Risk Factors for Hepatic Hydrothorax in Cirrhosis Patients with Ascites - A Clinical Cohort Study

Background The risk factors for hepatic hydrothorax are unknown. Methods We used data from three randomized trials of satavaptan treatment in patients with cirrhosis and ascites followed for up to 1 year. We excluded patients with previous hepatic hydrothorax or other causes for pleural effusion. The candidate risk factors were age, sex, heart rate, mean arterial pressure, diuretic-resistant ascites, a recurrent need for paracentesis, diabetes, hepatic encephalopathy, International Normalized Ratio, creatinine, bilirubin, albumin, sodium, platelet count, use of non-selective beta-blockers (NSBBs), spironolactone, furosemide, proton pump inhibitors, and insulin. We identified risk factors using a Fine and Gray regression model and backward selection. We reported subdistribution hazard ratios (sHR) for hepatic hydrothorax. Death without hepatic hydrothorax was a competing risk. Results Our study included 942 patients, of whom 41 developed hepatic hydrothorax and 65 died without having developed it. A recurrent need for paracentesis (sHR: 2.55, 95% CI: 1.28-5.08), bilirubin (sHR: 1.18 per 10 mu mol/l increase, 95% CI: 1.09-1.28), diabetes (sHR: 2.49, 95% CI: 1.30-4.77) and non-use of non-selective beta-blockers (sHR: 2.27, 95% CI: 1.13-4.53) were risk factors for hepatic hydrothorax. Development of hepatic hydrothorax was associated with a high mortality-hazard ratio of 4.35 (95% CI: 2.76-6.97). Conclusions In patients with cirrhosis and ascites, risk factors for hepatic hydrothorax were a recurrent need for paracentesis, a high bilirubin, diabetes and non-use of NSBBs. Among these patients with cirrhosis and ascites, development of hepatic hydrothorax increased mortality fourfold.