Table 2 Log-fold changes in cytokine levels by disease group compared to control.

From: Circulating cytokine levels in systemic sclerosis related interstitial lung disease and idiopathic pulmonary fibrosis

Cytokine

Disease

Log-fold Estimate

2.5 Percentile

97.5 Percentile

Holm’s p-value

Post Hoc Power

Effect Size \({{\varvec{\eta}}}_{{\varvec{p}}}^{2}\) 95% CI

ADAMTS13

IPF

0.090

− 0.237

0.418

0.0363

0.9998

0.22

dcSSc ILD

− 0.026

− 0.536

0.014

  

(0.09, 0.63)

dcSSc NoILD

− 0.068

− 0.360

0.224

   

lcSSc ILD

− 0.400

− 0.124

0.150

   

lcSSc NoILD

− 0.383

− 0.662

− 0.104

   

Eotaxin-1

IPF

0.753

0.185

1.321

0.0011

0.9999

0.28

dcSSc ILD

1.088

0.610

1.566

  

(0.13, 0.70)

dcSSc NoILD

0.829

0.320

1.338

   

lcSSc ILD

0.805

0.330

1.281

   

lcSSc NoILD

0.617

0.133

1.101

   

IL-6

IPF

1.879

0.398

3.359

0.0072

0.9998

0.25

dcSSc ILD

2.259

1.013

3.505

  

(0.11, 0.67)

dcSSc NoILD

0.937

− 0.390

2.263

   

lcSSc ILD

1.977

0.738

3.217

   

lcSSc NoILD

0.349

− 0.913

1.610

   

MIG/CXCL9

IPF

0.468

− 0.298

1.233

0.0368

0.9838

0.22

dcSSc ILD

0.524

− 0.120

1.168

  

(0.09, 0.63)

dcSSc NoILD

0.264

− 0.421

0.950

   

lcSSc ILD

1.056

0.415

1.696

   

lcSSc NoILD

0.707

0.055

1.359

   
  1. Estimates, confidence limits, Holm’s p-value, and effect size with bootstrapped 95% confidence interval (CI) are from a multiple regression linear model using R software of the base-2 logarithm of each cytokine level from peripheral blood for differences among categories with: continuous covariates of age and disease duration; categorical covariates of ever smoking, sex, antifibrotic medication (nintedanib or pirfenidone), anti-autoimmune medication (azathioprine, cyclophosphamide, mycophenolate mofetil, or rituximab); and interaction between age and sex. The post hoc power for a test at the 5% level of significance was estimated using PASS software in an analysis-of-covariance model with 7 covariates with group log-folds, proportion of variance explained by the multiple regression model, and residual error as estimated by R software.
Back to article page