Risk Factors for Thromboembolic Events in Patients With Dialysis-Dependent CKD: Pooled Analysis of Phase 3 Roxadustat Trials in Japan.

INTRODUCTION

Thromboembolic events have occurred in clinical trials of roxadustat. This post hoc analysis explored potential factors related to thromboembolic events in dialysis-dependent patients treated with roxadustat in four phase 3 clinical trials in Japan.

METHODS

Thromboembolic events with onset before and after week 12 were evaluated. Baseline risk factors for thromboembolic events were investigated by Cox regression analyses. Nested case-control analyses using conditional logistic models with matched pairs of case-control data explored relationships between thromboembolic events and laboratory parameters.

RESULTS

Of the 444 patients, 56 thromboembolic events were observed in 44 patients during ≤ 52 weeks of treatment. The proportion of venous and arterial thromboembolic events gradually increased after week 12. Baseline risk factors included hemodialysis (vs peritoneal dialysis), advanced age (≥ 65 years), shorter dialysis vintage (< 4 months), and history of thromboembolism. The absence of concomitant intravenous or oral iron therapy (including ferric citrate) was associated with thromboembolic events before week 12 (hazard ratio 11.25; 95% confidence interval [CI] 3.36-37.71; vs presence). Case-control analysis revealed that low average transferrin saturation (< 10%; unadjusted odds ratio [OR] 6.25; 95% CI 1.52-25.62; vs ≥ 20%), high average transferrin level (≥ 2.5 g/L; unadjusted OR 4.36; 95% CI 1.23-15.39; vs < 2.0 g/L), and high average roxadustat dose (≥ 150 mg; unadjusted OR 5.95; 95% CI 1.07-33.16; vs < 50 mg) over the previous 8 weeks before the event onset were associated with thromboembolic events after week 12. However, adjustment for iron status extinguished the significant relationship between roxadustat dose and events. Multivariate case-control analysis showed that increased transferrin from baseline (≥ 1.0 g/L; adjusted OR 7.85; 95% CI 1.82-33.90; vs < 0.5 g/dL) and decreased mean corpuscular volume (< - 2 fL; adjusted OR 5.55; 95% CI 1.73-17.83; vs ≥ 0 fL) were associated with increased risk of thromboembolic events.

CONCLUSION

In addition to established risk factors, iron deficiency may be related to thromboembolic events. Graphical Abstract available for this article.

TRIAL REGISTRATION

NCT02780726, NCT02952092, NCT02780141, NCT02779764.