Intravascular volume in cirrhosis. Reassessment using improved methodology.

Previous studies of blood volume (BV) in cirrhosis have either not adjusted BV properly for body size; determined plasma volume from the dilution of labeled albumin 10-20 min postinjection, when some extravascular redistribution has already occurred; and/or not used the correct whole body-peripheral hematocrit ratio (0.82) in calculating whole BV from plasma volume and the peripheral hematocrit. We measured BV with attention to these considerations in 19 patients with cirrhosis and reexamined the determinants of vascular volume and the relationship between vascular volume and sodium retention. BV was calculated as plasma volume (determined from extrapolated plasma activity of intravenously injected [131I]+albumin at time 0) divided by (peripheral hematocrit X 0.82). The result was expressed per kilogram "dry" body weight, determined by subtracting the mass of ascites (measured by isotope dilution; 1 liter = 1 kg) from the actual body weight of nonedematous patients. Measured and expressed in this way, BV correlated strongly with esophageal variceal size (r = 0.87, P less than 0.05), although not with net portal, right atrial, inferior vena caval, or arterial pressure, and was significantly greater in patients with sodium retention as compared to patients without sodium retention. The principal modifier of vascular volume in cirrhosis is vascular capacity, which is probably mainly determined by the extent of the portasystemic collateral circulation. Increased vascular volume in patients with sodium retention as compared to patients without sodium retention supports the "overflow" theory of ascites formation.