The implications of frailty for the health care sector gain increasing attention in the context of ageing populations. Frailty is a complex clinical syndrome, which describes a decrease of physiological reserves that come along with an increased vulnerability to stressors [1]. Although frailty increases with aging, and is usually regarded as geriatric syndrome, it reflects a functional impairment beyond the chronological age [2]. For example, single severe organic disease or high cumulative comorbidity also contribute to the development of frailty [3].

Potential medical stressors where an increased vulnerability of frail people might become particularly relevant are surgery or interventional procedures. In patients undergoing percutaneous coronary intervention or cardiac surgery, frailty was associated with higher mortality, vascular and bleeding complications and renal failure [4‐7]. These procedural complications translated into increased hospitalization costs in frail patients [8, 9]. Transcatheter valve interventions such as transcatheter aortic valve implantation [10] or transcatheter mitral valve repair (TMVR) were developed to decrease medical stress for patients [11, 12]. As consequence, this means that frailty is a frequent clinical reason to decide for a transcatheter approach and against surgical valve therapy and almost every second patient undergoing TAVI or TMVR is frail [13, 14].

So far it is unclear whether “low stress” procedures like TMVR have an impact on the complication rate, recovery from complications and resulting treatment costs in highly vulnerable patients. We have recently demonstrated that the risk of major cardiac or vascular complications was not increased in frail compared to non-frail patients undergoing TMVR [15]. However, major complications are overall very rare after TMVR and frailty might affect health care utilization through minor procedural or procedure independent complications following TMVR. The knowledge of determinants of treatment costs and underlying causes is of major relevance for all health care stakeholders. The aim of this study was to examine the association of frailty status with hospital related costs in consecutive patients undergoing TMVR with MitraClip at our institution.

Consecutive adult patients undergoing TMVR using the MitraClip system (Abbott Vascular, Santa Clara, California, USA) at our institution between May 2014 and October 2016 were eligible for this retrospective analysis of prospectively assessed data on frailty in patients undergoing TMVR. Treatment decision for TMVR with MitraClip was taken by a multidisciplinary heart team after preoperative risk assessment based on objective risk scores, clinical parameters and morphological suitability for MitraClip according to current guidelines [16]. All patients underwent TMVR in general anesthesia. Extubation in the cath lab is pursued in every patient. By default, all patients are transferred to the cardiology intermediate care unit (IMC) for at least 24 h after the procedure. If medically indicated, monitoring on the IMC can be extended to more than 24 h, and patients could be transferred to the intensive care unit (ICU) post-procedurally if extubation was unsuccessful or in the case of severe haemodynamic compromise. Written informed consent for participation was obtained from all patients included in this study. The study was approved by the local ethics committee of the University of Cologne (reference 14-116, German Registry of Clinical Trials registration DRKS00006194).

Frailty status using the Fried criteria was prospectively assessed before the MitraClip procedure as reported earlier [15]. Briefly, the five components of the frailty syndrome (unintentional weight loss, weakness, exhaustion, slowness and low physical activity) were assessed, and 1 point was scored for each criterion met to specification. Patients meeting at least 3 of the 5 criteria were classified as frail. All treating physicians were blinded to the frailty status. For confirmatory analysis we used a second definition of frailty based on disability in Instrumental Activities of Daily Life (IADL) as recommended by the Mitral Valve Academic Research Consortium [17]. Disabilities in IADL have been shown to correlate with physical frailty in elderly people [18, 19]. Since a validated threshold of IADL for frailty definition is lacking, we pragmatically defined a frail status by the group median of the IADL score which corresponds to a disability in more than two IADL domains. Prespecified baseline clinical characteristics were extracted from the patient record. The reason for a prolonged IMC/ICU stay which we defined as more than two days was assessed retrospectively from records using predefined categories: (a) postprocedural minor or major bleeding according to MVARC (Mitral Valve Academic Research Consortium); (b) an acute kidney injury requiring haemodialysis; (c) an acute infection with antibiotic treatment; (d) haemodynamic instability which required inotropes or vasopressors, or arrhythmias leading to haemodynamic compromise; (e) respiratory failure with need of mechanical or non-invasive ventilation and (f) in absence of medical reasons logistic causes, meaning no vacancy to transfer the patient to the normal ward. All cost and reimbursement data were provided by the department of controlling of the University Hospital of Cologne. Total hospital costs were calculated as the sum of all incurred costs during hospital stay for elective TMVR from the day of hospital admission to discharge, including personnel expenses, material costs and costs for infrastructure. Personnel expenses for physician and nursing services, overhead costs for medication, other medical supply and medical and not-medical infrastructure were time based allocated to a patient. Revenues were calculated according to the German Diagnosis Related Groups (G-DRG) system on a case-based lump sum.

Mean and standard deviation or median and interquartile range (where data were not normally distributed) was used for continuous variables and frequency with percentage for categorical variables. t test (and Welch’s t test in case of unequal variances) or Mann–Whitney U test or Chi²-test was used for comparison between groups. All cost variables showed a highly skewed distribution and were described as median and interquartile range. However, we also present the difference in mean costs of the two groups which is interesting from the economic point of view. Since cost distribution was skewed, reciprocal transformation was used to achieve normal distribution. Univariable and multivariable linear regression analyses were applied to identify significant preoperative predictors of total costs minus implant costs. The statistical analyses were performed using IBM® SPSS® Statistics 25 (SPSS Inc., Chicago, Illinois, USA).

Frailty is a leading cause for high surgical risk and the decision for a TMVR approach in patients with mitral regurgitation. Although about half of patients undergoing TMVR are frail [14, 15, 20], data on the impact of frailty on procedural health care utilization is lacking. Mean total costs without implant expenses were 32% higher in frail patients, and the association between frailty and costs persisted after excluding high cost outliers and was consistent across different definitions of frailty. Importantly, frailty was the only significant predictor of total costs when considering baseline patient characteristics including sex, age, comorbidities, NYHA functional class and history of previous cardiac surgery. The increase in hospital costs in frail patients was primarily the result of a prolonged stay on the IMC/ICU caused by minor medical complications.

Our data extend existing evidence on the role of frailty for adverse outcomes in various health care settings. A body of evidence exists on the effect of frailty on prolonged recovery time after surgical procedures; in particular, a higher rate of institutionalization, increased mortality [21, 22] and the disproportionally high health care utilization associated with frailty in industrialized countries overall [23, 24]. Here, we show the necessity of prolonged postprocedural IMC/ICU and total hospital stay and associated increase in costs in the TMVR population. TMVR is a low stress intervention particularly developed for frail patients. In the setting of transcatheter aortic valve implantation [10], which is another minimal invasive procedure, frailty was a cost driver due to a prolonged length of stay, a high rate of complications and institutional discharge [25]. However, these results cannot be directly translated to the setting of TMVR. First, the burden of heart failure is higher in the TMVR population. Patients with severe aortic stenosis undergo a TAVI soon after symptoms arise, whereas mitral regurgitation (in particular secondary MR) is commonly associated with left ventricular failure and consequently these patients receive appropriate treatment (TMVR) after having already suffered from long-term symptomatic heart failure. Second, procedural complications are rare in the setting of TMVR and most often are not critical, whereas TAVI has the potential of early life-limiting complications, which occur more frequently in frail patients [26].

We have shown recently that MitraClip procedure shows similar technical and procedural efficacy in frail and non-frail patients [15]. Major cardiac structural and vascular access site complications were also not different. So, what might be the reason for the increased length of stay on IMC/ICU and regular wards? Frail patients were older, had a worse NYHA class and were generally sicker as reflected by the higher EuroScore II. However, frailty was the only predictor of hospital costs and comorbidities were not. A detailed analysis of underlying reasons for the prolonged hospital stay is beyond the scope of this study and partly impossible in a retrospective design since total hospital stay is the result of a complex interplay between medial, social, functional and logistic determinants. However, frail patients caused higher costs for laboratory services and physiotherapy, which are based on direct patient level data. Although these cost centre groups account for only a small absolute cost difference, these findings suggest that frail patients for example need more support to regain mobility/physical capabilities than non-frail patients. In support of the latter, frail patients have substantially worse disease independent measures of physical functioning both before and after TMVR than non-frail patients [15]. Furthermore, our analysis of medical causes of prolonged IMC/ICU stay demonstrated more minor complications or complications not directly associated with the TMVR procedure. Clinically severe complications such as haemodynamic compromise and dialysis might be generally more frequent in frail patients due to the higher baseline cardiac and extracardiac morbidity. For instance, baseline glomerular filtration rate was lower in frail patients which increases the risk of postprocedural renal failure. Prolonged respiratory failure and mechanical ventilation can be also explained by reduced cardiopulmonary reserve in frail patients but might also be attributable to muscle weakness and sarcopenia which are typical characteristics in frail patients [21]. We do not know whether the incidence of complications such as minor bleeding or infections overall are higher, or whether these complications are only more relevant for the clinical course of frail compared to non-frail patients. Such complications are usually only documented if clinically relevant or manifest and as such could not be systematically assessed in our retrospective analysis. However, considering the pathophysiological concept of frailty as increased vulnerability to stressors it seems plausible that mild bleeding anemia or inflammatory response might contribute to delayed recovery of frail patients and in consequence to longer hospital stay and higher costs [27].

Our findings have several important implications. First, from a methodological point of view cost analysis provides complementary data of patient outcomes to primary medical data. This is of major interest since frequency of complications might not differ in frail and non-frail patients but might have different impact on outcome and health care utilization. For instance, in patients undergoing TAVI there was a significant interaction between frailty status and access site on clinical outcome demonstrating that increased stress of transapical access was only of prognostic relevance in vulnerable, frail patients [28]. Second, frailty has a substantial impact on reimbursement of TMVR patients. Although on average the revenues exceeded the costs in frail and non-frail patients, the reimbursement was not covering costs in only 7% of non-frail patients and in 22% of frail patients. This knowledge is valuable for a hospital’s total cost calculation and resource allocation planning. Given the expected increase of frailty in the aging population, it should be considered to integrate frailty as a diagnosis determining case complexity and hence magnitude of revenue within the DRG based health care system. Our findings must not be interpreted as to deny TMVR in frail patients because of increased costs. In contrast, we observed a more pronounced symptomatic benefit of frail compared to non-frail patients after TMVR indicating a similar net cost-effectiveness. Third, physical frailty is regarded a potentially reversible health state. As such, the implementation of a frailty screening prior intervention and a tailored approach to disabilities of frail patients with early interventions during or even prior to the hospital stay may reduce length of stay and total costs in frail patients [29]. A positive impact of physical exercise, nutritional supplement and cognitive training on functional outcome has been reported, and geriatric-specific care protocols are discussed to shorten hospital length of stay in frail patients [30]. As TMVR compared to medical therapy alone has shown to result in a lower rate of hospitalization for heart failure and lower all-cause mortality within 24 months of follow-up in highly selected patients [31], further research should evaluate the cost-effectiveness or cost-utility of TMVR with regard to frailty.

Frailty is associated with a mean 32% increase of hospital costs in patients undergoing TMVR, which is primarily the result of a prolonged recovery and increased vulnerability to complications. These findings are valuable for a hospital’s total cost calculation and resource allocation planning. Since frailty is regarded a potentially reversible health state, preventive interventions may help reduce costs in frail patients.