Background: The relationship between Barrett's oesophagus and colorectal cancer and other extra-oesophageal malignancies (EOM) has been a matter of controversy. These relationships have therefore been examined in a prospective study design in the General Practice Research Database.
Methods: Cohorts of patients having Barrett's oesophagus (n=1677), oesophagitis (n=6392), and simple reflux (n=6328), and a standard reference cohort representing the general population in the UK (n=13,416) were selected. The last three cohorts were matched to the Barrett's cohort by general practice, age and sex. Incident outcomes occurring beyond the first year of the follow-up were used for analyses. Hazard ratios and 95% confidence intervals were calculated using Cox's proportional hazards regression. The associations with cataract and oesophageal cancer were explored for comparison.
Results: Incident cases of 567 EOM (including 74 colorectal cancers), 448 cataract and 43 oesophageal cancers were used in the final analysis. The relative risks for colorectal cancer compared to the standard reference cohort were 1.16 (0.42-3.21) in the Barrett's cohort, 1.39 (0.76-2.54) in the oesophagitis cohort, and 0.93 (0.45-1.90) in the simple reflux cohort. The corresponding relative risks in the Barrett's cohort were 1.29 (0.90-1.85), 1.60 (1.10-2.32), and 10.56 (5.07-21.99) for EOM, cataract and oesophageal cancer, respectively.
Conclusions: The risk of colorectal cancer was not higher in any of the Barrett's oesophagus, oesophagitis, or reflux cohorts compared to the general population. The explanations for the modest increase in the risk of EOM and cataract in the above cohorts are unclear but they may be mediated by ascertainment bias or shared risk factors.