You mention that the lower doses of the novels were not studied.  I agree that 75mg of Dabi was not studied but the 15mg of Xarelto was studied in renal patients in Rocket - correct?  I know that it was studied at a clearance of 30 but moved to a clearance of 15 - 50 based on pharmacology - I am also not sure on the Apixaban trial.

All the trials (RELY, ROCKET< ARISTOTLE) excluded CrCl < 30.  You are correct that ROCKET-AF did adjust the dose for CrCl 30 - 49.

From the paper: "Patients were randomly assigned to receive fixed-dose rivaroxaban (20 mg daily or 15 mg daily in patients with a creatinine clearance of 30 to 49 ml per minute) or adjusted-dose warfarin (target international normalized ratio [INR], 2.0 to 3.0)."

http://www.nejm.org/doi/full/10.1056/NEJMoa1009638#t=articleTop

 My contention is that some of the hazard and subsequent bad press may be use of these agents in patients with CrCl that is not calculalted by prescribers and therefore may be low especially in the elderly (I have no data for this).  For now I'm sticking with refraining from these agents with CrCl < 40.

 Thanks for posting.

I comment as a patient who started Pradaxa the very week that it was approved for non-valvular afib stroke prevention.  To me, a major factor that is overlooked in making these comparisons is the [misapplied?] prohibition against taking any form of Vitamin K (principally Vitamin K1) with warfarin.  In the long run, that can hasten the onset of osteopenia and its progression to osteoporosis.  At this point, I am playing catch-up by taking hefty doses of Vitamin K2, but also Vitamin K1, as I am osteopenic in my left femoral neck.  I do indeed experience dyspepsia with Pradaxa, but taking it with meals helps in that regard.  Sometimes, I also need to take a swig of Maalox at bedtime.  

Dear Dr Bilazarian, 

It is early days I bet you'd answer, but for paroxysmal AF patients with spaced out episodes, would it make sense to adopt a "pill-in-the-pocket" approach, thus covering the stroke risk acutely?

Thanks. 

Dear Dr. Bilazarian,

 

The data in Rocket-AF for patients with CrCl <50 was very robust.  21% of the patients in the Rocket-AF trial had a CrCl <50.  The mean age of these patients with moderate renal impairment was 79 and the average CHADS2 score was 3.7.  Clearly, this is an advantage for Rivaroxaban and I don't understand why you would limit your practice to patients with CrCl >40.

Rivaroxaban should be taken with food, but doesn not have to be taken with the evening meal.

Rivaroxaban having different doses is an advantage because it has multiple indications and an approved and clinically proven renal dose.  How could this be a disadvantage?

Rivaroxaban had a statistically significant advantage over Warfarin for the final composite end point when analyzing the on-treatment population.

Also, you might spend a little more time understanding how to turn off your camera feed.

Your anacronym is not worth commenting on. 

Regards,

DHJ 

DHJ, thanks for posting and your clarifying comments:

You say that "Rocket-AF for patients with CrCl <50 was very robust.  21% of the patients in the Rocket-AF trial had a CrCl <50."   => I think that this is true but not complete and one of my points is that the CrCl < 30 population was excluded so it is more correct to say that the data is robust for the CrCl 30 - 50 population.  From the paper "Patients were randomly assigned to receive fixed-dose rivaroxaban (20 mg daily or 15 mg daily in patients with a creatinine clearance of 30 to 49 ml per minute)" http://www.nejm.org/doi/full/10.1056/NEJMoa1009638#t=articleTop - I don't think this is a subtle point for patient safety.

 You say "The mean age of these patients with moderate renal impairment was 79 and the average CHADS2 score was 3.7." =>  I completely agree that the strength of ROCKET AF is that the patients had much higher CHADS2 scores than the other trials.  I made this point when the data was first released.  http://blogs.theheart.org/private-practice/2011/11/8/adopting-rivaroxaban-xarelto-the-pros-and-the-cons   The accompanying slide set has the trial comparisons and shows how similar and lower risk RELY and ARISTOTLE patient populations are and how much higher risk the ROCKET-AF patient population is.  This has been a point of controversy which I dont really have a conviction about.  Although its useful to know that this agent is non inferior to warfarin in these higher risk patients some argue that it would have been expected that the drug should have been able to show superiority because of the higher risk population - this is primarily a meaningful thought experiment for the indirect comparisons between the new agents.

You say  "Clearly, this is an advantage for Rivaroxaban and I don't understand why you would limit your practice to patients with CrCl >40"  +> We dont have any data that Rivaroxaban is superior or an efficacy or safety "advantage".  I think it has shown itself to be solidly non-inferior.  The point I made about the CrCl < 40 is the one I most hope will drive comment amongst practicing clincians on this comment board.  I am not sure what to do in this regard.  Patients in any practice are not followed as closely as patients in clincial trials (frequent labs and visits).  I often see patients with serum creatinine at the upper end of the normal range bump up during mild illnesses (URI, UTI, diuretic adjustment) and have chosen, arbitrarily a cushion of CrCl with the hope that these declines in renal function will not result in significant changes in drug elimination and therefore hazard to the patient. Is this too big a cushion or too small a cushion or is this not even reasonable?

You say "Rivaroxaban should be taken with food, but does not have to be taken with the evening meal."  =>  The package insert and trial conduct both have the "evening with meal" instruction.  I have asked the company for guidance on this and gotten the standard answer that this is what the packege insert says.  Can clincians and patient be confident that they will derive the effciacy and safety seen in ROCKET-AF if they take it in the morning without food?  If so, that would be great to know. Compliance is always better with once daily in AM administration. I don't know what you based your comment on but hope you'l l share what you know.

You say "Rivaroxaban having different doses is an advantage because it has multiple indications and an approved and clinically proven renal dose.  How could this be a disadvantage?"  => I think there is is a potential for prescriber confusion.  Flexibility is of course very desirable and I have lamented the FDA's decision not to give us dosing flexibility with Dabigatran (Pradaxa).

You say "Also, you might spend a little more time understanding how to turn off your camera feed."  Guilty

Your anacronym is not worth commenting => acronyms work for me and might work for others

 Thanks for posting.  sb

Dr. Bilazarian,

Thank you for your thorough response.

Rivaroxaban per the PI and available Pk/PD analyses is more readily absorbed in the proximal GI. It should be administered with food, at the 15mg and 20mg doses, in order to slow motility and increase absorption.  Taken without food, at the 15mg and 20mg, Rivaroxaban has a bioavailability of 66%, but taken with food AUC and Cmax increase by 39% and 76%, respectively.  Taken with or without food, the 10mg dose for DVT prophylaxis is estimated to be 80% to 100%.

In your primary analysis, multiple doses was a disadvantage.  Yet, in your reply, you lament the fact that the FDA did not 'give us dosing flexibility with Pradaxa'.  This seems to be a contradiction.

Clearly, the high risk patient population in Rocket-AF gives me much greater comfort in prescribing Rivaroxaban to patients.  It's a once a day med, with good tolerability.  I take confidence in the superior results seen vs. warfarin when compared to patients on-treatment.

Recent VA recommendations for use with regards to Dabigatran and Rivaroxaban provide very clear and valuable patient inclusion and exclusion parameters. 

Thus far, Rivaroxaban seems to have good formulary support, and I feel that it has multiple indications to be a positive.

For patients with a CHADS2 of 0-2, Dabigatran or Rivaroxaban, overall, appear to be equally yoked.  However, for older sicker patients, those with CHADS3, I prefer Rivaroxaban or remain with warfarin. 

One glaring issue with Rivaroxaban is an absence of data with cardioversion, which the VA recommendations clearly outline.  Dabigatran actually does have some data regarding this important patient subset. 

Your anacronym as 'RABI' is certainly reasonable and useful.

Sincerely,

DHJ 

Thanks Seth.

 The last three minutes was excellent.

 

CRABI is a great way to assess AF.

 Last year we stayed at Krabi in Thailand, never thought it would relate to AF.

 Keep blogging

 

Cheers 

I find it interesting that all though there is no statistcal difference in the number of MI's between Warfarin and Pradaxa you list it in the "Against" column for Pradaxa.  Why if there is no statistical difference in the Re-LY trial should this impact my decision?

CLS, Thanks for posting. This is an importnat question.  I'm not sure what to think about it.  The MI risk is small but seems to be consistent.  If real it would need to be included as a negative.  here's what I know:

1.  In RELY http://www.nejm.org/doi/full/10.1056/NEJMoa0905561#t=articleTop the rate in the 150 mg bid arm was 0.74 / yr and and 0.53 / yr with warfarin  p =0.048 RR = 1.38 (1.00-1.91).  The Absolute hazard is 0.21 / year.  NNH = 476

2. The authiors then updated the RELY paper with added events http://www.nejm.org/doi/full/10.1056/NEJMc1007378 and the rates were 0.81 vs 0.64/ year and p became NS at 0.12 RR = 1.27 (0.94-1.71)  Therefore not statisitically significant and should not be part of the decision and is considered a play of chance or not stisitically detectable in the patient population.

3. In Archives of Int Med earlier this year http://archinte.jamanetwork.com/article.aspx?articleid=1108764 

The abstract says: A metaanlysis of Seven trials  (N = 30 514),  including 2 studies of stroke prophylaxis in atrial fibrillation, 1 in acute venous thromboembolism, 1 in ACS, and 3 of short-term prophylaxis of deep venous thrombosis.  Dabigatran was significantly associated with a higher risk of MI or ACS than that seen with agents used in the control group (dabigatran, 237 of 20 000 [1.19%] vs control, 83 of 10 514 [0.79%]; OR_M-H, 1.33; 95% CI, 1.03-1.71; P = .03). The risk of MI or ACS was similar when using revised RE-LY trial results (OR_M-H, 1.27; 95% CI, 1.00-1.61; P = .05) or after exclusion of short-term trials (OR_M-H, 1.33; 95% CI, 1.03-1.72; P = .03). Risks were not heterogeneous for all analyses (I² = 0%; P ≥ .30) and were consistent using different methods and measures of association.  Conclusions  Dabigatran is associated with an increased risk of MI or ACS in a broad spectrum of patients when tested against different controls. Clinicians should consider the potential of these serious harmful cardiovascular effects with use of dabigatran. 

 

The multiple editorials can be found here: http://archinte.jamanetwork.com/searchresults.aspx?q=uchino&t=&p=1&s=1&c=0  they have the usual criticisms about meta analyses and the fact hat RELY was so big compared to the smaller trials that the whole process was futlile.

Deep Breath!  At the end of the day I think the risk is very small and not clinically meaningful.  (Certianly a lower risk thatn NSAIDS and COX2 inhibitors)  I think for completelness its worth including it in the discussion.

 Thanks for asking.

sb 

Thanks for the reply Seth.  I didn't put much stock in the Uchino metanalysis since he compared apples and oranges with different comparartors and different pateint populations.  I read a post hoc analysis in Circulation by Hohnloser that seemed to be more relaible and really put that concern of MI's in perspective.  If I'm not mistaken the Dabigitran 150 arm started with more MI's in the babseline characteristics.  I think it's well documented now that those with previous MI's are more likely to have another MI.  It's reassuring to know that the FDA hasn't seen any reasons to include a black box or warning around MI's and Dabigitran.  Thanks again for all of the information.  With the new generation of meds your blogs make it easier to stay current.

Overall - NICE presentation! - THANK YOU for this blog!

My concern - "lower risk of stroke" was presented in "slam-dunk" fashion. In my view it is anything but. "Slam-dunk" is what the manufacturers have put their billions of dollars into - in the hope of encouraging marketing of their product.

My interpretation of the literature is different and more guarded. The ARISTOTLE Trial (NEJM 365:981,2011) was promoted as showing apixaban to be more effective than warfarin in preventing stroke, systemic embolism, bleeding and mortality. BUT - to get those results to achieve statistical significance - they needed no less than 1000 sites to get their 18,000 patients - with numbers like reduction in stroke from 5.04% with coumadin to 4.49% with apixaban. Yes - that's a relative 10% improvement - but a NNT (Number-Needed-to-Treat) ~ 200 patients in order to benefit one.  Similar small statistially significant differences were found in the other parameters stated, all of which sound much better with relative percentage reduction numbers than when absolute numbers of patients over the denominator of total patients in this study are used ....

This is THE BEST that it is going to get: i) manufacturer-supported study; ii) cards stacked in favor of apixaban (patients with renal insufficiency Cr <25 excluded) - AND - patients therapeutic on warfarin only 66% of the time (would seem excess bleeding clearly faciliated when warfarin is supratherapeutic - as often occurred in this manufacturer study .... ).

I am NOT saying warfarin is wonderful - because it is not. CREDIT to you because you clearly emphasize issues of cost, and favor continuing warfarin in patients doing well on that drug, those with renal problems, and those who don't want to be "early adaptors". BUT - Isn't this scenario familiar with "great results" in a manufacturer-supported study initially, only to see the blemishes come out one-by-one several billion-dollars-spent later after a period of use ... - and to me, results in Aristotle - while positive, are anything but "great" given small absolute number reductions, huge cost, GI upset, no monitoring (as a disadvantage), need for compliance given shorter-duration-of-action with no-way-to-monitor-if-patient-is-taking-the-drug - plus lack of reversibility.

I DO think the new blood thinners offer certain advantages - and for some patients they are a better choice than coumadin. THAT SAID - overpromotion and "slam-dunk-appearance" of "increased efficacy" makes me highly suspicious of the maker of the drug ...

Dr Grauer,

Great post - I agree with it all.  The only think I would add to your correct statement about the small benefit (in NNT) on the efficacy side is there is also a safety benefit (also small) and convenience benefits (difficult to quantify).

Thanks again for post, sb

Guy, Glad to hear of your suggestion.  Since Crab in Thai is ?? the translated acronym would be ??RABI not KRABI, sb

You disclosed participation in research trials, do you also receive honoraria for giving talks on any of the newer agents.

Thanks a lot for review.

I think that we could compare RE-LY and ARISTOTLE only, because ROCKET AF population is absolutely different.

And if we look closely for RE-LY subanalisys for more severe patients(

http://circ.ahajournals.org/content/123/21/2363) ,  we might suppose that result could be another, in case of patient's CHADS score would be higher in RE-LY.

Could you please comment on this?

Seth,

Nice job of working to help make difficult comparison recommendations.

Thank you.

Milton

Thanks for your question.  My disclosures are always complete.

I think and personally follow the standard that disclosures are complete - otherwise they are not disclosing.  If they are incomplete they would be called "partial" dislosures.

With the expected federal Sunshine Act there will be postings of data for individual physicians.  In my state this has already been in place for 2 years:

Mass. posts industry payments to health care providers 

http://www.boston.com/news/health/blog/2010/11/by_liz_kowalczy_2.html 

To see the detail for physicans in Massachusetts in 2010 (last year available you can check it out here) I am not listed because I have no payments.  See the list here:

http://www.mass.gov/eohhs/docs/dph/quality/healthcare/pcoc/top-20-manufacturers-detail-2010.pdf

or see the top 10 docs or payers at

http://www.mass.gov/eohhs/provider/licensing/programs/pharm-code-of-conduct/data/prepared-reports.html 

Milton, Thanks for the encouraging words.  sb

Dr. Feofanova, thanks for posting and for your encouraging words:

I agree with your assessement about the differences in the trials.  My blog on adopting Xarelto has a downloadable slideset and slide 11 has the data side by side data which ilustrates these trial similarity and differences.  Get it here at http://blogs.theheart.org/private-practice/2011/11/8/adopting-rivaroxaban-xarelto-the-pros-and-the-cons  

The RE-LY assessment you cite is a comparison between those less than 75 vs those > 75.

 The data comparing CHADS2 groups is in ANNALS  http://annals.org/article.aspx?articleid=1033155

 The problem with comparing the subgroups for RE-LYis that the trial already has 3 arms so there are many less patients in the CHADS2 groups in the 150 mg dose arm than there is ROCKET-AF so making the comaprisons is very hazardous.

 The paper basically says in Table 2  that  the efficacy moving from CHADS2 of 0-1, 2, > 3 has RR reduction of dabigatran 150 bid compared to warfarin of 0.43, 0.54 and 0.85 respectively, but the CHADS2 > 3 gorup has CI of 0.16 to 1.53 so is therefore not significant.  Similarly the CI for major bleeding by risk scores is not significant so it is not helpful.  As I have said I am willing to make indirect comparisons between trials but dont feel like I can say that there is enough data to say that I would recommend to patients that there is one new agent better than others for patients with CHADS2 score > 3.  I can say that ROCKET-AF evaluated  patients with CHADS2 score > 2, greater than 3 and showed RIvaroxaban to be not inferior to Warfarin in efficacy or safety.

Two comments:

 1) We have dabigatran 110 mg BID available here;

2) ROCKET-AF studied a very short-half-life molecule which they (inappropriately) dosed once daily - they did not show superiority to very poorly controlled warfarin, with a low TTR in the warfarin group. My contention is that rivaroxaban must be dosed BID, based on consideration of its PK/PD profile. In fact, the preponderance of strokes and deaths which occurred shortly after stopping rivaroxaban and starting warfarin supports this contention directly. I don't know if the drug is tablet or capsule based and, if the former, whether it can be split by pharmacists, but if I prescribed it, that's the way I would go.

 3) The NNTs quoted above by your reader on July 13 were misleading, since they are for apixaban COMPARED AGAINST WARFARIN. In practice, an AF patient is either treated with anticoagulant, antiplatelet or nothing. Therefore, the NNT for a drug like apixaban needs to be computed against imputed placebo (or aspirin, or aspirin-clopidogrel), not warfarin. The gain in stroke risk and mortality in your patient is when you apply apixaban instead of not applying an anticoagulant, since it's the anticoagulant that reduces the risk. Since warfarin has a dramatically low NNT versus placebo in multiple trials (see Hart's meta-analyses), and apixaban is decidedly better than warfarin, we can infer clearly that apixaban has even lower NNT against placebo. It's just that it's not ethical anymore to do placebo-controlled trials of anticoagulation in AF, which means that NNT's from trials (apixaban vs warfarin, rather than apixaban vs placebo) are distinctly non-informative. 

I would like to clarify my comment #3 above, which may sound confusing. One of the readers posted that the NNTs of apixaban vs warfarin were trivial, and so there is no need to select apixaban. However, what he is not realizing is that in any individual patient, the first (primary) choice is to anticoagulate or not, and therefore the differential NNT is based on this decision. Apixaban is clearly better than warfarin (ARISTOTLE) or aspirin (AVERROES). By logical deduction, it must be better than placebo (since both warfarin and aspirin have demonstrated benefit against placebo): if A is better than W, AND, W is better than P, then A must be better than P. The NNT to be computed could and should be computed versus placebo. I would not be surprised based on the warfarin literature and the demonstrated superiority on stroke and mortality that the NNT for apixaban versus an imputed placebo would be under 10. That should dramatically capture notice, not some nonesensical incremental NNT vs warfarin (in practice, a patient gets one of these agents or they don't).