Boehringer NT ProBNP study

RE: Boehringer NT ProBNP study

Sarah Keir — Tue, 23 May 2017 18:14:25 GMT

[quote user="Anthony Dennison"]

[quote user="Niall Taylor"]Starting this thread was presumably intended to help examine the ethics and usefulness.[/quote]

Actually, I started it to see if anyone else was doing it (because I hadn't heard about it) and to discuss accuracy of proBNP. Been an interesting discussion though!

[/quote]

They go hand in hand.

RE: Boehringer NT ProBNP study

Lucy Fleming — Tue, 23 May 2017 17:46:37 GMT

[quote user="David Mills"] It just bemuses me that we pursue this test when there is a non-invasive, far more informative, one available and achievable in practice.[/quote]

From that point of view, and it depends on the individual animal, but some will find restraint for an echo much more stressful than for a blood sample. I wouldn't use pro-BNP alone to make treatment decisions. Certainly right now we are offering it primarily because it is free, but it's made me think about discussing it more with owners at health checks in future (perhaps for Dobermanns, not labs )

RE: Boehringer NT ProBNP study

Martin Atkinson — Tue, 23 May 2017 15:39:23 GMT

[quote user="David Mills"]It just bemuses me that we pursue this test when there is a non-invasive, far more informative, one available and achievable in practice. I wonder if its because we take pleasure/comfort in numbers because they create the illusion of objectivity?[/quote]Because it's easy, 80% is better than nothing - I've yet to find a false -ve, not everyone is experienced enough to confidently echo a heart - even if they can if you charge less than the Cardiopet test you're giving your services away, or the animal has to travel to a specialist who will charge a lot, lot more.

If the sample is drawn cleanly and spun immediately with the plasma put straight into an EDTA tube and frozen until it is convenient to post it to Idexx it should be reliable. A paper by Mike Herrtage et al showed that although renal function had some effect on the result of NT-proBNP it was marginal, not in the thousands or even hundreds that would influence a judgement based on the result.

I'm not saying its perfect, or using it a a weapon against you David, but I do feel you are dismissing it and the rationale behind it too easily.

RE: Boehringer NT ProBNP study

David Mills — Tue, 23 May 2017 14:35:22 GMT

[quote user="Lucy Fleming"] A 'normal' pro-BNP is suggestive of no heart disease and no follow up required [/quote]

Problem is with the test. The sensitivity at best is about 80% so 1/5 negatives are likely to be false (and this is based on selected experimental animals mainly - the in-field sensitivity is likely to be even lower). Then throw in the other things that can affect it (handling, transport, renal function, time of day of test, drugs etc etc) and rapidly the faith in it identifying anything must start to tail off. It just bemuses me that we pursue this test when there is a non-invasive, far more informative, one available and achievable in practice. I wonder if its because we take pleasure/comfort in numbers because they create the illusion of objectivity?

RE: Boehringer NT ProBNP study

David Mills — Tue, 23 May 2017 14:30:20 GMT

[quote user="Kate Richardson"]Wasn't for research as far as I was aware- we were offered free pro BNP tests for a period of time and as our practice likes to be pro-active with regards to preventative health care, we were keen to take up this offer. Dog's weren't part of a trial of any kind[/quote]

So the results aren't going to be used?

The "preventative healthcare" argument is a non-sequitur unless you would offer BNP to all these dogs anyway - but I suspect it's being offered primarily because it is free.

The test, at best, is probably about 80% specific and that's not taking into account the huge "equivocal" range present. The numbers needed to even pick out one dog are likely going to be very large.

An echocardiogram is non-invasive, and would not fall under ASPA. Taking blood samples that would not be taken otherwise (see above) very much does.

RE: Boehringer NT ProBNP study

Martin Atkinson — Tue, 23 May 2017 14:04:04 GMT

[quote user="Lucy Fleming"]I feel like we might all be coming at this from different directions.[/quote]I feel we're all coming at it from the same direction bar one.

RE: Boehringer NT ProBNP study

Lucy Fleming — Tue, 23 May 2017 13:00:52 GMT

[quote user="David Mills"]I still maintain radiography is the mainstay of chf diagnosis [/quote]

I feel like we might all be coming at this from different directions. In this context, we have offered the pro-BNP test to dogs that are considered to be at increased risk of developing DCM compared to the 'average' dog. A 'normal' pro-BNP is suggestive of no heart disease and no follow up required (or annual re-test? Or six monthly? That's a whole other discussion). An elevated pro-BNP might be suggestive of heart disease (or might be a false positive). For these cases, we've discussed following up with an echo to assess for cardiac enlargement (suggesting evidence of DCM, but not at this stage looking for heart failure) .

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 21:36:23 GMT

[quote user="Beats"]But not all cases of dyspneoa due to left-CHF present with all these radiographic signs being clear-cut. The example I gave of an acute chordae tendinae rupture without significant left atrial enlargement was meant to illustrate that. [/quote]

No of course but ruptured chordae tendinae is very rare IME even with MMVD and I've never seen one without valvular heart disease. Possibly endocarditis, but there are guidelines on the diagnosis of this. There are always surprises but I still maintain radiography is the mainstay of chf diagnosis in almost all cases. Of course we're talking about tests in isolation so clinical examination still maintains (metronomic tachycardia, murmur etc).

[quote user="Beats"]To me, it is often far from clear-cut[/quote]

There used to be a great resource from university of guelph vetgo.com. Seems to be down at the moment but loads of radiographs and explanations (this is not meant to be patronising it basically helped me pass my cardiology cert )

And yes, I hate vhs. Almost completely breed dependent.

RE: Boehringer NT ProBNP study

Kate Richardson — Mon, 22 May 2017 21:32:46 GMT

[quote user="David Mills"]Is it covered by ASPA? Given this is blood sampling for research rather than under veterinary direction?[/quote]

Wasn't for research as far as I was aware- we were offered free pro BNP tests for a period of time and as our practice likes to be pro-active with regards to preventative health care, we were keen to take up this offer. Dog's weren't part of a trial of any kind

RE: Boehringer NT ProBNP study

Beats — Mon, 22 May 2017 21:08:40 GMT

[quote user="David Mills"]

[quote user="Beats"]I don't quite follow, in a dyspnoeic patient is it always possible (to the suitably trained eye) to determine whether a patient needs furosemide'd or not based on radiography alone?[/quote]

Yes I believe so, and this is echo'd by several specialists.

The cardinal signs are:

1. enlarged LA

2. pulmonary vein congestion (artery:vein >/= 1)

3. evidence of interstitial (early oedema) or alveolar pattern or pleural effusion in cats only (the pleural veins empty into azygous vein c.f. the dog)

All three must be present to put the dyspnoea down to cardiac cause.

[/quote]

and then I'd add response to furosemide as number 4.

But not all cases of dyspneoa due to left-CHF present with all these radiographic signs being clear-cut. The example I gave of an acute chordae tendinae rupture without significant left atrial enlargement was meant to illustrate that. To me, it is often far from clear-cut and an assessment of the LA:Ao ratio by echo (and probably NT-proBNP) would help me. Even when I send rads to specialist they are not always clear cut (although I appreciate there is clearly going to be sampling bias of sending the less clear-cut ones to start with...)

You may not be a VHS fan either (I know I'm not), but here's a graphic comparing BNP with VHS (albeit in a somewhat suspect sampling population):

⦁ The test compares favourably to other single parameter assessments, so while it shouldn’t be relied upon as the sole means of determining the presence or absence of heart failure, it may perform better than a different single measurement like vertebral heart score (VHS). A properly handled proBNP result may be more predictive of the presence or absence of heart disease than a poorly-positioned, expiratory chest radiograph. Above are ROC curves indicating the predictive value of proBNP as compared to VHS in 280 clinical patients with known status as to presence or absence of heart disease:

Legend:

A—Receiver-operating characteristic curve displaying sensitivity and specificity of the use of plasma NT-proBNP concentrations to distinguish between dogs with and without clinical signs of cardiac disease (n = 1,134 dogs) classified by use of the CAH cardiac disease grading system. Sensitivity was 70% and specificity was 83% at a cutoff of 874 pmol/L. B—Receiver-operating characteristic curve displaying sensitivity and specificity of the use of VHS to distinguish between dogs with and without clinical signs of cardiac disease (a randomly selected subset of the dogs in panel A [n = 280]). Sensitivity was 56% and specificity was 85% at a cutoff of 11.5. The 45line represents the line of no discrimination.

Or put simply, “NT-proBNP provides improved sensitivity (approx 70% vs 56% [for VHS]) and similar specificity (83% vs 85% [for VHS]) for the detection of clinical signs of cardiac disease at the respective optimal cutoffs values.”

[quote user="David Mills"]There is generally reasonable agreement between observers.[/quote]

I work with 7 other vets and it's hard to get us to agree on anything, and we're all generally reasonable folks

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 21:03:18 GMT

[quote user="Kate Richardson"]

[quote user="David Mills"]

Completely different set of circumstances old fruit. The protect study was looking at those with heart disease, as defined by the inclusion criteria.

AFAIK this proBNP study is randomly selecting patients. There will be false positives because the test isn't great, and affected by quite a few variables. We know that the vast majority of 3 year old labs will not have clinically relevant heart disease (and even if they do, will BNP pick it up?) I imagine the vision is to start treating based on numbers. A far better thing to do would echo them for a definitive answer - who would suggest that??

[/quote]

We were asked as a practice to take part. We were given guidelines as to patient selection- essentially breeds predisposed to DCM ie large breed dogs, aged between 3 and 7 years of age. As a practice we were allowed to extend these guidelines to include older dogs and other breeds if we wanted to. The aim was to offer to measure pro BNP in healthy dogs with no clinical evidence of cardiac disease. Any dogs with an elevated proBNP were then to be advised to do an echo to assess if there were early changes due to DCM ie pre clinical. These dogs according to the study (on Dobermans? I believe) would then potentially benefit from treatment before clinical signs related to the DCM developed. So essentially screening for pre clinical DCM with a blood test rather than an echo. No risk, minimum cost screening with no obligation to follow on with an echo, however, most people interested in taking the offer up were the sort who would likely follow up with an echo if indicated.

[/quote]

Is it covered by ASPA? Given this is blood sampling for research rather than under veterinary direction?

RE: Boehringer NT ProBNP study

Kate Richardson — Mon, 22 May 2017 21:00:59 GMT

[quote user="David Mills"]

Completely different set of circumstances old fruit. The protect study was looking at those with heart disease, as defined by the inclusion criteria.

[/quote]

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 20:56:50 GMT

Yes I believe so, and this is echo'd by several specialists.

The cardinal signs are:

1. enlarged LA

2. pulmonary vein congestion (artery:vein >/= 1)

3. evidence of interstitial (early oedema) or alveolar pattern or pleural effusion in cats only (the pleural veins empty into azygous vein c.f. the dog)

All three must be present to put the dyspnoea down to cardiac cause.

[quote user="Beats"]Presumably inter-observer variability is likely to be much higher for echo than NT-pro-BNP when performed by average, non-cardio-interest, vet with non-fancy machine?[/quote]

Possibly. There is generally reasonable agreement between observers. And to diagnose likely CHF from echo then you're only looking at LA:Ao ratio. The cause is essentially irrelevant in terms of therapy (unless due to arrhythmia, which is very infrequent IME).

You can't diagnose CHF from echo, only radiography.

RE: Boehringer NT ProBNP study

Beats — Mon, 22 May 2017 20:46:53 GMT

[quote user="David Mills"]I disagree. Amongst most cardiologists, the general consensus is that you don't need an echo to diagnose CHF. It's a radiographic diagnosis. Same goes for BNP I suspect.[/quote]

I don't quite follow, in a dyspnoeic patient is it always possible (to the suitably trained eye) to determine whether a patient needs furosemide'd or not based on radiography alone? If you don't need an echo to diagnose CHF, what do you need it for? How do you know in a case with no radiographic enlargement of left atrium that it is CHF causing the white lungs? [I only ever try to echo patients that are dyspnoeic generally and don't have a cardio interest]

[quote user="David Mills"]Any test should be considered on it's own two feet, as it were, and compared to a gold standard (echo in this case for pre-clinical heart disease).[/quote]

Presumably inter-observer variability is likely to be much higher for echo than NT-pro-BNP when performed by average, non-cardio-interest, vet with non-fancy machine?

[quote user="David Mills"]

[quote user="Beats"]I'd go as far as to say that it's inclusion in a pre-anesthetic blood screen[/quote]

Which are normally pretty useless. God only knows what we're getting ourselves into.

[/quote]

I'm happy to tangent this, but agree it's out of place in this thread.

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 20:22:29 GMT

[quote user="Beats"]

A case of small breed dog with grade 3 murmur, presenting with dyspnea and white lungs, with no left atrial enlargement, a suspicion of acute chordae tendinae rupture and an equivocal response to furosemide. A "low" NT-proBNP would push me back to the drawing board to consider other differentials. A >3000 result would encourage me to persevere with the furosemide a bit longer. Of course by the time the result comes back it might all be a bit academic, but if I lived near the Idexx lab I might get more use out of this.

[/quote]

I disagree. Amongst most cardiologists, the general consensus is that you don't need an echo to diagnose CHF. It's a radiographic diagnosis. Same goes for BNP I suspect.

[quote user="Beats"]Compared to many parameters tested on a more regular basis such as ALKP, cholesterol, CK, amylase etc I think it probably fairs pretty well as a biochemical marker of disease.[/quote]

This is a category fallacy. Any test should be considered on it's own two feet, as it were, and compared to a gold standard (echo in this case for pre-clinical heart disease).

[quote user="Beats"]I'd go as far as to say that it's inclusion in a pre-anesthetic blood screen[/quote]

Which are normally pretty useless. God only knows what we're getting ourselves into.

RE: Boehringer NT ProBNP study

Beats — Mon, 22 May 2017 20:11:28 GMT

[quote user="David Mills"]An echo should be achievable in most practices and should give more information. Possibly I'm skewed given my cardio interest but I can't think of a single time I would use BNP.[/quote]

I have worked in clinics without ultrasound in the not too distant past. More commonly, the combination of ultrasound probe available and ultrasonographic skill to wield it would make me think I don't know if what the echo tells me is reliable - I would not consider it likely to be more reliable than the NT-proBNP (but the in-house ultrasound machine is cheaper to use and to hand so practically gets a lot more use).

I don't think NT-proBNP is a bad test. Compared to many parameters tested on a more regular basis such as ALKP, cholesterol, CK, amylase etc I think it probably fairs pretty well as a biochemical marker of disease. Indeed, I'd go as far as to say that it's inclusion in a pre-anesthetic blood screen (along with PT/APTT perhaps) would stand a better chance of being helpful than liver enzymes or urea [assuming these tests all costed an irrelevantly small amount of money].

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 19:31:04 GMT

[quote user="Lucy Fleming"]I guess to answer this question more directly - sounds like based on a pro-BNP level of 1300 in a Labrador - probably not. However, I wouldn't rule it out based on no murmur, pulse defecits or muffled heart sounds, as the idea of screening is to pick cases up in the pre-clinical stages, so there wouldn't be anything evident on clinical exam. For this case I would discuss with the owners that this may very well be a false positive result (especially given that Labs may have higher average levels than other breeds), discuss echo to look for evidence of DCM, and let them decide. I think DCM is slightly over-represented in the breed, but possibly only in certain lines? 3 years is quite young as well. [/quote]

DCM is one of the more difficult things to diagnose but there are normally some clues such as persistently raised HR, exercise intolerance, and enlarged heart on radiography.

We are in danger of needless diagnostics in these dogs with false positives, and may be doing harm, especially if insured and the premiums become unaffordable and if it gets heart disease then preventing treatment.

It just seems an incredibly irrational way of going about things to me - pharma research predicated on privately owned dogs using a poorly-referenced test.

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 19:27:16 GMT

[quote user="Niall Taylor"]And what about vaccination as an example of treating risk?[/quote]

Vaccination is different, for a few reasons. Partly it's a population measure, herd immunity etc. Partly it's protective rather than treating a disease. And we know (as much as we can know anything) that it works. We're not talking about medicating a dog for the rest of its life. We don't know that BNP works or what it tells us is reliable. Equally we don't know that drugs are protective at all stages of disease. This seems an incredibly odd study design to me.

[quote user="Niall Taylor"]If the numbers correlate with the echo results then wouldn't that be a reasonable approach?[/quote]

I personally don't think so. BNP is a seductively simple test but there are big grey areas. An echo should be achievable in most practices and should give more information. Possibly I'm skewed given my cardio interest but I can't think of a single time I would use BNP.

[quote user="Niall Taylor"]I thought there was a good case for statins given in advance of clinical problems.[/quote]

I think the jury is still out. Bit like the TB debate. Problem with treating risk in humans is the outcome measure of quality adjusted life years. Data easily skewed by outliers. There's probably a massive proportion of those treated to which it makes no difference whatsoever. There was an interesting conference in Oxford recently on "Too much medicine", programme and overview available here

https://www.phc.ox.ac.uk/events/too-much-medicine-exploring-the-relevance-of-philosophy-of-medicine-to-medical-research-and-practice/too-much-medicine-programme-180417-v3.pdf

[quote user="Niall Taylor"]Thanks, I'll see if I can get a copy.[/quote]

It's free here

http://www.bmj.com/content/348/bmj.g3725

[quote user="Niall Taylor"]Risk aversion?[/quote]

God help us all.

RE: Boehringer NT ProBNP study

Lucy Fleming — Mon, 22 May 2017 19:18:21 GMT

[quote user="Anthony Dennison"]Heart sounds fine on exam, not muffled, no murmur, no pulse deficits. Should I be concerned about a high proBNP value even if the dog is low risk for DCM and doesn't have a murmur?[/quote]

I guess to answer this question more directly - sounds like based on a pro-BNP level of 1300 in a Labrador - probably not. However, I wouldn't rule it out based on no murmur, pulse defecits or muffled heart sounds, as the idea of screening is to pick cases up in the pre-clinical stages, so there wouldn't be anything evident on clinical exam. For this case I would discuss with the owners that this may very well be a false positive result (especially given that Labs may have higher average levels than other breeds), discuss echo to look for evidence of DCM, and let them decide. I think DCM is slightly over-represented in the breed, but possibly only in certain lines? 3 years is quite young as well.

RE: Boehringer NT ProBNP study

Anthony Dennison — Mon, 22 May 2017 19:03:03 GMT

[quote user="Niall Taylor"]Starting this thread was presumably intended to help examine the ethics and usefulness.[/quote]

Actually, I started it to see if anyone else was doing it (because I hadn't heard about it) and to discuss accuracy of proBNP. Been an interesting discussion though!

RE: Boehringer NT ProBNP study

Niall Taylor — Mon, 22 May 2017 18:43:51 GMT

[quote user="David Mills"]Completely different set of circumstances...[/quote]

I know, but it occurred to me as an example of medicating to reduce risk in a pre-clinical animal in the context of a thread on heart disease. Hence my question. And what about vaccination as an example of treating risk?

[quote user="David Mills"]I imagine the vision is to start treating based on numbers. A far better thing to do would echo them for a definitive answer...[/quote]

If the numbers correlate with the echo results then wouldn't that be a reasonable approach?

[quote user="David Mills"]Critiques of EBM in the human world have shown a big pharma-funded trend for chasing risk (statins being a good example) because its cheap to do, and you can sell a huge amount of drugs... [/quote]

I take the point, but just because pharmacological firms sell lots of drugs as a consequence doesn't mean such a policy is necessarily flawed. I thought there was a good case for statins given in advance of clinical problems.

[quote user="David Mills"]And it seems vets are playing ball without a thorough examination of the ethics and usefulness of the whole she-bang[/quote]

Starting this thread was presumably intended to help examine the ethics and usefulness.

[quote user="David Mills"]see Greenhalgh et al "EBM: A movement in crisis?[/quote]

Thanks, I'll see if I can get a copy.

[quote user="David Mills"]Goodness knows why.[/quote]

Risk aversion?

Niall

RE: Boehringer NT ProBNP study

Lucy Fleming — Mon, 22 May 2017 17:32:01 GMT

[quote user="Martin Atkinson"]the vet at RVC suggested that Labs may have slightly higher normal NT-proBNPs than average[/quote]

Interestingly, I have been in touch with the cardiologist at Langford and they said the same, and have apparently just submitted a research abstract along those lines. I scanned a lab last week with a pro-BNP of 3000 and couldn't find any evidence of DCM, apparently that's not that surprising for the breed.

FWIW, we have been performing the sponsored pro-BNP tests, and discussing echo with owners of those with elevated results. I can't see a situation where I would start a clinically unaffected dog on meds based on a pro-BNP alone, but can see it as potentially being a useful screening test for at risk animals (based on breed). Given that the survival times after diagnosis with overt DCM are generally quite poor, I can see a value in actively screening for preclinical cases (though I'm not generally a fan of 'fishing' for abnormalities)

RE: Boehringer NT ProBNP study

Martin Atkinson — Mon, 22 May 2017 17:12:47 GMT

[quote user="David Mills"] I imagine the vision is to start treating based on numbers. A far better thing to do would echo them for a definitive answer - who would suggest that??[/quote]Because despite your reservations NT-proBNP is a useful test and a fraction of the cost of echo. If nothing else it alerts you to the possibility of cardiac disease then you can start performing additional investigations.

There may be a few false positives: one of my friends had a middle aged lab with a grade 2 systolic murmur and increased VHS and raised proBNP of 1687. I referred it for echo and that was normal and the vet at RVC suggested that Labs may have slightly higher normal NT-proBNPs than average. So we decided to test again in a year to see if there was a rising titre. Most of the dogs I've seen with cardiac failure ave had titres of 3000 or more. Sadly she died of unrelated causes in the meantime.

[quote user="David Mills"]Critiques of EBM in the human world have shown a big pharma-funded trend for chasing risk because its cheap to do, and you can sell a huge amount of drugs. This to me is what this study is indicative of. [/quote]Sorry, if as Arlo suggested, you can't give vets credit for looking at the evidence and making their own decisions on whether treatment is necessary. What is wrong with another arrow in your quiver to give you information.

RE: Boehringer NT ProBNP study

David Mills — Mon, 22 May 2017 16:13:26 GMT

[quote user="Niall Taylor"]I would be interested in your take on the PROTECT study and why you seem to think we shouldn't be medicating in order to reduce or delay the future risk of developing cardiac disease.[/quote]

Completely different set of circumstances old fruit. The protect study was looking at those with heart disease, as defined by the inclusion criteria.

Critiques of EBM in the human world have shown a big pharma-funded trend for chasing risk (statins being a good example) because its cheap to do, and you can sell a huge amount of drugs (see Greenhalgh et al "EBM: A movement in crisis?") This to me is what this study is indicative of. And it seems vets are playing ball without a thorough examination of the ethics and usefulness of the whole she-bang. Goodness knows why.

RE: Boehringer NT ProBNP study

Michael Woodhouse — Sun, 21 May 2017 01:08:14 GMT

[quote user="Martin Atkinson"]PS: are you going to one star me again (Mr Mills)?[/quote]

No, that was me.