BACKGROUND:Tricyclic antidepressants have been demonstrated in the laboratory to have anticancer properties. A recent study by our group also suggested a protective effect against development of colorectal cancer and glioma. This study aims to determine whether the anticancer action of tricyclics translates to improved survival in these cancers post-diagnosis. METHODS:A study using the General Practice Research Database examined whether tricyclic antidepressant exposure in the months following diagnosis of glioma or colorectal cancer would affect longer term all-cause mortality. Cox proportional hazards modelling adjusted for age, gender, smoking, body mass index, comorbidity, and diagnosed depression. RESULTS:A cohort of 1,364 glioma and 16,519 colorectal cancer patients were identified. There was a non-significant reduction in the hazard for glioma patients treated with tricyclics (HR = 0.83, CI = 0.53-1.30). This was mainly found in patients who were not previously exposed to tricyclics (HR = 0.56, CI = 0.26-1.18). In contrast, a significant increase in hazard was found for colorectal cancer (HR = 1.37, CI = 1.21-1.54). This was mostly in patients prescribed low-dose tricyclics (HR = 1.57, CI = 1.33-1.86). CONCLUSIONS:We have shown no significant mortality reduction in colorectal cancer or glioma patients treated with tricyclics. An apparent detrimental effect observed in colorectal cancer may be related to prescription of low-dose tricyclics in the management of pain related to disseminated cancer. We cannot rule out small effects or an effect that occurs exclusively at higher doses. Blinded clinical studies may therefore be the only method of determining efficacy in glioma patients.