Introduction
0
It’s common for critically ill patients to be screened for MI using troponin. Troponin levels are often positive. One thing leads to another, and before you know it the frail 99-year-old lady you admitted for a COPD exacerbation has an intracranial hemorrhage from the heparin drip she was put on because her troponin was 0.09.
0
The delivery of good medical care is to do as much nothing as possible (Law #13 of the House of God). This isn’t exactly true, though. We need an approach to evaluating and treating myocardial ischemia in the ICU which is thoughtful and thorough while avoiding over-diagnosis and iatrogenic harm. This blog will propose the following approach, using seven conceptual steps.
0
0
Step 1: Recognize that we know nothing about treating MI in ICU patients.
0
Our knowledge of MI treatment is based on large industry-sponsored trials of patients presenting to the hospital with a diagnosis of myocardial infarction (whom I will refer to “cardiology patients”). Such studies often exclude sicker patients with renal dysfunction or bleeding risk factors. There are no randomized controlled trials of MI therapy in critically ill patients with non-cardiacprimary problems such as sepsis, respiratory failure, or stroke (whom I will refer to “ICU patients”)(1). There will probably never be an industry-funded study of MI therapy in ICU patients, as pharma is far too clever to trial its drugs on this complication-prone, heterogeneous population.
0
The natural tendency when treating an ICU patient with MI is to apply knowledge obtained from cardiology patients. This is not valid because these are two very different groups of patients. ICU patients are typically at greater risk to bleed than cardiology patients due to pre-existing coagulopathies as well as renal dysfunction (which impairs platelet function and causes some anticoagulants to accumulate)(2). Meanwhile, ICU patients are often at lower risk of benefitting from aggressive treatment since MI is not their primary problem.
0
Step 2: Checking a troponin level on every ICU patient isn’t wise.
0
0
It’s a common practice to check a troponin level on every ICU patient. No one wants to be the physician responsible for missing a myocardial infarction. This practice is wrong for two reasons.
0
First, the significance of a test depends on pre-test probability, sensitivity, and specificity. One of the most common errors in diagnostic testing is ignoring the pre-test probability. We are obsessed with which tests are “good tests,” but inappropriate application can render even the best tests worthless. Consider, for example, a 40-year-old man being admitted to the ICU for renal failure with no symptoms of ischemia and a normal EKG. His pre-test probability for a myocardial infarction is low, perhaps 0.25%. Lets say that troponin has a sensitivity of 100% and specificity of 75%. If his troponin level is elevated, his post-test probability increases from 0.25% to 1%, so a “positive” value is 99 times more likely to be a false-positive than a true-positive. This illustrates that by using troponin to screen patients with very low pre-test probability, the significance of a positive test is degraded.
0
Second, testing everyone ignores the concept of the test threshold. Testing a patient with very low pre-test probability for MI is more likely to yield false-positives and lead to iatrogenic harm than find a true-positive and actually benefit the patient. The test threshold is the pre-test probability at which the likelihood of actually finding disease and helping the patient starts to outweigh the risk of finding a false-positive and causing iatrogenic harm. The test threshold for any disease is calculated based on risks of diagnostic testing and over-treatment (i.e., risk of cardiac catheterization or anticoagulation in someone who doesn’t have MI) compared to risks of under-treatment (i.e., missing an opportunity to treat a true MI). Surprisingly I’m unable to find any research defining what the test threshold is for MI, probably because it’s been so dogmatically engrained in our culture that we must never miss one. My guess is that the test threshold for MI around 1-2%, similar to the test threshold for pulmonary embolism (Kline, 2004). Regardless, the principle is that testing should only be pursued for patients with a high enough pre-test probability that they are more likely be helped than harmed by testing – not simply for any patient with a pre-test probability above zero.
0
Step 3: Clinical history and EKG is a reasonable way to screen for ischemia in ICU patients.
0
Although checking a troponin level on every ICU patient isn’t helpful, we must exercise due diligence. A reasonable approach is a clinical history combined with the admission EKG (3). An EKG is less likely to lead to iatrogenic harm than checking a troponin. If an EKG shows some nonspecific changes, the clinician can carefully weigh the benefits vs. risks of investigating further. If there is mild concern, a reasonable approach is often to repeat the EKG once electrolytes are normalized and the patient improves a bit. It is much easier to choose to ignore a nonspecifically abnormal EKG than to choose to ignore an equally nonspecific troponin elevation. The troponin value of 0.06 (!!) will remain stuck in the patients labs in red, and sooner or later somebody will feel compelled do something about it.
0
After performing a history and evaluating the EKG, if you are genuinely concerned about myocardial ischemia then it makes sense to check a troponin.
0
Step 4: Troponin elevation is often not due to MI at all.
0
As the sensitivity of troponin assays has increased, the specificity has decreased. Current troponin assays often detect troponin released from a variety of cardiac stresses other than myocardial infarction (e.g., congestive heart failure, pulmonary embolism, renal failure, sepsis, stroke, etc.). In some series, the majority of ICU patients have been found to have positive troponin. Such non-MI troponin elevations correlate with poor overall prognosis, but are not an indication of MI.
0
The first question when faced with a positive troponin is whether the patient is actually having a MI (4). International consensus guidelines define an MI as a dynamic rise and fall in troponin plus at least one of the following: EKG evidence of ischemia, symptoms of an MI, or a new wall motion abnormality. Unfortunately ICU patients are often unable to relate a history, so this distinction can be challenging.
0
Step 5: Most myocardial infarctions in the ICU are type-2 MIs.
0
If you determine that the patient is having a MI, the next question is whether it is a type-1 MI (plaque rupture) or a type-2 MI (demand ischemia). Cardiology patients presenting to the hospital with MI generally have type-1 MIs due to atherosclerotic plaque rupture. In contrast, ICU patients are almost always experiencing type-2 MIs. A variety of factors increase myocardial oxygen supply-demand mismatch in ICU patients (i.e., tachycardia, shock, vasopressors, inotropes, hypoxemia, anemia, hypertension). If such patients have underlying coronary disease, it’s common to develop a type-2 MI. This is not a reflection of plaque rupture, but rather that they have underlying, stable coronary stenoses which reduce the physiologic reserve of their coronary circulation.
0
Why do we care? The treatment for type-1 MI centers around anticogulation and percutaneous coronary intervention. The problem is an unstable coronary plaque, and this needs to be anticoagulated, stented open, bypassed with CABG, etc. In contrast, the treatment for type-2 MI centers around the addressing the patient’s underlying disease (i.e. sepsis, hypoxemia, anemia, etc.). The active problem in type-2 MI is not the coronary arteries, but rather other factors which are stressing the heart. Almost all data on MI therapy comes from studies of cardiac patients experiencing type-1 MI, and there is little data regarding how to treat type-2 MI. When treating type-2 MI, giving aspirin is reasonable but anticoagulation is not indicated.
0
Making the distinction between type-1 and type-2 MI can be challenging. This generally centers around a comparison of the amount of stress on the patient’s heart compared to the size of the myocardial infarction. For example, if an elderly COPD patient with anemia in septic shock on vasopressors develops a mild troponin elevation with transient T-wave inversion, this is probably a type-2 MI. The treatment centers around treating his septic shock and anemia. Alternatively, if the same patient presented with a mild COPD exacerbation and subsequently developed crushing chest pain with anterior ST elevation this would be more consistent with a type-1 MI, and urgent cardiac catheterization would be advisable.
0
Step 6: Heparin infusion is not beneficial unless the patient is scheduled for cardiac catheterization.
0
0
For anyone who is unfamiliar with this concept, I would strongly recommend reviewing the material on the NNT.com and the corresponding SmartEM podcast. The bottom line is that for patients with acute coronary syndrome (including type-1 NSTEMI), heparin stabilizes plaque and reduces infarction. However, as soon as the heparin is stopped, the rate of infarction rebounds and ultimately there is no benefit from heparin. The heparin delays infarction but the eventual rate of infarction is the same.
0
For a patient being treated with heparin as a bridge to definitive treatment with cardiac catheterization, this makes sense. However, if the patient is being medically managed with a short course of heparin and no catheterization, this exposes the patient to risks of hemorrhage from heparin without any long-term benefit.
0
Unfortunately the AHA/ACC guidelines continue to recommend short-term heparin infusions for management of MI in patients not undergoing catheterization. This conclusion is inexplicably based on choosing re-infarction at one week as the primary outcome (5). If they had chosen to look at a later timepoint, their conclusions would have been different. The European Society of Cardiology Guidelines does recognize this, stating that “recurrence of events after interruption of unfractionated heparin explains why this benefit is not maintained over time, unless the patient is revascularized before interruption of unfractionated heparin.”
0
Step 7: Prophylactic dose fondaparinux is a reasonable approach to anticoagulate a type-1 MI patient who isn’t going for cardiac catheterization.
0
For ICU patients with a type-1 MI, there may be significant pressure to treat with anticoagulation even if the patient isn’t going for catheterization based on local norms and AHA/ACC guidelines. A reasonable approach to this situation is to treat the patient with a DVT prophylactic dose of fondaparinux (2.5 mg s.q. daily).
0
This is based on the OASIS-5 trial which randomized 20,078 patients with acute coronary syndrome to a therapeutic dose of enoxaparin (1 mg/kg BID) vs. a DVT prophylactic dose of fondaparinux. Patients treated with therapeutic enoxaparin had a significantly higher rate of death which was sustained over 180 days after randomization. Based on the 0.7% absolute difference in mortality, the number needed to kill was 143, that is for every 143 patients treated with therapeutic enoxaparin there was one excess death. Analysis of the data suggested that this difference reflected increased bleeding complications experienced in the enoxaparin group.
0
My interpretation of this study is not that fondaparinux is necessarily a wonder drug (at a prophylactic dose its unclear that it was doing much), but rather that therapeutic enoxaparin is probably harmful. Regardless, this study led the AHA/ACC and other guidelines to add prophylactic dose fondaparinux as a treatment option for acute coronary syndrome. In fact, in the 2011 guidelines “suggest an anticoagulant preference for these patients treated with a noninvasive strategy in the order of fondaparinux, enoxaparin, and unfractionated heparin (least preferred)(6).” European Society of Cardiology and British NICE guidelines similarly recommend fondaparinux as the preferred anticoagulant.
0
Fondaparinux 2.5 mg sq. daily is a good drug to use for DVT prophylaxis. It has the advantage of not causing HITT, and it is an accepted option in the 2012 ACCP guidelines. Thus, for a patient with type-1 NSTEMI being managed with a noninvasive strategy, 2.5 mg fondaparinux is adequate to cover both DVT prophylaxis and MI anticoagulation, simultaneously satisfying the ACCP and AHA/ACC guidelines. Unlike a therapeutic heparin infusion, this dose of fondaparinux is unlikely to harm the patient provided that renal function is adequate.
0
Conclusions
0
0
0
Using this approach, we should be able to reduce the amount of harm experienced by ICU patients due to heparin anticoagulation for MI. This starts with judicious measurement of troponin only in patients for whom MI is a genuine concern. When a positive troponin occurs, care must be taken to sort out whether the patient is experiencing a non-MI troponin elevation, type-1 MI, or type-2 MI. Neither type-2 MI nor non-MI troponin elevation are indications for anticoagulation. For patients with type-1 MI, treatment depends on whether the patient is a candidate for cardiac catheterization. For patients with type-1 MI being medically managed without catheterization, there is no long-term benefit from heparin anticoagulation. DVT prophylactic doses of fondaparinux may be a reasonable approach to such patients which satisfies AHA/ACC and ACCP guidelines, simultaneously anticoagulating for MI and providing DVT prophylaxis. Ultimately only a very small minority of ICU patients will actually benefit from a heparin infusion (patients with type-1 MI who are scheduled for cardiac catheterization).
0
0
The practice of screening critically ill patients with troponin is ill-conceived and runs contrary to fundamental principles of diagnostic medicine such as Bayes theorem and the test threshold. Administering heparin infusions to patients who aren’t pending catheterization isn’t supported by the evidence. Until these practices are curbed, critically ill patients will continue to be needlessly harmed by heparin infusions. Coronary disease is a very real problem in critically ill patients, but just as we must be alert to the threat of coronary disease we must be equally alert to the threat of iatrogenic harm in these patients.
0
Notes
(1) Of course this applies to critically ill patients regardless of their geographic location (ED, ICU, OR, etc.). The term “ICU patients” is used for the sake of brevity.
(2) A while ago at Genius General Hospital I presented a morbidity and mortality case about a patient with renal dysfunction who presented with a myocardial infarction, was treated with an eptifibatide infusion and subsequently had a lethal intracranial hemorrhage. Eptifibatide is renally cleared, and all of the studies of this drug had excluded patients with renal failure. This was an eye-opening case for me about the dangers of applying information from a study to patients who would have been excluded from that study.
(3) Note that the only reason to get the admission EKG isn’t to look for ischemia. An admission EKG is useful to evaluate cardiac structure and function (chamber dilation, prior infarction, etc.) and also to obtain a baseline for comparison later in the patient’s hospital course.
(4) In a perfect world the troponin wouldn’t be ordered unless there was other evidence to support a diagnosis of MI. However, in reality even if you’re trying to avoid a screening troponin it will often get ordered anyway. Therefore, it is important to know how to interpret these troponin elevations in patients without other features of MI.
(5) The rationale behind the AHA/ACC guideline can be found on figure 10, page e474 of the 2011 focused update. Although this is not the latest guideline, it is the most recent guideline to discuss their rationale for utilizing heparin in detail.
(6) This preference is absent in the 2012 ACCF/AHA guideline. However, fondaparinux remained listed as a Class I recommendation for noninvasive management of acute coronary syndrome. The European Society of Cardiology Guidelines (2011) suggest fondaparinux 2.5 mg sq. daily with a IA recommendation noting that it is “recommended as having the most favorable efficacy-saftey profile with respect to anticoagulation.” The British NICE guidelines also suggest fondaparinux as first-line therapy.
Latest posts by Josh Farkas (see all)
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
- PulmCrit wee: Why I like central lines for GI bleed resuscitation - March 13, 2024
EXCELLENT PRESENTATION OF THE COMMON DAY TO DAY PROBLEM FOR CARDIOLOGISTS