Demographic, clinical, and laboratory features that predict underlying malignancy in patients with dermatomyositis (DM) are poorly known. We conducted a retrospective study in all adult patients with a definite (n = 75) or probable (n = 32) diagnosis of DM according to Bohan and Peter criteria or with amyopathic DM (n = 14) who were referred to 2 departments during a 13-year period. The diagnosis of malignancy-associated DM was retained if DM occurred in a context of recently diagnosed malignancy or if a malignancy was diagnosed during the 5 years following the diagnosis of DM. The Kaplan-Meier method was used to assess the cumulative incidence rates of underlying malignancy during the first 5 years of DM. Factors associated with malignancy in patients with DM were identified by Cox proportional hazards models. During the study period, 121 patients fulfilled the inclusion criteria (median age, 52 yr; range, 19-77 yr; women: 70%). For 29 of them, the diagnosis of malignancy-associated DM was retained. The cumulative incidence rate of malignancy was 21 +/- 4% and 28 +/- 5%, 1 year and 5 years after the diagnosis of DM, respectively. The median duration of follow-up of the 92 patients with no malignancy diagnosed was 36 months (range, 1-140 mo). In multivariate analysis, independent factors associated with an underlying malignancy in patients with DM were an age at diagnosis >52 years (hazard ratio [HR], 7.24; 95% confidence interval [CI], 2.35-22.31), a rapid onset of skin and/or muscular symptoms (HR, 3.11; 95% CI, 1.07-9.02), the presence of skin necrosis (HR, 3.84; 95% CI, 1.00-14.85) or periungual erythema (HR, 3.93; 95% CI, 1.16-13.24), and a low baseline level of complement factor C4 (HR, 2.74; 95% CI, 1.11-6.75). Lastly, low baseline lymphocyte count (<1500/mm(3)) was a protective factor of malignancy (HR, 0.33; 95% CI, 0.14-0.80). Taken together, these data may help physicians focus on a group of patients who might benefit from extensive evaluation for malignancy.