snap cPLI diagnostic value

Re: snap cPLI diagnostic value

Kate Richardson — Sat, 24 Mar 2012 16:43:38 GMT

I know I am nit picking! But statements such as that can become fact very easily in some people's minds.

Re: snap cPLI diagnostic value

Graham Bilbrough — Sat, 24 Mar 2012 12:02:01 GMT

Hi, Kate

I do see your point, but I think the statement is still reasonable in the context of a fax-shot requiring brevity. I don't think I wrote the document and I don't know what the rest of it says. However, I am sure we could find numerous peer-reviewed papers with statements like "anaemia was ruled-out with a PCV". To be more precise, if the patient was not dehydrated... Or, "the Total T4 concentration excluded hypothyroidism". To be more precise, if we'd checked for auto-antibodies...Or, "the cat's total T4 concentration ruled-out hyperthyroidism". Well, how do we know that cat's total T4 has not increased dramatically despite it still being within the normal reference interval?

IDEXX has a team of diplomats in internal medicine and clinical pathology that do a great job of providing more detail on a case-by-case basis, whether the testing is done in-clinic or a the reference lab.

Thanks

Graham

Re: snap cPLI diagnostic value

Kate Richardson — Sat, 24 Mar 2012 11:04:06 GMT

Hi Graham,

[quote user="Graham Bilbrough"]IDEXX cannot (and does not) claim that Spec/SNAP cPL performs perfectly. For example, I have seen patients, that after many years of pancreatitis, have a normal [cPL]serum. [/quote]

Our practice received a fax from Idexx last month as did many people I would expect where it states in bold

"60%+ of the Spec cPL and fPL tests that were sent directly to the Reference Laboratory without the prior running of a SNAP were negative, therefore ruling out pancreatitis"

So does or doesn't Idexx claim the Spec/SNAP cPL performs perfectly? You are saying, and I agree, that dogs can have pancreatitis but have a normal cPL, the fax says that if they have a negative cPL they don't have pancreatitis? Or am I reading it wrong I think the last few words of the sentence in bold should read therefore making pancreatitis unlikely

Re: snap cPLI diagnostic value

Graham Bilbrough — Mon, 20 Feb 2012 08:29:33 GMT

Normal 0 false false false EN-GB X-NONE X-NONE MicrosoftInternetExplorer4

Good morning, all

I kept away from the Internet on Sunday and the conversation has moved on considerably--I am getting behind with my replies! I think I can answer most with this summary:

1. Ultrasound (by comparison to histology) will do a great job, with the right equipment, and in the hands of an expert. This is rarely available unless the dog is submitted to a referral practice. Despite this, all of the referral (diplomat status) centres that I am aware of (I can only quote from the UK) make use of the cPL assays, presumably because the sensitivity of their ultrasound is less than perfect.

2. There is no blood test that performs better than Spec/SNAP cPL. Lipase has a horrible specificity. IDEXX cannot (and does not) claim that Spec/SNAP cPL performs perfectly. For example, I have seen patients, that after many years of pancreatitis, have a normal [cPL]serum. I postulate that there is so little pancreatic tissue left that, even if what remains is inflamed, insufficient pancreatic specific lipase if released to the circulation to "cross that bar". But, don't we see the similar issues with cirrhotic livers and liver enzymes?

3. Owners respond differently (rightly or wrongly) when you present them with a diagnosis, rather than a clinical suspicion.

4. SNAP cPL is designed to give you as much confidence “in the normal" as you would get with Spec cPL. In doing so, it dramatically reduces the turn-around-time (you can have the result in mind whilst you are looking at your in-clinic haematology and biochemistry reports) and the average price to the pet owner.

5. All the major reference labs (I can only quote for the UK) pass on samples to IDEXX Laboratories for Spec cPL. I think it is reasonable to assume they see benefit to the assay.

6. I quoted the Tivedi et al paper as this is independent of the Texas A&M group.

All the best

Re: snap cPLI diagnostic value

Christopher Saul — Mon, 20 Feb 2012 00:28:40 GMT

[quote user="Kate Richardson"]

[quote user="robloxley"]There's no perfect test, so yes adding clinical signs, ultrasound etc. is the way to improve the reliability of the diagnosis.[/quote]

And that is the crux of it IMO

[/quote]

I second that

Re: snap cPLI diagnostic value

Kate Richardson — Sun, 19 Feb 2012 11:00:57 GMT

[quote user="robloxley"]There's no perfect test, so yes adding clinical signs, ultrasound etc. is the way to improve the reliability of the diagnosis.[/quote]

And that is the crux of it IMO.

Re: snap cPLI diagnostic value

Rob Loxley — Sun, 19 Feb 2012 10:28:35 GMT

[quote user="Jane Adams"]Caroline Mansfield is our head of medicine here at Melbourne Uni and does not rate cPLI as a diagnostic aid for pancreatitis.[/quote]

She may want to get IDEXX to change their marketing material, as this is from one of the spec cPL brochures: “Certainly, it [cPLI test] is the single test that is most useful for the diagnosis of acute pancreatitis...” Caroline Mansfi eld, BSc, BVMS, MVM, DECVIM-CA2

[quote user="Jane Adams"]In which case serum lipase, along with clinical signs, is still more reliable than cPLI[/quote]

Certainly having a read round, cPLI is reported to perform better than lipase, but correct me if I'm wrong. A lot of the papers supporting cPLI come from the Texas A&M group that developed the cPL test, but they're in major peer reviewed journals?
There's no perfect test, so yes adding clinical signs, ultrasound etc. is the way to improve the reliability of the diagnosis.

Re: snap cPLI diagnostic value

jd2008 — Sun, 19 Feb 2012 09:50:43 GMT

[quote user="robloxley"]

[quote user="Jane Dunnett"]Ultrasonography is the most reliable diagnostic test in combination with clinical signs[/quote]

Depending on the ultrasonographer... For lots of primary practices a blood test may perform better (if still not perfectly).

[/quote] In which case serum lipase, along with clinical signs, is still more reliable than cPLI

Re: snap cPLI diagnostic value

Rob Loxley — Sun, 19 Feb 2012 09:47:44 GMT

[quote user="Jane Dunnett"]Ultrasonography is the most reliable diagnostic test in combination with clinical signs[/quote]

Depending on the ultrasonographer... For lots of primary practices a blood test may perform better (if still not perfectly).

Re: snap cPLI diagnostic value

jd2008 — Sun, 19 Feb 2012 01:52:14 GMT

Caroline Mansfield is our head of medicine here at Melbourne Uni and does not rate cPLI as a diagnostic aid for pancreatitis. Ultrasonography is the most reliable diagnostic test in combination with clinical signs.

She has done a large amount of research into pancreatitis and is currently researching other biochemical markers. I'll dig out some of her more recent published papers and post references on here if anyone's interested.

Jane

Re: snap cPLI diagnostic value

Kate Richardson — Sat, 18 Feb 2012 14:35:16 GMT

Personally I don't use the snap test, because if I suspect the dog has pancreatitis, I treat it as pancreatitis, and would prefer to wait for a quantitative PLI result as the speed at which I can obtain a result doesn't change my treatment plan, so I feel it adds little extra value. Idexx can usually turn around a PLI result within 48hours.

Re: snap cPLI diagnostic value

Anonymous — Sat, 18 Feb 2012 14:26:48 GMT

[quote user="Mark Holmes"]

[quote user="Virginia Campbell"]

[quote user="Graham Bilbrough"]pet owners appreciate a diagnosis, rather than a strong suspicion, and warning of the potential sequelae e.g. diabetes mellitus and exocrine pancreatic insufficiency.[/quote]

I like the test for this reason; it's additional evidence to make me want to put the dog on a low fat food for life (as long as the pancreatitis wasn't just caused by one massive bin raid in an otherwise slim, healthy, non-schnauzer with no previous history of vomiting) and helps to impress upon the owner the importance of not feeding fatty scraps or big meals

[/quote]

Yet all those things would apply even if the dog doesn't have pancreatitis.

[/quote]

Agreed.

I also think it's important to separate discussion of acute pancreatitis and chronic pancreatitis - I believe the 2 syndromes to be possibly quite different. I'm not sure there's much evidence for low fat life-time diets for acute pancreatitis for instance. It's the clinical utility (or not) of the Spec CPL test in acute pancreatitis cases that I'm interested in determining, I don't think there is any evidence that it is of value in cases of chronic pancreatitis?

Re: snap cPLI diagnostic value

Mark Holmes — Sat, 18 Feb 2012 14:00:27 GMT

[quote user="Virginia Campbell"]

[/quote]

Yet all those things would apply even if the dog doesn't have pancreatitis.

Re: snap cPLI diagnostic value

Virginia Campbell — Sat, 18 Feb 2012 13:28:21 GMT

Re: snap cPLI diagnostic value

Anonymous — Sat, 18 Feb 2012 11:00:46 GMT

I would suggest that the specificity of SpecCPL at the >400 cut-off is perhaps 60-80% (i.e. ability to confirm pancreatitis) and thought that the sensitivity at the <200 cut-off was perhaps much better (maybe 90-95%) (i.e.ability to rule out pancreatitis), but that Trivedi et al. paper seems to call that into question also:

"I think at the 200 cut-off point, the Spec CPL had a sensitivity for moderate/severe pancreatic inflammation of 71% (with hopelessly wide confidence intervals of 29%-96%) based on picking up 5 out of 7 cases.

At the same 200 cut-off point, the sensitivity for mild pancreatitic inflammation was 43% (CI=30%-57%)." [quoted from earlier post I made re their results]

When making comparisons to standard lipase tests, it is important not to then take a single cut-off point with these, otherwise the comparison is fairly meaningless (like comparing those values given on SpecCPL with SpecCPL sens and spec with a single cut-off of 300 say). Unfortunately, I don't think this work has sensibly been approached. What is the specificity of standard lipase at a cut-off value of e.g. 2 x the upper reference limit and what is the sensitivity at e.g. half the upper reference limit? Comparisons based on single cut-off points for quantitative tests can rather miss the point of clinical utility.

Designing a study to assess the clinical utility of pancreatic inflammation biomarkers is of course full of difficulty. What is the gold-standard one is comparing to? Are we interested in the presence of pancreatic inflammation or in the 'clinical syndrome of pancreatitis'? How will this test alter treatment/diagnostic decisions etc? But even ignoring these limitations in the studies performed to date, I still struggle to see a clear benefit of SpecCPL over lipase (and/or experienced ultrasonography - another can o' worms...)

Re: snap cPLI diagnostic value

Anonymous — Sat, 18 Feb 2012 10:36:52 GMT

[quote user="robloxley"]My understanding is that the spec cPL is fairly sensitive and specific, definitely more than lipase or TLI, and probably more so than an experienced ultrasonographer,[/quote]

That was my understanding also, based on the promotional material at the time of the test launch...

I had perhaps misunderstood the difference between educational material from a lab and promotional material in a format I'm more familiar with from drug-companies (i.e. read the small print etc.)

When I looked for actual published data a couple of years later, I was somewhat less convinced that SpecCPL was actually significantly more useful than lipase or experienced ultrasonography.

Re: snap cPLI diagnostic value

Rob Loxley — Fri, 17 Feb 2012 23:26:46 GMT

[quote user="Christopher Saul"]In trying to sum up this conversation briefly, Graham, would it be fair to say that the SNAP PLI test has a very high sensitivity (and hence low false negative rate, making it a good rule out test) - though slightly lower specificity[/quote]

I think you need to define your standard: AFAIK the work for the in-clinic snap cPL test is just against the laboratory spec cPL test. One then needs to look at how well that test performs to see how well the snap will perform in 'making a diagnosis' (as per the Trivedi paper).

My understanding is that the spec cPL is fairly sensitive and specific, definitely more than lipase or TLI, and probably more so than an experienced ultrasonographer, and therefore has utility in determining pancreatitis cases, but isn't the be-all-and-end-all? Graham may correct me if my understanding is wrong.

As to running the snap test in house, my personal view is that it gives you some extra, rapid information, while waiting for the lab result, and the pricing structure from Idexx makes it economically sound.

Re: snap cPLI diagnostic value

Anonymous — Fri, 17 Feb 2012 19:35:54 GMT

I think this is how it is marketed - but I'm not sold yet that it does have a very high sensitivity. As far as I can tell (and there is a fair chance that I've got totally the wrong end of the stick...) the independent study Graham pointed me to demonstrated that at a cut-off of 200 for the SpecCPL, 2 out of 7 cases of moderate/severe pancreatitis were missed - based on that data it seems a less useful rule-out test than might have been suggested?

Re: snap cPLI diagnostic value

Anonymous — Fri, 17 Feb 2012 19:30:45 GMT

[quote user="Graham Bilbrough"]Perhaps if I send you mine, you could send me yours![/quote]

Thanks. My email address is d.beattie@gortlandsvets.com

The one I mentioned on SNAP vs SPEC CPL and gave reference to was performed by idexx staff I think - so you should be able to get a copy. If not, then let me know. It was maybe done quite a while ago with a long lag-time to journal publication in 2011.

[quote user="Graham Bilbrough"]If you or another vet (preferably one regularly submitting Spec cPLs) would like to do another "field study", I can fund the SNAP cPL devices.[/quote]

Can't say fairer than that! But from the sounds of it you've already done that study and have that data, so I might be quite satisfied on seeing it.

[quote user="Graham Bilbrough"]Turning to an earlier comment on the sensitivity of the test for _mild_ pancreatitis, I would argue much of it comes down to how you define severity. [/quote]

Re-read my post - I was more interested in the moderate/severe pancreatitis cases (or lack thereof).

[quote user="Graham Bilbrough"]I would suggest that either test (SNAP or Spec) does a very good job "spotting" dogs with clinically significant pancreatitis.[/quote]

Suggestion noted, but I'll need slightly more empirical evidence of the 'very good job' if I'm going to be able to interpret the test helpfully in a clinical setting and therefore have more inclination to use it - the way I look at it is I'm an assistant and to suggest to my boss that I wish to stock the SNAP CPL tests, I'd need to be able to give a rational reason for this, and I'm still struggling.

Re: snap cPLI diagnostic value

Christopher Saul — Fri, 17 Feb 2012 17:26:55 GMT

[quote user="Graham Bilbrough"]We (IDEXX Laboratories) recommend a Spec cPL, for quantification, only if you find an abnormally high concentration of cPL using the SNAP test. If you find a normal concentration with the SNAP device, you can stop there.[/quote]

In trying to sum up this conversation briefly, Graham, would it be fair to say that the SNAP PLI test has a very high sensitivity (and hence low false negative rate, making it a good rule out test) - though slightly lower specificity, in the same way as, say, SNAP FIV/FeLV ???

Re: snap cPLI diagnostic value

Graham Bilbrough — Fri, 17 Feb 2012 14:31:26 GMT

Hi, David

The study I have has an n of 70 dogs, but it does sound like you might have been reading an earlier draft. Perhaps if I send you mine, you could send me yours! My study was performed by veterinary professionals (mix of nurses and vets in general practice).

If you or another vet (preferably one regularly submitting Spec cPLs) would like to do another "field study", I can fund the SNAP cPL devices.

Turning to an earlier comment on the sensitivity of the test for _mild_ pancreatitis, I would argue much of it comes down to how you define severity. I would suggest that either test (SNAP or Spec) does a very good job "spotting" dogs with clinically significant pancreatitis. It is very similar to using bile acids or so many other analytes. If you declare an animal has hepatic dysfunction at 100umols/litre (post-prandial), you will miss mild hepatic dysfunction. If you set your threshold at 50units (as some clinical pathologists do), you will find relatively poor specificity-lots will be dogs on NSAIDS with livers that work! It is the usual compromise.

All best wishes

Graham

Re: snap cPLI diagnostic value

Graham Bilbrough — Fri, 17 Feb 2012 14:13:03 GMT

Hi, Martin

If the sample spot was the same intensity as the reference spot, then the SNAP was reporting an abnormally high concentration of cPL and the recommendation would be quantify with Spec cPL. I would argue that the screening test worked. But, given that the value at the reference lab was so high, I would expect you to say the sample spot was of relatively high intensity. Do you still have any sample? Did you notify our technical services team?

Did you send the SNAP device in with your submission? Let me have the reference number (by email) and I will investigate if you missed the discount.

Thanks

Graham

Re: snap cPLI diagnostic value

Martin Atkinson — Fri, 17 Feb 2012 13:29:32 GMT

[quote user="Graham Bilbrough"]

The SNAP cPL device correletes very well with Spec cPL. We (IDEXX Laboratories) recommend a Spec cPL, for quantification, only if you find an abnormally high concentration of cPL using the SNAP test. If you find a normal concentration with the SNAP device, you can stop there. The pricing structure at IDEXX is such that even if you do both SNAP cPL and Spec cPL for a given patient, it is cheaper than doing Spec cPL alone.

[/quote]

I have to say that I would not like to rely on it as the colour change can be subtle sometimes. Recently had a dog with acute abdominal pain the SNAP test colour change was barely distinguishable from the normal spot and the lab sample Spec cPL came back as >1000. As regards cost I'll quote your observation here Graham, the last one I sent for spec cPLi along with the SNAP kit certainly wasn't discounted that much.

Re: snap cPLI diagnostic value

Anonymous — Fri, 17 Feb 2012 10:50:32 GMT

[quote user="Graham Bilbrough"]After 20 dogs or so, you will have great confidence that a normal result on the SNAP cPL would give a value of less than 200units on the Spec cPL.[/quote]

I may be looking at the wrong concordance study, as the study in the paper I quoted in the last post involves 49 dogs I think (not 70), of which about 27 dogs tested 'normal' on at least one occasion with a SNAP cPL. Each dog was tested 3 times with a SNAP CPL. On 2 of those batches of SNAP CPL, 23 out of 25 'SNAP-normal dogs' had SpecCPL <200, while on 1 of those batches 24 out of 27 'SNAP-normal dogs' had SpecCPL <200. I gather, I think, from the discussion that those discrepant results all had SpecCPL <270, and that one serum sample was consistently discordant across all 3 lots of the SNAP tests, while discordant results for the other 4 samples occurred only once per lot.

So, from those results it appears that, in strict laboratory testing conditions, a 'normal' SNAP reliably indicates at least a SpecCPL <270, and more than likely this will be <200. [Of course I'm still struggling a little as to how useful konwing the SpecCPL is <200 (let alone <270) is]

I may have missed who was interpreting the SNAP tests, but I've assumed it was a lab-trained technician who is used to inerpreting colour changes in a lab setting etc?

A further interesting study could be performed if you analysed the data from all the vets sending in for SpecCPL who have already performed a SNAP CPL (which I understand from your earlier post would be most vets as it is cheaper to do it this way). This would assess the in-clinic test in 'field' conditions. I would be interested to see what % of 'normal' SNAP CPL tests as measured in veterinary clinics had a SpecCPL <200 over, say, 200 samples done this way.

Re: snap cPLI diagnostic value

Anonymous — Fri, 17 Feb 2012 09:47:51 GMT

[quote user="Graham Bilbrough"]For an independent review of the utility of Spec cPL, I can point you at Trivedi et al 2011 (Stanley Mark's group) in JVIM.[/quote]

Thanks, Graham. I'd skim-read this just yesterday after looking to see what new had been published. Arguably the best study to date of its kind to date I suspect.

I think at the 200 cut-off point, the Spec CPL had a sensitivity for moderate/severe pancreatic inflammation of 71% (with hopelessly wide confidence intervals of 29%-96%) based on picking up 5 out of 7 cases.

At the same 200 cut-off point, the sensitivity for mild pancreatitic inflammation was 43% (CI=30%-57%).

Even if I assume a >96% correlation between a 'normal' SNAP CPL and a SpecCPL <200, that doesn't exactly suggest a 'normal' result is likely to rule-out pancreatitis, so I still struggle to see much value to the clinical utility of the SNAP test?

[quote user="Graham Bilbrough"]The concordance study between SNAP and Spec was with 70 dogs.[/quote]

If that is:

Performance validation and method comparison of an in-clinic enzyme-linked immunosorbent assay for the detection of canine pancreatic lipase. Beall, M. J.; Cahill, R.; Pigeon, K.; Hanscom, J.; Huth, S. P Journal of Veterinary Diagnostic Investigation 2011 Vol. 23 No. 1 pp. 115-119

then I can access it, but have not done so yet.

[quote user="Graham Bilbrough"]On a slightly cheekier note, I could suggest you speak to those reference labs that buy our SNAP cPL devices and use those to provide results. I cannot give you their names, but you'll find them![/quote]

Outrageous! Whatever has happened to the ethics of commercial diagnostic laboratories...