Type 2 Perioperative Myocardial Infarction: Can We Close Pandora’s Box?

In this issue of the Anesthesiology, Helwani et al.¹  address the etiology, classification, and prognosis of perioperative myocardial infarction (MI), a subject of investigation for more than 75 yr now. A unique feature of their study is use of the most contemporary analytic paradigm for diagnosis available, specifically categorizing type 1 versus type 2 MI according to the “Universal Definition” discussed below.

A brief historical review may help to orient the reader to the rapidly expanding field of what constitutes MI, its subtypes, and treatment strategies. The perioperative setting has long frustrated cardiologists and perioperative clinicians given its potential for profound, transient physiologic alterations not encountered in ambulatory medical settings that may trigger de novo myocardial damage or exacerbate existing pathology.

Isolated case reports of perioperative MI first appeared in the early 1930s, and the first case series systematically evaluating it was reported by Master et al.²  in 1938 based on review of 35 cases of MI at Mt. Sinai Hospital in New York between 1931 and 1938 that followed any surgical procedure. MI was defined by either postmortem exam or “by electrocardiogram.” Of note, the authors excluded 13 additional cases in which a diagnosis was suspected but could not be confirmed by either of the two criteria. Approximately 52% of cases occurred within 3 days of surgery with a mortality of 66%. The authors note the difficulty in diagnosing MI given that the minority of patients present with typical chest pain (40%) and that other conditions, specifically pulmonary embolism, may present with similar electrocardiogram findings. Of the 19 cases that underwent autopsy, all were noted to have “sclerosis and narrowing of the coronary arteries,” while acute occlusion was noted in only 15 cases. It was not until 1952 that another report by Wroblewski and LaDue³  from the Memorial Hospital in New York, smaller (15 cases) but including only cases after major surgery, was reported. Although no autopsy data were presented, mortality was similarly high at 40%, and pain was only present in 27% of patients. These authors hint at the potential hazards of sinus tachycardia and its relation to shortening of diastole and mention that isolated ST segment depression or ventricular hypertrophy were present only in some cases. They opine that “coronary occlusion . . . may be precipitated by alteration in coronary hemodynamics” but also note that “although coronary occlusion may conceivably occur postoperatively in the absence of any pathological alteration in the coronary vessels, such a possibility seems highly unlikely.” By the mid to late 1970s, more precise diagnoses of MI based on the creatine kinase-muscle/brain biomarker and categorization of MI by transmural (presence of ST-segment elevation) or nontransmural (ST-segment depression only) electrocardiogram patterns were the focus of studies.⁴ 

The cardiology community has long grappled with how to categorize MI using either anatomic (e.g., acute complete or partial occlusion of epicardial coronary vasculature) or more recently, etiologic approaches (e.g., considering multiple physiologic pathways leading to myocardial damage). In 2000, the first Task Force of the European Society of Cardiology and the American College of Cardiology/American Heart Association for the Definition of Myocardial Infarction proposed “redefining MI,” moving away from the existent World Health Organization definition with its high specificity but low sensitivity (e.g., two or more of symptoms, enzyme rise, or Q waves) in favor of the groups contention that “any amount of myocardial necrosis caused by ischemia should be labeled an infarct.”⁵ Two subsequent revisions (2007 and 2012), now termed the “Universal Definition,” refined this framework into five categories: type 1 (spontaneous MI due to a coronary abnormality resulting in intraluminal thrombus impeding coronary blood flow), type 2 (MI secondary to ischemic imbalance in which a condition other than coronary artery disease contributes to an imbalance between supply and demand), type 3 (cardiac death due to MI even in the absence of biomarker or electrocardiogram evidence), types 4A and 4B (MI associated with percutaneous coronary intervention or stent thrombosis), and type 5 (MI associated with cardiac surgery) and myocardial injury (changes based on serial troponin testing without other clinical manifestations).⁶ 

Codification of the new type 2 category in 2007 led to considerable controversy akin perhaps to the opening of Pandora’s box. Pathologically, it encompasses a wide spectrum, including patients with known coronary artery disease who may have stable but flow-limiting stenosis, abnormalities of vasomotion of a coronary artery (vasospasm or endothelial dysfunction), as well as physiologic “supply–demand imbalance” in a normal vasculature. The “noncardiac” causes of supply–demand imbalance mentioned include coronary embolism, tachy- and bradyarrhythmias, anemia, respiratory failure, and hypo- and hypertension with or without left ventricular hypertrophy. The Universal Definition specifically excludes damage due to other causes such as myocarditis or trauma. It is left up to the treating clinician to decide whether there was a significant supply–demand imbalance, a potentially daunting task. Coupled with clinical introduction in Europe of high sensitivity troponin assays, an escalating number of cohort analyses attempting to categorize the epidemiology and prognostic implications of type 2 MI have been reported. Given the lack of specific physiologic criteria in the 2012 Universal Definition report, Danish members of that group developed specific values based on literature review for a variety of common clinical scenarios, applying them retrospectively to a large cohort of unselected inpatients at a tertiary university hospital with at least one elevated troponin measurement.⁷  They reported that approximately 25% of all MIs were type 2, that anemia, respiratory failure, and tachyarrhythmias were the most common causes; that 50% of all cases were from departments “outside of cardiology”; and that 50% of patients did not have significant coronary artery disease. A retrospective review of 475 inpatients using International Classification of Diseases, Ninth Revision coding of a secondary type cardiac event reported that the most common associated settings for type 2 MI or myocardial injury alone were noncardiac surgery (38%), anemia or bleeding requiring transfusion (32%), sepsis (31%), tachyarrythmia (23%), hypotension (22%), respiratory failure (23%), and severe hypertension (8%).⁸  Notably, inpatient mortality (5 to 6%) was no different between MI and myocardial injury alone, and only 43% of such patients received aspirin or statin therapy at discharge.

Recent studies evaluating patients presenting to the emergency room with signs of acute coronary syndrome report widely varying frequencies of type 2 MI (2.8, 7.3, 19.2, and 34.5%).^9–12  Subjects were older, were more often female, and had more comorbidities than type 1 MI patients. Outcomes were similar or worse, and substantially fewer received intervention or secondary prevention strategies.

Two recent studies with prospective long-term follow-up support not only the gravity of type 2 MI but also of myocardial injury as defined by the Universal Definition. The Catheter Sampled Blood Archive in Cardiovascular Disease (CASABLANCA) study followed 1,251 patients undergoing either coronary or peripheral arterial angiography at a single U.S. tertiary center for a median of 3.4 yr reporting a 12.2% incidence of type 2 MI that was frequently recurrent.¹³  Older age, along with a large number of comorbid conditions, were again predictive. Rates of major adverse cardiovascular events were similar to those reported for type 1 MI and all-cause or cardiovascular mortality rates were double to triple than those without. Chapman et al.¹⁴  report 5-yr follow-up of 2,122 consecutive patients with type 1 (55.2%) or type 2 (20.2%) or myocardial injury (24.6%) based on at least one elevated troponin measurement at a tertiary center in Scotland (excluding patients admitted for elective procedures). Statistical analysis considered the concept of “competing risks” taking into account the possibility that a larger percentage of noncardiovascular death in type 2 MI patients (due to underlying comorbidity) reduces the number of patients at risk for having a cardiovascular event (relative to those with type 1 MI in which these events are commonly recognized). The most common causes of type 2 MI were arrhythmia (19.1%), pneumonia (13.5%), heart failure (12.4%), and fracture (4.2%) with slightly varying frequency for myocardial injury. All-cause mortality for type 2 MI (62.5%) or injury (72.4%) was significantly higher than for type 1 MI (36.7%) primarily related to noncardiovascular causes, although similar rates of major adverse cardiovascular events were noted. The presence of coronary artery disease independently predicted major adverse cardiovascular events in either type 2 MI or injury, and the authors suggest that detection of coronary artery disease in this group should be a priority.

Despite publication of the Universal Definition groupings, it is notable that not all publications in the cardiology literature use this system, particularly those evaluating non-ST-elevation MI. That said, the choice of either an early invasive treatment strategy, consisting of aggressive medical therapy coupled with coronary angiography within 24 to 72 h (often with percutaneous coronary intervention of stenotic lesions) versus a more conservative “ischemia-driven strategy” with coronary angiography only for persistent angina or inducible ischemia on noninvasive stress testing remains a subject of controversy.¹⁵ 

Further complicating matters, additional MI or injury categories have been proposed. Readers of this journal are likely aware of the Myocardial Injury after Noncardiac Surgery (MINS) grouping popularized by the Vascular Events in Noncardiac Surgery Patients Cohort Evaluation (VISION) investigators, defined as isolated troponin elevation within 30 days of noncardiac surgery after exclusion of nonischemic etiologies, a classification that has not yet been adopted widely.¹⁶  Myocardial Infarction with Nonobstructive Coronary Arteries (MINOCA), a grouping that may include patients with type 1 (albeit rarely) and more commonly type 2 MI, has been recently proposed by a European Society of Cardiology working group.¹⁷ 

The study of Helwani et al.¹  focuses on a select group of 146 patients who underwent coronary angiography within 30 days of surgery with a clinical diagnosis of acute coronary syndrome (ST-elevation MI, non-ST-elevation MI, or unstable angina) from 2008 to 2015. Patients were identified retrospectively using their hospital billing database and were matched with records in the National Cardiovascular Data Registry CathPCI Registry, which captures demographic and clinical variables, along with findings at catheterization.¹⁸  An independent review of the films by two blinded coinvestigators was performed scoring for features of “complexity” defined by the presence of either an intraluminal filling defect, plaque ulceration, or two or more minor features (fissuring, irregular margins/overhanging edges, or hazy appearance). A patient with any of the three features was considered to have a type 1 MI. Diagnosis of type 4B MI was categorized (by virtue of a preexisting stent in place), and the remainder of cases were categorized as type 2 MI. The predominant clinical presentation was that of non-ST-elevation MI (80.1%), 14.4% were considered ST-elevation MI, and 5.5% had unstable angina. Given the inherent selection bias of this cohort, electrocardiogram changes of ischemia were present in nearly all patients. Based on angiography, type 1 MI was diagnosed in 25%, type 4B MI was diagnosed in 2%, and type 2 MI was diagnosed in the remaining 73%. No significant correlation was noted between the presenting clinical syndrome and the Universal Definition MI type. Of particular interest is the finding of absent or nonobstructive coronary artery disease in 27% of patients in the type 2 group (35% of the total cohort) but also in 5% of the type 1 patients. Stress-induced cardiomyopathy was coded in 10% of patients. Treatment strategies varied with 33% of type 1 patients receiving revascularization (coronary artery bypass graft surgery in all but one patient, which is quite unusual given the popularity of percutaneous coronary intervention), whereas nearly all of the type 2 patients received medical management alone. A retrospective secondary outcome analysis demonstrated 30-day and 1-yr mortality rates of 7 and 14%, respectively, with no difference between the three MI types but worse outcome in non-ST-elevation MI patients when compared to ST-elevation MI patients. The authors conclude that most perioperative MIs are due to demand ischemia, including many with nonobstructive disease (or stress-induced cardiomyopathy) and that measures to optimize perioperative supply–demand imbalance are most likely to be beneficial.

It is of interest that a similar classification of angiographic complexity was used in a smaller cohort of 56 patients reported by Hanson et al.¹⁹  using slightly different exclusion criteria (based on presence or absence of a “culprit lesion,” which biases the findings toward type 1 MI). They conclude that 59% of patients had type 1 perioperative MI and 41% type 2. Although type 1 patients had a higher incidence of ST elevation (53% vs. 23%), higher peak troponin levels (15.3 vs. 5.3 ng/dl), and greater decrease in intraoperative mean blood pressure, in-hospital mortality and length of stay were not significantly different. This smaller study, which presents data not considered by the present report, poses an interesting counterpoint and also emphasizes the gravity of type 2 MI on outcome.

How MI is coded in the medical record can have significant impact on an individual patient as well as hospital ratings and reimbursement. Goyal et al.²⁰  have presented a perspective on how best to use the new International Classification of Diseases, Tenth Revision classification code for type 2 MI (121.A1), stressing that it should only be used as a secondary diagnosis. They recommend that administrative agencies exclude this from Diagnostic Related Groups and acute MI performance measures and, of particular interest to the perioperative period, that a specific code should be created for elevated cardiac biomarker without MI.

The present study emphasizes the role of supply–demand imbalance in the pathogenesis of perioperative MI. Others have stressed its role in the isolated biomarker elevation (myocardial injury) with or without other contributing factors. The contemporary cardiology studies cited above suggest that type 2 MI and myocardial injury alone in the vulnerable patient have similar adverse prognostic implications. The senior author of the present study has recently engaged in a literature debate with members of the Universal Definition working group arguing against attempting to distinguish between type 2 MI and myocardial injury.²¹  Based on how we think about damage to other major organs, citing the example of how commonly we use the term acute kidney injury but rarely if ever would say renal infarction unless there was bona fide vessel occlusion (although the reverse argument can be made for watershed infarction in the brain), Nagele suggests that the two should be unified as “acute myocardial injury,” leaving only type 1 MI (excluding types 3, 4, and 5 from the discussion).²¹ With the pending introduction of high sensitivity troponin assays in the United States, the potential exists for a major increase in the diagnosis of type 2 MI in all areas of clinical medicine. Although an argument can be made that this could lead to more vigorous screening for occult “culprit” coronary artery disease, this hypothesis has not been tested, and given the character of many of these patients (e.g., older with multiple comorbidities), a favorable risk benefit ratio is not assured. In the corresponding viewpoint, representatives of the Universal Definition group acknowledge the confusion the term has caused and specifically note that they will consider the proposition of Dr. Nagele at the next meeting of the working group, which will likely be this year.²²  Thus, perhaps things will be simpler on the diagnostic front shortly. However, for our increasingly elderly and sick patients, that may be little consolation because how we define something that leads to their potential demise is of little value to an individual patient. Working harder to define and adequately treat supply–demand imbalance has to remain the primary focus for clinicians.

Helwani et al.¹  provide a strikingly different snapshot in time of perioperative MI than that of the first pioneers in this arena. The widespread availability of troponin biomarkers, 12 lead electrocardiograms, and widespread promotion of primary and secondary prevention strategies reducing the incidence of type 1 MI likely account for this shift. Despite this shift, the recommendations of Wroblewski and LaDue³  remain eerily prophetic: “Coronary artery occlusion can possibly be prevented if surgical or hemorrhagic shock is avoided, anoxia prevented, operative time minimized, non-cardiovascular complications prevented or controlled, and if detailed attention is directed toward all other preoperative and postoperative details.” This idealistic prescription for success is increasingly being visualized by a new generation of clinician researchers just as mortality has improved after perioperative MI. Although not all patients can be saved given overwhelmingly severe comorbidities, we now know that likely the biggest bang for the buck in this arena is in dealing with supply–demand imbalance. Although acronyms may continue to dazzle and confuse us, we appear closer to closing the Pandora’s box opened by the Universal Definition.