OBJECTIVES: To assess resource utilization from a cohort of patients with BD and perform a cost estimation of the first manic episode (FME) at the Instituto Nacional De Psiquiatría Ramón de la Fuente (INP). METHODS: Data from BP diagnosed patients between 2003-06 who received medical assistance at INP were prospectively collected according to the following inclusion criteria: 1) received medical care at emergency room or through premed service for the first time, 2) BD-diagnosed according to Structure Clinical Interview SCD-I, and 3) already had a FME. Young Mania Rating Scale (YMRS) and Psychotic Symptoms Scale were applied. Only direct medical costs were considered (from the hospital perspective) using a bottom up approach (hospitalization, drug treatment, laboratory tests and psychiatrist visit). Clinical data and resource utilization were obtained from individual medical records. Unitary costs were obtained from the Ministry of Health and adjusted to December 2006. Results are expressed in Mexican pesos (MXN). A 1000-iteration bootstrap was performed to obtain standard deviations (SD). Mean cost determinants were obtained using a generalized linear model (GLM). RESULTS: 52 BD patients were included, 38.46% males, mean age 27.44 (+10.16). Average BD initiation age was 24.44 (+8.90), with 12.63 school years (+2.75). 53.85% had a previous mania episode. 21.15% tried to commit suicide. Initial and final YMRS scores were 29.05 (+8.76) and 4.63 (+3.47), respectively. 98% of patients required hospitalization, with an average stay of 34.66 days (±10.93). The most utilized treatment drugs were benzodiazepines (86.54%). Mean cost (MC) was $74,676.16 (±$3,469.36, $68,278.36 - $81,525.87 95% CI). Hospitalization represented 91.44% of MC. GLM's statistically-significant MC determinants were age (p<0.039), school years (p<0.038), hospitalization days (p<0.0001), comorbidity (p<0.0001) and new antipsychotic drug use (p<0.004). CONCLUSION: BD's FME treatment demands high resource quantities at INP, thus savings should be enhanced.