Pulmonary function tests, interstitial lung disease and lung function decline in outpatients with classic and clinically amyopathic dermatomyositis

Dear Editor, Interstitial lung disease (ILD) is common among patients with both classical dermatomyositis (DM) and clinically amyopathic dermatomyositis (CADM), patients with classical skin manifestations without weakness.1,2,3 ILD prognosis is thought to be worse in patients with CADM, but this observation may be due primarily to an increased frequency of rapidly progressive ILD.2,4,5,6,7 The objectives of this study were to compare the frequency of pulmonary function test (PFT) abnormalities and ILD in patients with DM and CADM, and to assess the frequency of worsening PFTs.

Charts of 151 patients in a prospective cohort of outpatients with DM presenting or referred to the dermatology department at the University of Pennsylvania were retrospectively reviewed through October 2014. This identified 128 patients with possible, probable or definite adult DM based on the Bohan and Peter criteria,8 or CADM as per Sontheimer,9 excluding juvenile DM or overlap syndromes.

PFTs were routinely ordered at the initial visit as part of standard practice, and repeat testing was typically yearly or more frequently as clinically indicated. Abnormal PFTs were defined by forced vital capacity (FVC), total lung capacity (TLC) or lung diffusion capacity for carbon monoxide (DLCO) < 80% of the predicted value. Significant decline in PFTs was defined as a 15% decline in absolute values of DLCO or 10% decline in FVC or TLC at the last available PFT.10 To identify ILD, computed tomography (CT) scans within 1 year of the first abnormal PFTs were blindly reviewed by an experienced radiologist (W.T.M.). We used t‐tests, Wilcoxon rank‐sum tests and Fisher's exact tests to compare patients with DM vs. CADM with abnormal PFTs, and we also compared patients with and without ILD.

PFTs were available in 69 of 76 (91%) patients with DM and 47 of 52 (90%) patients with CADM. Abnormal PFTs were present in 34 of 69 (49%) with DM and 20 of 47 (43%) with CADM (P = 0·48). The characteristics of patients with DM vs. CADM with abnormal PFTs were similar aside from creatine kinase levels (Table 1). DLCO, FVC, TLC and PFT patterns at baseline were also similar. The most common pattern was an isolated reduction in DLCO.

Table 1

Characteristics of patients with classic dermatomyositis (DM) vs. clinically amyopathic dermatomyositis (CADM) with abnormal pulmonary function tests

	Classic DM (n = 34)	CADM (n = 20)	P‐value
Age (years)	52·9 ± 11·4	58·2 ± 13·2	0·13
Female	25 (74)	18 (90)	0·18
White	29 (85)	18 (90)	1·0
Disease duration at baseline (years)	1·5 (0·6–7·2)	1·8 (0·9–4·0)	0·94
Malignancy within 5 years of diagnosis	6 (18)	2 (10)	0·70
History of rapidly progressive ILDa	2 (6)	2 (10)	0·62
Fever associated with DM	3 (9)	3 (15)	0·66
Raynaud phenomenon	5 (15)	4 (20)	0·71
Inflammatory arthritis	7 (21)	5 (25)	0·74
Mechanic's hands	9 (26)	10 (50)	0·14
Ulcerating gottron	2 (6)	0	0·27
Echocardiogram with pulmonary hypertension (PASP ≥ 35 mmHg)	0/20 (0)	3/12 (15)	0·04
Creatine kinase at baseline, U L−1	138 (70–269)	52 (29–97)	< 0·001*
Aldolase at baseline, U L−1	4·4 (3·1–5·9)	4·8 (4·4–6·3)	0·30
Haemoglobin at baseline, g dL−1	13·6 (12·2–14·2)	13·5 (12·6–14·1)	0·71
ANA ≥ 1 : 160	14/29 (41)	6/18 (30)	0·37
Anti‐Jo‐1 antibodies	4/21 (12)	2/13 (10)	1·0
Anti‐SSA antibodies	3/22 (9)	1/14 (5)	1·0
Anti‐RNP antibodies	1/10 (3)	0/9 (0)	1·0
Anti‐MDA5/CADM140	0/0 (0)	2/2 (10)	–
DLCO% predicted	74 (64–79)	67 (58–72)	0·08
FVC% predicted	81 (71–91)	88 (78–97)	0·18
TLC% predicted	86 (77–98)	87 (70–98)	0·60
ILD present on CT scan	9/23 (39)	6/16 (38)	0·92

Values are the mean ± SD compared with the Student t‐test with equal variances, median (interquartile range) compared with the Wilcoxon rank‐sum test for skewed variables, or n (%) compared with Fisher's exact test. Percentages represent column totals. ILD, interstitial lung disease; PASP, pulmonary artery systolic pressure; ANA, antinuclear antibodies; MDA5, melanoma differentiation‐associated gene 5; DLCO, lung diffusion capacity for carbon monoxide; FVC, forced vital capacity; TLC, total lung capacity; CT, computed tomography. *Statistically significant after Bonferroni correction for multiple comparisons. ^aNo patients had rapidly progressive ILD during the study period (rapidly worsening symptoms over 3 months requiring oxygen or hospitalization), but several had a history of this presentation.

Baseline CT was available for review in 39 of 54 (72%) patients with abnormal PFTs. ILD was present in nine of 23 (39%) with DM vs. six of 16 (38%) with CADM (P = 0·92). Patients with ILD had lower FVC, DLCO and TLC than subjects without ILD (Table 2). All PFT patterns were represented in patients with ILD, with six of 15 (40%) having restriction at baseline. Fourteen of 15 (93%) patients with ILD had an abnormal DLCO < 80% predicted. Lower DLCO cut‐offs had higher specificity for ILD, but with substantial loss in sensitivity. Five of 15 (33%) patients with ILD had neither cough nor dyspnoea documented at baseline.

Table 2

Association between pulmonary function tests (PFTs), pulmonary symptoms and interstitial lung disease (ILD) among patients with abnormal PFTs and chest computed tomography available for review

	No ILD (n = 24)	ILD (n = 15)	P‐valuea
PFT pattern
Isolated low DLCO	11 (46)	5 (44)	0·52
Restriction	6 (25)	6 (40)	0·48
Obstruction	4 (17)	0	0·15
Mixed	0	2 (13)	0·14
Indeterminate	3 (12)	2 (13)	1·0
DLCO% predicted	70 (64–78)	62 (54–70)	0·06
FVC% predicted	81 (76–96)	76 (68–86)	0·08
TLC% predicted	90 (74–99)	77 (72–84)	0·02
Dyspnoea	8 (33)	10 (67)	0·06
Cough	5 (21)	5 (33)	0·46
No dyspnoea or cough	14 (58)	5 (33)	0·19

Values are the median (interquartile range) compared with the Wilcoxon rank‐sum test, or n (%) compared with Fisher's exact test. DLCO, lung diffusion capacity for carbon monoxide; FVC, forced vital capacity; TLC, total lung capacity. ^aNo P‐values met statistical significance after Bonferroni correction for multiple comparisons.

At the last follow‐up, PFTs declined in four of 15 (27%) patients with ILD: two of six (33%) with CADM and two of seven (29%) with DM. Among patients with ILD, three of four (75%) with a decline in PFTs had a history of rapidly progressive ILD, compared with one of 11 (9%) without a decline in PFTs.

Abnormal PFTs are common in patients with DM, and ILD is present in a subset, with similar rates in classic DM and CADM. ILD can occur with any PFT pattern, including an isolated reduction in DLCO, and in asymptomatic patients.

More than half of patients with abnormal PFTs had no CT evidence of ILD. Only a small proportion had other potential explanations for abnormal PFTs. Pulmonary hypertension and anaemia were uncommon, and muscle disease was typically well controlled. The subsequent development of ILD in one patient without ILD at baseline highlights the importance of repeat imaging if PFTs decline or symptoms progress.

Overall, one‐quarter of patients with ILD had a decline in PFTs, and there was one death among 15 patients with ILD during several years of follow‐up. Most patients received aggressive immunosuppression. Notably, this relatively favourable prognosis applies to patients with chronic ILD followed in the outpatient setting. Patients with rapidly progressive lung disease would not have been captured in this outpatient cohort during their initial pulmonary presentation. In addition, patients with usual interstitial pneumonia, severe ILD or certain autoantibodies that may have been under‐represented may have poorer prognosis.

Larger prospective studies that include patients with CADM are needed to inform the prognosis of patients with chronic ILD. In previous studies, clinical deterioration was often dominated by the subset of patients with rapidly progressive ILD, with most deaths occurring within the first year.5  11 The number of patients with ILD in this study makes it too small to examine reliably any prognostic factors, but it may be informative for future studies. PFTs declined at similar rates in CADM and DM, and it may be that autoantibodies are more important prognostically for ILD than the phenotype of muscle disease.

In summary, abnormal PFTs are common in patients with DM and CADM. ILD may be present with any abnormality in PFTs even in the absence of symptoms, but abnormal PFTs should not be assumed to represent ILD. Further studies to identify prognostic factors for patients with chronic ILD are needed.

Author notes

Funding sources: M.D.G. is supported by the National Institutes of Health (NIH 5T32AR007442‐28). V.P.W. is supported by the Department of Veterans Affairs (Veterans Health Administration, Office of Research and Development, Biomedical Laboratory Research and Development) and by the National Institutes of Health (NIH R21 AR066286). The funders had no involvement in the study design, data collection, data analysis, manuscript preparation and/or publication decisions.

Conflicts of interest: none declared.