Muscle Biopsy Findings in Combination With Myositis-Specific Autoantibodies Aid Prediction of Outcomes in Juvenile Dermatomyositis

Abstract

Objective: Juvenile dermatomyositis (DM) is a rare and severe autoimmune condition characterized by rash and proximal muscle weakness. While some patients respond to standard treatment, others do not. This study was carried out to investigate whether histopathologic findings and myositis-specific autoantibodies (MSAs) have prognostic significance in juvenile DM.

Methods: Muscle biopsy samples (n = 101) from patients in the UK Juvenile Dermatomyositis Cohort and Biomarker Study were stained, analyzed, and scored for severity of histopathologic features. In addition, autoantibodies were measured in the serum or plasma of patients (n = 90) and longitudinal clinical data were collected (median duration of follow-up 4.9 years). Long-term treatment status (on or off medication over time) was modeled using generalized estimating equations.

Results: Muscle biopsy scores differed according to MSA subgroup. When the effects of MSA subgroup were accounted for, increased severity of muscle histopathologic features was predictive of an increased risk of remaining on treatment over time: for the global pathology score (histopathologist's visual analog scale [hVAS] score), 1.48-fold higher odds (95% confidence interval [95% CI] 1.12-1.96; P = 0.0058), and for the total biopsy score (determined with the standardized score tool), 1.10-fold higher odds (95% CI 1.01-1.21; P = 0.038). A protective effect was identified in patients with anti-Mi-2 autoantibodies, in whom the odds of remaining on treatment were 7.06-fold lower (95% CI 1.41-35.36; P = 0.018) despite muscle biopsy scores indicating more severe disease. In patients with anti-nuclear matrix protein 2 autoantibodies, anti-transcription intermediary factor 1γ autoantibodies, or no detectable autoantibody, increased histopathologic severity alone, without adjustment for the effect of MSA subtype, was predictive of the risk of remaining on treatment: for the hVAS global pathology score, 1.61-fold higher odds (95% CI 1.16-2.22; P = 0.004), and for the total biopsy score, 1.13-fold higher odds (95% CI 1.03-1.24; P = 0.013).

Conclusion: Histopathologic severity, in combination with MSA subtype, is predictive of the risk of remaining on treatment in patients with juvenile DM and may be useful for discussing probable treatment length with parents and patients. Understanding these associations may identify patients at greater risk of severe disease.

Figure 1

Distributions and correlations of total biopsy scores and histopathologist's visual analog scale (hVAS) global pathology scores in patients with juvenile dermatomyositis. A and B, The distribution of total biopsy scores (A) and hVAS scores (B) was determined across subgroups of patients with myositis‐specific antibodies (MSAs) or no detectable autoantibody (nil) (n = 69). Factorial analysis of variance using the Kruskal‐Wallis test was performed to analyze the distribution of these scores. There was a significant main effect of MSA subtype on the hVAS score (χ² [df4] = 20.0, P = 0.0005; n = 69): for anti–melanoma differentiation–associated gene 5 (anti‐MDA5) vs. anti–Mi‐2, P = 0.0001; for anti‐MDA5 vs. anti–nuclear matrix protein 2 (anti–NXP‐2), P = 0.007; for anti‐MDA5 vs. anti–transcription intermediary factor 1γ (anti–TIF‐1γ), P = 0.04; for anti‐MDA5 vs. no detectable autoantibody, P = 0.03. There was also a significant main effect of MSA subtype on the total biopsy score (χ² [df4] = 20.4, P = 0.0004; n = 69): for anti‐MDA5 vs. anti–Mi‐2, P = 0.0009; for anti‐MDA5 vs. anti–NXP‐2, P = 0.0006; for anti‐MDA5 vs. anti–TIF‐1γ, P = 0.01; for anti‐MDA5 vs. no detectable autoantibody, P = 0.04. Symbols indicate individual patients; bars show the median. C, Correlation of total biopsy scores and hVAS scores was determined by Spearman's rank correlation analysis (n = 101) (expressed as R values with 95% confidence intervals).

Figure 2

Longitudinal generalized estimating equations (GEE) modeling of treatment status over time according to MSA subgroups and hVAS global muscle pathology scores or total biopsy scores. Forest plots depict odds ratios with 95% confidence intervals for being on treatment, estimated using GEE models fitted with MSA subgroups and either hVAS scores (A) or total biopsy scores (B) as predictors. The no detectable autoantibody group (nil) was used as the reference category. See Figure 1 for other definitions.

Figure 3

Longitudinal generalized estimating equations (GEE) models of the association between muscle biopsy scores and long‐term treatment status (on or off medication over time) in patients with anti–NXP‐2 autoantibodies, patients with anti–TIF‐1γ autoantibodies, and patients with no detectable autoantibody. A and B, Forest plots depict odds ratios with 95% confidence intervals for being on treatment, estimated using GEE models fitted with either the hVAS scores (A) or the total biopsy scores (B) as predictors. C and D, The predicted probability of being off treatment at 5 years postdiagnosis is plotted as a function of either the hVAS scores (C) or the total biopsy scores (D), derived from the GEE models. Dotted lines represent the 95% confidence intervals. The median values for the time from onset to diagnosis (median 0.214 years) and for the time from diagnosis to biopsy (median 0.0602 years) were used in the calculations of predicted probabilities. See Figure 1 for other definitions.