Although fluid analysis usually is the first step toward identifying t
he cause of pleural effusion in patients with cirrhosis and ascites, t
here are no available data on the reliability of this approach. theref
ore, we retrospectively evaluated hematologic and biochemical paramete
rs from pleural fluid analysis in 21 patients with hepatic hydrothorax
(with proven peritoneal-pleural communication) and 6 patients with pr
imary pleural disease (2 with tuberculosis, 3 with parapneumonic effus
ion, and 1 with empyema). The criteria developed by Light were diagnos
tic of pleural ''exudate'' in only one of six patients with primary pl
eural disease. concentrations of leukocytes, total protein (TP), album
in, and lactic dehydrogenase (LDH) in both fluids were measured and pl
eural fluid-to-ascites ratios of these measurements were calculated. O
nly ratio values for leukocytes and TP were higher in the group of pat
ients with primary pleural disease compared with those with hepatic hy
drothorax. Ratio values for leukocytes and TP overlapped between both
groups during baseline conditions and during episodes of spontaneous b
acterial peritonitis and pleuritis. We conclude that pleural fluid ana
lysis has limited diagnostic efficacy in the patient with cirrhosis. D
ata collected by other methods-clinical and radiologic-should assist i
n arriving at the correct diagnosis.