ABSTRACT
INTRODUCTION: Pretreatment plasma homovanillic acid (HVA) levels have been reported to be a correlate of clinical response to typical antipsychotics for schizophrenic, bipolar manic, and mixed groups of psychotic patients. Biological markers of clinical response to antipsychotics could be useful for optimizing drug treatment. METHOD: Thirty-one consenting acute inpatient subjects between ages 19 and 66 years with a DSM-III-R clinical diagnosis of bipolar disorder, manic with psychotic features were entered into this double-blind study and were randomly assigned to receive either haloperidol 25 mg/day or haloperidol 5 mg for the 3-week study. Subjects also received one of the following concomitant medications: standard lithium, lorazepam 4 mg/day, or placebo. RESULTS: The primary multiple regression analysis, including all subjects on both haloperidol doses, yielded a significant main effect for pretreatment plasma HVA (n=31, F=5.7, P=0.025), indicating that higher pretreatment plasma HVA was predictive of better clinical response. In addition, the interaction between haloperidol dose and pretreatment plasma HVA was also significantly associated with clinical response (F=12.59, P=0.0015). When the two haloperidol doses were analyzed separately, we found that pretreatment plasma HVA was only correlated with clinical response in the low haloperidol 5 mg/day group (n=18, F=11.73, P=0.0038) and was unrelated to clinical response to the high haloperidol 25 mg/day group. LIMITATIONS: The sample size was small. Results may have been confounded by prior antipsychotic treatment and concomitant use of lithium or lorazepam. DISCUSSION: These results suggest that pretreatment plasma HVA could be useful for dosing antipsychotics. Patients with high plasma HVA levels would be good candidates for low-dose treatment because they are more likely to improve on such a dose, while patients with low plasma HVA levels might warrant more rapid dosage escalation.