1Division of Gastroenterology, Hepatology and Nutrition, Virginia Commonwealth University and McGuire VA Medical Center, Richmond, Virginia.
2Department of Biostatistics, Virginia Commonwealth University and McGuire VA Medical Center, Richmond, Virginia.
3Department of Psychiatry, Virginia Commonwealth University and McGuire VA Medical Center, Richmond, Virginia.
4Division of Gastroenterology, Hepatology and Nutrition, Virginia Commonwealth University and McGuire VA Medical Center, Richmond, Virginia. Electronic address: firstname.lastname@example.org.
BACKGROUND & AIMS:
Covert hepatic encephalopathy (CHE) impairs quality of life (QOL) and can be difficult to diagnose. Patient-administered methods that do not require specialized tests or equipment might increase rates of detection. We performed a longitudinal study to determine whether demographic data and responses to a validated QOL questionnaire, the Sickness Impact Profile (SIP), can identify patients with CHE.
Patients with cirrhosis without prior overt HE were recruited from outpatient liver clinics at Virginia Commonwealth University Medical Center, from August 2008 through February 2012. We performed cognitive tests on 170 patients (mean age, 55 y; mean model for end-stage liver disease score, 9; 50% with hepatitis C-associated and 11% with alcohol-associated cirrhosis). Patients were also given the SIP questionnaire (136 questions on 12 QOL topics, requiring a yes or no answer) at enrollment, 6 months, and 12 months. The proportion of patients that responded "yes" to each question was compared between those with and without CHE. Patient variables (non-cognitive), demographics (age, education, sex, alcoholic etiology), and SIP questions that produced different responses between groups were analyzed by logistic regression and receiver operating characteristic analyses.
Based on cognitive test results, 93 patients (55%) had CHE when the study began. They had a higher proportion of "yes" responses to 54 questions on the SIP questionnaire, across all categories. We developed a formula to identify patients with CHE based on age, sex, and responses to 4 SIP questions (a SIP CHE score). Baseline SIP CHE scores >0 identified patients with CHE with 80% sensitivity and 79% specificity. Of the 98 patients that returned for the 6-month evaluation, 50% had CHE (the SIP CHE identified these with 88% sensitivity). Of the 50 patients that returned for the 12-month evaluation, 32% had CHE (the SIP CHE score identified these with 81% sensitivity).
We developed a system to identify patients with CHE based on age, sex, and responses to 4 SIP questions; this formula identified patients with CHE with >80% sensitivity over a 12-month period after the initial enrollment. Patient-administered CHE screening strategies that do not include specialized tests could increase detection of CHE and improve therapy.