Management of low-risk chest pain is, by reasonable conjecture, one of the greatest failings of Emergency Medicine and the medical profession in general. Whether driven by true altruism or by risk-management and zero-miss strategies based on the ACC/AHA guidelines, many, many, many patients are admitted and subjected to provocative testing.
And almost none of those patients are ultimately, correctly, diagnosed with the feared disease – acute coronary syndrome.
This is a prospective, observational evaluation of patients admitted for chest pain observation at a single academic center in Rhode Island. Over the course of ~2 years, 3,543 patients were admitted for an initial evaluation of chest pain after initial negative cardiac biomarkers in the Emergency Department. Approximately half of patients underwent stress testing.
Of 1,754 stress tests, there were 29 positives. Of those, 9 were false positives. Stratified by pretest probability, none of the patients with a “low probability” Diamond & Forrester Score had a true positive test. Only 1% of patients admitted and stressed with “intermediate probability” D&F Score ultimately proved to have true positive tests. Even with “high probability”, 5% of all stress tests performed were true positives.
The author of this article means to specifically reduce stress testing in the “low probability” cohort. this is a reasonable proposal to skim off a small percentage of tests. However, he misses asking the better question – how do we reduce use in our “intermediate probability” cohort, which constituted 85% of admissions with just a 1.7% yield for ACS? We need to seriously address the outdated and inefficient notion admission and testing for these patients is the ideal strategy – and that probably starts by tossing our current guidelines out the window.
“The Association Between Pretest Probability of Coronary Artery Disease and Stress Test Utilization and Outcomes in a Chest Pain Observation Unit”
http://www.ncbi.nlm.nih.gov/pubmed/24730402
Ryan,
Love the title of this post.
These guidelines have always struck me as a bit odd. They are predominently created by cardiologists with their own self interest involved. I would love to write guidelines that have the potential to make a lot of my car payments.
In the end, low and intermediate risk chest pain is going to be very difficult for the Emergency Physicians in the USA. Even with the most conservative approaches, a very small number of patients will slip through the cracks and have a bad outcome. We know that missed MI is the single biggest dollar payout for litigation in the USA. With the current malpractice ystem in the USA, I doubt this will change for the majority of ED docs (although some states like Texas and Georgia have changed the threashold for medical malpractice cases to gross negligence).
Therefore I would suggest you all move to Australia or New Zealand where we generally have a more common sense and pragmatic approach to the evaluation of chest pain. Good luck…
I quote Ryan:
1- "– how do we reduce use in our "intermediate probability" cohort, which constituted 85% of admissions with just a 1.7% yield for ACS?"
2- " We need to seriously address the outdated and inefficient notion admission and testing for these patients is the ideal strategy – and that probably starts by tossing our current guidelines out the window."
For 1 :
An "intermediate probability" population of chest pain patients only has ACS or MI in 1.7% ?
This obviously is a very very low probability population. Almost a "no risk" population.
Or did I miss something the authors say ?
As for 2:
The guidelines say "A high degree of suspicion and recognition of atypical presentations is important, because a significant number of patients present with “anginal equivalents” rather than chest pain. These symptoms include jaw, neck, or arm discomfort; dyspnea; nausea; vomiting; diaphoresis; and unexplained fatigue. These are seen more frequently in the elderly, women, and diabetic patients. Sharp, stabbing, or reproducible pain reduces but does not exclude the likelihood of ACS.
Pleuritic chest pain is consistent with a pulmonary condition, musculoskeletal disease, or pericarditis. However, the Multicenter Chest Pain Study found that 22% of patients presenting with symptoms described as sharp or stabbing pain (13% with pleuritic pain and 7% with pain reproduced on palpation) were eventually diagnosed with ACS.11 The National Heart Attack Alert Program recommends that patients with any of the aforementioned presenting symptoms should be assessed immediately and referred for rapid evaluation.
Isn't this the source of the problem ?
Hi Ryan,
Nice article! I also read your "Predicting TIMI's Mortality" in ACEP NOW with great interest, as I am working on developing CDS tools connected to our EMR, for our group. Which of the newer risk stratification tools you feel are ideal or close to ideal for ED care in the US?
Thanks.
Sorry about the late reply – had been busy working and preparing for #SAEM14.
I've started using the HEART score, just as I mention in ACEP Now. It's simple, it uses readily available information from the Emergency Department, and it predicts out to six weeks. It's also got, so far, the largest validation population. I'm curious to see how the EDACS rule turns out, although, that's a little more complicated – and will definitely require an online calculator to track. I'm a huge proponent of addressing risks/benefits with patients, not just because it's the right thing to do, but happily most patients make sensible choices to forego admission/observation.
No – using the Diamond and Forrester tool to assess for "intermediate" risk (which, in America, is basically everyone who isn't zero-risk) – the yield of a stress test is 1.7%. One of the reasons that doesn't match up with the D&F numbers is the D&F numbers are for CAD – and a stress will only pick up flow-limiting stenoses, which would be the minority of even advanced CAD. The problem is many-fold, and stress/echo/CCTA all give us different bits of information without truly giving us all the information. And, that doesn't even get into plaque histology, which is probably the true risk determinant of future acute coronary syndrome ….
Don't joke too much – give it another decade, and my family and I might happily take you up on that AUS/NZ offer ….