Comparative analyses of muscle MRI and muscular function in anti-synthetase syndrome patients and matched controls: a cross-sectional study

The study cohort has been described elsewhere [23]. Briefly, the cohort included 68 ASS patients and 67 healthy age and sex-matched controls examined at Oslo University Hospital between September 2011 and June 2014. All examinations were done concomitantly for each participant. The ASS cohort was largely unselected, and included all of the aaRS-positive myositis cases identified in a recent population-based myositis study from south-east Norway [24]. ASS was defined as follows: a positive test of an anti-aaRS antibody; ILD according to the American Thoracic Society [25]; and/or probable/definite PM/DM according to the Bohan and Peter criteria [26].

Serum anti-histidyl tRNA synthetase (Jo-1) and anti-SSA autoantibodies were detected by automated ELISA (EliA; Phadia, Freiburg, Germany). Anti-threonyl (PL-7), anti-alanyl (PL-12), anti-glycyl (EJ) and anti-isoleucyl (OJ) tRNA synthetases and anti-Ro52 antibodies were detected by line blot assay (Euroline Myositis kit; Euroimmune Laboratory, Luebeck, Germany).

All MRI scans were obtained on 1.5-T units (Siemens, Erlangen, Germany) with coronal and axial T1-weighted spin echo and short T1 inversion recovery (STIR) sequences of the thigh muscles. Participants were asked not to exercise 24 hours before the examination. The muscles were evaluated in the anterior, posterior and medial compartments and included muscles involved in hip flexion/abduction and knee flexion/extension. Two musculoskeletal radiologists (EK, EM), unaware of the participant’s status, analyzed the MRI scans together. If any disagreement occurred, consensus was reached by discussion. Muscle edema in each compartment was evaluated in terms of extent and signal intensity, both variables using a grading scale (0–3). For the extent of edema, grade 0 was defined as no edema, grade 1 as <33.3% (minor extent), grade 2 as >33.3% to <66.6% (moderate extent) and grade 3 as >66.6% (major extent). The signal intensity of the edema was graded as 0 = no signal, 1 = low, 2 = moderate and 3 = high signal intensity. With this method, the maximum muscle edema score for each patient was 36 (18 for each thigh). Fatty replacement was scored according to the grading system suggested by Gouttalier et al. [27]; 0 = no fat, 1 = fatty streaks, 2 = muscle greater to fat, 3 = muscle equal to fat, 4 = muscle less to fat and 5 = muscle totally replaced by fat (i.e., maximum possible score 30). A Gouttalier score ≥ 2 in at least one compartment was considered pathological. Fascial edema was defined as an abnormal thick deep fascial layer with increased STIR signal, scored as present (1) or not present (0-as was muscle volume reduction. The latter was only noted when there were obvious findings: asymmetry and/or disproportionate muscle compartments. Two arbitrary scores were derived from this assessment: (A) a total edema score with a maximum score of 42 consisting of three components—edema extent (18), edema intensity (18) and presence of fascial edema (6); and (B) a total damage score (maximum score 36) including fatty replacement (30) and presence of muscle volume reduction (6). Adding A and B gave a total MRI score of 42 + 36 = 78.

Muscle strength and muscle endurance were evaluated by two experienced physiotherapists. For muscle strength, the manual muscle test (MMT) ad modum Kendall et al. [28] was used—tested on 14 muscles (MMT14, maximum score 140) on the participants’ dominant side. The MMT14 included neck flexion/extension and peripheral and proximal muscle groups in the upper and lower extremities. For direct comparison with the MRI thigh scans, we derived a MMT4 score (maximum score 40) which included four muscles involved in hip flexion/abduction and knee flexion/extension. A slightly modified Functional Index test (FI2) [29] was used to evaluate muscle endurance, using 30 instead of 60 repetitions of neck flexion (maximum score 330).

Plasma creatinine kinase levels (P-CK) were measured in units per liter and were considered pathological if >2 times upper limit of normal (2ULN). For the ASS patients, CK values at the time of diagnosis were collected retrospectively from medical records.

IBM SPSS, version 22.0 was used for statistical analyses. Normal distributed variables’ descriptive data were presented as mean and standard deviation (SD) and non-normal distributed variables were presented as median with range. Differences in continuous variables were evaluated with Student’s t test (using paired samples T-test) for normally distributed data, and with Wilcoxon and Mann–Whitney U tests for non-normally distributed data. Associations between categorical variables were detected by the chi-square test. Correlations between MRI scores, CK values, MMT scores and FI2 scores were evaluated by Pearson correlation, and between MRI scores and subgroups of ASS by Spearman rank correlation. Multiple linear regression analyses were used to evaluate possible associations between clinical variables and MRI muscle outcome.

The ASS study cohort included 45 women and 23 men (53 with anti-Jo1, six with PL-7 and nine with PL-12 antibodies) with mean (SD) age of 47 years (13.8) at diagnosis. Median disease duration was 71 months (range 6–362). Altogether, 54 of the 68 patients (79%) had been diagnosed with myositis by their treating physician prior to study inclusion; 27 with PM, 25 with DM and two with hypomyopathic/amyopathic DM. For comparison, 97% were diagnosed with ILD [23]. The 54 ASS patients with a prior myositis diagnosis had longer median disease duration (76 months, range 6–232) than the 14 cases without myositis (22 months, range 7–230; p < 0.001). ASS patients positive for anti-Jo1 had myositis more often than the Jo1-negative patients (p < 0.001; Additional file 1: Figure S1). At study inclusion, 52/68 patients (75%) were on at least one disease-modifying drug (± oral steroids), 10 patients were on prednisolone as monotherapy and six patients did not use any immune modulating treatment.

All controls (N = 67) and 66/68 ASS patients underwent MRI of the thigh muscles (Fig. 1). Two patients were unable to complete the examination due to claustrophobia. Defined MRI abnormalities were identified in 65% of the patients and 22% of the controls (p < 0.001). In 26% of the patients there were concomitant edema and muscle damage changes, compared with 1.5% of the controls (p < 0.001). Muscle edema was more frequent in patients (38%) than in controls (12%) (p < 0.001), as was fatty replacement (42% versus 4%; p < 0.001) (Fig. 2). The magnitude of changes differed between patients and controls; muscle edema score > 4 was present in 20 patients (p < 0.001), but only in three of the controls; and fatty replacement score ≥ 6 was seen in 40 patients and 11 controls (p < 0.001). Fascial edema was present in 29% of the ASS patients and in 6% of the controls (p < 0.001). Sixteen of the 19 ASS patients with fascial edema had known myositis. Muscle volume reduction was evident in nine patients (14%) and five controls (7%) (p < 0.250). In four patients and five controls the MRI changes were asymmetric. Total MRI score ranged from 0 to 47 in the ASS patients, compared to 0 to 22 in the controls (Fig. 3a). Total MRI score ≥ 10 was found in 31 patients and four controls (p < 0.001). Total edema score ranged from 0 to 38 and from 0 to 16 in the ASS patients and controls, respectively, and the range of total damage score was 0–18 in the patients and 0–8 in the controls (Fig. 3b, c). Compared with the controls, the ASS patients scored significantly higher in all three scores (total MRI, total edema and total damage) (Table 1 and Fig. 3).

Muscle MRI changes were most common in the posterior compartment followed by the anterior and medial compartments (Fig. 4). Muscle edema was most prominent in the anterior compartment, evident in 1/3 of the patients, while fatty replacement was seen mostly posteriorly (in 36% of the patients) (Fig. 4). Fascial edema was distributed almost equal in the three compartments in about 21% of the patients.

In ASS patients previously diagnosed with myositis, MRI changes were identified in 35/53 patients (66%); 10/53 (19%) had a total MRI score = 0. Mean (SD) total MRI scores were 15.9 (14.67) in the myositis group and 5.5 (3.57) in the no myositis group (p < 0.001) (Table 2). Unexpectedly, 5/13 of the ASS patients in the no myositis group had MRI changes consistent with muscle inflammation and/or muscle damage. Muscle strength and endurance did not differ between the groups (Table 2). When dividing the ASS patients into anti-Jo-1-positive and anti-Jo-1-negative groups, the results of MRI abnormalities, CK levels and muscle strength corresponded to those of the myositis/no myositis groups (Table 2).

The only MRI score item that differed between PM and DM subsets was muscle edema, where DM patients scored significantly higher than PM patients (p < 0.039; Table 2). Fascial edema was equally frequent in PM and DM (present in eight patients in each group).

Subgroup segregation by treatment showed that patients with ongoing steroid treatment (N = 57) had lower scores of muscle edema (p < 0.014) and, consequently, total edema score (p < 0.044) than patients without steroids (N = 9; data not shown). No correlation was found between steroid treatment, muscle volume reduction or total damage score.

The six patients without any immune suppressive treatment all had known myositis and normal CK values at study inclusion. These six patients had a median (range) total MRI score of 26.5 (0–39) with a median (range) total edema score of 23.5 (0–27) (data not shown).

Plasma CK analyses were performed in 67 patients and 65 controls. No significant difference was found in median (range) CK levels between the groups (Table 1). Only 4/67 ASS patients and 1/65 controls had CK levels > 2 × ULN. Levels of CK correlated with total edema score in the ASS patients (r = 0.501, p < 0.01) but not in the controls (data not shown). Still, 23% of the ASS patients with a normal CK had total edema score > 18. All of these patients had established myositis. Retrospective CK data from the time of the ASS diagnosis was available in 55 patients. At diagnosis these patients had a median CK level of 308 (range 20–14,900), compared to 95 (range 24–1344) at study inclusion (p < 0.002). The CK values at diagnosis did not correlate with total edema score or total damage score.

The ASS group had significantly lower MMT14, MMT4 and FI2 scores than the matched controls (Table 1). In the ASS patients, significant correlations were seen between total damage score, MMT14 (r = –0.340, p < 0.01), MMT4 (r = –0.344, p < 0.01) and the FI2 test (r = –0.484, p < 0.01). None of the three muscle strength tests correlated with the total edema score.

In multiple linear regression analyses, the total edema score was associated with pathological CK levels (p < 0.001) and with anti-Ro52 and anti-Jo1 positivity (p < 0.016 and p < 0.036, respectively). No association was seen with age, gender, disease duration, CK levels at diagnosis or ongoing treatment. The total damage score in ASS was associated with age (p < 0.002) and disease duration (p < 0.015).