Evidence-Based Diabetes Management May 2015
Real-World Evidence Mounts for Rivaroxaban
A steady stream of follow-up research suggests that the novel oral anticoagulant (NOAC) rivaroxaban will fulfill the promise it showed in trials and provide a superior combination of stroke prevention, safety, and convenience for patients with nonvalvular atrial fibrillation (AF) and a wide variety of comorbidities, such as type 2 diabetes mellitus (T2DM).
Subgroup analysis indicates that rivaroxaban compares as well to war-farin (in terms of both safety and ef-ficacy) in small patient segments as it compares to the older drug in AF patients as a whole.1
Studies of clinical usage, moreover, have demonstrated an advantage that trials did not: the relative ease of riva-roxaban use, compared with warfarin, reduces the percentage of patients who give up on anticoagulation therapy and use less effective stroke prevention strategies.2
These same studies have raised some issues that should be noted by physicians who treat both AF and T2DM. Overall, however, they contain good news for the many patients who have both conditions. (Estimates of co-morbidity vary, but research suggests that T2DM may double a patient’s odds of developing AF,3 and that about 17% of patients with AF have diabetes.4
“Substudies have shown that the beneficial effects of rivaroxaban are consistent with the results of the over-all study, irrespective of the history of previous stroke or transient ischemic attack, heart failure, myocardial infarc-tion, diabetes or moderate renal dysfunction,” wrote the authors of a recent article in Expert Review of Cardiovascular Therapy. “Moreover, recent data have confirmed the efficacy and safety of ri-varoxaban in real-life practice.”5
Such findings have helped NOACs like rivaroxaban take market share from warfarin, a vitamin K antagonist that has long been a mainstay of AF treat-ment because it is a very potent anti-thrombotic agent.
The efficacy of warfarin does vary, not only among different types of AF patients with different comorbidities, but also among different individual pa-tients who appear medically relevant. That said, a meta-analysis of studies with more than 28,000 patients calculated that the drug reduced stroke risk by an average of 64%.6
Unfortunately, warfarin is a famously difficult medicine to use. Its narrow therapeutic window offers patients little protection against stroke at inter-national normalized ratios (INRs) below 2.0, and produces unacceptable bleed-ing risks at INRs above 3.0. Worse, the dose needed to keep patients inside this narrow therapeutic window varies from patient to patient, and the dose needed to keep any given patient within this range varies from time to time. The only way to keep patients in the therapeutic zone is to perform repeated blood tests throughout the patient’s life and make adjustments as needed.
Patients must submit not only to this testing regime but also to the dietary and medical restrictions needed to prevent interactions between warfarin and vari-ous other drugs and foods. Indeed, there is some evidence that warfarin interacts with α-glucosidase inhibitors that are sometimes used to treat diabetes.7
These drawbacks, along with the fear of intracranial hemorrhages and other major bleeds, have always led many patients to refuse warfarin treatment and many others to abandon the drug after using it for a short period. Estimates vary, but some suggest that half of all patients who would benefit from warfarin never try it8 and that half of those who try it discontinue treatment within 5 years.9
Such decisions often prove fatal. Even in the absence of significant comor-bidities, AF is associated with a 5-fold increase in the risk of ischemic stroke. Certain common comorbidities increase the risk further still: prior stroke/tran-sient ischemic attack (relative risk [RR], 2.5; 95% CI, 1.8-3.5), hypertension (RR, 2.0; 95% CI, 1.6-2.5), diabetes mellitus (RR, 1.7; 95% CI, 1.4-2.0), and increasing age (RR, 1.5 per decade; 95% CI, 1.3-1.7).10
These numbers illustrate the lifesaving potential of any drug that could match (or exceed) the efficacy and safe-ty of warfarin without all the difficul-ties that discourage patient adherence. The phase 3 trial that led the FDA to ap-prove rivaroxaban in late 2011 indicated that the newcomer had the potential to be just such a drug.
Researchers who randomized 14,264 AF patients between the 2 drugs and followed them for an average of 707 days found that the rates of stroke or systemic embolism in the intention to treat analysis were 2.1% per year for ri-varoxaban and 2.4% per year for warfa-rin (hazard ratio [HR], 0.88; 95% CI, 0.74-1.03; P <.001 for non-inferiority, P = .12 for superiority).11
Rivaroxaban also boasted several other obvious advantages over the older medication. It came in just 2 doses 15 mg for people with impaired kidney function and 20 mg for everyone else and thus required no monitoring or dose adjustments. It placed no restric-tions on what patients could eat or what other medications they could use. It was even associated with a far lower risk of intracranial hemorrhage (HR, 0.67; 95% CI, 0.47-0.93; P = .02).11
Yet despite all those advantages, the trial indicated that patients were just as likely to stop taking rivaroxaban as they were to stop taking warfarin. Discontinuation rates were 23.7% for rivaroxaban and 22.2% for warfarin.11
Now, a large study of real-world rivaroxaban use concludes that the discontinuation rates reported by the phase 3 trial says more about trial design than actual patient behavior.
That study which took data from an ongoing, prospective, noninterventional registry that is following more than 1204 rivaroxaban users in Germany found that about 15% of patients discontinued rivaroxaban at some point in the first year of treatment and only about 3% of pa-tients discontinued it at any subsequent point.2 This compares very favorably to tients (as opposed to those in clinical trials), which have found first-year discontinuation rates as high as 30% and 3-year discontinuation rates of 50%.9,11
Another recent study a retrospective matched-cohort study of American patients found that treatment persistence at 6 months was 81.5% for rivaroxaban and just 68.3% for warfarin.14 “Even more importantly, the persistence rates for rivaroxaban reported by Laliberte ́ et al, and confirmed in our study, are substantially higher than those re-ported for [vitamin K antagonists] in daily care settings,” wrote the authors of the German study in Europace.
Dramatically higher persistent rates could translate into substantially fewer strokes overall, and the persistence figures were not significantly different for most of the subgroups the German researchers broke out from the total cohort. Patients with renal impairments, prior strokes, arterial hypertension, and prior discontinuation of warfarin were about as likely as the general population to stick with rivaroxaban.
The only subgroup that was significantly less likely than average to discontinue treatment consisted of patients with unusually low body mass index (HR, 0.74; 95% CI, 0.55-0.99; P = .04).
The only 2 subgroups that were sig-nificantly more likely than average to discontinue treatment consisted of pa-tients with diabetes (HR, 1.35; 95% CI, 1.03-1.77; P = .03) and patients with transient ischemic attack, stroke, or system-ic embolism (HR, 1.41; 95% CI, 1.08-1.85; P = .01). The most common reasons for rivaroxaban discontinuation among the cohort as a whole were bleeding (30%), other side effects (24.2%), and resump-tion of sinus rhythm (9.9%).
The paper’s lead author, Jan Beyer-Westendorf, MD, declined to specify why patients with diabetes were more likely than others to stop taking rivaroxaban, but future papers from his study group should help physicians decide whether to try such patients on rivaroxaban or some alternative. “We are planning to look at several subgroups including diabetic pa-tients,” he wrote in an e-mail. “However, we are still in the follow-up phase [of the Dresden NOAC study] and perform only a few interim analyses. Therefore, we do not have answers at present.”
The years since rivaroxaban’s initial approval have also brought a steady stream of information about its safety and effica-cy in AF patients with various comorbidities. Most of that information has been mined from the drug’s phase 3 trial data, and most of it has been encouraging.
To start with the 2 populations that were unusually likely to discontinue rivaroxaban in the German study those with diabetes and those with a history of stroke analyses suggested that both groups might fare marginally better with rivaroxaban than warfarin. The risk of stroke or systemic embolism was very nearly (but not quite) significantly lower among diabetic AF patients who took ri-varoxaban rather than warfarin (HR, 0.74; 95% CI, 0.54-1.01; P = .055) and overall bleeding risks were nearly identical.13
In patients with prior strokes, rivaroxaban performed slightly better than warfarin but, again, it was close enough that it may have been by chance (HR, 0.94; 95% CI, 0.77-1.16). The same was true when researchers calculated the risk of intracranial hemorrhage (HR, 0.74; 95% CI, 0.47-1.15) and fatal bleed-ing (HR, 0.54; 95% CI, 0.29-1.00).15