Serum survivin predicts responses to treatment in active rheumatoid arthritis: a post hoc analysis from the SWEFOT trial

Abstract

Background: The identification of biomarkers that predict optimal and individual choices of treatment for patients with rheumatoid arthritis gains increasing attention. The purpose of this study was to investigate if the proto-oncogene survivin might aid in treatment decisions in early rheumatoid arthritis.

Methods: Serum survivin levels were measured in 302 patients who completed the Swedish pharmacotherapy (SWEFOT) trial at baseline, 3, 12, and 24 months. Survivin levels > 0.45 ng/mL were considered positive. Based on the survivin status, core set outcomes measuring disease activity, functional disability, as well as global health and pain were evaluated after methotrexate (MTX) monotherapy at 3 months, and at 12 and 24 months of follow-up. Treatment of non-responders was randomly intensified with either a combination of disease-modifying antirheumatic drugs (triple therapy: MTX, sulfasalazine, and hydroxychloroquine) or by adding antibodies against tumor necrosis factor (anti-TNF).

Results: Antirheumatic treatment resulted in an overall decrease of serum survivin levels. Survivin-positive patients at baseline who initially responded to MTX had a higher risk of disease re-activation (OR 3.21 (95% CI 1.12-9.24), P = 0.032) and failed to improve in their functional disability (P = 0.018) if having continued on MTX monotherapy compared to survivin-negative patients. Ever-smokers who were survivin-positive were less likely to respond to MTX than those who were survivin-negative (OR 1.91 (1.01-3.62), P = 0.045). In survivin-positive patients, triple therapy led to better improvements in disease activity than did MTX + anti-TNF. At 24 months, survivin-positive patients randomized to anti-TNF had a higher risk of active disease than those randomized to triple therapy (OR 3.15 (1.09-9.10), P = 0.037).

Discussion: We demonstrate for the first time that survivin is a valuable serologic marker that can distinguish drug-specific clinical responses in early rheumatoid arthritis through the pragmatic clinical setting of the care-based SWEFOT trial. Although treatment response cannot solely be attributable to survivin status, per protocol sensitivity analyses confirmed the superior effect of triple therapy on survivin-positive patients.

Conclusions: Survivin-positive patients have poor outcomes if treated with MTX monotherapy. A decrease of survivin levels during treatment is associated with better clinical responses. For survivin-positive patients who fail MTX, triple therapy is associated with better outcomes than anti-TNF therapy.

Trial registration: WHO database at the Karolinska University Hospital: CT20080004 ; ClinicalTrials.gov: NCT00764725, registered 1 October 2008.

Figures

Fig. 1

Analysis profile. Serum levels of survivin were measured from samples of 302 patients who completed the 24-month follow-up of the SWEFOT trial. Measurements at baseline resulted in the survivin-positive (> 0.45 ng/mL) and survivin-negative patient groups. Measurements at 3, 12, and 24 months identified patients who decreased or increased their serum survivin levels at any time point over 24 months. Four groups of patients were formed and compared: patients positive for survivin (PP), or negative for survivin (NN) on all testing occasions; and patients survivin-positive at baseline who converted negative (PN), or survivin-negative patients who converted positive (NP). After 3 months of methotrexate (MTX) monotherapy, patients still with active disease were randomized to triple therapy (MTX + sulfasalazine (SSZ) + hydroxychloroquine (HCQ)) or to anti-TNF (MTX + infliximab). *Patients with missing samples at 12 and 24 months (n = 15), who were not randomized (n = 2), or who changed their survivin status on several occasions (n = 15) were excluded from analysis

Fig. 2

Changes of serum survivin levels during the SWEFOT trial. Serum levels of survivin were measured in 302 patients enrolled in the trial at baseline, where 114 patients were survivin-positive (survivin > 0.45 ng/mL, dashed line), and the remaining 188 patients were survivin-negative. In total, a decrease of survivin levels was observed at 3, 12, and 24 months of antirheumatic treatment. The number of samples available for the analysis is indicated at each time point. Bolded lines indicate median levels of the survivin-positive at baseline group (open circles) and the survivin-negative converting to positive group at any time point over 24 months (filled rhombi). A total of 52 (46 %) of the survivin-positive patients converted to negative over 24 months. Comparison between the time points was done by the Wilcoxon signed-rank test, and the P values for the total patient cohort are indicated

Fig. 3

Core set clinical parameters in the survivin subgroups of the SWEFOT trial. At baseline, all patients were treated with methotrexate (MTX). a At 3 months, the patients with a disease activity score (DAS28) < 3.2 proceeded on MTX monotherapy. b The patients with a DAS28 > 3.2 were randomized to triple therapy with MTX + sulfasalazine (SSZ) + hydroxychloroquine (HCQ), c or to anti-TNF therapy (MTX + infliximab). The intention-to-treat analysis was conducted for clinical parameters, including the DAS28, functional disability (Health Assessment Questionnaire, HAQ), pain perception with the visual analog scale (pain-VAS), and the patient’s global assessment of disease activity (PtGA-VAS). Boxes represent the 25th to 75th percentile of the group, and horizontal lines within the boxes indicate median values. Four groups of patients were compared: patients positive for survivin (PP), and negative for survivin (NN) at baseline and over 24 months; patients positive for survivin at baseline who converted negative (PN), and negative for survivin who converted positive (NP) over 24 months. Comparisons of the absolute values were done by Mann–Whitney U tests

Fig. 4

Prevalence of active disease among survivin-positive or survivin-negative patients in the SWEFOT trial. The prevalence of active disease (disease activity score, DAS28 > 3.2) among patients who were survivin-positive or survivin-negative at baseline is presented for the groups treated with methotrexate (MTX) monotherapy and the groups randomized to triple therapy (TT) or to anti-TNF therapy. Comparisons were done by Pearson’s χ2 or Fisher’s exact tests, and odds-based estimates (odds ratio, OR) and 95 % confidence intervals (CI) are indicated. Two patients with no available DAS28 at 3 months were excluded from the analysis