Abstract
Background
Frailty is associated with a greater risk of readmission after cardiovascular procedures. However, the impact of frailty on readmission rates and outcomes after Impella mechanical circulatory support (MCS) remains unknown. We aimed to explore the impact of frailty on readmission outcomes in patients who received Impella MCS.
Methods
Using the National Readmissions Database, patients aged 65 years and older who received Impella MCS between January 2016 and December 2020 were identified. Frailty was determined by the Hospital Frailty Risk Score (HFRS), which stratifies patients into 3 frailty risk categories as low (<5), intermediate (5-15), and high (>15), with intermediate- and high-risk groups defined as frail. The impact of frailty on short-term (within 30 days) and midterm (31-180 days) readmission rates and in-hospital outcomes was assessed.
Results
Of the 16,289 patients identified in the 30-day cohort, 8647 (53.1%) were identified as frail (HFRS ≥5) and 2185 (13.4%) had an unplanned readmission at 30 days. After adjusting for age, sex and comorbidities, frailty status (HFRS ≥5) was associated with a greater risk of 30-day readmission (odds ratio [OR] 1.27, 95% confidence interval [CI] 1.17-1.37), death (OR 2.0, 95% CI 1.22-3.30), major adverse events (OR 1.73, 95% CI 1.29-2.33), length of stay >4 days (OR 1.80, 95% CI 1.44-2.26) and greater hospitalization expenditures (OR 1.44, 95% CI 1.17-1.80) during readmission. Of the 6497 patients identified in the 31-180-day cohort, 3521 (54.2%) were considered frail and 1809 (27.8%) experienced unplanned readmissions. An HFRS ≥5 was associated with a greater risk of readmission (OR 2.10, 95% CI 1.88-2.34), in-hospital death (OR 3.02, 95% CI 1.33-6.86), length of stay >4 days (OR 1.66, 95% CI 1.29-2.14), and greater hospital expenditures (OR 1.36, 95% CI 1.05-1.75) during 31-180-day readmission.
Conclusions
Frailty is common among patients undergoing Impella MCS and is associated with higher rates of readmission and adverse outcomes during readmission.
Résumé
Contexte
La fragilité est associée à un risque plus élevé de réadmission après une intervention cardiovasculaire. Cependant, son effet sur les taux de réadmission et l’issue après l’implantation d’une pompe Impella en tant que dispositif d’assistance circulatoire mécanique (ACM) demeurent inconnus. Nous avons tâché de déterminer l’effet de la fragilité sur la réadmission chez les patients ayant reçu une pompe Impella en tant que dispositif d’ACM.
Méthodologie
Nous avons recensé dans la base de données NRD (Nationwide Readmissions Database) les patients âgés de 65 ans et plus chez qui une pompe Impella a été implantée entre janvier 2016 et décembre 2020. La fragilité a été évaluée en fonction du score HFRS (Hospital Frailty Risk Score), qui stratifie les patients en trois catégories de risque de fragilité (faible [< 5], intermédiaire [5 à 15] et élevé [> 15]), et les groupes à risque intermédiaire et élevé étaient considérés comme fragiles. L’effet de la fragilité sur les taux de réadmission à court terme (dans les 30 jours) et à moyen terme (dans les 31 à 180 jours) et les résultats pendant l’hospitalisation ont été évalués.
Résultats
Sur les 16 289 patients recensés dans la cohorte de 30 jours, 8 647 (53,1 %) ont été considérés comme fragiles (score HFRS≥5) et 2 185 (13,4 %) ont connu une réadmission imprévue à 30 jours. Après ajustement pour l’âge, le sexe et les affections concomitantes, la fragilité (score HFRS ≥ 5) a été associée à un risque plus élevé de réadmission dans les 30 jours (rapport de cotes [RC] : 1,27; intervalle de confiance [IC] à 95 % : 1,17-1,37), de décès (RC : 2,0; IC à 95 % : 1,22-3,30), d’effets indésirables majeurs (RC : 1,73; IC à 95 % : 1,29-2,33), de séjour d’une durée > 4 jours (RC : 1,80; IC à 95 % : 1,44-2,26) et de frais d’hospitalisation plus importants (RC : 1,44; IC à 95 % : 1,17-1,80) pendant la réadmission. Parmi les 6 497 patients réadmis dans les 31 à 180 jours, 3 521 patients (54,2 %) ont été considérés comme fragiles et 1 809 (27,8 %) ont connu une réadmission imprévue. Un score HFRS ≥ 5 a été associé à un risque plus élevé de réadmission (RC : 2,10; IC à 95 % : 1,88-2,34), de décès à l’hôpital (RC : 3,02; IC à 95 % : 1,33-6,86), de séjour d’une durée > 4 jours (RC : 1,66; IC à 95 % : 1,29-2,14) et de frais d’hospitalisation plus importants (RC : 1,36; IC à 95 % : 1,05-1,75) pendant la réadmission dans les 31 à 180 jours.
Conclusions
La fragilité est fréquente chez les patients chez qui une pompe Impella est implantée en tant que dispositif d’ACM. Elle est associée à des taux plus élevés de réadmission et de résultats défavorables pendant la réadmission.
The use of Impella (Abiomed, Danvers, MA) microaxial pump for mechanical circulatory support (MCS) in cases of cardiogenic shock and high-risk percutaneous coronary intervention (PCI) has grown exponentially.1 Although previous randomized data did not demonstrate a benefit of Impella over intra-aortic balloon pump (IABP), a more recent randomized trial has indicated that Impella may reduce all-cause mortality at 6 months compared with standard of care in patients with ST-elevation myocardial infarction (STEMI) and shock, but at the cost of significantly higher rates of complications such as bleeding, limb ischemia, renal failure, and sepsis.2
Frailty is a state of increased vulnerability to stressors as a result of cumulative decline in multiple physiological systems.3 The prevalence of frailty increases with age and is associated with worse clinical outcomes among patients with acute coronary syndrome and those undergoing PCI.4, 5, 6, 7, 8, 9, 10 Elderly patients have more complex coronary anatomy and represent a growing proportion of patients undergoing high-risk PCI.11, 12, 13
Frailty has also been associated with hospital readmission among several health conditions.14, 15, 16, 17 However, data assessing readmissions after MCS with Impella amid frail patients are scarce. In this analysis, we aimed to determine the impact of frailty on 30- and 180-day readmission rates and outcomes in patients who received MCS with an Impella device.
Methods
The authors declare that all supporting data are available within the article and its online supplementary files. The Western University Health Science Research Ethics Board has exempted the requirement for consent, as well as the necessity for approval from the Ethics Committee or Institutional Review Board for this study. This decision stems from the fact that all data used in this research originates from the National Readmissions Database (NRD), which is a publicly accessible and deidentified administrative database.
Data source
The NRD was used to identify patients who had undergone Impella MCS between January 2016 and December 2020. The NRD is a nationally representative, all-payer health care database, which is part of the Healthcare Cost and Utilization Project (HCUP). This database provides information on hospital readmissions representing close to 60% of the total US population and hospitalizations.
Study population
The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes 5A0221D (assistance with cardiac output using Impella pump, continuous) was used to identify patients aged 65 years and older who had received Impella MCS. Unique patient identifier codes were used to track patients across hospitals during a calendar year and to identify the initial (index) admission and subsequent readmissions.
Two different data sets were created for the unplanned readmissions analysis. For the 30-day readmission data set (Supplemental Fig. S1), patients were excluded if they were discharged in December to ensure there was 30 days of follow-up. For the 31-180-day data set (Supplemental Fig. S2), patients were excluded if they were discharged between July and December to allow for 6 months of follow-up. Additionally, patients were excluded if they died during their index admission, if they had missing data related to mortality, elective status, length of stay (LOS), or discharge month, or if they received concomitant extracorporeal membrane oxygenation (ECMO) or IABP. Patients were also excluded if they had a planned readmission. Patient and hospital demographics were collected from the NRD.
Each discharge record in the NRD includes up to 40 diagnoses, which were used to identify the comorbid conditions necessary to calculate the Charlson Comorbidity Index (CCI) and the Elixhauser Comorbidity Score (ECS).18, 19, 20, 21 The CCI includes 17 conditions, each assigned a specific weight, which are summed to determine the total CCI score (Supplemental Table S1). The ECS consists of 30 conditions, which were converted into a point system for this analysis (Supplemental Table S2).
The Hospital Frailty Risk Score (HFRS) was used to identify frail individuals. This score has been initially developed in the United Kingdom22 and thereafter validated to identify the prevalence and related outcomes of frail individuals undergoing transcatheter structural heart procedures and PCI.7,8,23, 24, 25 The HFRS score was calculated using ICD-10 diagnostic codes (Supplemental Table S3) that were present on discharge data gathered during the index hospitalization. Patients were then stratified into frailty risk categories, including low-risk (<5), intermediate-risk (5-15), and high-risk (>15). Patients in the intermediate- and high-risk categories were defined as frail,22 thereby used for the models.
Study outcomes and measurements
The primary outcome was unplanned readmission and hospital death during the readmission episode among frail vs nonfrail patients. Secondary outcomes included major adverse events (MAEs), LOS, and cost of hospital stay during readmission episodes. Unplanned readmission is defined as the proportion of nonelective hospital admissions for any cause within 30 days or 31-180 days after discharge from the index hospital stay. The occurrence of MAE during the readmission encounter was defined as the composite of cardiac complications (myocardial infarction [MI], cardiogenic shock, arrhythmia, and cardiac arrest), bleeding and vascular complications, and stroke. The components of MAE were identified using ICD-10-CM codes (Supplemental Table S4).26 The causes of readmission were considered by the first diagnosis of readmission, which was based on Clinical Classification Software codes (Supplemental Table S5).26 This manuscript conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines,27 and the STROBE checklist is provided in Supplemental Table S6.
Statistical analysis
Qualitative variables were presented as numbers and percentages, whereas quantitative variables were expressed as mean ± standard deviation or median (interquartile range [IQR]), depending on the variable distribution. For the comparison of continuous variables, the Kruskal-Wallis test was used. The χ2 test was used for categorical variables with adjustments made for survey sampling design. Adjusted P values for each variable were computed accounting for sampling discharge-level weights, cluster, and strata, as recommended by AHRQ for survey-specific analysis. Variables with fewer than 11 counts for individual discharge records were not detailed, in accordance with the HCUP data use agreement. For the purposes of the descriptive analysis, we provide data on 3 categories; however, based on the definition of frailty, for the primary and secondary outcome measures, patients were categorized as frail (HFRS score ≥5) or nonfrail (HFRS <5). Readmissions at the 30th day and between the 31st-180th days were considered a binary outcome. To determine the association between frailty status and the primary and secondary outcomes, we first conducted a bivariate analysis for each outcome with single variable, then included the variables with a P value <0.1 in a multilevel multivariable model, along with the variables that were judged, a priori, to be clinically meaningful such as, among others, age and sex, while omitting those included for the HFRS calculation. The results are presented as odds ratio (ORs) with 95% confidence interval (CI) and considered significant at a P value threshold of <0.05. Statistical analysis was performed using R version 4.3.2.
Results
Study population
A total of 46,600 discharges were identified, among these, 30,311 patients who died during their index admission, were discharged in the month of December, had missing data, or received ECMO or IABP were excluded. Of the remaining 16,004 patients, the median age of the study population was 75 (IQR 70-81) years and 29% were women. According to the HFRS, 7357 individuals (46%) were classified as low risk, 7163 individuals (44.8%) as intermediate risk, and 1484 individuals (9.3%) as high risk of frailty. Therefore, a total of 8647 individuals (53.1%) were categorized as frail based on the HFRS criteria. The remaining baseline characteristics and in-hospital outcomes of this cohort are presented in Table 1.
Table 1.
Baseline characteristics according to 30-day readmission
| Characteristic | All patients (n=16,004) | Hospital Frailty Risk Score |
Adjusted P value∗ |
||
|---|---|---|---|---|---|
| Low (n=7357) | Intermediate (n=7163) | High (n=1484) | |||
| Patient characteristics | |||||
| Age, y | 75 (70-81) | 75 (70-81) | 75 (70-81) | 74 (69-80) | 0.17 |
| Women | 4660 (29) | 1956 (27) | 2230 (31) | 474 (32) | <0.001 |
| Nonelective admission | 11,799 (74) | 4392 (60) | 6059 (85) | 1348 (91) | 0.001 |
| Weekend admission | 2821 (18) | 944 (13) | 1527 (21) | 350 (24) | <0.001 |
| Median zip code income† | |||||
| 76-100th percentile | 2894 (18) | 1244 (17) | 1317 (19) | 333 (23) | <0.001 |
| 51-75th percentile | 3869 (25) | 1737 (24) | 1779 (25) | 353 (24) | |
| 26-50th percentile | 4440 (28) | 2063 (29) | 1967 (28) | 410 (28) | |
| 0-25th percentile | 4542 (29) | 2194 (30) | 1982 (28) | 366 (25) | |
| Expected primary payer | |||||
| Medicare | 14,084 (88) | 6441 (88) | 6341 (89) | 1302 (88) | 0.23 |
| Medicaid | 234 (1.5) | 104 (1.4) | 113 (1.6) | 17 (1.1) | |
| Private | 1142 (7.1) | 572 (7.8) | 458 (6.4) | 112 (7.6) | |
| Others | 534 (3.3) | 236 (3.2) | 246 (3.4) | 52 (3.5) | |
| Comorbidities | |||||
| Smoking | 5119 (32) | 2163 (29) | 2484 (35) | 472 (32) | <0.001 |
| Dyslipidemia | 11,275 (70) | 5425 (74) | 4892 (68) | 958 (65) | <0.001 |
| Hypertension | 13,966 (87) | 6276 (85) | 6370 (89) | 1320 (89) | <0.001 |
| Diabetes mellitus | 6958 (43) | 2924 (40) | 3343 (47) | 691 (47) | <0.001 |
| Previous myocardial infarction | 3504 (22) | 1677 (23) | 1546 (22) | 281 (19) | 0.01 |
| Cardiac arrhythmias | 7107 (44) | 2732 (37) | 3539 (49) | 836 (56) | <0.001 |
| Previous PCI | 3019 (19) | 1477 (20) | 1276 (18) | 266 (18) | 0.006 |
| Previous CABG | 1982 (12) | 1135 (15) | 714 (10) | 133 (9.0) | <0.001 |
| Valvular disease | 4723 (30) | 1960 (27) | 2302 (32) | 461 (31) | <0.001 |
| Previous cerebrovascular disease | 2062 (13) | 721 (9.8) | 997 (14) | 344 (23) | <0.001 |
| Fluid and electrolyte disorders | 42 (0.3) | <11 (0.07) | 30 (0.4) | <11 (0.5) | <0.001 |
| Peripheral vascular disease | 3180 (20) | 1324 (18) | 1536 (21) | 320 (22) | <0.001 |
| Chronic pulmonary disease | 4261 (27) | 1678 (23) | 2137 (30) | 446 (30) | <0.001 |
| Renal disease | 5952 (37) | 1638 (22) | 3502 (49) | 812 (55) | <0.001 |
| Obesity | 2706 (17) | 1054 (14) | 1334 (19) | 318 (21) | <0.001 |
| Peptic ulcer disease | 111 (0.7) | 27 (0.4) | 62 (0.9) | 22 (1.5) | <0.001 |
| Rheumatic disease | 410 (2.6) | 164 (2.2) | 205 (2.9) | 41 (2.8) | 0.01 |
| Liver disease | 456 (2.8) | 155 (2.1) | 234 (3.3) | 67 (4.5) | <0.001 |
| Cancer | 938 (5.9) | 365 (5.0) | 477 (6.7) | 96 (6.5) | <0.001 |
| Anemia | 4790 (30) | 1282 (17) | 2825 (39) | 683 (46) | <0.001 |
| Charlson Comorbidity Index | 3.6 ± 2.4 | 2.8 ± 2.1 | 4.1 ± 2.4 | 4.8 ± 2.6 | <0.001 |
| 0 | 820 (5.1) | 632 (8.6) | 169 (2.4) | 19 (1.3) | <0.001 |
| 1 | 2524 (16) | 1549 (21) | 842 (12) | 133 (9.0) | |
| 2 | 2647 (17) | 1622 (22) | 900 (13) | 125 (8.4) | |
| ≥3 | 10,013 (63) | 3554 (48) | 5252 (73) | 1207 (81) | |
| Elixhauser Comorbidity Score | 12.8 ± 7.6 | 10.0 ± 6.8 | 14.6 ± 7.1 | 17.8 ± 8.2 | <0.001 |
| ≤0 | 1045 (6.5) | 846 (11) | 185 (2.6) | 14 (0.9) | <0.001 |
| 1-5 | 1427 (8.9) | 916 (12) | 445 (6.2) | 66 (4.4) | |
| 6-10 | 3682 (23) | 2153 (29) | 1341 (19) | 188 (13) | |
| ≥11 | 9850 (62) | 3442 (47) | 5192 (72) | 1216 (82) | |
| Hospital Frailty Risk Score | 6.8 ± 5.7 | 2.2 ± 1.5 | 9.0 ± 2.7 | 19.4 ± 4.2 | --- |
| Diagnosis on admission | |||||
| STEMI | 2121 (13) | 825 (11) | 1055 (15) | 241 (16) | <0.001 |
| NSTEMI | 4763 (30) | 1735 (24) | 2556 (36) | 472 (32) | <0.001 |
| Chronic coronary syndrome | 4998 (31) | 3406 (46) | 1426 (20) | 166 (11) | <0.001 |
| Malignant arrhythmias/cardiac arrest | 781 (4.9) | 349 (4.7) | 365 (5.1) | 67 (4.5) | 0.99 |
| Heart failure and cardiogenic shock | 1763 (11) | 554 (7.5) | 991 (14) | 218 (15) | <0.001 |
| Hospital characteristics | |||||
| Resident in same hospital state | 14,742 (92) | 6745 (92) | 6609 (92) | 1388 (94) | 0.11 |
| Teaching status of hospital | |||||
| Metropolitan teaching | 12,702 (79) | 5636 (77) | 5834 (81) | 1232 (83) | <0.001 |
| Metropolitan nonteaching | 2794 (17) | 1430 (19) | 1144 (16) | 220 (15) | |
| Nonmetropolitan hospital | 508 (3.2) | 291 (4.0) | 185 (2.6) | 32 (2.2) | |
| Hospital urban-rural designation | |||||
| Large metropolitan areas with at least 1 million residents | 8716 (54) | 3642 (50) | 4114 (57) | 960 (65) | <0.001 |
| Small metropolitan areas with <1 million residents | 6780 (42) | 3424 (47) | 2864 (40) | 492 (33) | |
| Micropolitan areas | 494 (3.1) | 287 (3.9) | 181 (2.5) | 26 (1.8) | |
| Not metropolitan or micropolitan (nonurban residual) | 14 (0.09) | <11 (0.05) | <11 (0.06) | <11 (0.4) | |
| Hospital bed size | |||||
| Small | 1434 (9.0) | 762 (10) | 570 (8.0) | 102 (6.9) | <0.001 |
| Medium | 4149 (26) | 1928 (26) | 1820 (25) | 401 (27) | |
| Large | 10,421 (65) | 4667 (63) | 4773 (67) | 981 (66) | |
| In-hospital procedures | |||||
| PCI | 12,726 (80) | 6143 (83) | 5513 (77) | 1070 (72) | <0.001 |
| CABG | 890 (5.6) | 263 (3.6) | 502 (7.0) | 125 (8.4) | <0.001 |
| Year of Impella procedure | |||||
| 2016 (January-November) | 1853 (12) | 903 (12) | 838 (12) | 112 (7.5) | 0.03 |
| 2017 (January-November) | 2704 (17) | 1239 (17) | 1213 (17) | 252 (17) | |
| 2018 (January-November) | 3643 (23) | 1702 (23) | 1591 (22) | 350 (24) | |
| 2019 (January-November) | 4151 (26) | 1878 (26) | 1878 (26) | 395 (27) | |
| 2020 (January-November) | 3653 (23) | 1635 (22) | 1643 (23) | 375 (25) | |
| In-hospital complications | |||||
| Bleeding complications | 934 (5.8) | 293 (4.0) | 534 (7.5) | 107 (7.2) | <0.001 |
| Vascular complications | 284 (1.8) | 99 (1.3) | 147 (2.1) | 38 (2.6) | 0.002 |
| Stroke | 362 (2.3) | 24 (0.3) | 198 (2.8) | 140 (9.4) | <0.001 |
| Length of stay, median, d | 8 (3-14) | 4 (2-7) | 11 (7-17) | 21 (14-31) | <0.001 |
| ≤8 d | 8727 (55) | 5916 (80) | 2688 (38) | 123 (8.3) | <0.001 |
| >8 d | 7277 (45) | 1441 (20) | 4475 (62) | 1361 (92) | |
| Index cost (USD)‡ | 64,620 (46,532-92,104) | 52,852 (39,734-70,713) | 74,259 (54,347-104,957) | 104,109 (74,452-148,685) | <0.001 |
| Discharge destination | |||||
| Home (self-care) | 7805 (49) | 5181 (70) | 2416 (34) | 208 (14) | <0.001 |
| Short-term hospital | 510 (3.2) | 170 (2.3) | 284 (4.0) | 56 (3.8) | |
| Transfer to other institution§ | 3903 (24) | 701 (9.5) | 2437 (34) | 765 (52) | |
| Home health care | 3715 (23) | 1280 (17) | 1990 (28) | 445 (30) | |
| Against medical advice | 58 (0.4) | 23 (0.3) | 29 (0.4) | <11 (0.4) | |
Values are expressed as mean ± standard deviation, median (interquartile range), or counts (%) unless otherwise noted. Exact counts for variables with <11 patients are not detailed per the Healthcare Cost and Utilization Project data use agreement. Boldface indicates statistical significance (P < .05).
CABG, coronary artery bypass surgery; CI, confidence interval; MAE, major adverse event; NSTEMI, non–ST elevation myocardial infarction; OR, odds ratio; PCI, percutaneous coronary intervention; STEMI, ST-elevation myocardial infarction; USD, US dollar.
Adjusted P values for each variable were computed from adjusting sampling design by discharge-level weights, cluster, and strata.
Median zip code income was missing in 1.5%.
Index cost was missing in 0.3%.
Other institutions included skilled nursing facilities, intermediate care, or other types of facilities not elsewhere included.
Among intermediate- and high-HFRS patients, the primary diagnosis of admission was mostly related to acute coronary syndromes (non–ST elevation myocardial infarction [NSTEMI], STEMI, and heart failure or cardiogenic shock), whereas low-HFRS patients were more likely to present with chronic (stable angina) coronary syndromes. A higher burden of comorbidities was noted with increasing HFRS as evidenced by significantly higher CCI and ECS (P < 0.001 for both; Table 1).
The median LOS during the index admission was significantly longer with a higher degree of HFRS (P < 0.001). Hence, the index admission was also associated with a greater median cost as the HFRS category increased (P < 0.001). Intermediate- and high-risk frailty patients were less likely to be discharged home-self-care (P < 0.001; Table 1).
Regarding the 31-180-day readmission analysis, of the 46,600 identified discharges, 40,103 patients were excluded because they died during their index admission, or were discharged between July and December, had missing information, or received ECMO or IABP. Of the remaining 6497 patients, the median age of the population was 75 years (IQR 70-81), and women made up 28% of the total population. According to the HFRS, 2976 patients (45.8%) were classified as low-risk, 2898 individuals (44.6%) as intermediate-risk, and 623 individuals (9.6%) as high-risk. Consequently, 3521 (54.2%) of individuals were categorized as frail based on the HFRS criteria. The remaining baseline characteristics and in-hospital outcomes of this cohort are presented in Table 2.
Table 2.
Baseline characteristics according to 31-180-day readmission
| Characteristic | All patients (n=6497) | Hospital Frailty Risk Score |
Adjusted P value∗ |
||
|---|---|---|---|---|---|
| Low (n=2976) | Intermediate (n=2898) | High (n=623) | |||
| Patient characteristics | |||||
| Age, y | 75 (70-81) | 75 (70-81) | 75 (69-81) | 75 (70-81) | 0.26 |
| Women | 1816 (28) | 762 (26) | 860 (30) | 194 (31) | <0.001 |
| Nonelective admission | 4828 (74) | 1803 (61) | 2458 (85) | 567 (91) | <0.001 |
| Weekend admission | 1181 (18) | 401 (13) | 629 (22) | 151 (24) | <0.001 |
| Median zip code income† | |||||
| 76-100th percentile | 1192 (19) | 514 (18) | 529 (19) | 149 (24) | 0.006 |
| 51-75th percentile | 1559 (24) | 691 (24) | 720 (25) | 148 (24) | |
| 26-50th percentile | 1845 (29) | 841 (29) | 825 (29) | 179 (29) | |
| 0-25th percentile | 1796 (28) | 878 (30) | 780 (27) | 138 (22) | |
| Expected primary payer | |||||
| Medicare | 5673 (87) | 2581 (87) | 2541 (88) | 551 (88) | 0.42 |
| Medicaid | 88 (1.4) | 39 (1.3) | 42 (1.4) | <11 (1.3) | |
| Private | 489 (7.5) | 250 (8.4) | 193 (6.7) | 46 (8.3) | |
| Others | 247 (3.8) | 106 (3.6) | 122 (4.2) | 19 (3.4) | |
| Comorbidities | |||||
| Smoking | 2082 (32) | 861 (29) | 1011 (35) | 210 (34) | <0.001 |
| Dyslipidemia | 4563 (70) | 2190 (74) | 1969 (68) | 404 (65) | <0.001 |
| Hypertension | 5628 (87) | 2517 (85) | 2565 (89) | 546 (88) | <0.001 |
| Diabetes mellitus | 2752 (42) | 1147 (39) | 1316 (45) | 289 (46) | <0.001 |
| Previous myocardial infarction | 1408 (22) | 659 (22) | 626 (22) | 123 (20) | 0.63 |
| Cardiac arrhythmias | 2774 (43) | 1031 (35) | 1392 (48) | 351 (56) | <0.001 |
| Previous PCI | 1186 (18) | 611 (21) | 467 (16) | 108 (17) | <0.001 |
| Previous CABG | 798 (12) | 440 (15) | 294 (10) | 64 (10) | <0.001 |
| Valvular disease | 1805 (28) | 715 (24) | 887 (31) | 203 (33) | <0.001 |
| Previous cerebrovascular disease | 804 (12) | 289 (9.7) | 373 (13) | 142 (23) | <0.001 |
| Fluid and electrolyte disorders | 17 (0.3) | <11 (0.1) | <11 (0.3) | <11 (0.6) | 0.05 |
| Peripheral vascular disease | 1237 (19) | 510 (17) | 592 (20) | 135 (22) | 0.003 |
| Chronic pulmonary disease | 1722 (27) | 649 (22) | 865 (30) | 208 (33) | <0.001 |
| Renal disease | 2305 (35) | 606 (20) | 1360 (47) | 339 (54) | <0.001 |
| Obesity | 1068 (16) | 414 (14) | 525 (18) | 129 (21) | <0.001 |
| Peptic ulcer disease | 44 (0.7) | <11 (0.2) | 28 (1.0) | <11 (1.4) | <0.001 |
| Rheumatic disease | 169 (2.6) | 68 (2.3) | 87 (3.0) | 14 (2.2) | 0.25 |
| Liver disease | 167 (2.6) | 50 (1.7) | 84 (2.9) | 33 (5.3) | <0.001 |
| Cancer | 361 (5.6) | 138 (4.6) | 189 (6.5) | 34 (5.5) | 0.001 |
| Anemia | 1887 (29) | 494 (17) | 1107 (38) | 286 (46) | <0.001 |
| Charlson Comorbidity Index | 3.5 ± 2.4 | 2.7 ± 2.1 | 4.0 ± 2.4 | 4.8 ± 2.6 | <0.001 |
| 0 | 352 (5.4) | 270 (9.1) | 73 (2.5) | <11 (1.4) | <0.001 |
| 1 | 1100 (17) | 673 (23) | 366 (13) | 61 (9.8) | |
| 2 | 1109 (17) | 676 (23) | 384 (13) | 49 (7.9) | |
| ≥3 | 3936 (61) | 1357 (46) | 2075 (72) | 504 (81) | |
| Elixhauser Comorbidity Score | 12.5 ± 7.6 | 9.6 ± 6.7 | 14.4 ± 7.2 | 17.8 ± 8.2 | <0.001 |
| ≤0 | 451 (6.9) | 367 (12) | 79 (2.7) | <11 (0.8) | <0.001 |
| 1-5 | 609 (9.4) | 378 (13) | 205 (7.1) | 26 (4.2) | |
| 6-10 | 1542 (24) | 931 (31) | 528 (18) | 83 (13) | |
| ≥11 | 3895 (60) | 1300 (44) | 2086 (72) | 509 (82) | |
| Hospital Frailty Risk Score | 6.8 ± 5.8 | 2.1 ± 1.5 | 9.0 ± 2.8 | 19.3 ± 4.1 | <0.001 |
| Diagnosis on admission | |||||
| STEMI | 943 (15) | 387 (13) | 450 (16) | 106 (17) | 0.001 |
| NSTEMI | 1890 (29) | 675 (23) | 1016 (35) | 199 (32) | <0.001 |
| Chronic coronary syndrome | 1998 (31) | 1365 (46) | 568 (20) | 65 (10) | <0.001 |
| Malignant arrhythmias/cardiac arrest | 334 (5.1) | 140 (4.7) | 163 (5.6) | 31 (5.0) | 0.35 |
| Heart failure and cardiogenic shock | 668 (10) | 205 (6.9) | 373 (13) | 90 (14) | <0.001 |
| Hospital characteristics | |||||
| Resident in same hospital state | 5918 (91) | 2696 (91) | 2642 (91) | 580 (93) | 0.38 |
| Teaching status of hospital | |||||
| Metropolitan teaching | 5145 (79) | 2254 (76) | 2374 (82) | 517 (83) | <0.001 |
| Metropolitan nonteaching | 1134 (17) | 597 (20) | 443 (15) | 94 (15) | |
| Nonmetropolitan hospital | 218 (3.4) | 125 (4.2) | 81 (2.8) | 12 (1.9) | |
| Hospital urban-rural designation | |||||
| Large metropolitan areas with at least 1 million residents | 3535 (54) | 1487 (50) | 1640 (57) | 408 (65) | <0.001 |
| Small metropolitan areas with <1 million residents | 2744 (42) | 1364 (46) | 1177 (41) | 203 (33) | |
| Micropolitan areas | 208 (3.2) | 123 (4.1) | 77 (2.7) | <11 (1.3) | |
| Not metropolitan or micropolitan (nonurban residual) | <11 (0.2) | <11 (0.07) | <11 (0.1) | <11 (0.6) | |
| Hospital bed size | |||||
| Small | 577 (8.9) | 296 (9.9) | 229 (7.9) | 52 (8.3) | 0.45 |
| Medium | 1647 (25) | 757 (25) | 723 (25) | 167 (27) | |
| Large | 4273 (66) | 1923 (65) | 1946 (67) | 404 (65) | |
| In-hospital procedures | |||||
| PCI | 5135 (79) | 2478 (83) | 2200 (76) | 457 (73) | <0.001 |
| CABG | 413 (6.4) | 131 (4.4) | 231 (8.0) | 51 (8.2) | <0.001 |
| Year of Impella procedure | |||||
| 2016 (January-November) | 726 (11) | 356 (12) | 327 (11) | 43 (6.9) | 0.03 |
| 2017 (January-November) | 1082 (17) | 507 (17) | 473 (16) | 101 (16) | |
| 2018 (January-November) | 1509 (23) | 706 (24) | 653 (23) | 150 (24) | |
| 2019 (January-November) | 1731 (27) | 773 (26) | 783 (27) | 175 (28) | |
| 2020 (January-November) | 1450 (22) | 634 (21) | 662 (23) | 154 (25) | |
| In-hospital complications | |||||
| Bleeding complications | 377 (5.8) | 112 (3.8) | 221 (7.6) | 44 (7.1) | <0.001 |
| Vascular complications | 102 (1.6) | 31 (1.0) | 58 (2.0) | 13 (2.1) | 0.003 |
| Stroke | 145 (2.2) | <11 (0.3) | 89 (3.1) | 48 (7.7) | <0.001 |
| Length of stay, median, d | 8 (3-15) | 4 (2-7) | 11 (7-18) | 21 (14-31) | <0.001 |
| ≤8 d | 3466 (53) | 2369 (80) | 1048 (36) | 49 (7.9) | <0.001 |
| >8 d | 3031 (47) | 607 (20) | 1850 (64) | 574 (92) | |
| Index cost (USD)‡ | 65,188 (46,546-93,362) | 52,705 (39,431-70,949) | 75,251 (55,319-106,289) | 104,724 (75,540-147,610) | <0.001 |
| Discharge destination | |||||
| Home (self-care) | 3146 (48) | 2068 (70) | 996 (34) | 82 (13) | <0.001 |
| Short-term hospital | 206 (3.2) | 64 (2.2) | 120 (4.1) | 22 (3.5) | |
| Transfer to other institution§ | 1617 (25) | 297 (10) | 994 (34) | 326 (52) | |
| Home health care | 1499 (23) | 537 (18) | 775 (27) | 187 (30) | |
| Against medical advice | 21 (0.3) | <11 (0.3) | <11 (0.3) | <11 (0.6) | |
Values are expressed as mean ± standard deviation, median (interquartile range), or counts (%) unless otherwise noted. Exact counts for variables with <11 patients are not detailed per the Healthcare Cost and Utilization Project data use agreement. Boldface indicates statistical significance (P < .05).
CABG, coronary artery bypass surgery; CI, confidence interval; MAE, major adverse event; NSTEMI, non–ST elevation myocardial infarction; OR, odds ratio; PCI, percutaneous coronary intervention; STEMI, ST-elevation myocardial infarction; USD, US dollar.
Adjusted P values for each variable were computed from adjusting sampling design by discharge-level weights, cluster, and strata.
Median zip code income was missing in 1.5%.
Index cost was missing in 0.1%.
Other institutions included skilled nursing facilities, intermediate care, or other types of facilities not elsewhere included.
Intermediate- and high-risk frailty patients presented more often with primary diagnosis of NSTEMI, STEMI, and heart failure or cardiogenic shock on their index admission. In contrast, patients classified as low HFRS were more likely to present with chronic coronary syndrome. Comorbidity burden was higher with higher HFRS as evidenced by significantly higher CCI and ECS (P < 0.001 for both).
The LOS and the cost of index admission were also significantly higher with increasing HFRS category and, as the HFRS increased, patients were significantly less likely to be discharged home-self-care (P < 0.001 for both).
Thirty-day unplanned readmissions and outcomes
Of the 16,004 discharges, 2185 (13.7%) had an unplanned readmission at 30 days. Between discharge and 30 days, the 30-day timing of readmissions for frail and nonfrail patients is depicted in Figure 1A. The highest spike occurred on days 18 and 19 for frail patients and on days 10 and 12 for nonfrail patients.
Figure 1.
Timing of readmissions for frail and nonfrail patients at (A) 30 days and (B) 31-180 days.
At 30 days, the most common causes of unplanned readmissions at 30 days for patients classified as frail (HFRS ≥5) and nonfrail (low HFRS) were heart failure, infections, and bleeding (Fig. 2). The 30-day readmission rates were found to be highest in intermediate HFRS patients, followed by low-HFRS patients and then high-HFRS patients (P = 0.001; Table 3). The incidence of MAEs during readmission was significantly greater in the high-HFRS group (P = 0.006). Acute kidney injury was much higher among high-HFRS patients (P < 0.001). In-hospital mortality during readmission occurred in 8.2% of the population, and this mortality rate was nearly twice as high in the high-HFRS group (15%) compared to the intermediate-HFRS (7.6%) and low-HFRS groups (7.8%); however, this difference did not reach statistical significance (P = 0.09). As expected, the cost of readmission increased significantly with increasing frailty category (P < 0.001; Table 3).
Figure 2.
Causes of readmission at 30 days: (A) cardiac causes; (B) noncardiac causes.
Table 3.
Readmission rates and in-hospital outcomes during readmission
| Discharge to 30-d cohort | All patients (n=16,004) | Hospital Frailty Risk Score |
Adjusted P value∗ |
||
|---|---|---|---|---|---|
| Low (n=7357) | Intermediate (n=7163) | High (n=1484) | |||
| Readmission data | |||||
| 30-d readmission rate | 2185 (14) | 912 (12) | 1111 (16) | 162 (11) | 0.001 |
| MAE during readmission | 605 (28) | 220 (24) | 335 (30) | 50 (31) | 0.006 |
| Acute kidney injury | 804 (37) | 287 (31) | 431 (39) | 86 (53) | <0.001 |
| Need for transfusions | 296 (14) | 120 (13) | 151 (14) | 25 (15) | 0.57 |
| In-hospital death during readmission | 179 (8.2) | 71 (7.8) | 84 (7.6) | 24 (15) | 0.09 |
| Length of stay (d) | 4 (2-7) | 4 (2-6) | 4 (2-7) | 5 (3-9) | <0.001 |
| ≤4 d | 1201 (55) | 546 (60) | 575 (52) | 80 (49) | 0.008 |
| >4 d | 984 (45) | 366 (40) | 536 (48) | 82 (51) | |
| Readmission cost (USD) | 9775 (5457-19,886) | 9088 (5129-19,789) | 9987 (5740-19,463) | 11,547 (6265-22,387) | <0.001 |
| Discharge destination after readmission | |||||
| Home (self-care) | 774 (39) | 424 (50) | 319 (31) | 31 (22) | <0.001 |
| Short-term hospital | 35 (1.7) | 13 (1.5) | 18 (1.8) | <11 (2.9) | |
| Transfer to other institution† | 556 (28) | 137 (16) | 359 (35) | 60 (43) | |
| Home health care | 623 (31) | 263 (31) | 319 (31) | 41 (30) | |
| Against medical advice | 18 (0.9) | <11 (0.5) | 12 (1.2) | <11 (1.4) | |
| Discharge to 31-180-d cohort | All patients (n=6497) | Hospital Frailty Risk Score | Adjusted P value∗ |
||
| Low (n=2976) |
Intermediate (n=2898) |
High (n=623) |
|||
| Readmission data | |||||
| 31-180-d readmission rate | 1809 (28) | 614 (21) | 964 (33) | 231 (37) | <0.001 |
| MAE during readmission | 201 (11) | 65 (11) | 105 (11) | 31 (13) | 0.59 |
| Acute kidney injury | 585 (32) | 159 (26) | 338 (35) | 88 (38) | <0.001 |
| Need for transfusions | 196 (11) | 58 (9.4) | 103 (11) | 35 (15) | 0.08 |
| In-hospital death during readmission | 116 (6.4) | 29 (4.7) | 68 (7.1) | 19 (8.2) | 0.02 |
| Length of stay (d) | 4 (2-8) | 4 (2-6) | 4 (3-8) | 5 (3-9) | <0.001 |
| ≤4 d | 985 (54) | 369 (60) | 512 (53) | 104 (45) | <0.001 |
| >4 d | 824 (46) | 245 (40) | 452 (47) | 127 (55) | |
| Readmission cost (USD) | 10,386 (5621-20,526) | 10,128 (5297-20,200) | 10,385 (5661-20,932) | 10,921 (6786-21,333) | <0.001 |
| Discharge destination after readmission | |||||
| Home (self-care) | 758 (45) | 345 (59) | 358 (40) | 55 (26) | <0.001 |
| Short-term hospital | 18 (1.1) | <11 (1.0) | <11 (1.1) | <11 (0.9) | |
| Transfer to other institution† | 434 (26) | 92 (16) | 255 (28) | 87 (41) | |
| Home health care | 473 (28) | 137 (23) | 269 (30) | 67 (32) | |
| Against medical advice | <11 (0.6) | <11 (0.9) | <11 (0.4) | <11 (0.5) | |
Values are expressed as mean ± standard deviation, median (interquartile range), or counts (%) unless otherwise noted. Exact counts for variables with <11 patients are not detailed per the Healthcare Cost and Utilization Project data use agreement. Boldface indicates statistical significance (P < .05).
CI, confidence interval; MAE, major adverse events; OR, odds ratio; USD, US dollar.
Adjusted P values for each variable were computed from adjusting sampling design by discharge-level weights, cluster, and strata.
Other institutions included skilled nursing facilities, intermediate care, or other types of facilities not elsewhere included.
When analysing the outcomes based on frailty status (HFRS ≥5), these individuals exhibited a significantly higher risk of 30-day readmissions (OR 1.27, 95% CI 1.17-1.37), in-hospital death during readmission (OR 2.01, 95% CI 1.22-3.30), MAEs during readmission (OR 1.73, 95% CI 1.29-2.33), LOS >4 days (OR 1.80, 95% CI 1.44-2.26), and readmission cost greater than the median (OR 1.44, 95% CI 1.17-1.80; Table 4).
Table 4.
Multilevel multivariable logistic regression models adjusted for age and sex comorbidities not included in the Hospital Frailty Risk Score
| Outcome | Subject characteristic | Odds ratio (95% CI) | P value |
|---|---|---|---|
| 30-d readmission | |||
| 30-d readmission | HFRS continuous (1-point increase) | 1.01 (1.00-1.01) | 0.04 |
| HFRS ≥5 | 1.27 (1.17-1.37) | <0.001 | |
| In-hospital death during readmission | HFRS continuous (1-point increase) | 1.13 (1.08-1.18) | <0.001 |
| HFRS ≥5 | 2.01 (1.22-3.30) | 0.006 | |
| MAE during readmission | HFRS continuous (1-point increase) | 1.04 (1.01-1.07) | 0.02 |
| HFRS ≥5 | 1.73 (1.29-2.33) | <0.001 | |
| Length of stay >4 d | HFRS continuous (1-point increase) | 1.06 (1.04-1.09) | <0.001 |
| HFRS ≥5 | 1.80 (1.44-2.26) | <0.001 | |
| Hospital cost during readmission greater than the median cost | HFRS continuous (1-point increase) | 1.05 (1.02-1.07) | <0.001 |
| HFRS ≥5 | 1.44 (1.17-1.80) | <0.001 | |
| 31-180-d readmission | |||
| 31-180-d readmission | HFRS continuous (1-point increase) | 1.07 (1.06-1.08) | <0.001 |
| HFRS ≥5 | 2.10 (1.88-2.34) | <0.001 | |
| In-hospital death during readmission | HFRS continuous (1-point increase) | 1.08 (1.02-1.15) | 0.006 |
| HFRS ≥5 | 3.02 (1.33-6.86) | 0.008 | |
| MAE during readmission | HFRS continuous (1-point increase) | 0.99 (0.96-1.03) | 0.80 |
| HFRS ≥5 | 0.84 (0.50-1.41) | 0.51 | |
| Length of stay >4 d | HFRS continuous (1-point increase) | 1.05 (1.03-1.07) | <0.001 |
| HFRS ≥5 | 1.66 (1.29-2.14) | <0.001 | |
| Hospital cost during readmission greater than the median cost | HFRS continuous (1-point increase) | 1.03 (1.01-1.05) | 0.002 |
| HFRS ≥5 | 1.36 (1.05-1.75) | 0.02 |
Comorbidities used for adjustment that are not included in the HFRS include smoking, hypertension, diabetes, previous myocardial infarction, previous PCI, previous CABG, valvular disease, peripheral vascular disease, chronic pulmonary disease, obesity, rheumatic disease, liver disease, and cancer. Boldface indicates statistical significance (P < .05).
CABG, coronary artery bypass surgery; HFRS, Hospital Frailty Risk Score; MAE, major adverse event; PCI, percutaneous coronary intervention.
Thirty-one- to 180-day unplanned readmissions and outcomes
Of the 6497 discharges, 1809 (27.8%) experienced an unplanned readmission. The timing of readmissions between 31 and 180 days for frail and nonfrail patients is depicted in Figure 1B. The highest spike occurred on day 36 for frail patients and on days 32 and 33 for nonfrail patients. The most common causes of readmission for frail patients (HFRS ≥5) and nonfrail patients (low HFRS) were heart failure, infections, and MI (Fig. 3).
Figure 3.
Causes of readmission at 31-180 days: (A) cardiac causes; (B) noncardiac causes.
Individuals exhibiting a higher HFRS experienced significantly higher readmission rates (P < 0.001; Table 3). The occurrence of MAEs during readmission did not differ significantly between groups. However, those with higher HFRS showed higher incidence of acute kidney injury (P < 0.001). In-hospital mortality was also higher among patients with higher HFRS (P = 0.02). The cost of readmission also increased significantly with increasing HFRS category (P < 0.001; Table 3).
Individuals with frailty status (HFRS ≥5) were associated with significantly higher risks for 31-180-day readmission (OR 2.10, 95% CI 1.88-2.34), in-hospital death during readmission (OR 3.02, 95% CI 1.33-6.86), LOS >4 days (OR 1.66, 95% CI 1.29-2.14), and a hospitalization cost above the median (OR 1.36, 95% CI 1.05-1.75) (Table 4).
Post hoc analyses from multivariable logistic regression models adjusted for age and sex using the CCI (Supplemental Table S7) and ECS (Supplemental Table S8) did not improve prognostic utility but rather seriously increased imprecision as seen in concerning wider CIs around the point estimates.
Discussion
This large nationwide cohort study of readmissions after Impella MCS is the first to assess frailty status according to the HFRS and demonstrates several important findings. First, frailty is highly prevalent among patients undergoing MCS with Impella, with more than half of the population classified as either intermediate- or high-risk per the HFRS. Second, short- and intermediate-term readmissions after Impella MCS were common, occurring in 14% of patients at 30 days and 28% of patients between 31 and 180 days. Third, frailty as assessed by the HFRS was independently associated with an increased risk of readmissions at 30 and 31-180 days after discharge, in-hospital death, more prolonged LOS, and greater hospital expenditures during readmission episodes. Heart failure was the leading cause of readmissions accounting for 25% of cases.
Readmissions and outcomes
The impact of frailty on outcomes during readmission after Impella MCS has not been previously reported, and our study suggests that frailty, as assessed by the HFRS, is an independent risk factor for several adverse outcomes. Although not a clinical scoring system, the HFRS has been validated against clinical frailty scores such as the Rockwood and Fried scoring systems.22,28 Nonetheless, although the HFRS may be a useful tool for studying older populations using claims-based administrative codes, its agreement with measured frailty is low as reported by Gilbert et al.,22 with a κ statistic for agreement as low as 0.22 (95% CI 0.15-0.30) compared with a binary definition of frailty based on the Fried frailty phenotype criteria (3 or more items present).29 However, it is important to point out that the Fried phenotype assessment is a measure of physical frailty whereas the HFRS is a measure of deficit accumulation for frailty. Furthermore, the optimal tool for frailty assessment appears to be challenging considering that there are more than 40 available tools, all of which assess different areas, yet still lacking consensus for a unique scoring system.30
Routine clinical assessment of frailty may assist in recognizing patients at higher risk of worse clinical outcomes, readmissions, and complications during readmission while may also help target patients who could benefit from more intensive post-discharge care during follow-up.12 It may also aid in identifying those who may benefit from targeted interventions such as cardiac rehabilitation and exercise programs that have been shown to improve frailty and outcomes after invasive cardiac procedures.31, 32, 33, 34
Frailty, as assessed by the HFRS, has been demonstrated to be an independent predictor of 30-day readmission for a variety of cardiac conditions and procedures.10,15,24 In this study, individuals with a greater degree of frailty were at a significantly elevated risk of readmission after Impella MCS, AKI, prolonged LOS, and hospital expenditures.
Considering that a quarter of readmissions were heart failure related, our results highlight the importance of prescription of guideline-directed medical therapies and optimization during in-hospital stays to reduce readmissions. In addition, transition of care with close follow-up in heart function clinics are of paramount importance to ensure continuity of care and prevent readmissions.
The HFRS has been previously demonstrated to be associated with higher in-hospital mortality, with those with higher HFRS at the greatest risk.35,36 Therefore, it is possible that higher rates of in-hospital death during the index admission served as a competing risk factor by preventing the most vulnerable patients from readmissions.37 Moreover, patients in the high-HFRS group were significantly less likely to be discharged at home, with only 14% discharged into a self-care environment, whereas more than half were transferred to other institutions such as skilled nursing facilities and one-third receiving home health care.
Our results are consistent with previous reports, which show that higher HFRSs are associated with increased rates of readmissions among patients undergoing transcatheter structural heart procedures, PCI, and cardiac surgery.7,8,23, 24, 25,38 We found that 10% to 11% of patients were readmitted because of infections, 6.5% as a result of bleeding, and 2% for peripheral vascular disease. Furthermore, more than 30% of patients experienced AKI, and more than 10% require transfusions during readmission. Of note, the rate of adverse events during readmissions was markedly higher during the short-term (within 30 days) compared with mid-term (31-180 days, 28% vs 11%). This relates, in part, to the fact that several of the analysed events, such as stroke, vascular complications, and bleeding, are likely to be temporally associated with the patient’s index procedure and Impella insertion, therefore, are less likely to occur as time passes. Moreover, survival bias and selection of a “fitter” patient population may have also played a role.
Limitations
Our study must be interpreted within the context of its limitations of using an administrative database. It lacks granularity on important information such as concomitant medications during hospital admission and optimization of guideline-directed medical treatment after discharge, or even destination therapies, all relevant considering that one-fourth of the patients were readmitted because of heart failure. Because the cohort included patients with acute coronary syndromes, chronic coronary artery disease, supported high-risk PCI, as well as heart (advanced) failure therapy, generalization of the results is limited. Moreover, methods and completeness of revascularization, the indication for initiating MCS with Impella, the timing of insertion and duration of MCS, as well as details on hemodynamic variables and laboratory were not available. Furthermore, information pertaining to the specific type of Impella device (ie, 2.5, CP, or 5.0) and the method of access (percutaneous or surgical cut-down) was not available. The unavailability of left and right ventricular function data along with other relevant factors may have influenced the outcomes. However, the primary endpoint of this study focuses on readmission rates and outcomes during the readmission encounter rather than in-hospital outcomes of the index admission; hence, we included only individuals who were discharged following an initial admission where the indication for MCS with Impella took place. The NRD does not track patients who are readmitted to a hospital in a different state and does not track mortality data for patients who died outside of a hospital setting.
The HFRS also has limitations. The use of ICD-10 codes also means that the scoring system is subject to variation in documentation and coding which may lead to measurement errors. Furthermore, ICD codes do not adequately capture disease severity, functional capacity or reserve, cognitive decline, nutrition status, and psychologic measure or the need for support of activities of daily living or polypharmacy, all of which are critical components in the assessment of frailty.22,30 Furthermore, functional declining and deconditioning are often observed during admission and even more so in sicker patients.
Conclusion
Frailty is common among patients receiving MCS with an Impella device. It is independently associated with short- and mid-term readmission rates in addition to adverse outcomes during readmissions including in-hospital death. Heart failure was the leading cause of readmissions accounting for 25% of cases. Routine clinical assessment of frailty may assist in identifying patients who stand to benefit from more intensive post-discharge follow-up and enrolment in rehabilitation programs as well as heart failure clinics. This will become increasingly important as rates of complex procedures in the older adults continue to rise. Additional prospective data are also needed to further assess the impact of frailty on patients receiving MCS to better guide clinical decision making when managing this vulnerable group of patients.
Data Availability Statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Acknowledgements
The authors would like to thank the Healthcare Cost and Utilization Project (HCUP) and the HCUP Data Partners for providing the data used in the analysis.
Ethics Statement
The Western University Health Science Research Ethics Board has exempted the requirement for consent as well as the necessity for approval from the Ethics Committee or Institutional Review Board for this study. This decision stems from the fact that all of the data used in this research originates from the National Readmissions Database (NRD), which is a publicly accessible and deidentified administrative database.
Patient Consent
Not required.
Funding Sources
The authors have not declared a specific grant for this research from any funding agency in the public, commercial, or not-for-profit sectors.
Disclosures
Dr Damluji receives research funding from the Pepper Scholars Program of the Johns Hopkins University Claude D. Pepper Older Americans Independence Center funded by the National Institute on Aging (P30-AG021334); mentored patient-oriented research career development award from the National Heart, Lung, and Blood Institute (K23-HL153771); the National Institutes of Health (NIH) National Institute of Aging (R01-AG078153); and the Patient-Centered Outcomes Research Institute (PCORI). The remaining authors have no conflicts of interest to disclose.
Footnotes
See page 984 for disclosure information.
To access the supplementary material accompanying this article, visit CJC Open at https://www.cjcopen.ca/ and at https://doi.org/10.1016/j.cjco.2025.04.011
Supplementary Material
References
- 1.Amin A.P., Spertus J.A., Curtis J.P., et al. The evolving landscape of Impella use in the United States among patients undergoing percutaneous coronary intervention with mechanical circulatory support. Circulation. 2020;141:273–284. doi: 10.1161/CIRCULATIONAHA.119.044007. [DOI] [PubMed] [Google Scholar]
- 2.O'Neill W.W., Kleiman N.S., Moses J., et al. A prospective, randomized clinical trial of hemodynamic support with Impella 2.5 versus intra-aortic balloon pump in patients undergoing high-risk percutaneous coronary intervention: the PROTECT II study. Circulation. 2012;126:1717–1727. doi: 10.1161/CIRCULATIONAHA.112.098194. [DOI] [PubMed] [Google Scholar]
- 3.Clegg A., Young J., Iliffe S., Rikkert M.O., Rockwood K. Frailty in elderly people. Lancet. 2013;381:752–762. doi: 10.1016/S0140-6736(12)62167-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dou Q., Wang W., Wang H., et al. Prognostic value of frailty in elderly patients with acute coronary syndrome: a systematic review and meta-analysis. BMC Geriatr. 2019;19:222. doi: 10.1186/s12877-019-1242-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Collard R.M., Boter H., Schoevers R.A., Oude Voshaar R.C. Prevalence of frailty in community-dwelling older persons: a systematic review. J Am Geriatr Soc. 2012;60:1487–1492. doi: 10.1111/j.1532-5415.2012.04054.x. [DOI] [PubMed] [Google Scholar]
- 6.Murali-Krishnan R., Iqbal J., Rowe R., et al. Impact of frailty on outcomes after percutaneous coronary intervention: a prospective cohort study. Open Heart. 2015;2 doi: 10.1136/openhrt-2015-000294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kwok C.S., Lundberg G., Al-Faleh H., et al. Relation of frailty to outcomes in patients with acute coronary syndromes. Am J Cardiol. 2019;124:1002–1011. doi: 10.1016/j.amjcard.2019.07.003. [DOI] [PubMed] [Google Scholar]
- 8.Kwok C.S., Achenbach S., Curzen N., et al. Relation of frailty to outcomes in percutaneous coronary intervention. Cardiovasc Revasc Med. 2020;21:811–818. doi: 10.1016/j.carrev.2019.11.009. [DOI] [PubMed] [Google Scholar]
- 9.Heaton J., Singh S., Nanavaty D., et al. Impact of frailty on outcomes in acute ST-elevated myocardial infarctions undergoing percutaneous coronary intervention. Catheter Cardiovasc Interv. 2023;101:773–786. doi: 10.1002/ccd.30595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Iyengar A., Goel N., Kelly J.J., et al. Effects of frailty on outcomes and 30-day readmissions after surgical mitral valve replacement. Ann Thorac Surg. 2020;109:1120–1126. doi: 10.1016/j.athoracsur.2019.10.087. [DOI] [PubMed] [Google Scholar]
- 11.Rajani R., Lindblom M., Dixon G., et al. Evolving trends in percutaneous coronary intervention. Br J Cardiol. 2011;18:73–76. [Google Scholar]
- 12.Damluji A.A., Nanna M.G., Rymer J., et al. Chronological vs biological age in interventional cardiology: a comprehensive approach to care for older adults: JACC Family Series. JACC Cardiovasc Interv. 2024;17:961–978. doi: 10.1016/j.jcin.2024.01.284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Damluji A.A., Forman D.E., Wang T.Y., et al. Management of acute coronary syndrome in the older adult population: a scientific statement from the American Heart Association. Circulation. 2023;147:e32–e62. doi: 10.1161/CIR.0000000000001112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Patel A., Goodman S.G., Yan A.T., et al. Frailty and outcomes after myocardial infarction: insights from the CONCORDANCE Registry. J Am Heart Assoc. 2018;7 doi: 10.1161/JAHA.118.009859. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kundi H., Wadhera R.K., Strom J.B., et al. Association of frailty with 30-day outcomes for acute myocardial infarction, heart failure, and pneumonia among elderly adults. JAMA Cardiol. 2019;4:1084–1091. doi: 10.1001/jamacardio.2019.3511. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Davis-Ajami M.L., Chang P.S., Wu J. Hospital readmission and mortality associations to frailty in hospitalized patients with coronary heart disease. Aging Health Res. 2021;1 [Google Scholar]
- 17.Rubens M., Ramamoorthy V., Saxena A., et al. Association between frailty, 30-day unplanned readmission and mortality after hospitalization for heart failure: results from the nationwide readmissions database. J Aging Health. 2023;35:651–659. doi: 10.1177/08982643231152694. [DOI] [PubMed] [Google Scholar]
- 18.Sanjoy S., Choi Y.H., Holmes D., et al. Comorbidity burden in patients undergoing left atrial appendage closure. Heart. 2021;107:1246–1253. doi: 10.1136/heartjnl-2020-317741. [DOI] [PubMed] [Google Scholar]
- 19.Sanjoy S.S., Choi Y.H., Sparrow R.T., et al. Sex differences in outcomes following left atrial appendage closure. Mayo Clin Proc. 2021;96:1845–1860. doi: 10.1016/j.mayocp.2020.11.031. [DOI] [PubMed] [Google Scholar]
- 20.Sanjoy S.S., Choi Y.H., Sparrow R.T., et al. Outcomes of elderly patients undergoing left atrial appendage closure. J Am Heart Assoc. 2021;10 doi: 10.1161/JAHA.121.021973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sparrow R.T., Sanjoy S.S., Lindman B.R., et al. Racial, ethnic and socioeconomic disparities in patients undergoing transcatheter mitral edge-to-edge repair. Int J Cardiol. 2021;344:73–81. doi: 10.1016/j.ijcard.2021.09.037. [DOI] [PubMed] [Google Scholar]
- 22.Gilbert T., Neuburger J., Kraindler J., et al. Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391:1775–1782. doi: 10.1016/S0140-6736(18)30668-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kundi H., Popma J.J., Reynolds M.R., et al. Frailty and related outcomes in patients undergoing transcatheter valve therapies in a nationwide cohort. Eur Heart J. 2019;40:2231–2239. doi: 10.1093/eurheartj/ehz187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wang A., Ferro E.G., Song Y., et al. Frailty in patients undergoing percutaneous left atrial appendage closure. Heart Rhythm. 2022;19:814–821. doi: 10.1016/j.hrthm.2022.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Li Z., Wijeysundera H.C., Bagur R., et al. Performance of administrative database frailty instruments in predicting clinical outcomes and cost for patients undergoing transcatheter aortic valve implantation: a historical cohort study. Can J Anaesth. 2023;70:116–129. doi: 10.1007/s12630-022-02354-6. [DOI] [PubMed] [Google Scholar]
- 26.Kuchtaruk A.A., Sparrow R.T., Azzalini L., et al. Unplanned readmissions after Impella mechanical circulatory support. Int J Cardiol. 2023;379:48–59. doi: 10.1016/j.ijcard.2023.03.013. [DOI] [PubMed] [Google Scholar]
- 27.Vandenbroucke J.P., von Elm E., Altman D.G., et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4 doi: 10.1371/journal.pmed.0040297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Segal J.B., Chang H.Y., Du Y., et al. Development of a claims-based frailty indicator anchored to a well-established frailty phenotype. Med Care. 2017;55:716–722. doi: 10.1097/MLR.0000000000000729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Damluji A.A., Gerstenblith G., Segal J.B. Frailty measurement using administrative data in older patients with cardiovascular disease. JAMA Cardiol. 2020;5:967–968. doi: 10.1001/jamacardio.2020.1552. [DOI] [PubMed] [Google Scholar]
- 30.Ijaz N., Buta B., Xue Q.L., et al. Interventions for frailty among older adults with cardiovascular disease: JACC state-of-the-art review. J Am Coll Cardiol. 2022;79:482–503. doi: 10.1016/j.jacc.2021.11.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Herdy A.H., Marcchi P.L., Vila A., et al. Pre- and postoperative cardiopulmonary rehabilitation in hospitalized patients undergoing coronary artery bypass surgery: a randomized controlled trial. Am J Phys Med Rehabil. 2008;87:714–719. doi: 10.1097/PHM.0b013e3181839152. [DOI] [PubMed] [Google Scholar]
- 32.Bray N.W., Smart R.R., Jakobi J.M., Jones G.R. Exercise prescription to reverse frailty. Appl Physiol Nutr Metab. 2016;41:1112–1116. doi: 10.1139/apnm-2016-0226. [DOI] [PubMed] [Google Scholar]
- 33.Eichler S., Salzwedel A., Reibis R., et al. Multicomponent cardiac rehabilitation in patients after transcatheter aortic valve implantation: predictors of functional and psychocognitive recovery. Eur J Prev Cardiol. 2017;24:257–264. doi: 10.1177/2047487316679527. [DOI] [PubMed] [Google Scholar]
- 34.Tamuleviciute-Prasciene E., Beigiene A., Thompson M.J., et al. The impact of additional resistance and balance training in exercise-based cardiac rehabilitation in older patients after valve surgery or intervention: randomized control trial. BMC Geriatr. 2021;21:23. doi: 10.1186/s12877-020-01964-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kwok C.S., Zieroth S., Van Spall H.G.C., et al. The Hospital Frailty Risk Score and its association with in-hospital mortality, cost, length of stay and discharge location in patients with heart failure short running title: frailty and outcomes in heart failure. Int J Cardiol. 2020;300:184–190. doi: 10.1016/j.ijcard.2019.09.064. [DOI] [PubMed] [Google Scholar]
- 36.Sokhal B.S., Matetić A., Rashid M., et al. Association of frailty status on the causes and outcomes of patients admitted with cardiovascular disease. Am J Cardiol. 2023;192:7–15. doi: 10.1016/j.amjcard.2022.12.029. [DOI] [PubMed] [Google Scholar]
- 37.Damluji A.A., Ijaz N., Chung S.E., et al. Hierarchical development of physical frailty and cognitive impairment and their association with incident cardiovascular disease. JACC Adv. 2023;2 doi: 10.1016/j.jacadv.2023.100318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Strange J.E., Christensen D.M., Sindet-Pedersen C., et al. Frailty and recurrent hospitalization after transcatheter aortic valve replacement. J Am Heart Assoc. 2023;12 doi: 10.1161/JAHA.122.029264. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data relevant to the study are included in the article or uploaded as supplementary information.