Using H2[18O] tracer isotope dilution and corrected bromide space as standard reference techniques, we determined total body water and extracellular water in cirrhotic patients with (four men and four women) and without (seven men and six women) ascites and compared them with a normal control group (eight men and six women). These results were then compared with calculations of total body and extracellular water determined by the bioelectrical impedance analysis technique.

According to H2[18O] dilution, total body water was similar in cirrhotic patients without ascites and in controls (60.8% ± 2.1% vs. 60.3% ± 1.4% body wt), but was increased in patients with ascites (69.7% ± 1.2% body wt; p < 0.002). Correlation was excellent between the H2[18O] dilution and bioelectrical impedance measurements of total body water in controls and cirrhotic patients without ascites (r = 0.98; p < 0.0001). However, this correlation was poor in cirrhotic patients with ascites (r = 0.17; not significant). According to the bromide space, extracellular water (expressed as a percentage of total body water) was increased in cirrhotic patients with (57.8% ± 1.8%; p < 0.001) and without (44.0% ± 1.2%; p < 0.001) ascites compared with controls (36.6% ± 1.0%). A poor correlation (r = 0.41; p < 0.13) was seen for extracellular water measurements between the bromide space method and the bioelectrical impedance method, which failed to detect the differences among the three groups observed with the bromide space technique. Furthermore, bioelectrical impedance failed to detect any change in total body or extracellular water after paracentesis, with a degree of inaccuracy that increased linearly as the amount of ascitic fluid removed increased (r = 0.97; p < 0.001). All these intergroup comparisons remained the same, whether the analysis was of both men and women combined or for each gender individually. However, we saw differences between men and women in the control group and cirrhotic group without ascites.

These results demonstrate that abnormalities in water homeostasis and compartmentalization between intracellular (the difference between total body and extracellular water fluid) and extracellular water may exist in cirrhosis whether or not fluid accumulation is clinically evident. These data further indicate that alterations in the metabolically active body cell mass (as represented by intracellular water) in cirrhosis may occur independently of total body water and calculated fat-free body mass. In addition, gender is an important variable to control for in studies of this type. Finally, bioelectrical impedance alone is inaccurate for determining total body water and extracellular water in cirrhosis. The combined measurements of total body water by bioelectrical impedance and extracellular water by bromide dilution appear adequate for estimating fluid compartments in cirrhotic patients without ascites. However, if ascites is present, dilution techniques for both total body and extracellular water are necessary. (HEPATOLOGY 1991;14:1102–1111.)