Hepatic adverse effects of valproic acid: Management strategies

Abstract

INTRODUCTION:

Valproic acid (VPA) is a widely prescribed first-generation antiepileptic drug. It can induce different liver adverse effects (AEs) which range from mild to severe.

MATERIALS AND METHODS:

A retrospective study was conducted by collecting 9-year reports (from January 2015 to June 2023) of VPA-associated hepatic AEs. It was collected in the Regional Center of Pharmacovigilance of Sfax (Tunisia). Causality was evaluated with the French Bégaud et al. imputability method. Plasma VPA concentrations were measured using an automated enzyme immunoassay (INDIKO).

RESULTS:

We collected the eight cases of hepatic cytolysis with varying severity: Six had mild elevations of liver enzymes, one presented with moderately severe cytolysis and one case was fatal. There management strategies were guided by both transaminase levels and VPA concentrations. For those with alanine aminotransferase levels ≥5 times the upper limit of normal (ULN), we conducted the drug discontinuation. However, for patients with lower levels of transaminases, plasma VPA level measurement was recommended. The concentrations were within or near the ULN. Management consisted of either withdrawal or dose reduction, depending on associated clinical symptoms. Additional investigations to rule out nondrug-related causes performed in five cases were negatives. Most patients recovered favorable outcome after discontinuation or VPA dose reduction.

CONCLUSION:

Management of VPA-induced hepatic cytolysis requires integrating the severity of liver enzyme elevations, plasma drug level measurements, and the results of etiological investigation to guide the clinical decisions.

Introduction

Valproic acid (VPA) is a widely used first-generation antiepileptic drug (AED), indicated for epilepsy and certain psychiatric disorders.[1] Moreover, VPA can cause a range of adverse effects (AEs), including gastrointestinal, neurological, hematological, and hepatic.[1,2] Among these AEs, liver injury is frequent varying from mild, asymptomatic increases in transaminases to severe fulminant hepatitis.[3,4,5] Mild elevations of serum aminotransferases are the most frequent hepatic AEs associated with this AED.[6] However, serious life-threatening hepatitis can also occur.[7] Multiple mechanisms may explicate hepatotoxicity: It can be dose-dependent or idiosyncratic leading to acute liver failure.[8] These AEs are closely linked to the hepatic metabolism of VPA. This AED is metabolized by glucuronidation and mitochondrial β-oxidation with a smaller fraction (10% of the dose) processed through cytochrome P450 (CYP)-mediated oxidation.[9] Disruption of any of these three pathways contributes to hepatotoxicity. VPA, through CYP-mediated oxidation, can generate oxidative stress in hepatocytes with the formation of reactive metabolites including 4-ene-valproate. This metabolite impairs mitochondrial function and induces the accumulation of reactive oxygen species. Further, the inhibition of mitochondrial β-oxidation of fatty acids exacerbates oxidative stress.[7,10,11,12,13] In addition, VPA reduces L-carnitine availability, impairing so mitochondrial fatty acid transport and causing micro-vesicular steatosis.[14,15] In patients with L-carnitine deficiency, the metabolism of VPA deviate to microsomal oxidation producing toxic metabolites (2-n-propylpentanoic acid; 4-ene-valproate;…) which may interfere with the urea cycle and occasionally lead to hyperammonemia encephalopathy a rare but serious complication.[16] An immune-allergic mechanism, as the part of AED hypersensitivity syndrome, is infrequently reported with VPA. When present, liver injury can range from mild enzyme elevation to fulminant hepatitis.[7,17,18]

Reporting hepatic AEs to pharmacovigilance centers is crucial for establishing VPA imputability and guiding the management strategies. The objective of our study was to evaluate hepatic AEs related to VPA, reported to our Regional Center of Pharmacovigilance (RCPV) of Sfax (Tunisia), and to propose the management recommendations suitable for each case.

Materials and Methods

Study design and population

We conducted a retrospective descriptive study at the RCPV of Sfax (Tunisia) at a period of 9 years (from January 2015 to June 2023). The study included all reported cases of hepatic AEs related to VPA with an intrinsic imputability score of at least doubtful (≥I1). This score was calculated according to the French imputability method of Bégaud et al.[19] Cases with a score of I0 were considered as unrelated to VPA and were excluded from the series.

Data collection

From our pharmacovigilance records, we collected the following information for each patient: Demographics (age and sex); prescribed dose; treatment duration; time to onset of liver AEs; clinical presentation (symptoms); liver function tests (alanine aminotransferase [ALT], total bilirubin, alkaline phosphatase [ALP]); concomitant medications; results of etiological investigations (viral serologies, autoimmune markers, and imaging); plasma VPA level measurement; therapeutic management (drug discontinuation or dose adjustment), and the clinical outcome.

Causality study

All included cases were evaluated through the pharmacovigilance French method of Bégaud et al.[19] This method establishes the causal relationship between VPA and the observed liver injury. It is based on the specific evaluation criteria. The chronological (C) criteria: Evaluate the duration from drug administration to the onset of AEs as well as the time to their resolution after drug discontinuation. They result to these C scores: C0: Incompatible; C1: Doubtful; C2: Plausible, and C3: Probable. The semiological (S) criteria: Evaluate the tests suggesting drug-induced AEs as well as the clinical and laboratory investigations to exclude the alternative causes. They result to these S scores: S1: Doubtful; S2: Plausible; and S3: Probable. The intrinsic imputability (I) is determined by combining C and S scores: I0: Excluded; I1: Doubtful; I2: Plausible; and I3: Probable. Finally, extrinsic imputability (B): Evaluates the strength of bibliographic references based on the literature establishing a causal link between drugs and AEs. It includes: B0: Effect not documented; B1: Isolated reports; B2: Plausible (limited but suggestive data); and B3: Well-established effect.

In addition, the Roussel Uclaf Causality Assessment Method (RUCAM)[20] score was calculated for all patients, to standardize the evaluation of drug-induced liver injury (DILI). It includes the seven components: (1) Time to onset of the reaction; (2) course after drug withdrawal; (3) risk factors (age, alcohol use, and pregnancy); (4) concomitant hepatotoxic drugs; (5) exclusion of alternative causes; (6) known hepatotoxicity of the drug; and (7) response to rechallenge. RUCAM leads to these interpretations according to the calculated score: <0: Excluded; 1–2: Doubtful; 3–5: Possible; 6–8: Probable; and 8: Highly probable.

Both methods were applied to each patient. In fact, the imputability method of Bégaud et al.[19] was applied routinely for all the cases consulting our RCPV. While, RUCAM score was calculated specifically to standardize the DILI assessment. In cases where these two methods disagreed, the global clinical-pharmacovigilance evaluation was given the priority to the RUCAM score.

Classification of the severity of liver cytolysis

The severity of liver cytolysis was categorized following the US DILI Network guidelines.[21] The cases were classified as follows:

• Grade 1 (mild): Increases in ALT and/or ALP, total bilirubin <2.5 mg/dL, international normalized ratio (INR) <1.5, no required hospitalization

• Grade 2 (moderate): Elevated liver enzymes (ALT and/or ALP) accompanied by either total bilirubin ≥2.5 mg/dL or INR ≥1.5, without immediate organ failure

• Grade 3 (moderately severe): Elevation of liver enzymes with hyperbilirubinemia and/or coagulation abnormalities, requiring hospital admission or prolongation of stay

• Grade 4 (severe): Elevation in liver enzymes (ALT and/or ALP) associated with total bilirubin ≥2.5 mg/dL and hepatic failure (INR >1.5, ascites, or encephalopathy) or involvement of other organs due to DILI

• Grade 5 (fatal): Outcome resulting in death or liver transplantation due to DILI.

Plasma valproic acid monitoring

Plasma VPA level measurement was performed in patients with suspected overdose. Plasma concentrations were measured in the pharmacology laboratory of the Faculty of Medicine de Sfax (Tunisia) using an enzyme-linked immunoassay (INDIKO device). The samples were collected at least 5 halves-lives after treatment initiation or any dose change and 12 h after the last dose. The therapeutic range of plasma VPA concentrations was 40–100 μg/mL.

Statistical analysis

Given to the retrospective design of the study and the limited number of reported cases, the statistical approach was restricted to only descriptive analysis. All the variables were recorded in absolute values. Data management and analysis were performed using Microsoft Excel. Values were presented as the means ± standard deviation with minimum and maximum values when relevant. Due to the small size of study population, no comparative or inferential statistical tests were conducted.

Results

During the period of study, eight cases of hepatic cytolysis were collected with a mean age of 46.6 ± 11.7 years (range 24–62) with a predominance of males (6 men, 2 women; sex ratio 3). The severity of hepatic cytolysis was varied: Six patients were with mild enzyme elevations (mild group) (cases 2–7), one had moderately severe cytolysis (case 1), and one developed fulminant hepatitis resulting in death (case 8). Only case 2, among the mild group, had ALT levels ≥5 times the upper limit of normal (ULN). The interval between VPA initiation and the onset of cytolysis was suggestive in two patients and compatible in the other six cases [Table 1].

Table 1.

Severity, scores, and management strategies of valproic acid-related hepatic cytolysis and results of liver function tests

Case	Laboratory/clinical data	Severity	Onset	Resolution	Intrinsic imputability	RUCAM	Management strategies
1	ALT=97.2×ULN TB=2.65 mg/dL PT: Not recovered	Moderately severe	10 days	30 days	C3S2I3B3 (R +)	9	VPA contraindicated*
2	ALT=8.5×ULN TB=0.6 mg/dL PT=83%	Mild	7 days	10 days	C3S2I3B3	7	VPA contraindicated*
3	ALT=3.3×ULN TB=0.7 mg/dL	Mild	35 years	7 days	C2S2I2B3	8	Avoid VPA*
4	ALT=3×ULN TB=1 mg/dL	Mild	10 years	7 days	C2S2I2B3	4	Avoid VPA*
5	ALT=3×ULN TB=1.7 mg/dL	Mild	20 days	30 days	C2S2I2B3	5	Avoid VPA*
6	ALT=2.2×ULN TB=0.4 mg/dL	Mild	10 months	1 months	C2S2I2B3	4	Reduce VPA dose**
7	ALT=2×ULN TB=0.35 mg/dL	Mild	10 years	11 days	C2S2I2B3	6	Reduce VPA dose**
8	ALT=182.75×ULN TB=2.9 mg/dL PT=32%	Fatal	10 years	Death	C1S2I1B3	2

*Assess potential concomitant hepatotoxic drugs, **Evaluate concomitant hepatotoxic drugs, plasma VPA levels and liver function tests after dose adjustment. RUCAM score interpretation: <0=Excluded, 1–2=Doubtful, 3–5=Possible, 6–8=Probable, >8=Highly probable. B=Extrinsic imputability, B3=Well-documented effect of the drug in the literature, C=Chronological criteria, C1=Doubtful, C2=Plausible, C3=Probable, I=Intrinsic imputability, I1=Doubtful, I2=Plausible, I3=Probable, S=Semiological criteria, S2=Plausible, R=Re-administration, R +=Positive re-administration. ALT=Alanine aminotransferase, TB=Total bilirubin, ULN=Upper limit of normal, PT=Prothrombin time, VPA=Valproic acid, RUCAM=Roussel Uclaf Causality Assessment Method

Concomitant hepatotoxic medications were investigated in all patients. Paracetamol, at its therapeutic doses, was identified in case 7. Phenobarbital, a known CYP 450 inducer, was co-administered in case 3. These drugs were maintained at the same doses without impact on clinical evolution. In addition, other concomitant nonhepatotoxic drugs and those with no effect on VPA metabolism were also continued.

VPA was discontinued in cases 1, 2, and 8 because ALT exceeded five times the ULN. Liver function normalized in cases 1 and 2 after this discontinuation. Whereas case 8 had a fatal outcome. The resolution times for cytolysis are indicated in Table 1. Plasma VPA concentrations were measured in five patients with mild cytolysis (ALT 2–3 × ULN). In cases 3–5, plasma levels were within the therapeutic range, but the presence of gastrointestinal intolerance conduct to the AED withdrawal with a clinical improvement. In cases 6 and 7, plasma VPA levels were close to the ULN leading to a dose reduction (from 1500 mg/day to 1000 mg/day). After the dose adjustment, the concentrations of VPA were within the therapeutic range near its lower limit. The management strategies algorithm for dose adjustment (cases 3–7) is illustrated in Figure 1. Management strategies for all patients are provided in Table 1.

Figure 1.

Decision algorithm with alanine aminotransferase elevations of 2–3 upper limit of normal

Additional investigations to rule out nondrug-related AEs were found only in five cases. Their availability was dependent on the patients’ clinical presentation and available data in our pharmacovigilance records. Abdominal ultrasound was normal in cases 2 and 7, but diffuse hepatic steatosis was observed in cases 3 and 4. These findings and the results of viral and autoimmune hepatitis serology are summarized in Table 2.

Table 2.

Additional investigations performed in patients with valproic acid-related hepatic cytolysis

Type of liver injury	Investigation	Findings
Liver cytolysis	Abdominal ultrasound	Diffuse hepatic steatosis (cases 3, 4); normal (cases 2, 7)
	Immunological tests	Negative (cases 2, 3, 7)
	Viral hepatitis B and C serology	Negative (cases 2, 3, 7)
Fulminant hepatitis (PT=32%)	Viral hepatitis C serology	Negative (case 8)

PT=Prothrombin time

The causality scores using the French imputability method of Bégaud et al. and the RUCAM are presented in Table 1.

Discussion

In our study, VPA was implicated in only eight cases of hepatic cytolysis which may reflect underreporting of such AEs. In the literature, elevated transaminases have been documented in approximately 5%–10% of patients receiving VPA.[6,22] We reported six cases of mild cytolysis. Only one (case 2), among them, had a transaminase level ≥5 times the ULN requiring VPA discontinuation. Several authors recommend this limit as a criterion for stopping the incriminated drug.[23] Similarly, Meseguer et al.,[10] used this threshold for hepatic enzyme for diagnosing DILI. However, in our series, for five patients with more moderate transaminase elevation (ALT 2–3 × ULN) (cases 3–7), various strategies were employed including dose reduction or drug withdrawal after plasma VPA levels measurement. Two patients (cases 6 and 7) with asymptomatic plasma VPA level near the ULN benefitted from dose reduction resulting in the favorable outcome. In the literature, the measurement of plasma VPA levels and dose adjustments are essential, particularly when transaminases are elevated, even within the therapeutic range.[24,25] Although some reports permit continuation of VPA at the same dose in asymptomatic moderate elevations up to 2–3 × ULN.[2,26] Conversely, three patients (cases 3–5) with symptomatic ALT elevations (gastrointestinal intolerance), despite VPA level within the therapeutic range, required discontinuation of this AED. This follows the recommendations suggesting drug discontinuation or dose reduction in symptomatic liver function changes.[27] Only one patient (case 7) exposed to another hepatotoxic drug, paracetamol at its therapeutic dose. This drug may contribute to VPA hepatoxicity revelation. On the other hand, pharmacokinetic interactions were investigated. In fact, the case 3 described the concomitant use of phenobarbital, a CYP inducer. This drug is known to enhance VPA metabolism into potentially hepatotoxic metabolites such as 4-ene-valproate.[9] Evaluation of the possibility of nondrug causes of cytolysis was also performed. Abdominal ultrasound revealed diffuse hepatic steatosis in cases 3 and 4. This injury is due to VPA-induced inhibition of mitochondrial β-oxidation or due to another etiology like metabolic overload.[28] This type of hepatotoxicity can manifest after weeks to months and occasionally after years of treatment.[29] Abdominal ultrasound in case 7 was normal and autoimmune or viral hepatitis B/C were excluded in cases 3 and 7, excluding probably a nondrug related cause. The literature insists on the importance of excluding other etiologies when suspecting DILI.[28] In these five-mild cytolysis cases (ALT 2–3 × ULN) (cases 3–7), improvement in liver function after VPA discontinuation or dose reduction was in favor of VPA responsibility with plausible imputability scores. RUCAM scores were probable for cases 3 and 7 and possible for cases 4–6. In addition, three patients required VPA discontinuation: Case 1 (moderately severe cytolysis), case 2 (mild cytolysis: ALT ≥5 × ULN), and case 8 (fatal cytolysis). Case 1 had positive re-administration, a probable imputability score, and a RUCAM score of 9, reinforcing VPA’s responsibility. For case 2, the role of VPA was retained in view of suggestive onset and disappearance times and a negative etiological assessment (eliminating autoimmune hepatitis or hepatitis B/C with a normal liver ultrasound). In this case, the imputability score was plausible and the RUCAM score was 7. In case 8, the 10-year time to onset was compatible with VPA-related hepatotoxicity, but the outcome remained inconclusive due to patient death. While fulminant VPA hepatitis usually occurs within the months of treatment initiation,[22,30] delayed cases are reported.[31] Hepatitis C was negative in this patient, but further etiological workup was not possible. RUCAM scoring suggested doubtful VPA involvement (score 2).

These findings highlight the importance of pharmacological therapeutic monitoring for VPA. It is particularly crucial in children, patients on multiple medications, or those with liver failure.[32]

Most studies in the literature, are principally limited to case reports describing hepatic AEs associated with VPA.[7,10,16,33,34,35] Our analysis provides a more comprehensive view, detailing management strategies on a case-by-case hepatic cytolysis induced by this AED. Nevertheless, the small size of our population underscores the need for more collaboration with pharmacovigilance centers. Larger series, particularly incorporating pharmacogenetic evaluations of VPA metabolism, is useful to better characterize and prevent VPA hepatotoxicity.

Conclusion

In our study, VPA was associated with cytolytic liver injury that was principally mild but one case had a fatal outcome. Management strategies were guided by the severity of liver enzyme elevations, the presence of symptoms and plasma VPA concentrations. Mild, asymptomatic elevations allowed continuation with dose reduction of this AED with close monitoring. Whereas symptomatic and severe elevations required discontinuation. Thorough etiological assessment to exclude the alternative causes is essential. These findings underscore the importance of individualized therapeutic management and of regular liver function control in patients treated by VPA.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.