Abstract
Dabigatran is increasingly being used in clinical practice for the thromboprophylaxis in atrial fibrillation as a convenient therapy that needs no drug level monitoring. However, analysis of the data of the same clinical trial that led to the adoption of dabigatran in fixed-dosing regimens has indicated a small subgroup of patients that could be either over-treated, risking bleeding, or under-treated, risking embolism. Additional post-marketing data lends support to the favorable therapeutic profile of dabigatran but at the same time raises doubts about patient characteristics such as weight, age, renal function and their pharmacokinetic effects that, in some cases, could be serious enough to expose a minority of patients to risk. We will present a clinical case of a patient with an ischemic stroke while on dabigatran that was found with low dabigatran plasma levels and we will discuss the currently available data on the effects of inherent patient characteristics on dabigatran pharmacokinetics, the clinical impact of dabigatran plasma levels on safety and efficacy as well as the possibility of improving the risk–benefit profile of this agent by tailoring the dose for selected patient groups.
Introduction
Novel or non-vitamin K oral anticoagulants (NOACs) have been a recent therapeutic breakthrough for the prophylaxis of stroke and systemic embolism in patients with nonvalvular atrial fibrillation (NVAF). Three NOACs are currently in clinical use worldwide for NVAF: dabigatran, rivaroxaban and apixaban [Lip et al. 2012] and a fourth, edoxaban, has been approved recently by the Food and Drug Administration (FDA) for clinical use in the United States [FDA, 2015] and awaits approval by the European Medicines Agency (EMA) [EMA, 2015a]. Dabigatran is a direct thrombin inhibitor and the first NOAC that has been approved for NVAF in view of the results of RE-LY trial [Connolly et al. 2009] that provided evidence regarding the similar efficacy and better safety profile of dabigatran versus warfarin. The RE-LY study was a randomized open-label study that allocated patients with NVAF to dose-adjusted warfarin to target international normalized ratio (INR) 2–3, dabigatran 110 mg bid (dabi 110) and dabigatran 150 mg twice daily (dabi 150) irrespective of patient characteristics such as weight, age and renal function. Patients with severe renal failure (estimated creatinine clearance less than 30 ml/min) were excluded since dabigatran is primarily cleared by the kidneys. After a follow up of 2 years, both dabigatran doses were found to be noninferior to warfarin for the prevention of stroke and systemic embolism (which has been the primary efficacy outcome). The dabi 150 regimen was superior to warfarin for the primary efficacy outcome with similar bleeding risk, whereas dabi 110 was similar to warfarin for the primary efficacy outcome with lower bleeding risk. Absolute risk reductions were less than 1%/year. Patients on dabi 150 mg presented with a 0.56%/year absolute risk reduction in strokes compared with warfarin (annual absolute reduction of 0.28% for hemorrhagic strokes and 0.28% for ischemic strokes). Patients on dabi 110 mg had an absolute reduction of 0.65%/year for major bleeding. Subgroup analysis did not found any interaction of renal function or body weight with the treatment effect of dabigatran for both doses. However, there was no data for very obese patients nor for high glomerular filtration rate (GFR) values as all patients weighing over 100 kg were treated as a single group and so were patients with an estimated GFR over 80 ml/min.
Case presentation
We present the case of a 62-year-old female patient who presented with a recurrent ischemic stroke while being treated with dabi 150. The patient had her first stroke a year before and diagnostic evaluation disclosed NVAF as the underlying cause of cerebral infarction. She was started on warfarin that has been shifted to dabi 150 after a transient ischemic attack (TIA) 6 months later. Dabigatran seemed to confer a better secondary prevention of stroke and systemic embolism compared with warfarin in the RE-LY trial, a benefit that has been driven by fewer hemorrhagic strokes rather than fewer ischemic strokes; see Table 1. The patient weighed 116 kg and had a body mass index of 40 kg/m². She was on atorvastatin, atenolol and chlorthalidone for dyslipidemia and hypertension. It should be noted that the patient was seen in consultation 2 months before the index event and dabigatran levels were checked because of obesity which, according to our previous experience, could result in dabigatran failure [Safouris et al. 2014] as we will discuss later. Trough dabigatran levels were found to be very low (10 ng/ml) using Hemoclot, a diluted thrombin time (TT) coagulation test that has been validated for measuring dabigatran levels [Stangier and Feuring, 2012]. The patient was scheduled for a short follow-up appointment in order to discuss alternate oral anticoagulation options but did not present to consultation. At admission the patient had prolonged activated partial thromboplastin time (apTT; 41.6 seconds, normal <33 seconds, 1.26 upper limit of normal) and TT. TT, which is a much more sensitive and linear to dabigatran levels anticoagulation assay [Dager et al. 2012], was found to be equally elevated (89.5 seconds, normal <21 seconds), excluding patient noncompliance that constitutes a major cause of anticoagulation failure [Gallagher et al. 2008]. Diffusion weighted imaging showed two acute distinct ischemic lesions (Figure 1) excluding a lacunar infarction [Wardlaw et al. 2013], a possibility for any hypertensive patient, whereas full diagnostic work-up failed to identify additional to NVAF potential underlying causes of recurrent stroke. The medication list of the patient contained no drugs that could possibly lower dabigatran levels. We have calculated the creatinine clearance of the patient using the Cockroft–Gault formula and found it supraphysiologic at 146 ml/min (normal 88–128 ml/min). In conclusion, after excluding patient incompliance, other cause for stroke and drug interactions, we concluded that low dabigatran levels are probably the cause of dabigatran failure and low levels are caused by patient demographic characteristics. The patient suffered from the first stroke while being treated with warfarin, a fact that did not permit us to shift back to warfarin. Instead, we considered shifting the patient to a NOAC that would provide therapeutic levels at the approved dose.
Annualized rate of events in the RE-LY trial in patients that had already presented stroke or transient ischemic attack.
Diffusion weighted imaging of the brain showing two distinct recent ischemic lesions in the left cerebral hemisphere.
In 2013 a case report was published in New England Journal of Medicine concerning a 51-year-old patient weighing 153 kg that had an ischemic stroke while on dabi 150 because of infratherapeutic dabigatran levels [Breuer et al. 2013]. Infratherapeutic levels were attributed to extreme body weight. We have already proposed that supraphysiologic renal clearance is probably the mechanism to understand dabigatran failure in the very obese because of rapid dabigatran clearance [Safouris et al. 2014]. Almost 80% of dabigatran clearance is renal, in contrast to apixaban and rivaroxaban (27% and 35%, respectively) [Heidbuchel et al. 2013]. We have substituted dabigatran with rivaroxaban 20 mg once daily and after a week we have checked rivaroxaban trough plasma levels with DiXal [Douxfils et al. 2013] and found them to be well within the interquartile range for therapeutic peak and trough levels indicating effective anticoagulation (30 ng/ml, median rivaroxaban trough concentration 25–75th interquartile range 19–60 ng/ml) [Buller et al. 2008].
Weight, age and renal function: their effect on dabigatran levels and clinical outcome
In a post hoc analysis of the RE-LY trial dabigatran levels were inversely correlated with body weight, an association that has been considered of minor importance in the obese but could be relevant in the very obese [Patel et al. 2011]. In an analysis of the RE-LY data, the advantage of dabi 110 over warfarin (reduction of bleeding risk) became statistically insignificant for patients over 75; in the same patient group dabi 150 showed a tendency for greater risk of major bleeding. This was due to the fact that both doses were associated with higher risks of extracranial bleeding (mostly gastrointestinal bleeding) for patients older than 75 (annual absolute risk increase for extracranial major bleeding 0.66% for dabi 110 and 1.24% for dabi 150) [Eikelboom et al. 2011]. The authors tested the hypothesis of whether age-related renal function decline led to higher dabi levels that resulted in more bleeding but no significant interaction between renal function and major bleeding was found, suggesting other and yet unknown age-related factors that predispose older patients on dabi to gastrointestinal bleeding. On the other hand, another post hoc analysis of the RELY trial showed that no significant heterogeneity exists for the primary efficacy outcome across three subgroups of patients depending on renal function estimated with the Cockcroft–Gault formula (estimated GFR 30–50 ml/min, 50–80 ml/min and >80 ml/min) [Hijazi et al. 2014]. The RE-LY study was further scrutinized by transforming ischemic and hemorrhagic events to ischemic stroke equivalents; both dabi doses reduced ischemic stroke equivalents by 1%/year compared with warfarin; the higher dose by reducing ischemic stroke and the lower dose by decreasing major bleeding [Eikelboom et al. 2013]. The dabi 110 regimen seemed to provide greater overall benefit for patients at higher risk for ischemic stroke (as these patients are also at greater risk for bleeding) and dabi 150 greater benefit for those at lower risk for ischemic stroke.
Explaining the difference in Europe and USA dosing
On the grounds of solely one phase III trial, dabigatran has received approval for thromboprophylaxis in NVAF patients at risk for stroke with quite distinct dosing in the United States and in Europe. The FDA was the first to grant approval with an accelerated process due to the novel character of dabigatran (being the first oral anticoagulation approved for NVAF in more than half a century). The FDA only approved dabi 150 for clinical use considering that it was the only dose that showed superiority in preventing stroke and systemic embolism compared with warfarin and that dabi 110 is actually a less efficacious compromise that could be preferred by clinicians, leading to suboptimal thromboprophylaxis for patients [Unger, 2010]. Dabigatran 75 mg was granted approval for patients with severe renal failure based on pharmacokinetic modeling despite the fact that this dose has not been tested in any clinical trial for NVAF and that the pharmaceutical firm was clearly against this decision. Both choices of FDA have been heavily criticized [Moore et al. 2014].
The EMA followed a completely different approach and reached a quite different conclusion. First, the EMA demanded and received therapeutic levels for dabigatran. A proposed optimal trough concentration upper limit at 200 ng/ml was incorporated in the summary of the product characteristics [EMA, 2015b]. Finally, no compulsory testing for dabigatran levels has been demanded, which has been considered as a paradox by some [Moore et al. 2014]. EMA thought that dabi 110 had a place in clinical use and proposed the lower dose for patients older than 80 or at higher bleeding risk. High bleeding risk has not been formally defined, leaving this indication subject to the clinician’s judgment. HAS BLED score (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile International Normalized Ratio, Elderly, Drugs/Alcohol) has been proposed as a clinical tool to aid clinical decision making; a score equal or greater than 3 signifies increased bleeding risk [Apostolakis et al. 2012]. No other dose has been approved and the use of dabigatran has been restricted to patients with estimated GFR over 30 ml/min.
Drug interactions
Drug interactions with dabigatran are not considered a major clinical issue. Surely, compared with warfarin that has significant interactions with many drugs and foods, dabigatran has only a handful of major drug interactions. However, INR control is a means of adjusting to therapeutic levels the anticoagulant effect of warfarin thus correcting any interaction and, on the other hand, dabigatran presents a number of minor drug interactions that could have clinical repercussions in some patients. Heidbuchel and colleagues have published a list of interactions and recommendations for adapting NOAC dosing [Heidbuchel et al. 2013]. First of all, there are some drugs that are contraindicated while on dabigatran: dronedarone and antifungals (itraconazole, ketoconazole, posaconazole and voriconazole); these drugs can be used in the US with a reduced dabigatran dose at 75mg twice daily. Second, there is a group of drugs which, if indispensable, should deter the clinician from prescribing dabigatran: older antiepileptic drugs (carbamazepine, phenytoin and phenobarbital), rifampicin and St John’s wort all significantly reduce dabigatran levels. Third, concomitant verapamil use is allowed only with dabi 110; it should be noted that a patient over 80 years of age that should receive dabi 110 is prone to present high dabigatran levels if they are also on verapamil; moreover, creatinine clearance of 30–50 ml/min and concomitant verapamil treatment is an indication for dabi 75 in North America. Finally, there is list of medications that have minor interaction with dabigatran and the concerned clinician should stay up to date for new information as clinical use is expanding. Among these, amiodarone is of particular importance due to its common use in NVAF. Amiodarone has been reported to raise dabigatran levels by up to 60% [Liesenfeld et al. 2011] which is considered of minor importance and does not justify dose reduction but could be significant for patients with higher dabigatran trough concentrations.
Another issue of major importance is the pharmacodynamic interactions of dabigatran, concomitant medical treatment that predisposes to adverse events (bleeding) without effect on dabigatran levels. In a case series of bleeding events while on dabigatran, one common prescribing error has been starting dabigatran before INR falls below 2 when shifting from warfarin [Pfeilschifter et al. 2013]. In addition to prescribing errors during shifting, the most important pharmacodynamic interaction is concomitant antiplatelet use. It is known that adding aspirin on top of warfarin almost doubles and adding aspirin and clopidogrel more than triples bleeding risk [Hansen et al. 2010]. According to the European Heart Association guidelines [Lip et al. 2014b] adding an antiplatelet on top of anticoagulation is considered in NVAF patients after an acute coronary syndrome or percutaneous coronary intervention up to 12 months after the index event; on the contrary, double or triple antithrombotic treatment is discouraged in stable coronary artery disease. However, in clinical practice there is a tendency to prolong indefinitely double antithrombotic therapy with little or no benefit over oral anticoagulation and significantly greater bleeding risk [Olesen et al. 2012]. A post hoc analysis of the RE-LY study showed that dabi 110 plus antiplatelet treatment retained lower risk of major bleeding as compared with dabi 150 or warfarin plus concomitant antiplatelet therapy [Dans et al. 2013], rendering dabi 110 with antiplatelet therapy an attractive option when aspirin or clopidogrel is required in NVAF patients.
Dabigatran better in real-life than in RE-LY?
An important contribution to understanding the efficacy and risks of dabigatran came from the publication of the Danish nationwide cohort of patients on dabigatran [Larsen et al. 2013]. The Danish Registry of Medicinal Product Statistics was used to compare a cohort of patients with NVAF on dabigatran to an historical cohort of patients with NVAF on warfarin. In this clinical registry in a country with very efficient INR control (mean time in therapeutic INR range (TTR) 72% in the RE-LY study and more than 65% in clinical practice [Holm et al. 2002]) mortality was lower with both dabigatran doses. Gastrointestinal bleeding was not more frequent in patients on dabi 150 and was lower in the dabi 110 group. Compared with the RE-LY study group, the Danish cohort had a lower mean stroke risk. The real difference however between the two cohorts were the prescribing guidelines: in the RE-LY trial dabi 150 and dabi 110 were blindly prescribed regardless patient age and renal clearance whereas in Denmark the EMA recommendations were followed: dabi 110 was prescribed in patients over 80 years according to the EU license with local Danish guidelines setting a lower limit at 75 years; consequently, only 18% of patients on dabi 150 were over 75 and just 2% over 80. In conclusion, dabigatran not only preserved noninferiority to warfarin in a cohort with tightly controlled INR but was associated with reduced mortality. A possible explanation for this unexpected finding is the use of the lower dose for the elderly, according to the EU license and contrary to the RE-LY trial and the FDA approval.
The American post-marketing experience
Soon after the introduction of dabigatran in clinical practice in US, a number of post-marketing reports of bleeding stimulated heated discussions in medical publications as well as mainstream media questioning the benefit–risk profile of dabigatran. This controversy lead to a FDA initiative through Mini-Sentinel, a pilot post-marketing surveillance system. A preliminary report [FDA, 2012] and a publication showed that bleeding risk with dabigatran was actually lower than with warfarin both for intracranial and gastrointestinal hemorrhage, attributing the large number of reported bleeding events as a ‘a salient example of stimulated reporting’ [Southworth et al. 2013]. Controversy persisted [Institute for Safe Medication Practices, 2013] and so did post-marketing surveillance with Mini-Sentinel. A more recent (2014) update from the FDA, based on a much larger and older population than in the previous report, confirmed an increased risk of major gastrointestinal bleeding with dabigatran and a lower risk for ischemic stroke [FDA, 2014]. These results were consistent with the conclusions of the RE-LY trial. Moreover, dabigatran showed a statistically significant reduction in mortality (hazard ratio 0.86; 95% confidence interval [CI] 0.77–0.96) after adjustment for multiple confounding variables.
Does the safety and efficacy of dabigatran depend on drug levels?
Recently, a substudy of the RE-LY trial showed that dabigatran levels were highly variable and were related to renal function, age, weight and sex (female) [Reilly et al. 2014]. More specifically, higher dabigatran levels were strongly associated with bleeding risk (p < 0.0001) and lower levels had a weaker but still statistically significant correlation with ischemic stroke (p = 0.045). Cox regression analysis showed that the relative risk of major bleeding doubled at a trough concentration of 210 ng/ml compared with median trough levels (88 ng/ml) and that the relative risk of ischemic stroke or systemic embolism was increased by 50% at 28 ng/ml compared with median trough levels (59 ng/ml). The authors provided probability curves of clinical outcomes versus dabigatran plasma concentrations: a trough level at around 50 ng/ml for a 75-year-old patient (probably more in the older and even less in the younger patient group) seemed sufficient for prophylaxis and increasing levels did not confer any major additional benefit. In contrast, higher trough concentrations led to higher risk of bleeding with a steeper curve than ischemic events. Now, this point is of particular importance in selecting between the EMA and the FDA dosing recommendations: the ‘conservative’ EMA strategy seems to be equally efficacious and less dangerous than the ‘aggressive’ FDA strategy. This could also explain why the American post-marketing experience confirms the results of the RE-LY trial with the additional benefit for mortality, whereas in the Danish registry dabi 150 did not lead to more gastrointestinal bleedings than warfarin possibly by limiting its use in the elderly in favor of dabi 110. The EMA has reviewed 183 spontaneously reported cases of fatal bleeding on dabigatran for which posology, age and renal function were available and found 40 cases of patients on dabi 150 that would receive dabi 110 if EMA guidelines were followed (38 of them from the USA) [EMA, 2012]. Last but not least, a post hoc analysis of the RE-LY study data including 18,000 patients equally distributed to dabi 110, dabi 150 and warfarin, aimed to estimate the clinical benefit of applying the European prescribing guidelines to the RE-LY population [Lip et al. 2014a]. Only half of the RE-LY patients were treated according to EMA guidelines, one third of dabi 110 and two thirds of dabi 150 patients. ‘EU label simulated dabigatran treatment’ showed superior efficacy and safety compared with warfarin: there has been not a single efficacy or safety endpoint favoring warfarin; major gastrointestinal bleeding, myocardial infarction and ischemic stroke were similar and all the rest were significantly in favor of dabigatran (stroke and systemic embolism, hemorrhagic stroke, death, vascular death, major bleeding, life-threatening bleeding and any bleeding). Absolute risk reduction for dabigatran compared with warfarin was marginal with numbers needed to treat (for statistically significant hazard risk reductions) ranging from 45 for any bleed to 336 for hemorrhagic stroke. However, it is important to realize that increased gastrointestinal bleeding, which has been considered the Achilles heel of dabigatran, seems to be no longer a problem when EMA guidelines are applied.
Could we hope for even better results by tailoring dabigatran dose?
Interestingly, the EMA has already discussed this issue in a workshop in December 2014 [EMA, 2014]. Taking into account the reported dabigatran plasma concentrations in the RE-LY study, the workgroup raised the question of whether adjusting dabigatran dose to maintain a therapeutic level at 75–150 ng/ml or 75–100 ng/ml could provide a better benefit–risk ratio; the answer was that further research is necessary based on the fact that even unmonitored dabigatran works well. In the RE-LY trial, 10% of patients on dabi 110 had a trough plasma concentration lower than 28 ng/ml and another 10% were above 155 ng/ml. As for dabi 150, 10% were below 40 ng/ml and another 10% above 215 ng/ml [Reilly et al. 2014]. This means that one in ten patients on dabi (110 or 150) were at 50% increased risk for ischemic stroke and systemic embolism compared with the mean and one in ten patients on dabi 150 (but not dabi 110) were at double relative risk for major bleeding compared with those with mean trough dabigatran levels, as mentioned above. Since in the US only dabi 150 is available, as used in the RE-LY, it means that in up to 20% of patients on dabi 150 optimal anticoagulant effect is not achieved. As Reilly and colleagues point out, renal dysfunction is a risk factor for bleeding in anticoagulated patients but the relative risk of bleeding of dabigatran versus warfarin did not change in the RE-LY trial; it seems that low or high dabigatran levels are not by themselves a risk factor but in combination with old age, renal failure or low body weight may actually predispose to embolism or bleeding, respectively. On the other hand, dabi 110 is prescribed in Europe for the elderly patients and the aforementioned results of the RE-LY cannot be extrapolated; most probably the fact that dabi 150 is reserved for younger and dabi 110 for older patients means that the percentage of patients at risk (presenting with low or high trough levels) is much smaller. There is no data concerning dabigatran levels in a NVAF population on dabigatran according to the EMA guidelines.
A more elaborate effort has been undertaken by the manufacturer Boehringer Ingelheim comparing dabi 150 as prescribed in the United States with a hypothetical model in which dosing would be adapted to a trough plasma concentration of 90–140 ng/ml [Boehringer Ingelheim, 2012]. This model would require lowering the dose for more than half the patients on dabi 150, with half of them switching to dabi 110 (those with trough concentrations 90–140 ng/ml) and the other half to dabi 75 (trough concentrations >140 ng/ml). This adaptation could mean 20% less major bleeding (less than 1% absolute risk reduction per year) with no significant difference in stroke risk, which was considered of minor importance. Some argued that this difference is indeed important [Moore et al. 2014] but this is clearly an exaggeration since the adaptation of such a model would mean shifting half the patients from doses tested in clinical trials and checked in post-marketing registries with reassuring results to dosing schemas that are based on theoretical models in order to attain a marginal effect on the benefit–risk profile; on the other hand, a RE-LY 2 trial to test this hypothesis would be too costly and time consuming. It should be noted that this model has been discussed in 2012, well before the publication of the effect of dabigatran concentration on clinical outcomes. In the light of the aforementioned study, it is obvious that the real problem with this model is the proposed therapeutic limits that were most probably too narrow leading to dose adaptation in half the patient population. It would be very interesting to test a model for the 10% of patients in the RE-LY trial with very low trough levels and, even more importantly, the 10% of dabi 150 patients with very high trough levels, above 215 ng/ml.
What is the optimal management of patients with very low or very high trough dabigatran levels?
There are many different answers depending on the clinical scenario and the solution appears to be slightly easier for patients with high levels than those with low levels. Supposedly a patient on dabi 150 is found to present more than 200 ng/ml, he or she could easily be shifted to dabi 110 which is a dosage tested on clinical trials and available for clinical use in most countries (but not in the US) for NVAF. In fact, the EMA already proposes this solution despite the fact that it does not recommend regular testing as discussed above. So, when a validated test such as Hemoclot is available, elderly patients with moderate renal failure or low body weight that are under 80 years old and qualify for dabi 150 according to the EMA guidelines could benefit from Hemoclot testing and shifting to a lower dose if trough levels are found to be higher than 200 ng/ml. Supposedly the patient remains with high trough dabigatran levels despite being on dabi 110, the answer is slightly more complicated as shifting to a lower dose is not possible (dabi 75 for NVAF is only approved in the US) and there is no proof that shifting the patient to warfarin or another NOAC would result in lower bleeding risk; still such an attitude could be considered, especially for patients at high bleeding risk. The answer gets even more complicated for patients with low dabigatran levels. If they are on dabi 110, shifting to dabi 150 could be possible only if they are under 80 since age over 80 is a formal contraindication to dabigatran 150 for the EMA; otherwise they should be shifted to alternate oral anticoagulation. If they are on dabi 150, shifting to other oral anticoagulant is probably inevitable since there is no greater dabigatran dose approved for any indication worldwide. Even when dabi 300 mg bid was tested in a phase II trial in patients with mechanical valves (RE-ALIGN study) [ClinicalTrials.gov identifier: NCT01505881], a total of 8% of the study population failed to attain trough levels higher than 50 ng/ml. This was partially due to the fact that median age in RE-ALIGN was much lower than in RE-LY (56 versus 71 years) and median creatinine clearance was more than 105 ml/min, explaining why a high percentage of patients presented low levels of a primarily renally cleared molecule. Still, dose escalation would not necessarily mean optimal trough levels in every patient. Needless to say, when shifting from dabigatran to other oral anticoagulation, ideally it should be to an oral anticoagulant that its anticoagulant effect or its levels could be measured. A simplified approach of the aforementioned proposals is presented in Figure 2.
Schematic representation of a proposed dosing algorithm for dabigatran based on the EU licence. If on verapamil, dabigatran 150 mg (dabi 150) is not recommended. The upper limit for eGFR at 120 ml/min is arbitrary; a supraphysiologic GFR should prompt the clinician to check for dabi levels. In ‘grey zones’, e.g. eGFR 30–80 ml/min and age <75 or eGFR 50–120 ml/min and age 75–80, decision should individualized based on patient’s bleeding risk. If on concomitant antiplatelet therapy or HAS BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile International Normalized Ratio, Elderly, Drugs/Alcohol) score >2, opt for the lower dose, especially if on amiodarone. Further accuracy necessitates measuring Ctrough using a validated method such as Hemoclot especially for patients with very low or very high body weight. In case Ctrough < 48 ml/min, escalate dosing and recheck dabigatran levels or, if on dabi 150, consider alternate oral anticoagulation. In case Ctrough > 200 ng/ml, lower dosing and recheck dabi levels or, if on dabi 110, consider alternate oral anticoagulation. For discussion, see the text.
What are the arguments against measuring dabigatran levels?
First of all, in contrast to warfarin, there are no widely available sensitive and reliable laboratory methods for measuring the anticoagulant effect of dabigatran. Could activated partial thromboplastin time (apTT) provide a solution? In a dose-guiding trial, 10% of atrial fibrillation patients on dabi 150 had an apTT prolongation more than twice the upper limit of normal [Stangier et al. 2005] suggesting increased bleeding risk [van Ryn et al. 2010]. Still, there is a total absence of studies correlating apTT with bleeding risk in patients on dabigatran and it is a well-known fact that apTT is a crude measure of dabigatran’s anticoagulant effect and that is imprecise especially at high dabigatran concentrations [Stangier et al. 2008]. A recent systematic review on the different measurements of the anticoagulant activity of NOACs highlighted the different apTT results depending on the reagent used and proposed that laboratories should calibrate their apTT methods with dose–response studies [Cuker et al. 2014]. For example, the least sensitive apTT reagents required a dabigatran concentration as high as 400 ng/ml in order to produce a twofold apTT prolongation [Harenberg et al. 2012]. Prothrombin time shows an even greater variability depending on the method used [Helin et al. 2013]. A diluted thrombin time, Hemoclot [Stangier and Feuring, 2012], is commercially available and presents a linear response to dabigatran plasma levels but availability is an issue. In a survey in Australia and New Zealand only 1.5% of laboratories reported using it [Favaloro et al. 2013].
Conclusions
With the accumulation of data from clinical registries, case reports and post hoc analyses of the RE-LY trial, clinicians can be reassured that dabigatran is a safe and effective warfarin alternative for NVAF patients. Using both dabigatran doses tested in the RE-LY and reserving the lower dose for the elderly and those at bleeding risk appears to be a step forward from the RE-LY-based recommendations proposed by the FDA. Time will show whether a step further forward will be undertaken by tailoring dabigatran dose or precluding dabigatran from patients that present with very low or high plasma concentrations. Based on patient characteristics, the prototypical patient at risk for bleeding from dabigatran is a frail elderly patient with low body weight and moderate renal insufficiency and the patient at risk for embolism is a very obese younger patient with supraphysiologic renal clearance. For the time being the real challenge is to detect such patients based on their age, weight and renal function or, when available, through laboratory testing. Then an adaptation of dabigatran dose or shifting to other oral anticoagulation could assure optimal anticoagulant effect in the entirety of the NVAF patient spectrum. These steps could eventually lead from noninferiority of dabigatran compared with warfarin to clear superiority and greater benefit for NVAF patients.
Take home messages
More than 80% of patients on dabigatran in the RE-LY trial had optimal dabigatran plasma levels; in this particular patient group dabigatran could have had a better risk–benefit profile than warfarin.
EMA dabigatran prescribing guidelines are preferable to those proposed by FDA as they probably maintain efficacy with lower bleeding risk and most probably expand the percentage of patients with optimal dabigatran levels.
Patients at the extremes of dabigatran plasma concentration and either of advanced age or at the extremes of body weight or creatinine clearance could be at risk for hemorrhagic or thrombotic events; these patients could benefit from dose modification or other oral anticoagulation but no formal recommendations exist.
Low availability of dabigatran concentration measurement testing in clinical practice is the main obstacle in identifying patients at the extremes of plasma concentrations. Patient characteristics can point to the patients at risk: elderly low-weight patients with moderate renal failure are at risk of excessive dabigatran exposure and bleeding; younger obese patients with high creatinine clearance are likely to have subtherapeutic dabigatran levels and therefore are at risk of embolism.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr Georgios Tsivgoulis has been supported by European Regional Development Fund - Project FNUSA-ICRC (grant number CZ.1.05/1.1.00/02.0123) for drafting and revising the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.