Valproate—is withdrawal worth it?

Abstract

Morbidity and Mortality Risks Associated With Valproate Withdrawal in Young Adults with Epilepsy

Mbizvo GK, Bucci T, Lip GYH, Marson AG. Brain. 2024;147(10):3426-3441. doi: 10.1093/brain/awae128. PMID: 38657115

Valproate is the most effective treatment for idiopathic generalized epilepsy. Current guidance precludes its use in women of childbearing potential, unless other treatments are ineffective or not tolerated, because of high teratogenicity. This risk was recently extended to men. New guidance will limit use both in men and women aged <55 years, resulting in withdrawal of valproate from men already taking it, as occurs for women. Whether there are risks of personal harm (including injury or death) associated with valproate withdrawal has not yet been quantified for men or women on valproate, meaning clinicians cannot reliably counsel either sex when discussing valproate withdrawal with them, despite that this concern may be at the forefront of patients’ and clinicians’ minds. We assessed whether there are any morbidity or mortality risks associated with valproate withdrawal in young men and women. We performed a retrospective cohort study of internationally derived electronic health data within the TriNetX Global Collaborative Network. Included were men and women aged 16–54 years with ≥1 epilepsy disease or symptom code between 01/12/2017-01/12/2018 and ≥2 valproate prescriptions over the preceding two years (01/01/2015-30/11/2017). 5-year propensity-matched risks of mortality and a range of morbidity outcomes were compared between those remaining on versus withdrawn from valproate during the 01/12/2017-01/12/2018 recruitment period, regardless of whether switched to another antiseizure medication. Survival analysis was undertaken using Cox-proportional hazard models, generating hazard ratios (HRs) with 95% confidence intervals (CIs). 8991 men and 5243 women taking valproate were recruited. 28% of men and 36% of women were subsequently withdrawn from valproate. Valproate withdrawal was associated with significantly increased risks of emergency department attendance (HRs overall: 1.236 (CI 1.159-1.319), men: 1.181 (CI 1.083-1.288), women: 1.242 (CI 1.125-1.371)), hospital admission (HRs overall: 1.160 (CI 1.081-1.246), men: 1.132 (CI 1.027-1.249), women: 1.147 (CI 1.033-1.274)), falls (HRs overall: 1.179 (CI 1.041-1.336), men: 1.298 (CI 1.090-1.546)), injuries (HRs overall: 1.095 (CI 1.021-1.174), men: 1.129 (CI 1.029-1.239)), burns (HRs overall: 1.592 (CI 1.084-2.337)), and new-onset depression (HRs overall 1.323 (CI 1.119-1.565), women: 1.359 (CI 1.074-1.720)). The risk of these outcomes occurring was 1-7% higher in those withdrawn from valproate than in those remaining on valproate. Overall, valproate withdrawal was not associated with increased mortality. These results may help patients and clinicians have a more informed discussion about personal safety when considering valproate withdrawal.

Commentary

Valproate has been approved for use in epilepsy since 1983, and was later approved for the treatment of mania associated with bipolar disorder (1995) and for migraine prevention (1996). It is considered the gold standard for treatment of idiopathic generalized epilepsy. That said, it is used with extreme caution in women of childbearing potential, largely due to its association with teratogenicity and implications for neurodevelopment. New guidance put forth by medical regulators in the UK, limits the use of valproate in both women and men < 55 years of age. Cautions in use were extended to men after identification of a 5.6% to 6.3% risk for neurodevelopmental disorders and for transgenerational gene expression alterations in the offspring of men taking valproate. ¹ In addition, reduced male fertility was reported and noted to be reversible upon valproate discontinuation. However, there are also consequences to stopping valproate. Previous studies have demonstrated worsening seizure control when valproate was withdrawn, regardless of whether or not patients were switched to another ASM. ² Taking this into account, more information and guidance is needed to help direct clinical decision making regarding the potential risks of withdrawal of valproate as well as the implications for alternative therapy.

The article highlighted in this commentary is the largest known study assessing the risks and implications of valproate withdrawal in people with epilepsy (PWE), looking at associated risk of increased morbidity and mortality in both men and women, regardless of whether or not they were switched to another ASM. ³ It was an international cohort study that consisted of 14 412 PWE, aged 16 to 54 years (mean age 27 years), on valproate that was drawn from a pool of 112 healthcare organizations globally between 2008 and 2023. The participants had a predominant diagnosis of generalized epilepsy. Of the 14 412 PWE on valproate, 4436 13 (31%) were subsequently withdrawn from treatment and 9976 (69%) remained on treatment; 6501 (72%) men remained on treatment and 3363 (64%) women remained on treatment. Of the 107 women taking valproate who subsequently became pregnant during recruitment, 63% (68) were withdrawn from treatment and 36% (39) remained on valproate. Outcomes were collected for 5 years, starting a day after the first epilepsy disease/symptom code appeared without and with co-prescribed valproate.

The results demonstrated that mortality was not increased for men (HR 1.117, CI 0.879-1.419) nor women (HR 1.013, CI 0.769-1.335) withdrawn from valproate. Men and women withdrawn from valproate were at significantly increased risk of experiencing one or more ED visits (male HR 1.181, CI 1.083-1.288, female HR 1.242, CI 1.125-1.371), one or more hospital admissions (male HR 1.132, CI 1.027-1.249, female HR 1.147, CI 1.033-1.274), one or more injuries (HR 1.095, CI 1.021-1.174), and of developing new-onset depression (HR 1.323, CI 1.119-1.565) over the subsequent 5 years. When looked at separately by gender, increased risk of new-onset depression remained for women but not for men (female HR 1.359, CI 1.074-1.720, male HR 1.057, CI 0.837-1.336). There was no evidence of increased risk of self-harm or suicide attempts in those withdrawn from valproate. For the remaining risks assessed, falls and burns were 1% to 2% higher in the group withdrawn from valproate. The data did not show a difference in one or more seizure-coded healthcare consultations between those withdrawn from versus remaining on valproate. Overall, the data showed no evidence to support a statistically significant increased risk of death in PWE withdrawn from valproate.

In summary, the findings in this study demonstrated that PWE withdrawn from valproate were at increased risk of experiencing one or more ED visits, hospital admissions, falls, injuries, burns, and new-onset depression. Of these, risks of ED and/or hospital admissions were the highest, around 7% more in those withdrawn from valproate, and similar for men and women. This is the first time these wider morbidity outcomes have been assessed in PWE within a context of valproate withdrawal for men and women. Given this is the largest study of valproate withdrawal to date, drawing from data from around the world, it allows for more generalizable conclusions to be made from real-world clinical practice. These findings will help facilitate a broader conversation about the potential risks and benefits of valproate use and discontinuation with patients. This study is also the first dedicated large-sample assessment of valproate withdrawal in men. These findings help inform discussions assessing a broader range of withdrawal risks beyond seizures alone and can help clinicians and patients have more comprehensive discussions about safety implications when considering valproate withdrawal.

A major limitation of this study includes the possibility that not all breakthrough seizures were captured by healthcare coding. A possible inaccurate assessment of seizure recurrence is demonstrated by the fact that there were increases seen in ED visits, hospital admissions, falls, injuries, and burns with no associated increase in seizure-coded healthcare consultations. This may be explained by the fact that recurrent seizures in everyday practice often do not result in urgent health care encounters. The fact that seizure recurrence could have been missed or not clearly measured and thus underestimated, is an important point which is critical to patient and provider decision-making concerning risks of valproate withdrawal. Also, regarding implications on mortality, worsening mortality after valproate withdrawal was not found to be statistically significant. However, it's important to note that data sourced from EHRs alone may underestimate deaths due to some occurring outside of healthcare information networks. In addition, direct head-to-head comparisons of outcomes between withdrawal of valproate and withdrawal of other ASMs was not provided as well as recommendations for alternative treatment options. It would be helpful to have data comparing morbidity and mortality between different ASMs switched to from valproate to better understand which alternatives may be safest and most effective. Medication fill data was also unavailable, making it more difficult to ascertain whether participants were fully adherent to their medications as prescribed. That said, the best marker for adherence is plasma drug levels and detectable plasma valproate levels were provided at the last reading taken during the 5-year follow-up for 97.4% of participants remaining on valproate.

This publication reinforces the importance of addressing potential risks of valproate withdrawal in both women and men and emphasizes the need for more informed discussions. While valproate should not be the first-line option in young women of childbearing potential, it is still used, and for some, it's the only drug that provides significant seizure control. The risk of never offering this medication to patients who have poorly controlled epilepsy warrant discussion as well. This commentary enforces our need to gather as much information as possible to help make joint patient/provider informed choices. All patients should be counseled not only about the risks to fertility and increased teratogenicity while remaining on valproate, but also on potential harms associated with valproate withdrawal, including the risks of increased ED attendance/hospital admissions, falls, injuries, burns, and new-onset depression. Further studies are needed to fully understand whether mortality increases in association with valproate withdrawal, as a difference was not found to be statistically significant. Additionally, studies are needed to clarify the rate of seizure increase or recurrence following valproate withdrawal. Overall, this study highlights the necessity and complexity of shared decision-making surrounding valproate withdrawal in young men and women, emphasizing that there are risks in both directions that need to be thoroughly considered.