Prognostic value of ECG changes in non-ST elevation ACS patients

From the Division of Cardiology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada

Many studies have shown that ST-segment depression on admission electrocardiograms (ECGs) provides timely and independent prognostic insights for patients with non—ST-segment elevation acute coronary syndromes (ACS).1-3 Recent evidence from the Platelet IIb/IIIa Antagonist for the Reduction of Acute Coronary Syndrome Events in a Global Organization Network (PARAGON-A) study has extended this by demonstrating a quantitative relationship between the magnitude and distribution of ST-segment depression and short- and long-term outcomes.4 The admission ECG does not reflect the dynamic nature of myocardial ischemia, however, because it is limited to a single point assessment. Little is known about the prognostic utility of combining the discharge and admission ECGs to predict short- and long-term outcomes in patients with non–ST-segment elevation ACS. Therefore, we undertook the current study to (1) examine the prevalence of ST-segment depression on admission and discharge ECGs; (2) determine whether ST changes at discharge provide prognostic value for predicting death and recurrent myocardial infarction (MI) at 6 months in addition to admission ECGs; and (3) determine the prevalence and prognostic implications of Q-waves on admission and discharge ECGs.

Patients and methods

Our study included 918 patients from the PARAGON-B Troponin T substudy who had non—ST-segment elevation ACS and survived to hospital discharge free of confounding factors.5 All patients had evidence of cardiac ischemia, such as elevated creatine kinase-MB or troponin levels or electrocardiographic changes.

ECG analysis. All 12-lead ECGs were evaluated centrally by two readers at the ECG core laboratory at the Canadian VIGOUR Centre who were blinded to outcomes. Four patient groups were identified according to the ST status of admission and discharge ECGs. Group 1 included patients who had no ST-segment depression on admission or discharge; group 2 included patients who had no ST-segment depression on admission but developed new ST-segment depression at discharge; group 3 included those

who had ST-segment depression on admission with normalization by discharge; and group 4 included those with ST-segment depression on admission and discharge. The median interval between admission and discharge ECGs was 8 days (4 to 12 days for 25th and 75th percentiles, respectively).

Statistical analysis. Data are presented as medians with 25th and 75th percentiles for continuous variables. The Kruskal-Wallis test was used for group comparisons, and a Fisher’s exact test or chi-square test was employed for categorical variables. Kaplan-Meier survival estimates were used to compare the time to the first occurrence of the end point among the groups. A multivariate logistic analysis and Cox proportional hazards regression model were performed to determine how baseline characteristics and ECG variables affected outcomes. All tests were two-sided and had a 5% level of significance.

Results

Of the 918 patients studied, 542 (59%) had ST-segment depression on admission and 376 (41%) did not (table 1). Their 6-month death rates were 4.4% and 0.8% (P = .002), respectively. By discharge, the 35 patients (9.3%) who did not have ST-segment depression on admission had developed ST-segment depression, whereas 320 (59%) of those with admission ST-segment depression had normalized ST segments.

The patients’ key baseline characteristics, the in-hospital cardiac invasive procedures they were treated with, and their clinical outcomes according to the ST status of admission and discharge ECGs are shown in table 2. Patients who had new ST-segment depression by hospital discharge were more likely to have a history of congestive heart failure (CHF), a previous MI, and prior revascularization. Those with persistent ST-segment depression at discharge were older, had higher troponin T levels, and were more likely in Killip class II or higher on admission. The in-hospital coronary angiography rate for these patients was low (45.4%), yet they had a greater frequency of double and triple vessel disease compared with the other groups.

At 6 months, patients who had new ST-segment depression at discharge had more recurrent MIs

than those who never had ST-segment depression (20.6% versus 7.4%;

P = .018) and higher rates of death and recurrent MI (20.6% versus 8.3%; P = .03). Patients with persistent ST-segment depression on discharge had higher rates of death

(6% versus 0.9%; P = .001), recurrent MI (16.3% versus 7.4%; P = .002), and death and recurrent MI (20% versus 8.3%; P < .001) at 6 months compared with those who never had

ST-segment depression.

We also performed a Kaplan-Meier analysis to compare the rates of death and death and recurrent MI at 6 months according to the patients’ ST status at baseline and at discharge (figure). Patients with new ST-segment depression at

discharge were not included in this survival analysis because of their small sample size. Mortality rates differed significantly at 6 months between patients with no ST-segment depression on either admission or discharge and those with persistent ST-segment depression (P = .001) and those with normalized ST-segment depression at discharge (P = .038). The adjusted relative risk of 6-month mortality among patients with normalized ST-segment depression compared with those without ST-segment depression was 3.38 (95% confidence interval [CI], 0.914—12.3); comparing those with persistent ST-segment depression at discharge with those with no ST-segment depression, the adjusted relative risk was 5.18 (95% CI, 1.45–18.5; P = .011). Patients with persistent ST-segment depression had significantly higher rates of death and recurrent MI compared with patients with no ST-segment depression on both admission and discharge ECGs and those with normalized ST-segment depression by discharge (all P ≤ .005). At 6 months, the adjusted relative risk of death and recurrent MI among patients with normalized ST-segment depression compared with those without ST-segment depression was 1.20 (95% CI, 0.72–2.14); among patients with persistent ST-segment depression at discharge compared with those without

ST-segment depression, the adjusted relative risk was 2.58 (95% CI,

1.56—4.27; P < .001).

Q-wave status of admission and discharge ECGs. Table 3 shows the prevalence of Q-waves on admission

and discharge ECGs.6 On admission, 662 patients (72%) did not have

Q-waves, whereas 256 (28%) patients had Q-waves. Discharge ECGs demonstrated 320 patients (35%) had Q-waves and 53 (6%) lost their

Q-waves. Of the 249 patients who had a previous MI, 136 (54.6%) did not have admission Q-waves. Compared with those without a Q-wave or with Q-waves that disappeared

at discharge, patients with new

Q-waves or persisting Q-waves on discharge had more death (4.8% versus 1.9%; P = .021), recurrent MI (13.8% versus 8.3%; P =.011), and death and recurrent MI (16.4% versus 9.6%; P = .005) at 6 months. Patients with discharge Q-waves also had higher troponin T levels (0.081 ng/mL; range, 0—0.44 ng/mL) compared with those without discharge Q-waves (0.03 ng/mL; range, 0–0.21 ng/mL; P = .001). The presence of discharge Q-waves was highly significant in predicting

6-month death (2.52; 95% CI, 1.14-5.56; P = .021) and death and recurrent MI (1.83; 95% CI, 1.22—2.75;

P = .003) after adjusting for key baseline characteristics and status of ST-segment depression.

Discussion

Identifying patients with non—ST-segment elevation ACS who are at increased risk is crucial if we are to wisely allocate limited health resources toward appropriate intervention. The ECG remains a simple, noninvasive tool to detect evolutionary changes from admission to discharge.

Prognostic value of admission and discharge ECGs. Our findings indicate an association between ST-segment depression on the admission ECG and adverse short- and long-term outcomes. Patients with persistent ST-segment depression on discharge were older with more

prior MI and heart failure; their Killip class on admission and higher troponin T levels suggest a greater risk. Although there was a greater frequency of multivessel disease in this population, the frequency of angiography was lower. The poor prognosis of these patients suggests, but does not prove, that a more aggressive approach could be rewarding. This finding is consistent with previous studies showing the importance of ST-segment depression on the admission ECG, including our study.7-9

Our study provides new insights into the evolutionary ST-segment changes that occurred in 39% of our patients between admission and discharge. Most patients (90%) had normalization of ST-segment depression at discharge. We provide novel in-formation on the relationship between the discharge ECG and long-term outcomes in non—ST-segment elevation ACS. Schechtman and colleagues showed that ST-segment depression on the discharge ECG was an independent predictor of poor prognosis in non–ST-segment elevation ACS patients.10 Our results support and extend these findings

to a more comprehensive population. Important differences between the two studies include our larger sample and broader inclusion criteria. The mechanism by which persistent ST-segment depression relates to a worse prognosis is unclear. It could signal residual areas of ischemic myocardium, hibernating myocardium, or continuing silent myocardial ischemia.11-13

Prognostic value of Q-waves on discharge ECGs. The study by Kleiger and colleagues showed that Q-waves could develop during hospitalization or at discharge in patients with non—ST-segment elevation MI, the majority (70%) developing within 3 days.14 Our study provides new information on the value of Q-waves on the discharge ECG for predicting outcomes. Even after exclusion of in-hospital recurrent MI, those with discharge Q-waves had higher 6-month mortality rates. Because this finding was present at discharge as opposed to early after admission, it is unlikely that it reflects a missed ST-segment elevation MI or a temporal ECG lag. Most of our patients who had a previous MI (54.6%) did not have Q-waves on the admission ECG. This finding is consistent with those of Marcus and colleagues, who found that 42% of patients who had a previous acute MI showed a regression of Q-waves.15

Conclusion

Our study highlights the utility of adding a discharge ECG to the care plan of ACS patients to craft an appropriate follow-up strategy.