By Michael Crawford, MD, Editor

SYNOPSIS: An analysis of the Nationwide Readmission Database revealed one-fifth of transcatheter aortic valve replacement patients are readmitted a median of 31 days after discharge. Medical comorbidities are the most common reason.

SOURCE: Tripathi B, Nerusu LA, Sawant AC, et al. Transcatheter aortic valve implantation readmissions in the current era (from the National Readmission Database). Am J Cardiol 2020;130:115-122.

The rate of transcatheter aortic valve replacement (TAVR) is increasing exponentially in the United States. Hospitals are concerned Medicare will start scrutinizing readmissions following this procedure. This report derived from the Healthcare Cost and Utilization Project (HCUP) Nationwide Readmission Database (NRD) from 2016-2017 is of interest.

Time to readmission was calculated by subtracting the length of stay of the index TAVR admission from the interval between two admissions. From 73,784 TAVR admissions, 16,343 were readmitted within 90 days, which is the endpoint of this study. The secondary endpoints were the predictors of readmission, etiology of readmission, and in-hospital outcomes by using ICD-10 codes for the primary readmission diagnosis. Comorbidities were determined in a similar fashion, and NRD variables were used to identify demographic characteristics of the patients. Hospitalization costs were derived from HCUP data adjusted for inflation. Most TAVR patients were octogenarians (61%) and men (55%). About 44% were discharged within 48 hours, and only 31% stayed longer than five days. However, these patients recorded a higher readmission rate (47% vs. 33%; P < 0.001).

The median time to readmission was 31 days. Congestive heart failure was the most common reason for readmission (77%). Readmitted patients exhibited a higher prevalence of medical comorbidities, were more likely to have been discharged to a skilled nursing facility (SNF) compared to non-readmitted patients (22% vs. 12%; HR, 1.58; P < 0.001), and record a length of stay longer than two days (67% vs. 53%; HR, 1.4; P < 0.001). Non-cardiac conditions accounted for 64% of readmissions, most commonly infections (13%), GI complications (7%), neurological complications (6%), and pulmonary complications (6%).

Multivariate predictors of 90-day readmission included transapical approach (HR, 1.19), age older than 90 years (HR, 1.22), diabetes (HR, 1.15), heart failure (HR, 1.17), atrial fibrillation/flutter (HR, 1.39), prior stroke (HR, 1.15), prior pacemaker/defibrillator (HR, 1.09), anemia (HR, 1.13), chronic obstructive pulmonary disease (HR, 1.26), chronic kidney disease (HR, 1.33), liver disease (HR, 1.24), acute kidney injury (HR, 1.2), and major bleeding (HR, 1.16). The cost of index hospitalization for those readmitted was higher than those not readmitted ($57,066 vs. $52,204; P < 0.001). The authors concluded that one out of five TAVR patients is readmitted within 90 days, mostly for non-cardiac causes.

COMMENTARY

The Hospital Readmissions Reduction Program (HRRP) has been using 30-day readmission rates for certain diagnoses as a performance measure, penalizing hospitals with higher rates. Also, Medicare’s bundled payment care initiative makes hospitals assume care for 90 days for certain diagnoses, and penalizes hospitals with high costs, which include the cost of readmissions within 90 days. Cardiac conditions targeted by HRRP (e.g., myocardial infarction and percutaneous coronary interventions) have exhibited a decrease in readmission rates and costs. In August 2020, voluntary outcome reporting for TAVR started, which includes volume of cases as well as risk-adjusted hospital and 30-day mortality. Bundled payments for TAVR are sure to follow. Currently, there are more than 25,000 TAVRs performed in more than 400 centers in the United States. As more low-risk surgical patients undergo TAVR, this number is likely to grow exponentially. Thus, this analysis of TAVR readmissions is of interest because it sheds light on the magnitude and potential causes of TAVR readmissions. Also, the NRD database is relevant because it represents 58% of hospitals in the United States, includes all payors, and covers 36 million discharges.

The Tripathi et al study demonstrated that one-fifth of TAVR patients are readmitted within a median of 31 days, and noncardiac conditions account for nearly two-thirds of the readmissions. The strongest predictors of readmission are discharge to a SNF (HR, 1.58) and length of stay of the index hospitalization (HR, 1.4). Also, readmitted patients carried a higher burden of comorbidities and were more likely to experience a serious complication, which prolonged hospital stay and resulted in SNF placement. The top five diagnoses for readmission were infections, along with GI, neurologic, pulmonary, and bleeding complications. Efforts to reduce readmissions will need to use a multidisciplinary team approach that includes internists to manage these patients after discharge. Also, judicious patient selection may be necessary to shorten lengths of stay and prevent patients from likely readmission for their comorbid conditions. In addition, patients with appropriate anatomy for a successful procedure and meticulous procedural technique will be required to reduce complications. Hail Mary TAVRs likely will be discouraged.

Because this work was based on administrative data, that represents a major limitation. There were no patient-level clinical data or data on socioeconomic status, education levels, or race. These variables likely would affect readmission rates and would need to be considered in planning readmission prevention efforts. Also, death outside the hospital, which would be a competitor to readmissions, is unknown. Finally, the NRD includes only 21 states, so the results may not apply to all U.S. locales. Very elderly, frail patients with multiple comorbidities and TAVR procedural complications are at the highest risk for readmission. Future Medicare payment programs may necessitate a more selective approach to such patients.