Clinical selection criteria can predict futile intervention in patients referred for percutaneous endoscopic gastrostomy insertion

BackgroundPercutaneous endoscopic gastrostomy (PEG) placement is performed in a patient group with high mortality in the short and medium term. For a significant proportion of patients, the procedure provides no increase in survival. There are no standardised assessment tools available to determine the clinical appropriateness of PEG placement, nor any to predict clinical outcome. AimThe study aims to determine whether clinical assessment, by a trained dietitian, of the appropriateness of PEG placement is predictive of mortality in the short and medium terms. MethodsA prospective audit was undertaken of all requests for PEG placement at a single large, publicly funded Australian tertiary hospital. The clinical appropriateness of each request was assessed by a trained dietitian, and data on age, sex, reason for referral, comorbidities and satisfaction of assessment criteria were collected, and patient outcome and survival were compared for all patients according to whether a PEG was inserted or not. Main outcome measures were mortality at 30 and 150 days after referral. ResultsDuring the period 2005-2008, 198 patients were referred for PEG; 94 were assessed as appropriate referrals, 104 as inappropriate. Eighty-four patients who underwent gastrostomy, after being assessed as appropriate, had significantly reduced mortality at 30 days (96.4% vs 74.6%, P < 0.0001) and 150 days (82.1% vs 57.9%, P = 0.0001) compared with all other patients. Patients who received PEG despite contrary advice had no significant survival advantage, at 30 days or 150 days, over patients who did not receive PEG. ConclusionThe application of selection criteria by trained assessors improves patient selection for PEG insertion and predicts mortality at early and later time points, by identifying patients unlikely to benefit from PEG. The group of patients who received a gastrostomy despite an adverse assessment had no mortality benefit - in these patients, the procedure may have been futile.