Clinically Amyopathic Dermatomyositis: Title and subTitle BreakWhat Can We Now Tell Our Patients?

I first started caring for patients with dermatomyositis (DM) skin disease under the guidance of James N. Gilliam, MD, while serving as an immunodermatology research fellow in the Department of Dermatology at UT Southwestern Medical Center in Dallas, Texas, in 1976. Among the patients I encountered working with him were individuals who for atypically prolonged periods (>6 months) displayed hallmark cutaneous lesions of DM as an isolated clinical finding. That such patients had not developed the expected systemic manifestation of proximal muscle weakness during this time frame made them unusual. At that time, such patients were commonly referred to by dermatologists as having “DM sine myositis.” Such patients were not then recognized by nondermatologists at that time as having a form of DM because they did not meet the Bohan classification criteria for DM.^{1 - 2}

Jim Gilliam at that time believed that most such patients were destined to develop clinically significant systemic manifestations of DM, especially myositis. As such, we typically treated these individuals from the outset with aggressive systemic immunosuppressive therapy with the hope of preventing the appearance of such systemic manifestations.

At that time there was virtually no MEDLINE-cited published literature on the concept of DM sine myositis. When asked, I struggled to know what to tell such patients about their illness and to answer the questions they posed concerning their prognosis. As time went by and I gained more personal clinical experience in this area, I began to question whether our approach at UT Southwestern to this clinical problem was the correct one.

In 1990, with the help of a very bright and determined UT Southwestern dermatology resident, Rebecca Euwer, MD, I began a scholarly quest to better educate myself in this area in order to better serve my patients. I say “scholarly” effort because this line of endeavor was then and continues to be a labor of love for me. I have never received research funding for the study of any aspect of DM skin disease (but not because of the lack of applying for such funds). However, this line of patient-oriented, clinically applied curiosity has rewarded me as much as any of my other “funded” research efforts over the years.

The initial fruit of my collaboration with Rebecca Euwer in this area was the publication in 1991 of a retrospective case series of 6 patients with “amyopathic DM.”³ I had preferred to use the traditional designation of “DM sine myositis” in the title of that article already knowing how violently the dermatology community can react to the proposal of a new name for an old disease (ie, our earlier presentation of the clinical designation “subacute cutaneous lupus erythematosus”). It was Dr Euwer who was insistent that we use a new designation for this subgroup of patients with DM to draw much needed attention to this clinical entity.

As I mentioned earlier, Dr Euwer was a resolute and persuasive resident. Thus, in this initial report we used a term coined earlier for this subgroup of patients by a highly regarded American academic rheumatologist, Carl Pearson, MD. In writings by Dr Pearson during the 1970s, he used the designation amyopathic DM to describe several patients whom he had personally cared for during his career, who displayed the hallmark cutaneous manifestations of DM for years without ever displaying clinically significant muscle weakness.⁴

Since then I have had the opportunity to pursue further scholarly efforts to better define the clinical significance of this pattern of DM presentation. These efforts were greatly facilitated by subsequent bright, energetic, and curious dermatology residents including Pedram Gerami, MD, Hobart Walling, MD, PhD, and Jennifer Lewis, MD, at the University of Iowa, Iowa City. In addition, international dermatology colleagues such as a Sachiko Miyagawa, MD, at Nara Medical University in Nara, Japan, have also contributed significantly. During these efforts the term amyopathic DM was morphed into the umbrella designation clinically amyopathic DM (CADM) for practical and political reasons.⁵ In parallel to our efforts during this time frame, others were pursuing this same scholarly quest. Within the United States, those individuals included Jeffrey Callen, MD, Joseph Jorizzo, MD, William James, MD, Nancy Olsen, MD, and Melissa Costner, MD. International contributors to the early characterization of CADM concept include Marzia Caproni, MD, Jean Revuz, MD, Jean-Claude Roujeau, MD, Carmen Herrero, MD, Fukumi Furukawa, MD, and Peter Dent, MD.

However, throughout this time frame I have intermittently paused to wonder whether the information I was then giving my new patients with CADM concerning their risks for systemic disease manifestations and their overall prognosis was any better than that which I was giving the new patients with CADM I had seen in the 1970s and 1980s. Until now, virtually everything that has been written by me and others concerning the CADM subset has been based on anecdotal, retrospective observations adulterated by various forms of bias (eg, patient selection bias, publication bias). Until now there have been no population-based studies of this clinical entity from which to more accurately draw conclusions. However, this is no longer the case. In this issue of the Archives, Bendewald and colleagues⁶ at the Mayo Clinic in Rochester, Minnesota, report the results of an analysis of data from the Rochester Epidemiology Project that address the incidence and prevalence of classic DM and CADM.⁶ In this editorial I will comment on their findings from the perspective of the impact of their work on the practical management of CADM.

Bendewald et al⁶ identified all patients with DM (both adult onset and juvenile onset) in Olmstead County, Minnesota, between 1976 and 2007, using the Rochester Epidemiology Project database. They then used this clinical cohort data set to determine the population-based incidence and prevalence of traditional (classic) and clinically amyopathic DM. While previous population-based studies of the epidemiology of classic DM have been presented,^{7 - 8} the work of Bendewald et al⁶ represents the first report of a population-based study of the epidemiology of CADM.

Bendewald et al⁶ identified 29 validated cases of DM over a 32-year period in Olmstead County, Minnesota. Of these cases, 6 (21%) were classified as CADM (3 with amyopathic DM and 3 with hypomyopathic DM). However, this research group, led by Mark Davis, MD, in the Department of Dermatology at the Mayo Clinic in Rochester, also identified 3 additional cases of CADM that evolved to classic DM during the study interval (mean follow-up period of 6.85 years). Thus, in reality, 9 (31%) of the 29 patients with DM identified in this study population over a 3-decade interval experienced the hallmark cutaneous manifestations of DM for 6 months or longer without experiencing the most common systemic manifestations of DM: clinically significant muscle weakness and interstitial lung disease. This CADM/classic DM ratio (31%) was similar to but somewhat higher than what has been reported in most observational studies (10%-20%). The age- and sex-adjusted incidence of CADM was 2.08 per 1 000 000 persons. Most of the patients identified in this study presented with CADM as adults, similar to earlier observational studies. There was a marked female predominance and white race in the patients with CADM identified in this study, also similar to prior observational studies.

This study also identified higher annual rates of CADM in the final third of the study interval compared with the first third, suggesting the possibility that CADM might be increasing in frequency. However, the term amyopathic DM was first used in the MEDLINE-cited literature to designate this subset of DM in 1991, a year which occurred in the middle of the Bendewald et al⁶ study interval. Thus, case ascertainment bias cannot be excluded as the cause for this apparent increase in CADM age- and sex-adjusted incidence over the past 30 years. If this is the case, one wonders what all those CADM cases in the past were otherwise diagnosed as. Because patients with CADM have photosensitive interface dermatitis and a positive ANA test result, I suspect that many could be found under diagnoses such as “poikilodermatous cutaneous lupus erythematosus.”

First, we can tell our patients that this population-based study further documents that they are not alone. The study by Bendewald et al⁶ further solidifies the clinical concept of CADM as a distinctive subtype of DM that is more common than many have allowed in the past. In this study, CADM was twice as common as juvenile DM, a long recognized subtype of the idiopathic inflammatory myopathies. Any future reclassification of DM and the other idiopathic inflammatory myopathies will be remiss if it does not include an appropriate niche for both juvenile-onset and adult-onset CADM.

On becoming familiar with the concept of “skin-only” DM, the first thing a patient with newly diagnosed CADM wants to know is their risk for later developing the internal manifestations of classic DM including muscle weakness and interstitial lung disease. Data from the study by Bendewald et al⁶ suggest that approximately 1 in 3 patients with CADM are at risk for such outcomes. This level of risk is similar to that identified in recent systemic reviews of the published anecdotal literature on juvenile-onset and adult-onset CADM.^{9 - 10}

Patients also want to know if they prove to be the 1 in 3 of the unlucky patients with CADM who later develop systemic disease, at what time interval might this be expected to happen. The time intervals between the onset of DM skin disease activity and initial clinical manifestations of systemic disease in the 3 patients with CADM that evolved to classic DM identified in the study by Bendewald et al⁶ were not presented. In 14 evaluable cases of adult-onset CADM identified from a review of the literature published in 2006, this interval was 15 months to 6 years.⁹ Unfortunately, there are not enough published data available to comment on such intervals for interstitial lung disease.

It is now recognized that 20% to 25% of patients with adult-onset classic DM are at risk for developing a diverse array of internal malignant conditions. Such risk is especially high in older individuals, and according to other population-based studies, this risk disappears altogether within 3 years of the onset of classic DM. While a risk for internal malignancy has been suggested to be increased in patients with adult-onset CADM, there has been debate concerning the degree of such risk in comparison with adult-onset classic DM.

Internal malignancy was present in 2 of the 9 patients with CADM (22%) in the study by Bendewald et al.⁶ However, the temporal relationship between the onset of malignancy and the onset of CADM in these 2 cases could be questioned with respect to the significance of such associations. In 1 of these patients, esophageal cancer developed 12 years after the onset of CADM (this was 1 of the 3 patients with CADM that ultimately evolved to classic DM). This time-to-malignancy-diagnosis interval is much longer than previous population-based studies would accept as a significant association (ie, 3 years).¹¹

In the study by Bendewald et al,⁶ patients with classic DM had a higher risk of internal malignancy than patients with CADM (odds ratio, 4.61). However, this difference was not statistically significant. As I previously noted anecdotally, to my knowledge there has been only 1 case of internal malignancy occurring within 3 years of appearance of DM skin disease in more than 100 patients with CADM whom I have personally cared for in over 3 decades of academic clinical practice.^{5 ,9} The available data suggest that like juvenile-onset classic DM, juvenile-onset CADM carries no risk of internal malignancy. However, until the question of the true risk of internal malignancy in patients with adult-onset CADM is determined by larger population-based studies, I, like other practitioners, currently will continue to suggest that all adults presenting with hallmark skin lesions of DM undergo internal malignancy screening at the time of presentation and repeatedly on an annual basis for a minimum of 3 years after diagnosis.

Another question that often arises in the management of both adult-onset classic DM and CADM is what type and extent of internal malignancy screening is needed? The traditional internal malignancy screen in patients with DM includes a comprehensive medical history, physical examination, and laboratory screen with comprehensive follow-up on any identified abnormalities. Sex- and age-related imaging studies are also included (eg, chest radiography, barium enema/colonoscopy, mammography). In addition, transvaginal color Doppler ultrasound screening is now recognized as having incremental value in detecting ovarian cancer.¹² Subsequently, tumor markers were added to the screening process including CA-125. Within the general population, screening with such markers has little clinical utility in detecting cancer. However, within a population at increased risk of malignancy such as those with adult-onset classic DM, they can have clinical utility.¹³ One could speculate that screening for genetic markers for specific tumor types such as BRCA-1 and BRCA-2 for breast and ovarian cancer might also have clinical utility in the setting of adult-onset DM.

With the advent of modern sensitive imaging techniques, some practitioners have begun to preemptively order total body computed tomography and/or magnetic resonance imaging scans during the initial phase of malignancy screening for DM. However, a recent review by authors who are highly conversant with these issues continues to recommend only traditional age- and sex-directed internal malignancy screens.¹¹ However, some question the clinical utility and cost-effectiveness of internal malignancy screening in any type of DM. In a recent review of the relationship between musculoskeletal disorders (including DM) and internal malignancy from a rheumatology group in Israel, it was stated

It is generally accepted that an extensive search for malignancy in most patients with a recent-onset musculoskeletal disorder of unknown etiology is not cost effective, and thus not to be recommended unless a patient presents additional findings suggestive of malignancy.^{14 (p103)}

I currently tell my adult patients presenting with CADM that they might or might not be at risk for internal malignancy for up to 3 years after their DM skin disease first appeared. However, to be on the safe side, I advise them that they should have a thorough internal malignancy screen, which I suggest should be carried out by their primary care physicians. In my recommendations to their primary care physicians in this regard, I outline what I think would be an appropriate screen for underlying malignancy. This includes a complete medical history, a review of systems, a sex- and age-directed physical examination, and chest radiography. If the patient is female, I also recommend a mammogram and a transvaginal color Doppler pelvic ultrasound examination to image the ovaries. In addition, I recommend screening laboratory examinations to include a complete blood cell count with differential, a serum chemistry screen, urinalysis, and testing for CA-125 tumor marker. I have traditionally deferred to the judgment of the primary care physician whether additional surveillance imaging studies are obtained. It appears that in many such cases, abdominal and pelvic computed tomographic scans are currently obtained for both male and female patients.

Interstitial lung disease has been reported to occur in 10% to 40% of classic DM patient cohorts. As with adult-onset classic DM, interstitial lung disease has been anecdotally reported to occur in patients with adult-onset CADM. Some such patients have experienced rapidly progressive and fatal lung disease. A high percentage of such reports have involved Japanese patients with DM. Japanese ethnicity might confer an especially high risk of DM-induced interstitial lung disease. (Alternatively, the routine use of high-resolution computed tomographic scans by Japanese physicians in the initial workup of new patients with DM could contribute to the high rates of interstitial lung disease in Japan.) However, there has been concern that the risk of interstitial lung disease in CADM might have been exaggerated by reporter and publication bias.

Of the 9 patients with CADM in the study by Bendewald et al,⁶ 1 (11%) had interstitial lung disease. This patient was 1 of 3 in this study who had CADM that evolved to classic DM. None of the 6 patients with CADM in this study who had CADM that did not evolve to classic DM during the study interval experienced interstitial lung disease. The true risk of interstitial lung disease in patients with CADM awaits a larger population-based study. At this time, it seems prudent to advise all patients with DM to report to their physicians symptoms of dry cough and/or shortness of breath that last more than 3 days.

My patients with CADM are typically more concerned about the risk of later developing the traditional internal manifestations of muscle weakness or internal malignancy than interstitial lung disease. I tell my patients with CADM that their risk for developing interstitial lung disease appears to be quite low and may have been exaggerated by publication bias. As with internal malignancy, I am personally aware of only 1 patient with CADM for whom I have cared during the past 3 decades who developed a mild form of interstitial lung disease several years after the onset of her DM skin lesions. I instruct all my patients with DM to let me or their primary care physicians know if they notice a dry cough or shortness of breath that lasts longer than 2 to 3 days.

Unfortunately, survival data were not presented for the patients with CADM in the study by Bendewald et al.⁶ However, they did report that at the end of the 6.85-year mean study interval, 8 of the 29 patients with DM (28%) had died, with 6 of these deaths being attributable to metastatic cancer (21%). Thus, internal malignancy appeared to be the major determinant of survival in this study population. Malignancy was identified in only 1 of the 9 patients with CADM in this study (11%). Thus, it appears safe to conclude that patients with CADM for the most part enjoyed a better overall prognosis in this study compared with patients with classic DM. Likewise, only 1 in 9 patients with CADM (11%) in the study by Bendewald et al⁶ experienced interstitial lung disease. This rate is similar to that supported by prior observational studies of the CADM.^{9 - 10}

The patient population in the study by Bendewald et al⁶ was different from prior published cohorts of patients with CADM with respect to the presence of myositis-specific autoantibodies such as Jo-1. Of the 6 patients with CADM in this study, 3 (50%) were Jo-1 positive. This autoantibody specificity has been virtually absent from previously published cohorts of patients with CADM.^{9 - 10 ,15} Within the context of classic DM, Jo-1 (an autoantibody to the cytoplasmic enzyme, histidyl transfer RNA synthetase) can have prognostic significance in marking subgroups of patients at risk for certain patterns of systemic disease risk. Patients with DM are at risk for interstitial lung disease when they display the combined features of Jo-1 autoantibody, Raynaud phenomenon, arthritis, and the “mechanics hand” skin lesion (ie, the antisynthetase syndrome). It would have been interesting to know if Jo-1 had any prognostic value for progression to systemic disease in the 3 patients with CADM that evolved to classic DM in the study by Bendewald et al.⁶

Jo-1 autoantibody test results were reported in this study as units, implying that they were likely assayed by a more quantifiable solid-phase immunoassay as opposed to the technically more challenging immunodiffusion assays that were originally used to identify these antibodies. Solid-phase immunoassays tend to be more sensitive in identifying autoantibodies compared with immunodiffusion assays.¹⁵ This might have been a contributing factor to the higher incidence of Jo-1 antibody in patients with CADM in the study by Bendewald et al.⁶

It is unfortunate that data pertaining to the newer serological associations of DM were not included in the analysis by Bendewald et al.⁶ Recent work has suggested that autoantibodies to a 155-kDa antigen and a 140-kDa antigen are enriched in patients with CADM.^{16 - 18} The 155-kDa autoantibody has been suggested to also be a risk factor for internal malignancy in patients with adult-onset DM.¹⁹ In addition, the 140-kDa autoantibody has been suggested to confer increased risk for interstitial lung disease in patients with adult-onset CADM, especially those of Japanese ancestry.^{18 ,20} However, it is understandable that such data may not have been available to Bendewald et al⁶ for analysis, since these are relatively recently recognized autoantibody specificities. In addition, the research assays that have been used to identify and characterize these new autoantibodies, immunoprecipitation and Western blot, are not easily adaptable to clinical laboratories. However, progress has already been made in identifying some of these autoantibodies by solid-phase immunoassay techniques such as enzyme-linked immunosorbent assays.²⁰

This population-based study by Bendewald et al⁶ confirms earlier reviews of published anecdotal observations suggesting that approximately two-thirds of patients with CADM appear to never develop systemic manifestations of classic DM or internal malignancy. However, we do not yet have a mechanism for predicting which patients with CADM are at such risk. It would currently appear that if a patient with CADM goes for 10 years without developing such internal disease manifestations or associations, he or she is at extremely low risk for developing such disease complications after that.

This study is highly commendable on several levels; however, it would be naive to believe that it is not without fault. Their study population was biased with respect to racial/ethnic mix compared with the US population in general. As Bendewald et al⁶ have pointed out, only 3.8% of the population of Olmstead County, Minnesota, is African American compared with 12.8% for the United States overall. In addition, it has been previously pointed out that the hallmark cutaneous manifestations of DM can be more difficult to identify in heavily pigmented skin.²¹ As a result, it is possible that the incidence and prevalence of CADM might have been overestimated somewhat by Bendewald et al⁶ as a result of the white skin racial bias of their study population.

In addition, it would have been informative for Bendewald et al⁶ to have analyzed the 9 patients with CADM identified in their study as a group and the 3 patients with CADM that evolved to classic DM within this group as a subgroup. To have done so would have better addressed several clinically important questions, including, what were the time intervals between the onset of DM skin disease activity and the onset of systemic disease manifestations in the CADM-evolving-to-classic DM subgroup? Also, were the 3 patients in this subgroup different in some way from the 6 who did not? (For example, were there any clinical or laboratory differences in the 3 patients in this subgroup that might have prognostic value with respect to the later development of systemic manifestations of DM?) Such questions are vitally important to ascertaining prognosis in CADM.

Another limitation of this study is the small number of study subjects. Larger population-based studies will be needed to definitively resolve remaining questions about the epidemiology, management, and prognosis of CADM. Unfortunately, few defined patient populations are available to carry out such studies within the US health care system. In addition, it has been the personal experience of the author and his colleagues in the North American Rheumatologic Dermatology Society that studies such as this have traditionally and continue to be difficult to fund through the National Institutes of Health. As the findings of the population-based study by Bendewald et al⁶ have been in agreement with prior observational studies, perhaps efforts in this area should include larger, common-protocol, multicenter observational studies organized on a Web-based data entry platform. Accomplishing such cooperative, systemic studies is a goal of the North American Rheumatologic Dermatology Society (http://www.rheumaderm-society.org).

I must also admit that the lens through which I have viewed the clinical concept of CADM over the past 3 decades is somewhat clouded by bias. First, my personal clinical experience has been greatly biased toward patients with adult-onset CADM. In addition, I have examined patients with CADM throughout this time frame only within the context of dermatology departments at state university academic medical centers based predominately in the Southwestern and Midwestern parts of the United States. Medical practices at such tertiary care medical centers are inherently shaped by referral bias and the effects of time delay between initial disease presentation and clinical assessment.

Correspondence: Dr Sontheimer, Department of Dermatology, University of Utah School of Medicine, 30 N 1900 E, Ste 4A, 330, Salt Lake City, UT 84132 (richard.sontheimer@hsc.utah.edu).

Financial Disclosure: None reported.

Additional Contributions: I wish to express my appreciation to my department's current chairman, John J. Zone, MD, and its founding chairman, Gerald G. Krueger, MD, for creating, promoting, and continuing to support the encouraging academic environment that is conducive to the pursuit of unfunded clinical scholarship efforts such as this.