By Michael Crawford, MD, Editor

SYNOPSIS: Approximately 25% of atrial fibrillation patients in a large trial comparing rivaroxaban to warfarin developed worsening renal failure during follow-up. Those treated with rivaroxaban had less embolic events, but equal bleeding events compared to those on warfarin.

SOURCE: Fordyce CB, Hellkamp AS, Lokhnygina Y, et al. On-treatment outcomes in patients with worsening renal function with rivaroxaban compared with warfarin: Insights from ROCKET AF. Circulation 2016;134:37-47.

Physicians frequently end warfarin treatment if patients with non-valvular atrial fibrillation (AF) develop worsening renal function due to the fear of excessive bleeding risk. The liver predominantly metabolizes rivaroxaban, but one-third is renally excreted. However, clinicians know little about the bleeding risk of continuing rivaroxaban during worsening renal function. Thus, investigators from the Rivaroxaban once daily versus warfarin in AF (ROCKET-AF) trial evaluated the outcome of patients in this large trial presenting with worsening renal function, defined as a decrease in estimated creatinine clearance (CrCl) of > 20% compared to their baseline. By protocol, patients with baseline CrCl of 30-49 mL/min received 15 mg/day of rivaroxaban vs. 20 mg in those with a CrCl > 49 mL/min. Investigators did not change the rivaroxaban dose during the trial unless it fell below 30 mL/min. Patients with a high risk of bleeding were excluded. In 12,600 of the 14,200 patients, there was at least one creatinine after-study entry. These patients made up the study group. The primary efficacy outcome was stroke or systemic embolism, and the primary safety outcome was bleeding. During the course of the study, 26% of the rivaroxaban group and 27% of the warfarin group developed worsening renal function. There was no difference in the primary outcome or bleeding between the worsening renal function group and the rest of the patients. The worsening renal function group experienced a high incidence of vascular death (P = 0.026). Worsening renal function patients randomized to rivaroxaban experienced fewer of the primary outcome events compared to warfarin patients (1.5 vs. 3.3 events per 100 patients-year). Researchers did not observe this in the stable renal function patients. However, bleeding events were not different in those on rivaroxaban who suffered from worsening renal function. The authors concluded that in patients with non-valvular AF on oral anticoagulants who developed worsening renal function, rivaroxaban-treated patients experienced significantly fewer vascular embolic events compared to warfarin-treated patients, but bleeding events were not different between the two therapies.


It is well known that the new oral anticoagulants (NOACs) have to be dose-adjusted for renal function and are not recommended in patients presenting with severe renal impairment. Although warfarin does not need renal function adjustments to dose, the incidence of worsening renal function on warfarin seems higher than on NOACs. This is the first study that has studied the effects of worsening renal function in patients on a NOAC vs. warfarin. It showed more embolic and other vascular events with worsening renal function on both drugs, but there were significantly less on rivaroxaban compared to warfarin. This benefit of rivaroxaban was not at the expense of more bleeding events, which remained equivalent between the two treatment groups with worsening renal function. The authors believe that since about one-quarter of both treatment groups developed worsening renal function, this was an advantage for rivaroxaban therapy.

Secondary analyses of other trials pitting NOACs against warfarin have produced similar results, so this may be a class benefit of NOACs. Since most AF patients are elderly and often suffer from vascular disease, worsening renal function will be common. In fact, about 20% of all patients in the various NOAC trials featured a CrCl < 50 mL/min. The only caveat is that clinicians may need to adjust the NOAC dose as CrCl fails to maintain bleeding rates equivalent to or better than warfarin. Also, at some level of CrCl, most NOACs currently are contraindicated, usually < 15-30 mL/min. Additionally, almost no data exist on NOACs in patients presenting with end-stage renal disease. Thus, warfarin currently serves an important role in patients who need anticoagulation and are severely renally impaired.

Why did warfarin perform worse in comparison to NOACs in patients presenting with declining renal function? This study provides a potential answer, but in the discussion (not in the results) section of the paper. The authors stated that in the worsening renal function patients, warfarin-treated patients spent less time in the therapeutic INR range than that observed in stable renal function patients. Thus, better warfarin management could have eliminated this difference, but that is not always easy to achieve in all situations.

Currently, clinicians and pharmaceutical sales people compare one NOAC vs. warfarin trial results to another to decide which NOAC is better for certain endpoints, such as stroke and bleeding. However, this may not be accurate since each study contained somewhat different patients. It’s time for head-to-head NOAC studies.