Fig. 3.

Predictive performance of serum IGF1 levels for development of cardiovascular events in rheumatoid arthritis. a The table shows individual baseline characteristics of 12 patients, who developed CV events during the 5-year follow-up. Asterisk indicates the case of death. b The receiver operator characteristic (ROC) curves depict associations between CV events developed during the 5-year follow-up and age (red line) and absolute serum levels of IGF1 (blue line) at baseline in 184 female RA patients. AUC, area under the ROC curve. c The Kaplan-Meier curves show the development of new CV events in IGF1^hi (n = 88, red line) or IGF1^low (n = 96, blue line) groups. d The column diagrams show the frequency of treatment at baseline (BL, open columns) and at the 5-year follow-up (5Y, filled columns) in IGF1^low and IGF1^hi groups stratified by age, < 50 years (n = 68), 50–60 years (n = 53), > 60 years (n = 63). The p values are obtained by chi-square statistics. eCVR, CV risk estimated by the Framingham algorithm; DD, disease duration; IS, ischemic stroke; TIA, transitory ischemic attack; DVT, deep venous thrombosis; LE, lung embolism; MI, myocardial infarction; MTX, methotrexate, RTX, rituximab; ETC, etanercept; SLZ, sulfasalazine; AZA, azathioprine; TOCI, tocilizumab; OR, odds ratio; CI, confidence interval; T2D, type 2 diabetes