Interstitial lung disease in polymyositis and dermatomyositis: Longitudinal evaluation by pulmonary function and radiology

Authors


Abstract

Objective

To estimate predictors and long-term outcome of interstitial lung disease (ILD) in patients with polymyositis (PM) and dermatomyositis (DM).

Methods

We conducted a prospective study in which newly diagnosed PM/DM patients, regardless of clinical symptoms of pulmonary disease, were investigated with repeated chest radiography, high-resolution computed tomography (HRCT) of the lungs, and pulmonary function test (PFT). Clinical, radiologic, and lung function outcome was based on the last followup results.

Results

Twenty-three patients with a mean followup period of 35 months were included. Findings on radiographic examination and/or PFT compatible with ILD were recorded in 18 patients (78%). Patients with ILD had lower lung function, higher radiologic scores, and higher creatine kinase values than those without ILD. All patients were treated with high-dose glucocorticoids and other immunosuppressive agents. Two patients died due to ILD, both with active myositis. During the followup, total lung capacity (TLC) improved in 33%, remained stable in 39%, and deteriorated in 28%. Changes in TLC correlated only partially with HRCT findings, which persisted even after normalizing for lung function.

Conclusion

ILD associated with PM/DM is in most cases mild, chronic, and has a nonprogressive course during immunosuppressive treatment. PFT can be normalized during treatment with immunosuppressive therapy, even if radiologic signs of ILD persist. The course of ILD could not be predicted on the first examination. Therefore, myositis patients with ILD need careful evaluation of clinical features as well as PFT and radiologic features during followup.

INTRODUCTION

Polymyositis (PM) and dermatomyositis (DM) are multisystemic diseases characterized by inflammation of the striated muscles. Other organ systems are frequently involved, such as the heart, gastrointestinal tract, lungs, and skin in DM. Lung involvement in PM/DM may appear as inflammatory interstitial disease and/or as a complication to muscle weakness with consequences including aspiration pneumonia or ventilatory insufficiency.

Interstitial lung disease (ILD) is a negative prognostic factor associated with increased morbidity and mortality in patients with PM/DM (1–3). However, the clinical course of ILD in myositis is to a large extent unknown, and a high variability in the clinical course, response to treatment, and prognosis is suggested from previous reports (3, 4). Acute interstitial pneumonia (3, 5, 6), forced vital capacity <65% predicted (6), excess neutrophils in bronchoalveolar lavage fluid (3, 7), and lung histology such as diffuse alveolar damage (8) have been reported to predict a poor prognosis in patients with ILD associated with PM/DM.

In a previous study of patients newly diagnosed with PM/DM without prior selection based on clinical or laboratory presentation, we determined a high prevalence (65%) of ILD that in some cases occurred without any clinical symptoms (subclinical) (9). The present longitudinal study was initiated in order to evaluate the course of ILD in patients with PM/DM, and to identify factors predictive of the ILD course. Pulmonary function tests (PFTs), chest radiographs, and high-resolution computed tomography (HRCT) scans were evaluated in a well-defined unselected population of patients with myositis over a mean followup period of nearly 3 years, and the relationship between disease activity, disease duration, and lung affection was investigated.

PATIENTS AND METHODS

Study population.

Thirty-nine consecutive patients with recent-onset myositis based on the criteria by Bohan and Peter (10, 11) were identified between March 1998 and December 2002 at the rheumatology unit at Karolinska University Hospital, Stockholm, Sweden. Three patients with inclusion body myositis according to diagnostic criteria proposed by Griggs et al (12), 7 with other defined connective tissue diseases, 5 with a malignancy diagnosed within 1 year of the myositis diagnosis, and 1 with severe emphysema were excluded. Finally, 23 white patients with a diagnosis of definitive or probable PM/DM were included; the initial findings in 17 of these patients have previously been reported (9). The study was planned to include 1 enrollment visit and followup visits after 3, 6, and 12 months, and annually thereafter. At each visit, the patients were assessed with laboratory tests, chest radiography, HRCT, and PFT. Age, sex, disease duration, initial symptoms, respiratory symptoms (cough, dyspnea), and treatment were recorded.

All patients were treated, according to a routine protocol, with high doses of glucocorticoids, usually oral prednisolone (0.75 mg/kg/day), but occasionally intravenous (IV) methylprednisolone pulse therapy (750–1,000 mg/day for 3 days) and another immunosuppressive agent based on the decision by the treating physician. Changes in therapy were made when clinically indicated. Assessment of disease activity was conducted retrospectively from patients' medical charts and was based on patients' clinical history, physical examination, laboratory data, physician's global assessment, and the intention to change therapy.

Laboratory investigations.

Laboratory tests including erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) level, and serum creatine kinase (CK) were analyzed at the Department of Clinical Chemistry and autoantibodies including rheumatoid factor (RF), antinuclear antibodies (ANAs), anti-RNP, anticentromere, anti-Ro/SSA, anti-La/SSB, anti–topoisomerase I (Scl-70), anti-Sm, and anti–histidyl–transfer RNA synthetase (Jo-1) antibodies were analyzed at the Department of Clinical Immunology, both at Karolinska University Hospital, Stockholm. RF was measured by nephelometry, ANA by indirect immunofluorescence using HEp-2 cells, and the other autoantibodies by enzyme-linked immunosorbent assay and immunodiffusion.

Pulmonary function tests.

Static and dynamic lung volume measurements were performed using a body plethysmograph (Auto box 2800; Gould Electronics, Bilthoven, The Netherlands). Assessments of transfer factor for carbon monoxide, i.e., diffusing capacity for carbon monoxide (DLCO), were performed by the single-breath technique using the same spirometry system, and the values were adjusted for alveolar volume. The PFT results were expressed as the percentage predicted (13, 14); values <80% for total lung capacity (TLC) and DLCO were considered abnormal. According to an international consensus statement of the American Thoracic Society on the diagnosis and treatment of idiopathic pulmonary fibrosis (15), changes >10% in TLC and/or >15% in DLCO were considered to be significant, and were used as determinants of improvement or deterioration.

Radiology.

Chest radiography was performed with digital phosphor-plate technique (Fuji ECR 5501, Fuji, Tokyo, Japan) and included posteroanterior and lateral views. HRCT was performed using a Somatom Plus scanner (Siemens, Erlangen, Germany). All examinations were performed in maximal inspiration without intravenous contrast medium enhancement. Slice thickness was 2 mm every 20 mm including the entire lung. Images were reconstructed using a high spatial frequency algorithm (AB 7541). The evaluation of chest radiographs and HRCT was performed according to a modified grading system in which the score indicates presence of abnormalities on the radiographic examination and not the extent of the abnormalities (16–20).

The evaluation of chest radiography was performed for 3 zones in each lung: upper (above the carina), middle (from the carina to 4–5 cm below), and lower (from 4–5 cm below the carina to the diaphragm). The severity of radiographic changes (irregular opacities, consolidation, and honeycombing) was scored 0–3 (absent, mild, moderate, severe) for each zone, giving a total score from 0 to 54.

HRCT was evaluated at 5 levels (origin of brachiocephalic arteries, mid-aortic arch, carina, confluence of lung veins, and 1 cm above the right dome of the diaphragm). Findings consistent with inflammation were acinar nodules (diffusely delineated nodules <7 mm in diameter), ground-glass opacities (hyperattenuated areas in which the bronchi and vessels remain visible), and consolidation (areas with increased attenuation and obscuration of bronchial walls and vessels). Honeycombing (areas of small cystic spaces with thickened walls) and traction bronchiectasis (bronchial dilation due to traction by fibrous tissue) were considered consistent with fibrosis. A third category including linear opacities (thickened inter- and intralobular septa) was defined as neither inflammation nor fibrosis. Each finding was coded separately as present or absent in all 10 zones.

The chest radiographs and HRCT were assessed simultaneously by consensus of 3 experienced thoracic radiologists (UT, MB, and JV). The evaluations were performed in random order and on different occasions and the evaluators were unaware of the clinical data. Agreement between the radiologists was observed in all cases.

Interstitial lung disease.

ILD was defined as restrictive lung function impairment (TLC and DLCO <80% predicted) and/or radiologic signs consistent with ILD on chest radiograph or HRCT.

Statistical analyses.

Student's t-test was used for comparison of continuous variables (age, lung function, and time since onset of symptoms) between 2 groups, and analyses of variance were used for comparison of groups improving, deteriorating, or remaining unchanged. Maximum likelihood chi-square test was used for discrete variables (sex, PM/DM diagnosis, prevalence of pathologic laboratory tests, and symptoms). For long-term evaluation of ILD prevalence, last observation for lung function and radiology was carried forward.

RESULTS

Patients' characteristics.

Fourteen patients (9 women) diagnosed with PM and 9 patients (6 women) with DM with a mean ± SD age of 56.2 ± 14.8 years were included. The main clinical and laboratory characteristics are summarized in Table 1. Increased serum CK levels were noted in all patients.

Table 1. Clinical data (at enrollment) and cumulative treatment in 23 patients with PM/DM*
Patient no.Age, yearsSexDiag.SymptomsAntibodiesLung function, % predictedRadiology changesTreatment
MonthsTypeRFANASS-ASS-BJo-1TLCVCCRHRCTGC >10 mg/day, monthsOther IS
  • *

    PM = polymyositis; DM = dermatomyositis; RF = rheumatoid factor; ANA = antinuclear antibody; TLC = total lung capacity; VC = vital capacity; CR = chest radiograph; HRCT = high-resolution computed tomography; GC = glucocorticoids; IS = immune suppressive; M = muscle weakness/myalgia; R = rash; Dy = dysphagia; A = arthralgia/arthritis; Ra = Raynaud's phenomenon; C = cough; D = dyspnea; NA = none available; AZA = azathioprine; CYC = cyclophosphamide; MTX = methotrexate; CHL = hydroxychlorochine; CYS = cyclosporine; Ig = immunoglobulin.

149FemaleDM2M, R, Dy+808921.6AZA, CYC, MTX
260MalePM9M, A, Ra, C, D, Dy+++7477++6.7CYC, AZA,
342FemalePM12M, A, Ra+++++10398+5.4AZA
451FemalePM12M, D+++1341358.9AZA
551MaleDM1A, R++7153++7.6AZA
626MaleDM4R, D, Dy++110101NA6.8CHL
764FemalePM1M, D+++6268++<0.3CYC
865FemaleDM2M, R, D+9891+NA12.6AZA
969FemalePM3M, A, D+104112+NA5.5AZA, MTX
1054FemaleDM6M, A, R, Dy, D+12812012.4CYC, MTX
1174FemalePM6M, A, D, Dy+9476++12.0CYC, CYS, MTX
1274MalePM3M, A+++5360++3.5CYC
1380FemaleDM3M, A, R+NANA++8.4AZA
1453FemalePM24M, A, Ra, C, D10092++15.0MTX
1523FemalePM3M, Ra, C, D++9089+5.7CYC, CYS
1658MalePM2M, A, C+9087NA10.9AZA
1771FemaleDM1M, R, D, Dy119109+7.1AZA, MTX, Ig
1867MalePM3M+++79NA++24.7MTX, CYC
1941FemaleDM1M, A, R, C+NANA++8.7MTX, CYC
2049MalePM5M, A, D+7570++40.6MTX, AZA, Ig
2142FemalePM10M, A, Ra, C, D6763 +/+10.2AZA, CYC, CYS
2265FemalePM2M, A, C, D++6671 +/+12.6CYC
2364MaleDM1M, R+++8586 +/+13.9CYS, MTX

All patients were treated with high doses of glucocorticoids with slowly tapering doses during the first year. Additional immunosuppressive agents administered according to the treating physician's decision are listed in Table 1. In some patients, combination therapy with ≥2 immunosuppressive agents was used.

Two patients died 10 days and 18 months, respectively, after diagnosis. In both cases the cause of death was progressive respiratory insufficiency with radiologic signs of ILD. At the time of death, disease activity was considered to be high despite treatment with high doses of glucocorticoids and IV cyclophosphamide in both patients. At autopsy, pneumonia, focal pulmonary hemorrhage, and acute lung infarction were observed in one case, and chronic alveolitis in the other.

In the remaining 21 patients, disease activity and functional status according to physician's global assessment had improved at the time of followup evaluation compared with the initial presentation. Serum CK and serum CRP levels had returned to normal in all tested patients. Relapses with temporary elevated serum CK levels were noted during the followup period in 3 patients. At the final evaluation, 15 patients were still receiving glucocorticoid treatment with a prednisolone dosage of 2.5–30 mg/day, and 14 were treated with other immunosuppressive agents such as azathioprine (n = 5), cyclophosphamide (n = 3), methotrexate (n = 5), cyclosporine (n = 2), and IV immunoglobulin monthly or every 3 months (n = 2). Complete remission without need of further immunosuppressive treatment was achieved in 1 patient.

PFTs.

PFTs were available in 21 patients; 2 declined to undergo the investigation. Repeated PFTs were performed in 18 patients (2 had died, 3 declined investigation). The changes in TLC and DLCO between the first and last followup examination are shown in Figure 1.

Figure 1.

Results of initial and followup total lung capacity (TLC) and diffusing capacity for carbon monoxide (DLCO) measurements and total chest radiography score in patients with polymyositis and dermatomyositis. Each line represents 1 individual patient.

Eight patients had restrictive lung function impairment at the initial investigation (TLC <80% predicted) with a mean TLC of 68% predicted (range 53–79%), and they all had reduced DLCO (mean 50% predicted, range 37–58% predicted). Followup investigation in 6 of these patients showed that TLC and DLCO had become normal in 4 patients, and DLCO had normalized in 2. TLC and DLCO remained unchanged in 1 and deteriorated in 1.

Seven patients had isolated reduction of DLCO (<80%). After correction for the reduced lung volume (DLCO corrected for alveolar volume), there was no significant difference in DLCO between patients with restrictive lung impairment and those with isolated reduction of DLCO (mean 59% and 62%, respectively). Followup data were available for 6 of these patients; 1 developed a restricted pattern during the observation time. In 4 patients the PFTs were unchanged. One patient had developed >10% reduction of TLC but DLCO increased during the same period, although it was still <80% of the predicted value. Isolated reduction of DLCO was not associated with pulmonary hypertension as assessed by echocardiography.

Six patients had normal TLC and DLCO at the first examination. TLC had deteriorated at followup in 2 of these patients (but still remained within the normal range), 1 with and 1 without concomitant decreased DLCO. In 1 patient, TLC and DLCO both remained stable. The other 3 had increased TLC or DLCO.

The clinical and physiologic characteristics of patients whose TLC improved (n = 6), remained unchanged (n = 7), or deteriorated (n = 5) are listed in Table 2. The patients whose TLC improved had initially lower values of forced expiratory volume in 1 second (FEV1), TLC, and residual volumes compared with those who deteriorated or remained unchanged. The mean duration of disease prior to onset of treatment was shorter and the mean values of serum CK and ESR were higher in patients who improved, although the differences did not reach statistical significance. There was no significant difference between these 3 groups regarding age, sex, diagnosis, prevalence of autoantibodies, or radiographic scores. Anti–Jo-1 antibodies were detected in 3 of 7 patients who improved and 1 of the patients who remained unchanged but in none of the patients who deteriorated. Those who improved in TLC or DLCO were all treated with a combination of immunosuppressive agents (glucocorticoids plus cyclophosphamide or azathioprine) from the time of myositis diagnosis.

Table 2. Baseline characteristics in patients with polymyositis and dermatomyositis in relation to change in TLC*
 ImprovedUnchangedDeterioratedP
Mean ± SDNo.Mean ± SDNo.Mean ± SDNo.ANOVA
  • *

    TLC = total lung capacity; ANOVA = analysis of variance; CT = computed tomography; FEV1 = forced expiratory volume in 1 second; VC = vital capacity; MVV40 = maximum voluntary volume; DLCO = diffusing capacity for carbon monoxide; DLco/VA = diffusing capacity for carbon monoxide corrected for alveolar volume; RV = residual volume; CK = creatine kinase; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate.

Age, years55.33 ± 6.65651.14 ± 16.89754.80 ± 18.3850.859
Symptoms, months3.50 ± 3.0265.86 ± 5.2178.60 ± 8.7150.378
Chest radiograph score2.83 ± 2.9962.00 ± 2.7772.75 ± 2.0640.838
CT inflammation score5.40 ± 3.7857.33 ± 5.5065.60 ± 8.3250.846
CT fibrosis score3.40 ± 2.7053.50 ± 5.2161.60 ± 2.6150.677
FEV1, % predicted73.83 ± 15.08693.71 ± 13.01798.80 ± 11.6950.015
VC, % predicted74.50 ± 13.17695.86 ± 22.20794.40 ± 17.1350.108
TLC, % predicted76.00 ± 8.27699.43 ± 22.147102.20 ± 18.2850.042
MVV40, % predicted77.33 ± 17.39687.00 ± 14.93779.60 ± 6.0250.453
DLCO, % predicted62.00 ± 20.98670.14 ± 17.13777.20 ± 28.2050.527
DLCO/VA, % predicted67.83 ± 13.50665.29 ± 11.01769.20 ± 17.5450.883
RV, % predicted86.33 ± 19.256105.33 ± 25.416133.50 ± 12.7940.013
CK, μkat/liter42.44 ± 35.85523.92 ± 27.78638.20 ± 38.4340.640
CRP, mg/liter28.75 ± 28.49410.40 ± 6.19525.00 ± 21.4040.373
ESR, mm37.33 ± 38.21322.67 ± 11.59315.60 ± 14.9850.455

Seven patients had undergone repeated PFT between 11 and 19 weeks after the initial PFT. Changes in TLC established during this period persisted at the last PFT evaluation 54–242 weeks after the initial evaluation. The same tendency was also evident for DLCO.

Radiographic imaging.

Chest radiographs were available for 22 patients. Investigation was not performed in 1 patient because HRCT showed no changes. Changes compatible with ILD were recorded in 15 (68%) patients. Repeated chest radiography was performed in 17 patients (2 had died, 4 declined investigation or a repeat test was not clinically indicated according to the treating physicians) 15–238 weeks after the initial examination. The radiographic findings are listed in Table 3. The number of normal chest radiography examinations had decreased at followup. Irregular opacities were the most frequent finding in both the first examination and at followup. Irregular opacities and honeycombing were predominantly located in the lower lung zones.

Table 3. Frequency of chest radiography and HRCT findings in patients with polymyositis or dermatomyositis at the first and followup evaluations*
FindingInitial imageFinal image
  • *

    Values are the number or the number (percentage). HRCT = high-resolution computed tomography.

Chest radiography2217
 Irregular opacities13 (59)14 (82)
 Consolidation8 (36)4 (24)
 Honeycombing2 (9)1 (6)
 Normal7 (32)3 (18)
HRCT1915
 Inflammation  
  Nodule5 (26)7 (47)
  Ground glass10 (63)8 (53)
  Consolidation7 (26)4 (27)
 Linear opacities  
  Septal lines14 (74)15 (100)
  Subpleural lines/bands8 (42)6 (40)
  Peribronchovascular thickening10 (53)8 (53)
 Fibrosis  
  Traction bronchiectasis4 (21)2 (13)
  Honeycombing4 (21)4 (27)
 Normal4 (21)0 (0)

HRCT examination was available in 19 patients. HRCT was not performed in 4 because it was not considered clinically indicated or because the patients declined the examination. Changes predominantly consistent with inflammation were noted in 11 patients and those consistent with fibrosis in 2. Mixed inflammation and fibrosis was shown in 2 patients. HRCT showed no abnormalities in 4 patients. Followup examination 12–238 weeks after the initial examination was performed in 15 of the patients (Figure 2). None of the patients with changes consistent with ILD in the initial HRCT experienced a complete resolution at followup investigation. Of the 11 patients with predominantly inflammatory findings at the first examination, 3 had developed fibrotic changes, 2 had mixed inflammatory and fibrotic changes, and 4 remained unchanged at the last examination; 1 patient had died and 1 had no followup examination. Three of the 4 patients with normal HRCT at the first examination underwent followup investigation; 1 had developed fibrotic changes, 1 inflammatory changes, and 1 linear opacity changes.

Figure 2.

Initial and followup high-resolution computed tomography scores in patients with polymyositis or dermatomyositis.

ILD.

At the time of diagnosis, 18 (78%) of the patients had signs of ILD. Seven (33%) of 21 fulfilled the criteria for restrictive ventilatory impairment by PFT. Of those with ILD diagnosed by HRCT (15 [79%] of 19 cases), findings suggestive of ILD were apparent in 12 (80%) of 15 by chest radiography and in 8 (57%) of 14 by PFT. Furthermore, none of the patients with normal HRCT had signs of ILD by chest radiography or reduced TLC values.

Patients with or without ILD did not differ regarding sex, age, myositis diagnosis, disease duration, or symptoms. Patients with ILD had higher levels of serum CK (44.5 μkat/liter versus 3.7 μkat/liter) before treatment than those without ILD. There was no difference between the groups regarding occurrence of autoantibodies, but all patients with anti–Jo-1 antibodies had ILD and a lower TLC than those without such antibodies.

During the first 6 months after diagnosis, the proportion of patients with restrictive lung function impairment decreased while the proportion of patients with ILD by chest radiography and HRCT increased (Figure 3). At the last evaluation, 3 patients had reduced TLC, <80% of the predicted value, and 3 patients had no changes on chest radiography, but all of the patients had changes on HRCT.

Figure 3.

Prevalence of interstitial lung disease (ILD) in patients with polymyositis or dermatomyositis according to different criteria during followup. N indicates the number of subjects under followup, last observation carried forward. HRCT = high-resolution computed tomography; TLC = total lung capacity.

DISCUSSION

This longitudinal study of ILD in patients with newly diagnosed PM/DM demonstrates the following clinically important results. First, ILD in PM/DM was in most cases chronic but mild and nonprogressive during immunosuppressive treatment. Second, PFTs improved in some patients who were treated with a combination of high doses of glucocorticoids and other conventional immunosuppressive agents in the early stages of disease, and in a few patients PFT even normalized despite persisting radiographic changes. However, there were still 2 fatal cases. Third, the course of ILD could not be predicted at the time of diagnosis, but changes in TLC that had occurred after 11–19 weeks of therapy persisted at long-term followup. Fourth, anti–Jo-1 autoantibodies may not only be predictive of ILD but could also have a positive prognostic value.

We found a high prevalence of ILD (78%) in this study of consecutively recruited patients with PM/DM, similar to what we previously reported in 17 of theses patients at diagnosis (9). ILD is associated with restrictive lung function impairment, which is defined as a reduced TLC. Because DLCO is commonly reduced in patients with ILD, often at an earlier stage of the disease than TLC, we included this variable in the assessments and definition of ILD. Also, forced vital capacity (FVC) is commonly reduced in ILD; however, because low FVC is not specific for restrictive lung diseases but is also reduced in obstructive lung diseases, we did not include this variable in the definition of ILD. Low TLC in patients with myositis can in some cases also be a consequence of respiratory muscle weakness, and assessments of maximal inspiratory pressure (MIP), maximal expiratory pressure (MEP), reduced maximal voluntary ventilation, and increased residual volume (RV) without decreased FEV1/FVC ratio would have been more specific methods to distinguish reduced lung volumes due to respiratory muscle weakness from ILD. However, measurement of MIP and MEP was not performed in all patients, and therefore we cannot totally exclude influence of muscle weakness on the reduced lung volumes in some patients.

Patients with improved TLC, defined according to an international consensus, had lower FEV1, TLC, vital capacity, RV, shorter symptom duration, and higher ESR at the initial investigation than those with unchanged or deteriorated TLC. This may suggest that these patients had a higher degree of inflammation and fewer irreversible changes, which could explain the favorable course. Changes in TLC that had occurred within 11–19 weeks of therapy persisted at long-term followup. This reveals a prognostic value of trends in serial PFT and suggests that the effects of immunosuppressive treatment on pulmonary function could be detected early when serial investigations are used. However, this needs to be confirmed in future studies.

Although restrictive lung function impairment accompanied by reduced DLCO was the most frequently observed finding at the time of diagnosis, an isolated reduction of DLCO was observed in one-third of the patients. Pulmonary hypertension (pulmonary artery systolic pressure >35 mm Hg at rest [21]) as a cause of isolated reduced DLCO was excluded by echocardiography. At followup, only 1 patient with isolated DLCO had developed restrictive lung impairment and another patient had developed a significant reduction in TLC, but still within the normal range. Thus, a reduced DLCO without changes in lung volume did not predict a progression to restrictive lung function impairment. However, larger studies with longer followup are needed to evaluate the clinical significance of an isolated decrease in DLCO in patients with myositis.

Chest radiography and HRCT are essential for evaluation of ILD. Although most of the patients showed changes on chest radiography as well as HRCT, the chest radiography findings were subtle and difficult to interpret, whereas the HRCT changes were more specific for ILD. None of the patients with normal HRCT had signs of ILD by chest radiography or reduction of TLC value. This finding is consistent with a higher sensitivity of HRCT compared with chest radiography and PFT to detect signs of ILD. Still, the PFTs add important information in addition to HRCT as the PFTs only partially correlated with HRCT scores and pattern. PFT is also a valuable assessment tool to evaluate the clinical consequences of changes in HRCT. The clinical significance of abnormalities on HRCT in combination with normal PFTs is currently unclear and longer followup data are needed to clarify this issue.

As in previous reports (4, 22, 23), linear opacities, ground-glass opacities, and peribronchovascular thickening were the most common HRCT abnormalities at the initial imaging. An improvement in HRCT findings was apparent in some of our patients, but in other patients progression of HRCT abnormalities was observed despite immunosuppressive treatment. Thus HRCT was a sensitive method to detect changes in extent and potential reversibility over time when a validated scoring system was used by experienced radiologists.

Anti–Jo-1 antibodies were only recorded in patients who improved or remained stable and in no patients who deteriorated, although the difference was not statistically significant. Conflicting data have been reported on the prognostic role of anti–Jo-1 autoantibodies (24, 25). Our findings suggest that anti–Jo-1 autoantibodies may not only be predictive of ILD but may also have a positive prognostic value. Further investigations are needed for a definite conclusion.

The optimal treatment for patients with ILD-associated PM/DM is not known. Our study was not designed as a clinical trial. Azathioprine, cyclophosphamide, and methotrexate were the medications most often used by our patients. Changes in therapy were made during followup when clinically indicated according to the physician's decision. For this reason effects of different treatment regimens could not be evaluated, and treatment intervention trials are needed to determine differences in these therapies. Our present findings indicate that even during treatment with immunosuppressive agents, signs of ILD on HRCT may remain, despite remission of muscle involvement. This could indicate that the underlying pathogenic mechanisms that cause muscle symptoms and ILD might not always be the same.

The clinical significance of asymptomatic ILD in patients with myositis could not be addressed by our study because none of these patients developed clinical symptoms during immunosuppressive therapy that was required due to the muscle inflammation, and it would not be ethical to investigate the natural course of asymptomatic ILD in patients with clinical signs of myositis.

There are several limitations of our study, the major one being the limited number of patients, reflecting the rareness of the disease, and the length of the followup period. Nonetheless, to our knowledge, this is the largest observational study with prospective data collection in consecutive patients with myositis regardless of clinically manifest lung symptoms. Despite the directives in the study protocol, PFT and radiographic investigations were not always performed at the same session or at the planned followup times, mainly due to the treating physician's ambitions for effective care of the patients. Thus the number of patients who had been examined with combined PFT and radiologic investigations at different time points varied, making statistical calculations problematic. For evaluation of outcome, we chose to compare the latest PFT with the initial PFT.

In conclusion, due to the high prevalence of ILD associated with PM/DM, we recommend that patients with PM/DM should be investigated using a PFT, chest radiography, and HRCT of the lungs to identify patients with ILD early in the course of their disease, because both ILD and muscle involvement could influence activities of daily living and health-related quality of life. When ILD is present, PFT should be repeated during followup as an outcome measure of immunosuppressive treatment because series of tests are needed to estimate prognosis. Although the course of ILD may vary despite immunosuppressive treatment, in most cases the prognosis was good, with PFTs stabilizing, improving, or even normalizing after immunosuppressive therapy had been initiated. In most cases improvement in ILD seemed to parallel improved muscle performance. This may not be the case in some patients. In particular, a subset of patients with DM with minor or no muscle involvement seem to have a rapidly progressive ILD. Finally, although the number of patients in our study was low, our data indicate that anti–Jo-1 autoantibodies may not only be predictive of ILD, but could also have a positive prognostic value.

AUTHOR CONTRIBUTIONS

Dr. Fathi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Fathi, Tylen, Jorfeldt, Tornling, Lundberg.

Acquisition of data. Fathi, Tylen, Jorfeldt, Lundberg.

Analysis and interpretation of data. Fathi, Vikgren, Boijsen, Tylen, Jorfeldt, Tornling, Lundberg.

Manuscript preparation. Fathi, Vikgren, Tylen, Jorfeldt, Tornling, Lundberg.

Statistical analysis. Fathi, Tornling.

Ancillary