Background
Frailty is a phenotype of decreased physiologic reserve resulting in a vulnerability to adverse health outcomes upon confrontation by stressors. The phenotype was originally defined and validated in older populations by Fried et al. as the presence of at least 3 of 5 components: exhaustion, unintentional weight loss, low physical activity, slow walking speed, and poor grip strength [
1]. Among older adults, frailty prevalence ranges from 7 to 15% and is associated with increased risk of mortality, incident disability, hospitalizations, and falls [
1‐
5]. Among end stage renal disease (ESRD) patients, the prevalence of the Fried frailty phenotype is 5 to 7-fold higher than in community-dwelling older adults [
6] and is associated with higher mortality, more hospitalizations, more falls, poor cognitive function, and poor health-related quality of life [
6‐
9]. Among ESRD patients who undergo kidney transplantation (KT), the Fried frailty phenotype is associated with worse health related quality of life and physical function at the time of KT and longer post-KT hospital stay, delayed graft function, early hospital readmission, immunosuppression intolerance, and increased mortality [
6,
9‐
15].
Despite the importance of the Fried frailty phenotype in risk stratification of clinical outcomes in ESRD patients, its predictive ability is moderate (area under the receiver operative curve = 0.70; c statistic = 0.75) [
11,
12] suggesting that the phenotype may be missing some valuable information for ESRD patients. Because the Fried frailty phenotype was developed in community-dwelling older adults, some components may not be applicable to ESRD patients, and there may be aspects of physiologic reserve in ESRD patients that are not characterized by this phenotype. Supporting this hypothesis, we have previously found that not all 5 components contribute to the value of the Fried phenotype in predicting mortality risk among KT recipients [
16].
A more accurate measure of frailty among ESRD patients will improve risk prediction, aid in clinical decision-making, develop targeted patient counseling, and identify interventions to improve physiologic reserve. Both ESRD patient and clinician opinions are needed to identify which of the 5 Fried components are mechanistically and biologically plausible, and which ESRD-specific components should be added or substituted in the phenotypic description of frailty to improve characterization of physiologic reserve in this population.
The primary goal of this study was to ascertain the opinions of experts who care for patients with ESRD and of ESRD patients pursuing kidney transplantation about the usefulness of the Fried frailty phenotype, and which novel components may further characterize frailty among ESRD patients. Additionally, we sought to elicit the opinions of these clinicians and patients on the effectiveness of and patient interest in interventions to improve frailty in ESRD and transplant patients. The Delphi method [
17], an iterative survey design that builds consensus among experts through a series of structured surveys, has been used to elicit opinions on frailty among geriatrics experts [
18]. In this study, we used the Delphi method to evaluate clinician consensus on the constituent components of frailty in ESRD and contrast those opinions with those of ESRD patients.
Discussion
In this study to elicit opinions about frailty in ESRD patients, there was consensus among clinicians who care for this population, and moderate consensus among surveyed adults with ESRD, that ESRD patients are more likely to be frail than healthy adults. There was some discordance between self-identification and measurement of frailty status, and participants who identified as frail were more likely to think that adults with ESRD are more likely to be frail, regardless of their own measured frailty status.
Four of the five Fried frailty components (exhaustion, low activity, slow walking speed and grip strength) were thought to be relevant to adults with ESRD, and clinicians were more likely to believe those four components were relevant. There was only moderate consensus among both clinicians and patients that unintentional weight loss characterizes frailty in this population, likely given the weight fluctuations and fluid shifts that are common in ESRD patients. Other possible ESRD-specific frailty components that were identified included history of falls, physical decline, and poor cognitive impairment. Interestingly, the novel components identified are less likely to be measures of kidney disease, but rather are downstream effects of ESRD. Finally, there was consensus among the clinicians that intradialytic foot peddlers and prehabilitation prior to KT would be effective interventions for reducing the burden of frailty in this vulnerable population.
Prior research has shown that frailty, as measured by the Fried phenotype, is an important predictor of ESRD and transplant patient outcomes [
5‐
13]. A prior study has identified 67 different instruments for the study of frailty among older adults [
20]; however, none of these instruments are specific to patients with ESRD. Our findings suggest that physicians and patients believe that exhaustion, low physical activity, slow walking speed, and poor grip strength were relevant to ESRD patients but that weight loss was possibly not relevant.
Interventions such as intradialytic activities and prehabilitation have been proposed to improve frailty among adults with ESRD [
21,
22]. Our findings are consistent with a previous qualitative study of hemodialysis patients that reported patient support for intradialytic exercise [
21]. We extended these findings to include ESRD patient and clinician opinions on whether intradialytic exercise using a foot peddler would improve physiologic reserve. Additionally, we included clinician and patient opinions on prehabilitation prior to transplantation; there was consensus among both groups that prehabilitation would help ESRD patients and make them less frail. This perception is consistent with studies that demonstrate that frailty is associated with physical functioning in older adults [
1].
Clinicians had higher levels of consensus on all issues relating to frailty in ESRD than did patients. Given the discordance between frailty status and self-perception of frailty found in this study, the lower levels of consensus among patients might be related to a lack of awareness of frailty and its prevalence in ESRD.
A strength of this study was the two round Delphi study design of clinicians caring for ESRD patients, which uses both qualitative and quantitative methods to elicit opinion and facilitate consensus. Additionally, this is the first study to directly survey ESRD patients on their opinions regarding what makes them frail. This study’s main limitation was the single-center design with respect to sampling patients and clinicians at Johns Hopkins. Perspectives of clinicians and patients from one center may not be generalizable. The wording of the survey items, particularly those questions relating to clinician perceptions of acceptability of interventions to improve frailty, may have elicited an affirmative response.
Nephrologists, transplant surgeons and geriatricians may all have incomplete perspectives on frailty in ESRD given the limited populations with whom they work. However, all clinician participants had experience treating adults with ESRD, as this was an inclusion criterion for this study. Despite these limitations, 85% of clinicians who treat adults with ESRD at Johns Hopkins were represented in the study, and the sample size was consistent with or larger than prior Delphi studies [
17,
23].
As the target population for this study was ESRD patients who were pursuing kidney transplantation, patients included in this study were all under evaluation for kidney transplantation. With an 80.6% response rate, we believe this study is representative of ESRD patients pursuing kidney transplantation at our center. Patients being evaluated for transplantation might be concerned that their appearing frail would affect their listing for transplantation, which could bias their responses. However, patients were assured that none of their responses would affect any aspect of their care during evaluation, including their access to transplantation. Furthermore, the questions in this survey are part of a longer 45-min assessment consisting of 252 questions. Thus the likelihood of responses being biased by a desire to appear not frail should be mitigated. Additionally, ESRD patients were not asked to suggest potential novel ESRD-specific frailty components.
The Fried frailty measure is just one of many measures that could be used in an ESRD setting. However, in a recent review of 67 different frailty assessment instruments, the Fried phenotype was the most cited instrument [
20]. The Fried phenotype is commonly used in ESRD research [
24‐
26] and its biological basis is well established [
27]. Given its prevalence, particularly in ESRD research, we sought to assess its appropriateness in an ESRD setting.
This study is one step in the process of refining the definition of frailty in ESRD. This process should also include structured reviews and qualitative studies to elicit further themes relating to frailty in ESRD. Any new measure of frailty in ESRD patients should be thoroughly validated before informing clinical decisions. Likewise, any new components added to this measure should be further studied and operationalized prior to its inclusion in a definition of frailty.
In a given population, the clinical usefulness of a definition of frailty depends on the relevance of its components to the physiology of that population. Among ESRD patients and clinicians who care for them, there is consensus that many Fried frailty components measure decline of physiologic reserve in ESRD patients; however, the complex physiologic issues experienced by patients with ESRD may make other aspects of physiologic reserve, like history of falls or physical decline, relevant in this population. An ESRD-specific measure of frailty could improve risk prediction in ESRD and transplant patients, and the patient and clinician opinions elicited in this study should inform the development of such a measure.