Abstract
Virta LJ, Kälviäinen R, Villikka K, Keränen T
Eur J Neurol 2018;25:869–874. Documented teratogenic effects of valproate (VPA) prompted
restrictions of its use in females of childbearing age in 2014. We
investigated possible annual changes in the outpatient use of VPA in
Finland during 2008–2016 with a special focus on women. We identified all outpatients with VPA purchases between 2008 and
2016 categorizing users due to epilepsy, bipolar disorder or
miscellaneous indications. Temporal trends in the annual prevalence
rates of VPA use were estimated using Poisson regression
analyses. Between 2012 and 2016, the prevalence of VPA use among women aged
15–44 years decreased by 19%, from 50/10 000 to 40/10 000
(prevalence rate ratio, 0.81; 95% confidence intervals,
0.77–0.91; P < 0.001). The use of VPA
due to epilepsy decreased significantly in females aged 15–24
and 25–34 years and that due to bipolar disorders decreased
significantly in females aged 25–34 and 35–44 years.
The use of VPA in the miscellaneous indication group decreased by
32% after 2014 in females aged 15–44 years and, most
strikingly, by 56% among those aged 15–25 years. In women
with epilepsy, the use of VPA increased among those over the age of
44 years. The rates of female VPA users with childbearing potential have
decreased in all three major indication groups in Finland during
recent years, especially after the European Medicines Agency
restrictions were published in 2014. However, it stillremains open
to question as to whether the practice of VPA use follows current
guidelines. A special concern is the relatively high prevalence of
off-label use of VPA in fertile-aged females.Background and Purpose
Methods
Results
Conclusions
Commentary
The paper by Virta and colleagues represents one of several papers from different regions of the world that are reporting on changes in VPA use in women of childbearing age. The impetus for this topic seems to be the restrictive stance on the use of VPA by the European Medicines Agency. Interestingly, this is not a new concept. The potential negative effect of VPA on the fetus has been known since the 1980s with systematic data being made available from pregnancy registries since the 1990s.
Virta et al. used Finland's extensive prescription record system to report on trends in valproate use in women of childbearing age. This study uses two national register systems: 1) a nationwide Population Register covering permanent residents, containing information of sex, live births, deaths, immigration and emigration, as well as 2) a national health-insurance database—the Drug Purchase Register—covering all prescribed drug purchases reimbursed by the Social Insurance Institute since 1994. These data sources allowed inclusion of over 5 million residents enabling examination of records of almost 32,000 outpatients exposed to valproate in a particular year over a 9-year period (2008–2016). Valproate is sold only by prescription in Finland, and this database contains all valproate dispensed to outpatients. Although data were available for the entire 9-year period, some key findings reported included only a 4-year period (2012–2016).
Patients were divided into three groups: epilepsy, bipolar disorder, or miscellaneous indication. Although the particular use of valproate in the latter group could not be identified, it is assumed that possible uses include psychiatric indications other than bipolar disorder, migraine, or only received valproate for the short-term, less than 3 months, which is the amount of time covered by one prescription. Overall findings show a decrease in use of VPA in children and adolescents overall (<14 years) regardless of sex. Increases are reported among men and women over the age of 44. Almost half of all VPA use was indicated for epilepsy. For the particular age groups, the prevalence of VPA use decreased among women ages 15–44 years. In women with epilepsy ages 15–44 years, use of VPA decreased by 12% from 2012. The bipolar disorder and miscellaneous indication category experienced larger decreases in this age group (22% and 32%, respectively) with the largest drop of 56% being reported for 15–24 years in the miscellaneous group. Significant drops in VPA use with epilepsy were noted for 15–24 and 25–34 years, as well as those with bipolar disorder, dropping at 25–34 and 35–44 years.
A particular strength of this paper is the availability of records from two databases in Finland that provides systematic information on approximately 5.5 million people, representing treatment for patients receiving medications for any indication. The inclusion of such a large and inclusive dataset from one specific geographic area helps approach the true prevalence of VPA use in a population. For instance, although the currently available pregnancy registries provide a wealth of information, they are voluntary, and, therefore, a true accounting of the whole population is not possible. This is not an inferior approach; however, results from any source will have inherent biases and strengths depending on how data are collected and analyzed. The availability of pregnancy registries provides a longitudinal recording of different aspects of AEDs used during pregnancy. Recent analysis from the EUROP pregnancy registry enabled exploration of VPA over 17 years in 7555 prospective pregnancies on monotherapy of eight different AEDs. In the study by Tomson et al. (6), valproate exposure resulted in the highest prevalence (10.3%) for congenital malformations.
The opinion of many is not to use VPA in women of childbearing age if possible. The first malformation linked to valproate in humans was in reported in the early 1980s (3, 4). The effects seem to be dose related, although the critical dose threshold is not clear (6, 7). Although malformations may be detected early within the first year of life, other neurodevelopmental issues may not be apparent until later. In utero exposure to VPA has been associated with decreases in children's IQ compared to other AEDs (2). These results incentivized the European Medicines Agency to strengthen warnings for VPA use in women of childbearing age and recommend new measures in 2018 to avoid VPA exposure in pregnancy.
To summarize, most clinicians would likely agree that VPA should be avoided in women of childbearing age on the basis of several decades of data. One approach is to not initiate VPA therapy in this population, as a woman cannot predict if or when she might become pregnant. A caveat to this is that if VPA is the only medication that works for a specific patient, its use may not be avoidable. In addition, these cautions may not be relevant to women who may not wish to or are unable to become pregnant (e.g., hysterectomy). One must also remember that epilepsy is not the only indication for VPA treatment. If we review the current outcome data for children born to mothers receiving VPA, there are a few take-away messages: If possible, try to avoid VPA in childbearing age women, but if a woman needs VPA, then identify the lowest possible dose.