DOACs – advances and limitations in real world

Stroke prevention in non-valvular atrial fibrillation

Anticoagulation therapy is necessary to prevent stroke, systemic embolization and all-cause mortality in patients with AF. THE CHADS2 score (Cardiac Failure, Hypertension, Age, Diabetes, and Stroke [double]) was widely used in risk stratification to identify patients who will benefit from anticoagulation [2]. However, even within the score “0”, the risks of stroke ranged from 0.84 to 3.4 per year, hence missing out on those with increased risks who would have benefitted from anticoagulation. This gap was addressed with the CHA2DS2-VASc score (congestive heart failure, hypertension, age???75 years [doubled], diabetes mellitus, stroke [doubled], vascular disease, age 65–74 years, sex category [female]) which can better identify truly low risk AF patients, who are unlikely to benefit from antithrombotic therapy [3]. CHASDS2-VASc score is currently the preferred tool for risk stratification for stroke risk in AF patients.

Currently 4 DOACs available are dabigatran, rivaroxaban, apixaban and edoxaban have each shown similar efficacy and safety when compared to warfarin [4–7]. A meta-analysis of the phase III trials of these four DOACs showed a consistent favourable risk-benefit profile across a wide range of patients with significant reductions in stroke or systemic embolism, intracranial haemorrhage, and mortality but increased gastrointestinal bleeds when compared with warfarin [8]. Coupled with their convenient usage as fixed dose oral medications without the need for frequent laboratory tests and dose adjustments, DOACs have emerged as the preferred treatment option in some guidelines [9]. Large scale real world data of these drugs is has become available with increasing use of DOACs in routine care.

Dabigatran, in a review of 9 publications, involving more than 200,000 AF patients over 5 years [10], showed that data for dabigatran in ‘real world’ clinical practice were largely replicative of the main findings in the RE-LY phase III trial [4]. In particular, both dabigatran doses at 150 mg and 110 mg twice daily were associated with lower major extra-cranial bleeding rates than warfarin in patients less than75 years old and similar event rates in those above 75 years old. The 110 mg dose was associated with lower and similar gastrointestinal (GI) bleeding rates, and the 150 mg dose yielded similar and higher GI bleeding rates in patients less than 75 years and more than 75 years old, respectively. A more recent systematic review and meta-analysis studied 348 750 patients (56.65 % warfarin, 40.2 % dabigatran-150 mg and 3.2 % dabigatran-110 mg) in routine care and this is 20 times the size of RE-LY patient population [11]. It included heterogeneous study cohorts regarding history of stroke, hypertension, or diabetes mellitus and did not exclude patients with severe renal impairment (creatinine clearance ?30 mL/min), active liver disease, or conditions associated with an increased risk of bleeding. Patients receiving dabigatran-110 mg in routine clinical practice, tended to be older than patients in the RE-LY trial on this dose. In pooled analyses, dabigatran-150 mg was similar to warfarin in preventing stroke (hazard ratio, 0.92; 95 % confidence interval, 0.84–1.01; P?=?0.066) and a significantly lower risks of intracranial bleeding (0.44; 0.34–0.59; P??0.001). However risks of GI bleeding was significantly higher than warfarin (1.23; 1.01–1.50; P?=?0.041), particularly in studies of older versus younger populations (median/mean age, ?75 versus 75 years; ??=?1.53; 95 % confidence interval, 1.10–2.14; P?=?0.020). Again the findings were consistent with the RE-LY trial. Another observational study provided some insights on the use of Dabigatran 75 mg dose which was not studied in the RE-LY trial but approved in the USA for use in the renal impaired patients with CrCl 15-30 ml/min [12]. The use of dabigatran 75 mg was associated with significantly reduced risk of intracranial haemorrhage and similar rates of stroke, bleeding and mortality compared to warfarin. Interestingly, majority of patients on dabigatran 75 mg twice daily appeared not to have severe renal impairment as only 33 % had a diagnosis of chronic kidney disease, and 20 % of these with severe renal impairment, thus suggesting a possible off-label use of the 75 mg dose in many patients in the real world. While not advocating off label dose reduction, the observational data are nonetheless reassuring. Other findings noted from such real world observations include the observations that new starters of warfarin has a higher bleeding risk when compared to new starters of dabigatran, warfarin experienced switchers or patients remaining on warfarin [13], higher bleeding rates in the first 90 days of treatment in elderly new starters of dabigatran or warfarin [12, 14] and the higher bleeding risk with renal impairment regardless of which oral anticoagulant [15–17].

Rivaroxaban, in the Xantus prospective observational study for 6784 AF patients across 311 centres in Europe reported a lower thrombotic and bleeding rates for Rivaroxaban compared to its Rocket phase III clinical trial [18]. While the phase III ROCKET AF trial did not include patients with CHADs score of 0-1, Xantus had 12.7 % of patients had a CHA2DS2-VASc score of either 0 or 1. Generally, patients in Xantus had lower stroke risks, with a mean CHADS2 score of 2.0 and 19.0 % experiencing prior stroke/TIA or SE, compared with 3.5 and 55 % respectively in the Rocket AF trial. The overall bleeding incidence of 2.1 per 100 patient-years in Xantus was notably lower than 3.6 per 100 patient-years reported in Rocket AF. Similarly fewer major GI bleeds and ICH were observed in Xantus when compared to Rocket-AF. Recent analysis from the Dresden Registry [19] showed the overall rates of stroke and systemic embolism at 2.03/100 patient-years in the intention-to-treat analysis and 1.7/100 patient-years in the on-treatment analysis which were considerably lower than those in the ROCKET AF trial [5]. In addition, event rates for patients receiving 20 mg OD (1.25/100 patient-years), was considerably lower than patients on 15 mg OD (2.7/100 patient-years). Bleeding complications associated with rivaroxaban was addressed in a meta-analysis of 9 studies involving 51,533 patients in real world [20]. It showed the mean pooled rates of any major bleeding, major GI bleeding or ICH with rivaroxaban were 3.32, (95 % CI¼2.28–4.25); 2.41, (95 % CI¼1.25–3.56) and 0.40, (95 % CI¼0.17–0.74) events/100 patient-years. The pooled real-world rates of these bleeding rates largely mirrored those reported for rivaroxaban in the phase 3 ROCKET AF trial [5]. However, there were significant variability and heterogeneity variability in major bleeding rates across the studies. Five studies were retrospective claims analyses and identified bleeding using International Classification of Diseases–9/10 codes, while four were prospective registry studies and identified bleeding clinically using the International Society on Thrombosis and Haemostasis (ISTH) definition. Major bleeding rates and major GI bleeding rates as per 100 patient-years in studies that relied on claims were (2.86 to 12.79) and (2.53 to 9.5) respectively and these are substantially higher than (0.96 to 3) and (0.19 to 0.9) as reported in prospective studies using clinical identification. Such differences underscore the substantial heterogeneity across the studies and the inherent weaknesses associated with retrospective and observational studies. Nonetheless, the pooled rates of major bleeding with rivaroxaban estimated were generally low and consistent with those reported in ROCKET AF. This finding of a lower major bleed rate than Rocket AF was also seen in a large US study of electronic medical records of 27 467 patients (2.9 events per 100 patient-years) [18] as well as an earlier report from Dresden NOAC Registry involving 1200 AF patients treated with rivaroxaban(3.1 events per 100 patient-years) [21, 22].

Besides these published large scale real world comparisons of a single DOAC versus warfarin, evidence relating to the overall effectiveness and safety of all oral anticoagulant drugs used in clinical practice is emerging. An observational nationwide cohort study in Denmark had involved 61, 678 patients with non-valvular AF who were naïve to oral anticoagulants and had no previous indication for valvular AF or VTE [23]. The study population was distributed according to treatment type with 57 % warfarin, 21 % dabigatran 150 mg, 20 % on rivaroxaban 20 mg, and 10 % apixaban 5 mg. The baseline characteristics of patients in apixaban and rivaroxaban has more previous strokes, systemic embolism vascular disease and bleeding while dabigatran patients were younger and less renal impaired, warfarin has more patients with vascular disease hypertension, renal impairment, COPD and cancer. During 1 year follow-up, when compared to warfarin, annual rates of ischaemia strokes and systemic embolism were significantly lower for rivaroxaban (hazard ratio 0.83 (95 % confidence interval 0.69 to 0.99), while not significantly different for dabigatran and apixaban.(hazard ratios of 2.8 % and 4.9 % respectively) The mortality risk was significantly lower with apixaban (5.2 %) and dabigatran (2.75 %) when compared with warfarin (8.5 %), but not with rivaroxaban (7.7 %). No significant difference was found between DOACs and warfarin for ischaemic stroke. The bleeding endpoints for rivaroxaban 5.3 % was comparable to warfarin 5 %, while apixaban 2.3 % and dabigatran 2.4 % were both lower than warfarin. The risks of death, any bleeding, or major bleeding were significantly lower for apixaban and dabigatran compared with warfarin. This real world study concluded that all three DOACs seem to be safe and effective alternatives to warfarin in a routine care setting.

Real-life studies have their inherent weaknesses such as non-controlled and heterogeneous patient groups, uncontrolled influence of non-compliance, other concomitant medications and co-morbidities. However, they provide a wealth of data and insight into how DOACs are used in the real world. Despite the reassuring real world data on use of DOACs in routine care, the benefits of DOACs are not applicable to all patients. As shown in a smaller scale study involving 468 patient with AF from the UZ Brussel Stroke Registry, it was found that less than half of real life patients are eligible for therapy with one of the DOACs [24]. Reasons for non-eligibility include concomitant use of antiplatelet agents with apixaban, impaired renal function in dabigatran, concomitant use of rifampicin and anti-fungal drugs and presence of valvular heart diseases. More data are also required for AF patients on DOACs undergoing cardioversion or ablation. There are ongoing trials addressing some of these issues and their results together with more real world data can add more clarity to these limitations.