Causes of death in a cohort of early stage colorectal cancer patients treated at a regional centre in Australia.

Abstract

Introduction: Australia has amongst the highest incidence of colorectal cancer in the world. The Australian state of Victoria has amongst the best survival rates nationally and internationally. However, inequalities in CRC survival for cancer patients living in rural and regional areas persist. In particular, there is an absolute difference of 10% in 5-year survival rates of patients diagnosed between 2005 – 2009 in Metropolitan Melbourne compared to the Grampians region in regional Victoria. These differences are also apparent in those with early stage colorectal cancer. We conducted an analysis in a regional cohort of early stage patients to explore potential contributing factors for increased mortality. Methods: A cohort of patients with early stage colorectal cancer diagnosed from 2005 - 2009 at Ballarat Health Services, Victoria, who did not receive adjuvant chemotherapy, were identified via the Victorian Cancer Registry. Demographic, tumour and treatment data were extracted from medical records. Survival data was obtained from the Victorian Cancer Registry, with correlation with hospital records to identify cause of death. Results: A total of 123 patients were included. The median age was 73 years. 21% of patients (n = 26) were diagnosed after presenting as an emergency. Most tumours were T3 (47%) and 93% were node negative. Patients lived a median distance of 13km from the hospital, with 30% living alone. 5.7% were employed at diagnosis. The median Charlson score at diagnosis was 2, and the median ECOG performance status of 1. 69% of patients were either overweight (BMI 25.0–29.9) or obese (BMI >30). As of December 2016, 63 of the 123 patients had died. The median time from surgery to death was 56 months. Colorectal cancer was identified as the cause of death in 12 of 63 deaths. The majority of deaths (51 out of 63) were not related to cancer, with heart failure (n = 8), chronic obstructive pulmonary disease (n = 6), and stroke (n = 6) as the most common causes. 11 of the 63 deaths occurred within 12 months of surgery. The median Charlson score for these patients was 2, 45% were overweight or obese, and the median ECOG of 2 was poorer than the whole cohort. 7 of the 11 deaths were in patients who had presented as emergencies, but only 2 of these were directly attributable to cancer (perforation and sepsis). The other causes of death were hospital-acquired pneumonia (n = 2), end stage airways disease (n = 2), ischaemic heart disease (n = 2). Conclusion: The majority of deaths in this cohort of early stage colorectal cancer were non-cancer related. The Charlson index and the median ECOG performance status for the whole cohort were not particularly high. Patients who presented as an emergency accounted for a majority of early deaths, however this was not directly related to malignancy in the majority. Further investigation of causes of death in the first-year post resection may lead to strategies to reduce or prevent such deaths. To further elucidate factors contributing to non-cancer mortality, data from a contemporaneous cohort of early stage patients from a Victorian metropolitan centre will be analysed and compared.