Valproate for Alcoholics With Bipolar Disorder

To the Editor: Valproate was recently approved by the Food and Drug Administration for the treatment of acute mania associated with bipolar disorder. A rare and potentially fatal complication from valproate therapy is hepatotoxicity (1). This complication may be especially important in alcoholics because of the risk of preexisting liver disease as a result of excessive alcohol consumption. Additionally, chronic alcohol use may cause lower white blood cell and platelet counts, which may also complicate the use of valproate with alcoholics. We report on liver function as well as platelet and white blood cell counts in 20 patients with bipolar disorder and alcoholism who received valproate therapy for as long as 2 years.

The charts of 20 patients (12 men and eight women) with comorbid bipolar disorder and alcohol abuse/dependence who had been prescribed valproate were reviewed, and results of the following laboratory tests were evaluated: alkaline phosphatase, alanine aminotransferase, lactate dehydrogenase, γ-glutamyltransferase, aspartate aminotransferase, total bilirubin, WBC count, and platelet count. All patients had baseline test results and at least one set of follow-up laboratory test results for comparison. The patients were followed for an average of 5 months. Their average age was 38.7 years (SD=8.5). The patients had an average valproate level of 69.9 mg/liter (SD=15.8), with an average daily dose of 1562.5 mg (SD=499.1).

Laboratory test results were divided among the following time frames for comparison: baseline, less than 2 months after valproate treatment was started, 2 to 8 months, and more than 8 months. All laboratory test results were within normal range at baseline except the γ-glutamyltransferase. There were no statistically significant changes from baseline test results for any of the liver transaminases or the WBC count. There was a statistically significant decrease in platelet count from an average of 286.6×10³/µl at baseline to 229.5×10³/µl at follow-up; the decrease was evident by 1 month. In no case, however, did platelet counts fall below the normal range.

None of the patients had evidence of preexisting liver disease when the valproate regimen was initiated. Most of these individuals received treatment for bipolar disorder and substance use, and many decreased their substance use significantly during the review period. However, in a subgroup of eight individuals identified through patient progress notes who continued to drink during valproate therapy, there were also no significant elevations in their level of liver transaminases. Individuals with alcoholism in the current project exhibited elevated γ-glutamyltransferase levels, suggesting that they may have had some degree of alcoholic fatty liver disease. Despite this effect, individuals did not continue to develop higher elevations of liver transaminase levels. In fact, the level of liver transaminases tended to decrease over time.

From these data, it appears that moderate doses of valproate (with an average blood level of approximately 70 mg/liter) in alcoholics without significant impairment of liver function do not cause significant adverse effects on WBC count, platelet count, or liver transaminase level.