We identified 4230 consecutive patients with an intermediate or high likelihood of CAD who underwent SPECT between January 1, 2003 and March 1, 2011. The indication for the test was the presence of ischemic equivalent, consisting of symptoms or ECG findings consistent with CAD. Patients with previous myocardial infarction or revascularization were considered to have known CAD and were excluded from this study, as were patients with valvular heart disease or primary cardiomyopathy. Patients with a non-diagnostic ECG but with underlying ECG abnormalities, including left bundle branch block, and those with pacemakers were also excluded, as were patients unable to exercise. Of the remaining patients (1520), 60 (3.9%) were lost to follow-up. The study conformed to the ethical standards of the Declaration of Helsinki.

Subjects underwent technetium (Tc)-99m sestamibi rest-stress gated SPECT MPI with either a 1- or 2-day protocol. All patients performed a symptom-limited treadmill exercise test using standard protocols. During exercise, 12-lead ECG recording was performed at each minute of stress with continuous monitoring. At near maximal exercise, a weight-adjusted dose of Tc-99m sestamibi was injected and exercise continued for 1 min after injection. Tc-99m sestamibi SPECT imaging was begun 30 min after isotope injection. SPECT MPI was performed using an elliptical 180° acquisition for 64 projections at 20 s per projection. No attenuation or scatter correction was used. After filtered back-projection, short-axis, vertical, and horizontal long-axis tomograms were generated. SPECT data were analyzed by two imaging cardiologists with regard to the presence of reversible and/or fixed perfusion defects in short-axis and horizontal and vertical long-axis slices, as well as on polar maps, with a commercially available quantitative gated and perfusion SPECT software package, which also enabled assessment of left ventricular ejection fraction (LVEF). In addition, MPI scans were assessed semiquantitatively by segmental visual interpretation, using short-axis and vertical long-axis tomograms divided into 20 segments. Each segment was scored by consensus of two experienced observers using a five-point scoring system (0 = normal, 1 = equivocal, 2 = moderate, 3 = severe reduction of tracer uptake, and 4 = absence of detectable uptake). A summed stress score (SSS) was obtained by adding the scores of the 20 segments of the stress sestamibi images; SSS <3 was considered normal.

Cigarette smoking was defined as active smoking in the last six months. Diabetes was defined as increased fasting plasma glucose of ≥126 mg/dl, HbA1c of ≥6.5%, or undergoing treatment with insulin or hypoglycemic agents. Hypertension was defined as systolic/diastolic blood pressure >140/90 mmHg or current use of any antihypertensive medication. Dyslipidemia was defined as total cholesterol ≥220 mg/dl, triglycerides ≥150 mg/dl, high-density lipoprotein (HDL) cholesterol <40 mg/dl, or current use of lipid-lowering drugs. These four variables (cigarette smoking, diabetes, hypertension and dyslipidemia) were considered as cardiovascular risk factors, together with the presence of previous myocardial infarction or of peripheral arterial disease. Age >45 years (males) or >55 years (females) was also considered a cardiovascular risk factor. An ECG stress test was considered positive if there was a horizontal or downsloping ST-segment depression of ≥0.1 mV at 0.08 s after the J-point.

Events were defined as cardiac death (confirmed by review of death certificate, hospital chart, or physician's records), non-fatal myocardial infarction (documented by appropriate cardiac enzyme and electrocardiographic changes), revascularization or admission due to new-onset heart failure. Events at one year and during follow-up were analyzed. The median follow-up was 16.8 months (interquartile range: 14.0-39.6 months).

The statistical analyses were performed with SPSS (Statistical Package for the Social Sciences), version 17.0. Categorical or dichotomous variables are expressed as absolute values and percentages, and were compared with Pearson's chi-square test. Continuous variables with normal distribution are expressed as mean ± standard deviation, and the Student's t test was used for comparisons between groups. Continuous data from more than two groups were compared with ANOVA. A binomial logistic regression model was used to evaluate the independent contribution of ECG response and MPI to the risk of events at one year. Age in years, number of cardiovascular risk factors, LVEF as a continuous variable, ECG response and MPI results were included in the model in backward stepwise analysis. C statistics, as a measure of discriminative power, were calculated for predicting cardiac events in the first year after the test. A Cox proportional hazard analysis was carried out to assess the independent role of both tests (ECG and SPECT) for predicting mortality during follow-up using the same variables as in the binomial logistic regression. Adjusted odds and hazard ratios and 95% confidence intervals (CI) were presented. Kaplan-Meier curves were constructed to evaluate prognosis during follow-up of different groups based on ECG response and MPI results, and the results were analyzed using the log-rank test. Receiver operating characteristic (ROC) curve analysis was performed to establish the diagnostic value of both tests (ECG and SPECT) to predict events. A p value <0.05 was considered statistically significant.

A total of 1460 patients (777 female; mean age 62.6±11.4 years) were included in the study. Table 1 shows their baseline clinical characteristics. Ischemic ECG changes were observed during stress testing in 271 patients (18.5%) and 362 patients (24.7%) had positive MPI results (Table 2).

During the follow-up period (1-92 months), 224 (15.3%) of the total population presented cardiac events. Compared to patients who were still alive at the end of follow-up, those who died had a higher incidence of positive ECG response and positive MPI results (Table 3). There were no statistically significant differences between groups (with and without events) in terms of other variables (age and gender, cardiovascular risk factors and LVEF). The hazard ratios (HR) of ECG changes and MPI results to predict events were 1.506 (95% CI: 1.113-2.039, p=0.008) and 10.481 (95% CI: 7.799-14.080, p=0.001), respectively (Figure 1). For comparison of different groups (based on ECG and MPI results), survival curves were constructed for the four groups (Figure 2). As can be seen, the prognosis of patients with positive MPI was clearly worse than that of patients with negative MPI, independently of ECG changes during stress. In multivariate analysis (Table 4), adjusted for age, gender, number of cardiovascular risk factors and LVEF, MPI proved an independent long-term prognostic factor (HR 10.536, 95% CI: 7.759-14.308; p=0.001), but not ECG changes (HR 1.356, 95% CI: 0.994-1.850; p=0.055).

Figure 1.

Temporal trends in the occurrence of cardiac events in the groups with positive and negative ECG response (A) and SPECT MPI result (B).

(0,2MB).

Figure 2.

Kaplan-Meier survival curves according to ECG and SPECT findings.

(0,11MB).

At one year after testing, 127 (8.7%) of the total population presented events. Patients with ECG changes during exercise testing were at greater risk of events at one year than those without (OR: 1.768, 95% CI: 1.169-2.675; p=0.007), as were patients with positive SPECT (OR 13.684, 95% CI: 8.851-21.155; p=0.001). The area under the curve (AUC) for MPI results to predict events at one year (0.787: 95% CI: 0.743-0.831) was significantly larger than for ECG changes (0.549: 95% CI: 0.495-0.604) (p=0.002). The combined model of ECG changes and MPI results increased the AUC for the prediction of events to 0.802 (95% CI: 0.760-0.844), without reaching statistical significance compared to the AUC for MPI results (p=0.691). In multivariate analysis (Table 5), adjusted for age, gender, number of cardiovascular risk factors and LVEF, both ECG changes and MPI results proved independent predictors of 1-year outcomes (OR 1.737, 95% CI: 1.091-2.767; p=0.020 for ECG changes, and OR 13.670, 95% CI: 8.691-21.503; p=0.001 for MPI result). The C statistic of the model was 0.811.

Normal MPI with a positive ECG in the exercise test was observed in 190 patients (13.0%) and the opposite (positive MPI and negative ECG) in 281 patients (19.2%). No statistically significant differences were found between the clinical characteristics of either of these groups compared to the rest of the study population. Figure 3 shows the survival curves in the discrepant groups adjusted by gender and hypertension. Of the 1098 patients (75.2%) with negative SPECT, the presence of ECG changes during stress testing was not associated with more events at one year (OR 1.857, 95% CI: 0.810-4.257; p=0.244) or more events throughout the follow-up period (HR 1.214, 95% CI: 0.646-2.282; p=0.547). The same was true for the 362 patients (24.8%) with positive SPECT (OR for events at one year 1.479, 95% CI: 0.866-2.524, p=0.151, and HR for follow-up events 1.203, 95% CI: 0.848-1.705; p=0.307). However, in the latter group (positive MPI) we found that non-hypertensive patients (n=110) with ischemic ECG changes during stress testing (n=20) had a worse prognosis (percentage of events 70% vs. 37.8% in hypertensive patients with positive MPI and ischemic ECG changes) (HR for positive ECG response 1.937, 95% CI: 1.030-3.642, p=0.009).

Figure 3.

Survival curves in the discrepant groups adjusted by gender and hypertension.

(0,23MB).

In addition we focused on patients with negative MPI and positive ECG. In this subgroup, 11 (6.3%, 0.75% of the total study population) had cardiac events during follow-up. Of these, seven were undergoing cardiac catheterization, and multivessel disease was detected in all except one (who had proximal left anterior descending disease). In these patients, the ECG response was markedly positive (widespread ST depression ≥2 mm).

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that they have followed the protocols of their work center on the publication of patient data and that all the patients included in the study received sufficient information and gave their written informed consent to participate in the study.

The authors have obtained the written informed consent of the patients or subjects mentioned in the article. The corresponding author is in possession of this document.