Spinal cord injury (SCI) is a unique condition that markedly alters body composition, raising the possibility of having undefined effects on the traditional assumptions for body composition. In order to determine appropriate methods for the analysis of body composition in this population, 12 subjects with tetraplegia were studied for absolute weight as fat and percent fat by the following methods; bioelectrical impedance (BIA), dual energy X-ray absorptiometry (DEXA), total body potassium (TBK), total body water (TBW), and four anthropometric methods: Durnin and Womersley (DUR), Jackson and Pollack (J and P), Sloan (SLN) and Steinkamp et al (STK). The eight methods were compared with the mean of all means (24.4 +/- 2.2% fat), which was assumed to be the best estimate of percent fat. Four methods: BIA, DEXA, TBW and STK were not significantly different, while TBK, DUR, J&P and SLN were significantly (P < 0.009) different from the mean of the means. Using only the non-significantly different methods, repeat computation revealed TBW to have the smallest difference from the mean (0.1 +/- 2.8%), and DEXA to have the strongest correlation with the line of identity (r = 0.96, P < 0.0001). In conclusion, BIA, DEXA, TBW, and STK are equally valuable for estimating fat in those with tetraplegia.