Introduction Decision making regarding resection of colorectal cancer (CRC) is difficult in the very elderly. Risks associated with surgical intervention are high. Patients are more likely to be frail and have comorbidites which lead to death or debility in a shorter time frame than that dictated by an unresected CRC. Outcomes of operatively managed patients are well documented but those of the non-operatively managed less so. We aim to determine the outcomes of non-operative management in the very elderly with CRC to better inform clinician decision making and discussions with patients regarding treatment options.

Method All patients aged 80 or more diagnosed with colorectal cancer between 01/04/2010 and 31/03/2014 were identified from a prospectively populated database in a single NHS Trust. Details of demographics, cancer site, stage and management were extracted. Primary outcomes were one-year mortality and, in those managed without resection, reasons for non-operative management. Two-tailed fisher’s exact test was used to compare mortality rates in the non-operative group (s) compared to the operative group.

Results 1240 patients were diagnosed with CRC over this four year period, of whom 292 (154 male, 138 female) were over 80 years at the time of diagnosis (median 84.5, range 80–100). Of these, 141 (48.3%) underwent formal resection, 15 (5.1%) local resection, and 136 (46.6%) no resection. One year mortality in those who had a resection was 21.8% compared to 69.1% in those managed non-operatively (p < 0.0001). Reasons for non-operative management were inoperable tumour (2.9%), metastatic disease (39.7%), patient choice (7.3%) and frailty (37.8%). Mortality rates in these four subgroups were as follows; inoperable 100% (p = 0.003), metastatic disease 90% (p < 0.0001), patient choice 30% (p = 0.694) and frailty 46% (p = 0.001).

Conclusion It is no surprise that overall, resectional surgery improves 1 year survival. Analysis of the subgroups however, does raise some interesting questions. Resectional surgery is almost certainly inappropriate in metastatic or locally advanced disease but it is of particular note that the difference in 1 year mortality in the other three subgroups is not that marked. This does raise two questions; is patient selection for resective surgery optimal and to what extent is the CRC actually the determinate of survival in this patient cohort. Assessment of fitness for surgery in this Trust is currently relatively subjective, it is possible that a more objective assessment (such as cardiopulmonary exercise testing) would improve patient selection. Whether this would also change the proportion resected is unknown. To understand the extent to which CRC is the determinate of survival, information is needed about cause of death, further work is ongoing to establish that.

Disclosure of interest None Declared.