Body composition data are important for adequate monitoring of growth and nutritional status in infants. 18O Isotope dilution techniques (ID18-O) are widely used to estimate total body water (TBW) and calculate fat-free mass (FFM). A problem of isotope dilution is an underestimation of TBW by the extrapolation to t = O approach and an overestimation of TBW by the plateau approach. Using total body electrical conductivity (TOBEC) as the reference technique we validated the extrapolation approach by 149 measurements (boys, n = 76; girls, n = 73) in 50 healthy infants aged 1-12 mo. TOBEC-derived FFM and fat mass were in excellent agreement with Fomon's reference data. Strictly linear relationships with slopes not significantly different from one were found between FFM estimated by TOBEC (FFMTOBEC) and FFM estimated by ID18-O (FFM18-O) (r = 0.98 and residual SD = 0.29 for boys, r = 0.98 and residual SD = 0.32 for girls). FFM18-O was slightly but significantly lower than FFMTOBEC, the difference being on average 0.18 (+/- 0.24) kg for girls and 0.08 (+/- 0.21) kg for boys (i.e. respectively 4 (+/- 4.5)% (p < 0.0001) and 1.5 (+/- 3.9)% (p = 0.004) of FFMTOBEC). We conclude that ID18-O using the extrapolation to t = O approach is suitable for TBW and FFM estimations in groups of infants. Due to the considerable measurement error of ID18-O (estimated at approximately 6%), individual TBW18-O and FFM18-O estimates should be considered with some caution.