We enrolled patients who were admitted to our cardiovascular center with acute coronary syndrome (ACS) between January 2007 and December 2008, or those who were referred to the cardiac rehabilitation clinic after undergoing a PCI. The patients participated in a 6-8 week hospital based exercise after discharge. Then they were advised to participate in a community-based self-exercise program. We enrolled patients that could be followed up for 9 months after ACS onset. Our study was a prospective cohort study. The exclusion criteria included patients with prior experience with PCI or CABG; a left ventricular ejection fraction (LVEF) of less than 40%, uncontrolled arrhythmia, hypertension, or diabetes; as well as other issues that would preclude a potential patient from participation in the program.

We followed 51 patients, and among these, the follow up CAG positive group included 14 patients and the CAG negative group included 38 patients. There was no significant difference in the gender distribution, mean age, rate of patients who were diagnosed with ST elevation myocardial infarction (MI) or non ST elevation MI or unstable angina (UA), stent insertion site distribution among the coronary artery segments, past history (which included hypertension, diabetes mellitus, and dyslipidemia), body mass index, LVEF at the initial cardiac rehabilitation (CR), or the sort of cardiac drugs between the two groups (Table 1).

Subject patients received CAG at the onset of ACS and nine months afterward. They also received ETT at the start of the CR program and several days before receiving a follow-up CAG. All of the subject patients had been informed about the purpose, method and potent complication of a CAG and ETT before these procedures. Consent was received from the study subjects prior to beginning the study and then we proceeded with the study.

The CAG was conducted in the following fashion. A standard image acquisition was performed using ≥2 angiographic projections of the stenosis. Intracoronary nitroglycerin was administered to provide maximum coronary vasodilation. At follow-up, a repetition of the identical angiographic projections of the lesion was performed. The lesion length was defined as the axial extent of the lesion that contained a shoulder-to-shoulder lumen reduction by 20%. Selected images for the analysis were identified by the use of angiographic projections that demonstrated stenosis in an unforeshortened view, minimized by the degree of vessel overlap, and displayed stenosis in its "sharpest and tightest" view. The target lesion was defined as the stented segment plus a 5 mm segment proximal and distal to the stent or progression of the native coronary vessel. Angiographic follow-up was routinely performed at 6-9 months after the index procedure unless an earlier angiography was required for clinical reasons. At 6-9 months, angiographic restenosis was measured as an in-segment late loss of the stented coronary artery in addition to newly developed native coronary stenosis via quantitative coronary angiography using CAAS 5.9 (Pie Medical Imaging B.V., Maastrichitt, Netherlands).

We defined angiographic positive cases as the binary angiographic restenosis which was defined as either the progression of a percent diameter stenosis up to 50% or more or a newly developed native coronary stenosis of 75% or more at the qualifying angiographic follow-up. The late luminal loss per stent was defined as the minimal luminal diameter (MLD) immediately after the procedure minus the MLD at a 6-9 month follow-up. Two experienced cardiologists independently interpreted the images blinded. Disagreement was resolved by reanalysis and consensus.

Subject patients received ETT at the start of the CR and at nine months after that. A real-time recording 12-channel ECG Q4500 (Quinton Instrument Co., Boston, USA), a respiratory gas analyzer QMC (Quinton Instrument CO., Boston, USA), an automatic blood pressure and pulse monitor Model 412 (Quinton Instrument Co., Boston, USA), and a treadmill Medtrack ST 55 (Quinton Instrument Co., Boston, USA) were included in the test. The VO_2max was measured with a respiratory gas analyzer, and the maximum heart rate, the stable heart rate, and the myocardial oxygen demand (MVO₂) were estimated using an ECG, and an automatic blood pressure and pulse monitor. The MVO₂ was calculated by multiplying the systolic blood pressure and heart rate as the rate pressure product (RPP). The submaximal MVO₂ was measured at the end of stage 3 of the modified Bruce protocol.

We evaluated the usefulness of the following parameters in predicting restenosis. Based on the results of ETT, we measured: criterion A as the sum of the changing rate of the ST segment at lead II, III, aVF, V4, V5, V6 (ΣΔST index); criterion B as the sum of the ΔHR adjusted index of ΔST at the same leads (ΣΔST/ΔHR index); criterion C as the sum of the ΔRate pressure product (RPP) adjusted index of ΔST at same leads (ΣΔST/ΔRPP index). We compared these criteria with the results of follow up CAGs. We then identified the predictive power of each index of ETT for restenosis after PCI.

The data was statistically analyzed with SAS Enterprise Guide 4.1 (4.1.0.471). The Wilcoxon rank sum test was used to compare gender, age, the rate of patients who were diagnosed as ST elevation myocardial infarction (MI) or non ST elevation MI or unstable angina (UA) with a past history which included hypertension, diabetes mellitus, dyslipidemia, body mass index, LVEF at the initial cardiac rehabilitation (CR), a sort of drug between the positive CAG group and the negative CAG group. The Wilcoxon signed rank test was used to compare the criterion for A, B, C in each group at baseline and 9 months. It was also used to compare the resting heart rate (HR), the maximal HR, the VO_2max, the maximal and the submaximal myocardial oxygen demand (MVO₂) in each group at baseline and 9 months. A p-value of <0.05 was considered statistically significant.

In the negative CAG group, criterion A at baseline and 9 months was not statistically significant. However, in the positive CAG group, criterion A was 5.56±2.59 at baseline and increased to 9.85±5.20 at 9 months. It was statistically significant (p<0.05) (Table 2).

In the negative CAG group, criterion B was 10.1±7.33 at baseline and decreased 9.3±5.6 at 9 months. It was not statistically significant. However, in the positive CAG group, criterion B was 8.74±3.43 at baseline and increased to 13.7±5.20 at 9 months. It was statistically significant (p<0.05) (Table 2).

In the negative CAG group, criterion C was 0.10±0.13 at baseline and decreased 0.06±0.04 at 9 months. And it was statistically significant (p<0.05). However, in the positive CAG group, criterion C was 0.11±0.08 at baseline and decreased to 0.10±0.08 at 9 months. It was not statistically significant (Table 2).

There was a significant difference in the submaximal rate pressure product (RPP_submax) between the two groups. However, the maximal HR, the maximal RPP and the VO_2max increased more significantly in the negative CAG group at baseline and 9 months than they did in the positive CAG group (p<0.05) (Table 3).