OBJECTIVE Patients with alcoholic hepatitis frequently have moderate or severe malnutrition. Dietary protein intake may be restricted in these patients because of concurrent hepatic encephalopathy. To further evaluate the relationship between dietary protein intake and hepatic encephalopathy in alcoholic hepatitis, we evaluated prospectively gathered data from a study of 136 placebo-treated patients with moderate or severe alcoholic hepatitis conducted at eight Department of Veterans Affairs Medical Centers. METHODS Physical examination, laboratory tests, and grade of hepatic encephalopathy were recorded at entry and every seventh day for the first 28 days of study. Average daily protein intake was calculated from dietary evaluation obtained by a registered dietitian at entry and again three times a week. RESULTS Sixty-three percent of patients had hepatic encephalopathy at entry. Hepatic encephalopathy decreased over time. Time dependent regression analysis found low protein intake, along with high blood urea nitrogen (BUN) and high serum creatinine, to be independently associated with worsening hepatic encephalopathy. Similar analysis found low BUN and less malnutrition at entry into the study to be independently associated with improved hepatic encephalopathy. Higher protein intake was associated with improved hepatic encephalopathy in univariate (p = 0.01), but not multivariate, analysis. CONCLUSIONS In patients with alcoholic hepatitis who can be treated with standard anti-encephalopathy medications (e.g., lactulose and neomycin), low protein intake is associated with worsening hepatic encephalopathy while a higher protein intake correlates with improvement in hepatic encephalopathy.