(287) Paternal exposure to valproate and the risk of Neuro-Developmental Disorders in offspring: a population-based cohort study in Denmark, Norway and Sweden

Background: Valproate (VPA) is approved to treat epilepsy and bipolar disorder. Risk of neurodevelopmental disorders (NDD) and congenital malformations (CM) in offspring exposed to VPA in utero is well-known. However very few studies focused on the risk of NDD or CM in offspring whose father was exposed to antiepileptic drugs at time of conception.

Objectives: To investigate the risk of NDD (composite of all subtypes) in offspring paternally exposed to VPA vs lamotrigine or levetiracetam (LTG/LEV) in monotherapy at time of conception.

Methods: Offspring born between 1997 and 2018, 2007-2019, 2010-2019 to father exposed to VPA or LTG/LEV within 3 months before child conception were identified in population-based registries with family linkage, in Denmark (DK), Sweden (SE) and Norway (NO), respectively. Outcomes (ICD-10 codes) were observed up to 12 years of age. Country-specific hazard ratios (HRs) of NDD were estimated using Cox PH models, and later combined using meta-analysis. NDD subtypes were not assessed. Odds-ratios (OR) for CM were estimated using logistic regression in DK and NO only, due to database constraints.

Results: 5721 offspring (DK: 793 and 1157, SE: 930 and 1425, NO: 398 and 1018 in VPA and LTG/LEV groups respectively) were included in the comparative analyses. Cumulative risks of NDD in offspring were in DK: 5.6% and 3.2%, SE: 5.6% and 2.5%, NO: 4.0% and 2.3% in VPA and LTG/LEV groups respectively. Propensity Score weighted (PSW) adjusted (a)HRs (95% CI) showed no significant higher risk of NDD in the VPA vs LTG/LEV group (DK: 1.34 (0.79-2.25), SE: 1.54 (0.95-2.51), NO: 1.76 (0.83-3.71)). However, the combined HR showed a significantly higher risk in the VPA group: PSW-aHR= 1.50 (1.09-2.07). Length of follow-up was unbalanced between the 2 groups: respectively 50% and 34% of offspring had ≥8 years of follow-up in the VPA and LTG/LEV groups. Hence, offspring of the VPA group had a higher probability of being diagnosed with NDD.
1161 offspring (DK: 259 and 389; NO: 169 and 344 in VPA and LTG/LEV groups respectively) were included in the comparatives analyses. Crude ORs were in DK: 0.62 (0.37, 1.04), NO: 1.06 (0.62, 1.82), combined OR: 0.81, (0.48, 1.36), I²=49.6%. Only crude ORs were estimated, PS model did not converge in NO.

Conclusions: A borderline statistically significant higher risk of NDD associated with paternal exposure to VPA vs LTG/LEV at time of conception was observed when combining the results from the 3 countries. Due to limitations, especially the difference in follow-up time between the 2 groups, these findings should be interpreted with caution. No difference in risk of CM was found between the 2 paternal exposure groups. However, findings only based on crude estimates, potentially biased and affected by heterogeneity, should also be interpreted with caution.