Based on Master Nickson's comments on the PE debate, you could argue this would be an acceptable paradigm. Using Wells as your entry forces gestalt into the equation. Since Wells' low risk arguably gets you somewhere between 1-6% in ED populations, PERC should be acceptable.
Update:
- PE DX by Jeff Kline
- Clinical Guidelines from ACP include intermediate d-dimers, age-adjusted d-dimer (Ann Intern Med 2015;163:701)
- Likely pretest prob patients are ruled out by neg CTA (Safety of multidetector computed tomography pulmonary angiography to exclude pulmonary embolism in patients with a likely pretest clinical probability. J Thromb Haemost. 2017 Jun 2. doi: 10.1111/jth.13746.)
- Normalization of Vital Signs does not reduce probability of PE (https://coreem.net/journal-reviews/vs-normalization/)
- I actually now use Modified Geneva score
Editor-in-Chief, at EMCrit.org
An ED Intensivist from NY.
Professor
Nassau University Medical Center
No conflicts of interest (coi).
Professor
Nassau University Medical Center
No conflicts of interest (coi).
Latest posts by Scott Weingart, MD FCCM (see all)
- EMCrit Wee (392.5) – Naughty or Nice? Bad Behavior in Healthcare with Liz Crowe, PhD - January 15, 2025
- EMCrit 392 – All Things Defibrillation with Sheldon Cheskes - January 10, 2025
- EMCrit 391 – Pericardiocentesis and Tamponade Temporization - December 27, 2024
Ok Weingart, i’ll bite. Fueled by the fact that i am post shift, it is 3 am and the dogs woke me up and i have had a beer. This algorithm is completely nuts! Maybe i am an ignorant slut but…….. PERC was studied and validated as a stand alone rule to obviate the need for testing AT ALL. Unless Jeff Kline has done a bunch of new, large studies incorporating PERC + dimer or PERC + Wells or PERC + wells+ dimer then all of these new algorithms are just more mental masturbation strewn with unproven assertions and leading… Read more »
Mike,
You lost me there. PERC negative in the diagram = stop w/u; no d-dimer or imaging.
PERC use already acknowledges that we are not trying to get to zero risk. The diagram is built-in acceptance of the 2% acceptable miss threshold.
Tell me my friend where we have parted ways on our journey????
Hi Scott I’m an rn. Just wondering. Stable looking patient. Absolutely no distress. Reports some recent sob and generalized weakness (has a hard time getting things off the shelf at his job as a dishwasher. Dry cough. Hr 107. 35 male. 101.4. 1 episode of blood tinged sputum “a little bit”. Was worked up for pneumonia. Cxr normal. WBC nl. Resident was suspicious for pe. Now is like “ok ctpa for pe”. Is it ethical to send this patient for a ctpa without sending a d-dimer? I did not protest but it seems like the chances of this pt dying… Read more »
Perhaps more succinctly: is it ever ok to scan a stable patient without a d dimer?
You are partly right Mike.
PERC can be used in this setting but after the Well’s moderate arm. Well’s low is low enough, we do not need to double check this one, but if you get a moderate just double check with PERC- if it passes PERC do not bother with a D-dimer.
Personaly I do the opposite- if someone fails PERC, I go on to do a Well’s to see if I can stop doing a D-dimer. This was very elegantly suggested by Dave Newman in his PE talk on smartem.org
Martin,
PERC can absolutely not be used after Wells moderate, not sure where you got that. As to Wells low being low enough, that is merely conjecture at this stage. David’s assertion very well may turn out to be true (most of them are), but the evidence is contradictory and you certainly are unlikely to get below 2% unless you are including a bunch of pts who should not get a PE work-up in your cohort.
Your assertion that a moderate Wells can be ruled out with PERC is dangerous and not even suggested by David.
If I understand it all correctly, PERC was intended to rule out someone who is low risk for PE by clinical gestalt. Adding Wells to the mix permits an objective way to call a patient low-risk. I completely agree with the point Dr. Jasumback that we should not try to find and treat every single PE, however I do not think that we need a randomized control trial to validate the algorithm above because it makes sense and was built into PERC as clinical gestalt (…aren’t prediction rules simply objective clinical gestalt anyway)?
Haney
Haney, absolutely correct! In fact most people are not using the PERC as it was validated, b/c they are supposed to first ask themselves the question of whether they believe the chances of PE are <15% before they even think about using PERC. Unfortunately as Nickson and other listeners have pointed out, this is still not objective b/c the only thing we are using Wells for is to get the gestalt piece.
s
S,
Sober and well-rested, and your diagram looks good. in short, PERC helps you find those patients whose pre-test prob of PE is below the d-dimer test threshold. also, of important note, it was validated on patients who stated “Shortness of breath” was the reason they came to the ED (or, of equal importance to cc), so applying it to every misc “chest pain” patient is a bit of an EBM stretch… Although seems to be done on a routine basis. interested in your thoughts…
Matt
EMCC
Matt, do you mean Wells or PERC for SOB. Wells has been done on a ton of different ED populations, and in all the low risk is <6% (safe for PERC territory). PERC has also been looked at in a few, though less, ED populations; though as Dr. Newman points out nobody has yet done the study where you use PERC and send the patient home based on it and then follow these patients rather than just collecting PERC and seeing if they have a PE or not.
s
Updated the diagram with an important first question
Scott, I think it was around the 2nd beer that things might have gone south, but perhaps not…… More succinctly this time, then. My understanding of PERC is that PERC negative gets you to no testing. I’d have to go back an look at the derivation and confirmation studies. But basically I think it went something like, I think this pt COULD have a PE. If PERC- then no testing. But now we’ve added stuff. So now it takes WELLS low + PERC – to get to no testing. And I don’t know that that assertion has been tested. With… Read more »
PERC was validated by first asking the clinician what their gestalt is for PE. If you do Wells’ first, all this adds is it forces you to answer this gestalt question. Nothing else in Wells’ is different than PERC (even malignancy alone doesn’t take you away from low). No need for validation, this is just a way to get people to do what they are supposed to do any way. Not sure of the smell test, but high sens d-dimer is probably superior to most imaging we have available. The numbers are there in countless studies. ACEP’s clinical policy will… Read more »
One more thing… What this illustrates is the problem in all of clinical medicine, the a priori point estimate of probability of disease. This is not, and can never be, objective. Why, because you have to start somewhere. In your new algorithm, the first box illustrates that. What is that driver, that symptom complex, that gut instinct that gets you to “This patient might have a PE” ? It is from there that you have to generate a point estimate of disease probability, appropriately use a tool of some sort that has been verified in a population (but most likely… Read more »
I don’t look at it as a problem, I look at it as the reason for our existence. If everything could be protocolized, what is the point of physicians?
Scott, I like the thought process and algorithm, but Newman has made the point that PERC and Wells are mutually exclusive and can be used separately; if either is negative the workup can stop. However, I generally use them together as well, though I might move PERC above Wells in the algorithm. i.e.–PERC negative => stop the workup, PERC Positive => Apply Wells criteria as you have written. This makes a little more sense from a binary perspective. I don’t think you need the first question either, since that is already addressed in the clinical gestalt portion of PERC. I… Read more »
Only reason for the first question is that some have started using PERC and d-dimer as SCREENS for PE, which actually increases imaging rate instead of decreasing it. This is one point that Dr. Newman and I agree on. If you don’t think the patient has a PE, don’t do PERC, Wells, or d-dimer…
I agree that inappropriate use of D-Dimer (and medical decision make rules in general) is a travesty in medicine, but for the purposes of building an algorithm I’d put clinical gestalt in there only once. I think PERC uses a suspicion of <15% which is probably adequate for the first question.
JB
Don’t mean to belabor this, but nope. First question asks you is your pretest < 2% before starting a work-up--if so stop. PERC gestalt question asks you, "Ok, it is >2%, but is it < 15%?" Two very different questions.
I think it gets hard to quantify gestalt, but I see your point and that does make more sense in the context of the algorithm. . .thanks for the forum.
JB
Maybe I misunderstood PERC (and I haven’t gone back and looked at the validation study) but I thought it was stand alone for the low risk pt (<12-15% ), and that's where my hearburn lies. I have always found that the clinical gestalt and Wells rules aligned closely enought that I never applied Wells formally to get to that low risk number. I have also undestood that the low risk pt (<12-15%) with negative d-dimer got the post-test probability low enough to stop testing. As for the clinical assesment question, I agree, that is the reason for our existence as… Read more »
Mike,
Absolutely true, you have no need of Well’s before PERC, you can just answer the gestalt question. When you put Wells next to PERC what you will find is the only thing putting it before PERC does is makes you answer the gestalt question. There is no other effect.
Good d-dimer will take your low and
intermediate patient
to a post-test below test threshold.
as to essentials, sounds good!
And by the way, don’t you people have lives? It’s the 4th of July for Gods sake. Go hold a firecracker or something. (just don’t do it in far northern California, I don’t have a hand surgeon either)
At 3 am there were no comments on this now we are at what, 20? We need lives..
oh well, sober, moderately well rested and off to the 4p-1a shift, yippeee
Mike
I’m working; I don’t know of the others’ excuses, but remember most of our audience is not in the States and some probably rue the 4th as loss of domain.
Scott,
Reference for the Intermediate Probability Pt? I’ve been using dimer for low prob pts. But not brave enough to get into those I’d call intermediate i.e. >12-15% pretest probability
Mike
Here are just a few from a quick search, there are quite a few more out there:
Thromb Haemost 1998;79:32–7.
Arch Intern Med 2001;161:447–53.
Hematology. 2009 Oct;14(5):305-10.
AJR Am J Roentgenol. 2009 Aug;193(2):425-30
Systematic Reviews
Thromb Haemost. 2009 May;101(5):886-92.
S,
I meant PERC, but mea culpa, while the initial PERC validation was limited to those SOB pts, indeed the larger prospective MCT validation included almost half with cc of CP…
Thnx, lead on,
Matt
(less rested, less sober… Happy 4th!)
Hi Scott et al
I love it when you Yanks get drunk and have a fight!
I had questions about imaging and your proposed algorithim, please. What imaging do you refer to?
And in the Wells criteria HIGH RISK in your algorithim you proceed straight to imaging. If that is reported as negative, what would you do next? Perform a sens D Dimer and back track? Or rely on your gestalt and ignore the imaging report?
thanks and appreciate the debate here for my own learning
Hi Minh and the rest of the EMCrit crowd, This what we do at Charlies once we’ve decided we need to do imaging (i.e. low/int PTP and positive D-dimer OR high PTP based on Wells with D-dimer not indicated): CTPA negative • Rules out PE if low or intermediate PTP on Wells score • If high PTP do CTV or leg USS – if negative then NO PE CTPA positive • PE regardless of low/ int/ high PTP Certain patients we don’t (usually) do CTPAs on: • Female <45y • Pregnant • CTPA contraindicated These patients have VQ SPECT (ideally… Read more »
HI all I have a neologism for you, one I constantly employ – lets call it “schizogestalt”. Definition: Schizo`ges`talt This is when you have a patient referred by helpful nearby doctor who orders a blunder-bust of blood tests including a D-Dimer – which is weakly positive, then sends the patient in for a ?PE workup to ED. The patient looks great, healthy, and maybe has a few risk / Well’s points. However, it all hangs on the big 3 points for: “PE is most likely diagnosis”. So here is what you do – you split your brain in half and… Read more »
The minefield that is VTE…..so many different pathways so many different views. Should keep researchers in a job for many a year to come. At the same time will mortality and morbidity from VTE actually reduce, or we we just find more small PE that would never cause harm?
“At the same time will mortality and morbidity from VTE actually reduce, or we we just find more small PE that would never cause harm?”
This was in the discussion but the Weiner paper in May 2011 in Archives of Internal Medicine answers that pretty definitely as the latter:
http://www.ncbi.nlm.nih.gov/pubmed/21555660
Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med. 2011 May 9;171(9):831-7.
great reference, Trueg!
Great stuff guys enjoyed watching the video till the 60 min mark and my cellphone bandwidth allowance ran out! I’m on holidays and away from the joys of broadband! My two cents (or maybe more than that…) I think Scott’s added question (do you REALLY think it’s a PE) in the image is key – the thing with all the PE studies is that someone has to suspect a PE as a possibility to even get into the study and have the 50 point data sheet used. So the factors that go into why a doc enrolls them in the… Read more »
First, thank you all for the incredible comments! To answer some of the questions posed in the thread: imaging We use CTPA for pretty much everyone, unless their renal function is crap. My personal belief (and I think Jeff Kline’s as well) is that even in high risk, a well done CTPA negative = done. PIOPED II unfortunately got it totally wrong. I believe an angiogram is no better than CTPA at this point, and a negative CTPA probably predicts pts with a good prognosis without treatment. I think it is clever to get d-dimer in the high risk pts… Read more »
thanks Scott et al I am not convinced by your advice to perform a D Dimer in the high risk patient with a negative CTPA. It would seem more logical to do the cheaper , radiation free test first in the high risk group. What would you do if you had a +ve D Dimer and -ve CTPA in a high risk patient? Could you not assess for Wells criteria and if high risk then do the D Dimer. If positive you treat? If negative then you are done? With safer anticoagulation options already for things such as AF and… Read more »
Minh, D-dimer is a non-specific test: positives mean virtually nothing. Certainly not enough to move even a high risk patient above the “treatment” threshold. High risk patient with Neg CTPA and + d-dimer needs a dvt study, whether it is bedside or formal ultrasound or CT leg veins based on the current evidence. Giving a false + dx of PE has ramifications far in excess of the initial anti-coag. Every time this patient comes to an ED with chest pain, sob, or tachycardia they are going to wind up with a PE w/u. They will have problems getting life insurance… Read more »
Scott, I admire your energy for teaching and debate when you are up posting at midnight! I don;t think the issue is as black and white as it may initially appear. Andy is right that this is a grey area of clinical practice. Take the use of gestalt in the decision making. This is so non standardised that its use is hard to define. In essence we are saying gestalt=clinical experience and that as we know is not something uniform nor standard. It is as you say the reason why we still need human brains in point of care medicine.… Read more »
Minh-of course we will always need to be clinicians in the gray zones. I’m not sure how I am contradicting myself. The nature of a sensitive, non-specific test is that a negative can have dramatic effects on your w/u, while a positive doesn’t affect your probability and you continue the w/u as if the test did not exist. If there was no such thing as d-dimer, you would need to proceed to further testing after a negative ct in a high risk patient (at least in the vision of the US organizations), so you do the same in the case… Read more »
thanks Scott
What if you were in a remote area, like Antarctica? You got someone with High risk Wells presentation and positive D Dimer? Do you wait for the CTPA before starting anticoagulation?
Sure you can get a portable USS for DVT but if that is negative do you still wait for the CTPA in the tachycardic, tachypnoeic, pregnant patient? Even though it will take 3 days to retrieve them to a CTPA machine? And by that time the PE signs may have resolved on imaging?
2 things – why would you even bother with a d-dimer in this high risk patient in Antarctica? And secondly, in general the selection of people to head South tends to avoid sending pregnant women…as for what happens when folks get there, well, I have heard some stories…
OKay good point, my use of pregnant patient was not the best example. My point really is what would you do if you had to wait a few days, weeks to get your CTPA in that high risk patient. Is it so vital to your diagnostic workup that you simply cannot do anything until you get that CTPA?
Just watched the last 30 mins of the video, i now see how my comments are kind of redundant!
If you’re so afraid of getting sued in the US over non-diagnosis/treatment of PE then you can do the trial here in Ireland as we’re almost a 3rd world country these days and can’t afford all the CT scans so we’d love a reason not to do them!
Sounds great and reasonable. Although the starting questionI think is really the key point. Nonetheless if you really (really!) start thinking about PE in a let’s say young patient healthy looking shound’t you answer yes to the tricky question: PE is the first diagnosis or equally likely? I guess everyone has felt uncomfortable with this criteria of the Wells’ score in the very beginning of his residency, since the question is cloudy enough to be a test of self confidence. With time I’ve started thinking it was subtly included to solicit cliniacians to wonder whether it is really worth to… Read more »
Mattia,
That is exactly what I used to advocate. Low risk geneva should take you below 15%, which seems perfect for the PERC rule application–it was beautiful! Until Hugli’s article (pubmed link) showed that low risk geneva negative PERC still has a post-test prob that is too high.
Definitely have missed that pubblication and maybe quite a lot relevant others. I’ll go through it eagerly, and eventually leave the kids with any excuse to grandma more frequently.
I’ve read all the comments now and I apologize since my post was largely inapt beeing the answer
already adressed by Nickson.
One more thought if you allow me: do we all have the perception that a certain amount of episodes of small PE occurs probably with no or subtle symptoms all life long. As if it might be a sort of physiological process as much as atherosclerosis is for the non pulmonary circle? After all centuries of natural bleeding selection should have favoured thrombophilic profiles.
Consequently
A small amount of small PE that cause some pleuritic discomfort whether or not accompanied by mild dispnea could be tolerated as just overnatural events without any anticoagulation therapy?
Scott Keep going with the PE brain-storming it’s just great!
Mattia
Thanks Mattia!
In response to Mattia’s penultimate comment above, the AHA’s March 2011 paper speaks to the concept of CTEPH (Chronic Thromboembolic Pulmonary Hypertension) which appears to be fairly nebulously attributed to “one or more” previous PEs, despite stating that “up to 63% of patients with CTEPH were not previously aware of having had a PE.” I have a few doubts about the breadth of this link to begin with, and one wonders if there is a bit of a backlash in the “establishment” to the idea that perhaps we shouldn’t be panicking so much about finding every last tiny little subclinical… Read more »
Pretty late to this discussion but I’ll throw in my 0.02…I think this diagram is the way that we all have been practicing for a while…however I love the first step…”Do you think it’s a PE” and if the answer is no, stop there. That way we don’t even mess with Well’s or PERC or anything else- we just say NO and move on. I haven’t been explicitly using Well’s to risk stratify people before using PERC but in reality I have been doing that on a subconscious level. I agree with the use of the d-dimer- if you have… Read more »
How can you use two rules in a row? Its similar to using likelihood ratio in series, and that has never been validated, yet it makes sense.
Andy which two rules in a row do you mean?
If you use the wells criteria to lower your pretest prob to less then 15% then using the perc to then again lower your pretest prob im not sure thats kosher? i mean the wells is essentially plugging a prior pretest prob and using the wells to generate a post test prob which then becomes your next pretest prob right? Is this like using the wells as a negative likelihood ratio and then again using the PERC as another -LR and i did not think LR tests used in series had ever been validated… however it makes theoretical sense.
Scott, im not sure if i am making myself clear, you are using the wells criteria as a predictive rule to generate a negative predictive value , and i understand you can choose which rule you want to use. You could use either the perc or the wells and which ever one got you a lower value …yes you could use that! however i am not sure you can use one to generate a post test prob and then call it a pretest prob for another test ( the PERC )…?
https://emcrit.org/blogpost/a-debate-on-pe-decision-rules/
AMAZING DEBATE!!! this answers ALL my questions!! Thanx for posting this.
I totally agree on this one. I was about to post a similar comment. I guess what Martin ment to say is that in Kline’s 2008 validation study even just PERC negative patients independently from clinical judgment had very good outcomes (only about 1% had VTE/death at 45 days). I guess this is what David Newman might have mentioned in his podcast. I think though this is not surprising since the profile of a PERC negative patient almost matches the one of a Wells low score one. Almost inevitably low risk patient HAD NO CLOTS!
I know this is an old thread but… I use the algorithm at the top of the page and it makes sense to me. I also know that we still overinvestigate. But I had a clinical scenario that still made me a little uneasy (there’s Gestalt creeping in again…) when I chose NOT to investigate. Well’s score 1.5 or 4.5 depending on the subjective question (I would go for 4.5), high sensitivity D-dimer negative. Stop investigating. Easy. However, this was a young patient, on the OCP, with a long haul flight (UK to Australia) four weeks before, with a story… Read more »
if your gestalt is high, spin, spin, spin
Here’s my (not really) modified version: http://mdaware.blogspot.com/2013/09/how-to-pe.html