Data on body composition in patients with insulin-dependent diabetes mellitus (IDDM) are scarce. Dual X-ray absorptiometry (DXA) scanning has proved useful for this purpose in other groups of patients. We tested the validity of the DXA scanner for the determination of fat-free mass (FFM) and fat mass in IDDM patients and control subjects, as compared to other reference methods, i.e. total body potassium by K-40 whole body counting (TBK), total body water by tritiated water (TEW), bioelectrical impedance analysis (BIA) and 24-h urinary creatinine excretion (Ucrea). A total of 13 healthy controls, 5 males and 8 females, aged 34.2 years +/- SD 10.4, and 11 IDDM patients, 5 males, 6 females, aged 28.1 years +/- 7.3, diabetes duration 4.2 +/- 2.9 (1.0-9.9), were examined. The patients had no long-term diabetic complications and they had normal ophthalmoscopy and urine albumin excretion. The agreement between FFM estimated by DXA and the other methods, expressed as mean difference +/- 2 SD was; for DXA vs. TBK, 0.09 +/- 6.26 and 0.50 +/- 5.26 kg for controls and IDDM patients respectively; DXA vs. TEW, -2.07 +/- 2.56 and -1.07 +/- 4.58 kg; DXA vs. Ucrea, -2.62 +/- 8.02 and 2.00 +/- 10.0 kg; DXA vs. BIA, -7.90 +/- 8.92 and -7.85 +/- 2.32 kg. The results obtained with BIA were significantly different from the other methods for both control subjects and IDDM patients. In conclusion, DXA scanning is a precise and valid method for estimation of fat-free mass in IDDM patients.