The safety of early enrollment (<2 weeks after hospital discharge) into cardiac rehabilitation (CR) after recent coronary artery bypass graft (CABG) surgery or heart valve surgery (HVS) has not previously been assessed and has important policy implications. Consequently, we performed a detailed review of all clinical adverse events within 6 months of hospital discharge. We compared early and late attendees for patients undergoing CABG surgery or HVS and included patients with myocardial infarction (MI) as an additional control group. We analyzed 112 patients undergoing CABG surgery, 69 patients undergoing HVS, and 59 patients with MI. Median time (interquartile range) from hospital discharge to CR enrollment was 10.5 (8 to 15), 12 (8.5 to 21), and 9 days (7 to 14), respectively. There was no difference in major event rates between early and late enrollees (17% vs 17%, respectively, log-rank p = 0.98) or by diagnosis (15%, 16%, and 22% for CABG surgery, HVS, and MI, respectively; log-rank p = 0.50). Sternal instability and wound infection rates were similar. CR-related adverse events trended toward increased event rates in surgical and early enrollees, but of 44 events, only 3 were exercise related, none resulted in permanent harm, and 41 (93%) were managed in CR without need for emergency services. In conclusion, it appears that a policy of encouraging early enrollment into CR in patients with a recent open heart surgery seems unlikely to harm patients when careful individualized assessment and exercise prescription take place within the bounds of an established CR program.
Highlights
- •
Early enrollment into cardiac rehabilitation benefits medical patients.
- •
Safety concerns persist about early enrollment in surgical patients.
- •
Surgical patients enrolled at a median of 10 days after hospital discharge.
- •
Major adverse event rates were similar between groups.
- •
Routine early enrollment is encouraged for postoperative patients.
To address potential safety concerns of early cardiac rehabilitation (CR) participation, we measured the safety of early enrollment (<2 weeks) in patients with coronary artery bypass graft (CABG) surgery or heart valve surgery (HVS). We hypothesized that early CR enrollment would not be associated with increased adverse clinical events.
Methods
We evaluated all patients who underwent either CABG surgery or HVS and enrolled in the Mayo Clinic CR program in Rochester, Minnesota, from May 2009 to August 2012, with follow-up through February 2013. We then limited our sample to residents of Olmsted County, Minnesota, who had been hospitalized at St Mary’s Hospital immediately preceding their enrollment into outpatient CR. All patients included had previously given permissions for retrospective chart review. This study was reviewed and approved by the Mayo Clinic Institutional Review Board.
Our program has a long-established policy of early enrollment for all patients, regardless of qualifying diagnosis, the need for stress testing, or referring physician permissions. Hospitalized patients are assessed by inpatient CR staff, and an appointment is made before hospital discharge, typically allowing just a few days between hospital discharge and CR orientation. At CR enrollment, all patients are assessed for safety and suitability for exercise training, with most patients beginning exercise on either the first or the second session of CR. After 4 to 6 sessions, patients usually begin high-intensity aerobic interval training.
Patients who underwent open heart surgery are prescribed lower extremity exercises such as walking or stationary cycling. Upper extremity exercises for patients who have undergone a sternotomy are restricted to general mobility, stretching, and upper extremity lifting of <10 pounds for 6 weeks after the operation. Patient symptoms or complaints are evaluated by CR nursing staff, where nurses can observe the patient, refer the patient to see their primary physician, refer to the daily supervising CR physician, or refer to the emergency department.
We evaluated all adverse events occurring within 6 months of hospital discharge through the individual review of medical and program records. All reviews were performed by 2 experienced CR staff members (KJD and KPR) using a standard template and definitions. Borderline events were discussed between staff members for internal consistency.
We defined 3 separate outcome categories: major clinical, surgical-related, and CR-related adverse events. We defined major adverse clinical events to be death, myocardial infarction (MI), repeat revascularization (percutaneous coronary intervention [PCI]), stroke, hospital readmission, placement of pacer device, repeat surgery, or sudden cardiac death. Surgical-related adverse events consisted of wound infection, sternal instability, or blood transfusion. Infection was divided into all-cause and deep sternal infections. All-cause infection was defined as infection of sternal or skin wounds or the use of an oral outpatient antibiotic, which could additionally include use for pneumonia and urinary tract infections. Deep sternal infections were defined as readmission to the hospital because of an infection that required surgical debridement or intravenous antibiotics.
We defined CR-related events as adverse signs or symptoms that required direct management during or immediately after a session of CR. These were premature stopping of CR exercise due to safety concerns, fall with exercise, presyncope with exercise, unplanned medication administration, unplanned electrocardiography, unplanned blood draw in CR, or emergency department evaluation after a CR exercise session. We then divided CR-related events by their association with exercise. This allowed us to distinguish between true exercise-related harm and adverse events that were detected in CR but were unrelated to exercise, such as obtaining an electrocardiogram for a patient presenting to CR with new-onset atrial fibrillation.
Our primary predictor variable was the interval between hospital discharge and the first session of CR. We divided each group (stratified by diagnosis) by median time to enrollment, with patients attending on or before the median date in the “early” group. The “late” group served as the first control group. To compare the adverse event rates of surgical patients to those reported in the medical literature, we also included a second control group of patients with MI who enrolled in CR during the same period. We believed patients with MI would be more closely comparable to patients who underwent CABG and HVS than other patient groups that attend CR. Given the time-intense nature of the chart review, we randomly selected 60 patients with MI to roughly match the number of “late” CABG patients.
Descriptive statistics were applied, with count and percentage used for categorical data, means and SDs for normally distributed continuous variables, and medians and interquartile ranges for skewed data. Groups were tested with chi-square, analysis of variance, Student t tests, and Wilcoxon ranked sum testing as appropriate.
Outcomes analysis was performed using time to first safety event and log-rank statistical testing. Raw adverse events were not tested for statistical significance because some patients experienced multiple adverse events (many of which were interrelated, such as having presyncope, obtaining an electrocardiogram, and stopping exercise early). In cases of multiple events in the same patient, only the first event within each category (major, surgical related, and CR related) was used. Log-rank statistical testing was used in all cases. Patients were censored at 6 months from hospital discharge if no clinical event had taken place.
Our primary outcome was time to first major event after enrollment in CR, and our secondary outcome was time to first CR-related event after enrollment in CR. As an additional secondary outcome, we also report time to major event after hospital discharge regardless of when a patient enrolled into CR. All statistical tests were 2 tailed, and significance was set at α = 0.05. We used JMP 9.0.2 (SAS Institute, Cary, North Carolina).
Results
We identified 111, 69, and 182 sequential residents of Olmsted County with CABG, HVS, and MI respectively, who enrolled in CR during the study period. Of the 60 randomly selected patients with MI, 1 patient initially labeled as MI was discovered to have been treated with CABG and was subsequently analyzed in that group. Patient characteristics and enrollment times are reported in Table 1 .
| Baseline Characteristic | Overall (N=240) | By Diagnosis | p-value by Diagnosis | By Enrollment Time | p-value by Enrollment Time | |||
|---|---|---|---|---|---|---|---|---|
| CABG (N=112) | Valve (N=69) | MI (N=59) | Early Enrollment (N=127) | Late Enrollment (N=113) | ||||
| Age (years ± SD) | 67 ± 12 | 68 ±10 | 65 ±14 | 71 ± 12 | 0.01 | 65 ± 12 | 70 ± 12 | 0.002 |
| Male | 173 (72%) | 101 (90%) | 34 (49%) | 38 (64%) | < 0.001 | 80 (71%) | 93 (73%) | 0.67 |
| Non-Hispanic white | 219 (91%) | 103 (92%) | 60 (92%) | 56 (92%) | 0.26 | 113 (89%) | 106 (94%) | 0.58 |
| Currently employed | 77 (32%) | 36 (32%) | 23 (33%) | 18 (31%) | 0.92 | 53 (42%) | 24 (22%) | 0.001 |
| Hypertension (n= 233) | 180 (77%) | 87 (78%) | 40 (65%) | 53 (90%) | 0.05 | 86 (70%) | 94 (85%) | 0.009 |
| Diabetes Mellitus (n = 233) | 81 (35%) | 46 (41%) | 19 (31%) | 16 (27%) | 0.14 | 36 (30%) | 45 (41%) | 0.08 |
| Former smoker (n = 94) | 45 (47%) | 24 (60%) | 7 (28%) | 14 (48%) | 0.16 | 26 (48%) | 19 (48%) | 0.71 |
| Current smoker (n = 94) | 7 (7%) | 2 (5%) | 3 (12%) | 2 (7%) | 0.16 | 3 (6%) | 4 (10%) | 0.71 |
| Body mass index (kg/m 2 ± SD) | 29 ± 6 | 30 ± 6 | 28 ± 5 | 29 ± 5 | 0.04 | 29 ± 5 | 29 ± 6 | 0.96 |
| Total sessions attended, n [IQR] | 23.5 [11, 35] | 28.5 [15, 35] | 19 [7, 35] | 18 [10, 35] | 0.04 | 24 [12, 35] | 22 [9, 35] | 0.52 |
| Time to enrollment (days [IQR]) | 10 [8,16] | 10 [8, 15] | 12 [8.5, 20] | 9 [7,14] | 0.02 | 8 [7, 9] | 16[13, 24] | <0.001 |
Raw adverse events are found in Table 2 . There were 85 major events in 65 patients (27%). One patient died (0.4%) from metastatic cancer 5 months after enrolling in CR following CABG.
| All | By Diagnosis | By Enrollment Time | ||||
|---|---|---|---|---|---|---|
| CABG | Valve | MI | Early Enrollment | Late Enrollment | ||
| Total Patients at Risk | 240 | 112 | 69 | 59 | 127 | 113 |
| MAJOR, TOTAL | 85 | 41 | 26 | 18 | 32 | 51 |
| •Death | 1 | 1 | 0 | 0 | 0 | 1 |
| •Myocardial infarction | 2 | 2 | 0 | 0 | 0 | 2 |
| •Repeat revascularization | 6 | 5 | 1 | 0 | 2 | 4 |
| •Stroke | 2 | 1 | 1 | 0 | 2 | 0 |
| •Hospital readmission | 66 | 26 | 23 | 17 | 24 | 40 |
| •Device placement | 5 | 3 | 1 | 1 | 3 | 2 |
| •Repeat surgery | 2 | 2 | 0 | 0 | 1 | 1 |
| •Sudden cardiac death | 1 | 1 | 0 | 0 | 0 | 1 |
| SURGICAL RELATED, TOTAL | 24 | 15 | 7 | – | 9 | 13 |
| •Infection, all cause | 16 | 10 | 6 | – | 7 | 9 |
| •Deep sternal infection | 3 | 2 | 1 | – | 1 | 2 |
| •Sternal instability | 6 | 5 | 1 | – | 2 | 4 |
| •Blood transfusion | 0 | 0 | 0 | – | 0 | 0 |
| CARDIAC REHAB RELATED – TOTAL | 44 | 29 | 8 | 5 | 30 | 15 |
| •Emergency room after CR | 3 | 1 | 0 | 2 | 2 | 1 |
| •Early stop of exercise session | 19 | 12 | 4 | 3 | 10 | 9 |
| •Fall | 0 | 0 | 0 | 0 | 0 | 0 |
| •Syncope/pre-syncope | 11 | 9 | 2 | 0 | 9 | 2 |
| •Unplanned electrocardiogram | 10 | 6 | 4 | 0 | 8 | 2 |
| •Unplanned medication use | 1 | 1 | 0 | 0 | 1 | 0 |
| •Unplanned blood draw | 0 | 0 | 0 | 0 | 0 | 0 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
