Abstract

Aims

Patients with non-valvular atrial fibrillation (AF) and renal insufficiency are at increased risk for ischaemic stroke and bleeding during anticoagulation. Rivaroxaban, an oral, direct factor Xa inhibitor metabolized predominantly by the liver, preserves the benefit of warfarin for stroke prevention while causing fewer intracranial and fatal haemorrhages.

Methods and results

We randomized 14 264 patients with AF in a double-blind trial to rivaroxaban 20 mg/day [15 mg/day if creatinine clearance (CrCl) 30–49 mL/min] or dose-adjusted warfarin (target international normalized ratio 2.0–3.0). Compared with patients with CrCl >50 mL/min (mean age 73 years), the 2950 (20.7%) patients with CrCl 30–49 mL/min were older (79 years) and had higher event rates irrespective of study treatment. Among those with CrCl 30–49 mL/min, the primary endpoint of stroke or systemic embolism occurred in 2.32 per 100 patient-years with rivaroxaban 15 mg/day vs. 2.77 per 100 patient-years with warfarin [hazard ratio (HR) 0.84; 95% confidence interval (CI) 0.57–1.23] in the per-protocol population. Intention-to-treat analysis yielded similar results (HR 0.86; 95% CI 0.63–1.17) to the per-protocol results. Rates of the principal safety endpoint (major and clinically relevant non-major bleeding: 17.82 vs. 18.28 per 100 patient-years; P = 0.76) and intracranial bleeding (0.71 vs. 0.88 per 100 patient-years; P = 0.54) were similar with rivaroxaban or warfarin. Fatal bleeding (0.28 vs. 0.74% per 100 patient-years; P = 0.047) occurred less often with rivaroxaban.

Conclusion

Patients with AF and moderate renal insufficiency have higher rates of stroke and bleeding than those with normal renal function. There was no evidence of heterogeneity in treatment effect across dosing groups. Dose adjustment in ROCKET-AF yielded results consistent with the overall trial in comparison with dose-adjusted warfarin.

See page 2347 for the commentary on this article (doi:10.1093/eurheartj/ehr344)

While most patients with non-valvular atrial fibrillation (AF) benefit from anticoagulation to prevent ischaemic stroke and systemic embolism, those with renal dysfunction face high risks of both thromboembolism and bleeding during antithrombotic therapy.1–3 In observational studies, anticoagulant therapy is frequently not administered in patients with AF and renal dysfunction4,5 based on the concern that bleeding may outweigh the potential benefit. A key question is whether reliable anticoagulation without excessive bleeding risk can be achieved in patients with reduced renal function.

Rivaroxaban is a direct factor Xa inhibitor with predictable pharmacokinetics and a rapid onset and offset of action after oral administration; it is metabolized predominantly (approximately two-thirds) by the liver, with approximately one-third excreted unchanged in the urine.6,7 The Rivaroxaban Once-daily, oral, direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation (ROCKET-AF) compared the oral factor Xa inhibitor rivaroxaban with warfarin for prevention of all stroke and systemic embolism in 14 264 patients with AF.8 The trial was designed to reflect the comorbidities often seen in clinical practice, and as a result, the risk profile of the randomized population was higher than in other recent trials of patients with AF.9–11 Rivaroxaban was non-inferior to warfarin, superior to warfarin while patients were receiving randomized treatment, and resulted in similar overall rates of bleeding, but reduced intracranial and fatal bleeding compared with warfarin.8

In subjects with moderate renal insufficiency [creatinine clearance (CrCl) 30–49 mL/min at baseline, 20.7% of the trial cohort], the dose of rivaroxaban was reduced from 20 to 15 mg daily based on extensive pharmacodynamic data and pharmacokinetic modelling. This pre-specified secondary analysis assessed the risks and benefits of the lower dose of rivaroxaban compared with warfarin in the high-risk cohort of patients with moderate renal insufficiency. It examined the extent to which the findings were consistent with those seen in patients with preserved renal function and in the trial overall.

Methods

The design and main results of the ROCKET-AF trial have been reported previously.8,9 In brief, this was a multicentre, randomized, double-blind, double-dummy, event-driven trial comparing fixed-dose rivaroxaban (20 mg daily or 15 mg daily in patients with  CrCl 30–49 mL/min at baseline) with adjusted-dose warfarin [target international normalized ratio (INR) 2.0–3.0] for prevention of all stroke (ischaemic or haemorrhagic) or systemic embolism.9 To maintain blinding, the HemoSense point-of-care device was used to generate real INR or sham values. The doses of warfarin and matching placebo tablets were adjusted based on these values.12

Renal function and dose assignment

Patients with a CrCl <30 mL/min were excluded from ROCKET-AF. Creatinine clearance was determined by the Cockcroft–Gault formula.13 Based upon prior studies, patients with moderate renal insufficiency (CrCl 30–49 mL/min) have maximal serum concentrations of rivaroxaban 25–30% higher than those with preserved renal function.14,15 Pharmacokinetic models projected that a 25% dose reduction would lead to similar exposure and trough levels in patients with moderate renal insufficiency. The decision about this dose was also informed by analysis of the bleeding risks in the older population with renal dysfunction and associated co-morbidity. Patients were assigned study drug dose based upon CrCl during the screening visit: 20 mg daily or 15 mg daily in patients with CrCl 30–49 mL/min at baseline. There were no dose adjustments post-baseline for changing CrCl. However, patients with CrCl <30 mL/min were required to discontinue study drug.

Study participants

Patients with electrocardiographically documented non-valvular AF at moderate-to-high risk of stroke were recruited at 1178 participating sites in 45 countries. Elevated stroke risk was indicated by a history of stroke, transient ischaemic attack, or systemic embolism or at least two of the following risk factors: heart failure or left ventricular ejection fraction ≤35%, hypertension, age ≥75 years, or diabetes mellitus (CHADS2 score ≥2). Complete inclusion and exclusion criteria have been published. Those with a high risk for bleeding (including prior intra-cerebral bleeding, surgical trauma within 30 days, gastrointestinal bleeding within 6 months) were excluded.9

Outcomes

The primary efficacy endpoint was the composite of all stroke (both ischaemic and haemorrhagic) and systemic embolism. Secondary endpoints included the composite of stroke, non-central nervous system systemic embolism, cardiovascular death, and myocardial infarction, and individual components of the composite endpoints. The principal safety endpoint was the composite of major and non-major clinically relevant bleeding events.16 Bleeding events involving the central nervous system meeting the definition of stroke were adjudicated as haemorrhagic strokes and included in both the primary efficacy and safety endpoints. An independent clinical events committee applied the protocol definitions and adjudicated all suspected stroke, systemic embolism, myocardial infarction, death, and bleeding events contributing to the pre-specified efficacy and safety endpoints.

Statistical analysis

The statistical analysis plan specified testing the efficacy and safety of the randomized comparison in patients with moderate renal dysfunction for consistency with the overall trial results. By design, ROCKET-AF was a non-inferiority trial, and the primary analysis was performed in the per protocol on treatment population.9 In this pre-specified analysis, rivaroxaban patients are classified as 15 mg or 20 mg based on the first dose of study drug. Event rates by treatment arm and renal dysfunction group are presented as per 100 patient-years (%/year). The homogeneity of treatment effects on the occurrence of the primary efficacy and safety endpoints across subgroups of renal dysfunction status (30–49 mL/min and ≥50 mL/min) was tested with a treatment by renal dysfunction interaction. The interaction term, including randomized treatment and renal dysfunction (treatment*renal impairment), was included as a covariate in all of the Cox models.

Estimates and two-sided 95% confidence intervals (CIs) for the hazard ratio (HR) (rivaroxaban vs. warfarin) within levels of renal dysfunction based on the previous model are presented in the per-protocol, on-treatment population for the efficacy endpoints and in the safety, on-treatment population for the safety endpoints. All statistical analyses were performed with SAS software (version 9.2, SAS Institute, Cary, NC, USA).

Results

Patient characteristics

Of the 14 264 patients randomized with AF, 2950 (20.7%) had moderate renal impairment (CrCl 30–49 mL/min) at enrollment. Those randomized with moderately impaired renal function had a median age of 79, a mean CHADS2 score of 3.7 ± 1, 62% were previously treated with a vitamin K antagonist, and 36% were taking aspirin. As shown in Table 1, patients with moderately impaired renal function were older, had higher CHADS2 scores, higher prevalence of heart failure, peripheral vascular disease and prior myocardial infarction. Additionally, compared with those with CrCl >50 mL/min, patients with moderately impaired renal function had lower body mass indices, less frequent history of stroke or transient ischaemic attack, and were less likely to be diabetic (Table 1). Among those patients with CrCl 30–49 mL/min randomized to warfarin, the median time in therapeutic range was 57.7 (25th/75th percentiles: 42.2–69.9. Among those patients with CrCl ≥50 mL/min randomized to warfarin, the median time in therapeutic range was 58.0 (25th/75th percentiles: 43.1–70.8).

Table 1

Baseline patient characteristics randomized to rivaroxaban 15 mg or warfarina

Characteristic CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n= 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Age, median (25th, 75th), years 79 (75, 82) 79 (75, 83) 71 (63, 76) 71 (63, 76) 
Female, n (%) 811 (55.0) 825 (55.9) 2008 (35.6) 1999 (35.4) 
Body mass index, median (25th, 75th), kg/m2 25.1 (22.7, 28.0) 25.2 (22.8, 27.9) 29.2 (26.1, 33.0) 28.9 (26.0, 32.7) 

 
Blood pressure, median (25th, 75th), mm Hg 
 Systolic 130 (120, 140) 130 (120, 140) 130 (120, 140) 130 (120, 140) 
 Diastolic 80 (70, 82) 80 (70, 82) 80 (71, 86) 80 (72, 86) 

 
Clinical presentation, n (%) 
 Type of atrial fibrillation 
  Persistent 1211 (82.2) 1231 (83.4) 4560 (80.9) 4516 (80.1) 
  Paroxysmal 245 (16.6) 215 (14.6) 997 (17.7) 1052 (18.7) 
  Newly diagnosed/new onset 18 (1.2) 30 (2.0) 80 (1.4) 72 (1.3) 

 
Prior aspirin use, n (%) 529 (35.9) 552 (37.4) 2049 (36.4) 2060 (36.5) 
Prior vitamin K antagonist use, n (%) 924 (62.7) 904 (61.3) 3507 (62.2) 3548 (62.9) 

 
Clinical risk factors 
 CHADS2 score, mean ± SD 3.68 (1.00) 3.67 (1.01) 3.42 (0.91) 3.41 (0.92) 
  2, n (%) 130 (8.8) 134 (9.1) 793 (14.1) 797 (14.1) 
  3, n (%) 594 (40.3) 595 (40.3) 2453 (43.5) 2555 (45.3) 
  4, n (%) 426 (28.9) 419 (28.4) 1661 (29.5) 1577 (28.0) 
  5, n (%) 267 (18.1) 267 (18.1) 663 (11.8) 612 (10.9) 
  6, n (%) 56 (3.8) 48 (4.0) 67 (1.2) 99 (1.8) 

 
Prior TIA/stroke or systemic embolism, n (%) 738 (50.1) 725 (49.1) 3167 (56.2) 3160 (56.0) 
Congestive heart failure, n (%) 973 (66.0) 964 (65.3) 3484 (61.8) 3468 (61.5) 
Hypertension, n (%) 1352 (91.7) 1360 (92.1) 5067 (89.9) 5100 (90.4) 
Diabetes mellitus, n (%) 468 (31.8) 492 (33.3) 2401 (42.6) 2319 (41.1) 
Prior myocardial infarction, n (%) 276 (18.7) 302 (20.5) 902 (16.0) 977 (17.3) 
Creatinine clearance, median (25th, 75th), mL/minb 42 (37–46) 42 (37–46) 75 (62–94) 74 (61–92) 
Peripheral vascular disease, n (%) 107 (7.3) 115 (7.8) 292 (5.2) 322 (5.7) 
Chronic obstructive pulmonary disease, n (%) 169 (11.3) 177 (11.9) 582 (39.0) 563 (37.8) 
Characteristic CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n= 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Age, median (25th, 75th), years 79 (75, 82) 79 (75, 83) 71 (63, 76) 71 (63, 76) 
Female, n (%) 811 (55.0) 825 (55.9) 2008 (35.6) 1999 (35.4) 
Body mass index, median (25th, 75th), kg/m2 25.1 (22.7, 28.0) 25.2 (22.8, 27.9) 29.2 (26.1, 33.0) 28.9 (26.0, 32.7) 

 
Blood pressure, median (25th, 75th), mm Hg 
 Systolic 130 (120, 140) 130 (120, 140) 130 (120, 140) 130 (120, 140) 
 Diastolic 80 (70, 82) 80 (70, 82) 80 (71, 86) 80 (72, 86) 

 
Clinical presentation, n (%) 
 Type of atrial fibrillation 
  Persistent 1211 (82.2) 1231 (83.4) 4560 (80.9) 4516 (80.1) 
  Paroxysmal 245 (16.6) 215 (14.6) 997 (17.7) 1052 (18.7) 
  Newly diagnosed/new onset 18 (1.2) 30 (2.0) 80 (1.4) 72 (1.3) 

 
Prior aspirin use, n (%) 529 (35.9) 552 (37.4) 2049 (36.4) 2060 (36.5) 
Prior vitamin K antagonist use, n (%) 924 (62.7) 904 (61.3) 3507 (62.2) 3548 (62.9) 

 
Clinical risk factors 
 CHADS2 score, mean ± SD 3.68 (1.00) 3.67 (1.01) 3.42 (0.91) 3.41 (0.92) 
  2, n (%) 130 (8.8) 134 (9.1) 793 (14.1) 797 (14.1) 
  3, n (%) 594 (40.3) 595 (40.3) 2453 (43.5) 2555 (45.3) 
  4, n (%) 426 (28.9) 419 (28.4) 1661 (29.5) 1577 (28.0) 
  5, n (%) 267 (18.1) 267 (18.1) 663 (11.8) 612 (10.9) 
  6, n (%) 56 (3.8) 48 (4.0) 67 (1.2) 99 (1.8) 

 
Prior TIA/stroke or systemic embolism, n (%) 738 (50.1) 725 (49.1) 3167 (56.2) 3160 (56.0) 
Congestive heart failure, n (%) 973 (66.0) 964 (65.3) 3484 (61.8) 3468 (61.5) 
Hypertension, n (%) 1352 (91.7) 1360 (92.1) 5067 (89.9) 5100 (90.4) 
Diabetes mellitus, n (%) 468 (31.8) 492 (33.3) 2401 (42.6) 2319 (41.1) 
Prior myocardial infarction, n (%) 276 (18.7) 302 (20.5) 902 (16.0) 977 (17.3) 
Creatinine clearance, median (25th, 75th), mL/minb 42 (37–46) 42 (37–46) 75 (62–94) 74 (61–92) 
Peripheral vascular disease, n (%) 107 (7.3) 115 (7.8) 292 (5.2) 322 (5.7) 
Chronic obstructive pulmonary disease, n (%) 169 (11.3) 177 (11.9) 582 (39.0) 563 (37.8) 

SD, standard deviation; TIA, transient ischaemic attack.

aThe safety analysis population included 14 236 patients; however, 9 patients assigned to warfarin had no creatinine clearance data, leaving 14 227 patients as analysed here.

bUsing the Cockcroft–Gault formula.13

CHADS2 Score: One point assigned for: congestive heart failure, hypertension, age over 75, diabetes; 2 points for prior stroke or transient ischaemic attack.

Efficacy outcomes

The primary efficacy analysis for ROCKET-AF showed that stroke or systemic embolism occurred in 429 patients, 188 on rivaroxaban (1.71% per year), and 241 on warfarin (2.16% per year) (HR 0.79; 95% CI 0.66–0.96; P < 0.001 for non-inferiority). In those with moderate renal dysfunction (CrCl 30–49 mL/min), rates of stroke and systemic embolism (the principal efficacy analysis) were higher, regardless of treatment received, than for patients with CrCl >50 mL/min (Table 2). For the randomized comparison in the per-protocol population, there were 2.32 per primary outcome events (stroke or systemic embolism) per 100 patient-years with rivaroxaban 15 mg/day compared with 2.77 per 100 patient-years with dose-adjusted warfarin (HR 0.84; 95% CI 0.57–1.23). These findings were consistent with those seen in patients with CrCl >50 mL/min (1.57 per 100 patient-years and 2.00 per 100 patient-years) (Table 2, Figure 1). As shown in Figure 2, these relationships were similar in the intention-to-treat analysis (ITT). By ITT, the primary event rate was 2.95 per 100 patient-years with rivaroxaban 15 mg/day compared with 3.44 per 100 patient-years with dose-adjusted warfarin (HR 0.86; 95 % CI 0.63–1.17).

Table 2

Primary trial endpoint: stroke and non-CNS embolism in the per protocol population

Clinical endpoint CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
P-value for interaction 
 Rivaroxaban 15 mg (n = 1474)a Warfarin (n= 1476)a Hazard ratio (95% CI), rivaroxaban vs. warfarin Rivaroxaban 20 mg (n= 5637)a Warfarin (n= 5640)a Hazard ratio (95% CI), rivaroxaban vs. warfarin 
Principal efficacy endpoint (stroke and systemic embolism) 2.32 2.77 0.84 (0.57–1.23) 1.57 2.00 0.78 (0.63–0.98) 0.76 
Stroke, systemic embolism, vascular death 4.64 4.83 0.96 (0.73–1.27) 2.76 3.32 0.83 (0.70–0.98) 0.38 
Stoke, systemic embolism, MI, vascular death 5.58 6.54 0.85 (0.67–1.09) 3.55 4.16 0.85 (0.73–0.99) 0.98 

 
Stroke 
 Ischaemic 1.98 1.78 1.11 (0.71–1.73) 1.20 1.34 0.90 (0.69–1.16) 0.41 
 Haemorrhagic 0.29 0.52 0.56 (0.21–1.51) 0.26 0.42 0.62 (0.37–1.03) 0.88 
 Undetermined 0.05 0.09 0.51 (0.05–5.67) 0.07 0.10 0.68 (0.24–1.90) 0.84 
Clinical endpoint CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
P-value for interaction 
 Rivaroxaban 15 mg (n = 1474)a Warfarin (n= 1476)a Hazard ratio (95% CI), rivaroxaban vs. warfarin Rivaroxaban 20 mg (n= 5637)a Warfarin (n= 5640)a Hazard ratio (95% CI), rivaroxaban vs. warfarin 
Principal efficacy endpoint (stroke and systemic embolism) 2.32 2.77 0.84 (0.57–1.23) 1.57 2.00 0.78 (0.63–0.98) 0.76 
Stroke, systemic embolism, vascular death 4.64 4.83 0.96 (0.73–1.27) 2.76 3.32 0.83 (0.70–0.98) 0.38 
Stoke, systemic embolism, MI, vascular death 5.58 6.54 0.85 (0.67–1.09) 3.55 4.16 0.85 (0.73–0.99) 0.98 

 
Stroke 
 Ischaemic 1.98 1.78 1.11 (0.71–1.73) 1.20 1.34 0.90 (0.69–1.16) 0.41 
 Haemorrhagic 0.29 0.52 0.56 (0.21–1.51) 0.26 0.42 0.62 (0.37–1.03) 0.88 
 Undetermined 0.05 0.09 0.51 (0.05–5.67) 0.07 0.10 0.68 (0.24–1.90) 0.84 

aEvent rates per 100 patient-years of follow-up.

Figure 1

Efficacy events in the per-protocol (on-treatment) population.

Figure 1

Efficacy events in the per-protocol (on-treatment) population.

Figure 2

Efficacy endpoints according to the intention to treat.

Figure 2

Efficacy endpoints according to the intention to treat.

As shown in Table 2, the individual and composite secondary efficacy endpoints were consistent in those with moderately impaired renal function and those with CrCl >50 mL/min (P> 0.4 for all interactions). For each of these randomized comparisons, there was no evidence that the treatment effect was different in the impaired renal function subgroup compared with the overall trial.

Safety outcomes

The principal safety endpoint (major and clinically relevant non-major bleeding) occurred more frequently in those with renal insufficiency than in those without, regardless of randomized treatment assignment. However, there was no excess bleeding on rivaroxaban compared with warfarin (Table 3, Figure 3). There was no excess in the principal safety endpoint (HR 0.98; 95% CI 0.84–1.14) or in the individual bleeding outcomes in those treated with rivaroxaban 15 mg/day compared with dose-adjusted warfarin (Table 3). Furthermore, in those with moderate renal insufficiency, critical organ bleeding (HR 0.55; 95% CI 0.30–1.00) and fatal bleeding (HR 0.39; 95% CI 0.15–0.99) were less frequent with rivaroxaban. The lower rate of fatal bleeding was consistent with the findings in those with preserved renal function (HR 0.55; 95% CI 0.32–0.93) (Table 3).

Table 3

Bleeding rates by treatment group rivaroxaban vs. warfarin

Clinical endpoint CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
P-value for interaction 
 Rivaroxaban 15 mg (n = 1474)a Warfarin (n= 1476)a Hazard ratio (95% CI), rivaroxaban vs. warfarin Rivaroxaban 20 mg (n= 5637)a Warfarin (n= 5640)a Hazard ratio (95% CI), rivaroxaban vs. warfarin 
Primary safety endpoint 17.82 18.28 0.98 (0.84–1.14) 14.24 13.67 1.04 (0.96–1.13) 0.4496 

 
Major bleeding 4.49 4.70 0.95 (0.72–1.26) 3.39 3.17 1.07 (0.91–1.26) 0.4800 
 Hb drop 3.76 3.28 1.14 (0.83–1.58) 2.54 2.03 1.25 (1.03–1.52) 0.6456 
 Transfusion 2.34 2.00 1.17 (0.77–1.76) 1.49 1.16 1.28 (0.99–1.65) 0.7066 
 Clinical organ 0.76 1.39 0.55 (0.30–1.00) 0.83 1.13 0.74 (0.55–0.99) 0.3866 
 Fatal bleeding 0.28 0.74 0.39 (0.15–0.99) 0.23 0.43 0.55 (0.32–0.93) 0.5302 

 
Intracranial haemorrhage 0.71 0.88 0.81 (0.41–1.60) 0.44 0.71 0.62 (0.42–0.92) 0.5065 
Clinical endpoint CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
P-value for interaction 
 Rivaroxaban 15 mg (n = 1474)a Warfarin (n= 1476)a Hazard ratio (95% CI), rivaroxaban vs. warfarin Rivaroxaban 20 mg (n= 5637)a Warfarin (n= 5640)a Hazard ratio (95% CI), rivaroxaban vs. warfarin 
Primary safety endpoint 17.82 18.28 0.98 (0.84–1.14) 14.24 13.67 1.04 (0.96–1.13) 0.4496 

 
Major bleeding 4.49 4.70 0.95 (0.72–1.26) 3.39 3.17 1.07 (0.91–1.26) 0.4800 
 Hb drop 3.76 3.28 1.14 (0.83–1.58) 2.54 2.03 1.25 (1.03–1.52) 0.6456 
 Transfusion 2.34 2.00 1.17 (0.77–1.76) 1.49 1.16 1.28 (0.99–1.65) 0.7066 
 Clinical organ 0.76 1.39 0.55 (0.30–1.00) 0.83 1.13 0.74 (0.55–0.99) 0.3866 
 Fatal bleeding 0.28 0.74 0.39 (0.15–0.99) 0.23 0.43 0.55 (0.32–0.93) 0.5302 

 
Intracranial haemorrhage 0.71 0.88 0.81 (0.41–1.60) 0.44 0.71 0.62 (0.42–0.92) 0.5065 

aEvent rates per 100 patient-years of follow-up.

Figure 3

Safety endpoints.

Figure 3

Safety endpoints.

Bleeding events were also examined by site of haemorrhage as defined in Table 4. In patients with moderate renal insufficiency, rivaroxaban-treated patients had more frequent gastrointestinal bleeding (4.1 vs. 2.6%; P = 0.02). While limited by smaller sample size, there were no significant differences in other bleeding sites between rivaroxaban 15 mg and dose-adjusted warfarin. Haemorrhagic stroke was less frequent with rivaroxaban than with warfarin for those with CrCl >50 mL/min (HR 0.58; 95% CI 0.36–0.94), and the finding in those with moderate renal insufficiency was consistent (HR 0.58; 95% CI 0.23–1.47) (Figure 2).

Table 4

Bleeding sites

 CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n = 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Major bleeding, n (%) Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up 
Gastrointestinal (upper, lower, and rectal) 2.88 1.77 1.79 1.12 

 
Intracranial 0.71 0.88 0.44 0.71 
 Intraparenchymal 0.43 0.55 0.31 0.48 
  Non-traumatic 0.38 0.51 0.28 0.47 
  Traumatic 0.05 0.05 0.03 0.01 
 Intraventricular 0.19 0.28 0.10 0.26 
 Subdural hematoma 0.28 0.46 0.09 0.19 
 Subarachnoid 0.14 0.28 0.04 0.09 
 Epidural hematoma 0.00 0.00 0.00 0.01 

 
Macroscopic haematuria 0.05 0.18 0.28 0.19 
Bleeding associated with non-cardiac surgery 0.24 0.42 0.15 0.19 
Intraocular/retinal 0.05 0.18 0.18 0.22 
Intra-articular 0.00 0.23 0.18 0.17 
Epistaxis 0.19 0.09 0.10 0.13 
 CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n = 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Major bleeding, n (%) Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up Event rates per 100 patient-years of follow-up 
Gastrointestinal (upper, lower, and rectal) 2.88 1.77 1.79 1.12 

 
Intracranial 0.71 0.88 0.44 0.71 
 Intraparenchymal 0.43 0.55 0.31 0.48 
  Non-traumatic 0.38 0.51 0.28 0.47 
  Traumatic 0.05 0.05 0.03 0.01 
 Intraventricular 0.19 0.28 0.10 0.26 
 Subdural hematoma 0.28 0.46 0.09 0.19 
 Subarachnoid 0.14 0.28 0.04 0.09 
 Epidural hematoma 0.00 0.00 0.00 0.01 

 
Macroscopic haematuria 0.05 0.18 0.28 0.19 
Bleeding associated with non-cardiac surgery 0.24 0.42 0.15 0.19 
Intraocular/retinal 0.05 0.18 0.18 0.22 
Intra-articular 0.00 0.23 0.18 0.17 
Epistaxis 0.19 0.09 0.10 0.13 

Adverse events according to randomized treatment were similar in those with CrCl 30–49 mL/min compared with those with CrCl >50 mL/min (Table 5). In those with CrCl 30–49 mL/min, the adverse event rates were similar between those randomized to rivaroxaban and warfarin.

Table 5

Adverse events according to renal function and randomized treatment

Adverse eventb, n (%) CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n= 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Total patients with treatment-emergent adverse eventsa 1248 (84.7) 1281 (86.8) 4543 (80.6) 4520 (80.1) 
 Epistaxis 150 (10.2) 121 (8.2) 571 (10.1) 488 (8.7) 
 Peripheral oedema 115 (7.8) 120 (8.1) 320 (5.7) 324 (5.7) 
 Dizziness 110 (7.5) 118 (8.0) 323 (5.7) 330 (5.9) 
 Nasopharyngitis 76 (5.2) 84 (5.7) 345 (6.1) 369 (6.5) 
 Cardiac failure 104 (7.1) 120 (8.1) 293 (5.2) 299 (5.3) 
 Bronchitis 94 (6.4) 90 (6.1) 302 (5.4) 326 (5.8) 
 Dyspnoea 83 (5.6) 102 (6.9) 297 (5.3) 292 (5.2) 
 Diarrhoea 84 (5.7) 96 (6.5) 295 (5.2) 300 (5.3) 
 Cough 81 (5.5) 74 (5.0) 262 (4.6) 279 (4.9) 
 Back pain 58 (3.9) 74 (5.0) 280 (5.0) 272 (4.8) 
 Upper respiratory tract infection 62 (4.2) 70 (4.7) 274 (4.9) 255 (4.5) 
 Headache 68 (4.6) 70 (4.7) 256 (4.5) 291 (5.2) 
 Arthralgia 73 (5.0) 69 (4.7) 228 (4.0) 262 (4.6) 
 Haematuria 47 (3.2) 58 (3.9) 249 (4.4) 183 (3.2) 
 Urinary tract infection 72 (4.9) 105 (7.1) 221 (3.9) 216 (3.8) 
Adverse eventb, n (%) CrCl 30–49 mL/min
 
CrCl ≥50 mL/min
 
 Rivaroxaban 15 mg (n= 1474) Warfarin (n= 1476) Rivaroxaban 20 mg (n= 5637) Warfarin (n= 5640) 
Total patients with treatment-emergent adverse eventsa 1248 (84.7) 1281 (86.8) 4543 (80.6) 4520 (80.1) 
 Epistaxis 150 (10.2) 121 (8.2) 571 (10.1) 488 (8.7) 
 Peripheral oedema 115 (7.8) 120 (8.1) 320 (5.7) 324 (5.7) 
 Dizziness 110 (7.5) 118 (8.0) 323 (5.7) 330 (5.9) 
 Nasopharyngitis 76 (5.2) 84 (5.7) 345 (6.1) 369 (6.5) 
 Cardiac failure 104 (7.1) 120 (8.1) 293 (5.2) 299 (5.3) 
 Bronchitis 94 (6.4) 90 (6.1) 302 (5.4) 326 (5.8) 
 Dyspnoea 83 (5.6) 102 (6.9) 297 (5.3) 292 (5.2) 
 Diarrhoea 84 (5.7) 96 (6.5) 295 (5.2) 300 (5.3) 
 Cough 81 (5.5) 74 (5.0) 262 (4.6) 279 (4.9) 
 Back pain 58 (3.9) 74 (5.0) 280 (5.0) 272 (4.8) 
 Upper respiratory tract infection 62 (4.2) 70 (4.7) 274 (4.9) 255 (4.5) 
 Headache 68 (4.6) 70 (4.7) 256 (4.5) 291 (5.2) 
 Arthralgia 73 (5.0) 69 (4.7) 228 (4.0) 262 (4.6) 
 Haematuria 47 (3.2) 58 (3.9) 249 (4.4) 183 (3.2) 
 Urinary tract infection 72 (4.9) 105 (7.1) 221 (3.9) 216 (3.8) 

Based on safety population.

aEvents that started on or after the first dose of study medication and up to 2 days after the last dose of study medication.

bFifteen most frequent treatment-emergent adverse events based on the rivaroxaban group.

Discussion

Among patients with AF, renal dysfunction is common and progressively increases with older age.17 As reflected in the ROCKET-AF trial, such patients also demonstrate complex co-morbidity, including congestive heart failure, prior hypertension, and diabetes. ROCKET-AF differs from a number of prior trials in AF, both in design (double-blind vs. open-label) and in the risk profile of the included patients (ROCKET-AF mean CHADS2 score = 3.5 vs. RE-LY mean CHADS2 score = 2.1). Due to differences in both study design and the selected population, caution must be exercised when comparing across trials.18 Consistent with prior observations, this study demonstrates that those with renal dysfunction are at increased risk for stroke and embolic events and, irrespective of anticoagulant administered, they are also at increased risk for bleeding events. For the randomized comparison of rivaroxaban vs. warfarin, the findings in the patients with moderate renal dysfunction were consistent with the overall trial. Specifically, the reduced dose of rivaroxaban preserved the treatment effect of warfarin without increasing bleeding and with fewer fatal bleeds.

Rivaroxaban 15 mg dose selection in moderate renal insufficiency

A single dose of rivaroxaban inhibits thrombin generation for 24 h and prolongs the pro-thrombin time in a dose-dependent manner.6,19 In prior studies of prophylaxis of venous thromboembolism, once-daily and twice-daily regimens differed only in rivaroxaban trough concentrations, not total exposure as measured by the area under the curve.20,21 In post-operative deep vein thrombosis prevention studies, once-daily dosing led to similar efficacy as twice-daily dosing with less bleeding.22,23 Accordingly, ROCKET-AF employed once-daily dosing of rivaroxaban. While rivaroxaban is predominantly metabolized by the liver, approximately one-third is cleared renally.24 In patients with moderate renal insufficiency (CrCl 30–49 mL/min), there is a moderate but relevant effect on rivaroxaban exposure; maximal serum concentrations are 25–30% higher14 and pharmacokinetic models projected that a 25% dose reduction would lead to similar exposure and trough levels in patients with moderate renal insufficiency. Additionally, the higher intrinsic risk of bleeding in patients with moderate renal impairment combined with the comorbidities of an elderly AF population also favoured a lower dose. Therefore, based upon the aggregate data, patients in ROCKET-AF with moderate renal insufficiency were treated with a 25% dose reduction (15 mg daily).9 Major and clinically relevant non-major bleeding occurred more frequently in those with renal insufficiency than in those without, regardless of randomized treatment assignment, but there was no excess bleeding with rivaroxaban compared with warfarin. There was no excess in the individual bleeding outcomes in those treated with rivaroxaban 15 mg/day compared with dose-adjusted warfarin, and critical organ bleeding (HR 0.55; 95% CI 0.30–1.00) and fatal bleeding (HR 0.39; 95% CI 0.15–0.99) were less frequent with rivaroxaban. The lower rate of fatal bleeding was consistent with the findings in those with preserved renal function. Clearance of rivaroxaban (mean terminal elimination half-life up to 9 h) is more rapid than for wafarin and standard medical and surgical measures are recommended, if necessary, to control bleeding (as set out in the protocol of the study).8 No additional measures were required among the patients in the ROCKET AF trial to control bleeding. The study protocol specified that bleeding may be managed (after attention to other potential causes of bleeding, including concomitant therapies) by infusion of fresh frozen plasma and, if necessary, factor concentrates.8 Although bleeding is not more frequent than with warfarin, the treatment of bleeding due to rivaroxaban merits additional study.

Implications of the findings

Irrespective of geographic location, observational studies reveal that older patients with AF and those with renal dysfunction are under-treated with anticoagulants.5,25,26 Concerns about bleeding and falls among frail patients with AF have often led to the decision not to treat with anticoagulants, despite the known risks of stroke and embolism. In the US-based NABOR program, almost half of the patients at high risk for stroke were not treated with an anticoagulant, and a similar proportion of moderate-risk patients were untreated.27 Previous studies have shown that bleeding increases with CHADS2 score (19.5 major bleeds per 100 patient-years for CHADS2 of 3; 23.5 per 100 patient-years for CHADS2 of 4: compared with 7.2 per 100 patient-years overall).28 Similarly, in this US-based study, managing warfarin anticoagulation was more challenging in those with congestive heart failure.28,29 Previous randomized trials of anticoagulants in AF have focused on lower-risk patients (CHADS2 median of approximately 2) than those randomized in the ROCKET-AF trial (CHADS2 median >3) and seen in unselected clinical practice.10,11,30 Those with moderate renal insufficiency have higher CHADS2 values than patients without renal dysfunction (Table 1), hence the importance of testing the randomized comparison of rivaroxaban and warfarin in a population that included a substantial cohort with moderate renal insufficiency (one in five study patients). This study has shown that it is possible to anticoagulate patients with moderate renal dysfunction, without excessive bleeding.

The strengths of this study are that the double-blind, randomized comparison was conducted in a wide range of countries and different health care systems. This is relevant to the translation of trial evidence into clinical practice. Although those with moderate renal insufficiency are a substantial cohort within the ROCKET-AF trial, this analysis was not powered to demonstrate non-inferiority or superiority for the comparison of rivaroxaban vs. warfarin. Nevertheless, the findings show consistency between the subgroup with renal dysfunction and those without renal dysfunction.

Conclusion

Patients with AF and moderate renal impairment respond favourably to a reduced dose of rivaroxaban (15 mg once daily) compared with warfarin and in a fashion similar to those with preserved renal function given 20 mg of rivaroxaban daily. Although we were unable to demonstrate non-inferiority or superiority for the comparison of rivaroxaban vs. warfarin in patients with moderate renal insufficiency (CrCl 30–49 mL/min), the absence of heterogeneity between overall results and this subgroup suggests that the dose adjustment of rivaroxaban in patients with moderate renal insufficiency yields results consistent with the overall trial. Rivaroxaban preserved the benefit of warfarin in preventing stroke and systemic embolus, and produced lower rates while on treatment. Bleeding rates with the reduced dose of rivaroxaban were similar to those on warfarin therapy, and there were fewer fatal bleeds with rivaroxaban.

Funding

This work was supported by grants from Johnson & Johnson Pharmaceutical Research & Development L.L.C. (Raritan, NJ) and Bayer HealthCare Pharmaceuticals (Berlin, Germany). An international executive committee, led by Drs Califf and Fox, designed the trial and was responsible for oversight of study conduct, retained independent ability to analyse and present the data, and take responsibility for the accuracy and completeness of data analyses. The Duke Clinical Research Institute (Durham, NC) coordinated the trial, managed the database, and performed the primary analyses for this manuscript independent of the sponsors.

References

1
Fang
MC
Go
AS
Chang
Y
Borowsky
LH
Pomernacki
NK
Udaltsova
N
Singer
DE
A new risk scheme to predict warfarin-associated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) study
J Am Coll Cardiol
 , 
2011
, vol. 
58
 (pg. 
395
-
401
)
2
Pisters
R
Lane
DA
Nieuwlaat
R
de Vos
CB
Crijns
HJ
Lip
GY
A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: The Euro Heart Survey
Chest
 , 
2010
, vol. 
138
 (pg. 
1093
-
1100
)
3
Vazquez
E
Sanchez-Perales
C
The HAS-BLED score and renal failure
Chest
 , 
2011
, vol. 
139
 (pg. 
1248
-
1249
author reply 1249–1250
4
Johnston
JA
Cluxton
RJ
Jr
Heaton
PC
Guo
JJ
Moomaw
CJ
Eckman
MH
Predictors of warfarin use among Ohio medicaid patients with new-onset nonvalvular atrial fibrillation
Arch Intern Med
 , 
2003
, vol. 
163
 (pg. 
1705
-
1710
)
5
Piccini
JP
Hernandez
AF
Zhao
X
Patel
MR
Lewis
WR
Peterson
ED
Fonarow
GC
Quality of care for atrial fibrillation among patients hospitalized for heart failure
J Am Coll Cardiol
 , 
2009
, vol. 
54
 (pg. 
1280
-
1289
)
6
Kubitza
D
Becka
M
Wensing
G
Voith
B
Zuehlsdorf
M
Safety, pharmacodynamics, and pharmacokinetics of BAY 59–7939—an oral, direct factor Xa inhibitor—after multiple dosing in healthy male subjects
Eur J Clin Pharmacol
 , 
2005
, vol. 
61
 (pg. 
873
-
880
)
7
Kubitza
D
Becka
M
Roth
A
Mueck
W
Dose-escalation study of the pharmacokinetics and pharmacodynamics of rivaroxaban in healthy elderly subjects
Curr Med Res Opin
 , 
2008
, vol. 
24
 (pg. 
2757
-
2765
)
8
Patel
MR
Mahaffey
KW
Garg
J
Pan
G
Singer
DE
Hacke
W
Piccini
JP
Becker
R
Nessel
CC
Paolini
JF
Berkowitz
SD
Fox
KAA
Califf
RM
Rivaroxaban versus warfarin in nonvalvular atrial fibrillation
N Engl J Med
 , 
2011
9
ROCKET AF Study Investigators
Rivaroxaban—once daily, oral, direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation: Rationale and design of the ROCKET AF study
Am Heart J
 , 
2010
, vol. 
159
 (pg. 
340
-
347
)
10
Connolly
SJ
Ezekowitz
MD
Yusuf
S
Eikelboom
J
Oldgren
J
Parekh
A
Pogue
J
Reilly
PA
Themeles
E
Varrone
J
Wang
S
Alings
M
Xavier
D
Zhu
J
Diaz
R
Lewis
BS
Darius
H
Diener
HC
Joyner
CD
Wallentin
L
Dabigatran versus warfarin in patients with atrial fibrillation
N Engl J Med
 , 
2009
, vol. 
361
 (pg. 
1139
-
1151
)
11
Connolly
SJ
Eikelboom
J
Joyner
C
Diener
H-C
Hart
R
Golitsyn
S
Flaker
G
Avezum
A
Hohnloser
SH
Diaz
R
Talajic
M
Zhu
J
Pais
P
Budaj
A
Parkhomenko
A
Jansky
P
Commerford
P
Tan
RS
Sim
K-H
Lewis
BS
Van Mieghem
W
Lip
GYH
Kim
JH
Lanas-Zanetti
F
Gonzalez-Hermosillo
A
Dans
AL
Munawar
M
O'Donnell
M
Lawrence
J
Lewis
G
Afzal
R
Yusuf
S
Apixaban in patients with atrial fibrillation
N Engl J Med
 , 
2011
, vol. 
364
 (pg. 
806
-
817
)
12
Albers
GW
Diener
HC
Frison
L
Grind
M
Nevinson
M
Partridge
S
Halperin
JL
Horrow
J
Olsson
SB
Petersen
P
Vahanian
A
Ximelagatran vs warfarin for stroke prevention in patients with nonvalvular atrial fibrillation: a randomized trial
JAMA
 , 
2005
, vol. 
293
 (pg. 
690
-
698
)
13
Cockcroft
DW
Gault
MH
Prediction of creatinine clearance from serum creatinine
Nephron
 , 
1976
, vol. 
16
 (pg. 
31
-
41
)
14
Kubitza
D
Becka
M
Mueck
W
Halabi
A
Maatouk
H
Klause
N
Lufft
V
Wand
DD
Philipp
T
Bruck
H
Effects of renal impairment on the pharmacokinetics, pharmacodynamics and safety of rivaroxaban, an oral, direct factor Xa inhibitor
Br J Clin Pharmacol
 , 
2010
, vol. 
70
 (pg. 
703
-
712
)
15
Mueck
W
Lensing
AW
Agnelli
G
Decousus
H
Prandoni
P
Misselwitz
F
Rivaroxaban: population pharmacokinetic analyses in patients treated for acute deep vein thrombosis and exposure simulations in patients with atrial fibrillation treated for stroke prevention
Clin Pharmacokinet
 , 
2011
 
in press
16
Schulman
S
Kearon
C
Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients
J Thromb Haemost
 , 
2005
, vol. 
3
 (pg. 
692
-
694
)
17
Soliman
EZ
Prineas
RJ
Go
AS
Xie
D
Lash
JP
Rahman
M
Ojo
A
Teal
VL
Jensvold
NG
Robinson
NL
Dries
DL
Bazzano
L
Mohler
ER
Wright
JT
Feldman
HI
Chronic kidney disease and prevalent atrial fibrillation: the chronic renal insufficiency cohort (CRIC)
Am Heart J
 , 
2010
, vol. 
159
 (pg. 
1102
-
1107
)
18
Connolly
SJ
Ezekowitz
MD
Yusuf
S
Eikelboom
J
Oldgren
J
Parekh
A
Pogue
J
Reilly
PA
Themeles
E
Varrone
J
Wang
S
Alings
M
Xavier
D
Zhu
J
Diaz
R
Lewis
BS
Darius
H
Diener
HC
Joyner
CD
Wallentin
L
RE-LY Steering Committee and Investigators
Dabigatran versus warfarin in patients with atrial fibrillation
N Engl J Med
 , 
2009
, vol. 
361
 (pg. 
1139
-
1151
)
19
Harder
S
Graff
J
Hentig
NV
Misselwitz
F
Kubitza
D
Zuehlsdorf
M
Wensing
G
Mueck
W
Becka
M
Breddin
HK
Effects of BAY 59-7939, an oral, direct Factor Xa inhibitor, on thrombin generation in healthy volunteers
Pathophysiol Haemost Thromb
 , 
2004
, vol. 
33
 
Suppl. 2
pg. 
97
  
Abstract PO078
20
Agnelli
G
Gallus
A
Goldhaber
SZ
Haas
S
Huisman
MV
Hull
RD
Kakkar
AK
Misselwitz
F
Schellong
S
Treatment of proximal deep-vein thrombosis with the oral direct factor Xa inhibitor rivaroxaban (BAY 59-7939): the ODIXa-DVT (Oral Direct Factor Xa Inhibitor Bay 59-7939 in Patients With Acute Symptomatic Deep-Vein Thrombosis) study
Circulation
 , 
2007
, vol. 
116
 (pg. 
180
-
187
)
21
Buller
H
Deitchman
D
Prins
M
Segers
A
Efficacy and safety of the oral direct factor Xa inhibitor apixaban for symptomatic deep vein thrombosis. The Botticelli DVT dose-ranging study
J Thromb Haemost
 , 
2008
, vol. 
6
 (pg. 
1313
-
1318
)
22
Eriksson
BI
Borris
LC
Dahl
OE
Haas
S
Huisman
MV
Kakkar
AK
Muehlhofer
E
Dierig
C
Misselwitz
F
Kalebo
P
A once-daily, oral, direct factor Xa inhibitor, rivaroxaban (BAY 59-7939), for thromboprophylaxis after total hip replacement
Circulation
 , 
2006
, vol. 
114
 (pg. 
2374
-
2381
)
23
Piccini
JP
Patel
MR
Mahaffey
KW
Fox
KA
Califf
RM
Rivaroxaban, an oral direct factor Xa inhibitor
Expert Opin Investig Drugs
 , 
2008
, vol. 
17
 (pg. 
925
-
937
)
24
Kubitza
D
Becka
M
Voith
B
Zuehlsdorf
M
Wensing
G
Safety, pharmacodynamics, and pharmacokinetics of single doses of BAY 59-7939, an oral, direct factor Xa inhibitor
Clin Pharmacol Ther
 , 
2005
, vol. 
78
 (pg. 
412
-
421
)
25
Ogilvie
IM
Newton
N
Welner
SA
Cowell
W
Lip
GY
Underuse of oral anticoagulants in atrial fibrillation: a systematic review
Am J Med
 , 
2010
, vol. 
123
 (pg. 
638
-
645
)
26
Robert-Ebadi
H
Le Gal
G
Righini
M
Use of anticoagulants in elderly patients: practical recommendations
Clin Interv Aging
 , 
2009
, vol. 
4
 (pg. 
165
-
177
)
27
Waldo
AL
Becker
RC
Tapson
VF
Colgan
KJ
Hospitalized patients with atrial fibrillation and a high risk of stroke are not being provided with adequate anticoagulation
J Am Coll Cardiol
 , 
2005
, vol. 
46
 (pg. 
1729
-
1736
)
28
Hylek
EM
Evans-Molina
C
Shea
C
Henault
LE
Regan
S
Major hemorrhage and tolerability of warfarin in the first year of therapy among elderly patients with atrial fibrillation
Circulation
 , 
2007
, vol. 
115
 (pg. 
2689
-
2696
)
29
Witt
DM
Delate
T
Clark
NP
Martell
C
Tran
T
Crowther
MA
Garcia
DA
Ageno
W
Hylek
EM
Outcomes and predictors of very stable INR control during chronic anticoagulation therapy
Blood
 , 
2009
, vol. 
114
 (pg. 
952
-
956
)
30
Le Heuzey
JY
Breithardt
G
Camm
J
Crijns
H
Dorian
P
Kowey
PR
Merioua
I
Prystowsky
EN
Schwartz
PJ
Torp-Pedersen
C
Weintraub
W
The RecordAF study: design, baseline data, and profile of patients according to chosen treatment strategy for atrial fibrillation
Am J Cardiol
 , 
2010
, vol. 
105
 (pg. 
687
-
693
)

Supplementary data

Comments

0 Comments