Introduction Insertion of a percutaneous endoscopic gastrostomy (PEG) in patients with dysphagia to enable provision of artificial nutrition and hydration is a widely recognised and commonplace technique. Although there are numerous small outcome studies highlighting overall risk of complications post insertion, there are very little data which reflect disease specific outcome following PEG insertion. We retrospectively evaluated disease specific 30-day mortality in a large cohort of patients undergoing PEG insertion.

Method A 10 year retrospective evaluation of all PEGs inserted at University Hospitals Coventry and Warwickshire was performed with patients identified from the Nutrition Team database. All patients between Sept 2004 and Aug 2014 were included. All patients who died within 30-days of PEG insertion were identified. Patient notes, electronic records, and death certificates (where available) were evaluated with particular attention to: demographics, primary diagnosis, indication for PEG, and co-morbidities.

Results A total of 1518 PEGs were inserted over 10 years (59% male, 41% female, age range 17–99 yrs; mean age 67). Overall 30-day mortality was 7.6% (n = 115). Procedure related mortality was 1% (n = 19), due to: immediate aspiration post procedure (0.2%); peritonitis (0.6%); cardiac event (0.1%); and ‘PEG pulled out’ (0.3%). Indications for PEG included: CVA (38%, n = 579); head and neck (H&N) cancer (29%, n = 435); neurosurgical/ traumatic brain injury (12%, n = 186); neurodegenerative disease (10%, n = 154); learning disability (2.8%); cancer (1.8%); dementia (1.1%); renal (0.8%) post-cardiothoracic surgery (0.5%); and other (4%).

Thirty day mortality by disease group was highest in patients with Dementia (23.5%), Cancer (non-H&N) (18.5%), Learning disability (14%), and CVA (12.3%). Lowest 30-day mortality was seen in H&N cancer (2.5%), and neurosurgical/neurodegenerative disease (4.4%).

When comparing 30-day mortality in the stroke cohort in the first 5 years, compared to second 5 years, there was a reduction in mortality of almost 50% (15.3 vs 7.9%). This is despite similar age range and number of procedures. Cause of death in CVA patients in the first 5 year period was aspiration pneumonia in 73% vs. 58% in the second period.

Conclusion Our audit highlights a significant variation in disease related 30-day mortality following PEG insertion in the largest published cohort (to our knowledge). This supports the need to consider disease specific factors when assessing and consenting patients for PEG insertion with respect to expected risks and outcome. Our findings emphasise the need for a robust, prospective, multi-centre audit of PEGs to clarify disease related outcomes in the UK.

Disclosure of interest None Declared.