Am J Clin Dermatol DOI 10.1007/s40257-015-0120-1

SYSTEMATIC REVIEW

Cancer Risk in Dermatomyositis: A Meta-Analysis of Cohort Studies Jeannette M. Olazagasti • Pedro J. Baez David A. Wetter • Floranne C. Ernste



Ó Springer International Publishing Switzerland 2015

Abstract Background An association between dermatomyositis (DM) and cancer has been reported since 1916; however, estimates of the associated risk vary widely. For cost-effectiveness reasons it might be important to elucidate the degree of overall cancer risk in DM. Objective The aim of this systematic review was to investigate the association of cancer in DM by performing a meta-analysis of cohort studies. Data Sources A systematic literature search of PubMed, Ovid MEDLINE, EMBASE, Web of Science, Scopus, and the Cochrane Collaboration was conducted without language restriction, to 1 May 2014. Study Selection Inclusion criteria included cohort studies assessing overall cancer risk in DM. Two reviewers independently performed the study selection. Inter-rater reliability for inclusion decisions was quantified using

Electronic supplementary material The online version of this article (doi:10.1007/s40257-015-0120-1) contains supplementary material, which is available to authorized users. J. M. Olazagasti (&)  P. J. Baez Mayo Graduate School, Mayo Clinic, 200 1st St SW, Rochester, MN 55905, USA e-mail: [email protected] D. A. Wetter Department of Dermatology, Mayo Clinic, Rochester, MN, USA F. C. Ernste Department of Medicine, Division of Rheumatology, Mayo Clinic, Rochester, MN, USA

Cohen’s j statistic. Disagreements were resolved by discussion. Data Extraction and Synthesis Desired variables were extracted from eligible studies independently by two investigators and disagreements were resolved by discussion. Quality of the selected studies was assessed using a modification of a recently employed system designed with reference to Meta-analysis Of Observational Studies in Epidemiology (MOOSE), Quality Assessment Tool for Systematic Reviews of Observational Studies (QATSO), and Strengthening the Reporting of Observational Studies in Epidemiology (STROBE). Summary estimates were derived using a random-effects model. Main Outcome(s) and Measure(s) Main outcome was the calculated relative risk of developing cancer after diagnosis of DM compared with the general population, estimated as the age- and sex-adjusted standardized incidence ratio (SIR). We hypothesized a priori that the relative risk would be higher in patients diagnosed with DM. Results A total of 1,272 articles were initially identified but only ten studies met the inclusion criteria. Selected studies included seven population-based and three hospitalbased DM cohorts that ranged from 49 to 1,012 patients and had mean follow-up times from 3.7 to 10.4 years. The pooled SIR for the incidence of overall cancer in DM patients was 4.79 (95 % confidence interval 3.71–5.87) with significant heterogeneity (I2 = 85.8 %). However, the heterogeneity had no substantial influence on the pooled SIR for overall cancer in DM according to the sensitivity analysis. Conclusions Compared with the general population, DM patients are at a significantly increased risk for developing cancer. Understanding the magnitude of this risk is highly relevant toward assisting healthcare providers in clinical decision making, such as screening DM patients for cancer.

J. M. Olazagasti et al.

Key Points Compared with the general population, dermatomyositis (DM) patients have about a fivefold increased risk for developing a malignancy. The increased risk of lymphatic/hematopoietic malignancies was especially notable since these malignancies were associated with the highest standardized incidence ratio, which was more than 22-fold higher than in the general population. Understanding the magnitude of cancer risk is highly relevant toward assisting healthcare providers in clinical decision making, such as screening DM patients for malignancy.

1 Introduction Dermatomyositis (DM) is a systemic autoimmune disease characterized by progressive, symmetrical muscle weakness and distinctive cutaneous lesions. Although an association between DM and cancer has been reported since 1916 by Stertz [1], an accurate assessment of the degree of this risk has not been elucidated since estimates of the associated risk vary widely in the literature [2–15]. Some studies have reported that the risk of cancer after diagnosis of DM is not significant [2, 12, 13], while others have found more than a fivefold increased risk of cancer in DM patients [4, 5, 7, 8, 10]. Several methodological factors have hindered resolution of this problem. First, as DM is a rare disease many studies have been hampered by small sample sizes, and they are usually conducted at large tertiary referral hospitals, resulting in case selection and referral biases. Second, most studies utilized the old Bohan and Peter criteria [16, 17], which may lack sensitivity in identifying DM patients and specificity in distinguishing DM from polymyositis (PM). Furthermore, it fails to define subtypes of DM, such as amyopathic DM. Third, while there have been epidemiological studies assessing cancer risk in DM, some have not compared this risk with a control group or the general population [18–23]. At the present time, observational studies with comparison groups, such as case-control studies, where the comparison group is a non-disease group, and cohort studies where the comparison group is the general population, are the best evidence available. Case-control studies assessing cancer risk in DM are sparse and have failed to reach statistical significance, likely due to the lack of power from small sample sizes. In contrast to case-

control studies, the number of cohort studies is far superior, probably because of the simplicity of comparing the risk of cancer in DM with that of the general population. However, the age- and sex-standardized incidence ratios (SIRs) calculated in these cohort studies are inconsistent [2–11]. The aim of our study was to examine the risk of cancer in patients with DM compared with an age- and sex-matched general population. To assess this risk, we performed a meta-analysis of cohort studies.

2 Methods 2.1 Data Sources and Searches An expert medical librarian with extensive meta-analytical experience collaborated to design the search strategies (PJE). Six different databases (PubMed, Ovid MEDLINE, EMBASE, Web of Science, Scopus, and the Cochrane Collaboration) were searched without language restriction for the terms ‘dermatomyositis’ or ‘myositis’ combined with ‘cancer’, ‘malignancy’, ‘neoplasm’, ‘tumor’, ‘tumour’, or ‘carcinoma’, to 1 May 2014. We also used the medical subject heading terms ‘dermatomyositis& and ‘neoplasms&for our search of the MEDLINE database. In addition, the references in the identified or related articles were then manually reviewed (JMO and PJE) in the search for other relevant citations. 2.2 Study Selection Studies were eligible for inclusion if they were DM cohort studies that reported the main outcome of interest, which was the calculated relative risk of cancer compared with the general population, generally estimated as the age- and sex-adjusted SIR. The SIR was the selected outcome of interest as it provides a point estimate of relative risk and is accompanied by a 95 % confidence interval (CI). In terms of time to development of cancer after diagnosis of DM, studies were considered for inclusion only if the patients included were diagnosed with cancer concurrently (within 2 years) or after diagnosis of DM. Studies that included patients whose diagnosis of cancer predated their diagnosis of DM by more than 2 years were excluded from our meta-analysis. Studies regarding cancer mortality, case controls, case reports, case series, and review articles were excluded from this study. We excluded case controls to avoid heterogeneity regarding estimates of risk (case-control studies provide an odds ratio, in contrast to cohort studies, which provide an SIR). Furthermore, case-control studies were excluded since the control group in these studies may not necessarily be representative of the general population.

Cancer Risk in Dermatomyositis

Also, if a study met our inclusion criteria, but had a patient population that overlapped with a similar study, then we included only the study with the longer follow-up. Two reviewers (JMO and PJE) independently performed the study selection. Inter-rater reliability for inclusion decisions was quantified using Cohen’s j statistic. Any disagreement was resolved by discussion and consensus. 2.3 Data Extraction and Quality Assessment of the Studies To further improve reliability, the following variables were strictly extracted from eligible studies by two investigators (JMO and PJE): authors&name(s), publication year, country where the study was conducted, type of cohort study (population-based vs. hospital-based), period of follow-up, patient demographics, definition of DM, numbers of patients studied, numbers of cancers observed, expected numbers in a matched background population, and/or observed-to-expected cancer ratios with 95 % CIs. Disagreements were resolved by discussion and consensus. We contacted the authors of the original studies to request any unpublished data. If the authors did not reply, we used the available data for our analyses. Two investigators (JMO and PJE) independently assessed the quality of the selected studies using a modification of a recently employed system [24, 25]. This system was designed with reference to the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) [26], the Quality Assessment Tool for Systematic Reviews of Observational Studies (QATSO) [27], and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) [28]. The criteria included the following: (1) representativeness of the exposed cohort; (2) diagnosis of DM based on the Bohan and Peter criteria [16, 17] or muscle pathology; (3) use of an inception cohort; (4) cancer outcomes assessed using medical records; (5) adjustment or stratification for age and sex; and (6) other relevant adjustment or stratification (such as for race, smoking, alcohol use, and immunosuppressive treatments). Disagreements were resolved through discussion and consensus. 2.4 Data Synthesis and Analysis SIRs with 95 % CIs for overall cancer in the combined (male and female) population were pooled using a randomeffects model. SIRs with 95 % CIs for overall cancer in a male- and female-only population were also pooled in the same method. Organ-specific cancer risks were combined in the same method only when data from two or more studies were available for a given type of cancer. In studies where the desired SIR was not specifically reported, it was calculated from the observed number of malignancies in

the DM cases and the expected number of malignancies in the general population presented in those studies (SIR = number of observed malignancies per number of expected malignancies), and a corresponding 95 % CI was determined assuming that the frequency of observed cases followed a Poisson distribution. Heterogeneity was assessed by means of Cochran&s Q statistic and the test of inconsistency (I2). I2 values of 25, 50, and 75 % were defined as low, moderate, and high estimates, respectively. The following pre-planned subgroup analyses were conducted for the primary outcome (overall cancer risk in the combined DM population) to assess potential confounding and explore heterogeneity: (1) type of cohort study (population-based vs. hospital-based); (2) sample size (\100 DM patients vs. C100 DM patients); (3) patients included (inclusion of children vs. non-inclusion of children); and (4) follow-up time (\5 years vs. C5 years). A study was considered to include children if the cohort included patients less than 20 years of age. Differences between subgroups were analyzed using the test of subgroup differences, and the results were expressed using p values. A sensitivity analysis was also performed for the primary outcome in order to verify the influence on the pooled estimate of any single study. Begg’s funnel plot and Egger&s regression asymmetry test were used to assess the possibility of publication bias. Stata statistical software version 12.0 (StataCorp LP, College Station, TX, USA) was used for all analyses. Unless otherwise stated, p values \0.05 were considered significant.

3 Results 3.1 Search Results and Study Characteristics A total of 1,272 articles from six different databases (PubMed, Ovid Medline, EMBASE, Web of Science, Scopus, and the Cochrane Collaboration) were initially identified as containing the specified search terms. A flow diagram for retrieval and inclusion of studies is shown in Fig. 1. Most of the articles were found in the Web of Science database (n = 423), PubMed database (n = 386), and EMBASE database (n = 222); 423 of the articles that were found were duplicates and were therefore excluded. After the initial screening of titles and abstracts, a further 799 articles were excluded and 50 full-length articles were selected for detailed analysis on the basis of their title or abstract. Thirty-two papers were observational studies that failed to meet the inclusion criteria, such as studies of prevalence and studies of incidence without a comparison group. Of the remaining papers, three were excluded since they had patient populations that seemed to overlap with

J. M. Olazagasti et al. Fig. 1 Study selection process

the patient population of a similar study [15, 29, 30]. In these cases only the study with the longer follow-up was included. In two instances [31, 32], the reported SIR included both DM and PM patients. The authors of these studies were therefore contacted via email on two occasions to verify the SIR corresponding to DM solely, but they were excluded after no response. As per protocol, three studies were excluded since they were case-control studies [12–14]. Eventually, ten articles met all the inclusion criteria [2–11]. Inter-rater reliability for inclusion decisions was adequate (j = 0.96). In most studies that met the inclusion criteria, the reported SIR of overall cancer was for the combined (male and female) population. Therefore, SIRs for the combined population were calculated in the two studies that solely

presented SIRs stratified by sex [6, 11]. Also, in terms of time to development of cancer after diagnosis of DM, most of the studies that met the inclusion criteria solely included cancers diagnosed after DM diagnosis [2, 4, 5, 7–11]. Only two studies included cancers that preceded DM diagnosis along with those that occurred after. To et al. [3] included cancers that were diagnosed 1 year before DM diagnosis, and Antiochos et al. [6] included cancers diagnosed 2 years before DM diagnosis. The characteristics of the selected studies are outlined in Table 1. These studies included seven population-based and three hospital-based DM cohorts that ranged from 49 to 1,012 patients and had mean follow-up times from 3.7 to 10.4 years. Four studies were conducted in Europe, three in Asia, two in Australia, and one in the US. Five studies

Cancer Risk in Dermatomyositis Table 1 Characteristics of cohort studies of cancer incidence in patients with DM included in the meta-analysis References

Country

Type of cohort

Limaye et al. [2]

Australia

Population-based

To et al. [3]

China

So et al. [4]

Korea

Total no. of patients with DM

Children included

Mean age at DM diagnosis

Sex (M/F)

Mean/ median follow-up (years)

Overall cancer [O/E]; SIR (95 % CI)

49

NA

NA

14/35

8

7/3.23;

Hospital-based

125

No

NA

NA

4.7

44/11.28;

Hospital-based

98

No

47.1 ± 15.4

40/58

3.70

23/1.62;

2.17 (0.86–4.46) 3.9 (2.8–5.5) 14.2 (9.0–21.3) Chen et al. [5]

Taiwan

Population-based

1,012

Yes

41.79 ± 18.96

315/697

5.09

95/18.58; 5.11 (5.01–5.22)

Antiochos et al. [6]

USA

Hospital-based

61

No

56.7 ± 15.0

15/46

NA

17/3.5; 4.86 (2.83–7.78)

Buchbinder et al. [7]

Australia

Population-based

85

No

51.7 ± 16.83

38/47

5.3

17/2.7;

Stockton et al. [8]

Scotland

Population-based

286

Yes

NA

97/189

NA

50/6.49;

6.2 (3.9–10.0)

Airio et al. [10]

Finland

Population-based

71

Yes

NA

NA

8.7

7.7 (5.7–10.1) 19/2.9;

Chow et al. [9]

Denmark

Population-based

203

Yes

48.8

NA

5

31/8.16;

Sigurgeirsson et al. [11]

Sweden

Population-based

392

Yes

47 ± 19.25

145/247

10.4

61/21.03;

6.5 (3.9–10) 3.8 (2.6–5.4) 2.9 (2.22–3.73) NA not available, DM dermatomyositis, M male, F female, O observed cases, E expected cases, SIR standardized incidence ratio, CI confidence interval

included children in their population of DM, while the rest included solely adults, with the exception of one study where this remained unclear [2]. 3.2 Assessment of Study Quality eTable 1 (see electronic supplementary material [ESM]) presents the results of the quality assessment. All studies met the criteria for representativeness of the exposed cohort, use of an inception cohort, appropriate cancer ascertainment, and adjustment for age and sex. However, none of the studies adjusted for other factors. In terms of diagnostic criteria for DM, six studies used the Bohan and Peter criteria [3–6, 10, 11] and two studies used muscle pathology [2, 7], findings characteristic of DM. The two remaining studies [8, 9] used the International Classification of Diseases code for identification of patients with DM but it was unclear which diagnostic criteria for DM were used. 3.3 Overall Cancer Risk in Dermatomyositis (DM) The pooled SIR for the overall risk of cancer in combined male and female DM patients was 4.79 (95 % CI

3.71–5.87), with the ten studies showing high heterogeneity (I2 = 85.8 %). A forest plot of the SIRs is shown in Fig. 2. In subgroup analyses (Table 2), stratification based on type of cohort (population-based or hospital-based) resulted in a pooled SIR of 4.62 (95 % CI 3.36–5.88) for population-based studies and 6.28 (95 % CI 2.67–9.89) for hospital-based studies; however, the difference between the subgroups was not statistically significant (p = 0.60). Stratification based on sample size resulted in a pooled SIR of 5.92 (95 % CI 3.10–8.75) in studies with a sample size \100 DM patients and 4.50 (95 % CI 3.21–5.78) in studies with a sample size C100 DM patients; the difference was not statistically significant (p = 0.28). Stratification based on whether or not children were included in the study resulted in a pooled SIR of 4.87 (95 % CI 3.46–6.28) for studies that included children and 6.04 (95 % CI 3.37–8.71) for studies that did not include children; the difference was not statistically significant (p = 0.32). Stratification based on follow-up time resulted in a statistically non-significant pooled SIR of 8.60 (95 % CI 1.46–18.65) for studies with a follow-up time\5 years and a statistically significant pooled SIR of 4.18 (95 % CI

J. M. Olazagasti et al.

Fig. 2 Combined (male and female) SIRs of overall cancer in dermatomyositis, with 95 % CIs. The size of the boxes is proportional to the weight (1/SEM) of each study. CI confidence interval, SEM standard error of the mean, SIR standardized incidence ratio Table 2 Subgroup analyses of overall cancer risk in the combined DM population Subgroup

I2 (%)

Test for subgroup differences (p value)

4.62 (3.36–5.88)

88.5

0.60

6.28 (2.67–9.89)

80.8

No. of studies

Total patients with DM (n)

SIR (95 % CI)

Population-based

7

2,098

Hospital-based

3

284

\100 DM patients

5

364

5.92 (3.10–8.75)

78.9

C100 DM patients

5

2,018

4.50 (3.21–5.78)

90.9

5

1,964

4.87 (3.46–6.28)

90.4

4

369

6.04 (3.37–8.71)

73.8

8.60 (-1.46 to 18.65)

90.3

4.18 (2.85–5.50)

89.3

Type of cohort study

Sample size

Patients included Inclusion of children Non-inclusion of children

0.28

0.32

Follow-up time, years \5

2

223

C5

6

1,812

0.42

DM dermatomyositis, SIR standardized incidence ratio, CI confidence interval

2.85–5.50) for studies with a follow-up time C5 years. However, a test for subgroup differences found this difference in estimates to be non-significant (p = 0.42). A sensitivity analysis was performed to assess the stability of the meta-analysis of the overall risk of cancer in combined male and female DM patients (see ESM eFig. 1). When any single study was deleted, the corresponding pooled SIRs were not substantially altered. The statistically similar results indicated the stability of the meta-analysis. Begg’s funnel plot for the overall risk of cancer in combined male and female DM patients was symmetric in

appearance, indicating lack of potential publication bias (see ESM eFig. 2). Results of Egger’s regression asymmetry test results were not significant (p = 0.97). The latter test indicates absence of significant publication bias. 3.4 Overall Cancer Risk in DM Based on Sex Of the ten selected studies, only six separately reported the SIR for overall cancer in male and female DM patients. The pooled SIR for the overall risk of cancer in male DM patients was 5.36 (95 % CI 5.19–5.53) with high

Cancer Risk in Dermatomyositis Table 3 Pooled SIRs for cancers among patients with dermatomyositis Type of cancer

Sex of included population

No. of studies

Pooled SIR

95 % CI

All cancers

Male and female

10

4.79

(3.71–5.87)

All cancers

Male only

6

5.36

(5.19–5.53)

All cancers

Female

6

5.09

(4.95–5.22)

Lung

Male and female

5

19.74

Ovary

Female

5

5.39

Breast

Female

5

3.52

(3.28–3.75)

Colon

Male and female

4

4.13

(3.77–4.49)

Lymphatic and hematopoietic

Male and female

3

22.72

Stomach

Male and female

3

1.03

(18.91–20.58) (4.65–6.13)

(20.37–25.07) (0.83–1.24)

Prostate

Male

2

4.90

(-2.55 to 12.36)

Bladder Cervix

Male and female Female

2 2

4.05 3.28

(3.47–4.62) (2.91–3.66)

Pancreas

Male and female

2

3.07

(2.46–3.68)

Esophagus

Male and female

2

3.06

(2.45–3.67)

SIR standardized incidence ratio, CI confidence interval

heterogeneity (I2 = 88.7 %), while the pooled SIR for the overall risk of cancer in female DM patients was 5.09 (95 % CI 4.95–5.22) with moderate heterogeneity (I2 = 61.9 %) (Table 3). 3.5 Organ-Specific Cancer Risk in Patients with DM Significant increases were observed in the risk of cancer of the lymphatic and hematopoietic system [SIR 22.72 (95 % CI 20.37–25.07)], lung [SIR 19.74 (95 % CI 18.91–20.58)], ovary [SIR 5.39 (95 % CI 4.65–6.13)], colon [SIR 4.13 (95 % CI 3.77–4.49)], bladder [SIR 4.05 (95 % CI 3.47–4.62)], breast [SIR 3.52 (95 % CI 3.28–3.75)], cervix [SIR 3.28 (95 % CI 2.91–3.66)], pancreas [SIR 3.07 (95 % CI 2.46–3.68)], and esophagus [SIR 3.06 (95 % CI 2.45–3.67)] (Table 3). However, there was no increase in the risk of stomach or prostate cancer.

4 Discussion This study revealed an association between DM and an increased risk of cancer, particularly of cancers of the lymphatic/hematopoietic system (including non-Hodgkin lymphoma, leukemia, and multiple myeloma), lung, ovary, colon, bladder, breast, cervix, pancreas, and esophagus. We detected an increased risk of cancer in patients with DM compared with the general population. Although one study [2] did not demonstrate such an increase in the risk of cancer after diagnosis of DM, the rest of the studies yielded results similar to ours. Furthermore, Yang et al. recently published a meta-analysis on cancer risk in PM and DM patients. Interestingly, our findings are consistent with their

results regarding overall cancer risk in DM. In the metaanalysis of Yang et al. [33], the pooled SIR of overall cancer in patients with DM was 5.50 (95 % CI 4.31–6.70), which is very close to our pooled estimate. An underlying mechanism for the association between DM and cancer remains inconclusive. One theory is known as the crossover autoimmunity. Myositis autoantigen expression is increased in several cancers known to be associated with DM but not in their related normal tissues, demonstrating that tumor cells and undifferentiated myoblasts are antigenically similar [34]. Therefore, it is proposed that the association between DM and cancer may be due to an autoimmune response directed against cancer which cross-reacts with regenerating muscle cells, and then enables a feed-forward loop of tissue damage and antigen selection [34]. The increased risk of lymphatic/hematopoietic malignancies which clustered non-Hodgkin lymphoma, leukemia, and multiple myeloma, was especially notable since these malignancies were associated with the highest SIR, which was more than 22-fold higher than in the general population. All the studies [5, 8, 9] reporting these types of malignancies followed the trend seen in our meta-analysis. In a case series of 18 patients with DM, approximately half (56.3 %) developed a lymphatic/hematopoietic malignancy, with lymphoma being the most frequent cancer [35]. Interestingly, an increased risk of lymphatic/hematopoietic malignancies has been also suggested for other autoimmune diseases, including primary Sjo¨gren syndrome (pSS), rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). A meta-analysis looking at the association between lymphoma and these autoimmune diseases reported a high risk of lymphoma development for pSS [SIR 18.8

J. M. Olazagasti et al.

(95 % CI 9.5–37.3)], moderate risk for SLE [SIR 7.4 (95 % CI 3.3–17.0)], and lower risk for RA [SIR 3.9 (95 % CI 2.5–5.9)] [36]. Also prominent was the increased risk of lung cancer since this cancer was associated with the second highest SIR, which was approximately 20-fold higher than in the general population. The five studies [5, 8–11] included in this assessment followed the trend seen in the meta-analysis. Unfortunately, the studies included in this metaanalysis did not adjust for cigarette smoking, a known independent risk factor for lung cancer. Studies evaluating the frequency of smoking habits among DM patients are sparse. Nevertheless, a recent study assessing the frequency of metabolic syndrome in DM patients reported that the frequency of smoking in DM patients was similar to that of the controls [37]. The primary strength of this study was its reliance on the relative risk compared with the general population, which is the age- and sex-adjusted SIR. While there are several cohort studies assessing the incidence of cancer in patients with DM, some do not compare this incidence with the general population, and thus were excluded from our metaanalysis. Also, although we excluded three case-control studies since the odds ratio is a different estimate of risk and these studies did not differentiate PM from DM, one of these studies [13] demonstrated an increase in the risk of overall cancer which was in accordance with our results [12, 14]. Our study also has several potential limitations. First, the observational nature of the investigations included in the meta-analysis is prone to bias of various sources. Specifically, the data in most studies were retrospectively retrieved from record linkage between various healthcare databases which may not contain detailed clinical information. As a consequence, relevant confounding factors such as smoking status could not be considered. Second, there was significant heterogeneity among the studies in terms of data sources and populations examined; however, the heterogeneity had no substantial influence on the pooled SIR for overall cancer according to the sensitivity analysis. Third, time to development of cancer was evaluated in a meta-analysis but three of the four studies included analyzed both PM and DM together. Fourth, not all studies reported organ-specific cancer SIRs; therefore, some of the meta-analyses performed to evaluate organspecific cancer risk in patients with DM contained a limited number of studies. Regarding subgroups of DM, amyopathic DM could not be assessed in this study since the majority of cohort studies assessing cancer risk in DM use the Bohan and Peter [16, 17] criteria, which does not include amyopathic DM. Similarly, juvenile DM could not be included in a separate meta-analysis since only one of the ten studies that

met the criteria for inclusion separately analyzed cancer risk in juvenile DM [6]. However, since some of the studies included in our main meta-analysis included children, we stratified by inclusion or non-inclusion of children, and found that the pooled SIR for studies that included children was lower than that of the studies that did not include children. However, the difference between these two subgroups was not statistically significant. Screening DM patients for cancer is common practice but unfortunately there is no consensus at the current time regarding what type of screening and to what extent should the screening be performed. Conventional cancer screening in patients with DM includes a comprehensive medical history, complete physical examination, and laboratory tests (complete blood count and serum chemistry panel) with appropriate follow-up on any identified abnormalities [38, 39]. Sex- and age-related imaging studies are also included (e.g. chest radiography, thoracoabdominal computed tomography (CT) study, barium enema/colonoscopy, transvaginal ultrasonography, and mammography) [38, 39]. Tumor markers such as CA-125, CA19-9, carcinoembryonic antigen, and prostrate-specific antigen are generally also included in the screening process [38, 39]. Newer imaging techniques such as whole-body [18F] fluorodeoxyglucose positron emission tomography (FDGPET)/CT have been used to detect occult malignant disease in paraneoplastic conditions [40]. However, one study reported that the performance of FDG-PET/CT for diagnosing cancer in patients with myositis (DM and PM) was comparable to that of broad conventional screening [39]. Also, antibodies to transcription intermediary factor-1c (antip155, anti-155/140, anti-p155/140) have been associated with cancer in adult patients with DM [41–46]. Nevertheless, the usefulness of these autoantibodies for cancer screening patients with myositis remains to be elucidated.

5 Conclusions DM is associated with an increased risk of cancer, particularly of the lymphatic/hematopoietic system, lung, ovary, colon, bladder, breast, cervix, pancreas, and esophagus. Understanding the magnitude of this risk is highly relevant toward assisting healthcare providers in clinical decision making, such as screening DM patients for cancer. Future studies are needed that can evaluate the clinical utility of the available cancer screening assessments in DM patients. Acknowledgments The authors are indebted to Patricia J. Erwin for her assistance in performing the search for studies. We also thank Victor M. Montori, M. Hassan Murad, and Colin P. West for their suggestions.

Cancer Risk in Dermatomyositis Funding/support This study was supported in part by Clinical and Translational Science Award grant number UL1 TR000135 from the National Center for Advancing Translational Sciences, a component of the National Institutes of Health (NIH). The funding sponsor was not involved. This study’s contents are solely the responsibility of the authors and do not necessarily represent the official view of the NIH.

16. 17. 18.

Conflict of interest Jeannette M. Olazagasti, Pedro J. Baez, David A. Wetter, and Floranne C. Ernste declare that they have no conflicts of interest. 19.

References 1. Stertz G. Polymyositis. Berl Klin Wochenschr. 1916;53:489. 2. Limaye V, Luke C, Tucker G, Hill C, Lester S, Blumbergs P, et al. The incidence and associations of malignancy in a large cohort of patients with biopsy-determined idiopathic inflammatory myositis. Rheumatol Int. 2013;33(4):965–71. 3. To CH, Mok CC, So H, Yip ML, Ying SKY. Standardized incidence ratios and predictors of malignancies in 215 Southern Chinese patients with inflammatory myopathies. Int J Rheum Dis. 2012;15:117. 4. So MW, Koo BS, Kim YG, Lee CK, Yoo B. Idiopathic inflammatory myopathy associated with malignancy: a retrospective cohort of 151 Korean patients with dermatomyositis and polymyositis. J Rheumatol. 2011;38(11):2432–5. 5. Chen YJ, Wu CY, Huang YL, Wang CB, Shen JL, Chang YT. Cancer risks of dermatomyositis and polymyositis: a nationwide cohort study in Taiwan. Arthritis Res Ther. 2010;12(2):R70. 6. Antiochos BB, Brown LA, Li Z, Tosteson TD, Wortmann RL, Rigby WF. Malignancy is associated with dermatomyositis but not polymyositis in Northern New England, USA. J Rheumatol. 2009;36(12):2704–10. 7. 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. 2001;134(12):1087–95. 8. 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. 2001;85(1):41–5. 9. Chow WH, Gridley G, Mellemkjar L, McLaughlin JK, Olsen JH, Fraumeni JF Jr. Cancer risk following polymyositis and dermatomyositis: a nationwide cohort study in Denmark. Cancer Causes Control. 1995;6(1):9–13. 10. Airio A, Pukkala E, Isomaki H. Elevated cancer incidence in patients with dermatomyositis: a population based study. J Rheumatol. 1995;22(7):1300–3. 11. Sigurgeirsson B, Lindelof B, Edhag O, Allander E. Risk of cancer in patients with dermatomyositis or polymyositis: a populationbased study. N Engl J Med. 1992;326(6):363–7. 12. Lakhanpal S, Bunch TW, Ilstrup DM, Melton LJ 3rd. Polymyositis-dermatomyositis and malignant lesions: does an association exist? Mayo Clin Proc. 1986;61(8):645–53. 13. Manchul LA, Jin A, Pritchard KI, Tenenbaum J, Boyd NF, Lee P, et al. The frequency of malignant neoplasms in patients with polymyositis-dermatomyositis. A controlled study. Arch Intern Med. 1985;145(10):1835–9. 14. Lyon MG, Bloch DA, Hollak B, Fries JF. Predisposing factors in polymyositis-dermatomyositis: results of a nationwide survey. J Rheumatol. 1989;16(9):1218–24. 15. Kuo CF, See LC, Yu KH, Chou IJ, Chang HC, Chiou MJ, et al. Incidence, cancer risk and mortality of dermatomyositis and

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

polymyositis in Taiwan: a nationwide population study. Br J Dermatol. 2011;165(6):1273–9. Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med. 1975;292(7):344–7. Bohan A, Peter JB. Polymyositis and dermatomyositis (second of two parts). N Engl J Med. 1975;292(8):403–7. Bendewald MJ, Wetter DA, Li X, Davis MD. Incidence of dermatomyositis and clinically amyopathic dermatomyositis: a population-based study in Olmsted County, Minnesota. Arch Dermatol. 2010;146(1):26–30. Lee SW, Jung SY, Park MC, Park YB, Lee SK. Malignancies in Korean patients with inflammatory myopathy. Yonsei Med J. 2006;47(4):519–23. Neri R, Simone B, Iacopetti V, Iacopetti G, Pepe P, d’Ascanio A, et al. Cancer-associated myositis: a 35-year retrospective study of a monocentric cohort. Rheumatol Int. 2014;34(4):565–9. Wakata N, Kurihara T, Saito E, Kinoshita M. Polymyositis and dermatomyositis associated with malignancy: a 30-year retrospective study. Int J Dermatol. 2002;41(11):729–34. Zhang W, Jiang SP, Huang L. Dermatomyositis and malignancy: a retrospective study of 115 cases. Eur Rev Med Pharmacol Sci. 2009;13(2):77–80. Andras C, Ponyi A, Constantin T, Csiki Z, Szekanecz E, Szodoray P, et al. Dermatomyositis and polymyositis associated with malignancy: a 21-year retrospective study. J Rheumatol. 2008;35(3):438–44. Ni J, Qiu LJ, Hu LF, Cen H, Zhang M, Wen PF, et al. Lung, liver, prostate, bladder malignancies risk in systemic lupus erythematosus: evidence from a meta-analysis. Lupus. 2014;23(3):284–92. Onishi A, Sugiyama D, Kumagai S, Morinobu A. Cancer incidence in systemic sclerosis: meta-analysis of population-based cohort studies. Arthritis Rheum. 2013;65(7):1913–21. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) Group. JAMA. 2000;283(15):2008–12. Wong WC, Cheung CS, Hart GJ. Development of a quality assessment tool for systematic reviews of observational studies (QATSO) of HIV prevalence in men having sex with men and associated risk behaviours. Emerg Themes Epidemiol. 2008;5:23. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596): 1453–7. Huang YL, Chen YJ, Lin MW, Wu CY, Liu PC, Chen TJ, et al. Malignancies associated with dermatomyositis and polymyositis in Taiwan: a nationwide population-based study. Br J Dermatol. 2009;161(4):854–60. Hill CL, Zhang Y, Sigurgeirsson B, Pukkala E, Mellemkjaer L, Airio A, et al. Frequency of specific cancer types in dermatomyositis and polymyositis: a population-based study. Lancet. 2001;357(9250):96–100. Azuma K, Yamada H, Ohkubo M, Yamasaki Y, Yamasaki M, Mizushima M, et al. Incidence and predictive factors for malignancies in 136 Japanese patients with dermatomyositis, polymyositis and clinically amyopathic dermatomyositis. Mod Rheumatol. 2011;21(2):178–83. Maoz CR, Langevitz P, Livneh A, Blumstein Z, Sadeh M, Bank I, et al. High incidence of malignancies in patients with dermatomyositis and polymyositis: an 11-year analysis. Semin Arthritis Rheum. 1998;27(5):319–24. Yang Z, Lin F, Qin B, Liang Y, Zhong R. Polymyositis/dermatomyositis and malignancy risk: a meta-analysis study. J Rheumatol. 2015;42(2):282–91.

J. M. Olazagasti et al. 34. Casciola-Rosen L, Nagaraju K, Plotz P, Wang K, Levine S, Gabrielson E, et al. Enhanced autoantigen expression in regenerating muscle cells in idiopathic inflammatory myopathy. J Exp Med. 2005;201(4):591–601. 35. Marie I, Guillevin L, Menard JF, Hatron PY, Cherin P, Amoura Z, et al. Hematological malignancy associated with polymyositis and dermatomyositis. Autoimmun Rev. 2012;11(9):615–20. 36. Zintzaras E, Voulgarelis M, Moutsopoulos HM. The risk of lymphoma development in autoimmune diseases: a meta-analysis. Arch Intern Med. 2005;165(20):2337–44. 37. de Moraes MT, de Souza FH, de Barros TB, Shinjo SK. Analysis of metabolic syndrome in adult dermatomyositis with a focus on cardiovascular disease. Arthritis Care Res (Hoboken). 2013;65(5):793–9. 38. Sontheimer RD. Clinically amyopathic dermatomyositis: what can we now tell our patients? Arch Dermatol. 2010;146(1):76–80. 39. Selva-O’Callaghan A, Grau JM, Gamez-Cenzano C, VidallerPalacin A, Martinez-Gomez X, Trallero-Araguas E, et al. Conventional cancer screening versus PET/CT in dermatomyositis/ polymyositis. Am J Med. 2010;123(6):558–62. 40. Berner U, Menzel C, Rinne D, Kriener S, Hamscho N, Dobert N, et al. Paraneoplastic syndromes: detection of malignant tumors using [(18)F]FDG-PET. Q J Nucl Med. 2003;47(2):85–9. 41. Fiorentino DF, Chung LS, Christopher-Stine L, Zaba L, Li S, Mammen AL, et al. Most patients with cancer-associated dermatomyositis have antibodies to nuclear matrix protein NXP-2 or

42.

43.

44.

45.

46.

transcription intermediary factor 1. Arthritis Rheum. 2013;65(11):2954–62. Trallero-Araguas E, Rodrigo-Pendas JA, Selva-O’Callaghan A, Martinez-Gomez X, Bosch X, Labrador-Horrillo M, et al. Usefulness of anti-p155 autoantibody for diagnosing cancer-associated dermatomyositis: a systematic review and meta-analysis. Arthritis Rheum. 2012;64(2):523–32. Trallero-Araguas E, Labrador-Horrillo M, Selva-O’Callaghan A, Martinez MA, Martinez-Gomez X, Palou E, et al. Cancer-associated myositis and anti-p155 autoantibody in a series of 85 patients with idiopathic inflammatory myopathy. Medicine. 2010;89(1):47–52. Kaji K, Fujimoto M, Hasegawa M, Kondo M, Saito Y, Komura K, et al. Identification of a novel autoantibody reactive with 155 and 140 kDa nuclear proteins in patients with dermatomyositis: an association with malignancy. Rheumatology (Oxford). 2007;46(1):25–8. 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. 2007;66(10):1345–9. Targoff IN, Mamyrova G, Trieu EP, Perurena O, Koneru B, O’Hanlon TP, et al. A novel autoantibody to a 155-kd protein is associated with dermatomyositis. Arthritis Rheum. 2006;54(11): 3682–9.

Cancer risk in dermatomyositis: a meta-analysis of cohort studies.

An association between dermatomyositis (DM) and cancer has been reported since 1916; however, estimates of the associated risk vary widely. For cost-e...
447KB Sizes 1 Downloads 8 Views