We validated the determination of extracellular water (ECW) by the bioelectrical impedance method (BIA), using the RJL manufacturer-supplied equation and the equation of Lukaski in 34 growth hormone-deficient (GHD) patients before and after replacement with recombinant human growth hormone (rhGH), using the bromide-dilution (Br-) method as the reference method. At baseline, paired comparisons showed no significant differences between mean ECW assessed by the Br- method and by BIA using both equations in the placebo and rhGH group. At 6 months, no significant difference was found between ECW assessed by the Br- method and by BIA using the manufacturer-supplied equation, both in the placebo and rhGH group. High coefficients of reliability (0.88-0.98) and high correlations (0.79-0.98; p < 0.001) were found between ECW assessed by the Br- method and by BIA applying two different regression equations in the placebo and rhGH groups at baseline and after 6 months. Mean differences in ECW assessed by the Br- method and BIA were between 2.6 and 4.1% of the mean ECW determined by the Br- method at baseline and between 2.7 and 8.5% after 6 months. Multiple comparisons showed a significantly greater ECW assessed by the Br- method than by BIA using the equation of Lukaski (p < 0.0001). This difference was found in the placebo and rhGH replacement group. The difference changed with time (p = 0.005), indicating that the usefulness of the formula of Lukaski seems to be affected by rhGH replacement therapy. We conclude that the noninvasive BIA technique might be a useful method to predict ECW in GHD patients.(ABSTRACT TRUNCATED AT 250 WORDS)