BACKGROUND: PEG feeding is not recommended for short-term use because the 30-day mortality after PEG placement is substantial. The primary aim of this study was to prospectively identify factors predictive of survival in patients referred for PEG placement. METHODS: All patients for whom gastroenterology consultation was sought for feeding PEG placement were prospectively studied. Demographic data, Charlson comorbidity index, and functional status were recorded at entry. After PEG placement, patients were followed for up to 12 months. RESULTS: Of the 67 patients for whom consultation was requested, 58 were eligible for the study and 50 underwent PEG placement. The 7-day and 30-day mortality rates in the PEG placement group were 4% and 20%, respectively. In multivariate analysis, only the Charlson index > or =4 was associated with decreased survival time (relative hazard = 2.9: 95% CI [1.20, 7.21], p = 0.019). Median survival in patients with Charlson comorbidity index > or =4 was significantly shorter than that in patients with Charlson index < 4 (p = 0.013). CONCLUSIONS: A Charlson comorbidity index > or =4 was significantly associated with shorter patient survival after initial consultation. Careful consideration of predictive factors of survival may improve patient selection for feeding PEG placement.
BACKGROUND:PEG feeding is not recommended for short-term use because the 30-day mortality after PEG placement is substantial. The primary aim of this study was to prospectively identify factors predictive of survival in patients referred for PEG placement. METHODS: All patients for whom gastroenterology consultation was sought for feeding PEG placement were prospectively studied. Demographic data, Charlson comorbidity index, and functional status were recorded at entry. After PEG placement, patients were followed for up to 12 months. RESULTS: Of the 67 patients for whom consultation was requested, 58 were eligible for the study and 50 underwent PEG placement. The 7-day and 30-day mortality rates in the PEG placement group were 4% and 20%, respectively. In multivariate analysis, only the Charlson index > or =4 was associated with decreased survival time (relative hazard = 2.9: 95% CI [1.20, 7.21], p = 0.019). Median survival in patients with Charlson comorbidity index > or =4 was significantly shorter than that in patients with Charlson index < 4 (p = 0.013). CONCLUSIONS: A Charlson comorbidity index > or =4 was significantly associated with shorter patient survival after initial consultation. Careful consideration of predictive factors of survival may improve patient selection for feeding PEG placement.