SURGERY FOR NON-MALIGNANT POLYPS IS ASSOCIATED WITH SIGNIFICANT AVOIDABLE MORBIDITY AND MORTALITY. FINDINGS FROM A LARGE WESTERN COHORT

Introduction
The majority of non-malignant colorectal polyps (NMCPs) can be safely and efficiently endoscopically resected. Despite this, surgical resection is still frequently performed resulting in avoidable morbidity, mortality and health system costs. We aimed to understand the prevalence, time trends and outcomes of surgical resection of NMCPs in order to improve the management of these lesions in Australia.

Methods
A linked dataset provided by the Centre for Health Record Linkage (CHeReL), including all admissions for surgery in patients aged 18 and over in New South Wales (NSW) between July 2007 and December 2017 was analysed. Australian Classification of Health Interventions (ACHI) codes were used to filter the data set for colorectal surgery and International Classification of Diseases (ICD) codes were used to identify cases with a primary diagnosis of NMCP or colorectal cancer (CRC). Emergency surgery, inflammatory bowel disease and total colectomy were excluded. Trends over two time periods (2008-2012 and 2013-2017) were compared. Major adverse events and mortality within 30 days were analysed. Logistic regression was used to estimate risk ratios and 95% confidence intervals.

Results
36,257 surgeries for NMCP and CRC were performed in NSW in the study period. 4181 (11.4%) were performed for NMCP. Excluding the rectum, NMCPs accounted for 16.4% of surgery for neoplasia. NMCP surgery rates were higher in private (13.2%) compared to public hospitals (10.1%) (p=<0.001). Between 2008-2012 and 2013-2017 there was no difference in volumes of surgery for NMCPs (2008-2012,1968 cases (11.5%); 2013-2017, 1948 cases (11.4%); p=0.774). (Graph 1.) Major adverse events were encountered in 15.8% of NMCP surgery (661/4181). 30-day mortality was 4.7% (195/4181). Predictors of major adverse events included age, male sex, presence of a comorbidity, rectal surgery, American society of anaesthesiologists physical classification system (ASA) IV and laparotomy (Table 1). Full cover health insurance was associated with fewer adverse events (OR 0.58, p=0.03, 95% CI 0.46-1.32). Mean length of stay in the total cohort was 6.9 days, this increased to 13.9 days in cases where a major adverse event occurred.

Conclusion
Surgery for NMCPs occurs frequently in Australian hospitals and is associated with significant morbidity and a 30 day mortality rate of 4.7%. Surgery rates have not declined over time despite the proven scientific efficacy and safety of endoscopic resection from multiple independent large volume studies with no mortality. Endoscopic resection pathways are required within public and private hospitals to improve outcomes for patients with NMCPs