Improving Risk Stratification for Patients With Type 2 Myocardial Infarction

Background Despite poor cardiovascular outcomes, there are no dedicated, validated risk stratification tools to guide investigation or treatment in type 2 myocardial infarction. Objectives The goal of this study was to derive and validate a risk stratification tool for the prediction of death or future myocardial infarction in patients with type 2 myocardial infarction. Methods The T2-risk score was developed in a prospective multicenter cohort of consecutive patients with type 2 myocardial infarction. Cox proportional hazards models were constructed for the primary outcome of myocardial infarction or death at 1 year using variables selected a priori based on clinical importance. Discrimination was assessed by area under the receiving-operating characteristic curve (AUC). Calibration was investigated graphically. The tool was validated in a single-center cohort of consecutive patients and in a multicenter cohort study from sites across Europe. Results There were 1,121, 250, and 253 patients in the derivation, single-center, and multicenter validation cohorts, with the primary outcome occurring in 27% (297 of 1,121), 26% (66 of 250), and 14% (35 of 253) of patients, respectively. The T2-risk score incorporating age, ischemic heart disease, heart failure, diabetes mellitus, myocardial ischemia on electrocardiogram, heart rate, anemia, estimated glomerular filtration rate, and maximal cardiac troponin concentration had good discrimination (AUC: 0.76; 95% CI: 0.73-0.79) for the primary outcome and was well calibrated. Discrimination was similar in the consecutive patient (AUC: 0.83; 95% CI: 0.77-0.88) and multicenter (AUC: 0.74; 95% CI: 0.64-0.83) cohorts. T2-risk provided improved discrimination over the Global Registry of Acute Coronary Events 2.0 risk score in all cohorts. Conclusions The T2-risk score performed well in different health care settings and could help clinicians to prognosticate, as well as target investigation and preventative therapies more effectively. (High-Sensitivity Troponin in the Evaluation of Patients With Suspected Acute Coronary Syndrome [High-STEACS]; NCT01852123)

T he universal definition of myocardial infarction differentiates type 1 myocardial infarction due to atherosclerotic plaque rupture and intracoronary thrombosis from type 2 myocardial infarction secondary to oxygen supply and demand imbalance, without atherosclerotic plaque rupture and thrombosis, typically in the setting of another acute systemic or cardiovascular illness. [1][2][3] Outcomes in patients with type 2 myocardial infarction are poor, with more than two-thirds of patients dead at 5 years. 4,5 Although it is understood that noncardiovascular death is prevalent in this population, who are often older and have a higher burden of comorbidity, recent studies suggest that the absolute rates of cardiovascular events are similar to those with type 1 myocardial infarction. [6][7][8] Early risk stratification is important to inform prognosis and to guide management in patients with acute coronary syndrome. The GRACE (Global Registry of Acute Coronary Events) 2.0 score 9,10 and the TIMI (Thrombolysis In Myocardial Infarction) score 11 are recommended in international guidelines and used widely in clinical practice. [12][13][14] However, these scores were developed before the introduction of high-sensitivity cardiac troponin assays and the classification of myocardial infarction according to mechanism. Type 2 myocardial infarction is a more heterogeneous condition than type 1 myocardial infarction. Few risk stratification tools have been optimized for this patient group, and none are recommended by current guidelines.
We aimed to derive and validate a new risk stratification tool for use in patients with type 2 myocardial infarction to determine the likelihood of future myocardial infarction or death that could assist clinicians in the targeting of further investigation and secondary prevention.   Table 1). All other deaths were classified as noncardiovascular. In the multicenter, international validation cohort, the cause of death during follow-up was obtained from the patient's hospital notes, the family physician's records, and the national registry on mortality. Cardiovascular death was defined as death from myocardial infarction, stroke, heart failure, or sudden cardiac death, or death within 7 days of cardiovascular intervention.

All index admission myocardial infarction events
were excluded in all 3 cohorts.
STATISTICAL ANALYSIS. Baseline characteristics were summarized for patients with type 2 myocardial infarction enrolled in the derivation and validation cohorts. Continuous variables are described by using mean AE SD or median (IQR) as appropriate; categorical variables are described as frequencies and percentages. Where data were missing in the derivation cohort, this was assumed to be at random after visual inspection, and multiple imputation using chained equations was performed (Supplemental Figure 1).
Survival analysis was performed by using Cox proportional hazards models for the primary and secondary outcomes separately. For the secondary outcome, noncardiovascular deaths were censored to account for competing risks in a cause-specific model.  Although the T2-risk score was derived by using hs-cTnI, we have modeled the relationship between hs-cTnT and hs-cTnI concentration using linear regression, allowing the user to employ the T2-risk score when either assay is available (Supplemental  year (Supplemental Figure 3). For the primary outcome, the T2-risk score displayed good discrimination (AUC: 0.76; 95% CI: 0.73-0.79) (Figure 2A). For the secondary outcome, discrimination was similar  Figure 3A). Further assessment using bootstrapped bias-corrected calibration curves produced similar results (Supplemental Figure 4). We  Table 4).  Figure 4).
The characteristics of these patients are presented in Supplemental Table 5. These thresholds were applied to both validation cohorts without recalibration. In in the high-risk category and 4% (9 of 252) in the lowrisk category (Supplemental Figure 6).  Table 3.

DISCUSSION
We have derived and externally validated a risk stratification tool to guide prognostication in patients with type 2 myocardial infarction. We show that routinely recorded clinical variables can be used to    The T2-risk score facilitates identification of patients at highest risk of future myocardial infarction or death. Although we trained the T2-risk score for this outcome to enable a direct comparison with the GRACE 2.0 score, heart failure is also an important outcome for patients with type 2 myocardial infarction, and it is likely that risk prediction could be further refined with cardiac imaging. In patients with type 2 myocardial infarction, coronary disease and left ventricular dysfunction are highly prevalent. In a prospective cohort study enrolling patients with type 2 myocardial infarction, coronary artery disease was identified in 68% of patients and was obstructive in 30%. 31 Importantly, this was previously unrecognized in more than one-half, with fewer than one-third of death. Although this approach has been validated, this may lead to overprediction of secondary outcomes. 33 We did not evaluate the performance of the T2-risk model in patients with acute nonischemic myocardial injury, but these patients share similar characteristics and outcomes, and it would be of interest to evaluate in future studies. The derivation and validation cohorts used the hs-cTnI and hs-cTnT assay, respectively, but we modeled the relationship in patients with both concentrations measured and applied a linear regression correction, revealing robust performance. We used multiple imputation for missing variables in the derivation cohort that were presumed to be at random, and showed in a sensitivity analysis that similar performance was observed in the complete data set without imputation. We were not able to incorporate all physiological parameters nor include phenotypic information known to influence prognosis. Although all index and outcome events were adjudicated in the derivation cohort and the multicenter validation cohort, ICD-10 coding was used for outcomes in the single-center validation cohort, which could contribute to misclassification and underestimate T2-risk model performance.
Finally, although our risk stratification tool seems to delineate risk well, prospective evaluation is required if it is to be used to guide treatment decisions in practice.