Electroconvulsive therapy (ECT) is an important treatment for catatonia. We aimed to study the response rate of catatonia treated with ECT and its clinical correlates in a large sample of inpatients. The ECT parameters of all patients (n = 63) admitted with catatonia between the months of January and December 2007 were examined. The number of ECTs administered, seizure threshold, failure to achieve adequate seizures and clinical signs pertaining to catatonia were analyzed. Response was considered as complete resolution of catatonic symptoms with Bush Francis Catatonia Rating Scale (BFCRS) score becoming zero. ECT was mostly started after failed lorazepam treatment except in 6 patients where ECT was the first choice. Patients who responded in 4 ECT sessions were considered fast responders (mean session number for response is 4 sessions) and response with 5 or more ECTs was considered slow response. Fast responders had significantly lower duration of catatonia (19.67 ± 21.66 days, P = 0.02) and higher BFCRS score at presentation (17.25 ± 6.21, P = 0.03). Presence of waxy flexibility and gegenhalten (22.60% vs. 0%, P = 0.01) predicted faster response, whereas presence of echophenomena (3.2% vs. 24.0%) predicted slow response. The response rate to catatonia appears to be associated with the severity and duration of catatonia, and the presence of certain catatonic signs.