The overall and temporal association of cancer with polymyositis and dermatomyositis.

OBJECTIVE

To summarize the association of cancer with adult polymyositis (PM) and dermatomyositis (DM) and to test whether myositis precedes or follows cancer diagnosis, with the latter suggesting that myositis is paraneoplastic.

METHODS

Using primary patient data, we performed an analysis of all published case control and cohort studies evaluating the association of myositis and malignancy. We evaluated the overall association of cancer within 10 years of myositis diagnosis and separately assessed cancer occurrence in periods 5 years before and 5 years after myositis diagnosis.

RESULTS

Four case control or cohort studies met inclusion criteria providing 1078 myositis cases (565 PM and 513 DM) with a comparable number of controls. One hundred fifty-three myositis patients had cancer within 10 years of diagnosis. Despite reports of no myositis association with cancer, each study's data suggested an association of both PM and DM with cancer. Overall odds ratio (OR) for the association of cancer with DM = 4.4 (9.5% CI 3.0, 6.6) and for PM, OR = 2.1 (95% CI 1.4, 3.3). Cancer risk was high both before and after DM diagnosis, but for PM, cancer risk was increased only after myositis diagnosis.

CONCLUSION

DM is associated with cancer, and our results suggest that DM is often paraneoplastic. The high rate of malignancy only after PM diagnosis is consistent with cancer detection bias, but may suggest joint susceptibility of affected individuals to both cancer and myositis.