In electroconvulsive therapy (ECT), electrical dosage is determined using 'fixed-dose', 'age-based' dose, or empirical titration methods. Estimation of initial seizure threshold (IST) has been claimed to be imperative for suprathreshold dosing. This systematic review aimed to determine common levels of IST, to define cut-off values for high IST, and to summarize reported IST associated factors. Medline and PsycINFO were searched from 1966 to January 2008 and relevant references were cross-checked. Subject headings including ECT, seizure threshold, dosage, and dosing were used. All articles reporting on levels of IST and/or associated factors were included. Of 395 potentially relevant reports, 46 studies on 70 samples concerning 3,023 patients were selected. Nine samples (n = 306 patients) without available standard deviation and four samples (n = 275 patients) treated with mixed electrode placement were excluded. Meta-analysis was done on 30 unilaterally treated samples (n = 1,326 patients) and 27 bilaterally treated samples (n = 1,116 patients). In unilateral ECT, weighted mean of IST was 68.2 milliCoulombs (mC; 95% CI 63.2-73.3 mC), and in bilateral ECT 111.6 mC (95% CI 103.7-119.4 mC). Calculated cut-off values for high IST were 121 mC for unilateral ECT and 221 mC for bilateral ECT. According to the literature, male gender and use of bilateral electrode placement appeared to increase IST most prominently. In conclusion, calculated electrical doses for 'suprathreshold' right unilateral ECT and for 'moderate above threshold' bilateral ECT, using commonly reported IST levels, were in the same though narrower ranges as provided in 'fixed-dose' and 'half-age' based strategies, respectively.