As transcatheter aortic valve implantation (TAVI) becomes more routinely used, the recommended duration of monitoring after uncomplicated TAVI remains indeterminate. Retrospective analysis suggests that discharge within 72 hours is safe, but prospective data are largely lacking. We therefore prospectively assess the feasibility and safety of early discharge (within 72 hours) after transfemoral TAVI using Edwards SAPIEN-XT and SAPIEN-3 prostheses. Patients undergoing elective transfemoral TAVI were assessed prospectively for early discharge home. Feasibility and safety (death or repeat hospitalization within 30 days of discharge) of early discharge were assessed. Causes for failure of early discharge were assessed by prospective data collection and multivariate analysis. Of 130 patients, 76 (59%) were discharged early. Death or repeat hospitalization within 30 days occurred only in 4 cases (5%) among patients who discharged early: repeat hospitalization within 30 days was required in 3 early-discharge patients (4%), and there was a single death at 30 days. By multivariate analysis, factors associated with delayed discharge were blood transfusion (hazard ratio 13.85, 95% CI 1.61 to 119.40, p = 0.017) and pacemaker implantation (hazard ratio 4.47, 95% CI 1.34 to 14.26, p = 0.012). In conclusion, early discharge after elective transfemoral TAVI with SAPIEN-XT/SAPIEN-3 prostheses is safe and attainable in a large proportion of patients, with no evident compromise in safety. Factors associated with failure of early discharge are postprocedural blood transfusion and permanent pacemaker implantation.
Transcatheter aortic valve implantation (TAVI) has seen rapid adoption and evolution since the first-in-man TAVI was performed in 2002. Continued developments have included reductions in catheter diameters and routine use of preprocedural computed tomography for both valve sizing and assessment of peripheral vasculature, in addition to increasing institutional and operator experience. This has contributed to a growing popularity of the “minimalist” approach to transfemoral TAVI, which is feasible using local anesthesia and fluoroscopic guidance. As procedures become less invasive, the duration of monitoring required after elective transfemoral TAVI remains indeterminate. Early discharge (within 72 hours of TAVI) is an attractive option with multiple potential benefits including the well-being of frail or elderly patients (in whom prolonged unnecessary hospitalization may be detrimental), in addition to cost-saving benefits. Our center previously reported on the feasibility and safety of early discharge in a retrospective study, suggesting that discharge within 72 hours of uncomplicated transfemoral TAVI was safe and possible in 36% of patients. Independent predictors of early discharge included preexisting pacemaker and the absence of acute kidney injury. Prospective data, however, are largely lacking in this area. In our present study, we aimed to prospectively evaluate the safety and feasibility of early discharge after elective transfemoral TAVI using Edwards SAPIEN-XT or SAPIEN-3 valves. We also sought to identify specific factors associated with early discharge failure or success.
Methods
The study population consisted of all consecutive patients electively implanted with the Edwards SAPIEN-XT or SAPIEN-3 prosthesis through transfemoral approach from January 2014 to January 2015 at the Hospital Charles Nicolle, University Hospital of Rouen. The study was approved by the local ethics committee. Individual patient assessment for TAVI was performed by the local Heart Team based on current guidelines. Patients and their relatives were explicitly informed about the discharge and follow-up protocol before the index TAVI procedure, with an understanding that early discharge to home would be targeted. All patients provided written informed consent and were included in national TAVI registries.
Transfemoral TAVI was performed with local anesthesia as per techniques previously reported, and all patients were monitored for at least 24 hours after procedure in the intensive care unit. Daily electrocardiograms were recorded, and a transthoracic echocardiogram was obtained before discharge. Each patient was deemed a candidate for early discharge at the completion of their procedure, and daily visits were made by the primary operator to assess appropriateness for discharge. Criteria for early discharge eligibility were the absence of major periprocedural complications, conduction abnormalities, or social issues. A follow-up outpatient visit was routinely performed 1 month after procedure, including clinical evaluation, electrocardiogram, and transthoracic echocardiogram.
In patients exhibiting conduction disturbances post-TAVI, the decision to proceed with permanent pacemaker (PPM) implantation was guided by local protocol and practices. Patients with new persistent or recurrent high-grade atrioventricular (AV) block were generally implanted within 24 to 48 hours of TAVI. Patients with other new conduction disorders were monitored for a minimum of 24 to 48 hours before rendering decisions about pacemaker appropriateness.
Complications were classified using Valve Academic Research Consortium 2 criteria. Length of hospital stay was considered to be from the time of TAVI to discharge home. Safety end point was a composite of all-cause mortality or repeat hospitalization within 30 days. Other end points were the individual components of all-cause mortality at 30 days or repeat hospitalization within 30 days.
Qualitative variables are expressed as percentages and quantitative variables as mean ± SD. Patients were analyzed in 2 groups for comparison purposes: those discharged within 72 hours of TAVI versus those discharged >72 hours after procedure. The Student t test was used to compare numerical variables between groups. The chi-square or Fisher’s exact tests were used to compare quantitative variables. Causes for failure of early discharge were assessed by prospective data collection and multivariate analysis. Using variables with p values <0.1 in univariate analysis, a multivariate logistic regression model was used to explore factors associated with failed early discharge. Analyses were performed using SPSS, version 17.0 (SPSS Inc., Chicago, Illinois).
Results
Of 153 patients treated with transfemoral TAVI during the study period ( Figure 1 ), 4 patients were in cardiogenic shock immediately before undergoing TAVI, and 8 were already hospitalized for refractory congestive heart failure, these patients were excluded from further analysis. Similarly, 11 patients were further excluded for the following reasons: CoreValve device used (5 patients), procedural failure to deploy valve (3 patients), intraprocedural death (1 patient), nonresident of France with inability to follow-up (1 patient). Finally, 1 patient with a mechanical mitral valve was expected to require a longer hospitalization to ensure therapeutic international normalized ratio levels before discharge, this patient was excluded from the study. Ultimately, 130 patients undergoing elective transfemoral TAVI were prospectively included.
Baseline characteristics of the study population are presented in Table 1 . The average age of the study population was 84.7 years with a nearly equal proportion of men and women. The mean logistic EuroScore was 15%, mean left ventricular ejection fraction was 63%, and most patients were in New York Heart Association class II (33%) or III (61%) heart failure at the time of TAVI. There were no significant differences in baseline characteristics between the early and nonearly discharge group.
| Parameters | Total population (N=130) | Early discharge (N=76) | Delayed discharge (N=54) | p value |
|---|---|---|---|---|
| Age, mean (years) | 84.7 ± 5.4 | 84.4 ± 5.8 | 85.4 ± 4.8 | 0.28 |
| Men | 68 (52%) | 36 (47%) | 32 (59%) | 0.18 |
| Body mass index (kg/m 2 ) | 27.1 ± 4.4 | 27.1 ± 4.2 | 27.1 ± 4.8 | 0.98 |
| Diabetes mellitus | 37 (28%) | 23 (30%) | 14 (26%) | 0.59 |
| Hypertension | 107 (82%) | 65 (86%) | 42 (78%) | 0.25 |
| Dyslipidemia | 76 (59%) | 48 (63%) | 28 (52%) | 0.20 |
| Tobacco use | 35 (27%) | 20 (26%) | 15 (28%) | 0.85 |
| Atrial fibrillation | 48 (37%) | 24 (32%) | 24 (44%) | 0.13 |
| Peripheral vascular disease | 8 (6%) | 7 (9%) | 1 (2%) | 0.09 |
| History of myocardial infarction | 13 (10%) | 9 (12%) | 4 (7%) | 0.41 |
| Percutaneous coronary intervention | 25 (19.2%) | 15 (20%) | 10 (19%) | 0.86 |
| Permanent pacemaker | 19 (15%) | 12 (16%) | 7 (13%) | 0.65 |
| History of stroke | 7 (5%) | 3 (4%) | 4 (7%) | 0.39 |
| Previous balloon aortic valvuloplasty | 7 (5%) | 2 (3%) | 5 (9%) | 0.10 |
| Logistic EuroScore (%) | 15.3 ± 8.5 | 15.7 ± 8.8 | 14.7 ± 8.0 | 0.54 |
| New York Heart Association class ≥II | 128 (99%) | 75 (99%) | 53 (98%) | 0.19 |
| Creatinine clearance (ml/min) | 47.3 ± 19.4 | 49.7 ± 21.0 | 44.0 ± 16.5 | 0.10 |
| Left bundle branch block | 13 (10%) | 10 (13%) | 3 (6%) | 0.15 |
| Right bundle branch block | 13 (10%) | 7 (9%) | 6 (11%) | 0.72 |
| Left ventricular ejection fraction (%) | 63.0 ± 13.1 | 62.7 ± 13.9 | 63.4 ± 14.1 | 0.78 |
| Aortic valve gradient, mean (mmHg) | 47.1 ± 14.3 | 47.2 ± 14.4 | 47.0 ± 14.2 | 0.93 |
| Pulmonary artery systolic pressure (mmHg) | 37.7 ± 14.0 | 37.6 ± 14.2 | 37.9 ± 14.0 | 0.95 |
Procedural characteristics and complications are summarized in Table 2 . A higher overall proportion of SAPIEN-XT prostheses were used (62%) than those of SAPIEN-3 (37%). Major complications were mostly vascular (n = 12; 9%) and frequently requiring blood transfusion (n = 11; 9%). Significant aortic regurgitation (≥grade II) occurred in 23 (18%) patients and PPM was required in 17 (13%) patients. Average time to discharge for all patients was 4.0 ± 2.7 days. A total of 76 patients (59%) were discharged within 72 hours: 55 patients (42% of the study population) within the first 48 hours, and 4 patients (3%) within 24 hours of TAVI. The distribution of length of stay after TAVI is shown in Figure 2 . Rates of early discharge were similar between SAPIEN-XT and SAPIEN-3 valves (58% vs 58%, p = 0.70).
| Parameters | Total population (N=130) | Early discharge (N=76) | Delayed discharge (N=54) | p value |
|---|---|---|---|---|
| Valve in valve | 6 (5%) | 4 (5%) | 2 (4%) | 0.6 |
| Contrast use (mL) | 163.4 ± 57.1 | 163.5 ± 53.0 | 163.3 ± 63.0 | 0.9 |
| Fluoroscopy time (minutes) | 19.2 ± 8.9 | 17.9 ± 6.8 | 20.1 ± 10.9 | 0.06 |
| Valve type, | 0.6 | |||
| Sapien 3 | 48 (37%) | 28 (37%) | 20 (38%) | |
| Sapien XT | 81 (62%) | 47 (62%) | 34 (63%) | |
| Direct Flow | 1 (1%) | 0 | 1 (2%) | |
| Mean time to discharge (days) | 4.0 ± 2.7 | 2.2 ± 0.5 | 6.5 ± 2.6 | <0.0001 |
| Vascular complications | ||||
| Minor | 25 (19%) | 15 (20%) | 10 (19%) | 0.86 |
| Major | 12 (9%) | 1 (1%) | 11 (20%) | <0.0001 |
| Vascular stent | 16 (12%) | 8 (11%) | 8 (15%) | 0.46 |
| Bleeding | ||||
| Minor | 11 (9%) | 6 (8%) | 5 (9%) | 0.78 |
| Major | 10 (8%) | 1 (1%) | 9 (17%) | 0.001 |
| Life-threatening | 3 (2%) | 0 | 3 (6%) | 0.07 |
| Blood transfusion | 11 (9%) | 1 (1%) | 10 (9%) | 0.001 |
| Neurologic complications | ||||
| Transient ischemic attack | 0 | 0 | 0 | 1 |
| Minor stroke | 0 | 0 | 0 | 1 |
| Major stroke | 1 (1%) | 0 | 1 (2%) | 0.42 |
| Myocardial infarction | 1 (1%) | 0 | 1 (2%) | 0.42 |
| Acute kidney injury | 0.74 | |||
| Stage 1 | 3 (2%) | 2 (3%) | 1 (2%) | |
| Stage 2 | 1 (1%) | 0 | 1 (2%) | |
| Stage 3 or 4 | 0 | 0 | 0 | |
| Permanent pacemaker | 17 (13%) | 5 (7%) | 11 (20%) | 0.009 |
| Aortic regurgitation, grade ≥II/IV | 23 (18%) | 14 (18%) | 9 (17%) | 0.80 |
| Decline in hemoglobin (g/dL) | 1.6 ± 1.0 | 1.2 ± 0.7 | 2.2 ± 1.0 | <0.0001 |
| Increase in creatinine (μmol/L) | 8.3 ± 40.2 | 1.8 ± 30.8 | 17.4 ± 49.3 | 0.029 |
| Troponin T peak (ng/L) | 194.3 ± 203.0 | 164.6 ± 122.4 | 236.0 ± 275.7 | 0.048 |
The principal reasons for failure of early discharge were electrical conduction disturbances (n = 23; 18%) and major vascular complications (n = 12; 9%). The conduction disturbances observed during or immediately after TAVI included: high-grade AV block (12 patients), new complete left bundle branch block (7 patients), new left bundle branch block and first-degree AV block (1 patient), and sinus node dysfunction (2 patients). The final patient had a nonspecified conduction disturbance. Of the 12 patients with new high-grade AV block, 8 received PPM with an average time to implantation of 1.9 days (range 0.5 to 5 days). The remaining 4 patients with high-grade AV block experienced resolution of their conduction disorder and did not receive a PPM before discharge. Only 2 patients with new left bundle branch block (LBBB) underwent PPM implantation: a patient with new LBBB and new first-degree AV block received a PPM at day 3 after TAVI, and another patient with concomitant rapid atrial fibrillation and new LBBB underwent PPM implantation and AV nodal ablation at day 6 after TAVI.
In 6 patients (5%), social or disposition issues were responsible for delayed discharge. Other causes of delayed discharge were congestive heart failure (3 patients; 2%), acute renal insufficiency (2 patients; 2%), myocardial infarction (1 patient; 1%), cerebral vascular accident (1 patient; 1%), valve embolization into the left ventricle (1 patient; 1%) and tamponade requiring percutaneous drainage (1 patient; 1%). Febrile episodes, delirium, or drug-induced skin reaction accounted for the remaining 6 (5%) patients.
Factors associated with early discharge were assessed by univariate and multivariate analysis ( Table 3 ). Six parameters with p <0.1 by univariate analysis were included in the multivariate analysis: new PPM, blood transfusion, creatinine clearance, change in creatinine, fluoroscopy time, and peak troponin. Of these, blood transfusion and implantation of a pacemaker were the only factors associated with delayed discharge (hazard ratios 13.85 and 4.47, respectively). There were no predictors for successful early discharge discernable.
