Serum type I interferon score as a disease activity biomarker in patients with diffuse cutaneous systemic sclerosis: a retrospective cohort study.

backgroundType I interferon (IFN) pathway activation has been associated with severe systemic sclerosis. We aimed to examine the association of serum IFN scores with disease activity and outcomes in two cohorts of patients with diffuse cutaneous systemic sclerosis.

methodsIn this retrospective cohort study, we included adult (aged >18 years) patients with diffuse cutaneous systemic sclerosis enrolled in the US Prospective Registry of Early Systemic Sclerosis (PRESS; incident cohort) or in the UK observational cohort (Stratification for Risk of Progression in Scleroderma [STRIKE]; prevalent cohort) registries and healthy controls (volunteers). Sera were analysed by Myriad-Rules Based Medicine's Luminex xMAP Technology Multiplex Assay (Austin, TX, USA), and IFN scores were generated using concentrations of CCL2, CCL8, CCL19, CXCL9, CXCL10, and CXCL11. Patients were classified as IFN-high (vs IFN-low) when mean sum of the natural logarithm of the six chemokines was greater than (or within) two standard deviations of healthy controls. The main outcome measures were the baseline and the minimal clinically important differences for modified Rodnan Skin Score, forced vital capacity (FVC), diffusing capacity for carbon monoxide (DLCO), and Health Assessment Questionnaire-Disability Index at 12 months. A person with lived experience of systemic sclerosis was involved in writing the report.

findingsPatients with diffuse cutaneous systemic sclerosis in the PRESS incident cohort were recruited between April 1, 2012, and Jan 1, 2019, and healthy controls and patients in the STRIKE prevalent cohort were recruited between Dec 1, 2014, and Dec 1, 2018. IFN scores were generated for 110 patients in the incident cohort (mean age 50·2 years [SD 15·0], 76 [69%] women and 34 [31%] men, 87 [79%] White) and 72 patients in the prevalent cohort (mean age 51·7 years [SD 10·9], 50 [69%] women and 22 [31%] men, 64 [89%] White), and 32 healthy controls (mean age 47·0 years [SD 12·4]; 19 [59%] women and 13 [41%] men; 24 [75%] White). 50 (45%) of 110 patients in the incident cohort and 27 (38%) of 72 patients in the prevalent cohort were classified as IFN-high. In the incident cohort, patients classified as IFN-high had worse baseline disease compared with patients classified as IFN-low, as assessed by mean predicted FVC (72·0% [SD 18·9] vs 85·3% [18·5]; p=0·0028), DLCO (56·8% [SD 21·6] vs 76·6% [25·3]; p=0·0008), and median Health Assessment Questionnaire-Disability Index (1·4 [IQR 0·8-2·0] vs 0·8 [0·4-1·5]; p=0·0033). Differences in FVC and DLCO persisted at last follow-up (median 34 months [IQR 19·8- 54·0] for FVC and median 34 months [IQR 22·5- 54·0] for DLCO). In the prevalent cohort, patients classified as IFN-high had a shorter median disease duration (2·2 years [IQR 0·7-8·2] vs 5·0 years [1·9-10·0]; p=0·035) compared to those classified as IFN-low, and worse 12-month lung outcomes independent of baseline FVC or immunosuppression (5% relative worsening of FVC in nine [39%] of 23 patients with IFN-high vs seven [17%] of 41 patients with IFN-low; p=0·051). Moreover, cumulative 5-year mortality was 24·9% (95% CI 14·9-39·7) for IFN-high versus 8·6% (3·6-19·9) for IFN-low (p=0·052).

interpretationSerum IFN score assessment for patients with diffuse cutaneous systemic sclerosis could identify patients with high disease activity who are more likely to have worse 12-month prognosis and overall survival.

fundingNational Scleroderma Foundation, National Institutes of Health National Institute of Arthritis and Musculoskeletal and Skin Diseases Rheumatic Disease Research Core Centers, National Institute of Health Research Leeds Biomedical Research Centre, and Kennedy Trust for Rheumatology Research.