Cardiac surgery as a means of reversing frailty – the CURE-Frailty Trial a prospective cohort study Authors: V-A. Rudas, A. Lassnigg, A. Fischer-Hammerschmied, S. Ryz, J. Tichy, R. Ristl, J. Lanzerstorfer, M. H. Bernardi Published: 2026 ABSTRACT BackgroundThere is significant clinical and pathophysiological overlap between frailty and cardiovascular disease. Reversibility of frailty has previously been discussed; therefore, we hypothesised that alleviation of heart failure symptoms through cardiac surgery can reduce frailty. We aimed to examine frailty before and after cardiac surgery, as well as factors influencing frailty improvement.MethodsWe conducted a prospective observational single centre study in patients undergoing elective open or interventional cardiac surgery. Frailty was assessed using the Clinical Frailty Scale (CFS). Follow-up was performed at the time of hospital discharge, as well as 6 and 12 months postoperatively. Change in CFS between baseline and final follow-up was the primary endpoint of the study and was analysed using a paired Wilcoxon rank-sum test. Additionally, a logistic regression analysis was conducted to identify factors influencing improvement of frailty.ResultsSeventy-six patients were analysed in this trial. 1 year after cardiac surgery, median CFS decreased from 4 to 3 (P < 0.001). Frailty improved in 35.5%, increased in 5.2% and remained stable in 38.2% of participants. Multivariate logistic regression analysis showed a significant association between cardiac rehabilitation and improvement of frailty.ConclusionsCardiac surgery has the potential to reverse frailty, with cardiac rehabilitation being an important factor in the postoperative improvement of frailty.Trial registrationClinicalTrials.gov (ID: NCT04875923), registered on 21/04/2021.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12877-026-07764-5. Background The Frailty Syndrome is a multifactorial condition associated with older age. In individuals aged 65 years or older, prevalence is around 10%, as opposed to more than 25% in patients aged over 85 years [1]. Additionally, with the presence of cardiovascular disease, frailty prevalence rises to more than 50% [2]. One of the main clinical implications of frailty is increased vulnerability to external stressors, such as infections or surgical procedures [3]. Following surgery or admission to the intensive care unit (ICU), for example, frailty is associated with increased mortality, higher rates of adverse events, as well as a higher probability of discharge into institutional care. Furthermore, quality of life can be negatively influenced by frailty [4, 5]. Therefore, evaluation of frailty is an important part of delivery of medical care and can provide guidance in clinical decision making [6]. Among the key features of frailty are unintentional weight loss, reduced strength and physical activity, as well as the feeling of exhaustion [7]. These symptoms, however, overlap closely with those of cardiovascular disease (CVD). Furthermore, both conditions seem to promote the development of the other. For example, while a lack of physiological reserve makes frail patients prone to cardiac decompensation [2], heart failure leads to physical inactivity and sarcopenia due to reduced exercise capacity [8]. Frailty has been shown to be a dynamic process, with transitions in both directions possible [9]. As potential treatment options, exercise and dietary measures have been discussed [10]. Due to the overlap of CVD and frailty, a possible reduction of frailty through alleviation of disease symptoms can be hypothesised. To date, longitudinal studies examining cardiac surgery as potential treatment for frailty are scarce. Therefore, we aimed to analyse whether preoperative frailty improves after cardiac surgery. Additionally, we wanted to identify factors influencing frailty improvement and study the effects of cardiac surgery on functional status and quality of life. Methods Study design and setting The CURE-Frailty Trial was conducted as a prospective observational cohort study. Patients were enrolled between April and September 2021. The aim of this study was to examine frailty before and after cardiac surgery, as well as factors influencing frailty improvement. Participants We included patients undergoing elective open and interventional cardiac surgery at our centre. We excluded patients younger than 60 years, to ensure a sufficient prevalence of frailty in our study sample to examine potential frailty improvement. Further exclusion criteria were patients undergoing emergency procedures or non-cardiac surgery such as pulmonary thrombendarteriectomy, as well as pregnant women or patients who declined to participate. Data collection Case report forms were used to record baseline patient data [age, sex, body mass index (BMI), European system for cardiac operative risk evaluation (EuroSCORE) II, recent doctor’s visits, hospitalizations and long-term medication], laboratory results (serum hemoglobin, serum creatinine, albumin and proBNP), surgical factors [diagnosis, planned procedure, surgical access, duration of anaesthesia, surgery, cardiopulmonary bypass (CPB) and aortic cross-clamp], unplanned intra- or postoperative use of assist devices [left ventricular assist devices (LVAD), extra-corporal membrane oxygenation (ECMO), or renal replacement therapy (RRT)]. Frailty was primarily assessed using the Clinical Frailty Scale (CFS) [11] and defined as a CFS of 4 or greater. The CFS was chosen as our primary frailty marker for two reasons. First, it is a widely validated score, in both routine and critical care settings. It is a quick and simple measure that does not need extensive prior training and can therefore be applied by experienced clinicians and study nurses alike. After obtaining a general impression of the patient and assessing mobility and dependence in activities of daily living, a frailty level ranging from 1 (very fit) to 9 (terminally ill) is assigned. The scale is supported by pictograms and brief descriptions for each category, facilitating a rapid and standardised assessment (Supplemental Fig. 1). Furthermore, part of the follow-up was conducted via telephone, for which the clinical frailty scale has been shown to be feasible as well, along with the SF-36, the precursor of the SF-12 [12]. Second, the CFS is the standard tool for frailty assessment during routine preanesthesia consultations at our centre, thus making it possible to compare our study sample with the general surgical population at our centre. Hand grip strength, Five-Times-Sit-To-Stand-Test (FTSST) [13] and the Essential Frailty Toolset (EFT) [14] were measured additionally to provide a more comprehensive frailty assessment. Hand grip strength was used to assess strength in the upper extremities, while the FTSST examines the lower extremities. Both tests are easy to conduct with minimal use of space and utensils, while carrying minimal risk of overexertion for the participants. The EFT was assessed additionally, since the information needed to score this test (i.e. serum hemoglobin and albumin, FTSST and cognitive function) were already available. Furthermore, functional status and quality of life were assessed using the Functional Independence Measure (FIM) [15] and the Short Form 12 Patient Questionnaire (SF-12) [16], which was scored according to the corresponding scoring manual [17]. Patients were enrolled within 10 days before surgery and follow-up took place at hospital discharge, as well as 6 and 12 months postoperatively. At baseline and hospital discharge, data was collected in person, while the long-term follow-up was conducted via telephone. Outcomes The primary outcome of this study was the change in postoperative CFS one year after cardiac surgery compared to the preoperative baseline. Secondary outcomes consisted of surgical outcome parameters [30-day and overall mortality, ICU and hospital length of stay (LOS), duration of mechanical ventilation (MV) and incidence of postoperative delirium], as well participation in and duration of postoperative cardiac rehabilitation. Furthermore, factors influencing postoperative improvement of frailty were investigated. Statistical analysis The sample size was planned for a comparison of pre- and post-surgery CFS using a paired t-test. As minimal relevant difference we considered a mean improvement by 0.5 points. A difference of 1 point on the Clinical Frailty Scale (CFS) represents a meaningful change in a patient’s functional status, as each category reflects distinct levels of mobility and independence. We therefore considered that smaller average changes may still be clinically relevant at the group level. Accordingly, we defined a minimal clinically important difference of 0.5 points as a pragmatic threshold. This reflects a scenario in which a substantial proportion of patients experience a meaningful improvement (e.g., a 1-point reduction in CFS), while others remain stable. For example, an average change of 0.5 points could correspond to 50% of patients improving by one CFS level and 50% showing no change. Given the absence of established thresholds for minimal clinically important differences for the CFS, this definition was chosen as a clinically plausible and conservative estimate. Under a conservative scenario with 50% CFS values remaining stable and 50% of CFS values changing by not more than 2 points, the standard deviation of the difference would be at most 1.41. Under these assumptions, a sample size of 64 patients was required to reject the null hypothesis of no mean pre- versus post-surgery difference at the 5% significance level with 80% power. To account for up to 15% dropouts, we aimed to include 80 patients. Basic descriptive statistics were used to describe baseline demographic patient data. Shapiro–Wilk test was used to test continuous data for normal distribution, according to which data was presented as mean and standard deviation, or median and interquartile range. Normally distributed variables were analysed using Student’s t-test, while non-normally distributed variables were analysed using Wilcoxon signed-rank test. Categorical data was expressed as counts and percentages and was analysed using the χ2-test. Furthermore, survival analysis was performed using a log rank test. To analyse factors influencing improvement of CFS, univariate logistic regression analysis was performed. Factors showing a significant effect were subsequently included into a multivariate logistic regression analysis. To investigate factors associated with the change in frailty, a linear regression analysis was performed using the change in CFS between baseline and 12-month follow-up as dependent variable. The level of statistical significance was set at a P-value of < 0.05. Statistical analysis and creation of plots were performed using the R Project for Statistical Computing version 4.3.1 (Copyright (C) 2023 The R Foundation for Statistical Computing; http://www.R-project.org/), including the following packages and libraries: compareGroups, ggpubr, alluvial, RColorBrewer, survival, survminer and dplyr. Results General information Eighty-eight individuals undergoing elective cardiac surgery were screened for eligibility for the study. One patient was excluded due to a language barrier interfering with the planned follow-up via telephone, while seven patients declined to participate. Consequently, 80 patients were included in the study. Due to four dropouts (three not operated within 10 days of enrollment, one not operated at all), 76 patients were included in the final analysis. During follow-up, eight individuals passed away and another eight were lost to follow-up. Therefore, a complete follow-up is available in 60 cases (Fig. 1).Fig. 1Strobe diagram of recruitment process Strobe diagram of recruitment process Mean age was 74.3 ± 7.3 years, and 28 (36.8%) individuals were female. The median CFS was 4 [IQR 3–4]. Using a cutoff of CFS ≥ 4, 44 (57.9%) individuals were classified as frail. Frail patients showed significantly higher EuroSCORE II levels (1.39% [IQR 0.96–2.46] vs. 2.82% [IQR 1.72–5.59], P = 0.001), as well as lower haemoglobin, lower albumin and higher proBNP levels (13.7 g/dl ± 1.59 vs. 12.4 g/dl ± 1.74, P = 0.001; 43.6 g/l [IQR 42.1–45.0] vs. 42.0 g/l [IQR 39.2–44.0], P = 0.020; 470 pg/ml [IQR 206–1126] vs. 1818 pg/ml [IQR 339–2722], P = 0.008, respectively). Furthermore, significantly more frail patients underwent interventional procedures without the use of a CPB (17 [38.6%] vs. 1 [3.12%], P = 0.001). Detailed information can be found in Table 1.Table 1Baseline demographicsNon-Frail (N = 32)Frail (N = 44)P-ValueAge (y)72.8 ± 6.8675.4 ± 7.510.123Female12 (37.5%)16 (36.4%)1.000Male20 (62.5%)28 (63.6%)BMI (kg/m2)25.9 [24.0;27.6]26.5 [25.0;28.5]0.328EuroSCORE II (%)1.39 [0.96;2.46]2.82 [1.72;5.59]0.001Baseline Laboratory Results Hemoglobin (g/dl)13.7 ± 1.5912.4 ± 1.740.001 Creatinine (mg/dl)0.98 [0.73;1.11]1.06 [0.88;1.28]0.072 Albumin (g/l)43.6 [42.1;45.0]42.0 [39.2;44.0]0.020 proBNP (pg/ml)470 [206;1126]1818 [339;2722]0.008Preoperative medical care and medication Recent doctor’s visits (n)4.5 [2.75;8.0]10.0 [4.75;15.0]0.016 Recent hospitalization (yes)23 (71.9%)32 (72.7%)1.000 Long-term medication (n)5.5 [4.0;27.25]9.9 [6.75;11.0]< 0.001Type of surgery CABG10 (31.2%)4 (9.09%)0.031 Valve15 (46.9%)15 (34.1%)0.374 Combined5 (15.6%)8 (18.2%)1.000 Interventional1 (3.12%)17 (38.6%)0.001 Other1 (3.12%)0 (0.00%)0.421Surgical Access Sternotomy25 (78.1%)23 (52.3%)0.039 Hemisternotomy5 (15.6%)3 (6.82%)0.270 Thoracotomy1 (3.12%)1 (2.27%)1.000 Transapical0 (0.00%)1 (2.27%)1.000 Femoral1 (3.12%)16 (36.4%)0.002Surgical characteristics Surgery using CPB31 (96.9%)27 (61.4%)0.001 Reoperation0 (0.0%)10 (22.7%)0.004 Anesthesia Duration (min)302 [274;380]302 [97.5;404]0.388 Surgery Duration (min)236 [213;314]232 [70.5;327]0.330 CPB (min)122 [89;136]143 [110;198]0.032 AoCC (min)73 [59;95]84 [72.5;115]0.239Values presented as number (%), median [IQR], or mean ± SD; after conducting a Shapiro–Wilk normality test, p-values were calculated using t-test for normally distributed variables and Wilcoxon rank-sum test for non-normally distributed variables. Categorical variables were compared using the χ2-testAbbreviations: BMI Body Mass Index, EuroSCORE II European System for Cardiac Operative Risk II, proBNP pro-b-type natriuretic peptide, CABG coronary artery bypass graft, CPB cardiopulmonary bypass, Redo surgery after prior cardiac surgery, AoCC aortic cross-clamp Table 1Baseline demographicsNon-Frail (N = 32)Frail (N = 44)P-ValueAge (y)72.8 ± 6.8675.4 ± 7.510.123Female12 (37.5%)16 (36.4%)1.000Male20 (62.5%)28 (63.6%)BMI (kg/m2)25.9 [24.0;27.6]26.5 [25.0;28.5]0.328EuroSCORE II (%)1.39 [0.96;2.46]2.82 [1.72;5.59]0.001Baseline Laboratory Results Hemoglobin (g/dl)13.7 ± 1.5912.4 ± 1.740.001 Creatinine (mg/dl)0.98 [0.73;1.11]1.06 [0.88;1.28]0.072 Albumin (g/l)43.6 [42.1;45.0]42.0 [39.2;44.0]0.020 proBNP (pg/ml)470 [206;1126]1818 [339;2722]0.008Preoperative medical care and medication Recent doctor’s visits (n)4.5 [2.75;8.0]10.0 [4.75;15.0]0.016 Recent hospitalization (yes)23 (71.9%)32 (72.7%)1.000 Long-term medication (n)5.5 [4.0;27.25]9.9 [6.75;11.0]< 0.001Type of surgery CABG10 (31.2%)4 (9.09%)0.031 Valve15 (46.9%)15 (34.1%)0.374 Combined5 (15.6%)8 (18.2%)1.000 Interventional1 (3.12%)17 (38.6%)0.001 Other1 (3.12%)0 (0.00%)0.421Surgical Access Sternotomy25 (78.1%)23 (52.3%)0.039 Hemisternotomy5 (15.6%)3 (6.82%)0.270 Thoracotomy1 (3.12%)1 (2.27%)1.000 Transapical0 (0.00%)1 (2.27%)1.000 Femoral1 (3.12%)16 (36.4%)0.002Surgical characteristics Surgery using CPB31 (96.9%)27 (61.4%)0.001 Reoperation0 (0.0%)10 (22.7%)0.004 Anesthesia Duration (min)302 [274;380]302 [97.5;404]0.388 Surgery Duration (min)236 [213;314]232 [70.5;327]0.330 CPB (min)122 [89;136]143 [110;198]0.032 AoCC (min)73 [59;95]84 [72.5;115]0.239Values presented as number (%), median [IQR], or mean ± SD; after conducting a Shapiro–Wilk normality test, p-values were calculated using t-test for normally distributed variables and Wilcoxon rank-sum test for non-normally distributed variables. Categorical variables were compared using the χ2-testAbbreviations: BMI Body Mass Index, EuroSCORE II European System for Cardiac Operative Risk II, proBNP pro-b-type natriuretic peptide, CABG coronary artery bypass graft, CPB cardiopulmonary bypass, Redo surgery after prior cardiac surgery, AoCC aortic cross-clamp Baseline demographics Values presented as number (%), median [IQR], or mean ± SD; after conducting a Shapiro–Wilk normality test, p-values were calculated using t-test for normally distributed variables and Wilcoxon rank-sum test for non-normally distributed variables. Categorical variables were compared using the χ2-test Abbreviations: BMI Body Mass Index, EuroSCORE II European System for Cardiac Operative Risk II, proBNP pro-b-type natriuretic peptide, CABG coronary artery bypass graft, CPB cardiopulmonary bypass, Redo surgery after prior cardiac surgery, AoCC aortic cross-clamp Postoperative change in frailty At baseline, 44 (57.9%) patients were classified as frail, and 32 (42.1%) as non-frail. One year later, 21 (27.6%) patients were frail and 39 (51.3%) non-frail. Eight patients (10.5%) passed away and another eight patients were lost to follow-up. We found a significant change of CFS at all time points compared to baseline. While frailty was significantly worse at hospital discharge (CFS 4 [IQR 3–4] vs. CFS 4 [IQR 3–5], P < 0.001), follow-up at 6 and 12 months showed a significant improvement (CFS 4 [IQR 3–4] vs. CFS 3 [IQR 2–4], P < 0.001 and CFS 4 [IQR 3–4] vs. CFS 3 [IQR 2–4], P < 0.001), respectively) (Fig. 2a). Differences in frailty trajectory could be seen according to baseline frailty status. While both non-frail and frail patients showed a significant improvement after 6 months (CFS 3 [IQR 2.75–3] vs. CFS 2.5 [IQR 2–3], P = 0.035 and CFS 4 [IQR 4–6] vs. CFS 4 [IQR 3–5], P < 0.001, respectively) and 12 months (CFS 3 [IQR 2.75–3] vs. CFS 2 [IQR 2–3], P = 0.033 and CFS 4 [IQR 4–6] vs. CFS 4 [IQR 3–5], P < 0.001, respectively), only non-frail patients showed a significant increase in CFS at hospital discharge (CFS 3 [IQR 2.75–3] vs. CFS 3 [IQR 3–4], P < 0.001 and CFS 4 [IQR 4–6] vs. CFS 5 [IQR 4–6], P = 0.073, respectively) (Fig. 2b).Fig. 2a Frailty status according to CFS at baseline, hospital discharge, 6 months and 12 months after surgery. *: P < 0.001 compared to baseline [38]; b Frailty distribution at each time point, grouped according to baseline frailty status. *: P < 0.05; **: P < 0.001 compared to baseline a Frailty status according to CFS at baseline, hospital discharge, 6 months and 12 months after surgery. *: P < 0.001 compared to baseline [38]; b Frailty distribution at each time point, grouped according to baseline frailty status. *: P < 0.05; **: P < 0.001 compared to baseline Transitions between levels of frailty Frailty improved in 27 (35.5%) participants, of which 11 (14.5%) individuals initially presented as non-frail, and 16 (21.1%) as frail. Of these 16 frail patients, 13 (17.1%) managed to switch to the non-frail cohort. Frailty increased in four (5.2%) participants (3 non-frail, 1 frail). Two (2.6%) non-frail individuals became frail during the observation period. 29 (38.2%) patients (15 frail, 14 non-frail at baseline) remained stable in their frailty status (Fig. 3). No significant differences could be seen in frailty dynamic according to baseline frailty status.Fig. 3Alluvial plot showing transition between levels of frailty on the CFS from baseline to 12 months after surgery; LTFU: Lost to Follow-up Alluvial plot showing transition between levels of frailty on the CFS from baseline to 12 months after surgery; LTFU: Lost to Follow-up Surgical outcome parameters No study participant in either the frail or non-frail cohort passed away within 30 days of the surgery. One (3.12%) non-frail and seven (15.9%) frail patients died within 12 months after surgery (P = 0.071). Among patients undergoing open heart surgery, one (3.12%) non-frail and three (11.1%) frail patients died. (P = 0.230), while four (23.5%) frail and zero non-frail patients died following interventional procedures (P = 0.611). Overall, mortality occurred predominantly within the first three months of follow-up. Among these patients, one showed improvement in frailty, four worsened, one remained unchanged, and two were not discharged and therefore not assessable longitudinally. Furthermore, frail patients had longer ICU stays following open heart surgery than non-frail patients. (2 days [1;5.5] vs. 1 day [1;2]; P = 0.047). Notably, significantly more non-frail individuals had completed a cardiac rehabilitation program one year after surgery [24 (75.0%) vs. 18 (40.9%); P = 0.007]. Functional status and quality of life Compared to baseline, FIM scores were significantly worse at hospital discharge (125 [IQR 123.8–126] vs. 124 [IQR 120–126], P = 0.017) and significantly better 6 months later (125 [IQR 123.8–126] vs. 126 [IQR 125–126], P = 0.001). After 12 months, no significant differences in FIM scores could be shown (125 [IQR 123.8–126] vs. 126 [IQR 124.8–126], P = 0.212). SF-12 scores are separately calculated as a physical component summary (PCS) and a mental component summary (MCS). The PCS remained stable at hospital discharge (39.15 [IQR 30.14–48.34] vs. 37.32 [IQR 30.68–44.24], P = 0.259), but improved significantly after 6 and 12 months (39.15 [IQR 30.14–48.34] vs. 48.79 [IQR 41.30–53.43], P < 0.001 and (39.15 [IQR 30.14–48.34] vs. 49.86 [IQR 40.54–54.40], P < 0.001, respectively). No significant change in MCS could be shown over the course of the follow-up (52.83 [IQR 47.42–58.39] vs. 55.70 [IQR 46.58–61.11] at discharge, P = 0.131; 55.85 [IQR 51.20–58.06], at 6 months, P = 0.165; and 54.81 [IQR 49.40–58.10], at 12 months, P = 0.180). Improvement in PCS at 12 months was significantly associated with the stay at rehabilitation facility, while improvement in MCS was associated with higher grip strength in the multivariate regression analysis (OR 9.56, CI [2.99–37.64], P < 0.001; and OR 1.09, CI [1.01–1.19], P = 0.041; Supplemental Table 2 and 3). Regression analysis for improvement of frailty To investigate which factors influence frailty improvement, a logistic regression analysis was performed (Table 2). In the univariate analysis, significant influences could be shown for the EFT-Score (OR 0.54, CI [0.32–0.86], P = 0.017), as well participation in and duration of cardiac rehabilitation (OR 6.38, CI [2.20–21.65] P = 0.001 and OR 1.48, CI [1.05–2.32] P = 0.043, respectively). Significant factors were consequently analysed in a multivariate logistic regression model. The variable “stay at a rehabilitation facility” was excluded, as it represents a dichotomized form of “rehabilitation facility length of stay (LOS)” and an inclusion of both variables would introduce multicollinearity and reduce the reliability of effect estimates. Therefore, the more informative continuous variable was used. The number of weeks spent at a rehabilitation facility were the only factor that persisted to have a significant influence in the multivariate regression analysis (OR 1.51, CI [1.05–2.43], P = 0.047).Table 2Logistic regression analysis for improvement of frailtyLogistic Regression AnalysisOdds RatioCI (95%)P-ValueUnivariate Analysis CFS (baseline)0.910.62–1.340.646 Frailty1.090.42–2.880.858 Age (y)0.980.92–1.050.593 Age ≥ 80 years0.430.11–1.390.184 Sex (male)2.140.79–6.320.147 BMI (kg/m2)1.010.91–1.130.795 EuroSCORE II0.940.81–1.060.346 Hemoglobin (g/dl)1.150.88–1.530.309 Creatinine (mg/dl)0.750.20–2.380.693 Albumin (g/l)1.140.99–1.350.093 proBNP (pg/ml)0.990.99–1.000.522 Recent doctor’s visits (n)1.020.98–1.070.401 Recent hospitalization (yes)0.650.23–1.850.411 Long-term medication (n)0.920.80–1.040.179 CABG1.010.28–3.310.987 Valve1.740.67–4.600.253 Combined1.160.32–3.930.808 Interventional0.430.11–1.390.184 OtherNANA0.992 Sternotomy1.270.48–3.490.631 Hemisternotomy1.100.21–4.880.902 ThoracotomyNANA0.992 TransapicalNANA0.992 Femoral0.480.12–1.550.247 Surgery using CPB2.300.72–8.900.184 Reoperation1.250.29–4.820.751 Anaesthesia Duration (min)1.000.99–1.000.391 Surgery Duration (min)1.000.99–1.000.361 CPB (min)1.000.99–1.010.371 AoCC (min)1.000.99–1.020.189 Grip Strength dominant Hand (kg; baseline)1.040.99–1.090.161 FTSST (sec; baseline)1.050.99–1.140.151 FIM (points; baseline)1.081.00–1.220.094 EFT (points; baseline)0.540.32–0.860.017 SF-12 PCS (points; baseline)1.000.96–1.040.868 SF-12 MCS (points; baseline)1.050.99–1.110.105 Stay at rehabilitation facility6.382.20–21.650.001 Rehabilitation facility LOS (weeks)1.481.05–2.320.043 Hospital LOS (days)1.000.96–1.030.906 ICU LOS (days)1.010.95–1.070.732 Duration MV (h)1.000.99–1.010.394 Postoperative Delirium1.200.15–7.710.847Multivariate Analysis Rehabilitation facility LOS (weeks)1.511.05–2.430.048 EFT (points; baseline)0.630.35–1.050.086Univariate and multivariate logistic regression analysis for improvement of frailty. Cells labelled “NA” had too few observations and did not yield proper odds ratios and confidence intervalsAbbreviations: AoCC aortic cross-clamp, BMI Body Mass Index, CFS Clinical Frailty Scale, CPB cardiopulmonary bypass, EFT essential frailty toolset, EuroSCORE II European System for Cardiac Operative Risk II, FIM functional independence measure, FTSST five-time-sit-to-stand-test, ICU intensive care unit, LOS length of stay, MV mechanical ventilation, proBNP pro-b-type natriuretic peptide, SF12-PCS Short Form 12 Physical Componen Summary, SF-12 MCS Short Form 12 Mental Component Summary Table 2Logistic regression analysis for improvement of frailtyLogistic Regression AnalysisOdds RatioCI (95%)P-ValueUnivariate Analysis CFS (baseline)0.910.62–1.340.646 Frailty1.090.42–2.880.858 Age (y)0.980.92–1.050.593 Age ≥ 80 years0.430.11–1.390.184 Sex (male)2.140.79–6.320.147 BMI (kg/m2)1.010.91–1.130.795 EuroSCORE II0.940.81–1.060.346 Hemoglobin (g/dl)1.150.88–1.530.309 Creatinine (mg/dl)0.750.20–2.380.693 Albumin (g/l)1.140.99–1.350.093 proBNP (pg/ml)0.990.99–1.000.522 Recent doctor’s visits (n)1.020.98–1.070.401 Recent hospitalization (yes)0.650.23–1.850.411 Long-term medication (n)0.920.80–1.040.179 CABG1.010.28–3.310.987 Valve1.740.67–4.600.253 Combined1.160.32–3.930.808 Interventional0.430.11–1.390.184 OtherNANA0.992 Sternotomy1.270.48–3.490.631 Hemisternotomy1.100.21–4.880.902 ThoracotomyNANA0.992 TransapicalNANA0.992 Femoral0.480.12–1.550.247 Surgery using CPB2.300.72–8.900.184 Reoperation1.250.29–4.820.751 Anaesthesia Duration (min)1.000.99–1.000.391 Surgery Duration (min)1.000.99–1.000.361 CPB (min)1.000.99–1.010.371 AoCC (min)1.000.99–1.020.189 Grip Strength dominant Hand (kg; baseline)1.040.99–1.090.161 FTSST (sec; baseline)1.050.99–1.140.151 FIM (points; baseline)1.081.00–1.220.094 EFT (points; baseline)0.540.32–0.860.017 SF-12 PCS (points; baseline)1.000.96–1.040.868 SF-12 MCS (points; baseline)1.050.99–1.110.105 Stay at rehabilitation facility6.382.20–21.650.001 Rehabilitation facility LOS (weeks)1.481.05–2.320.043 Hospital LOS (days)1.000.96–1.030.906 ICU LOS (days)1.010.95–1.070.732 Duration MV (h)1.000.99–1.010.394 Postoperative Delirium1.200.15–7.710.847Multivariate Analysis Rehabilitation facility LOS (weeks)1.511.05–2.430.048 EFT (points; baseline)0.630.35–1.050.086Univariate and multivariate logistic regression analysis for improvement of frailty. Cells labelled “NA” had too few observations and did not yield proper odds ratios and confidence intervalsAbbreviations: AoCC aortic cross-clamp, BMI Body Mass Index, CFS Clinical Frailty Scale, CPB cardiopulmonary bypass, EFT essential frailty toolset, EuroSCORE II European System for Cardiac Operative Risk II, FIM functional independence measure, FTSST five-time-sit-to-stand-test, ICU intensive care unit, LOS length of stay, MV mechanical ventilation, proBNP pro-b-type natriuretic peptide, SF12-PCS Short Form 12 Physical Componen Summary, SF-12 MCS Short Form 12 Mental Component Summary Logistic regression analysis for improvement of frailty Univariate and multivariate logistic regression analysis for improvement of frailty. Cells labelled “NA” had too few observations and did not yield proper odds ratios and confidence intervals Abbreviations: AoCC aortic cross-clamp, BMI Body Mass Index, CFS Clinical Frailty Scale, CPB cardiopulmonary bypass, EFT essential frailty toolset, EuroSCORE II European System for Cardiac Operative Risk II, FIM functional independence measure, FTSST five-time-sit-to-stand-test, ICU intensive care unit, LOS length of stay, MV mechanical ventilation, proBNP pro-b-type natriuretic peptide, SF12-PCS Short Form 12 Physical Componen Summary, SF-12 MCS Short Form 12 Mental Component Summary Similar to the logistic regression results, both participation in and duration of cardiac rehabilitation were found to have a significant influence on change in frailty in the linear regression model (β = 0.41, 95% CI [0.07–0.76], P = 0.021; and β = 0.13 per week, 95% CI [0.03–0.23], P = 0.01, respectively; Supplemental Table 1). Subgroup analysis of change in frailty according to completion of rehabilitation To further assess the influence of cardiac rehabilitation on frailty improvement, a subgroup analysis was conducted. Patients who underwent cardiac rehabilitation showed a significant improvement in CFS levels 12 months postoperatively (CFS 3 [IQR 3–4] vs. CFS 3 [IQR 2–3], P < 0.001). Patients who did not undergo rehabilitation showed no significant improvement of CFS 12 months postoperatively (CFS 4 [IQR 4–5] vs. CFS 4 [IQR 3–4], P = 0.480). Most patients (n = 69, 91%) were discharged home prior to starting cardiac rehabilitation, which generally began approximately two months after hospital discharge. Seven patients (9%) were not discharged directly home: five were transferred to other hospitals and two to geriatric or remobilization facilities. One transferred patient died shortly after discharge. All remaining patients completed cardiac rehabilitation programs. Subgroup analysis of change in frailty according to performed procedures Separate analyses were performed for patients undergoing open-heart surgery and catheter-based interventions. Frailty increased significantly in patients undergoing open-heart surgery at hospital discharge, but not in patients undergoing interventional procedures (CFS 3 [IQR 3–4] vs. CFS 4 [IQR 3–4], P < 0.001 and CFS 5 [IQR 4–6] vs. CFS 5.5 [IQR 4–6], P = 0.103, respectively). In patients undergoing open-heart surgery, frailty improved significantly at 6 months (CFS 3 [IQR 3–4] vs. 3 [IQR 2–4], P < 0.001) and 12 months postoperatively (P < 0.001). Patients undergoing interventional procedures also demonstrated significant improvement at 6 months (CFS 5 [IQR 4–6] vs. 4.5 [IQR 3–6], P = 0.025) and 12 months (CFS 5 [IQR 4–6] vs. 5 [IQR 3–5.5], P = 0.045). In subgroup-specific regression analyses, cardiac rehabilitation was significantly associated with frailty improvement only in patients undergoing open-heart surgery, but the number of observations in the interventional subgroup may have been too small to detect significant associations. Discussion The findings of this study have shown that frailty is a dynamic condition and changes in level of frailty are possible in both directions. Cardiac surgery, furthermore, has the potential to reverse frailty in individuals suffering from cardiovascular disease. Additionally, this study revealed the strong effect of rehabilitation on postoperative physical improvement. To date, only two studies have been published examining the effect of cardiac surgery on postoperative frailty [18, 19]. A significant reduction of frailty could be shown in both cases. Relatively short follow-up periods of three to six months, however, give little insight into long-term outcomes and persistence of frailty improvement. Our study has shown a significant reduction of frailty at six and twelve months after cardiac surgery. More than a third of patients managed to improve their frailty level, with a notable share of frail patients presenting as non-frail one year after surgery. Furthermore, we could show that improvement of frailty was likely irrespective of baseline frailty status or whether patients had undergone open or interventional cardiac procedures. With the rapid aging of the surgical patient population, the number of frail individuals undergoing cardiac interventions are increasing. While the majority of patients report a postoperative improvement in quality of life [20, 21], up to 19% of elderly patients regret the decision to undergo cardiac surgery [22]. With the lack of optimal prediction tools to identify patients likely to improve, factors influencing postoperative health trajectories need to be investigated. To our knowledge, this is the first study to investigate factors influencing the postoperative reduction of frailty. Our results showed that cardiac rehabilitation had the strongest effect on improvement of frailty, with a 51% in probability to reduce frailty with every further week spent at a rehabilitation facility. Other preoperative patient characteristics, such as baseline frailty status, age, sex or laboratory findings did not have an influence on frailty improvement in our sample, hereby further stressing the importance of cardiac rehabilitation. Frail patients, however, are generally less likely to participate in cardiac rehabilitation [23], which could also be seen in our sample: only 41% of frail patients participated in cardiac rehabilitation, as opposed to 75% of non-frail patients. Cardiac rehabilitation aims to prevent immobility and further decline. Still, worries regarding reduced exercise capacity and safety often prevent consistent referral to rehabilitation programs [24], even though a positive impact on frailty and sarcopenia has previously been proven [25, 26]. Apart from a potential reduction of frailty, rehabilitation programs can improve survival, as well as cardiac health and quality of life [27, 28]. In our study, an initial increase of frailty was observed at the time of hospital discharge. This is likely due to the invasive nature of most cardiac surgical procedures, which require long periods of convalescence, typically exceeding the duration of the hospital stay. Furthermore, our subgroup analysis only showed significant frailty improvement in patients who had undergone rehabilitation. This finding suggests that while cardiac surgery may provide the necessary physiological substrate for recovery, improvement of frailty appears to be primarily associated with participation in postoperative rehabilitation. This further stresses the need for transitional care, such as rehabilitation programs, to support patients from hospital discharge to an independent life at home [29]. Apart from rehabilitative measures, concepts of prehabilitation to counteract the negative effects of major surgery and increase preoperative physiological reserve are emerging [30]. Exercise, respiratory training, optimisation of nutritional status and lifestyle modifications to such as cessation of smoking or sleep improvement can also be of great benefit [31]. Prehabilitation have previously been shown to be safe and potentially accelerate recovery after surgery [32]. In addition, receptivity for postoperative rehabilitation in patients who underwent prehabilitation is substantial [33]. Our study further contributes to the recognition of patient-centred outcomes in cardiac surgery. With the advancement of treatment options and surgical techniques, as well as with the aging of our population, postoperative mortality after cardiac surgery is continuously decreasing [34]. Mortality alone, however, is not an adequate primary outcome parameter in the care of frail and elderly patients. Functional survival, symptom control and quality of life are often valued more than a mere extension of lifespan [35]. Therefore, medical care should ideally aim to improve, or at least the preserve, the present health status. Patients should furthermore be involved in the setting of treatment goals and expectations regarding potential outcomes should furthermore be managed accordingly, to promote informed consent and reduce decisional regret [36]. In any case, consistent postoperative care is of paramount importance and must be encouraged and facilitated wherever possible. Strengths and limitations First, this prospective study was carried out at a highly specialised tertiary cardiac surgery centre, which potentially limits generalizability of our results to smaller centres. Moreover, non-randomised prospective studies come with an inherent level of selection bias, due to the preselection to undergo cardiac surgery prior to inclusion into our study. In addition, without a non-surgical control group, surgery-related effects on frailty dynamics are impossible to distinguish from spontaneous changes in frailty in similarly aged individuals. While prior studies [37] have shown a survival benefit of cardiac surgery to a matched non-surgical population, especially in older patients, this does not allow conclusions regarding frailty trajectories and does not mitigate the above limitation. Second, inclusion of patients undergoing both open and interventional cardiac procedures resulted in a heterogeneous study population. As treatment allocation was not randomized a selection bias cannot be excluded. However, including both groups, provides additional insight into frailty distribution, its dynamics and postoperative outcomes and reflects real-world clinical decision-making. Third, follow-up via telephone at six and twelve months postoperatively is a potential limitation, since in-person interviews and physical tests are often of superior evaluative quality. However, the selection of questionnaires used were suitable to gain an adequate impression of the participants’ health status and have previously been validated for telephone interviews. Additionally, the CFS is a categorical tool and, while simple and widely used, may be limited in its ability to detect and interpret subtle changes in frailty over time. Fourth, since referral to cardiac rehabilitation is managed outside of our department, we had no influence or control over the referral process, potentially introducing selection bias. Lastly, the small sample size and the single-centre nature of the study represents important limitations. Despite accounting for up to 15% of dropouts, our study remains slightly underpowered due the number of patients either deceased or lost to follow-up. Furthermore, the limited number of events restricted the complexity of the multivariable analysis. Variable selection was therefore based on univariate significance, which may introduce bias and overlook relevant confounders. Conclusions Our study has shown a significant reduction of frailty status one year after cardiac surgery. These findings prove that cardiac surgery has the potential to reverse frailty and support the concept of frailty as a dynamic condition. Cardiac rehabilitation was shown to be the only benefactor in improvement of frailty in our sample. Therefore, awareness for the necessity of postoperative rehabilitation and supportive measures following discharge must be raised in caretakers and patients alike, especially in the care of frail and elderly patients undergoing cardiac surgery. Supplementary Information Supplementary Material 1. Supplementary Material 1.