An overview of frailty in kidney transplantation: measurement, management and future considerations

Abstract

The construct of frailty was first developed in gerontology to help identify older adults with increased vulnerability when confronted with a health stressor. This article is a review of studies in which frailty has been applied to pre- and post-kidney transplantation (KT) populations. Although KT is the optimal treatment for end-stage kidney disease (ESKD), KT candidates often must overcome numerous health challenges associated with ESKD before receiving KT. After KT, the impacts of surgery and immunosuppression represent additional health stressors that disproportionately impact individuals with frailty. Frailty metrics could improve the ability to identify KT candidates and recipients at risk for adverse health outcomes and those who could potentially benefit from interventions to improve their frail status. The Physical Frailty Phenotype (PFP) is the most commonly used frailty metric in ESKD research, and KT recipients who are frail at KT (~20% of recipients) are twice as likely to die as nonfrail recipients. In addition to the PFP, many other metrics are currently used to assess pre- and post-KT vulnerability in research and clinical practice, underscoring the need for a disease-specific frailty metric that can be used to monitor KT candidates and recipients. Although frailty is an independent risk factor for post-transplant adverse outcomes, it is not factored into the current transplant program risk-adjustment equations. Future studies are needed to explore pre- and post-KT interventions to improve or prevent frailty.

INTRODUCTION

Frailty is a syndrome characterized by diminished strength, endurance and reduced physiologic function, increasing an individual’s vulnerability for developing increased dependency and/or dying when confronted with a stressor [1]. The construct of frailty was first established by geriatricians and gerontologists who identified the need to distinguish the physiological age from the chronological age of older adults [2]. Metrics of frailty were developed to improve the ability to accurately identify the most vulnerable individuals in older populations by going beyond traditional risk factors such as age and comorbidity. In recent years there has also been a proliferation of research on frailty in nonelderly populations and in numerous medical subpopulations, including those with kidney disease and solid organ transplants [3, 4].

In kidney transplantation (KT), clinical care paradigms are adapting to an aging transplant candidate pool [5–7] and increasing waiting times [8, 9]. These trends underscore the need to accurately identify KT candidates and recipients who are at higher risk of adverse outcomes when facing health stressors, including the surgical and immunologic stressors of KT. Some experts suggest that a patient’s frailty status could inform decisions about the referral, evaluation and management of KT candidates as well as optimal rehabilitation plans after transplant surgery [10, 11].

In this review we summarize available tools to measure frailty. We then discuss the impact of frailty on access to KT and on morbidity and mortality before and after KT. Next, we consider topics relating to the immunosuppressive management of frail KT recipients and examine the most recent data on interventions to improve frailty. Finally, we emphasize key areas in which research is needed to improve the identification and clinical management of frailty in the KT patient population.

METHODOLOGY

In 2017, members from the American Society of Transplantation (AST) Kidney Pancreas Community of Practice (KPCOP) formed a Frailty Work Group with the goal of summarizing the current literature on frailty metrics, the impact of frailty on end-stage kidney disease (ESKD) and KT populations and potential interventions and areas for additional research. This project was part of a larger AST initiative to build consensus related to frailty measurement and care in solid organ transplantation [4]. We first conducted a PubMed search for literature on relevant topics in frailty, including search terms such as ‘frail elderly’, ‘frail instrument’ and ‘kidney transplant’. Our search strategy yielded 641 unique articles. We supplemented this search with searches in the EMBASE, Community Index to Nursing and Allied Health Literature and Cochrane databases. An updated search of the PubMed database was also performed in April 2019. Members of the KPCOP Frailty Work Group were divided into six subgroups of three to four individuals who reviewed the literature on a specific subtopic and created a summary. The information was shared with the KPCOP Frailty Work Group via a series of teleconferences and web-based communications from August 2017 to February 2018. The products of the subgroups were collated into a single harmonized manuscript by three primary authors (M.N.H., M.K.R. and K.J.W.) and two senior authors (M.M.D. and D.M.D.). A complete draft was circulated to the work group for feedback and revision, resulting in the final review manuscript. All authors had continuous access to the working documents to provide input, critical review and revisions.

IDENTIFYING FRAILTY IN KT CANDIDATES AND RECIPIENTS

Instruments to measure frailty

Although there is an agreement regarding the underlying conceptual framework of frailty, there is a low level of consensus regarding the constituent elements to be included in operational definitions of frailty [12, 13]. At least 67 different frailty scales have been used in population-based studies [14, 15], and there is similar heterogeneity in studies of patients with ESKD [16]. In Table 1, we summarize a nonexhaustive list of the available frailty instruments that have been applied to populations with chronic kidney disease (CKD), dialysis dependence and KT. Instruments were included if they have been applied to patients with ESKD in at least one study of prevalence and/or outcome prediction. As in the geriatric arena, the PFP, originally described by Fried et al. in 2001 [17], has emerged as the most commonly applied frailty assessment in studies of patients with ESKD [16]. The PFP includes five domains: weakness, slowness, unintentional weight loss, exhaustion and low physical activity. Individuals who meet three or more of these criteria are at high risk of adverse outcomes when faced with health stressors. A number of factors such as advancing age, comorbidities, polypharmacy and malnutrition contribute to this phenotype among individuals with ESKD, exacerbating vulnerability to illness and to treatment interventions such as dialysis, transplant surgery and immune therapy (Figure 1).

Table 1.

Summary of commonly utilized frailty and functional metrics in studies of dialysis and KT populations

Frailty indicator	Components	Limitations	Scoring	Populations studied
Clinical frailty scale [18]	8-point scale with brief descriptions based on clinical interview that takes into account: Mobility; energy; physical activity; function	Subjective Does not include multimorbidity but does include disability	From 1 (very fit) to 8 (very severely frail)	Incident dialysis [19]
Physical frailty phenotype [17]	Five components: Shrinking (i.e. unintentional weight loss); exhaustion; physical inactivity; weakness; slowness	Subjective and objective components	Score 0–5; 0 = Robust 1–2 = Intermediate ≥3 = Frail	KT [20−26] Prevalent hemodialysis [27, 28] Incident dialysis [29]
Groningen frailty indicator [30]	15 items: mobility [4]; self-rated physical fitness [1]; vision [1]; hearing; nourishment [1]; morbidity [1]; cognition [1]; psychosocial [5]	Subjective and objective components Includes morbidity and disability	Scores 0–15	Predialysis clinic [31] Predialysis and prevalent dialysis [32] Incident dialysis [29] Prevalent dialysis [33]
Tilburg frailty indicator [34, 35]	15 components: physical [8]; psychologic [4]; social [3]	Subjective Does not include morbidity or disability	Scores 0–15	Prevalent dialysis [33]
Frailty index [36, 37]	Deficit accumulation, including comorbid illness, poor health attitudes, signs of disease and self-reported disabilities; 40–70 deficits typically included	Components vary Includes multimorbidity and disability	0–1, scored as the number of deficits present divided by the total number assessed	Predialysis and prevalent dialysis [32]
Edmonton frail scale [38]	Eight items: cognition; general health status; functional independence; social support; medication use; nutrition; mood; continence; functional performance	Subjective and objective Includes morbidity and disability	Scores 0–17; >7 = frail	Prevalent dialysis [33]
FRAIL scale [39]	Five domains: Fatigue; resistance (ability to climb one flight of stairs); ambulation (ability to walk 1 block); illnesses (>5); loss of weight (>5%)	Subjective Includes morbidity and disability	Scores 0–5; 0 = not frail 1–2 = pre-frail ≥3 = frail	Prevalent dialysis [33]
1994 frailty measure or Strawbridge questionnaire [40]	16 items, including 4 domains: physical functioning [4]; nutritive functioning [2]; cognitive functioning [4]; sensory problems [6]	Subjective Does not include morbidity or disability	Score of ≥3 on any item = problem or difficulty with that domain; frail = difficulty in ≥2 domains	Prevalent dialysis [33]
SF-12 PCS	12-item assessment of physical functioning. Subscale of the Kidney Disease Quality of Life-36 instrument	Subjective	0–100 (lower score = worse PF)	Prevalent dialysis [41, 42] KT [43]
SPPB [44]	Three items: balance; chair standing; gait speed	Does not include multimorbidity or disability	Scores 0–12	KT [45, 46]
Timed up and go [47]	Standing from chair, walking a short distance, returning and sitting	Does not include multimorbidity or disability	Score in seconds	KT [48]
Gait speed [49]	Timed walking over a short distance	Does not include multimorbidity or disability	Score in seconds or meters per second	ESKD [50]

FIGURE 1.

Frail individuals are most vulnerable to the numerous health stressors of kidney disease.

Although there are several validated, self-reported instruments that are simpler to apply in clinical settings than the PFP [e.g. the Kidney Disease Quality of Life Short Form Physical Component Subscale (SF-12 PCS)] [51], there are potential trade-offs in utilizing assessments that may not directly assess physiologic reserve. Conversely, as the KT evaluation setting might make some patients reluctant to reveal the extent of their functional limitations to KT providers, a purely objective frailty instrument, such as the Short Physical Performance Battery [52], may be desirable. The vast number of available frailty metrics and the differences between them underscore the importance of developing more unified measures to assess vulnerability across the broad population of KT candidates and recipients.

In a national survey of 133 KT centers representing 79% of all adult KT candidates in the USA [53], there was substantial heterogeneity in the metrics used in clinical settings to assess pre-KT frailty (n = 18 distinct metrics). The majority of centers (67%) reported utilizing more than one frailty metric in their transplant evaluation process and the most common metric utilized by KT centers was a timed walk test (19%) (Table 1). The variability in KT center practices with respect to measurement and utilization of frailty instruments is likely a result of the lack of consensus in the transplant community about how frailty should be assessed. Accordingly, a recent AST consensus statement opined that organ system–specific frailty assessments are likely needed [4].

PRETRANSPLANT FRAILTY: PREVALENCE, RISK FACTORS AND OUTCOMES

Prevalence of frailty in populations with CKD and ESKD

People with CKD and ESKD have a high prevalence of frailty: there is 15–21% frailty prevalence in the CKD population versus 3–6% in the general population [54, 55]. Among dialysis-dependent individuals, the prevalence of frailty is likely higher, ranging from 14 to 73%, and is common among those <40 years of age (63%) [16, 56]. These data suggest that the prevalence of frailty is high among all KT candidates, including pre-emptive candidates and younger candidates.

Individuals who are referred for KT evaluation are likely to be healthier than the overall population of individuals with ESKD, and those who are selected for KT waitlists may be healthier still. A large multicenter study identified 18% of individuals as frail at the time of initial evaluation, while only 12% of individuals were identified as being frail among those who were ultimately listed for KT [57]. Furthermore, frailty status may change considerably from the time of listing to the time of KT. For example, in a single-center study of 569 adult KT candidates, 22% of the cohort was more frail at the time of KT than at the time of KT evaluation, whereas 24% were less frail at KT [58]. Approximately 20% of KT recipients are frail at the time of KT [59], and the frailty components most commonly observed in this population are weak grip strength (50%) and low physical activity (49%) [59]. Given the dynamic nature of frailty, periodic reassessments of frailty may be warranted prior to the surgical stressor of KT.

Risk factors and correlates for frailty

Many of the risk factors for frailty in potential KT candidates are not modifiable (Figure 2). For example, studies have consistently shown that KT candidates of advanced age and female sex are more likely to be frail than younger candidates and males, respectively [57, 60, 61]. However, other risk factors, such as obesity and low physical activity, might be modifiable. With respect to correlates of frailty, although higher comorbidity burden is also a risk factor for frailty among KT candidates, frailty can also occur in the setting of lower comorbidity burdens [62]. Diabetes and serum albumin concentration are also associated with frailty among prevalent dialysis patients [63] and individuals with CKD and ESKD who are frail are also more likely to have cognitive impairment and sarcopenia, or low muscle mass, than their nonfrail counterparts [64, 65].

FIGURE 2.

The continuum of frailty in kidney disease. The figure displays current knowledge on the risk factors, correlates and outcomes of frailty among individuals with kidney disease.

Among individuals with nondialysis-dependent CKD, the risk of frailty has an inverse relationship with CKD stage, as defined by cystatin-based glomerular filtration rate (GFR) calculations [66]. However, the association between CKD stage and frailty is attenuated when GFR is estimated using creatinine as opposed to cystatin C, a finding that is potentially explained by the relation of creatinine to muscle mass (i.e. lower serum creatinine may reflect sarcopenia). Therefore creatinine-based estimated GFR (eGFR) may overestimate actual GFR in frail people with sarcopenia, an important consideration given that waiting time for deceased donor KT (DDKT) cannot be accrued in the USA until individuals have eGFRs ≤20 mL/min/1.73 m².

Among those who are dialysis-dependent, it is unclear whether dialysis itself improves or worsens frailty. Multiple studies have shown a decline in functional status in older adults who initiate dialysis [67, 68]. In a longitudinal study that measured frailty in a dialysis cohort of 762 subjects, most subjects’ scores changed from year to year [63]. However, improvement in frailty was as common as the worsening of frailty. With respect to modality of renal replacement therapy, several recent studies have suggested that frailty and functional impairments are similarly prevalent among patients with ESKD who receive hemodialysis, peritoneal dialysis and conservative care [69–71]. Studies are needed to determine whether dialysis treatment–related interventions (e.g. parenteral nutrition and duration of treatment) could improve frailty.

Outcomes of frailty in populations with nondialysis-dependent CKD and dialysis dependence

Potential KT candidates with frailty are at high risk of multiple adverse health outcomes. In a cohort of 336 subjects with nondialysis-dependent CKD, the proportions of frail individuals who had impairment in at least one activity of daily living, instrumental activity of daily living and mobility were 15, 60 and 40%, respectively, compared with 5% (P = 0.009), 28% (P < 0.001) and 18% (P = 0.001), respectively, among those without frailty [72]. Frailty is also an independent risk factor for hospitalization [27, 56, 73] and doubles the risk of death among individuals with ESKD and KT [19, 27, 54, 56, 72–74]. Slow gait speed [50, 74], immobility [75] and poor physical function (PF) [76] have also been associated with a higher risk of death in both CKD and ESKD. In a study of 311 subjects with nondialysis-dependent CKD, the 6-min walk distance had the highest discriminative accuracy for 3-year mortality {area under the curve [AUC] 0.80 [95% confidence interval (CI) 0.70–0.90]}, followed by gait speed [AUC 0.78 (95% CI 0.70–0.86)] and timed up and go [AUC 0.74 (95% CI 0.64–0.84)]. Each of these physical performance tests had an AUC that was superior to commonly measured biomarkers of CKD, including eGFR, serum bicarbonate, hemoglobin, C-reactive protein and albumin [74]. Therefore, knowledge of frailty could inform shared decision-making about the risks and benefits of KT between potential KT candidates and their providers beyond the standard biomarkers that are commonly assessed and reviewed.

Frailty and access to KT

Among individuals who are being evaluated for KT, frailty is associated with reduced access to the waiting list and higher waiting list mortality. In a study of 7078 individuals who were evaluated at three transplant centers between 2009 and 2018, frail individuals were almost half as likely as nonfrail individuals to be placed on a KT waiting list [hazard ratio (HR) 0.62 (95% CI 0.56–0.69)] [57]. In another study of 128 individuals who were evaluated for KT, 30.4% of frail individuals were subsequently listed for KT, compared with 57.6% of nonfrail individuals [77]. Among those who are successfully wait-listed, frail KT candidates may be more likely to be inactive, less likely to receive KT [51] and more likely to die while wait-listed than nonfrail KT candidates [61]. Healthcare utilization might be another useful proxy for frailty in predicting waiting list outcomes: compared with listed KT candidates with no hospitalization days in the first year after listing, a study of 51,111 wait-listed individuals in the USA found that candidates with ≥15 hospitalized days had a >2-fold risk of subsequent waiting list mortality [HR 2.07 (95% CI 1.99–2.15)] [78].

POTENTIAL ROLES FOR FRAILTY METRICS IN PRE-TRANSPLANTATION SETTINGS

KT candidate assessment

Given the extent to which KT candidate selection relies on provider perceptions of patients, clinicians who rely solely on clinical acumen for health surveillance of their KT candidates may be likely to misclassify some patient subgroups as frail. In a study of 146 hemodialysis patients from a single dialysis facility, investigators assessed agreement between measured frailty using the Fried et al. criteria to nephrologists’ subjective ratings of patient frailty. Nephrologists were inaccurate in their ratings of frailty in 37% of cases, and older adults were the patient subgroup that was most likely to be misclassified as frail [79]. Misclassification of frailty could have large implications for access to KT, as individuals who are inaccurately deemed ‘too old, ill or frail’ to undergo KT may be less likely to receive transplant education or referral as a result [79–82]. To minimize frailty misclassification that may improperly restrict access to KT, nephrologists may consider augmenting their clinical assessments of potential KT candidates with direct and objective measures of frailty, particularly among older individuals with ESKD.

Unintended consequences: frailty assessments and transplant center practices

Knowledge of frailty in the KT evaluation and selection processes may help to promote individualized care of the most vulnerable patients, permitting timely interventions to improve functional status and listing outcomes [83–85]. However, concerns arise about the potential of unintended consequences when integrating frailty assessments into the KT evaluation process. Earlier in this article, we described evidence that frail individuals with ESKD may still receive a substantial survival benefit from KT compared with remaining on dialysis [51], and that frailty is potentially reversible with successful KT [20]. Indeed, data suggest that selected older patients and those with long dialysis exposures who receive KT are likely to have better survival and quality of life than similar patients who do not receive KT [86, 87]. However, among US KT programs that use a frailty metric during KT candidate assessments, 53% reported that they were less likely to list a frail KT candidate for transplant [53]. These findings have led to concerns that until transplant programs are no longer disincentivized from accepting high-risk KT candidates, individuals with ESKD who are assessed (accurately or inaccurately) to be frail will have reduced access to KT [11].

Longevity matching

Given the independent association between frailty and survival, another potential role of frailty metrics is to improve efforts to maximize utility in organ allocation. The revised US Kidney Allocation System (KAS) incorporated a continuous scale called the Estimated Post-Transplant Survival (EPTS) to facilitate allocation of the highest quality deceased donor organs to recipients who are expected to live the longest (i.e. ‘longevity matching’) [88]. Under the revised KAS, candidates with the longest predicted post-transplant survival (EPTS ≤20%) are prioritized for kidneys from donors ranked as ‘top 20%’ highest quality based on the Kidney Donor Profile Index. The EPTS is based on candidate age, duration of dialysis, diabetes and prior solid organ transplant status, and has a C-statistic of 0.69 (i.e. considered ‘good’ discriminatory ability) [89]. An important area for future research is to examine whether the inclusion of frailty may improve the current longevity matching paradigm.

POSTTRANSPLANT FRAILTY: EARLY AND LATE OUTCOMES

Early outcomes among frail KT recipients

Increasing evidence suggests that frail KT recipients are more vulnerable to the immediate surgical and immunologic stressors of KT than nonfrail recipients (Table 2). In a prospective cohort study of 183 KT recipients transplanted between 2008 and 2010, frailty was independently associated with a nearly 2-fold higher risk of delayed graft function (DGF) [adjusted relative risk (aRR) 1.94 (95% CI 1.13–3.36)] [21]. Frailty is also associated with a 2-fold higher risk of post-KT delirium [aOR 2.05 (95% CI 1.02–4.13)] [90]. These findings suggest that frail KT candidates are likely to require more support during their initial KT hospitalization.

Table 2.

Studies that have evaluated the association between frailty (and related constructs) and post-transplant outcomes

References	Design and participants	Frailty measure	Frailty distribution	Correlates of frailty	Outcomes
Garonzik-Wang et al. [21]a	Prospective cohort 183 KT recipients at 1 US center (December 2008–April 2010) 35% LDKT	Physical frailty phenotype defined as score ≥3	Frailty at KT: 25% (46/183)	Baseline demographics, diabetes prevalence and donor traits were similar in frail and nonfrail recipients	DGF: 30% versus 15% in frail versus nonfrail KT recipients Frailty was independently associated with twice the risk of DGF [aRR 1.94 (95% CI 1.13–3.36)]
McAdams-DeMarco et al. [22]a	Retrospective cohort 383 KT recipients at 1 US center (December 2008–December 2012) 39% LDKT	Physical frailty phenotype defined as score ≥3	Frailty at KT: 19% (72/383)	Frail sample includes more males (71% versus 58%, P = 0.046)	Frail KT recipients were more likely to experience EHR within 30 day (46% versus 28%, P = 0.005), regardless of age After adjusting for previously described registry-based risk factors, frailty predicted 61% higher risk of EHR [aRR 1.61 (95% CI 1.18–2.19)]
McAdams-DeMarco et al. [20]a	Prospective cohort 349 KT recipients at 1 US center (December 2008–March 2014) 37% LDKT	Physical frailty phenotype defined as score ≥3	Frailty over time: At KT: 20%; 1 month: 33%; 2 months: 28%; 3 months: 17%		Frailty scores typically worsen at 1-month post-KT (mean change +0.4, P < 0.001), and improved by 3 months post-KT (mean change −0.3, P = 0.04) Recipients frail at time of KT were more than twice as likely to improve in post-KT frailty than nonfrail recipients [aHR 2.55 (95% CI 1.17–3.82)]
Reese et al. [41]	Retrospective cohort 19 242 KT recipients in US with linked Fresenius dialysis records 10% LDKT	SF-36 PF subscale quartile (SF-12 PCS), administered on dialysis (before listing and time-updated)	Median (IQR) subscale score: 55 (35–80)		After KT, lower PF score was associated with shorter 3-year survival (84% versus 92% for the lowest versus highest function quartiles) Compared with dialysis, KT was associated with a statistically significant survival benefit by 9 months for patients in every SF-12 PCS quartile
Harhay et al. [43]	Retrospective cohort 8870 hemodialysis patients who received KT in US—dialysis records linked to UNOS and Medicare claims data 18% LDKT	SF-36 Physical Function subscale quartile (SF-12 PCS), administered on dialysis (before listing) Hospitalizations in year before KT as proxy for frailty	28% with ≥1 pre-KT hospitalization		Lower PF score was associated with 1.24-fold EHR risk compared with higher quartiles (aOR 1.08–1.43) More than one pre-KT hospitalization associated with 1.32-fold higher EHR risk [aOR 1.32 (95% CI 1.17–1.49)]
Alhamad et al. [91]	Retrospective cohort 489 KT recipients at one center (2000–14) 29% LDKT	6-min walk test at the time of transplant evaluation	25% short walk <1101 feet 50% medium walk 1101–1414 feet Long walk >1414 feet	Recipients with short 6-min walk were more likely female and had a history of diabetes and stroke	Short walk distance was not associated with short-term risk of graft failure, but was associated with longer-term risk Selection bias, as recipients with short walk distance would be denied for transplant unless they have other favorable characteristics predicting excellent short-term outcomes
McAdams-DeMarco et al. [92]a	Prospective cohort 443 KT recipients at 2 centers (May 2014–17) 35% LDKT	Physical Frailty Phenotype defined as score ≥2			Frail recipients had worse physical (P < 0.001) and kidney disease–specific HRQOL (P = 0.001), but similar mental HRQOL (P = 0.43) Frail recipients experienced significantly greater rates of improvement in physical HRQOL [frail: 1.35 points/month (95% CI 0.65–2.05); nonfrail: 0.34 points/month (95% CI −0.17 to 0.85); P = 0.02] and kidney disease–specific HRQOL [frail: 3.75 points/month (95% CI 2.89–4.60); nonfrail: 2.41 points/month (95% CI 1.78–3.04); P = 0.01], but no difference in mental HRQOL
McAdams-DeMarco et al. [59] a	Prospective cohort 663 KT recipients at one center (December 2008–August 2015) 39% LDKT	Physical frailty phenotype defined as score ≥2	Frailty at KT: 20%	Age was the only conventional factor associated with frailty However, factors rarely measured as part of clinical practice (e.g. HRQOL, IADL disability and depressive symptoms) were significant correlates of frailty	KT recipients with exhaustion and slowed walking speed [HR 2.43 (95% CI 1.17–5.03)] and poor grip strength, exhaustion and slowed walking speed [HR 2.61 (95% CI 1.14–5.97)] had increased mortality risk over an average of 3.1 years of follow-up
Lynch et al. [93]	Retrospective cohort 37 623 Medicare-insured KT recipients in the USRDS (January 2000–December 2010) %LDKT not reported	Hospitalization days within 1 year before KT as proxy for frailty	Hospitalization days in pre-KT year: 0: 51%; 1–7: 25%; 8–14: 11%; 15+: 13%	Patients with highest level of pre-KT hospitalization days were more commonly: women, previous transplant recipients, diabetic and those with CHF, atherosclerotic vascular disease and COPD	After adjusting for other baseline factors, pre-KT hospitalization days (1–7, 8–14, 15+) bore graded associations with readmission in the first year after KT [aHR 1.28 (95% CI 1.17–1.70)] and with mortality [aHR 1.42 (95% CI 1.20–1.70)] and all-cause graft loss [aHR 1.30 (95% CI 1.15–1.44)] >3 years post-KT
Lynch et al. [78]	Retrospective cohort 51 111 Medicare-insured KT recipients in the USRDS (January 2000–December 2010) 0% LDKT	Hospitalization days in first year after waitlisting as a proxy for frailty	Hospitalization days in first year after listing: 0: 46.9%; 1–7: 22.6%; 8–14: 11.7%; 15+: 18.7%	More hospitalization days associated with female sex, white race, previous transplant, diabetes, CHF, atherosclerotic vascular disease and COPD	After adjusting for other baseline factors, hospitalization days after listing (1–7, 8–14, 15+) bore graded associations with death after listing [aHR 1.49 (95% CI 1.24–2.07)] and with decreased likelihood of KT Model using wait-list hospitalization days alone had higher predictive accuracy for wait-list mortality than EPTS score

^aDistinct samples/analyses from the same cohort.

LDKT, living donor kidney transplantation; aHR, adjusted hazard ratio; USRDS, US Renal Data Service; IADL, independent activities of daily living.

Frail KT recipients and those with physical impairments are also likely to have higher healthcare utilization post-KT. For example, lower extremity impairment at the time of KT, objectively measured by the Short Physical Performance Battery (SPPB), is associated with a longer hospital length of stay following KT [45]. Frail KT recipients are more likely to experience early hospital readmission within 30 days of discharge from KT than nonfrail recipients [45.8% versus 28.0%; aRR for readmission among frail recipients 1.61 (95% CI 1.18–2.19)] [22]. Thus, interventions that can improve PF and frailty status prior to KT could have the potential to decrease the number of hospital days and readmissions post-KT and reduce costs.

Longer-term patient and graft survival outcomes in frail KT recipients

The long-term benefits of KT are not uniform or guaranteed, but rather vary based on factors including recipient age, comorbidities, the timing of transplantation and organ quality [7, 94, 95]. Independent of traditional risk factors, the PFP is associated with a 2.2-fold higher risk of mortality after KT, whereas intermediate frailty is associated with a 1.5-fold higher risk of mortality [23]. Similarly, lower extremity impairment, objectively measured by the SPPB, is associated with a 2.3-fold higher post-KT mortality risk and a 16% absolute increase in 5-year mortality post-KT [46]. With respect to self-reported PF, a retrospective cohort study of 19 242 US KT recipients with linked dialysis center records including SF-12 PCS scores found that lower SF-12 PCS scores were associated with reduced 3-year survival (84% versus 92% for the lowest versus highest quartiles) [41]. Nonetheless, KT was associated with a statistically significant survival benefit over dialysis by 9 months for patients in every PF quartile in this study. These results suggest that the survival advantage of KT persists across KT recipients of varying PF.

Post-KT changes in frailty status

KT itself is associated with dynamic changes in frailty status: patients are confronted with surgical and immunologic stressors, but also experience restored kidney function. In a prospective cohort of 349 KT recipients, investigators found that among recipients of all ages, frailty initially worsened in the first-month post-KT but then improved by 3 months post-KT [20]. Furthermore, KT recipients who were frail at KT were more likely than nonfrail recipients to show improvements in physiological reserve over time, suggesting that frailty is potentially reversible with KT [20]. Some evidence suggests that frail KT recipients receive outsized benefits from KT with respect to improvements in health-related quality-of-life (HRQOL). In a retrospective cohort study of 443 KT recipients at two US centers, frail recipients experienced significantly higher rates of improvement in physical HRQOL and kidney disease–specific HRQOL with KT than nonfrail recipients [92]. These studies suggest that carefully selected frail KT candidates can receive substantial benefits from KT.

IMAGING STUDIES, MORPHOMETRIC AGE AND KT OUTCOMES

Whereas dedicated frailty instruments are not a uniform feature of KT candidate assessments, KT candidates often undergo cross-sectional imaging as part of their preoperative workup. In the general surgical literature, preoperative imaging has allowed for the quantification of interrelated concepts including sarcopenia, morphometric age and fat composition as potentially modifiable predictors of postoperative outcomes [96–98]. Sarcopenia, as measured by low psoas muscle cross-sectional area and density on computed tomography (CT), is associated with higher waiting list mortality in KT candidates [99], but studies on postoperative outcomes are lacking. As opposed to chronologic age, older morphometric age, as quantified using CT-measured aortic calcifications, psoas muscle cross-sectional area and psoas muscle density, is associated with higher mortality among both kidney and liver transplant recipients [100, 101]. Taken together, these results suggest that morphometric measurements may have a role in frailty assessment.

POTENTIAL ROLES FOR FRAILTY METRICS IN POSTTRANSPLANTATION SETTINGS: GRADING TRANSPLANT PROGRAM PERFORMANCE

Given the strong associations between frailty and adverse post-KT outcomes, frailty assessments may augment posttransplant clinical care and help providers to identify KT recipients who require additional support. Furthermore, the independent association between frailty and adverse post-KT outcomes has important implications in the regulation of US transplant program performance. Currently, program performance is graded based on risk-adjusted computations of expected 1-year patient and graft survival, and programs face serious consequences when recipient death and/or graft loss rates ‘exceed expected’ [102]. Frailty is not one of the risk factors used to adjust outcome expectations and this may serve as a disincentive for transplant programs to accept frail candidates for KT. Given the data that suggest frail candidates can receive survival and HRQOL benefits from KT, an important question is whether incorporation of frailty-based risk scores into program-specific reports could reduce disincentives for transplant programs to select frail individuals who may benefit from KT [11].

The Organ Procurement and Transplantation Network (OPTN) already collects information on the Karnofsky Performance Score, a measure of self-reported or observed functional status that has been shown to enhance prediction of posttransplant survival [103]. However, the subjective nature of the Karnofsky Performance Score has rendered its inclusion in national risk adjustment (constructed by the Scientific Registry of Transplant Recipients with OPTN data) problematic, and it is no longer included in post-KT risk adjustment equations. Therefore, developing a consensus on objective instruments to measure frailty that can be widely implemented across transplant programs is an important goal for the future [104].

FRAILTY, AGING AND IMMUNE SYSTEM

The immunosenescence phenotype of aging and chronic disease

Akin to physiologic aging and chronic disease, frailty has been associated with alterations in the immune system [61, 105–107] and these changes may have important implications for graft survival and immunosuppressive management for frail KT recipients. Immunosenescence, a state of deteriorating and compromised immune response, has been studied in the context of aging and among older KT recipients [108]. Notably, physiologic aging appears to be linked to an imbalance of innate and adaptive immunity, with innate immune responses gaining prominence (Figure 3, Panel A). Chronic disease processes such as CKD may also enhance the state of immunosenescence. Studies of immune function in pediatric patients with CKD have shown premature T-cell aging, including a reversal of the CD4:CD8 ratio, reduced portions of native T-cells with an evidence of T-cell exhaustion and loss of CD28 expression [111]. Immunosenescence, as defined by immunophenotyping or measurement of telomere length, has been linked to broad changes of the immune response, increasing the risk of infection and malignancy in older adults [109, 112, 113]. However, it should be noted that the state of immunosenescence does not equate to absence of inflammation. In fact, senescent cells have been characterized as secreting a number of proinflammatory cytokines, chemokines, growth factors and proteases locally and contribute to the ‘inflammaging’ phenotype of the elderly [114]. Therefore, given the aging KT populations [7], research is needed to explore potential differences in the mode of action, dosing, metabolism and pharmacokinetics of immunosuppressants in the setting of immunosenescence.

FIGURE 3.

Physiology of immunosenescence: the aging immune system. (Panel A) Aging is associated with immunosenescence, resulting in alterations in the immune response. These alterations may require adjustment of immune therapy after KT [109] (Panel A redesigned with permission from Transplantation: November 2015-Volume 99-Issue 11- p 2258-2268, Copyright © 2015 Wolters Kluwer Health [110]). (Panel B) In addition to aging, frailty may also influence immune therapy risks after KT. Novel immune system biomarkers may permit individualization of immune therapy among vulnerable transplant recipients.

The inflammatory phenotype of frailty

As frailty can occur across the lifespan in ESKD, it may not always be associated with a state of immunosensecence. However, relative to nonfrail individuals with ESKD, those with frailty exhibit increased inflammatory markers such as C-reactive protein and interleukin-6, findings that are independently associated with higher mortality risk among ESKD patients [61]. The implications of frailty-related inflammation on the response to immunosuppressive treatment are unclear. Importantly, there are no definitive data that would support reduced immunosuppression in the context of frailty alone. In contrast, a study of 525 KT recipients showed that mycophenolate mofetil dose reduction was associated with a 5-fold higher risk of death-censored graft loss and that this association was not modified by frailty status [24]. Research is needed to explore whether interventions to improve frailty can also impact systemic inflammation and how changes in inflammation might influence KT outcomes.

New tools to tailor immunosuppressive therapy for frail KT recipients

Given the complex interplay between frailty and physiologic aging, developing a better understanding of the role of biomarkers such as inflammatory markers, cytokines, T-cell phenotypes and markers of senescent cells in targeting immunosuppression may facilitate improved management of frail KT recipients [105–107, 115]. A number of biomarkers have been associated with frailty, leading to interest in developing a biomarker-based frailty index [116]. However, studies of such an index in ESKD and KT populations have not yet been conducted. In addition, a number of noninvasive urine and blood biomarkers for acute rejection of kidney allografts have been studied and validated [116–119], but these have not been examined in the context of frailty. Validation and implementation of noninvasive biomarkers, such as urine and blood messenger RNA/micro-RNA profiles, T- and B-cell phenotypes, blood cytokine levels and cell-free donor-derived DNA, in the frail KT population could lead to the development of a personalized approach to immunosuppression for vulnerable KT recipients. Striking the right balance between the risks of rejection and graft loss with the risks of infection and malignancy (Figure 3, Panel B) is critical to improving the quality of life for frail KT recipients.

The interaction between aging, frailty and the immune system is complex. Identification of frailty status prior to KT offers a window of opportunity to change one of the variables in the equation. Whether improvement in frailty status alters the immune phenotype of young and older frail KT candidates and improves posttransplant outcomes needs to be studied. Therefore it will be critical to define measures that can reduce frailty and mitigate the deleterious immune consequences of frailty prior to KT, which may include interventions such as physical therapy, cognitive training and novel therapeutics, including senolytic agents [120].

OPPORTUNITIES TO INTERVENE IN PRE- AND POST-KT FRAILTY: STRUCTURED EXERCISE PROGRAMS, PREHABILITATION AND REHABILITATION

Interventions to reduce frailty in populations with CKD and ESKD are understudied, although data from interventions tested among frail older patients may be instructive. Interventions to reduce frailty in community-dwelling older adults are most often multidimensional [121] and include exercise training, nutritional supplementation or pharmaceutical agents [122]. They have focused on the reversible phenotypic frailty components (weakness, slowness and low energy expenditure) to delay functional decline and disability rather than to prime patients before a major stressor [123, 124]. Interventions that significantly reduce frailty among community-dwelling older adults include physical activity interventions and pre-emptive rehabilitation (i.e. prehabilitation) [125].

Exercise trials: data from populations with CKD and ESKD

It remains an open question as to whether exercise can improve overall vulnerability among CKD and ESKD patients. However, several randomized trials of patients with CKD and ESKD have demonstrated the potential benefits of physical exercise programs to prevent or reverse sarcopenia and improve PF [126–130]. A Cochrane review evaluating the effect of exercise on CKD and KT patients that included 45 randomized controlled trials showed that regular exercise improved physical fitness, cardiovascular dimensions, serum albumin and health-related quality of life [128]. A meta-analysis evaluating 41 trials that compared any regular exercise training for at least 8 weeks with sham or no exercise in CKD and ESKD showed any type of exercise significantly increased aerobic capacity and mid-thigh muscle area (four trials) but found no change in walking capacity [131]. Neither the Cochrane review nor the meta-analysis focused specifically on frail individuals. Small trials have demonstrated the potential for physical therapy programs to benefit frail patients with CKD and ESKD [83].

Prehabilitation before KT

Prehabilitation, or intensive exercise therapy prior to an elective surgical intervention, shifts the focus to optimization prior to surgery rather than rehabilitation after surgery [84]. In a recent survey, both clinicians (97%) and patients (94%) agreed that pre-KT prehabilitation could help patients undergoing KT and that prehabilitation could make ESKD patients less frail (clinicians 100%, patients 84%). Additionally, 97% of clinicians and 80% of patients agreed that patients would be interested in pre-KT prehabilitation [13]. In a pilot study [85], 18 KT candidates participated in weekly physical therapy sessions with at-home exercise. After 2 months, participants improved their physical activity by 64% (P = 0.004). These data suggest that prehabilitation is a promising intervention for KT candidates with frailty. However, larger studies with longer durations of follow-up are likely needed to determine whether exercise programs can improve pre- and peritransplant vulnerability to health stressors.

Posttransplant rehabilitation

Several studies have investigated the role of exercise therapy in ambulatory KT recipients. Two European centers report efficacy from a structured rehabilitation program post-KT [132, 133]. In a US trial (N = 97), an individualized home exercise regimen starting at 1-month posttransplant and monitored with regular phone follow-up improved peak oxygen uptake (a surrogate of cardiopulmonary fitness), muscle strength and self-reported physical functioning compared with usual care at 1 year [134]. The average age in the trial was low (40 ± 13 years in the exercise arm) and 43% of the cohort did not complete the exercise protocol. A subsequent UK trial recruited 60 older patients (age 55 ± 11 years in the exercise arm) within 1 year of KT and tested the effect of 12 weeks of supervised structured exercise classes twice per week for 12 weeks (aerobic versus resistance training versus usual care) and reported improvement in peak oxygen uptake attributable to both aerobic and resistance training compared with usual care [135]. Together, these studies suggest that KT itself improves cardiorespiratory fitness within the first-year posttransplant and that these improvements can be further augmented by aerobic and resistance exercise.

CONCLUSION

In this review we discussed the evidence that frailty is highly prevalent among individuals before and after KT, with implications for post-KT outcomes and the need for future research on interventions and access to KT (Table 3). Many tools exist that may assist clinicians in identifying KT candidates and recipients who are uniquely vulnerable to health stressors. However, research is needed to compare the discriminatory ability of existing frailty metrics for monitoring patient-oriented KT outcomes. A harmonized dynamic measure that captures decreased physiologic reserve in ESKD patients may be needed. Furthermore, the preponderance of evidence suggests that frailty is an independent and commonly unmeasured risk factor for numerous adverse outcomes among KT candidates and recipients, underscoring the urgent need to prospectively evaluate the impact of targeted frailty interventions (e.g. structured exercise, physical therapy and dietician support) on access to KT and post-KT outcomes. Finally, although pre- and post-KT outcomes among frail individuals are worse than outcomes among nonfrail peers, KT may still provide survival and quality-of-life benefits for many frail individuals compared with remaining on dialysis. Therefore we recommend that evidence of frailty should not be used to disqualify individuals from KT candidacy, but rather used to identify KT candidates that may require additional surveillance and support before and after KT.

Table 3.

Take-home points and recommended future research in frailty and KT