Efficacy of CPET Combined with Systematic Education of Cardiac Rehabilitation After PCI: A Real-World Evaluation in ACS Patients

The present single-center retrospective analysis was conducted to determine the effectiveness of one-time CPET in hospitals combined with systematic education of CR after stent implantation. From 1 April 2015 to 20 May 2017, 11,345 acute coronary syndrome (ACS) patients hospitalized for PCI were retrospectively screened; 536 patients were excluded because of incomplete information or inability/unwillingness to cooperate with telephone follow-up; 4001 patients received CR; and 6808 patients received routine postoperative education without CR (controls), as shown in Fig. 1. All data on baseline characteristics, medication, and clinical follow-up including Seattle Angina Questionnaire (SAQ) score were routinely collected using the cv-net clinical data collection system (Beijing Crealife Technology Co., Ltd.) by a specialized clinical research team. CR was performed at the doctor’s discretion after evaluation of the patient’s overall health and socioeconomic status, disease severity, and willingness to undergo CR. Before CR, patients signed an informed consent form expressing understanding of the risks of CPET. All patients’ clinical treatment and medications were optimized according to current PCI guidelines [10]. The Medical Ethics Committee of General Hospital of Northern Theater Command has approved our study and waived the need for patient consent. The data of patients in our study were not identifiable.

CPET was performed by CARDIOVIT AT-104 PC Ergo-Spiro (SCHILLER) in the Heart Rehabilitation Center at our department. CPET is a diagnostic method that reflects the indexes of human cardiopulmonary function during exercise with increasing load. During CPET, the subjects' static electrocardiogram and static lung function (vital capacity/maximum ventilation) were individually collected. The patients were asked to breathe calmly, exhale forcefully to the limit after three or four times, then inhale forcefully to the limit, and then exhale forcefully. Cardiopulmonary function at different power levels and test results of exercise electrocardiography/exercise pulmonary function and cardiopulmonary function were also collected. The cardiopulmonary test entails four phases: resting phase (patient sitting on a vehicle, stationary for 1–2 min); warm-up phase (patient begins to pedal, no power load, pedal’s speed at 40 times per minute, until the respiratory quotient steadily returns to about 0.85; lasts 2–3 min); exercise phase (patient begins to pedal according to the previously configured power scheme and decides whether to terminate the power load according to the termination indication); and the recovery phase (patient keeps bicycling, no power load, depending on the recovery of heart rate to determine the end of the experiment; lasts 2–3 min). The test results of CPET including maximal oxygen uptake, metabolic equivalent (MET), and anaerobic threshold will be assessed for patients in the CR group. The rating of perceived exertion (RPE) will also be assessed using the Borg scale after CPET, which will be combined with other signs or symptoms to evaluate the cardiopulmonary function and exercise capacity. The indications for termination of CPET in our study were in accordance with the scientific statement from the American Heart Association [11].

After CPET, systematic prescription and education of CR for post-discharge were performed, which comprise rehabilitation exercises, nutrition, psychological intervention, smoking cessation, and medications according to the guidelines for coronary heart disease rehabilitation/secondary prevention [12]. Rehabilitation exercise prescription was individualized according to the maximum MET value suggested by CPET. Detailed exercise types and MET values are listed (see Table S1 in the electronic Supplementary Material for details). Exercise intensity and duration were adjusted every 3–6 months. The exercise was advised to be halted if any of the following occurred: (1) chest pain, dyspnea, or dizziness during or after exercise; (2) heart rate fluctuation > 30 beats/min; (3) blood pressure > 200/100 mmHg or systolic blood pressure increase > 30 mmHg or decrease > 10 mmHg; (4) electrocardiogram monitoring during exercise showed ST segment depression ≥ 0.1 mV or elevation ≥ 0.2 mV; (5) severe arrhythmia occurred during or after exercise. Nutritional recommendations were individualized based on the principles of lower caloric intake, lower cholesterol intake, and balanced intake of fruits and vegetables. The psychological intervention was prescribed by a psychologist, while smoking cessation and medications were prescribed by clinical doctors. Finally, exercise type and volume of habitual exercise were routinely recorded by telephone follow-up after discharge.

The main outcome was quality of life as determined by SAQ scoring. The SAQ quantifies five domains, namely physical limitation (9 items), angina stability (1 item), angina frequency (2 items), treatment satisfaction (4 items), and quality of life (3 items), to assess angina and its impact on health and psychological status of the patients [13]. Degrees of physical limitation, angina stability, angina frequency, and treatment satisfaction from worst to best status ranged from 1 to 6 and quality of life from 1 to 5. SAQ scores overall and their components ranged from 0 denoting the worst to 100 denoting the best. Major adverse cardiovascular event (MACE) and its components of target vessel revascularization (TVR), myocardial infarction (MI), and cardiac death were recorded and analyzed [14]. MI [15] was defined as the presence of abnormal imaging findings of MI, clinical symptoms, or electrocardiographic changes combined with an increase in the creatine kinase myocardial band fraction above the upper normal limits or an increase in troponin-T/troponin-I to > 99th percentile of the upper normal limit. TVR was defined as subsequent revascularization of the target vessel by either PCI (additional stent or angioplasty) or coronary artery bypass grafting (CABG).

Via telephone interview, each patient completed the SAQ questionnaire and a detailed volume/type of habitual exercise questionnaire 1–3 months after discharge. Follow-up for clinical events was performed in 1, 3, 6, 12, 24, 36, 48, and 60 months. Questions related to physical activity we prescribed were completed by all patients enrolled for analysis. The intensity of physical activity was categorized as mild (< 3 METs), moderate (3–6 METs), and high intensity (> 6 METs) [16], as detailed in Table S1 in Supplementary Material. Cumulative exercise time was defined as the cumulative time of walking, light household activities, and exercise according to rehabilitation prescription.

Categorical variables are expressed by numbers (percentages) and were compared using the chi-square test or Fisher’s exact test. Continuous variables are expressed as mean ± SD or median and were compared using the independent t-test; 2805 CR participants were matched 1:1 to controls using propensity scores for a total of 5610 patients analyzed. A non-parsimonious model was used containing all variables in Tables 1 and 2. Pairs were matched by the nearest neighbor matching model. The covariate balance achieved was assessed by calculating the standardized differences in covariates between the two groups. A difference  < 10% suggested the appropriate balance between groups. All tests were two-tailed. A p value < 0.05 was considered significant. R software was used for propensity score matching and SPSS 24.0 for statistical analyses.

After exclusion of patients with incomplete information or unwilling/unable to participate in follow-up interviews, the total CR rate at our center was 37%, and the longest follow-up period was 963 days, while the shortest follow-up period was 184 days, with a median follow-up of 583 days (Fig. 2); 2805 patients of the 4001 patients who received CR were successfully matched to 2805 non-CR controls out of the 6808 controls eligible for matching (Fig. 1).

Baseline characteristics of patients before and after propensity score matching (PSM) are presented in Table 1. Before PSM, compared with controls, patients in the CR group had less severe conditions indicating that physicians were more likely to recommend CR for patients with less severe disease status. The patients in CR group were younger (61.34 ± 10.67 vs. 57.42 ± 9.46 years, p < 0.01); more often men (74% vs. 77.7%, p < 0.01); and presented less often with ST segment elevation acute myocardial infarction (STEM) (22% vs. 10.4%, p < 0.01) and non-ST segment elevation elevation ACS (NSTE-ACS) (15% vs. 9%, p < 0.01) but more often with unstable angina ( 63% vs. 80.7%, p < 0.01). Moreover, this group often had less burden of risk factors for coronary artery disease, i.e., hypertension, diabetes, previous PCI, and prior stroke (58.6 vs. 55.3%, p < 0.01; 28.8% vs. 22.3%, p < 0.01; 20.7% vs. 15.6%, p < 0.01; and 5.7% vs. 2.7%, p < 0.01, respectively) except for smoking, which was higher (47.8% vs. 57.7%, p < 0.01). After the education of CR, the smoking rate decreased (22.5% vs. 5.2%, controls vs. CR, p < 0.01). In terms of medications (Table 2), the usage rates of aspirin, ticagrelor, and ACEI did not differ significantly between CR and controls, while the use of clopidogrel, β blockers, and statins was higher in the CR group. During the treatment and follow-up, 610 patients in controls and 358 patients in the CR group crossed over between clopidogrel and ticagrelor. After propensity score matching, all baseline characteristics and medications were balanced (Tables 1, 2).

After PSM, both study groups were followed up after discharge (Table 3). Among controls, 956 patients did not have exercise, while in the CR group the no exercise number was 22. More patients in the CR group did a small amount of physical exercise vs. controls (2736 vs. 1849). In the CR group, 47 patients followed the exercise prescribed for rehabilitation relative to none among controls. Physical exercise type differed between groups (p < 0.05). Cumulative exercise time (hours/week) was higher in the CR group (8.22 ± 6.17 vs. 3.00 ± 1.65, P < 0.05). Both before and after PSM, scores for the five components of the SAQ were higher for the CR groups vs. controls (all p < 0.01; Table 4). After PSM, the physical limitation (69.59 ± 10.96 vs. 57.49 ± 7.19), anginal stability (80.50 ± 18.21 vs. 58.82 ± 11.95), anginal frequency (78.58 ± 11.07 vs. 67.14 ± 22.41), treatment satisfaction (82.33 ± 13.21 vs. 56.84 ± 21.61), and quality of life (68.69 ± 18.33 vs. 60.26 ± 17.13) in the CR group were all higher than those in the control group.

As Table 5 shows, at median 461 days' follow-up, incidences of adverse cardiac events in CR group vs. controls before and after PSM were respectively: for TVR, 0.4% vs. 0.8%, p = 0.005 and 0.5% vs. 0.7%, p = 0.222; for MI, 0.5% vs. 0.5%, p = 0.398 and 0.5% vs. 0.7%, p = 0.397; for death, 0.2% vs. 0.3%, p = 0.292 and 0.2% vs. 0.2%, p = 0.781; and for MACE, 0.8% vs. 1.2% and 0.8% vs. 1.3%. However, as shown in Kaplan-Meier curves of MACE before and after PSM in Fig. 3, before PSM, cumulative hazard of MACE in controls was higher than in CR group (landmark p = 0.006), while after PSM, cumulative hazard of MACE did not differ between groups (landmark p = 0.621).

There were no fatal adverse events among all 4001 CR patients undergoing CPET. Only one patient suffered from ventricular tachycardia with no disturbance of consciousness and resolution 1–2 min later. Events during CPET included: posterior sternal burning pain (n = 1); chest tightness and shortness of breath (n = 3); chest tightness (n = 7); chest pain (n = 4); palpitation (n = 1); palpitation and ventricular bigeminy (n = 1); bradycardia (n = 1); and ventricular tachycardia (n = 1).

The present study has the inherent limitations of its retrospective observational cohort by follow-up design. Moreover, propensity score matching was used to adjust for differences in baseline characteristics between CPET combined CR and controls, which may introduce statistical bias. The observation of MACE being different or not between CR and controls depending on the statistical method used suggests the possible presence of confounders not included in the analysis or differences in follow-up duration.