Dabigatran: how the drug company withheld important analyses
BMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g4670 (Published 23 July 2014) Cite this as: BMJ 2014;349:g4670Concerns over data in key dabigatran trial
Dabigatran, bleeding, and the regulators
The trouble with dabigatran
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Whatever happened to medical common sense?
Four new oral anticoagulants (NOAC) are available in Japan: Pradaxa (Prazaxa in Japan, anti-thrombin inhibitor, Dabigatran, Boehringer Ingelheim) and three anti-factor Xa inhibitor, Xarelto (Rivaroxaban, Bayer), Eliquis (Apixaban, Pfizer) and Lixiana (Edoxaban, Daiichi Sankyo).
NOAC is used instead of warfarin for the prevention of stroke in atrial fibrillation. Selling point of them is “one dose fits all approach” without monitoring and the titration like warfarin.
However a simple question arises; when hypertensive patients were given only one fixed dose drug and drug companies ensured that blood pressure would be normalized without monitoring blood pressure or diabetic patients were told that blood glucose level could surely be well controlled by the fixed dose oral glucose lowering agent without the adjustment by the monitoring of the glucose level or hemoglobin A1c, few believe these advertisements that appear too good to be true. Why did doctors, patients and regulators believe these drugs to be used without monitoring?
If the blood concentration of the drug varies in a considerable way like Dabigatran, there is a good likelihood of extremely higher and lower concentration, which lead to the bleeding and ischemic stroke. The pharmaceutical company should not sell NOAC and the authorities should not approve until the monitoring of the concentration and the activity can be easily estimated in the hospital.
In the Dabigatran case, DTT assay (the Hemoclot test) is available in EU. Thus, also in both US and Japan, the company should have started selling dabigatran after having prepared the assay kit of Dabigatran. The kit now should be equipped as rapidly as possible. The (pretreatment) guideline is useful for patients’ selection, however once Dabigatran treatment begins, there is no other way to prevent patients other than the monitoring. In addition, there is no available antidote, a neutralizing agent. Moreover, verapamil, which is frequently used drug for atrial fibrillation, increases the concentration.
All NOACs at first in the real clinical world should be monitored by the blood concentration or the activity. It should be mandatory at least for high risk patients (elderly patients, patients with impaired renal function or low body weight). After a while, if the monitoring is scientifically proven to be unnecessary, the drug without monitoring should be recommended. If this doesn’t happen, the reputation of drugs, which actually may do much good, will be jeopardized. In reality, what I feared the most may be happening.
Competing interests: No competing interests
1) Scientific grounds for major bleeding over aPTT 80s in RE-LY: Where did aPTT 80s come from?
Doctors in Japan have been told by BI (Boehringer Ingelheim) that aPTT 80s is the good index of Dabigatran for preventing major and life-threatening bleeding, and also that this had been obtained from RE-LY trial (N Engl J Med 2009;361:1139-51).
In fact, in the Japanese prescribing guide of Dabigatran (revised April 2013) made by BI, the following description is found; In the phase III international collaborative trial (RE-LY) including Japanese patients, lots of major bleedings were seen in patients at trough aPTT>80s. So, aPTT>80s in Dabigatran has been believed to be a very important guide. However, recently I have found the contradictory descriptions between the published paper and the official document. In Table 4 of Thromb Haemost 2012, 107; 838-847 including BI employees as authors, authors described that from the analysis of RE-LY and RE-NOVATEII, aPTT>80s (2-3 x baseline value) at trough increased the risk of bleeding. On the contrary, in CTD (common technical documents, 5.3.5.1-4. Page37, in Japanese) which was submitted by BI to the regulatory authority in each country, the following sentence was described: It was impossible to set an upper limit for aPTT to avoid bleeding events in RE-LY. This means that the analysis to determine the upper limit of aPTT to prevent bleeding events in RE-LY was a failure.
Recently, I asked BI (Germany) about the grounds for aPTT 80s. BI corporate medical team replied that after 150mg BID the 90th percentile of aPTT at trough was 76.4s, and it’s purely descriptive representation of the percentiles. The cut off at 90th percentile is commonly used by regulators for descriptive representation of data in label documents. The selected aPTT prolongation of 80s (2-3x baseline) is to some extent arbitrary. aPTT value of 80s is not a definition of therapeutic range. It does not consider efficacy, and it is not intended to define an aPTT value that optimizes benefit-risk in the overall patient population. The aPTT test has limited sensitivity and is not suitable for precise qualification of anticoagulant effect, especially at high plasma concentrations of Dabigatran. High aPTT values should be interpreted with caution. An analysis of RE-LY data submitted to authorities showed an increasing but non-significant trend between trough aPTT and major bleed rate. On the other hand, FDA analysis identified both ECT and aPTT at trough as significant predictors of life-threatening bleeds (FDA Briefing Information, Dabigatran Etexilate Mesylate Capsules, for the September 20, 2010 Meeting of the Cardiovascular and Renal Drugs Advisory Committee. page 27. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMateria....).
Figure 4 at page 27 shows aPTT and the probability of a life-threatening bleed within 1 year in RE-LY. The bars of the plot region represent the 10th to 90th percentiles. However, from this data, it is impossible to use aPTT 80s (2-3 x baseline value) at trough as a target to avoid bleeding events. As the percentile increases, the risk increases slowly but continuously. Indeed aPTT 80s is an important value in daily clinical practice, but this value has been used for heparin. aPTT 80s may not hold true for Dabigatran in real clinical practice. It may be lower than 80s. If BI wanted to use aPTT to avoid major bleeding, the determination of aPTT suitable for Dabigatran is necessary. Judging from no references and no explainable data in Thromb Haemost, the negative description in CTD, and FDA data, I think there is no definite basis employing 90th percentile aPTT (80s) as alarming point for major bleeding as written in the prescribing guide. I have no other choices but to say that aPTT>80s is not derived from RE-LY and is scientifically groundless. aPTT 90th percentile from RE-LY seems to have happened to be at around 80s.
I have asked BI three times about the contents of Thromb Haemost and CTD. However, I have heard nothing from BI.
I can understand that EMA, FDA, and PMDA did not employ a PTT and it is not surprising that there are sentences suggesting aPTT data as groundless evidences in the PMDA review report.
2) Why was the death rate after adverse events so high in Dabigatran compared with Warfarin and Rivaroxaban?
In Table 2 – Direct and manufacturer reports for 3 anticoagulants- 2012 (ISMP, QuarterWatch, October 17, 2013), Dabigatran adverse event cases were reported to be more likely to result in death (TABLE).
By chi-square test, Dabigatran death rate was highly significant. Is this a fact or something like data collection problem?
3) The plasma Dabigatran concentration is not a determinant for major bleeding, and the monitoring is not necessary - a misdirected official view of Japan Boehringer Ingelheim (JBI)
In P.325 of J Am Coll Cardiol (2014; 63:321-8), the relationship between major bleeding and trough plasma concentrations was analyzed by logistic regression analyses with and without covariates. With covariates, logistic regression was significant. C-statistic of the best-fit model was 0.74 and c-statistic without plasma concentrations was 0.68. The Hosmer-Lemeshow goodness-of-fit test was P=0.082.
JBI considers the role of the concentration is reflected in this slight difference, and the concentration doesn’t significantly contribute to the major bleeding. Therefore, they explain that the monitoring of the plasma concentration is not necessary.
However, I think this JBI’s remark seems to be incorrect for the following reasons. In the first place, c-statistic of the logistic-regression model is 0.715. This means the low predictive accuracy. Hosmes-Lemeshow goodness of fit which assesses the data fitting of the model is narrowly P=0.082. The results of c-statistic and Hosmes-Lemeshow goodness of fit show that major bleeding can’t be fully explained by this model. Thus, using this incomplete model, the discussion about the role of the plasma concentration for major bleeding seems to be inappropriate.
In the discussion of this paper (P.326), the following sentence is described; the effect of age on Dabigatran exposure is likely due to the decreasing renal function in the elderly. This means the age increases the plasma concentration. Thus, even if the concentration was excluded from the model, the effect of the concentration still remains through age. This might lead to the minute difference with or without the concentration. Diabetes also have an effect on the concentration by the impaired kidney. I may add I wonder whether the contribution of age will increase when the concentration is excluded from the model.
I don’t think the statistical interpretation of this model serves as a justification that monitoring of plasma concentration is of little avail and unnecessary.
BI insists that the merit of the fixed doses of Dabigatran over warfarin was proven by relative risk (RR) in RE-LY. However, this interpretation of RR is premised on the constancy of the Dabigatran concentration. Warfarin is controlled by INR between 2.0 and 3.0. In fact, from 50 to 0 ng/ml, ischemic stroke/SEE rapidly increases. From 200 to 800 ng/ml, major bleeding also exponentially increases. Therefore, for the moment, the most important purpose of the monitoring is to save both 0-10th percentile and 90th-100th percentile patients.
In addition, hypertension (high SBP) is the most important determinant factor of stroke (both bleeding and infarction). Hypertensive patients are expected to have higher risk for stroke. In the logistic regression analysis, I’m curious to know whether SBP had a significant effect or not.
Competing interests: No competing interests
We would like to respond to comments made by Dr. Yui in three rapid responses to the BMJ article by Cohen [1] on 4, 11 and 16 September 2014.
The author raises questions about why dabigatran is marketed without monitoring of plasma dabigatran concentrations, about the nature of the exposure–response relationships for efficacy and safety of dabigatran, and whether an optimal plasma trough dabigatran concentration could be determined.
The most important reason why dabigatran etexilate is labeled without monitoring is the fact that no monitoring was undertaken in the pivotal RELY trial [2], which formed the basis for registration with over 18,000 patients. In RE-LY, superiority to well controlled dose-adjusted warfarin was achieved for efficacy with the dabigatran etexilate 150 mg twice daily dose, and reduced bleeding was achieved with the dabigatran etexilate 110 mg twice daily dose. Both doses reduced significantly the intracranial bleed rate compared to warfarin. Any putative additional benefit in clinical outcomes with adjusted dabigatran dosing, based on monitoring on the basis of retrospective analysis, would be speculative.
In a recent publication [3], cited by Dr. Yui, it has been hypothesized that, for patients above the 90th or below the 10th percentile of trough plasma dabigatran concentrations for the 150 mg twice daily or 110 mg twice daily doses, respectively, dose adjustment might improve their benefit–risk ratio. Multiple analyses have shown that patient characteristics (such as age [3, 4] and renal function [5] – both of which are correlated with dabigatran plasma level) are critical factors for the benefit–risk assessment of dabigatran. This is already reflected by the inclusion of such factors in treatment labels. The additional impact of plasma levels remains a hypothesis. The favourable clinical profile of fixed doses of dabigatran in comparison to warfarin was reaffirmed by the FDA in one of the largest real-world analyses of its kind, comprising 134,000 patients above 65 years [6].
We understand that the author concludes that BI could not identify a relationship between dabigatran plasma concentration and efficacy based on the logistic regression analyses. The respective analyses have been performed by the FDA and by Boehringer Ingelheim (BI) and indicate an exposure–efficacy relationship of borderline significance [7]. The FDA analyzed the data using a Cox model which nearly reached significance. Although a relationship between exposure and efficacy was not identified when BI used a univariate logistic regression model (containing no explanatory variables beyond concentration), BI subsequently performed a multivariate logistic regression analysis which showed a significant relationship between exposure and efficacy (odds ratio 0.562, 95% CI 0.320, 0.987; P = 0.0448) [3].The minor difference between the P values is explained by the use of different methods. BI explored the data using several different methods. All analyses using different methods reached the same conclusion.
The relationship between exposure and safety and/or efficacy is well described in all submission documents provided to health authorities worldwide. BI never denied these relationships.
Based on decades of oral anticoagulation with vitamin K antagonists, with doses adjusted based on international normalized ratios (INRs), it may be attractive to define target plasma concentration ranges and adjust dosing based on plasma level monitoring for dabigatran. The author estimates the optimal concentration range of dabigatran for both efficacy and safety between 60 and 200 ng/mL. This may have been based on the concentration ranges where HR less than 1.0 in Figure 3 [3]. However, the HR = 1.0 lines in Figure 3 are relative to the responses at median concentrations. Using median concentration as a comparator is somewhat arbitrary and therefore is not suitable for deriving an optimal therapeutic range. Furthermore, if this estimation is based on Figures 2 and/or 3 in reference [3], this estimated range must be limited to patients with the same characteristics as in the graphs, namely a 72-year-old male with prior stroke and diabetes (Figure 2) and a 71-year-old patient with CHADS2 score of 2 (Figure 3).
Extrapolation to other patient types is not warranted. As can be seen in Figure 1 from the same paper [3], differences in age alone have marked effects on both safety and efficacy, and identical concentrations correspond to different risks and benefit–risk balances depending on age. It is this observation that led Reilly et al. [3, 8] to conclude that there is no single plasma concentration range that provides optimal benefit–risk for all patients.
The author suggests that BI should compare the responses observed at a range of dabigatran concentrations against warfarin. However, such an approach is not possible as no adequate comparison group within the warfarin-treated patients can be defined. In order to ensure a fair assessment of the effect of, for example, high dabigatran concentration relative to treatment with warfarin, the warfarin comparator group would need to be sufficiently large and match the dabigatran-treated group. The fixed-dose trough level of dabigatran depends on age, creatinine clearance and other variables, whereas the dose adjustment of warfarin based on INR implicitly adjusts for several, even unmeasured, impacting variables. Therefore, no adequate comparison group within the warfarin-treated patients can be defined for treatment with certain trough levels of dabigatran within the RE-LY database.
Joerg Kreuzer, Therapeutic Area Head Medicine Cardiovascular, Boehringer Ingelheim GmbH, Ingelheim, Germany
Paul Reilly, Global Team Lead, Clinical Development Cardiovascular, Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, CT, USA
Hugo Maas, Pharmacometrician, Boehringer Ingelheim GmbH & Co. KG, Biberach an der Riss, Germany
Herbert Noack, Expert Statistician, Boehringer Ingelheim GmbH, Ingelheim, Germany
1. Cohen D. Dabigatran: how the drug company withheld important analyses. BMJ 2014;349:g4670.
2. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361(12):1139-51.
3. Reilly PA, Lehr T, Haertter S, Connolly SJ, Yusuf S, Eikelboom JW, et al. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY Trial (Randomized Evaluation of Long-Term Anticoagulation Therapy). J Am Coll Cardiol 2014;63(4):321-8.
4. Eikelboom JW, Wallentin L, Connolly SJ, Ezekowitz M, Healey JS, Oldgren J, et al. Risk of bleeding with 2 doses of dabigatran compared with warfarin in older and younger patients with atrial fibrillation: an analysis of the randomized evaluation of long-term anticoagulant therapy (RE-LY) trial. Circulation 2011;123(21):2363-72.
5. Hijazi Z, Hohnloser SH, Oldgren J, Andersson U, Connolly SJ, Eikelboom JW, et al. Efficacy and safety of dabigatran compared with warfarin in relation to baseline renal function in patients with atrial fibrillation: a RE-LY (Randomized Evaluation of Long-term Anticoagulation Therapy) trial analysis. Circulation 2014;129(9):961-70.
6. FDA Drug Safety Communication: FDA study of Medicare patients finds risks lower for stroke and death but higher for gastrointestinal bleeding with Pradaxa (dabigatran) compared to warfarin – 13 May 2014. http://www.fda.gov/Drugs/DrugSafety/ucm396470.htm. Last accessed 17 September 2014.
7. Centre for Drug Evaluation and Research: Clinical Pharmacology Review NDA 22-512, Dabigatran. http://www.accessdata.fda.gov/drugsatfda_docs/nda/2010/022512Orig1s000Cl.... Last accessed 30 September 2014.
8. Reilly PA, Connolly SJ, Yusuf S, Eikelboom JW, Ezekowitz MD, Wallentin L; RE-LY Investigators. Reply: regarding the effect of dabigatran plasma concentrations. J Am Coll Cardiol 2014;63(25 Pt A):2885-6.
Competing interests: Employees of Boehringer Ingelheim
In Figure 3 (Paul A. Reilly et al., The Effect of Dabigatran Plasma Concentrations and Patient Characteristics on the Frequency of Ischemic Stroke and Major Bleeding in Atrial Fibrillation Patients. J Am Coll Cardiol 2014; 63: 321-8), Cox regression analysis of ischemic stroke/SEE and major bleeding versus trough plasma concentration of Dabigatran is presented. Predicted HR (hazard ratio) is shown in comparison with HR=1.0 at 59 or 88 ng/ml. The horizontal axis is the dabigatran concentration.
I’m curious to know what curves will be obtained if Dabigatran is compared with warfarin instead of HR1.0 Dabigatran. In other words, this shows HR of Dabigatran to warfarin as a function of plasma Dabigatran concentration for ischemic stroke/SEE and major bleeding.
I think this analysis is very important, because HR curve of dabigatran vs warfarin will show the concentration range of the superiority or the inferiority of Dabigatran to warfarin for ischemic stroke/SEE and major bleeding.
Competing interests: No competing interests
When such events occur, where pharmaceutical companies withheld important data for whatever reason, the regulatory authorities should consider a harsh penalty to stop this type of behaviour by the companies in future. I think the protection on the patent of that agent should be removed around the world, and that molecule should be open to generic developemnt companies immediately. A settlement of 450 M is nothing when the company is making over a B a year, from that agent. By removing the patent protection, the company is going to lose more money, and likely will be careful in future.
Competing interests: No competing interests
Regarding the confrontation between BMJ and Boehringer Ingelheim (BI)
BI’s assertion is as follows:
As compared with warfarin, the 100-mg dabigatran was associated with lower rates of major hemorrhage and similar rates of stroke and systemic embolism; the 150-mg dose of dabigatran was associated with a similar rate of major hemorrhage but with lower rates stroke and systemic embolism. As there is no single plasma concentration range that provides optimal benefit-risk for all patients, there is no need to monitor drug levels (JACC 2014:63:321-8). The safety compared with warfarin has also been confirmed in post-marketing analysis by the FDA. (FDA Drug Safety Communication: FDA study of Medicare patients finds risks lower for stroke and death but higher for gastrointestinal bleeding with Pradaxa (dabigatran) compared to warfarin – 13 May 2014. http://www.fda.gov/Drugs/DrugSafety/ucm396470.htm Last accessed 24 July 2014).
Surely, Pradaxa may be better than warfarin, but monitoring the concentration can further decrease bleeding events in patients at the extremes of trough concentrations, and could have saved lives. The extremes mean both 0-10th percentile and 90th-100th percentile of the plasma dabigatran concentration. Plasma concentration of Pradaxa varies widely and can’t be anticipated accurately. Therefore, monitoring is necessary to save patients in extremes, which leads to the decrease in both high risk of stroke and high risk of bleeding. In the draft version of JACC written by BI employee (unsealed court documents in Pradaxa case), the same point of view is described: patients with extremely low trough concentrations, e.g. below 35 ng/ml are at higher risk of stroke while trough concentrations above 250 to 300 ng/ml are associated with a higher risk of major bleeding. Based on the concentration ranges achieved with the two doses of dabigatran in RE-LY, less than 20% of patients would be expected to be at the extreme of the plasma concentration distribution. Monitoring of plasma concentrations or antithrombotic activity, e.g. aPTT, would be required to identify these patients. A dose adjustment could improve the benefit-risk ratio. Up to 20% of patients treated with 110 and 150 mg bid doses used in RE-LY may fall outside this range.
Thus, in conclusion, theoretically, monitoring (and titration) Pradaxa concentration is considered to be absolutely necessary for patients’ safety.
Competing interests: No competing interests
The reason why Boehringer Ingelheim (BI) sold dabigatran without monitoring plasma concentration is the key issue. BI explains that they could not obtain the optimal risk benefit concentration range of dabigatran from the RE-LY trial.
From the logistic regression model analysis, they could get a significant exposure-safety relation, but could not get the exposure-response relation.
On the contrary, FDA reanalyzed the data using the Cox model and they got the exposure-response relation with marginal P=0.056.
In Figure 3 (JACC, 2014; 63:321-8), the same result as FDA using Cox model is shown.
From both exposure-response and exposure-safety curve, the optimal range is estimated to be around 60-200 ng/ml.
Then, a question arises. We usually use the Cox model which contains time to event parameter instead of a logistic regression model which lacks this parameter.
I wonder why BI undertook a statistical analysis using a logistic regression model instead of the Cox model. BI presumably must have computed a Cox model analysis. Did BI present the result of the Cox model to the FDA?
Competing interests: No competing interests
The authors accuse Boehringer Ingelheim, the manufacturer of dabigatran, of having failed to share with regulators information on how monitoring plasma levels of the drug and subsequent dose adjustment could reduce the risk of major bleeds. Of the numerous misleading statements made regarding dabigatran etexilate in various articles in the BMJ, (1-5) published on July 23 and subsequently on August 11, 2014, we are compelled to address one that we consider particularly important to correct.
The article by Cohen (1) suggests that Boehringer Ingelheim withheld important information from regulators “about the potential benefits of monitoring anticoagulant activity and adjusting the dose to make sure the drug is working as safely and effectively as possible.” (1)
We would like to reiterate our conclusions on this subject, which had been provided to the BMJ well ahead of the publication of the manuscripts, and to clarify the misleading remarks made in the BMJ. (1-5)
The RE-LY trial demonstrated that dabigatran etexilate 150 mg bid without monitoring and without dose adjustments was superior to standard of care (monitored and dose-adjusted warfarin) in reducing the risk of ischaemic and haemorrhagic stroke in patients with atrial fibrillation, while having a numerically lower rate of major bleeding, and significantly lower rates of life-threatening, intracranial and any bleeding. (6,7)
Data on plasma levels of dabigatran from the RE-LY study, as well as respective PK modelling, were provided to the US and European regulatory authorities and included in their discussions when reviewing dabigatran in 2010 and 2011. While the benefits of dabigatran compared to warfarin were striking as detailed above, we explored potential opportunities to improve dabigatran’s benefit–risk profile even further. In 2012, our scientists performed preliminary exploratory simulations with PK modelling to understand whether dose adjustments based on plasma concentrations might further improve the efficacy and safety profile of dabigatran etexilate. However, after scientific and mathematical scrutiny was applied, the initial hypothesis from the PK model could not be supported. Rather, the information and conclusions reached using the initial PK model (information submitted to regulatory authorities with the initial application for approval) remained unchanged. Therefore, there were no new data or conclusions to share with regulatory authorities. This is contrary to the author’s suggestion that Boehringer Ingelheim withheld important analyses. (1)
A paper recently published by the RE-LY trial investigators documents the exposure–response relationship of dabigatran. (8) The aim of this analysis was to examine the association between plasma concentrations and efficacy and safety outcomes, and to identify patient factors that influenced this relationship. The results of this pre-specified sub-analysis showed that the risks of major bleeding, but also the risk of ischaemic stroke, correlated with dabigatran plasma concentrations. Whether, however, a single optimal concentration range for all patients could be specified was a subject of extensive discussions among the authors. Finally, they came to the conclusion that there is no single range that fits all patients because individual patient characteristics (e.g. age, renal function, prior stroke), put into perspective through medical judgement, are at least as important as the plasma concentrations in assessing benefit–risk and confound the association between drug levels and outcomes. (8,9)
The article by Cohen (1) suggests that determining dabigatran levels by anticoagulant monitoring may further reduce the risk of bleeding and improve outcomes. However, the issue is not simply about how to reduce major bleeds; it is about achieving a balance that maximizes the reduction of strokes and systemic embolisms while minimizing major bleeds. It may appear logical to monitor concentrations and adjust dosing as known from decades of oral anticoagulation with vitamin-K-antagonists (VKAs). However, in contrast to VKAs, which are affected by metabolism and numerous drug–drug and drug–food interactions, the plasma levels of dabigatran are primarily dependent on renal function.
Dabigatran has a direct mode of action on thrombin with different pharmacokinetics to the VKAs, which act indirectly through the vitamin-K-dependent clotting factors. For the overall benefit–risk assessment of dabigatran, the assessment of the patient’s renal function and consideration of the patient’s age is more relevant than the assessment of plasma concentration which is not considered to contribute clinically relevant added value.
In May 2014, the US Food and Drug Administration (FDA) reaffirmed the positive efficacy–safety profile of dabigatran in one of the largest real-world analyses of its kind comprising 134,000 patients above 65 years. (10)
Klaus Dugi, Chief Medical Officer, Boehringer Ingelheim GmbH, Ingelheim, Germany
Joerg Kreuzer, Therapeutic Area Head Medicine Cardiology, Boehringer Ingelheim GmbH, Ingelheim, Germany
1. Cohen D. Dabigatran: how the drug company withheld important analyses. BMJ 2014;349:g4670.
2. Cohen D. Concerns over data in key dabigatran trial. BMJ 2014;349:g4747.
3. Moore TJ, Cohen MR, Mattison DR. Dabigatran, bleeding, and the regulators. BMJ 2014; 349:g4517.
4. Charlton B, Redberg R. The trouble with dabigatran. BMJ 2014;349:g4681.
5. Cohen D. Re: Dabigatran: how the drug company withheld important analyses [Response, 11 August 2014]. BMJ 2014;349:g4670.
6. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361(12):1139-51.
7. Connolly SJ, Ezekowitz MD, Yusuf S, Reilly PA, Wallentin L; Randomized Evaluation of Long-Term Anticoagulation Therapy Investigators. Newly identified events in the RE-LY trial. N Engl J Med 2010;363(19):1875-6.
8. Reilly PA, Lehr T, Haertter S, Connolly SJ, Yusuf S, Eikelboom JW, et al. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY Trial (Randomized Evaluation of Long-Term Anticoagulation Therapy). J Am Coll Cardiol 2014;63(4):321-8.
9. Reilly PA, Connolly SJ, Yusuf S, Eikelboom JW, Ezekowitz MD, Wallentin L; RE-LY Investigators. Reply: regarding the effect of dabigatran plasma concentrations. J Am Coll Cardiol 2014;63(25 Pt A):2885-6.
10. FDA Drug Safety Communication: FDA study of Medicare patients finds risks lower for stroke and death but higher for gastrointestinal bleeding with Pradaxa (dabigatran) compared to warfarin – 13 May 2014. http://www.fda.gov/Drugs/DrugSafety/ucm396470.htm Last accessed 24 July 2014.
Competing interests: Employees of Boehringer Ingelheim
The issues surrounding whether dabigatran's advantages (no need for monitoring) have been overstated - deliberately or otherwise - are newsworthy and potentially scandalous.
What is of even greater concern, however, is the fact that this new anticoagulant-of-convenience cannot be reversed. There is no antidote. The time for the effects to wear off is often many days. In trauma management this is a major issue. We are used to patients on warfarin presenting with these problems. Warfarin can be reversed with fresh-frozen plasma or vitamin K and so we can usually control the life threatening haemorrhage or normalise the INR sufficiently to allow eg fracture neck of femur surgery.
As the use of dabigatran becomes more prevalent the number of serious complications is rising.
Is a monthly INR really so inconvenient? Not when compared to the inconvenience of bleeding that can only be controlled by haemodialysis.
Until such time as an antidote is available it is irresponsible to routinely prescribe this drug.
Competing interests: No competing interests
Re: Dabigatran: how the drug company withheld important analyses
We are responding to comments made by Dr. Yui in his rapid responses to the BMJ article by Cohen (1) on 15 October 2014.
The author raises questions and makes assertions regarding the scientific grounds of the 80 seconds aPTT value appearing in the Japanese package insert and other documents shared with Dr. Yui as part of our ongoing scientific exchange with the author.
First, we would like to correct the assertions made on our communication toward health care professionals in Japan. The author states that, “Doctors in Japan have been told by Boehringer Ingelheim that aPTT 80s is the good index of dabigatran for preventing major and life-threatening bleeding”, and that BI would promote aPTT >80 seconds as “very important guide”. On the contrary, in the Japanese PRADAXA labelling as well as in the promotional material we shared with the Japanese Ministry of Health, Labour, and Welfare regarding the limitations of anticoagulant assays, the description of the 80 seconds aPTT as an indicative value and not as a threshold was clearly highlighted. In the “Warnings” and “Important precautions” sections of the Japanese package insert it clearly reads, “since there are neither established indicators to accurately assess haemorrhage risks of this drug nor a neutralizer of its anticoagulant effects, patients should be sufficiently monitored for signs of haemorrhage or anaemia, etc. not just relying on blood coagulation levels during treatment with this agent. If any such signs are observed, appropriate measures must be immediately taken.” The author also omits to mention that the paragraph in the “Important precautions” section where the aPTT 80 seconds value appears starts with, “In patients with bleeding, aPTT measurement might serve as an indicator for excessive anticoagulation. In the phase III trial including Japanese patients, major bleeding occurrence was higher when trough aPTT exceeded 80 seconds”. The aPTT value of 80 seconds is mentioned, however, it is not meant to be a threshold or reference level but merely a descriptive, possibly indicative value.
We reaffirm that Nippon Boehringer Ingelheim communication is fully aligned with the approved labelling and intended to promote the most appropriate use of dabigatran in patients with NVAF.
The author also states that BI never explained the origin of this value (80 sec) and concludes that it is “groundless” and not derived from the RE-LY study. We object to this interpretation. The exploratory analysis “PK and PD profile of dabigatran and relations to efficacy and safety outcome events” was pre-specified. (2) It was based on blood samples available from approximately 9 000 patients for peak and 8 400 patients for trough measurements for each of the dTT, aPTT, and ECT assays. (3) This represents one of the largest PK/PD datasets available in the field of anticoagulation in a prospective trial. It is clearly stated in the EMA public assessment report, and the Japanese public assessment report which we shared with the author, that the aPTT of 80 seconds originates from RE-LY.
Despite this large dataset, we could not demonstrate a significant correlation between aPTT at trough and the occurrence of major bleeding events after applying logistic regression (the FDA found an association with “life-threatening bleeding”).
The author questioned why and how the “cut-off” at 80 seconds was selected. As reported by Dr. Yui, we explained that this represented the 90th percentile of the trough aPTT value observed in RE-LY. We explained that although the “80 sec” was not supported by robust evidence, it has been requested by some health authorities in order to have at least one indicative value to guide physicians in the instance of some specific conditions such as requirement for surgery or occurrence of bleeding. No therapeutic strategy derived from the 80-second “cut-off” is suggested in the Japanese label.
The analysis provided to Japanese health authorities in reference to the origin of the “80 seconds” is also purely descriptive (the crude rates of major bleeding in patients on-treatment was 3.4% and 5.1% in the aPTTtrough ≤80 and >80 seconds strata, respectively). No inferential statistics (significance tests) were applied since the calculation of a P-value would have been inappropriate, and as described elsewhere, misleading in this context. (4, 5)
Taken together, measurement of the aPTT (and/or ECT/dTT in institutions/countries where it is available) can help to guide physicians in certain clinical settings and patients.
Prof. Dr. Jörg Kreuzer, Therapeutic Area Head Medicine Cardiology, Boehringer Ingelheim GmbH, Ingelheim, Germany
1. Cohen D. Dabigatran: how the drug company withheld important analyses. BMJ 2014;349:g4670.
2. Ezekowitz MD, Connolly S, Parekh A, et al. Rationale and design of RE-LY: randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J 2009;157:805-10, 810.e1-2.
3. Reilly PA, Lehr T, Haertter S, et al. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY Trial (Randomized Evaluation of Long-Term Anticoagulation Therapy). J Am Coll Cardiol 2014;63:321-8.
4. Goodman SN. Toward evidence-based medical statistics. 1: The P value fallacy. Ann Intern Med 1999;130:995-1004.
5. Ludwig DA. Use and misuse of p-values in designed and observational studies: guide for researchers and reviewers. Aviat Space Environ Med 2005;76:675-80.
Competing interests: No competing interests