Biomarker patterns across skin-lung-gut burden in systemic sclerosis.
Systemic sclerosis (SSc) is a multisystem autoimmune disease in which skin, lung, and gastrointestinal involvement often co-occur. Whether increasing organ co-involvement burden is accompanied by coherent circulating biomarker changes remains unclear.
To characterize immune, inflammatory, and nutritional biomarker patterns across skin-lung-gut burden groups in SSc.
We conducted a retrospective cross-sectional study of 314 patients with SSc fulfilling the 2013 ACR/EULAR classification criteria. Three prespecified axes were used to define organ burden: skin axis, modified Rodnan skin score (mRSS) ≥14; lung axis, interstitial lung disease (ILD); and gut axis, broad gastrointestinal involvement (bGI). Patients were classified into four mutually exclusive groups according to the number of involved axes: 0-Axis, 1-Axis, 2-Axis, and 3-Axis. Clinical characteristics and biomarker distributions were compared across groups. Multinomial logistic regression was performed using 0-Axis as the reference group, with prespecified sensitivity analyses including ordinal and treatment-adjusted models, subtype-informed analyses, and axis-specific adjusted analyses. Unsupervised clustering was used as an exploratory complementary analysis.
Among 314 patients, 37 (11.8%) were classified as 0-Axis, 136 (43.3%) as 1-Axis, 118 (37.6%) as 2-Axis, and 23 (7.3%) as 3-Axis. Anti-Scl-70 positivity increased across burden groups, from 51.4% in the 0-Axis group to 71.9% in the 1-Axis group, 83.1% in the 2-Axis group, and 91.3% in the 3-Axis group (p<0.001). hs-CRP differed across groups and was higher in the 2-Axis group than in the 1-Axis group (9.60 [3.13-19.40] vs 4.50 [1.79-12.03] mg/L, p=0.012 overall). Albumin decreased across groups and was lower in the 3-Axis group than in the 0-Axis group (31.70 [27.10-33.60] vs 34.30 [31.80-37.50] g/L, p=0.005 overall). In adjusted multinomial joint models, Anti-Scl-70 positivity was associated with 2-Axis versus 0-Axis (OR 3.86, 95% CI 1.46-10.22, p=0.007) and 3-Axis versus 0-Axis (OR 12.78, 95% CI 1.86-87.91, p=0.010). Lower albumin was associated with 2-Axis versus 0-Axis (OR 0.52, 95% CI 0.31-0.89, p=0.017) and 3-Axis versus 0-Axis (OR 0.38, 95% CI 0.18-0.79, p=0.009). Higher total cholesterol was associated with 3-Axis versus 0-Axis (OR 2.81, 95% CI 1.11-7.13, p=0.030). hs-CRP showed an inverse association for 1-Axis versus 0-Axis (OR 0.60, 95% CI 0.37-0.98, p=0.039), but was not associated with 2-Axis versus 0-Axis (OR 0.66, 95% CI 0.42-1.04, p=0.074) or 3-Axis versus 0-Axis (OR 1.04, 95% CI 0.60-1.81, p=0.890) after joint adjustment. Unsupervised clustering identified three biomarker-defined clusters with differing centroid profiles and differing axis-burden distributions. Supplementary analyses supported the robustness of the principal Anti-Scl-70 and albumin signals while clarifying the contributions of subtype structure, lung-axis predominance, treatment exposure, and GI heterogeneity.
Increasing skin-lung-gut co-involvement burden in SSc was associated with structured circulating biomarker patterning. The most consistent adjusted signals were Anti-Scl-70 positivity and lower albumin in higher burden groups. These findings support a clinically anchored, biomarker-informed perspective on multisystem involvement in SSc, while remaining complementary to conventional subtype-based classification and requiring cautious interpretation.
To characterize immune, inflammatory, and nutritional biomarker patterns across skin-lung-gut burden groups in SSc.
We conducted a retrospective cross-sectional study of 314 patients with SSc fulfilling the 2013 ACR/EULAR classification criteria. Three prespecified axes were used to define organ burden: skin axis, modified Rodnan skin score (mRSS) ≥14; lung axis, interstitial lung disease (ILD); and gut axis, broad gastrointestinal involvement (bGI). Patients were classified into four mutually exclusive groups according to the number of involved axes: 0-Axis, 1-Axis, 2-Axis, and 3-Axis. Clinical characteristics and biomarker distributions were compared across groups. Multinomial logistic regression was performed using 0-Axis as the reference group, with prespecified sensitivity analyses including ordinal and treatment-adjusted models, subtype-informed analyses, and axis-specific adjusted analyses. Unsupervised clustering was used as an exploratory complementary analysis.
Among 314 patients, 37 (11.8%) were classified as 0-Axis, 136 (43.3%) as 1-Axis, 118 (37.6%) as 2-Axis, and 23 (7.3%) as 3-Axis. Anti-Scl-70 positivity increased across burden groups, from 51.4% in the 0-Axis group to 71.9% in the 1-Axis group, 83.1% in the 2-Axis group, and 91.3% in the 3-Axis group (p<0.001). hs-CRP differed across groups and was higher in the 2-Axis group than in the 1-Axis group (9.60 [3.13-19.40] vs 4.50 [1.79-12.03] mg/L, p=0.012 overall). Albumin decreased across groups and was lower in the 3-Axis group than in the 0-Axis group (31.70 [27.10-33.60] vs 34.30 [31.80-37.50] g/L, p=0.005 overall). In adjusted multinomial joint models, Anti-Scl-70 positivity was associated with 2-Axis versus 0-Axis (OR 3.86, 95% CI 1.46-10.22, p=0.007) and 3-Axis versus 0-Axis (OR 12.78, 95% CI 1.86-87.91, p=0.010). Lower albumin was associated with 2-Axis versus 0-Axis (OR 0.52, 95% CI 0.31-0.89, p=0.017) and 3-Axis versus 0-Axis (OR 0.38, 95% CI 0.18-0.79, p=0.009). Higher total cholesterol was associated with 3-Axis versus 0-Axis (OR 2.81, 95% CI 1.11-7.13, p=0.030). hs-CRP showed an inverse association for 1-Axis versus 0-Axis (OR 0.60, 95% CI 0.37-0.98, p=0.039), but was not associated with 2-Axis versus 0-Axis (OR 0.66, 95% CI 0.42-1.04, p=0.074) or 3-Axis versus 0-Axis (OR 1.04, 95% CI 0.60-1.81, p=0.890) after joint adjustment. Unsupervised clustering identified three biomarker-defined clusters with differing centroid profiles and differing axis-burden distributions. Supplementary analyses supported the robustness of the principal Anti-Scl-70 and albumin signals while clarifying the contributions of subtype structure, lung-axis predominance, treatment exposure, and GI heterogeneity.
Increasing skin-lung-gut co-involvement burden in SSc was associated with structured circulating biomarker patterning. The most consistent adjusted signals were Anti-Scl-70 positivity and lower albumin in higher burden groups. These findings support a clinically anchored, biomarker-informed perspective on multisystem involvement in SSc, while remaining complementary to conventional subtype-based classification and requiring cautious interpretation.