Inverse association of cancer with diagnoses of dementia in a large health claims case-control study

Abstract

Background

Previous studies reported an inverse association of cancer with dementia risk.

Objective

To study associations of cancer diagnoses with risk of a dementia diagnosis in a large case-control study.

Methods

We included patients with a dementia diagnosis and controls matched 1:2 for age, sex, region, and earliest year of outpatient treatment in a case-control study from health insurance claims of outpatient consultations. Exposure included seventeen cancer categories. We compared the annual prevalence trajectories of cancer categories between dementia cases and controls over a period of 10 years before the index date. We report unadjusted odds ratios (ORs) and ORs adjusted for established dementia risk factors.

Results

We included 1,686,759 patients with incident dementia and 3,373,518 matched controls. We identified four types of associations: Cancers with equal or lower prevalence in controls at year −10 and higher prevalence in controls after year 6 to 4 before the index date; with higher prevalence in controls at all time points; with a lower prevalence in controls up to three years before the index date and higher prevalence after that; and with equal prevalence in cases and controls until year −9 and a steep increase in controls after that. When using adjusted ORs the overall pattern of an inverse association of cancer with dementia was maintained.

Conclusions

Our data suggest that a mixture of causes contributes to the inverse association of cancer and dementia diagnoses, including selective survival, underdiagnosis of dementia after cancer and of cancer in dementia, but also shared biological factors.