Introduction High mortality rates have previously been reported following gastrostomy insertion, particularly among certain patient groups (e.g., dementia). With the increasing use of prophylactic gastrostomy for head and neck cancer (HNC), our group aimed to examine survival in this cohort compared to other referral indications and assess risk factors.

Methods Gastrostomy insertions were examined from two hospitals in Sheffield between 2004–2013. Data was prospectively collected from all referred patients including demographic data, biochemical parameters, referral indications and gastrostomy type. Statistical analysis was performed with Chi-squared or Fishers exact tests.

Results 1733 patients were included (1004 male, mean age=65). 30 day and 1 year mortality was 9.66% and 44.98% respectively. Indications for gastrostomy included; HNC (n = 591), neurological disease (n = 429), dysphagic stroke (n = 393), dementia (n = 9) and other (n = 311). The lowest mortality was seen in patients with HNC (30 day mortality = 5.2%, 1 year mortality = 32.6%). In comparison, 30 day mortality in all other groups was significantly higher (8.47% in neurological diseases, 15.86% in dysphagic stroke, 33.3% in dementia and 11.25% in ‘other’ indication, p < 0.01). Mortality was also significantly higher at one year (p < 0.01). There was no significant difference in mortality when comparing radiologically inserted and percutaneous endoscopic gastrostomies. Higher mortality rates were seen in patients aged 60 years or above at 30 days (OR 2.439 (1.666 – 3.731) p < 0.0001) and also at 1 year (OR 3.140 (0.268 – 0.600) p < 0.0001). Albumin less than 30 g/L was also associated with significantly higher 30 day (OR 4.486 (3.067 – 6.561) p < 0.0001) and 1 year mortality outcomes (OR 2.319 (1.830 to 2.939) p < 0.0001). In accordance with recent published data, our findings would support an elevated CRP (>5 mg/L) being a factor associated with 30 day mortality (OR 8.930 (1.199 to 66.51) p = 0.006).

Conclusion Referral indication for gastrostomy significantly impacts 30 day and 1 year mortality outcomes, with lowest rates demonstrated in patients with HNC. Identification of factors associated with mortality as seen in this study could help improve patient selection and be of relevance in the decision making process for gastrostomy.

Disclosure of Interest None Declared.