“In the Name of the Father: Risk to the Offspring From Paternal Exposure to Valproate During Conception”

Commentary

Epilepsy frequently starts early in life and, hence, it is likely to afflict individuals throughout reproductive age. The anatomic and behavioral impact of antiseizure medications (ASMs) to the fetus through the mother during pregnancy and lactation has been extensively studied. ¹ Yet, the impact of ASMs to the offspring through the father during spermatogenesis has drawn much less attention.^2-5

In addition to women of child-bearing age, ¹ Valproate (VPA) has been under the spotlight also for men with epilepsy. The same question is posed for a host of male patients without epilepsy, who are nevertheless prescribed ASMs for other indications (eg, migraines, mood stabilization, peripheral neuropathy, etc). After retrospective observational data derived from Scandinavian registries raised concerns for a heightened risk of neurodevelopmental disorders to children whose fathers used VPA around the time of conception, governing organizations, such as the European Medicines Agency (EMA) (https://www.ema.europa.eu/en/news/ema-review-data-paternal-exposure-valproate) ⁶ and the United Kingdom Medicines and Healthcare Products Regulatory Agency, (https://www.gov.uk/government/publications/valproate-review-of-safety-data-and-expert-advice-on-management-of-risks)^7,8 have placed the issue under scrutiny. So, should all these men use contraception to avoid fathering a child while on this medication or stop the medication during the time of spermatogenesis (ie, 3 months prior to conception) if they decide to become parents?

The current study ⁹ attempts to elucidate this issue. By using a population-based, healthcare database of men with filled prescriptions of VPA during spermatogenesis, the authors investigated the risk to the offspring for major congenital malformations in the first year of life and neurodevelopmental disorders (NDDs) at the end of follow up. With a median follow-up period of 10 years, no increased risk was identified comparing 1336 VPA exposed children to 1,234,017 VPA unexposed children.

This is a well conducted study on a large, nationally representative sample in Denmark. ⁹ Of note, registries from the same country have previously identified the risks related to exposure to VPA in child-bearing women. ¹⁰ There was careful ascertainment of both exposures (through Anatomical Therapeutic Chemical [ATC] codes) and outcomes (through International Classification of Diseases [ICD] codes), in addition to including evaluation of pertinent variables that could lead to confounding by indication (such as paternal and maternal epilepsy and/or psychiatric disorders), or skew the results (such as birth year as well as parental age and educational level). To strengthen the derived conclusions, a host of subanalyses were undertaken such a dose-response analysis (low vs high VPA dose exposure using 750 mg/d as the cut-off), siblings analysis (VPA exposed vs unexposed children in the same family), restriction analysis (VPA exposed vs unexposed children born only of fathers with epilepsy), matched active comparator analysis using a reproductive age “friendlier” ASM (Lamotrigine [LTG]) as control, and negative exposure control analysis (VPA exposed children close to conception vs children whose fathers used VPA long before the spermatogenesis period). Moreover, a sensitivity analysis excluding other potentially teratogenic drugs administered to the father during spermatogenesis or the mother during pregnancy was conducted. None of these subanalyses identified a worrisome signal. ⁹

On the other hand, reliance on ATC and ICD codes of large healthcare databases always carries some risk of missing or misclassifying exposures and outcomes. In spite of all good intents, the possibility of unmeasured confounding cannot be entirely excluded in any observational study. The lack of information on drug levels could not safeguard adherence or provide a dose-response analysis substantiated by them. Malformation of cortical developments leading to pregnancy terminations or those detected beyond the first year could not be captured. Although the follow-up period for NDDs discovery was nearly a decade long and despite the heightened vigilance in diagnosing these diseases over the course of the years, it is conceivable that some may have escaped timely observation. Some of the potentially harmful factors for brain development (eg, smoking and prenatal insults) were not registered, and, conversely, some of the protective ones (eg, folic acid intake by the mother and prenatal monitoring) were not available. Finally, unlike similar studies performed in mothers, issues related to false paternity are hard to account for, although that would likely affect the VPA exposed versus unexposed cohorts proportionally.

The issue of VPA use in men of child-fathering potential remains unresolved. Male reproductive toxicity could in theory occur through nongenetic (eg, due to the presence of the drug in seminal fluid), genetic (eg, chromosomal abnormality or gene mutation), or epigenetic mechanisms (eg, effect on gene expression). ¹¹ For VPA specifically, there is both animal ¹² and clinical ¹³ data that it causes reversible change in sperm count, morphology, motility and testicular cytoarchitecture. However, whether this translates to heightened risk to the offspring has not been conclusively shown. Despite the precautionary alerts derived from observational data (https://www.ema.europa.eu/en/news/ema-review-data-paternal-exposure-valproate ⁶ and https://www.gov.uk/government/publications/valproate-review-of-safety-data-and-expert-advice-on-management-of-risks),^7,8 recent nationwide population based studies, ⁵ including the current one, ⁹ alleviate such concerns. This controversy may stem from methodological differences between observational studies including study design, exposure and outcome ascertainment, adjustment for potential confounders (particularly those creating confounding by indication) and heterogeneous follow-up period. Until the jury is out, we cannot deprive all our male patients with epilepsy from a medication that is proven to work well, ¹⁵ particularly for certain indications such as generalized epilepsies, ¹⁴ but it would be reasonable to consider first a newer generation alternative with a better side-effect profile toward the reproductive system and beyond. That practice trend is already depicted in the current study where VPA use has remained relatively steady over the course of the two decades, while LTG use has exponentially increased. ⁹

Simplistically put, there appears to be smoke surrounding parental use of VPA during conception with regard to the adverse outcomes to the offspring, but a fire has not been decisively demonstrated. In the name of the father, there is need for more comprehensive, worldwide, and longitudinal assessment of the impact of paternal exposure to ASMs (including VPA) during spermatogenesis, encompassing epigenetic effects and the potential transgenerational risk. That would provide the direly needed guidance on risk minimization strategies to healthcare professionals and patients alike.