Correlation of cumulative ST elevation with left ventricular ejection fraction and 30-day outcome in patients with ST elevation myocardial infarctionV Kiron, PV George
Department of Cardiology, Christian Medical College, Vellore, Tamil Nadu, India
Keywords: ST elevation myocardial infarction, ST resolution percentage, sum ST elevation
Selevation myocardial infarction (STEMI) constitutes approximately 25% to 40% of all acute coronary syndromes and has a hospital mortality of 5% to 6% and long-term (1 year) mortality of up to 20%., Early risk stratification tools such as TIMI risk score, GRACE score, and so on have shown good correlation with short- and long-term mortality., Late risk stratification consists of an assessment of left ventricular function by echocardiography or ischemia evaluation by means of a coronary angiogram, stress testing, and imaging modalities such as cardiac magnetic resonance imaging (MRI). People with persistently depressed left ventricular ejection fraction (LVEF) are at high risk for arrhythmic and non-arrhythmic deaths.
STEMI is a medical emergency, and often, the only investigation available prior to reperfusion therapy is the electrocardiogram (ECG) which is obtained within 10 min of the patients' arrival. In most developing countries, investigations like echocardiography and cardiac MRI are not available at all centers managing STEMI. Therefore, it makes sense to prognosticate patients with STEMI based on ECG alone. Risk stratification based on ECG is possible by means of various scoring systems (e.g., Selvester score); however, they have multiple parameters to assess and are not practical for routine clinical use. We aim to calculate the sum of ST elevation (STE) in millimeters in the ECG and assess its correlation with the LVEF at discharge, and with the 30-day outcome. STE is easily reproducible at the bedside, and it can help in prognostication of patients presenting with STEMI in the hospitals where echocardiography for assessment of LVEF is not readily available.
This was a prospective cohort study conducted in a tertiary care referral hospital in South India. Patients with a diagnosis of STEMI as per the third universal definition of myocardial infarction (MI) by ACC/AHA-2012 were assessed. All patients aged 18 years and above who gave a valid written consent and presented within 24 hours from the index pain were included in the study, provided they were willing to undergo reperfusion therapy. Patients were excluded if they had overt contraindications for thrombolytic therapy and were not willing for angioplasty, had late presentation (beyond 24 hr) or unclear window period, received thrombolytic therapy elsewhere prior to referral to our center, resuscitated cardiac arrest, or those with baseline left bundle branch block (LBBB) on ECG.
At admission, all patients underwent a standard 12-lead ECG and routine blood investigations as per the institutional protocol. All patients were offered primary angioplasty as the preferred mode of revascularization, and those who were not willing for the same were offered thrombolysis. All patients underwent a repeat ECG 90 min after revascularization to assess the response to therapy. Patients also underwent echocardiography prior to discharge, and the LVEF was assessed. All patients were advised a clinical follow-up after 30 days, on an outpatient basis, or earlier, if symptomatic.
A standard 12-lead ECG was acquired using a Philips TC 20 Cardiograph (Koninklijke Philips N.V.) at a page speed of 25 mm/s and a gain setting of 10 mm/mV. In patients with inferior MI, additional posterior leads (V7-V9) were recorded. The ECGs were converted to PDF format and analyzed on the 4× zoomed image using a caliper. ST elevation (STE) was measured to the nearest 0.5 mm mark at the J point based on the Third Universal Definition of Myocardial Infarction 2012 guidelines. ST measurement at the J point also served as a good standardization with consistent reproducibility. The residual STE was measured in a repeat ECG 90 min after reperfusion to calculate the percentage of ST resolution (STR%).
A Phillips HD CX50 system with S5-1 MHz probe (Phillips Medical Systems, Andover, MA, USA) was used for the study. The LVEF was measured using biplane Simpson's method and Teichholz method.
Clinical endpoints during the hospital stay were the duration of stay, re-infarction, congestive cardiac failure, and death. These parameters were assessed once again during the 1-month follow-up after-study recruitment.
The primary objective was to assess the correlation between the STE at admission and the LVEF prior to discharge. The secondary objectives were to assess the correlation between the STR% and LVEF, and the correlation between both STE and STR with the 30-day mortality and 30-day MACE (major adverse cardiac events).
Sample size calculation
A pilot study was done with 40 patients (20 with anterior MI and 20 with inferior MI). The two parameters of interest were thought to have a linear correlation. It was hence decided to calculate the Pearson correlation coefficient to aid in sample size calculation. The Pearson correlation coefficient was −0.24 and −0.51 for anterior and inferior MI, respectively. The necessary sample size with a two-sided alpha error of 0.05 and a beta error of 0.20 was 132 patients with anterior MI and 28 patients with inferior MI.
Study approval and statistical analysis
Institutional Review Board (IRB) approval was sought prior to the commencement of the study (IRB minute no. 9137, 12/11/2014). All study variables were described using descriptive statistical methods. The continuous variables were compared using either independent two-sample t test or Mann–Whitney U test, as appropriate. Dichotomous variables were compared between the two groups using Chi-square or Fisher's exact test, as appropriate. The results were analyzed using SPSS version 18 (Chicago, IL, USA).
A total of 268 patients with a diagnosis of STEMI were assessed, of which 230 patients were finally enrolled [Figure 1]. The demographic and clinical profile of study participants is described in [Table 1] and [Table 2]. Very few patients belonged to Killip Class III and IV put together (9.1%). All the patients who died prior to ECHO after recruitment into the study belonged to Killip Class III and IV at admission.
Mean STE among the study population was 13.06 mm and the mean STR% was 55.3%. There was no significant correlation between the window period and STE (r = 0.054, P = 0.41). The STR%, however, showed a mild inverse correlation with the window period (r = −0.16, P = 0.015). The mean LVEF of the study population was 42.73% as assessed by biplane Simpson's method and 42.8% by Teichholz method [Table 3]. There was a good correlation between the various methods of assessment of ejection fraction. Patients with anterior MI had a statistically significant lower LVEF than those with inferior and inferoposterior MI (39.3% vs. 48.6% vs. 47.6%, P = 0.04). Mitral regurgitation was more common among those with inferior and inferoposterior wall STEMI.
There was a good inverse correlation (r = −0.639) between the STE and the LVEF [Figure 2]. The correlation was better in patients with inferior MI (−0.527) when compared to patients with anterior MI (−0.499) and inferoposterior MI (−0.47). An STE ≥15 mm had an odds ratio of 8.5 (3.94-18.34, P = 0.001) in predicting LVEF ≤35% with a sensitivity of 74% and a specificity of 72%. There was a moderate correlation between the STR% and the LVEF (r = 0.591, P ≤ 0.001). An STR ≤52.75% had an odds ratio of 8.06 (3.38-19.31, P = 0.001) in predicting LVEF ≤35% with a sensitivity of 65% and a specificity of 82% [Figure 3].
Based on the above analysis, we computed a formula as follows:
y = a + bx
where y = dependent variable “Sum of ST-Elevation”; x = independent variable “EF Biplane”; a = intercept point; and b = slope of the line. Sum of ST-Elevation = 37.34- 0.567 EF Biplane. This could be re-coded as follows:
LVEF = (37.34- sum ST elevation)/0.567
Using this formula, the approximate LVEF could be calculated from the admission STE in an emergency setting when echocardiography is not readily available. The formula has its inherent limitation of being applicable only in a narrow range of STE.
The strongest predictors of MACE and mortality at 30 days following STEMI were LV dysfunction on echocardiography and Killip Class ≥ III at presentation [Table 4]. An STE ≥15 mm is a good prognostic marker with odds ratios of 3.13 for MACE at 30 days and 6.54 for mortality at 30 days. An STR ≤52.75% also had a significantly increased risk for adverse outcome with an odds ratio of 3.17 and 5.67 for MACE at 30 days and mortality at 30 days, respectively.
The majority of our study patients were male (86%) resulting in underrepresentation of the female patients. Women constituted 20% of patients in the CREATE registry from India whereas 26% to 30% of patients in various international registries.,, This may affect the generalizability of our results to a standard population. Most patients in our study were relatively stable similar to the CREATE registry (85% Killip Class I) and the European registry (90% Killip Class I)., Over 90% of our patients presented within 12 hr window period as opposed to 65% in the CREATE registry. Owing to enhanced general awareness and increased health insurance coverage from the Government Health Insurance Scheme, numerous patients could undergo primary angioplasty. We found that 45% of patients underwent primary angioplasty which is significantly more than the 8% of patients as observed in the CREATE registry. An additional 16% of patients in our study underwent pharmacoinvasive or rescue angioplasty. A more recent data from the NIC (National Interventional Council) had an acute coronary syndrome setting angioplasty rate of only 30%.
There was no correlation between the STE and the window period. This is probably due to the fact that as the window period increased, Q waves would appear on the ECG and the STE may not increase further. There was a good inverse correlation of −0.64 between the STE at admission and the LVEF. This correlation was consistent across all types of STEMI and was best among patients with inferior MI as expected in our pilot study. Receiver operating characteristic (ROC) curve suggested a cutoff point of 15 mm STE which would predict an LVEF <35% with a sensitivity and specificity of over 70%.
An important aspect of this study was to derive a formula that would help in calculating the LVEF based on the STE on presentation ECG: LVEF = (37.34- STE)/0.567. A major limitation of this formula is its applicability to a very limited range of STE, mostly closer to the mean STE in the study (13 mm). Although this measurement is not very accurate, this could help in approximate estimation of the LVEF in a peripheral hospital where echocardiography is not readily available. We aim to prospectively analyze this equation in our institution and come up with its accuracy and further limitations in the future. A major drawback in this equation according to us is the sole reliance on the STE and not incorporating the other parameters which could potentially influence the LVEF, for example, Window period, type of MI, and so on. STR% also was found to have a good correlation to the LVEF (r = 0.59), and this trend persisted across all types of STEMI. Similar data could not be traced from any prior studies from India or abroad. Also considering the significant change in the therapeutic armamentarium, these data from the current era cannot be compared with the data from trials performed in the prethrombolytic/pre-interventional periods.
Observed mortality at hospital discharge in our study population was only 2.6% (6 patients) and the mortality at 1 month was 4.3% (10 patients) which is significantly lesser than the data from other registries. This observation resulted because most sick patients could not be recruited into the study.
Our study was not a blinded study which would lead to significant observer bias. The time delay between presentation and revascularization was not studied which could have affected the LVEF. The mortality rate was very less among the study participants, and therefore, significant correlations could not be drawn to the clinically meaningful hard endpoints (i.e., STE with mortality). The study population mostly consisted of patients in Killip Class I; therefore, its generalizability to entire STEMI spectrum is doubtful. The formula to calculate the LVEF from sum STE would be valid only on a limited range of STE. Parameters like window period and type of MI are not included in the formula calculating the LVEF. Although the presence of MR may result in underestimating the LV dysfunction, only 10 patients had significant MR in our study population, and hence, it was assumed that it did not affect the study outcome.
STE is a useful parameter on admission ECG in patients with STEMI. It has a good correlation with the LVEF and the hospital outcome (r = −0.64). STE cutoff of ≥15 mm has an odds ratio of 8.5 in predicting LVEF <35% and an odds ratio of 3.13 in predicting 30-day MACE. STR% post-revascularization also has a good correlation to the LVEF (r = 0.591). STR% ≤52.75% has an odds ratio of 8.06 in predicting LVEF <35% and an odds ratio of 3.17 in predicting 30-day MACE. In case of unavailability of an echocardiogram, STE on admission ECG could help predict the approximate LVEF by the following formula: (37.34 – STE)/0.567.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]