BACKGROUND: Many drugs are commonly administered according to total body weight or age basis. However, drugs are primarily distributed to the lean body mass. This study was undertaken to find out the best determinant for drug requirements during induction in elderly. METHODS: Forty-five ASA 1 to 3 male and female patients older than 65 years scheduled for elective surgery were divided into 3 groups and received thiopental sodium 62.5 mg/min (group T, n = 15), propofol 25 mg/min (group P, n = 15), or etomidate 5 mg/min (group E, n = 15) respectively. Kendall's tau test for correlations was used to describe the relationship between drug requirements for induction and total body weight, lean body mass determined by Weisburg's modification of Gubner's formula, ideal body weight calculated by Devine's method, and body surface area. RESULTS: Loss of consciousness was obtained with a thiopental sodium dose of 128.4 +/- 29.3 mg, propofol 59.8 +/- 13.6 mg, and etomidate 9.8 +/- 1.4 mg. Kendall's tau correlation test showed that requirements of drugs were related to lean body mass (group T, r = 0.490*, group P, r = 0.433*, group E, r = 0.493*, global P < 0.05) and ideal body weight (group P, r = 0.426*, group E, r = 0.434*, p < 0.05), but not to total body weight or body surface area. CONCLUSIONS: Our results indicate that drug requirements for induction of anesthesia correlate better with lean body mass and ideal body weight than with total body weight or body surface area in elderly patients.