Utility of esophageal gastroduodenoscopy at the time of percutaneous endoscopic gastrostomy in trauma patients

Abstract

Background: The utility of esophagogastroduodenoscopy (EGD) performed at the time of percutaneous endoscopic gastrostomy (PEG) is unclear. We examined whether EGD at time of PEG yielded clinically useful information important in patient care. We also reviewed the outcome and complication rates of EGD-PEG performed by trauma surgeons. Methods: Retrospective review of all trauma patients undergoing EGD with PEG at a level I trauma center from 1/01-6/03. Results: 210 patients underwent combined EGD with PEG by the trauma team. A total of 37% of patients had unsuspected upper gastrointestinal lesions seen on EGD. Of these, 35% had traumatic brain injury, 10% suffered multisystem injury, and 47% had spinal cord injury. These included 15 esophageal, 61 gastric, and six duodenal lesions, mucosal or hemorrhagic findings on EGD. This finding led to a change in therapy in 90% of patients; either resumption/continuation of H2-blockers or conversion to proton-pump inhibitors. One patient suffered an upper gastrointestinal bleed while on H2-blocker. It was treated endoscopically. Complication rates were low. There were no iatrogenic visceral perforations seen. Three PEGs were inadvertently removed by the patient (1.5%); one was replaced with a Foley, one replaced endoscopically, and one patient underwent gastric repair and open jejunostomy tube. One PEG leak was repaired during exploration for unrelated hemorrhage. Six patients had significant site infections (3%); four treated with local drainage and antibiotics, one requiring operative debridement and later closure, and one with antibiotics alone. Conclusion: EGD at the time of PEG may add clinically useful data in the management of trauma patients. Only one patient treated with acid suppression therapy for EGD diagnosed lesions suffered delayed gastrointestinal bleeding. Trauma surgeons can perform EGD and PEG with acceptable outcomes and complication rates.