depression and 6.4% obesity when entering pregnancy; 6% had diagnosis of gestational diabetes, 0.9% gestational hypertension and 1.1% preeclampsia; median number of primary (including prenatal) care visits was 17 (IQR 6-26). Rates of pregnancy outcomes were: 1.38% congenital anomalies (1.8% for LB, 4.1 for SB); 18.6% spontaneous abortion; 2.5% elective termination; 4.3% unknown abortion (spontaneous/ elective); 0.37% stillbirth; 73.9% live birth, and from them, 9.1% preterm birth and 8.8% low birth weight. Ongoing pregnancies and lost to follow up accounted for 0.24% and 0.07% of the cohort. Regarding drug use, we found that 56.4% of the cohort had at least 1 medicine dispensed (62.2% for LB). Median number of active ingredients dispensed was 1 (IQR 0-3). The most used ATC’s were: B03 antianemics 30%, J01 antibacterials 20% and N02 analgesics 18%.

Conclusions: The population-based nature of PregVal, the ability to reliably detect live births as well as pregnancy losses, and to link mothers with their offspring with long-term follow up, the availability of a large number of covariates and outcomes, and the possibility to estimate drug exposure and its timing accurately, make of PregVal a cohort with great potential for the generation of evidence on the use and safety of medications during pregnancy.