Predicting factors of malignancy in dermatomyositis and polymyositis: a case-control study

• Published in: British journal of dermatology (1951) Vol. 144; no. 4; pp. 825 - 831
• Main Authors: Chen, Y-J., Wu, C-Y., Shen, J-L.
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Science Ltd 01.04.2001; Blackwell; Oxford University Press
• Subjects: Adult; Age of Onset; Analysis of Variance; Autoimmune Diseases - complications; Biological and medical sciences; Case-Control Studies; Child; Connective Tissue Diseases - complications; dermatomyositis; Dermatomyositis - etiology; Female; Humans; Male; malignancy; Medical sciences; Middle Aged; Multivariate Analysis; Nasopharyngeal Neoplasms - complications; Neoplasms - complications; Paraneoplastic Syndromes - etiology; polymyositis; predictive factor; Prognosis; Retrospective Studies; Risk Factors; Sarcoidosis. Granulomatous diseases of unproved etiology. Connective tissue diseases. Elastic tissue diseases. Vasculitis; Sex Factors; Human; Immunopathology; Connective tissue disease; Skin disease; Dermatomyositis; Autoimmune disease; Case control study; Malignant tumor; Statistical study; Striated muscle disease; Polymyositis; Association; Systemic disease; Predictive factor
• ISSN: 0007-0963; 1365-2133
• DOI: 10.1046/j.1365-2133.2001.04140.x

Background An association between dermatomyositis (DM)/polymyositis (PM) and malignancies has been widely reported in the literature. The validity of extensive evaluation for malignancies in those patients has also been questioned for decades. Only limited papers regarding the signs of malignancy and the prognostic factors in DM/PM have been reported. Objectives To define the potential risk factors of concomitant neoplastic diseases in patients diagnosed as having DM/PM. Methods From 1 April 1983 to 30 June 1999, 147 patients were diagnosed as having probable or definite DM/PM at the Veterans General Hospital, Taichung, Taiwan. We excluded four patients who had preceding neoplastic diseases diagnosed before DM/PM, then retrospectively reviewed the data of the remaining 143 patients and subgrouped the cases as four main types: primary idiopathic DM, primary idiopathic PM, juvenile DM/PM and amyopathic DM (ADM). We next performed univariate analysis using logistic regression to evaluate the possible predictive factors for malignancies, such as mean age at onset, gender, manifestations at onset, association with other connective tissue diseases, initial skin presentations, complications and laboratory data. Then we chose the significant factors for multivariate analysis by logistic regression, to determine the independent risk factors of malignancies in DM/PM patients. Results Among the 143 patients, DM was the most common type (64%), followed by ADM (14%), juvenile DM/PM (13%) and PM (10%). The mean age at onset overall was 42·4 years. Other connective tissue diseases were present in 22% of all patients, especially PM (50%) and juvenile DM/PM patients (28%). Internal malignancies were present in 13% of patients, and most were associated with DM. Nasopharyngeal carcinomas (NPCs) were the most common tumours. Patients with primary idiopathic DM, with an older age at onset, higher serum creatine phosphokinase levels and male gender, had more chance of developing concomitant malignancies. Those associated with complications, especially interstitial lung diseases, had a lower risk of associated neoplasia. In multivariate analysis, an older age at onset (odds ratio 9·10) and male gender (odds ratio 4·06) were associated with greater risk of developing malignancies. Conclusions The two independent predictive factors for malignancy (P < 0·05) in patients with DM/PM were an older age at onset (> 45 years) and male gender. The primary idiopathic DM group was shown to have higher risk of developing internal malignancies, especially NPC. However, this was not identified as an independent predictive factor for concomitant neoplastic diseases in multivariate analysis. In addition, patients who had the complication of interstitial lung disease had a significantly lower frequency of malignancies (P < 0·001).