Percutaneous Endoscopic Gastrostomy in Cirrhosis Patients: A Comparative Cohort Study

Purpose: Cirrhosis is considered a relative contraindication for percutaneous endoscopic gastrostomy (PEG) with limited published data to support this position. The purpose of this study is to analyze outcomes of hospitalized cirrhotic patients who underwent PEG compared to cirrhosis patients without PEG and non-cirrhotic patients undergoing PEG. Methods: Patients were found using ICD codes from the UVAHS clinical database (1993–2005) to identify cirrhotic patients undergoing PEG placement and comparative control cohorts. Systematic chart review validated the diagnoses and other clinical characteristics. The study population outcomes were analyzed compared to a hospitalized cohort with cirrhosis and a cohort undergoing PEG placement without cirrhosis. Results: There were 26 patients with documented cirrhosis who underwent PEG placement. The most common etiology for cirrhosis was alcohol. The mean MELD score was 15.3 (95% CI, 12.4–18.2), INR was 1.4 (1.3–1.5), and platelet count was 161x1000/mcL (115–207). Seven (26%) patients had esophageal varices and 16 (61%) had a history of ascites. Ten patients (39%) died within 30 days of the procedure; 2 deaths were related to the procedure and 3 were possibly related to the procedure. No patient who survived more than 30 days after PEG died within the next 60 days. Nine of the 10 patients that died within 30 days of the procedure had a history of ascites. Ascites was an independent risk factor for death (RR for death = 2.1, 95% CI 1.1–3.7, p= 0.04). When compared to a cohort of 554 inpatients with cirrhosis that did not receive a PEG tube and to a separate cohort of 341 inpatients without cirrhosis that did undergo PEG placement, mortality for the cirrhosis PEG patients at 30 days was not statistically different from the general cirrhosis cohort (38.5% vs 23.1%, p= 0.10); however, 30 day mortality for the cirrhosis PEG patients was higher than the general PEG cohort (38.5% vs. 18.1%, p= 0.02). At 90 days after PEG, there were no statistical differences between any of the three cohorts in overall mortality (p>0.20). Conclusions: Our findings suggest hospitalized cirrhotic patients undergoing PEG placement have 30 day mortality rates similar to a cohort of hospitalized cirrhotic patients without PEG placement. While PEG placement in cirrhosis should be considered carefully, its' role in management of malnutrition in cirrhosis should be reconsidered. Prospective trials in the cirrhotic population are needed.