Abstract
Purpose:
Myositis-associated interstitial lung disease (MA-ILD) is associated with increased mortality, but no prognostic model exists in this population. The ILD-GAP index was developed to predict mortality risk across all subtypes of chronic ILD. The purpose of this study was to validate the ILD-GAP risk prediction model in patients with MA-ILD.
Procedures:
We completed a retrospective cross-sectional study of patients enrolled in the Johns Hopkins Myositis Center database between 2006 and 2017. Cumulative mortality rates were estimated using the Kaplan-Meier test. Model calibration was determined by using standardized mortality ratios of observed versus expected deaths.
Main Findings:
179 participants with MA-ILD were included. The mean baseline percent predicted forced vital capacity was 65.2 ± 20.6%, forced expiratory volume in the first second 65.4 ± 20.4%, and carbon monoxide diffusing capacity 61.6 ± 20.0%. Thirty-two participants died (17.9%). The ILD-GAP model had poor discriminative performance and calibration.
Conclusions:
The ILD-GAP risk prediction model is a poor predictor of mortality among individuals with MA-ILD. The identification of a better predictive model for MA-ILD is needed to help guide care in this patient population.
Keywords: myositis, lung diseases, interstitial, calibration, cross-sectional studies
INTRODUCTION
Myositis is a chronic heterogeneous group of connective tissue disorders associated with dermatologic, cardiovascular and pulmonary morbidity.1 The prevalence of myositis-associated interstitial lung disease (MA-ILD) is high,2 and has been associated with increased mortality in this population.3–6 Despite the known increased mortality among individuals with MA-ILD, risk prediction is challenging given the heterogeneity of disease presentation and progression. The GAP risk prediction model was initially developed for patients with idiopathic pulmonary fibrosis (IPF) as a prognostic guide.7 The ILD-GAP index was created to better predict mortality across all chronic ILD subtypes, including those with connective tissue disease (CTD).8 This index added a disease subtype variable to the GAP model and demonstrated good discrimination in a group of individuals with chronic ILD.8 However, neither the GAP model nor ILD-GAP index have been validated in individuals with MA-ILD. The purpose of this study was to validate the ILD-GAP risk prediction model in this population.
METHODS
We performed a retrospective cross-sectional study of consecutive patients enrolled in the Johns Hopkins Myositis Center database between January 2006 and July 2017. Autoimmune myositis cases included polymyositis (PM), dermatomyositis (DM), and clinically amyopathic dermatomyositis (CADM). Individuals were classified has having clinically evident MA-ILD if they met American Thoracic Society criteria for restriction or diffusing capacity deficits9 and had evidence of diffuse parenchymal lung disease on high-resolution computed tomography. All participants provided written informed consent; the present study was approved by the Johns Hopkins University IRB (NA_00007454).
Derivation and validation of the ILD-GAP index has been described previously.8 The predictor variables considered in this model include gender, age, lung physiology variables (forced vital capacity [FVC] and carbon monoxide diffusing capacity [DLCO]), and ILD subtype (IPF, CTD-associated ILD/idiopathic nonspecific interstitial pneumonia, chronic hypersensitivity pneumonitis, or unclassifiable ILD). The ILD-GAP index is calculated by combining points assigned to the variables listed above to obtain a total point score. The total point score is used to classify patients to stages, which correlate with predicted mortality.
All-cause mortality was determined and confirmed through the Social Security Death Index. Lung transplantation was treated as a competing risk. Follow-up time was defined as the date of diagnosis with autoimmune myositis to the date of death or lung transplantation or the last follow-up visit. Cumulative mortality rates were estimated using the Kaplan-Meier test. Model calibration was determined by using standardized mortality ratios of observed versus expected deaths. All calculations were performed using intercooled Stata 13 (StataCorp, College Station, TX).
RESULTS
This study included 179 individuals with MA-ILD. Most participants were women (68.2%) with a mean age of 57.8 ± 12.2 years, and 36.9% of participants were African-American. Ninety-five (53%) had DM, 79 (44%) PM, and 5 (3%) CADM. The mean body mass index was 29.1 ± 6.9 kg/m2. The mean percent predicted FVC was 65.2 ± 20.6%, forced expiratory volume in the first second (FEV1) 65.4 ± 20.4%, total lung capacity (TLC) 67.1 ± 18.2%, and DLCO 61.6 ± 20.0%. One hundred sixty-eight participants (93.9%) had autoantibodies with anti-Jo-1 being the most common. The median follow-up time was 9 years (range 0-24 years). There were 32 deaths; one participant underwent lung transplantation.
The ILD-GAP model had poor discriminative performance and calibration in this patient population (Table 1). The ILD-GAP model predicted higher mortality rates than that observed in this population at 1, 2 and 3-years.
Table 1.
One, two and three-year survival probabilities, by ILD-GAP Index
| ILD-GAP Index | 1-year survival | 2-year survival | 3-year survival | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| # at Risk | # of Events | Observed (95% CI) | Predicted | # at Risk | # of Events | Observed (95% CI) | Predicted | # at Risk | # of Events | Observed (95% CI) | Predicted | |
| 0-1 | 128 | 1 | 99% (94.6-99.9) | 96.9% | 126 | 2 | 98% (92.9-99.2) | 93.4% | 124 | 1 | 97% (91.9-98.8) | 89.8% |
| 2-3 | 39 | 0 | 100% | 91.2% | 39 | 0 | 100% | 82% | 39 | 1 | 97% (83.2-99.6) | 73.1% |
| 4-5 | 7 | 0 | 100% | 81.8% | 7 | 1 | 86% (33.4-97.9) | 65% | 7 | 1 | 86% (33.4-97.9) | 50.8% |
Observed = Observed survival in the current MA-ILD cohort
Predicted= Predicted survival based on the original ILD-GAP cohort derivation
DISCUSSION
This study of a well-characterized, diverse cohort of participants with MA-ILD showed poor discriminative performance of the ILD-GAP risk prediction model. There are several potential explanations for our findings. The ILD-GAP model was designed to assess three-year risk. In our study population, only 20% of events (death or lung transplantation) occurred within three years of diagnosis, with most occurring after five years. This may be related to differences in overall survival and or response to therapy in this patient population compared to patients with other forms of CTD-ILD. Staging systems derived in this population would need to consider mortality over a longer time period to accurately predict risk. Our study included a relatively large proportion of African-Americans. The racial/ethnic breakdown of the populations of the GAP model and ILD-GAP index are not publically available, but attempts to validate the GAP model among non-Western IPF patients have produced mixed results.10–12 Race may play a significant role in prognosis among individuals with ILD, specifically MA-ILD, and may be an important variable in future derivation scores. Finally, it may be inappropriate to so heavily weigh lung physiology variables in models designed to assess disease severity in patients with MA-ILD. Autoantibodies, including anti-Jo1, are known to correlate with increased risk of ILD among patients with myositis and may also correlate with disease activity.5,13 Creatinine kinase and aldolase, which correlate with myositis activity, may also be markers of ILD severity early in the course of disease.13,14 Inclusion of these biomarkers may increase the validity of a prognostic model for MA-ILD.
We were unable to develop a potential prognostic model due the relatively small number of events compared to potential predictor variables. A longitudinal collaborative multicenter study is needed to generate a risk prediction model in patients with MA-ILD.
CONCLUSIONS
In conclusion, The ILD-GAP risk prediction model is a poor predictor of mortality risk among individuals with MA-ILD. The identification of a better predictive model for MA-ILD is needed to help guide care in this patient population.
Neither the GAP model nor ILD-GAP index have been validated in individuals with myositis-associated interstitial lung disease (MA-ILD).
The ILD-GAP risk prediction model is a poor predictor of mortality among individuals with MA-ILD
The identification of a better predictive model for MA-ILD is needed to help guide care in this patient population.
ACKNOWLEDGEMENTS
ROLE OF FUNDING SOURCE
This research was supported by the Huayi and Siuling Zhang Discovery Fund, and the NIH grant K01HL135459.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
COMPETING INTERESTS
None
REFERENCES
- 1.Saketkoo LA, Ascherman DP, Cottin V, Christopher-Stine L, Danoff S K, Oddis CV. Interstitial Lung Disease in Idiopathic Inflammatory Myopathy. Curr Rheumatol Rev 2010;6(2):108–119. doi: 10.2174/157339710791330740 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fathi M, Lundberg IE. Interstitial lung disease in polymyositis and dermatomyositis. Curr Opin Rheumatol 2005;17(6):701–706. doi: 10.1097/01.bor.0000179949.65895.53 [DOI] [PubMed] [Google Scholar]
- 3.Woo JH, Kim YJ, Kim JJ, et al. Mortality factors in idiopathic inflammatory myopathy: Focusing on malignancy and interstitial lung disease. Mod Rheumatol 2013;23(3):503–508. doi: 10.1007/s10165-012-0673-2 [DOI] [PubMed] [Google Scholar]
- 4.Marie I Interstitial lung disease in polymyositis and dermatomyositis. Arthritis Rheum 2002;47(6):614 http://metalib.lib.ic.ac.uk:9003/sfx_local?sid=OVID%5Cnisbn=%5Cnissn=0004-3591%5Cnvolume=47%5Cnissue=6%5Cndate=2002%5Cntitle=Arthritis%26Rheumatism%5Cnatitle=Interstitiallungdiseaseinpolymyositisanddermatomyositis.%5Cnaulast=MarieI%5Cnspage=614. [DOI] [PubMed] [Google Scholar]
- 5.Mimori T, Nakashima R, Hosono Y. Interstitial lung disease in myositis: Clinical subsets, biomarkers, and treatment. Curr Rheumatol Rep 2012;14(3):264–274. doi: 10.1007/s11926-012-0246-6 [DOI] [PubMed] [Google Scholar]
- 6.Johnson C, Pinal-Fernandez I, Parikh R, et al. Assessment of Mortality in Autoimmune Myositis With and Without Associated Interstitial Lung Disease. Lung. 2016;194(5):733–737. doi: 10.1007/s00408-016-9896-x [DOI] [PubMed] [Google Scholar]
- 7.Ley B, Ryerson CJ, Vittinghoff E, et al. A multidimensional index and staging system for idiopathic pulmonary fibrosis. Ann Intern Med 2012;156(10):684–695. doi: 10.7326/0003-4819-156-10-201205150-00004 [DOI] [PubMed] [Google Scholar]
- 8.Ryerson CJ, Vittinghoff E, Ley B, et al. Predicting Survival Across Chronic Interstitial Lung Disease. Chest. 2014;145(4):723–728. doi: 10.1378/chest.13-1474 [DOI] [PubMed] [Google Scholar]
- 9.American Thoracic Society. American Thoracic Society / European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias. Am J Respir Crit Care Med 2002;165(2):277–304. doi: 10.1183/09031936.02.00492002 [DOI] [PubMed] [Google Scholar]
- 10.Kim ES, Choi SM, Lee J, et al. Validation of the GAP score in Korean patients with idiopathic pulmonary fibrosis. Chest. 2015;147(2). doi: 10.1378/chest.14-0453 [DOI] [PubMed] [Google Scholar]
- 11.Kondoh S, Chiba H, Nishikiori H, et al. Validation of the Japanese disease severity classification and the GAP model in Japanese patients with idiopathic pulmonary fibrosis. Respir Investig 2016;54(5):327–333. doi: 10.1016/j.resinv.2016.02.009 [DOI] [PubMed] [Google Scholar]
- 12.Lee SH, Kim SY, Kim DS, et al. Predicting survival of patients with idiopathic pulmonary fibrosis using GAP score: A nationwide cohort study. Respir Res 2016;17(1). doi: 10.1186/s12931-016-0454-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Solomon J, Swigris JJ, Brown KK. Myositis-related interstitial lung disease and antisynthetase syndrome. J Bras Pneumol 2011;37(1):100–109. doi: 10.1016/j.biotechadv.2011.08.021.Secreted [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bohan A, Peter J. Bohan A, Peter J Polymyositis and dermatomyositis. N Engl J Med 1975;292:344–347. doi: 10.1056/NEJM197502202920807 [DOI] [PubMed] [Google Scholar]