A Double-blind, Randomized, Placebo-controlled 4-week Study on the Efficacy and Safety of the Purinergic Agents Allopurinol and Dipyridamole in Acute Bipolar Mania.

Men and women ages 18 to 65 years, who were inpatients with a diagnosis of manic episode were recruited at the emergency room of Espirita Hospital of Porto Alegre (HEPA), Brazil (n=180) between September 2004 and 2006. The presence of manic episode was confirmed using SCID-I and the severity of episode was evaluated using the Young Mania Rating Scale (YMRS) and the Clinical Global Impression scale (CGI). Patients were required to present a score>22 on the YMRS at screening to be included. Subjects presenting rapid cycling, mixed episodes and comorbidities with axis I psychiatric disorders were not considered for inclusion. All subjects presented good physical health determined by physical exam, medical history, blood screening and electrocardiogram. Patients were randomly assigned to allopurinol 600 mg/day, dipyridamole 200mg or placebo as add-on medication lithium treatment for a 4-week, double-blind trial. The use of adjunctive antipsychotic agents (typical or second-generation drugs) was not allowed during the double-blind period. Adjunctive diazepam was used when necessary (maximum dose=20mg/day). Raters (psychiatrists) were trained together to establish reliability (YMRS =0.90). This study was approved by the Espirita Hospital Ethics Committee-IRB. Written informed consent was obtained from all patients and/or family member. Possible adverse events were monitored weekly during the follow-up period. Regarding primary outcome measures, we compared the percentage of responders according YMRS score decrease of at least 50% among allopurinol, dipyridamole and placebo groups, using fisher exact test. Also, the percentage of improvement from baseline to endpoint was obtained and compared among groups using analysis of variances (one-way ANOVA) with Duncan post-hoc test. Remission rates (YMRS scores < 12) were also analyzed. P values < 0.05 were considered statistically significant. Besides completers analysis, intention to treat and LOCF were employed to include data from drop-out patients who were evaluated at visit 3. Possible correlations were measured using Pearson test. Fischer exact test and X2 evaluated response and remission rates. Data are presented as mean ± standard deviation.