Understanding variations in survival for colorectal cancer in Europe: a EUROCARE high resolution study

Background-Marked differences in population based survival across Europe were found for colorectal cancers diagnosed in 1985-1989. Aims-To understand the reasons for these differences in survival in a new analysis of colorectal cancers diagnosed between 1988 and 1991. Subjects-A total of 2720 patients with adenocarcinoma of the large bowel from 11 European cancer registries (CRs). Methods-We obtained information on stage at diagnosis, diagnostic determinants, and surgical treatment (not routinely collected by CRs) and analysed the data in relation to three year observed survival, calculating relative risks (RRs) of death and adjusting for age, sex, site, stage, and determinants of stage. Results-Three year observed survival rates ranged from 25% (Cracow) to 59% (Modena), and were low in the Thames area (UK) (38%). Survival rates between registries for "resected" patients varied less than those for all patients. When age, sex, and site were considered, RRs ranged from 0.7 (95% confidence intervals (CI) 0.6-0.9) (Modena) to 2.3 (95% CI 1.9-2.9) (Cracow). After further adjustment by stage, between registry RR variation was between 0.8 (95% CI 0.6-0.9) and 1.8 (95% CI 1.5-2.2). Inter-registry RR differences were slightly reduced when the determinants of stage (number of nodes examined and liver imaging) were included in the model. The reduction was marked for the UK registries. Conclusions-The wide differences across Europe in colorectal cancer survival depend to a large extent on differences in stage at diagnosis. There are wide variations in diagnostic and surgical practices. There was a twofold range in the risk of death front colorectal cancer even after adjustment for surgery and disease stage.