Aspirin for chest pain is here forever, but should it be? The Centers for Medicare & Medicaid Services withdrew its aspirin rule on arrival last year. The federal government no longer sees the point of monitoring compliance or financially rewarding good compliance. A CMS spokesperson confirmed this to me in an email, writing, “We have determined that performance on this measure is so high and consistent that it is impossible to make meaningful distinctions in improvement. [T]The burden of reporting the measurement extracted from the graph is not justified by the value of its preservation.
How did we get so good at giving aspirin? Simple: we used too much.
I’m not talking about giving aspirin for traumatic chest pain, which I’ve seen done. (“Driver unbelted, no airbags. Minimal damage to vehicle. Said his chest hurt where he hit the steering wheel, so we loaded him up with aspirin 324mg.”) heart attack acute myocardial or cardiac chest pain probable. (CMS. June 30, 2021; https://bit.ly/31rtQae.) Interesting. A patient must have confirmed or even just probable heart pain to require aspirin treatment. How likely? How do we decide what is likely?
There are no regulations punishing clinicians for giving aspirin to treat acid reflux or what turns out to be a pneumothorax. The hospital administrator’s 10,000 foot view is fair to give it to everyone with a chest and pain in it. What could go wrong? Maybe a lot, according to my math, admittedly, on the back of the envelope. This shotgun approach, as I understand it, charges three aortic dissections with antiplatelet therapy for every life saved from myocardial infarction.
Number needed to treat
The evidence for giving aspirin to treat cardiac chest pain comes largely from the landmark ISIS-2 trial published over 30 years ago. (Lancet. 1988;2:349.) It really was a great trial. It randomized 17,187 patients at over 400 hospitals in over 16 countries and reported 97% follow-up thereafter. That would be a pretty impressive data management feat even today, let alone in the age of the computing power of the Commodore 64 and Apple Macintosh.
Patients with suspected or confirmed myocardial infarction – not with chest pain but actual MI – were assigned to one of four groups: intravenous thrombolytic therapy for one hour, aspirin for one month, both, or neither. (Again, this was in the 1980s, so they were using streptokinase, the big-haired fanny pack of thrombolytics.) Aspirin and streptokinase together reduced vascular deaths as well as heart attacks and cerebrovascular accidents. Considered separately, each also reduced mortality at five weeks, and the aspirin cohort recorded a 2.4% decrease in deaths. This cohort also saw 1% fewer reinfarctions and 0.3% fewer strokes, with no increase in intracranial bleeding, bleeding events or other harms.
This is often summarized as a number needed to treat (NNT) of 42: Treating 42 of a particular type of patient with aspirin saves a life, with no harm identified, right? Wrong. The NNT specifies a particular type of patient, one who has a myocardial infarction. Read ISIS-2 carefully and it seems that 56% of enrolled patients actually had STEMI ECGs! (STEMI at that time was not yet universally recognized as a pattern of acute coronary occlusion.) Treat 42 MIs with aspirin, save a life. Sounds great, but remember the CMS formulation is for chest pain patients with probable cardiac chest pain.
This is where it gets interesting: only about 4% of patients who present to the ER with chest pain experience an MI. (J Med Urgent. 2005;29:383.) In other words, we have to give aspirin to 25 patients to find an MI, to find one of the 42 MIs we have to treat. The NNT to save a life giving aspirin to all chest pains is not one in 42 but about one in 1050. (25×42 = 1050.) I wonder what else is going on when you give aspirin to 1050 patients with chest pain?
Maybe four percent of patients with chest pain in your emergency department have a heart attack, but even fewer, about 0.3 percent, have an aortic dissection. (J Med Urgent. 2005;29:383.) This is a dissection of 333 patients with chest pain. Giving a thousand doses of aspirin to treat chest pain saves the life of a patient with myocardial infarction, but also increases the risk of bleeding in three patients with aortic dissection. Here’s how good we are at what we do: saving a patient’s life, but wasting a full day for three heart surgeons.
Dr Bivensworks full-time in Massachusetts emergency services, including St. Luke’s in New Bedford and Beth Israel Deaconess in Boston. He is also EMS Medical Director.