Dermatomyositis (DM) is a rare systemic disorder with inflammatory myopathy and typical cutaneous involvement. Large studies report a 3- to 6-fold increase in the risk of malignancy in patients with DM.6,20,35 Schematically, ovarian or breast cancer in females and lung cancer in males are the main malignancies associated with DM.20 Because patients with DM associated with malignancy have a poorer prognosis,3,12,31 an extensive evaluation for cancer in patients with DM has been recommended.34 However, the benefit of extensive evaluation for malignancies in patients with DM has been questioned due to the financial cost and the iatrogenicity of such an extensive screening.7,8 The type of screening workup and the optimal interval for rescheduling are still debated. In this context, identifying a subgroup of DM patients with a markedly higher risk of malignancy would be useful. Only a few studies have focused on demographic, clinical, and laboratory markers associated with underlying malignancy in patients with inflammatory myopathy,3,9,17,23,28,31,34 and studies focusing on DM alone are uncommon.3 We conducted the current study to identify factors associated with an underlying malignancy in a large retrospective cohort of patients diagnosed with DM.
PATIENTS AND METHODS
We conducted a retrospective study in all adult patients with DM referred to the Department of Dermatology (Saint-Louis Hospital) and the Department of Internal Medicine (Saint-Antoine Hospital) of 2 French tertiary centers between January 1, 1995, and December 31, 2007. All medical records of patients coded during the study period as M33 (Dermatopolymyositis), M33.1 (Other dermatomyositis), M33.2 (Polymyositis), or M33.9 (Dermatopolymyositis, unspecified) following the International Classification of Diseases (ICD-10)42 were collected. Patients with a definite or probable diagnosis of DM according to Bohan and Peter criteria4,5 or patients with amyopathic DM were included in this retrospective inception cohort. Patients with polymyositis were excluded, as were patients who were referred for DM relapse during the period study or patients with DM associated with other connective tissue disorders. The diagnosis of malignancy-associated DM was retained if DM occurred in a context of recently diagnosed (<1 yr) malignancy or in a context of recently diagnosed recurrence of a known malignancy, or if a malignancy was diagnosed during the 5 years following the diagnosis of DM.
All data reported in this retrospective inception cohort study were based on hospital records. Baseline demographic data (age, sex), clinical data (DM associated or not with malignancy, duration of symptoms before the diagnosis of DM, cutaneous involvement, associated interstitial pneumonia), and laboratory data (leukocytes, lymphocytes, aspartate aminotransferase [AST], alanine aminotransferase [ALT], creatine phosphokinase [CPK], lactate dehydrogenase [LDH], antinuclear antibody, complement factors C3 and C4) were recorded.
Proportions and standard error were used as descriptive statistics for categorical variables. Continuous variables were described by median [range]. Kaplan-Meier survival analysis was used to assess the cumulative incidence rates of underlying malignancy during the first 5 years of DM. Malignancies diagnosed before DM were taken into account as if they were diagnosed concurrently with DM. Multivariate analysis was performed using the Cox proportional hazard model. The following variables were assessed as potential factors associated with underlying malignancy: sex, age, duration of symptoms at time of DM diagnosis, presence of Gottron papules, presence of heliotrope erythema, presence of periorbital erythema, presence of periungual erythema, presence of megacapillaries, presence of skin necrosis, leukocyte count, lymphocyte count, AST, ALT, CPK, LDH, antinuclear antibody, complement factor C3 and C4 levels. The proportion hazards assumption was verified both by graphical methods and by using time-dependent variables. We also verified the goodness of fit of our estimated statistical model using the Akaike information criterion.
Variables with a p value of less than 0.2 in univariate analysis were selected in the model for multivariate analysis. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. All statistical analyses were tested at the significance threshold α = 0.05 and were 2-tailed. Analyses were performed using SAS software (version 8.2, SAS Institute, Cary, NC).
During the study period, 134 DM patients were seen in our departments. Ten of them had a DM recurrence and were then excluded. Three other patients were excluded because they suffered concomitantly from other connective tissue disease. A total of 121 patients fulfilled the inclusion criteria and are reported here. Fourteen of them had amyopathic DM. Among the 107 patients with clinical muscle weakness, the diagnosis of DM was definite in 75 and probable in 32 according to Bohan and Peter criteria4,5 (Figure 1). For 33 of these patients, a muscular biopsy was not performed because the diagnosis of DM was unequivocal.
The baseline characteristics of the 121 patients are reported in Table 1. At the time of DM diagnosis, the median age of these patients was 52 years (range, 19-77 yr); 85 of the 121 (70%) were women. The median time elapsed between first skin and/or muscular complaints and the diagnosis of DM was 4 months (range, 0-48 mo).
In 29 of these 121 patients, a diagnosis of malignancy-associated DM was retained. The underlying malignancy was ovarian cancer (n = 7), lung cancer (n = 5), breast cancer (n = 5), head and neck cancer (n = 6), non-Hodgkin lymphoma (n = 2), bladder cancer (n = 1), prostate cancer (n = 1), seminoma (n = 1), and neuroendocrine tumor (n = 1). For 18 (62%) of these 29 patients, the underlying malignancy was diagnosed concurrently (that is, at the same time or within 3 mo) with DM, whereas it was diagnosed before DM in 2 patients (7%) and during the 5 years following the diagnosis of DM in 9 (31%) patients (Figure 2). The cumulative incidence rate of malignancy was 21 ± 4% at year 1, 25 ± 4% at year 3, and 28 ± 5% at year 5.
The median duration of follow-up of the 92 patients without malignancy was 36 months (range, 1-140 mo). Only 7 of these 92 patients were followed for less than 6 months, and 3 were followed for less than 3 months.
Factors Associated With Underlying Malignancy
In univariate analysis (Table 2), male sex, an older age at presentation, acute onset of cutaneous and/or muscular symptoms, periungual erythema, necrotic skin lesions, higher CPK and antinuclear antibody levels, and lower C4 and C3 levels were associated with underlying malignancy in patients with DM. Moreover, patients with DM not associated with malignancy had lower lymphocyte count than those with underlying malignancy.
Because a patient could not be included in these multivariate analyses if data were missing or unavailable on any of these variables, the analysis was performed using 109 of the original 121 patients. Missing data concerned complement factors levels (n = 11) and/or antinuclear antibody (n = 8) and/or lymphocyte count (n = 2). Independent factors associated with an underlying malignancy in patients with DM were an age at diagnosis >52 years, a rapid onset of skin and/or muscular symptoms, the presence of skin necrosis and periungual erythema, and a low baseline C4 level (Table 3). Lastly, a low baseline lymphocyte count was more frequently observed in patients without malignancy. The Akaike information criterion of our model with covariates was 205.9 compared with 232.1 for the model without covariates.
The results of the current study show that demographic, clinical, and laboratory factors are associated with an underlying malignancy in patients with DM. In this retrospective cohort of 121 patients, older age at onset, rapid onset of skin and/or muscular symptoms, and presence of necrotic skin lesions or periungual erythema were significantly more frequently observed in patients with malignancy-associated DM than in patients with DM not associated with malignancy. Moreover, patients with malignancy-associated DM had lower baseline C4 levels and higher baseline lymphocyte counts than those without malignancy.
We present one of the largest case series of DM published to date. Although this study was based on a retrospective chart review, we believe the study has several strengths. First, because of the limited number of clinical departments involved, investigational habits were quite homogeneous, as witnessed by the low number of missing data. However, although autoantibody profiles have been shown to be associated with the risk of underlying malignancy in DM patients,10 we were not able to consider this point in the present study, because of too many missing data. Second, because clinicians in these 2 referral departments were experienced in diagnosing DM, we are quite confident that misclassification bias regarding DM diagnosis has been avoided. Not all patients had muscular investigations however, because in 1 department invasive workup was performed only if the clinical signs of DM were not unequivocal. Third, misclassification bias regarding cancer diagnosis is limited because patients keep being followed in the same departments for their DM, and few patients were lost to follow-up; in addition, most malignancies associated with DM are generally diagnosed within the first months after the diagnosis of DM. Fourth, all patients with malignancy did not undergo a specific search for DM. However, because DM exhibits specific and usually pronounced dermatologic signs in a setting where dermatologists are experienced and easy to reach, we are confident that no cancer patients with undiagnosed DM would have escaped our attention. Moreover, a few patients in this cohort were referred by oncologists. Last, the large number of patients included in this cohort allowed us to analyze jointly multiple factors.
DM is a rare disease. Its annual incidence is about 5-10 per million persons.29,38,40 DM is considered to be associated with high morbidity and mortality rates, primarily related to life-threatening muscle weakness, cardiac and lung impairment, and infectious manifestations.12,26,27 Patients with DM associated with malignancy have a poorer prognosis than patients with DM not associated with malignancy.3,12,31 In the current study, 1 year after and 5 years after the diagnosis of DM, the cumulative incidence rate of malignancy was 21% and 28%, respectively. This finding is in agreement with previous reports. Indeed, in large series, an associated malignancy was shown in about 25%-42% of patients diagnosed with DM.6,20,35 As in other studies, we found that most malignancies were diagnosed concurrently with DM. In the large population-based study by Hill et al,20 standardized incidence ratios of malignancy were 13.5 [10.4-17.6] during the first year following the diagnosis of DM and dropped to 1.4 [range, 1.0-2.0] after 5 years. The mechanisms underlying the association between DM-and polymyositis-and cancer are incompletely understood. It is well accepted that both host genes and environmental factors are involved. The primary antigenic target in DM is the endothelium of the endomysial capillaries, through the C5b-9 membranolytic attack complex. In DM associated with cancer, tumor antigens might be involved and trigger the autoimmune process either through mimicry with endothelium antigens, or via bystander stimulation.11,30
Among the factors that we found associated with malignancy in DM patients, some have been consistently identified in previous studies (that is, age, sex, skin necrosis), others inconsistently (such as CPK level). In addition, we present new findings suggesting that low C4 levels and lymphopenia were associated-positively and negatively, respectively-with cancer risk.
In most of the previously published studies,3,9,17,28 an older age at onset and/or male sex were associated with the risk of underlying malignancy. In the largest of these studies, Chen et al9 found that an age at onset >45 years and male sex were associated with a 4- to 9-fold increased risk of malignancy in 143 patients with inflammatory myositis (adjusted odds ratios [ORs], 9.10; 95% CI, 2.03-40.74 for age and 4.06; 1.06-15.57 for male sex). Our results are in accordance with these findings. Moreover, in other series,3,13,17,21,28 as in ours, the presence of cutaneous necrosis was also suggested to be associated with an increased risk of malignancy. In the larger clinical series, Feldman et al13 found clinical evidence of vasculitis manifesting as dermal or subcutaneous nodules, periungual infarcts or digital ulceration in 7 of 76 patients. Two of 7 (29%) polymyositis/DM patients with clinical vasculitis had underlying malignancy, compared to 4 of 69 (6%) patients without clinical vasculitis. In a small anatomo-pathological study, Hunger et al21 found that when there was histopathological evidence of vasculitis in the skin biopsy specimen, the prevalence of neoplasia was greater; 4 of their 5 patients with vasculitis had a cancer, compared with 3 of their 18 patients without vasculitis (p < 0.05). In this context, the low baseline complement factor C4 level and the rapid onset of symptoms that we found in patients with malignancy-associated DM may also reflect an underlying histologic vasculitis, not systematically observed in cases of DM. Although it may seem uncommon, malignancy-associated vasculitis with low complement factor C4 has already been reported.18
On the other hand, the association between CPK level and the risk of underlying malignancy is more debated. In 1986, Fudman and Schnitzer14 reported 7 patients with DM and normal CPK levels, 3 of whom had malignancy. However, in a 1988 series of 26 patients, La Montagna et al23 found that CPK level was not a good marker for underlying malignancy in patients with DM. In that study, the authors found that 5 of 17 (30%) patients with an elevated CPK level had malignancy compared with 3 of 9 (33%) patients with a normal CPK level (p > 0.05). In a more recent article (2002), Sparsa et al34 described 33 patients with DM and 7 patients with polymyositis. They found that patients with higher levels of CPK were more likely to have cancer. Patients with malignancy (n = 16) presented with mean serum CPK level of 2840 U/L (range, 96-12,100 U/L) compared with 1346 U/L (range, 35-10,000 U/L) for the 24 patients without malignancy (p = 0.01). In a larger study9 where DM cases were pooled with other inflammatory myopathies, among the 143 patients reported, 74 had DM whereas 69 had polymyositis, juvenile DM, amyopathic DM, or DM associated with connective tissue disease. In that study, malignancies were diagnosed in 18 patients (16 of whom were DM patients) which represents 13% of the overall study population and 22% of the population of DM patients. A higher CPK level was significantly associated with malignancy in univariate analysis (OR, 5.16; 95% CI, 1.14-23.42). This higher risk was nearly significant in multivariate analysis (OR, 4.80; 95% CI, 0.85-27.23; p = 0.08). In the current study, a higher CPK level was observed in patients with cancer-associated DM but the difference did not reach statistical significance in multivariate analysis.
We showed that low complement C4 levels were associated with a risk of malignancy. The mechanisms underlying this association are unclear. An acquired complement fraction deficiency has been observed in patients with hematologic malignancies.15,16,33 In Sjögren syndrome, another connective tissue disease, low C4 levels are a warning sign for a high risk of associated lymphoma.22,32,37 In our series, however, most cancers originated from solid organs, and to our knowledge there is no universal mechanism that could account for low C4 levels in cancer, whatever its type. The second original finding is that we showed lymphopenia to be a factor negatively associated with cancer in DM patients. DM is an autoimmune disease and lymphopenia is a common finding in other connective tissue diseases, such a systemic lupus erythematosus or Sjögren syndrome. Absence of lymphopenia in cancer-associated DM might reflect different pathogenic mechanisms in both types of DM. In current study, lymphocyte counts were collected before the introduction of corticosteroids or other immunosuppressive drugs and therefore could not be therapeutically induced.
Our results help identify a group of DM patients with an increased risk for malignancy. How these results transfer to clinical practice is not straightforward, however. At the initial diagnosis of DM, all experts agree that a search for cancer must be conducted. The means and extensiveness of this workup is debated. Some have favored an extensive search for malignancy based on the fact that patients with DM carry a risk for developing cancer ranging from 6 to 12 times the risk in the general population,1,25,34 but others have disagreed.7
From a clinical perspective, it may be reasonable to advise that asymptomatic patients be screened for malignancies for which early detection and treatment have been shown to improve patient outcome (for example, cervical, breast, and colorectal cancers).24 Others have suggested that it may be reasonable to perform cancer screening in the population where factors associated with malignancy (for example, epidermal necrosis) are present.3,28,41 Another reason to screen asymptomatic patients for malignancy might be if successful treatment of malignant disease leads to improvement in DM. Indeed, DM is frequently noted to improve after cancer treatment, with recurrence of muscle weakness occurring at the time of relapse.2,19,36,39,41
Along with Callen,7 we think that the evaluation should begin with a careful history and physical examination and "standard" noninvasive, laboratory, and radiologic evaluation (that is, blood count, chest radiography, test for blood in the stool). Any abnormality should be thoroughly investigated. The evaluation should include tests that would be ordered for a "healthy" person of the same age, sex, and race as the newly diagnosed patient with DM (for example, it is recommended that persons aged older than 50 years have a colonoscopy). Moreover, because of the high prevalence of lung, colon, breast, and ovary cancer and lymphoma in DM patients,20,35 we suggest that computed tomographic scanning of the chest and abdomen should be performed in recently diagnosed DM patients. In women, ultrasonography of the pelvis and mammography screening seem justified. We think the benefit of systematic colonoscopy is more debatable. Lastly, although it appears that continued surveillance is necessary for these patients, the testing that should be done, beyond age-specific cancer screening, and the schedule of screenings are not clear. The current study does not provide direct evidence to substantiate any position. There is no evidence that systematic screening would lead to better outcomes.
In conclusion, DM patients who are older at the time of onset, exhibiting necrotic skin lesions and periungual erythema, and presenting with baseline higher lymphocyte count and lower C4 level are at increased risk of underlying malignancy. For these patients, we think an extensive search for cancer is warranted. Nevertheless, the optimal screening schedule in patients without a previous history of malignancy or in patients for whom the type or location of associated cancer is not suspected on clinical and routine background screening remains to be determined. In this context, it would be instructive to test the performance of new imaging techniques, such as positron emission tomography, in identifying underlying malignancy in adult patients with DM.
1. Airio A, Pukkala E, Isomaki H. Elevated cancer incidence in patients with dermatomyositis: a population based study. J Rheumatol
2. Barnes BE, Mawr B. Dermatomyositis and malignancy. A review of the literature. Ann Intern Med
3. Basset-Seguin N, Roujeau JC, Gherardi R, Guillaume JC, Revuz J, Touraine R. Prognostic factors and predictive signs of malignancy in adult dermatomyositis. A study of 32 cases. Arch Dermatol
4. Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med
5. Bohan A, Peter JB. Polymyositis and dermatomyositis (second of two parts). N Engl J Med
6. Buchbinder R, Forbes A, Hall S, Dennett X, Giles G. Incidence of malignant disease in biopsy-proven inflammatory myopathy. A population-based cohort study. Ann Intern Med
7. Callen JP. When and how should the patient with dermatomyositis or amyopathic dermatomyositis be assessed for possible cancer? Arch Dermatol
8. Callen JP. The value of malignancy evaluation in patients with dermatomyositis. J Am Acad Dermatol
9. Chen YJ, Wu CY, Shen JL. Predicting factors of malignancy in dermatomyositis and polymyositis: a case-control study. Br J Dermatol
10. Chinoy H, Fertig N, Oddis CV, Ollier WE, Cooper RG. The diagnostic utility of myositis autoantibody testing for predicting the risk of cancer associated myositis. Ann Rheum Dis
11. Dalakas MC, Hohlfed R. Polymyositis and dermatomyositis. Lancet
12. Danko K, Ponyi A, Constantin T, Borgulya G, Szegedi G. Long-term survival of patients with idiopathic inflammatory myopathies according to clinical features: a longitudinal study of 162 cases. Medicine (Baltimore)
13. Feldman D, Hochberg MC, Zizic TM, Stevens MB. Cutaneous vasculitis in adult polymyositis/dermatomyositis. J Rheumatol
14. Fudman EJ, Schnitzer TJ. Dermatomyositis without creatine kinase elevation. A poor prognostic sign. Am J Med
15. Fust G, Miszlay Z, Czink E, Varga L, Paloczi K, Szegedi G, Hollan SR. C1 and C4 abnormalities in chronic lymphocytic leukaemia and their significance. Immunol Lett
16. Fust G, Czink E, Minh D, Miszlay Z, Varga L, Hollan SR. Depressed classical complement pathway activities in chronic lymphocytic leukaemia. Clin Exp Immunol
17. Gallais V, Crickx B, Belaich S. [Prognostic factors and predictive signs of malignancy in adult dermatomyositis]. Ann Dermatol Venereol
18. Greer JM, Longley S, Edwards NL, Elfenbein GJ, Panush RS. Vasculitis associated with malignancy. Experience with 13 patients and literature review. Medicine (Baltimore)
19. Hidano A, Kaneko K, Arai Y, Kikuchi R. Survey of the prognosis for dermatomyositis, with special reference to its association with malignancy and pulmonary fibrosis. J Dermatol
20. Hill CL, Zhang Y, Sigurgeirsson B, Pukkala E, Mellemkjaer L, Airio A, Evans SR, Felson DT. Frequency of specific cancer types in dermatomyositis and polymyositis: a population-based study. Lancet
21. Hunger RE, Durr C, Brand CU. Cutaneous leukocytoclastic vasculitis in dermatomyositis suggests malignancy. Dermatology
22. Ioannidis JP, Vassiliou VA, Moutsopoulos HM. Long-term risk of mortality and lymphoproliferative disease and predictive classification of primary Sjogren's syndrome. Arthritis Rheum
23. La Montagna G, Manzo C, Califano E, Tirri R. Absence of elevated creatine kinase in dermatomyositis does not exclude malignancy. Scand J Rheumatol
24. Lipskie TL. A summary of cancer screening guidelines. Chronic Dis Can
25. Maoz CR, Langevitz P, Livneh A, Blumstein Z, Sadeh M, Bank I, Gur H, Ehrenfeld M. High incidence of malignancies in patients with dermatomyositis and polymyositis: an 11-year analysis. Semin Arthritis Rheum
26. Marie I, Hachulla E, Cherin P, Hellot MF, Herson S, Levesque H, Hatron PY. Opportunistic infections in polymyositis and dermatomyositis. Arthritis Rheum
27. Marie I, Hachulla E, Hatron P-Y, Hellot M-F, Levesque H, Devulder B, Courtois H. Polymyositis and dermatomyositis: short term and long term outcome, and predictive factors of prognosis. J Rheumatol
28. Mautner GH, Grossman ME, Silvers DN, Rabinowitz A, Mowad CM, Johnson BL Jr. Epidermal necrosis as a predictive sign of malignancy in adult dermatomyositis. Cutis
29. Oddis CV, Conte CG, Steen VD, Medsger TA Jr. Incidence of polymyositis-dermatomyositis: a 20-year study of hospital diagnosed cases in Allegheny County, PA 1963-1982. J Rheumatol
30. Pignone A, Fiori G, Del Rosso A, Generini S, Matucci-Cerinic M. The pathogenesis of inflammatory muscle disease: on the cutting edge among the environment, the genetic background, the immune response and the dysregulation of apoptosis. Autoimmun Rev
31. Ponyi A, Constantin T, Garami M, Andras C, Tallai B, Vancsa A, Gergely L, Danko K. Cancer-associated myositis: clinical features and prognostic signs. Ann N Y Acad Sci
32. Ramos-Casals M, Solans R, Rosas J, Camps MT, Gil A, Del Pino-Montes J, Calvo-Alen J, Jimenez-Alonso J, Micó ML, Beltran J, Belenguer R, Pallares L; GEMESS Study Group. Primary Sjogren syndrome in Spain: clinical and immunologic expression in 1010 patients. Medicine (Baltimore)
33. Schlesinger M, Broman I, Lugassy G. The complement system is defective in chronic lymphatic leukemia patients and in their healthy relatives. Leukemia
34. Sparsa A, Liozon E, Herrmann F, Ly K, Lebrun V, Soria P, Loustaud-Ratti V, Bouyssou-Gauthier ML, Boulinguez S, Bedane C, Jauberteau MO, Vidal E, Bonnetblanc JM. Routine vs extensive malignancy search for adult dermatomyositis and polymyositis: a study of 40 patients. Arch Dermatol
35. Stockton D, Doherty VR, Brewster DH. Risk of cancer in patients with dermatomyositis or polymyositis, and follow-up implications: a Scottish population-based cohort study. Br J Cancer
36. Sunnenberg TD, Kitchens CS. Dermatomyositis associated with malignant melanoma: parallel occurrence, remission, and relapse of the two processes in a patient. Cancer
37. Theander E, Henriksson G, Ljungberg O, Mandl T, Manthorpe R, Jacobsson LT. Lymphoma and other malignancies in primary Sjogren's syndrome: a cohort study on cancer incidence and lymphoma predictors. Ann Rheum Dis
38. Vargas-Leguas H, Selva-O'Callaghan A, Campins-Marti M, Hermosilla Perez E, Grau-Junyent JM, Martinez Gomez X, Vaque Rafart J. [Polymyositis-dermatomyositis: incidence in Spain (1997-2004)]. Med Clin (Barc)
39. Verducci MA, Malkasian GD, Friedman SJ, Winkelmann RK. Gynecologic carcinoma associated with dermatomyositis-polymyositis. Obstet Gynecol
40. Weitoft T. Occurrence of polymyositis in the county of Gavleborg, Sweden. Scand J Rheumatol
41. Whitmore SE, Rosenshein NB, Provost TT. Ovarian cancer in patients with dermatomyositis. Medicine (Baltimore)
42. World Health Organization. International Statistical Classification of Diseases and Related Health Problems. 10th Revision; Version for 2007. Available at www.who.int/classifications/apps/icd/icd10online/
. Accessed November 2008.