Predictors of Extended Postoperative Length of Stay After Uncomplicated Transcatheter Aortic Valve Replacement

Abstract: Background. The current study aims to identify predictors of extended postoperative length of stay (PLOS) after uncomplicated transcatheter aortic valve replacement (TAVR). Methods. Patients who underwent TAVR at a single center between June 2012 and June 2016 were analyzed. Patients were stratified by time into an early cohort (EC; 2012-2014) and current cohort (CC; 2015-2016). Those who had complications post procedure were excluded. The CC group was dichotomized based on its median PLOS. Factors associated with a longer PLOS were investigated by using multivariable logistic regression analysis. Results. Mean age of the 686 patients (299 in the EC group and 387 in the CC group) was 82 ± 8 years. PLOS in the CC group was significantly lower than in the EC group (4 days vs 6 days, respectively; P<.001). Median PLOS in the CC group was 2 days. Dichotomizing the CC group by median PLOS resulted in 148 patients (54%) ≤2 days vs 128 patients (46%) >2 days. Of these, PLOS was 1 day in 71 patients (26%) and 2 days in 61 patients (28%). Independent predictors of PLOS >2 days were non-transfemoral approach, non-elective admission, female sex, low mean transaortic gradient, presence of chronic renal failure, and pulmonary hypertension. Conclusion. Experience coupled with improvements in TAVR technology over the past few years have led to a significant decrease in PLOS after TAVR. In the current TAVR era, 1 out of every 2 patients stays for a day or two in the absence of perioperative adverse events.

J INVASIVE CARDIOL 2019;31(5):153-158.

Key words: low-risk TAVR, pulmonary hypertension, renal failure

Over the past decade, transcatheter aortic valve replacement (TAVR) has successfully evolved as an intervention of choice among severe aortic stenosis (AS) patients deemed inoperable and as a viable alternative to surgery in high operative risk patients.¹ Compared to patients treated with surgical aortic valve replacement (SAVR), TAVR patients reported better mortality and morbidity benefits in addition to significantly shorter length of hospital stay.² Length of hospitalization after TAVR has observed a progressive decrease over the years and several factors have been associated with this change.

These factors include: (1) the widespread adoption of the procedure leading to increasing operator experience; (2) advancements in TAVR technology resulting in reduction in catheter diameters, new-generation bioprosthetic valves, and the use of advanced imaging techniques for valve sizing; (3) a recent shift of TAVR indication to low-risk to intermediate-risk populations;^3-5 and (4) the motivation to reduce total hospital costs by shortening postoperative in-hospital stay.

In the current era of value-based health care, there is concerted demand to decrease postoperative length of stay among costly procedures like TAVR. Previously, higher TAVR costs were associated with the incremental cost of managing perioperative complications associated with an early learning phase.^6,7 Procedure-related complications accounted for an approximately 25% increase in non-implant related hospital costs and 2.4 days of in-hospital stay among patients enrolled in the early PARTNER (Placement of Aortic Transcatheter Valves) trials.⁷ Moreover, a significant number of patients are hospitalized for prolonged monitoring rather than valid clinical reasons. The increase in operator experience is expected to reduce perioperative morbidity rates, and this is likely to be reflected in a decreased postoperative length of stay (PLOS). New clinical algorithms developed based on patient-specific criteria that identify patients who may be suitable for early discharge can enhance early recovery and decrease wasteful use of medical resources.

The primary objectives of the present study are: (1) to investigate the clinical predictors of extended PLOS in uncomplicated TAVR cases; and (2) to describe the impact of evolving experience on PLOS at a high-volume TAVR center.

Methods

This study is a retrospective analysis of all patients who underwent TAVR for the treatment of severe AS at a single center between June 2012 and December 2016. Patients who fulfilled the criteria for AS per the American College of Cardiology/American Heart Association guidelines¹ were included in the study. Preoperative assessment was done by a heart team that included at least two cardiac surgeons and an interventional cardiologist. TAVR was proposed as a treatment option for patients who were deemed non-eligible for SAVR based on their overall surgical risk. The Society of Thoracic Surgeons (STS) risk score and frailty status of patients were two main elements used to risk stratify patients.

Frailty was assessed by using an index that represented the different domains of frailty previously described by Freid et al.⁸ Our method of assessing frailty has been previously described in detail.⁹ In brief, patients were assigned a score of 0 or 1 when deemed as non-frail or frail per the four domains of frailty. A combined score ranging from 0 to 4 that consisted of scores from frailty components was developed. Patients were classified as frail if their total frailty score was ≥3 out of 4 and non-frail if <3 out of 4.

Preoperative screening also included transesophageal echocardiography, cardiac catheterization, and thoracic and abdominal computer tomography. Patients with inadequate iliofemoral access based on imaging findings were recommended a non-transfemoral (TF) approach for TAVR. Among these patients, TAVR was performed via the transapical, transaortic, subclavian, and transaxillary routes. The final study population did not include patients who attempted TAVR but later had it converted to an open procedure.

Preoperative and postoperative clinical and echocardiographic patient data were collected and compared among patient groups. Baseline clinical and demographic and periprocedural data were also collected. Patient co-morbidities were obtained using the definitions provided by the STS data collection system. Perioperative morbidities were defined per criteria defined by the second version of the Valve Academic Research Consortium (VARC-2).¹⁰

Initially, the study population was divided into two groups according to treatment time. Since the TAVR program at our institution was started in June 2012, patients were stratified into an early cohort (EC: June 2012-December, 2014) and a current cohort (CC: January 2015-December 2016). PLOS was defined as the duration of stay after TAVR until discharge from the hospital. Demographics and baseline clinical characteristics were compared for both patient groups. The impacts of early experience and the evolving TAVR technology on PLOS were assessed by comparing PLOS for patient groups.

Procedure-related morbidity was investigated in all patients who were part of the CC group. Patients were excluded from the analysis if they experienced complications of any sort post procedure or died during the procedure. The distribution of PLOS was investigated among the final group and patients were dichotomized based on the median PLOS. PLOS was defined as extended if greater than the median PLOS for the final group. The primary outcome was extended PLOS and all factors associated with an extended PLOS were investigated using logistic regression models.

Statistical analysis. All continuous variables are reported as mean ± standard deviation and compared by using analysis of variance (ANOVA) or the Wilcoxon rank-sum test when noted. Categorical variables are reported as percentages and numbers and compared using Chi-square test or Fisher’s exact test. For logistic regression analysis, extended PLOS (PLOS >2 days) was chosen as the dependent variable. The following covariates were considered for selection: age; sex; body mass index; New York Heart Association (NYHA) class; frailty status; source of admission; previous history (hypertension, dyslipidemia, chronic lung disease, pulmonary hypertension, coronary artery disease, cerebrovascular accident, chronic renal failure, liver dysfunction, chronic atrial fibrillation, and permanent pacemaker placement); procedure approach; and echocardiographic parameters (left ventricular ejection fraction, mean aortic valve gradient, peak aortic valve gradient, moderate-severe mitral valve regurgitation, moderate to severe aortic valve regurgitation). Backward selection in a stepwise manner at a P entry of .25 and exit level of .10 was used to identify variables that were associated with an extended PLOS. Selected variables were placed in a multivariable logistic regression model to determine independent factors associated with an extended PLOS. Hosmer-Lemeshow Chi-square statistics were used to assess goodness of fit of the model. The beta estimates, their 95% confidence intervals (CI), and the standard error for each variable were reported. All statistical analyses were performed using JMP version 10.0 software (SAS Institute, Inc). A two-sided P-value <.05 was considered as statistically significant.

Results

Between June 2012 and December 2016, a total of 686 patients underwent TAVR for the treatment of AS. Table 1 shows the baseline demographics and clinical characteristics of all patients. Of the 686 patients who were studied, a total of 299 patients underwent TAVR between June 2012 and December 2014 and comprised the EC group, while 387 patients underwent TAVR from January 2015 to December 2016 and comprised the CC group. As expected, mean age and STS scores were significantly higher in the EC group than in the CC group (82.2 ± 7.7 years vs 80.9 ± 8.0 years [P=.04] and 8.6 ± 6.9% vs 6.9 ± 5.1% [P<.001], respectively). The EC group had more patients designated NYHA functional status III/IV than the CC group (P<.001). There were more elective cases in the CC group than in the EC group (76% vs 57%, respectively; P<.001). There were more TF cases in the CC group than in the EC group (86% vs 64%, respectively; P<.001). The incidences of comorbidities were similar in both groups, except for the history of chronic lung disease (CLD), wherein the EC group had a higher incidence of CLD than the CC group (34% vs 19%, respectively; P<.001). After excluding in-hospital deaths in both cohorts, PLOS was longer in the EC group than in the CC group (6 days vs 4 days, respectively; P<.001). In addition, the mean baseline aortic valve area in the EC group was significantly smaller than in the CC group (0.6 ± 0.2 cm² vs 0.7 ± 0.3 cm²; P=.02).

The CC group consisted of 387 patients. Of these, a total of 110 (28.4%) either had a complication or died before discharge and were therefore excluded from the final cohort. Excluded patients were analyzed separately, and their mean PLOS is given in Table 3. After exclusion, the median PLOS of the remaining group was found to be 2 days. One-hundred and twenty-eight patients stayed for >2 days while 148 patients were in the hospital for ≤2 days. The distribution of patients by PLOS after excluding patients with perioperative complications is illustrated in Figure 1.

Table 2 provides a comparison of baseline demographic and clinical characteristics of patients per duration of stay (≤2 days vs >2 days). Patients who stayed for ≤2 days had significantly lower STS scores, were non-frail, and had TAVR performed via TF approach. The PLOS >2 days group was more likely to have NYHA class III/IV functional status, and had a higher prevalence of pulmonary hypertension, liver dysfunction, and chronic renal failure. The number of patients with a previous history of balloon aortic valvuloplasty and percutaneous coronary intervention was also higher for patients who stayed >2 days. Moreover, the mean and peak aortic valve gradients were significantly lower for the PLOS >2 day group than in patients who stayed for fewer days.

A subanalysis of PLOS per complication type and frequency among patients in the CC group is provided in Table 3. Seven patients experienced respiratory failure perioperatively and their PLOS was extended by a mean duration of 17 days. The most common complication was conduction abnormalities, which occurred in 44 patients (40%). Their mean PLOS was 7 days, while patients who had a major bleeding stayed for 14 days on average. Overall, the mean PLOS of patients who had complications was 7 days, and this doubled for patients who experienced >2 perioperative complications.

Factors found to have significant univariable associations PLOS >2 days were procedure approach, STS score, NYHA class, source of admission, previous history of chronic renal failure, lung disease, pulmonary hypertension, mean and peak aortic valve gradients. On multivariable logistic regression analysis, in-patient status (P=.02), non-TF approach (P<.01), female sex (P=.02), history of chronic renal failure (P<.01), history of pulmonary hypertension (P=.01), and low mean aortic valve gradient (P<.01) were identified as independent predictors of a PLOS >2 days (Table 4).

Discussion

The current study presents an overview of a single-center experience with TAVR for the treatment of severe AS. After an initial period of learning coupled with an era of evolving technology, PLOS decreased significantly. During the post-learning era, more than half of uncomplicated TAVR cases stayed in the hospital for ≤2 days before discharge. Procedure-related comorbidities, however, extended the PLOS by almost 1 week. A non-TF TAVR approach, female sex, and non-elective cases were found to be associated with longer PLOS.

Evidence from recently published clinical trials and a meta-analysis have shown similar clinical outcomes in low-intermediate risk TAVR and SAVR patients; thus, the recent shift in TAVR indication to AS patients deemed to be of low or intermediate risk for SAVR.⁴ Findings from the present study confirm this evolving trend, as evidenced by the older age, higher STS risk scores, and smaller aortic valve areas seen in patients from the EC group. The significantly lower mean PLOS for the current cohort is an outcome of TAVR expansion to a more tolerable patient population. Factors such as younger age, lower surgical risk, and approach are known to result in a shorter PLOS.^4,13 Other factors include a higher procedural volume in the CC group than in the EC group. Operator experience in this sense contributes to the observed findings. There were also newer-generation TAVR valves (Sapien XT and Sapien S3; Edwards Lifesciences) used among patients in the CC group. Most EC cases were performed using the earlier-generation Sapien TAVR valve. The introduction of newer Sapien XT and Sapien S3 valves during the CC era explains the impact of evolving TAVR technology on PLOS.

The safety and feasibility of early discharge (≤3 days) post TAVR have been examined by previous studies.^11-13 In a report by Barbanti et al,¹¹ about 1 in 5 patients studied were discharged within 3 days post TAVR. Predictors of early discharge were year of procedure and pre-existing pacemaker, while baseline NYHA class IV and bleeding of any cause were associated with higher odds of discharge after 3 days. Similar findings were seen in data published from the nationwide inpatient database of 7321 TAVR cases in the United States between 2011-2012.¹³ The early discharge rate was 21%, and factors including year of procedure, TF approach, and prior balloon aortic valvuloplasty independently predicted early discharge. Female sex and in-hospital adverse effects were independently associated with late discharge. Expectedly, in-hospital adverse effects extended PLOS in the hospital after TAVR and a positive correlation between cost of hospitalization and TAVR PLOS was observed. The present study confirms the impact of procedure year on PLOS by the significantly lower PLOS in the CC group than in the EC group and a median 2-day PLOS after uncomplicated TAVR. In accordance with results from the nationwide sample, female sex and non-TF approach were associated with higher odds of longer PLOS after TAVR.¹³ In the current cohort, history of end-stage renal disease, pulmonary hypertension, lower mean aortic valve gradient, and non-elective procedures independently predicted extended duration of stay after uncomplicated TAVR.

Several observational studies have examined the incidence of perioperative adverse events and their impact on PLOS after TAVR.^7,14 In the PARTNER trial, a 49% incidence of periprocedural complications extended PLOS by 2.4 days on average. A German registry found significant associations between bleeding events, acute kidney injury, and strokes and longer PLOS after TAVR.¹⁴ In the current study, at least 1 complication extended PLOS by 5 days in the CC group; as with previous reports, major bleeding was associated with the longest stay after TAVR. Among patients who had >2 complications, PLOS was extended an average of 2 weeks.

A decrease in the number of uncomplicated TAVRs will lead to shortened hospital stays, thus ensuring optimal efficiency for this procedure. With the minimalist approach, Lauck et al¹⁵ proposed a risk-stratification model to optimize PLOS among TAVR patients. In their model, well-selected patients successfully underwent TAVR without general anesthesia and constant intraoperative monitoring with transesophageal echocardiography. Cases included in this study were performed in the hybrid operative room under general anesthesia. Lebehn et al¹⁶ recently examined the impact of conscious sedation alone on post-TAVR outcomes; they demonstrated that while conscious sedation results in significant TAVR cost savings, it has no significant effect on overall LOS after adjusting for other components of the minimalist strategy.

Study limitations. The limitations inherent to the present study include its retrospective nature and the lack of further cost analysis to objectively assess the impact of LOS on total TAVR costs. Subgroup analysis on patients who stayed longer after uncomplicated TAVR without tangible medical reasons would have been helpful in designing a risk-stratification model targeted at cost reduction. Since patients with adverse events of any type were excluded from the CC group, a general conclusion of PLOS in real-life scenarios cannot be made from these findings. The results are therefore limited to uncomplicated TAVR cases, but serve as a landmark to identify patients who may stay longer in-hospital without tangible medical reasons. 

Conclusion

Findings from the present study have shown that PLOS after TAVR has significantly decreased due to the expansion of procedure indication and advancements in technology. Better patient selection criteria will aid in achieving optimal efficiency for this evolving technology.

References

1. Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC guideline for the management of patients with valvular heart disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63:2438-2488.

2. Burrage M, Moore P, Cole C, et al. Transcatheter aortic valve replacement is associated with comparable clinical outcomes to open aortic valve surgery but with a reduced length of in-patient hospital stay: a systematic review and meta-analysis of randomised trials. Heart Lung Circ. 2017;26:285-295.

3. Arora S, Strassle PD, Ramm CJ, et al. Transcatheter versus surgical aortic valve replacement in patients with lower surgical risk scores: a systematic review and meta-analysis of early outcomes. Heart Lung Circ. 2017;26:840-845. Epub 2017 Jan 24.

4. Sardar P, Kundu A, Chatterjee S, et al. Transcatheter versus surgical aortic valve replacement in intermediate-risk patients: evidence from a meta-analysis. Catheter Cardiovasc Interv. 2017;90:504-515.

5. Arora S, Misenheimer JA, Jones W, et al. Transcatheter versus surgical aortic valve replacement in intermediate risk patients: a meta-analysis. Cardiovasc Diagn Ther. 2016;6:241-249.

6. Reynolds MR, Magnuson EA, Wang K, et al. Cost-effectiveness of transcatheter aortic valve replacement compared with standard care among inoperable patients with severe aortic stenosis: results from the placement of aortic transcatheter valves (PARTNER) trial (Cohort B). Circulation. 2012;125:1102-1109.

7. Arnold SV, Lei Y, Reynolds MR, et al. Costs of periprocedural complications in patients treated with transcatheter aortic valve replacement: results from the Placement of Aortic Transcatheter Valve trial. Circ Cardiovasc Interv. 2014;7:829-836.

8. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146-M156.

9. Chauhan D, Haik N, Merlo A, et al. Quantitative increase in frailty is associated with diminished survival after transcatheter aortic valve replacement. Am Heart J. 2016;182:146-154.

10. Kappetein AP, Head SJ, Genereux P, et al. Updated standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document (VARC-2). Eur J Cardiothorac Surg. 2012;42:S45-S60.

11. Barbanti M, Capranzano P, Ohno Y, et al. Early discharge after transfemoral transcatheter aortic valve implantation. Heart. 2015;101:1485-1490.

12. Durand E, Eltchaninoff H, Canville A, et al. Feasibility and safety of early discharge after transfemoral transcatheter aortic valve implantation with the Edwards SAPIEN-XT prosthesis. Am J Cardiol. 2015;115:1116-1122.

13. Mallikethi-Reddy S, Akintoye E, Telila T, et al. Transcatheter aortic valve implantation in the United States: predictors of early hospital discharge. J Interv Cardiol. 2017;30:149-155.

14. Kaier K, Reinecke H, Naci H, et al. The impact of post-procedural complications on reimbursement, length of stay and mechanical ventilation among patients undergoing transcatheter aortic valve implantation in Germany. Eur J Health Econ. 2018;19:223-228. Epub 2017 Feb 22.

15. Lauck SB, Wood DA, Achtem L, et al. Risk stratification and clinical pathways to optimize length of stay after transcatheter aortic valve replacement. Can J Cardiol. 2014;30:1583-1587.

16. Lebehn M, Hunter T, Clancy S, et al. Effects of anesthesia on cost and length of stay in patients undergoing transcatheter aortic valve replacement (TAVR) - results from a cohort study. J Am Coll Cardiol. 2016;68(18S):B278.

From the ¹Departments of Cardiothoracic Surgery and Cardiology, RWJ Barnabas Health-Newark Beth Israel Medical Center, Newark, New Jersey; and ²Department of Cardiology, Rush University Medical Center, Chicago, Illinois.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted November 3, 2018, and accepted December 26, 2018.

Address for correspondence: Alexis K. Okoh, MD, Cardiovascular Research Unit, RWJ Barnabas Health Heart Centers, Newark Beth Israel Medical Center, 201 Lyons Avenue, Suite G5, Newark, NJ 07112. Email: disciple951@gmail.com