ARTICLES Correlation of muscle biopsy, clinical course, and outcome in PM and sporadic IBM

Nizar Chahin, MD ABSTRACT Andrew G. Engel, MD Objective: To correlate muscle biopsy findings with prebiopsy and postbiopsy clinical course and response to therapy in (PM) and sporadic inclusion body (IBM). Address correspondence and Background: Existence of pure PM has recently been questioned; subsequently, the definition and reprint requests to Dr. Andrew G. Engel, Mayo Clinic, criteria for diagnosing PM were debated. Rochester MN 55905 Methods: Patient records, follow-up information, and muscle biopsies were analyzed in 107 pa- [email protected] tients whose biopsies were initially read as PM and IBM. Results: The patients fell into three groups by combined biopsy and clinical criteria: PM, 27 pa- tients; IBM, 64 patients; PM/IBM, 16 patients with biopsy diagnosis of PM but clinical features of IBM. For the three groups, the respective mean periods from disease onset to end of follow-up were 5.9, 8.5, and 9.6 years. Another was present in 4 of 27 PM, 8 of 64 IBM, and 1 of 16 PM/IBM cases. An autoimmune serologic marker occurred in one-third of each group. Nineteen PM patients had no associated autoimmune disease or marker. Nonnecrotic fiber invasion by mononuclear cells appeared in all IBM, 17 of 27 PM, and 13 of 16 PM/IBM patients. The density of both invaded fibers and cytochrome-c oxidase–negative fibers was higher in IBM and PM/IBM than in PM. Immunotherapy improved 22 of 27 PM patients but had only transient beneficial effects in 2 of 32 IBM and 1 of 14 PM/IBM patients. Conclusions: 1) Sixteen of 43 patients (37%) with biopsy features of polymyositis (PM) had clini- cal features of inclusion body myositis (IBM). 2) Absence of canonical biopsy features of IBM from clinically affected muscles of IBM patients challenges biopsy criteria for IBM, or the IBM markers appear late in some patients, or their distribution in muscle is patchy and restricted compared with that of the inflammatory exudate. 3) The muscle biopsy is a reliable instrument in the diagno- sis of PM and IBM in close to 85% of the patients. Errors of diagnosis in the remaining 15% can be avoided or reduced by combined evaluation of the clinical and pathologic findings. Neurology® 2008;70:418–424

GLOSSARY DM ϭ ; CK ϭ ; COX ϭ cytochrome-c oxidase; IBM ϭ inclusion body myositis; PM ϭ polymyositis.

Clinical or pathologic investigation of a disease mandates definition of the criteria by which the disease is diagnosed. However, an unambiguous definition of the diagnostic criteria for polymyositis (PM) has never been proposed or universally accepted. The simplest but unsatisfactory criterion for diagnosis of PM is presence of inflammatory changes in multiple skeletal muscles. The first report of an inflammatory muscle disease by Wagner in 1863 described a patient who likely had dermatomyositis (DM). Subse- quently, it was recognized that inflammatory can also occur without cutane- ous lesions.1 In 1956, nearly a century after Wagner’s report, Eaton reviewed clinical, EMG, and muscle biopsy findings in 41 patients who had subacutely evolving and mainly proximal . The EMG showed abnormal electrical irritability with low- amplitude, rapidly recruiting motor unit potentials in all patients, and muscle specimens Editorial, page 414

e-Pub ahead of print on September 19, 2007, at www.neurology.org. From the Department of Neurology, Mayo Clinic, Rochester, MN. Disclosure: The authors report no conflicts of interest.

418 Copyright © 2008 by AAN Enterprises, Inc. of 28 of 39 patients revealed inflammatory nonnecrotic muscle fibers. These investiga- exudates. Nearly one-half of Eaton’s pa- tors also excluded patients with unspeci- tients had cutaneous findings compatible fied myositis, IBM, , with DM or , and two had features suggesting muscular dystrophy, rheumatoid arthritis. Eaton commented incomplete clinical data, or no inflamma- that the term polymyositis was distin- tion in muscle. Of note, the definition of guished by the ambiguity surrounding it PM in this study was different from previ- and proposed reserving it for patients in ous definitions2,3,5,6 in that it excluded myo- whom muscle involvement predominates.2 sitides associated with collagen-vascular The clinical heterogeneity of PM and its diseases or markers. Subsequently, other frequent association with other collagen- investigators argued that they would have vascular diseases were also recognized by included cases of “unspecified myositis” Walton and Adams3 and Rowland4 in under the rubric of PM,18 that the diagnosis 1958, and by Pearson and Rose in 1960.5 of PM should not depend on the invasion In 1975, Bohan and Peter included der- of nonnecrotic muscle fibers by mononu- matomyositis and other collagen-vascular clear cells,19 that the criteria for the diagno- diseases under the rubric of PM. The diag- sis of PM are too rigid and have low nostic criteria proposed by Bohan and Pe- sensitivity,20 and that the final diagnosis ter were similar to those used by Eaton but must weigh the clinical as well as patho- also included elevation of serum enzymes logic features.20 of muscle origin.6 In the 1970s, after spo- To further investigate the controversial radic inclusion body myositis (IBM) had issues described above, we reviewed mus- emerged as a clinical and pathologic cle biopsy findings, prebiopsy and postbi- entity,7-9 it became apparent that previous opsy clinical records, and information clinical classifications of inflammatory my- available by patient contact from patients opathies were unsatisfactory, because IBM who had muscle biopsies at the Mayo evolved slowly and often involved distal Clinic between 1998 and 2003 interpreted muscles. The pathologic criteria for the di- to show PM or IBM. We asked the follow- agnosis of PM and IBM were later modi- ing questions: 1) At what frequency are au- fied by histologic studies that identified T toimmune diseases or markers present in cell–mediated muscle fiber injury10-12 and patients whose muscle biopsies are inter- by discovery of congophilic deposits in preted to show PM and IBM? 2) Is there a IBM.13 More recent communications ques- group of patients in whom PM is not asso- tioned the reliability of pathologic criteria for ciated with another autoimmune disease or the diagnosis of IBM because some patients marker? 3) Can the muscle biopsy reliably with clinical features of IBM lacked the distinguish between PM and IBM? canonical pathologic features of the disease even on repeated muscle biopsies.14-16 METHODS Patients. Between 1998 and 2003, muscle biopsies of 146 patients were interpreted as showing PM or A controversy has recently emerged IBM. Of these patients, 107 were available for follow-up. about the existence of PM as a distinct en- Information was obtained from patient records and from all tity. A retrospective study from The Neth- patients or their families by a questionnaire regarding the erlands analyzed 165 patients who had most recent clinical status, therapy received, progression of symptoms, distribution of muscle weakness, response to been diagnosed with myositis between therapy, and the cause of death if the patient had died. When 1977 and 1998 and found that only 5 ful- a response to the questionnaire was ambiguous, the patient filled the criteria of pure PM.17 In this or the family was contacted by telephone for clarification. study, the diagnosis of PM was restricted Sixty patients were also reevaluated on return visit. All stud- ies were in accord with the guidelines of the Mayo Institu- to patients with serum creatine kinase (CK) tional Review Board. elevation greater than twice above normal Studies of biopsy specimens. Muscle specimens were and presence of endomysial mononuclear processed for histochemical analysis as described.21 The ini- cells surrounding and preferably invading tial muscle biopsies were reviewed independently by the two

Neurology 70 February 5, 2008 419 Table 1 Patient material

PM IBM PM/IBM

No. of patients 27 64 16*

Age at time of biopsy, years† (range) 56.1 Ϯ 10.9 (28 to 75) 66.7 Ϯ 8.9 (39 to 85) 68.6 Ϯ 11.3 (45 to 86)

M:F 8:19 42:22 5:11‡

No. of patients who received immunotherapy 9120 before biopsy

Follow-up after biopsy, years 3.3 Ϯ 1.8 3.3 Ϯ 2.1 5.4 Ϯ 1.8

Symptoms before biopsy, years§ 2.6 Ϯ 2.6 5.2 Ϯ 3.0 4.2 Ϯ 2.6

Disease duration at last follow-up, years࿣ 5.9 Ϯ 4.4 8.5 Ϯ 5.1 9.6 Ϯ 4.4

Values are mean Ϯ SD. * Fourteen patients had clinical features of inclusion body myositis (IBM) before and two after the time of biopsy. † Lower in polymyositis (PM) than in IBM (p Ͻ 0.001) or PM/IBM (p ϭ 0.004). ‡ More women than men in PM/IBM than in IBM (p ϭ 0.02). § Shorter in PM than in IBM (p Ͻ 0.001) or PM/IBM (p ϭ 0.006). ࿣ Shorter in PM than in IBM or PM/IBM (p Ͻ 0.001).

authors. Assessment of each muscle biopsy included deter- RESULTS Demographics. Twenty-seven PM, 64 mination of the muscle fiber area, the number of nonnecrotic IBM, and 16 PM/IBM patients were available fibers invaded by mononuclear cells per unit biopsy area, for follow-up (table 1). At the time of diagno- and the number of cytochrome-c oxidase (COX)–negative fibers per unit biopsy area. Biopsy areas were measured dig- sis, the mean age of IBM and PM/IBM patients ex- itally in suitably enlarged images. ceeded that of PM patients by approximately a decade. The mean duration of symptoms before bi- Criteria for diagnosis. The clinical and pathologic crite- ria for PM consisted of 1) subacutely evolving predomi- opsy was approximately twice as long for patients nantly proximal limb muscle weakness with or without an with IBM and PM/IBM than for patients with PM. associated autoimmune disease or marker; 2) compatible Similarly, the mean total disease duration at the last EMG findings; 3) an inflammatory exudate present in en- follow-up was longer for patients with IBM and domysium, with or without invasion of nonnecrotic muscle PM/IBM than for patients with PM. Twice as many fibers by mononuclear cells, and with or without perimysial or perivascular ; and 4) absence of clinical or women than men were diagnosed with PM or PM/ pathologic features of DM, necrotizing , muscular IBM; conversely, twice as many men than women dystrophy, or overlap syndromes with myositis but without were diagnosed with IBM. Thus, PM/IBM patients endomysial inflammation. resembled IBM patients in age at onset and mean The clinical and pathologic criteria for IBM were 1) duration of symptoms before biopsy, but in this slowly evolving muscle weakness with selectively severe in- small group and unlike in the IBM group, women volvement of the flexor or quadriceps muscles or both,22-25 with or without an associated autoimmune disease were significantly more frequently represented than or marker24; 2) compatible EMG findings26; 3) fibers harbor- men. ing vacuoles rimmed by membranous material with SMI-31 reactivity of some of the vacuoles27,28; 4) an endomysial in- Distribution of weakness. All PM patients had flammatory exudate with invasion of nonnecrotic fibers by predominantly proximal limb muscle weak- mononuclear cells with or without perivascular or perimy- ness. In IBM, 55 patients had selectively severe sial inflammation10; and 5) congophilic deposits within mus- weakness of the quadriceps and finger flexor 13 cle fibers. muscles; 8 had selective weakness of the quadri- Using the above criteria, we encountered a third group of patients whose muscle specimens had pathologic features of ceps but not of the finger flexors, and 4 of these PM without canonical features of IBM, namely rimmed had developed finger flexors weakness on vacuoles or congophilic deposits, yet had slowly evolving follow-up. One patient had selective weakness weakness with preferential involvement of the finger flexor of the finger flexors but not of the quadriceps at and quadriceps muscles. We designated this group as the time of diagnosis and on follow-up. In 14 PM/IBM. PM/IBM patients, the distribution of weakness Statistics. Means of normally distributed values were com- from the start of illness resembled that ob- pared by the t test; medians of values not normally distrib- served in IBM; in 2 PM/IBM patients, selec- uted were compared by the Mann–Whitney rank-sum test. tively severe weakness of the finger flexor and Frequencies of a given observations in different disorders were compared by the Fisher exact probability test. All tests quadriceps muscles became apparent within 2 were two-tailed. years after the onset of symptoms.

420 Neurology 70 February 5, 2008 Table 2 Serum CK levels

PM (27) IBM (64) PM/IBM (16)

Mean Ϯ SD 2,199 Ϯ 3,078 546 Ϯ 571 570 Ϯ 383

Range 50 to 13,254 67 to 3,883 200 to 1,490

Median 1,015* 402 428

Patients with normal CK 4/27 17/63 1/16

* Higher in polymyositis (PM) than in inclusion body myositis (IBM) (p Ͻ 0.001) or PM/IBM (p ϭ 0.034). CK ϭ creatine kinase.

Myalgia. Prominent was reported by 12 of Muscle biopsy findings. Nonnecrotic muscle fibers 27 PM patients but only by 2 of 16 PM/IBM patients invaded by mononuclear cells were present in all and by none of the IBM patients. IBM biopsy specimens and in the majority of PM and PM/IBM biopsy specimens. The number of Serum CK levels. The mean serum CK level at the invaded fibers per unit biopsy area was signifi- time of diagnosis was threefold to fourfold higher in PM than in IBM or PM/IBM, and was similar cantly higher in IBM than in PM or PM/IBM (ta- for IBM and PM/IBM (table 2). Of note, 4 of 27 ble 5). 29-31 patients with PM, 17 of 63 with IBM, and 1 of 16 Consistent with previous studies, mito- with PM/IBM had CK levels that fell in the nor- chondrial abnormalities were frequently present mal range. Moreover, 6 of 27 PM patients had CK in our patients. COX-negative fibers were de- elevations less than twofold above the upper limit tected in muscle specimens of 63 of 64 patients of normal. Thus, 6 of the PM patients in our se- with IBM, 20 of 27 patients with PM, and all 16 ries would not have been diagnosed as having PM cases of PM/IBM. The number of COX-negative by the criteria used by the study from The fibers was significantly higher in IBM and PM/ Netherlands.17 IBM than in PM (table 5). Three PM/IBM patients underwent repeat Presence of autoimmune disease or markers. A clin- muscle biopsies of clinically affected muscles so ically manifest autoimmune disease was present that both distal and proximal muscles were sam- in 4 of 27 PM, 8 of 64 IBM, and 1/16 PM/IBM pled 1, 4, and 6 years after the initial biopsy find- patients (table 3). These proportions were not sig- ings. None of the repeat biopsies showed nificantly different. One-third of patients in each characteristic histologic features of IBM. of the three patient groups tested positively for one or more autoimmune markers (table 4). Of Response to immunotherapy.Twenty-two of 27 pa- note, 19 PM patients had no associated autoim- tients with PM improved by different modalities mune disease or marker. The values shown in ta- of immunotherapy; 2 of 32 with IBM and 1 of 14 ble 4 likely represent an underestimate of the with PM/IBM responded transiently to immuno- frequency of associated autoimmune markers be- therapy (table 6). The proportion of patients with cause not all patients were tested for all available autoimmune markers or disease responding to autoimmune markers. immunotherapy did not differ significantly from

Table 3 Association with other autoimmune diseases

PM (27) IBM (64) PM/IBM (16)

Total with other autoimmune disease* 48 1

Hashimoto thyroiditis 12†—

Celiac disease —2†—

Idiopathic thrombocytopenic purpura —— 1

Sjo¨gren syndrome 13 —

Rheumatoid arthritis 21 —

Ocular MG —1 —

* Polymyositis (PM) and inclusion body myositis (IBM) frequencies are not significantly different. † One IBM patient had both Hashimoto thyroiditis and celiac disease. MG ϭ myasthenia gravis.

Neurology 70 February 5, 2008 421 Table 4 Association with autoimmune serologic markers (positive test result/patients tested)*

PM (27) IBM (64) PM/IBM (16)

Total with immune markers† 8/25 18/51 5/14

ANA 4/24 8/49 1/14

SSA 2/19 8/40 0/13

SSB 1/19 6/40 0/13

Jo1 1/19 0/40 0/13

Rheumatoid factor 4/17 10/27 5/10

Monoclonal protein 0/14 3/29 2/11

Anti-AChR antibody 1/12 1/30 0/9

Thyroid antibodies 1/3 2/3 0/1

* More than one marker was detected in some patients. Not all patients were tested for all markers. † Frequencies in polymyositis (PM), inclusion body myositis (IBM), and PM/IBM are not significantly different. ANA ϭ antinuclear antibodies; SSA ϭ Sjo¨gren syndrome A; SSB ϭ Sjo¨gren syndrome B; AChR ϭ acetylcholine receptor.

that of patients without autoimmune markers or Netherlands,17 but we did not exclude patients disease. Among the PM patients, responsiveness with an associated autoimmune disease or posi- to immunotherapy was not significantly different tive test result for an autoimmune serologic according to presence or absence of invaded fibers marker, or patients with CK elevations less than in their biopsy specimens. twice above normal. Consistent with previous observations,14-16 we Mortality and association with malignancy. Four PM patients died of causes unrelated to PM. Five encountered a group of 16 PM/IBM patients with IBM patients died: 2 of , 1 clinical features of IBM but with muscle biopsies of unrelated cause, and 2 of unknown cause. Two that lacked rimmed vacuoles or congophilic inclu- PM/IBM patients died; 1 of aspiration pneumonia sion, canonical markers for IBM. These biopsy and 1 of an unrelated cause. None of the three specimens may have lacked IBM features owing categories of patients investigated in this study to sampling error. However, this seems unlikely had developed a malignancy during or within 2 because each biopsy specimen was sectioned at years preceding the period of observation. multiple levels, and in three patients repeat biop- sies 1, 4, and 6 years after the initial biopsy still DISCUSSION The present study confirms previ- did not show the canonical IBM markers. The ab- ously noted demographic, clinical, and pathologic sence of canonical biopsy features of IBM from features of IBM as regards age at onset, rate of clinically affected muscles of some IBM patients progression, distribution of weakness, and typical challenges biopsy criteria for IBM, or the IBM muscle biopsy.23,25,26,32-34 Our histologic criteria markers appear late in some patients, or their dis- for the diagnosis of PM (endomysial inflamma- tribution in muscle is patchy and restricted com- tion with or without invasion of nonnecrotic fi- pared with that of the inflammatory exudate. bers by mononuclear cells and with or without Our studies indicate that autoimmune diseases perimysial or perivascular inflammation) were associated with PM and IBM are present at a low like those proposed by the recent study from The frequency after a reasonable period of follow-up

Table 5 Muscle biopsy findings

PM (27) IBM (64) PM/IBM (16)

No. of patients with nonnecrotic fibers invaded by mononuclear cells* 17 64 13

Invaded fibers/cm2, mean, median† 14, 9 51, 46 28, 17

No. of patients with COX-negative fibers‡ 20 63 16

COX-negative fibers/cm2, mean, median§ 84, 15 243, 169 251, 91

* Frequency higher in inclusion body myositis (IBM) than in polymyositis (PM) (p Ͻ 0.001) or PM/IBM (p ϭ 0.005). † Median higher in IBM than in PM (p Ͻ 0.001) or PM/IBM (p Ͻ 0.001). ‡ Frequency higher in IBM (p Ͻ 0.001) and PM/IBM (p Ͻ 0.04) than in PM. § Median higher in IBM (p Ͻ 0.001) and PM/IBM (p Ͻ 0.005) than in PM. COX ϭ cytochrome-c oxidase.

422 Neurology 70 February 5, 2008 COX-negative fibers seems to be strong evidence Table 6 Effects of immunotherapy* against IBM or PM/IBM.

PM/IBM The markedly different responsiveness to im- PM (27) IBM (64) (16) munotherapy in PM vs IBM and PM/IBM are in Treated 27 32 14 keeping with previous observations in PM and Improved† 22 2‡ 1‡ IBM (reviewed in references 24 and 47). That re- Stabilized† 21 4 sponsiveness to immunotherapy in PM does not Worsened 0289 correlate with presence or absence of nonnecrotic

Response not known 31 0 fibers invaded by mononuclear cells indicates that

Untreated 0322 the autoaggressive inflammatory exudate in and of itself does not portend an unfavorable Improved 00 0 prognosis. Stabilized 04 0 In the 16 patients with PM/IBM, the biopsy Worsened 0261 findings did not indicate IBM, but the clinical Response not known 02 1 findings and the clinical course were like those in

* Therapy in polymyositis (PM), inclusion body myositis (IBM), IBM. Although the muscle biopsy is not infallible and PM/IBM consisted of (25, 25, 12), metho- for the diagnosis of IBM, it remains important for trexate (16, 8, 6), mycophenolate mofetil (10, 3, 2), azathio- excluding other possible diagnoses, such as distal prine (5, 8, 6), IV immunoglobulin (6, 10, 2), cyclosporin (0, 0, muscular dystrophies, myofibrillar myopathies, 3), and plasmapheresis (0, 1, 0). Twenty PM, 17 IBM, and 8 PM/IBM patients received more than one modality of and neurogenic disorders. Also, as for any muscle immunotherapy. disease, the final diagnosis of IBM must rest on † Proportion of improved patients is higher in PM than in IBM combined, clinical, EMG, and pathologic criteria. or PM/IBM (p Ͻ 0.001). ‡ Transient improvement.

Received March 29, 2007. Accepted in final form May 24, (table 4). Autoimmune serologic markers appear 2007 in one-third of patients in each group (table 4). The association of PM5,6,35 and of IBM36-46 with REFERENCES other autoimmune diseases or autoimmune sero- 1. Hepp P. Ueber einen Fall von acuter parenchymato¨ser logic marker has been previously documented. Myositis, welche Geschwu¨lste bildete und Fluctuation These findings do not establish that either PM or vorta¨uschte. Klin Wochenschr 1887;24:389–391. IBM are facets of another autoimmune disease; 2. Eaton LM. The perspective of neurology in regard to instead, they argue for an autoimmune contribu- polymyositis. Neurology 1954;4:245–263. 3. Walton JN, Adams RD. Polymyositis. Edinburgh: E & tion to their pathogenesis. S Livingstone, 1958. On the basis of our findings, we disagree with 4. Rowland LP. Muscular dystrophies, polymyositis and the notion that “pure” PM is very rare, or possi- other myopathies. J Chronic Dis 1958;8:510–535. bly nonexistent,17 because 19 of our PM patients 5. Pearson CM, Rose AS. Myositis: the inflammatory dis- had no associated autoimmune disease or marker. orders of muscle. Res Publ Assoc Res Nerv Ment Dis However, PM is indeed the least frequent entity in 1960;38:422–478. the group “idiopathic” inflammatory myopathies 6. Bohan A, Peter JB. Polymyositis and dermatomyositis. N Engl J Med 1975;292:344–347. that comprise PM, IBM, and DM. 7. Yunis EJ, Samaha FJ. Inclusion body myositis. Lab In- The muscle biopsy findings in our patients in- vest 1971;25:240–248. dicate that invasion of nonnecrotic fibers by 8. Carpenter S, Karpati G, Heller I, Eisen A. Inclusion mononuclear cells occurs less frequently in PM body myositis: a distinct variety of idiopathic inflam- than in IBM, where it is always present, and that matory . Neurology 1978;28:8–17. on the average the frequency of invaded fibers per 9. Carpenter S. Inclusion body myositis, a review. J Neu- ropathol Exp Neurol 1996;55:1105–1114. unit biopsy area are close to four times higher in 10. Arahata K, Engel AG. Monoclonal antibody analysis IBM than in PM. Similarly, 63 of 64 IBM and all of mononuclear cells in myopathies: I. Quantitation of PM/IBM but only 20 of 27 PM patients had COX- subsets according to diagnosis and sites of accumula- negative fibers. Also, on average, COX-negative tion and demonstration and counts of muscle fibers in- fibers per unit biopsy area were more frequent in vaded by T cells. Ann Neurol 1984;16:193–208. IBM and PM/IBM than in PM. These statistics 11. Engel AG, Arahata K. Monoclonal antibody analysis of mononuclear cells in myopathies: II. Phenotypes of are of no direct help in distinguishing PM from autoinvasive cells in polymyositis and inclusion body IBM or from PM/IBM with confidence, and do myositis. Ann Neurol 1984;16:209–215. not predict responsiveness to immunotherapy. 12. Arahata K, Engel AG. Monoclonal antibody analysis However, absence of both invaded fibers and of mononuclear cells in myopathies: III. Immunoelec-

Neurology 70 February 5, 2008 423 tron microscopy aspects of cell-mediated muscle fiber muscle of patients with inflammatory myopathies. injury. Ann Neurol 1986;19:112–125. Acta Neuropathol 1996;91:530–536. 13. Mendell JR, Sahenk Z, Gales L, Paul L. Amyloid fila- 31. Dahlbom K, Lindberg C, Oldfors A. Inclusion body ments in inclusion body myositis: novel findings pro- myositis: morphological clues to correct diagnosis. vide insight into nature of filaments. Arch Neurol 1991; Neuromuscul Disord 2002;12:853–857. 48:1229–1234. 32. Peng A, Koffman BM, Malley JD, et al. Disease pro- 14. Amato AA, Gronseth GS, Jackson CE, et al. Inclusion gression in sporadic inclusion body myositis: observa- body myositis: clinical and pathological boundaries. tions in 78 patients. Neurology 2000;55:296–298. Ann Neurol 1996;40:581–586. 33. Rose MR, McDermott MP, Thornton CA, et al. A pro- 15. van der Meulen MFG, Hoogendijk JE, Jansen GH, spective natural history study of inclusion body myosi- Velman H, Wokke JHJ. Absence of characteristic fea- tis: implications for clinical trials. Neurology 2001;57: tures in two patients with inclusion body myositis. J 548–550. Neurol Neurosurg Psychiatry 1998;64:396–398. 34. Tawil R, Griggs RC. Inclusion body myositis. Curr 16. Calabrese LH, Mitsumoto H, Chou SM. Inclusion Opin Rheumatol 2003;14:653–657. body myositis presenting as treatment-resistant poly- 35. Matsubara S, Mair WG. Ultrastructural changes in myositis. Arthritis Rheum 1987;30:397–403. polymyositis. Brain 2003;102:701–725. 17. van der Meulen MFG, Bronner IM, Hoogendijk JE, et 36. Riggs JE, Schochet SS, Gutmann L, McComas CF, al. Polymyositis: an overdiagnosed entity. Neurology Rogers JS. Inclusion body myositis and chronic im- 2003;61:316–321. mune thrombocytemia. Arch Neurol 1984;41:93–95. 18. Miller FW, Rider LG, Plotz PH, et al. Polymyositis: an overdiagnosed entity. Neurology 2004;63:402. Letter. 37. Gutmann L, Govindan S, Riggs JE, Schochet SS. Inclu- 19. Bradley WG. Polymyositis: an overdiagnosed entity. sion body myositis and Sjo¨gren’s syndrome. Arch Neu- Neurology 2004;63:402. Letter. rol 1985;42:1021–1022. 20. Hengtsman GJD, van Engelen BGM. Polymyositis: an 38. Lane RJM, Fulthorpe JJ, Hudgson P. Inclusion body overdiagnosed entity. Neurology 2004;63:402–403. myositis: a case with associated collagen vascular dis- Letter. ease responding to treatment. J Neurol Neurosurg Psy- 21. Engel AG. The muscle biopsy. In: Engel AG, Franzini- chiatry 1985;48:270–273. Armstrong C, eds. Myology. 3rd ed. New York: 39. Khraishi MM, Jay V, Keystone EC. Inclusion body my- McGraw-Hill, 2004:681–690. ositis in association with vitamin B12 deficiency and 22. Sekul EA, Chow C, Dalakas MC. Magnetic resonance Sjo¨gren’s syndrome. J Rheumatol 1992;19:306–309. imaging of the forearm as a diagnostic aid in patients 40. Soden M, Boundy K, Burrow D, et al. Inclusion body with sporadic inclusion body myositis. Neurology myositis in association with rheumatoid arthritis. 1997;48:863–866. J Rheumatol 1994;21:344–346. 23. Phillips BA, Cala LA, Thickbroom GW, Melsom A, 41. Dalakas MC, Illa I, Gallardo E, Juarez C. Inclusion Zilko PJ, Mastaglia FL. Patterns of muscle involve- body myositis and paraproteinemia: incidence and im- ment in inclusion body myositis: clinical and magnetic munopathologic correlations. Ann Neurol 1997;41: resonance imaging study. Muscle Nerve 2001;24:1526– 100–104. 1534. 42. Koffman BM, Rugiero M, Dalakas MC. Immune- 24. Mikol J, Engel AG. Inclusion body myositis. In: Engel mediated conditions and antibodies associated with AG, Franzini-Armstrong C, eds. Myology. 3rd ed. New sporadic inclusion body myositis. Muscle Nerve 1998; York: McGraw-Hill, 2004:1367–1388. 21:115–117. 25. Badrising UA, Maat-Schieman ML, van Houwelingen 43. Bouillot S, Coquet M, Ferrer X, Lagueny A, Leroy J-P, JC, et al. Inclusion body myositis: clinical features and Vital C. Inclusion body myositis associated with sarcoid- clinical course of the disease in 64 patients. J Neurol osis: report of 3 cases. Ann Pathol 2001;21:334–336. 2005;252:1448–1452. 44. Hengtsman GJD, Ter Laak HJ, van Engelen BGM, 26. Lotz BP, Engel AG, Nishino H, Stevens JC, Litchy WJ. Van Venrooij WJ. Anti-Jo-1 positive inclusion body Inclusion body myositis: observations in 40 patients. myositis with a marked and sustained improvement af- Brain 1989;112:727–747. ter oral prednisone. J Neurol Neurosurg Psychiatry 27. Askanas V, Alvarez RB, Mirabella M, et al. Use of anti- 2001;70:706. neurofilament antibody to identify paired-helical fila- ments in inclusion-body myositis. Ann Neurol 1996;39: 45. Wenzel J, Uerlich M, Gerdsen R, et al. Association of 389–391. inclusion body myositis with subacute cutaneous 28. van der Meulen MFG, Hoogendijk JE, Moons KGM, erythematosus. Rheumatol Int 2001;21:75–77. et al. Rimmed vacuoles and the added value of SMI-31 46. Badrising UA, Schreuder GMTh, Giphart MJ, et al. staining in diagnosing sporadic inclusion body myosi- Associations with autoimmune disorders and HLS tis. Neuromuscul Disord 2001;11:447–451. class I and II in inclusion body myositis. Neu- 29. Oldfors A, Larsson N-G, Lindberg C, Holme E. Mito- rology 2004;63:2396–2398. chondrial DNA deletions in inclusion body myositis. 47. Engel AG, Hohlfeld R. The polymyositis and dermato- Brain 1993;116:325–336. myositis syndromes. In: Engel AG, Franzini-Armstrong 30. Chariot P, Ruet E, Authier F-J, Labes D, Poron F, C, eds. Myology. 3rd ed. New York: McGraw-Hill, Gherardi R. Cytochrome c oxidase deficiencies in the 2004:1321–1366.

424 Neurology 70 February 5, 2008