It's contrary - comorbidity does not affect survival of South Africans with colorectal cancer: An analysis from the Colorectal Cancer in South Africa cohort.

BACKGROUND

The Colorectal Cancer South Africa (CRCSA) study is an observational cohort of patients with colorectal cancer (CRC) in Johannesburg, South Africa (SA). We found that the mean age at the time of CRC diagnosis was 56.6 years, consistent with studies from SA and sub-Saharan Africa. In high-income settings, comorbidity adversely affects CRC survival, and patients are substantially older at the time of CRC diagnosis. Given the younger age at CRC diagnosis in the CRCSA cohort, we hypothesised that comorbidity may be less prevalent and have little impact on CRC survival.

OBJECTIVES

To determine the prevalence of comorbidity and whether comorbidity adversely affects overall survival (OS) of CRC patients.

METHODS

Patients enrolled in the CRCSA study between January 2016 and July 2018 were included. The cohort comprised a convenience sample of adults with histologically confirmed CRC, treated at the University of the Witwatersrand Academic Teaching Hospital Complex. Demographic, clinical and histological variables were collected at baseline and participants were followed up for OS. The Charlson comorbidity index (CCI) scoring system was used to classify participants as 'no comorbidity' (CCI score 0) and '1 or more comorbidities' (CCI score ≥1). A descriptive analysis of the cohort was undertaken, while survival across comorbidity groups was compared by the Kaplan-Meier method and Cox proportional hazards (PH) regression models. Multivariable Cox PH regression was performed to examine the effect of comorbidity on survival (unadjusted) and then adjusted for variables.

RESULTS

There were 424 participants, and the mean (standard deviation) age was 56.6 (14.1) years (range 18 - 91). Only 19.1% of participants had ≥1 comorbidities, of which diabetes mellitus was most frequent (12.3%), followed by chronic obstructive pulmonary disease (4.7%) and cardiovascular disease (3.1%). There was no significant difference in unadjusted and adjusted risk of death for the group with ≥1 comorbidities compared with those with no comorbidity. However, an incidental finding showed a significantly increased risk of death for those receiving potentially curative treatment later than 40 days after CRC diagnosis.

CONCLUSIONS

In the CRCSA cohort from Johannesburg, comorbidity is uncommon, with no significant adverse effect on OS. If potentially curative treatment is initiated within 40 days of CRC diagnosis, OS could be improved. To fully understand the epidemiology of CRC in SA, population-based registries are essential, and future research should aim to identify health system failures that lead to delays in intervention beyond 40 days in patients with CRC.