Cardiac resynchronization therapy (CRT) and implantable cardioverter defibrillator (ICD) implantation improve morbidity and mortality in selected patients. Many centers still admit patients overnight. We evaluated the safety, feasibility, and cost savings of same-day CRT/ICD device implantation by performing a retrospective study of all consecutive elective CRT/ICD implants at a tertiary center from January 2009 to April 2013. All emergency and/or inpatient cases were excluded. Data were collected on baseline demographics, implantation indication, procedure details, complications (categorized as immediate [≤24 hours], short term [24 hours to 6 weeks], medium term [6 weeks to 4 months], and long term [>4 months]), and mortality (30 day and 1 year). Comparisons were made between those having planned same-day versus overnight stay procedures. A cost analysis was performed to evaluate cost savings of the same-day policy. A total of 491 devices were implanted during this period: 267 were elective (54 planned overnight, 213 planned same-day) of which 229 were CRT pacemakers or CRT defibrillators and 38 ICDs. There were 26 total overall complications (9.7%) with no significant differences between planned same-day versus planned overnight stay cohorts (9.4% vs 11.1%, p = 0.8) and specifically no differences in immediate, short-, medium-, and long-term complications at follow-up. The 30-day and 1-year mortality rates did not differ between the two groups. An overnight stay at our hospital costs $450 (£300); our cost saving during this period was $91,800 (£61,200). Same-day CRT/ICD implantation is safe, feasible, and associated with significant cost savings. It provides significant advantages for patients and health care providers, especially given the current financial climate.
Cardiac resynchronization therapy (CRT) and the implantable cardioverter defibrillator (ICD) are known to improve morbidity and mortality in selected patients. CRT has been shown to reduce mortality and hospitalization in selected patients with heart failure on optimal medical therapy. ICDs implanted for primary and secondary prevention have become the cornerstone in the prevention of sudden cardiac death in selected patients. Recent changes to international guidelines reflect the success of these devices and are now offered to a wider range of patients with implantation rates continuing to rise year on year. The cost-effectiveness of such procedures has been an important issue of discussion. Many centers in Europe and North America keep patients overnight after implantation, driven mainly by the assumed risk of device-related complications. With increasing health care costs and the current worldwide financial climate, the cost–benefit ratio of such a strategy is now being reconsidered. We have previously reported that same-day bradycardia pacemaker implantation is safe and cost-effective with significant cost savings for health care providers. The aim of the present study was to evaluate the safety and feasibility of this same-day policy by comparing outcomes with those routinely kept overnight; we also performed a cost-saving analysis for the period of study.
Methods
A retrospective study of all consecutive elective CRT pacemaker (CRTp), CRT defibrillator (CRTd), or ICD (single/dual chamber) implants performed at University Hospital Coventry, UK, from January 2009 to April 2013. This included all de novo and upgrade cases but excluded all having a pulse generator change only (as these patients are not usually admitted overnight). Elective implants were those seen in the outpatient clinic and subsequently admitted with a planned date. We excluded all who had a device implanted during an acute emergency hospital admission. Electronic patient records and case notes were reviewed for all cases. Data were collected on patient demographics, implant indication, procedure details, and outcomes. Patients were divided into 2 cohorts: planned same-day implants (following commencement of our same-day policy in 2010) and those planned to be kept overnight (patients implanted before same-day policy). Patients were seen routinely at 6 weeks and 4 months after the implant. Approval for the study was provided by our Local Audit and Research Department. The study applied the principles of the declaration of Helsinki. Complications were categorized overall and by time of occurrence: immediate (≤24 hours), short term (24 hours to 6 weeks), medium term (6 weeks to 4 months), and long term (>4 months, median 30 months, interquartile range 20 to 42 months). Mortality was examined at 30 days and 1 year after the procedure and classified as due directly to the device implant, cardiac cause (heart failure, myocardial infarction, arrhythmia, sudden cardiac death), and non-cardiac causes. Types of complication recorded were cardiac arrest, cardiac chamber perforation, tricuspid valve injury, pneumothorax or hemothorax, stroke, myocardial infarction, pericardial effusion or tamponade, major bleeding (defined as decrease in hemoglobin ≥20 g/l and/or blood loss requiring blood transfusion), wound hematoma (with or without reintervention), coronary sinus (CS) dissection or perforation (for CRT implants), procedure- or device-related infections, lead displacement, phrenic nerve stimulation postoperatively for CRT cases, and device erosion. Failed procedures due to patient factors (such as poor CS anatomy, poor patient tolerance) were also recorded. Both upgrade and de novo device implants were evaluated because of the defined increase in complication rates reported in upgraded devices.
All implants during the period of study were performed by a consultant cardiologist with specialist interest in EP/Devices alone or scrubbed with a Devices Fellow. Patients on oral anticoagulation were asked to discontinue it 3 days before operation and resume it the same day after operation. Conscious sedation (mostly midazolam 0.1 mg/kg intravenously) and local anesthesia (1% lidocaine) was used in all patients with a left deltopectoral groove incision in most. Venous access was mainly cephalic and axillary veins with subclavian vein used only if the first 2 routes failed. The right ventricular lead for most patients was implanted at the right ventricular apex and right atrial lead at the right atrial appendage. Patients in permanent atrial fibrillation did not have an atrial lead implanted. For left ventricular (LV) lead implantation, the CS was cannulated, and angiography was performed in all with the most lateral and basal/midcavity position favored; redundant leads were capped and left in situ. The pulse generator was placed into a prepectoral pocket. Defibrillation safety margin testing was not performed. All were given preprocedural intravenous antibiotics using 1-g flucloxacillin and 1.5-mg/kg gentamicin (maximum dose 150 mg) and 3 days after procedural oral antibiotics (flucloxacillin 500 mg 4 times a day). Patients with penicillin allergy were given intravenous teicoplanin 600 mg preoperatively, and oral doxycycline 200 mg/day for 3 days postoperatively. All had postimplant chest x-ray to check lead position and exclude pneumothorax. Patients were taken back to the ward and kept overnight for observation before our same-day protocol. Following the same-day protocol, patients were taken to our cardiology day-case unit and observed for 3 to 4 hours before being discharged if all checks were satisfactory. If for any reason there were concerns about same-day discharge, patients were admitted overnight and discharged the following day. Those discharged were instructed not to undertake strenuous physical activity for 1 week after and advised to keep the wound dry for 3 days postoperatively. In case of problems, patients were instructed to contact our day-case unit or cardiology ward immediately. Those living farther away were listed earlier than those living closer to allow timely discharge. Age and geography were not specifically used to exclude patients. The costing for an overnight stay in a cardiology bed was obtained from our finance department. We calculated actual cost savings that occurred during the period of study for those patients who were discharged home the same day only.
Statistical analysis was performed using Statistical Package for Social Sciences [SPSS], version 22.0 (IBM, Chicago, Illinois). Categorical variables were reported as numbers and percentages. Comparison analysis for categorical data was performed using the chi-square test. Where smaller than expected values were derived, Fisher’s exact test was performed. Continuous data underwent histogram plots and Kolmogorov–Smirnov tests of normality. Normally distributed data were reported as mean ± SD. Non-normally distributed data were reported as median ± interquartile range. Comparison analysis for independent normally distributed continuous data was performed using the unpaired t test. Independent continuous data that were not normally distributed underwent comparison analysis using the Mann–Whitney U test. A value of p < 0.05 was used for statistical significance.
Results
There were 491 patients who underwent a complex device procedure during the period of study ( Figure 1 ). Of these, 224 (53 CRTd, 40 CRTp, 131 ICD) were done during an acute hospital admission and were excluded from analysis leaving 267 elective procedures (114 CRTd, 115 CRTp, 38 ICD). Of these, 54 (20%) were planned overnight and 213 (80%) planned same-day cases. Baseline characteristics of the 2 cohorts are summarized in Table 1 with no significant differences noted. In the same-day cohort, there were 9 (4.2%) unplanned overnight hospital admissions: 8 CRT (CRTp = 5, CRTd = 1, CRTp upgrade = 2) and 1 single chamber ICD. Three were kept in because of a complication (pneumothorax, wound hematoma); the remaining were kept in at the operators discretion (late finish = 2, multiple co-morbidity/social reasons = 2, generally unwell after procedure with no specific complication = 2). These patients were analyzed as part of the same-day cohort when comparing groups; however, they were excluded from analysis of actual cost savings as they were admitted overnight.
| Characteristics | Total Cohort (n = 267) | Same-day (n = 213) | Overnight (n= 54) | P Value |
|---|---|---|---|---|
| Age (yrs) median ± SD | 73.0±15.0 | 73.0-±15.0 | 73.5±13.0 | 1.0 |
| Men | 196 (73%) | 157 (74%) | 39 (72%) | 0.8 |
| Device Type: | ||||
| CRT-D | 114 (43%) | 91 (43%) | 23 (43%) | |
| CRT-P | 115 (43%) | 89 (42%) | 26 (48%) | 0.3 |
| ICD | 38 (14%) | 33 (16%) | 5 (9%) | |
| NYHA class | ||||
| 1 | 23 (9%) | 19 (9%) | 4 (8%) | |
| 2 | 21 (8%) | 17 (8%) | 4 (8%) | 0.4 |
| 3 | 199 (78%) | 159 (78%) | 40 (76%) | |
| 4 | 13(5%) | 8 (4%) | 5 (9%) | |
| Ischemic etiology | 146 (56%) | 115 (56%) | 31 (57%) | 0.9 |
| Non-Ischemic etiology | 105 (42%) | 83 (43%) | 22 (42%) | 0.9 |
| Atrial fibrillation/flutter | 70 (26%) | 56 (26%) | 14 (26%) | 0.96 |
| Previous MI | 117 (47%) | 92 (47%) | 25 (48%) | 0.9 |
| CABG | 53 (21%) | 43 (22%) | 10 (19%) | 0.7 |
| Hypertension | 113 (42%) | 92 (49%) | 21 (47%) | 0.8 |
| Diabetes Mellitus | 57 (23%) | 49 (25%) | 8 (16%) | 0.2 |
| Chronic Kidney Disease | 76.0 (34%) | 57 (33%) | 19 (37%) | 0.6 |
| Electrocardiogram: ∏ | ||||
| QRS Duration | ||||
| <120msec | 35 (14%) | 28 (141%) | 7 (14%) | |
| 120-149msec | 48 (20%) | 42 (22%) | 6 (12%) | |
| ≥150msec | 161 (66%) | 125 (78%) | 36 (22%) | 0.3 |
| LBBB morphology | 144 (54%) | 111 (52%) | 33 (61%) | 0.4 |
| Echocardiography: ∏ | ||||
| LVEF ≤35% | 249 (94%) | 195 (92%) | 54 (100%) | 0.2 |
| Medication: ∏ | ||||
| Warfarin | 87 (38%) | 66 (37%) | 21 (43%) | 0.5 |
| Aspirin | 127 (56%) | 101 (57%) | 26 (53%) | 0.8 |
Successful device implantation (satisfactory lead and device implantation) occurred in 260 (97.3%) of the entire cohort. The 7 failed implants were all CRT (CRTp = 2, CRTp upgrade = 3, CRTd = 1, CRTd upgrade = 1) with reason for failure in all being inability to securely implant an LV lead through the CS. Reasons were CS occlusion (n = 2), inability to cannulate the CS because of difficult anatomy (n = 2), and no usable CS branches to place the LV lead (n = 3). Table 2 summarizes the complications rates of the 2 cohorts. There were no significant differences in the overall complication rate between the same-day versus overnight stay cohorts and no differences at any prespecified time points. The most common complication was LV lead displacement (n = 10): 3 successfully repositioned percutaneously and 7 had successful epicardial LV leads through a surgical minithoracotomy. Two developed diaphragmatic pacing immediately after operation because of LV lead microdisplacement: 1 had successful LV lead reprogramming, the other needed LV lead repositioning. One patient developed a small pneumothorax immediately after operation, which was managed conservatively and discharged home the next day. There were 4 immediate bleeding complications: 3 wound hematomas and 1 wound oozing; 3 of these patients were on warfarin (stopped preoperatively) and 1 on aspirin (not stopped preoperatively). All were treated conservatively with pressure dressings; no reintervention or blood transfusions were required. Two LV leads failed to capture at short-term follow-up (no macrodisplacement on chest x-ray): 1 had reprogramming of threshold and the other required LV lead repositioning. Two patients developed wound infections (1 short-term and 1 medium-term follow-up). Both were treated successfully with antibiotics and required no further intervention. Wound site pre-erosion occurred in 1 patient at short-term follow-up and required pulse generator reburial. There were no differences in complications between the 2 cohorts. Three patients (1.1%) died at ≤30 days of device implantation; 2 from cardiovascular causes unrelated to the procedure (1 acute myocardial infarction, 1 heart failure). There were 24 patients (9.0%) who had died 1 year after device implantation; 12 from cardiovascular causes: 8 heart failure, 2 sudden cardiac death, and 2 acute myocardial infarction. No deaths were device implantation related. Two deaths had no cause identified, from primary or secondary care medical records. A cause related to device implantation was ruled out. All mortalities ≤1 year after implant occurred in CRT implants only. There were no significant differences in 30-day or 1-year mortality rates between the 2 cohorts ( Table 2 ).
| Outcomes | Total Cohort (n=267) | Same-day (n = 213) | Overnight (n=54) | P Value |
|---|---|---|---|---|
| Failed Procedure | 7 (2.6%) | 7 (3.3%) | 0 | 0.4 |
| Unplanned overnight stay | 9 (4.3%) | |||
| Total Complications | 26 (9.7%) | 20 (9.4%) | 6 (11.1%) | 0.8 |
| Immediate (≤24hrs) | 9 (3.4%) | 7 (3.3%) | 2 (3.7%) | 1.0 |
| RV Lead Displacement (n) | 2 | 1 | 1 | |
| Diaphragmatic Stimulation (n) | 2 | 2 | 0 | |
| Pneumothorax (n) | 1 | 1 | 0 | |
| Hematoma (n) | 3 | 2 | 1 | |
| Wound bleeding (n) | 1 | 1 | 0 | |
| Short term (>24hrs-6 wks ) | 6 (2.2%) | 4 (1.9%) | 2 (3.7%) | 0.4 |
| LV Lead displacement (n) | 2 | 1 | 1 | |
| RA Lead displacement (n) | 1 | 0 | 1 | |
| LV Lead not capturing (n) | 1 | 1 | 0 | |
| Wound infection (n) | 1 | 1 | 0 | |
| Pre-erosion (n) | 1 | 1 | 0 | |
| Medium term (6 wks-4 mths) | 4 (1.5%) | 4 (1.9%) | 0 (0.0%) | 0.6 |
| LV Lead Displacement (n) | 2 | 2 | 0 | |
| RV Lead Displacement (n) | 1 | 1 | 0 | |
| Device Infection (n) | 1 | 1 | 0 | |
| Long term (>4mths) | 7 (2.6%) | 5 (2.3%) | 2 (3.7%) | 0.6 |
| LV Lead Displacement (n) | 6 | 5 | 1 | |
| LV Lead Not Capturing (n) | 1 | 0 | 1 | |
| Mortality: | ||||
| Mortality ≤ 30 days | 3 (1.1%) | 2 (0.9%) | 1 (1.9%) | 1.0 |
| Mortality ≤1 years | 24 (9.0%) | 17 (8.0%) | 7 (13.0%) | 0.5 |
There were no differences in baseline characteristics, implant data, complications, or mortality between the same-day versus overnight stay patients undergoing CRT implantation ( Table 3 ). Vascular access for the right atrial and ventricular leads was mostly cephalic vein (70%), for LV lead was predominantly axillary vein (55%) with the remainder through the cephalic (23%) or subclavian (21%) veins. There were no differences in the axial LV lead position at the midcavity level in both cohorts (79.4% vs 69.4%, p = 0.6). There were 50 (19%) cardiac device upgrades, all of which were CRT upgrades. The original device was either a dual-chamber pacemaker (n = 25, 50%), single-chamber pacemaker (n = 6, 12%), dual-chamber ICD (n = 4, 8%), or single-chamber ICD (n = 15, 30%). There were more failed procedures in the upgraded group compared with de novo implant group (8% vs 1.2%, p = 0.04). However, there were no significant differences in complications between the 2 groups. There was a trend toward higher 1-year mortality rate in the upgrade group compared with de novo group (18% vs 9%, p = 0.07).
