ORIGINAL RESEARCH

The Prevalence, Association, and Clinical

Outcomes of Frailty in Maintenance
Dialysis Patients
So-Young Lee, MD, PhD,* Dong Ho Yang, MD, PhD,* Eunah Hwang, MD, PhD,†
Seock Hui Kang, MD, PhD,‡ Sun-Hee Park, MD, PhD,§ Tae Woo Kim, MD, PhD,{
Duk Hyun Lee, MD, PhD,** Kisoo Park, MD, PhD,†† and Jun Chul Kim, MD, PhD‡‡

Objective: To investigate the clinical implications of frailty in chronic kidney disease patients undergoing maintenance hemodialysis
and chronic peritoneal dialysis.
Design: In this prospective study, all of the participants completed the Short Form of the Kidney Disease Quality of Life questionnaire,
Korean version, to determine their frailty phenotype. We also obtained blood chemistry and demographic data at enrollment. Data
regarding the history of hospitalization and death were collected during the follow-up period.
Subjects: We recruited 1,658 patients (1,255 maintenance hemodialysis and 403 chronic peritoneal dialysis) from multidialysis units
(n 5 27). We excluded patients who had been hospitalized in the previous 3 months.
Main Outcome Measures: Hospitalization and survival rate during study period.
Results: The participants’ mean age was 55.2 6 11.9 years old, and 55.2% were male. Among the participants, 34.8% were rated as frail
and 45.7% as prefrail. Multivariate analysis demonstrated significant associations of frailty with age, comorbidity, disability, unemployment,
higher body mass index, and a lower educational level. During the follow-up period (median 17.1 months), 608 patients (79 not frail, 250
prefrail, and 279 frail) were hospitalized, and 87 patients (10 not frail, 24 prefrail, and 53 frail) died (P , .001). Frailty was associated
with hospitalization (adjusted hazard ratio, 1.80; 95% confidence interval: 1.38-2.36) and mortality (hazard ratio, 2.37, 95% confidence in-
terval: 1.11-5.02).
Conclusion: The frailty phenotype was common even in, prevalent end-stage renal disease patients on dialysis, and was significantly
associated with higher rates of hospitalization and mortality.
Ó 2016 The Authors. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction According to annual reports from the United States

F RAILTY IS A biological syndrome characterized by

decreased reserves and resistance to stressors resulting
from cumulative decreases across multiple physiologic sys-
Renal Data System, 44.5% of end-stage renal disease
(ESRD) patients undergoing dialysis are aged 65 years and
older,9 and the prevalence per million continues to increase
tems.1-6 Frail individuals are therefore vulnerable to adverse more rapidly among older age groups.10 The Korean
outcomes, which can include disability, dependency, falls, nationwide ESRD patient registry reported that the mean
institutionalization, hospitalization, and death.1-7 The age of these patients has sharply increased from 55.2 years
concept of frailty was primarily developed and has evolved in 2005 to 60.3 years in 2014, and the proportion of people
to define geriatric syndromes, and chronological aging has with ESRD aged 65 years and older increased from 28% to
been positively correlated with the increasing prevalence of 40.7% during the same period.11 The incidence and preva-
the frail phenotype.1,8 lence of ESRD have also been rising rapidly, especially in

* ‡‡
Department of Internal Medicine, Division of Nephrology, CHA Bundang Department of Internal Medicine, Division of Nephrology, CHA Gumi
Medical Center, CHA University School of Medicine, Seongnam, South Korea. Medical Center, CHA University, Gumi, South Korea.
†
Department of Internal Medicine, Division of Nephrology, Keimyung Uni- Financial Disclosure: The authors declare that they have no relevant financial
versity School of Medicine, Daegu, South Korea. interests.
‡
Department of Internal Medicine, Division of Nephrology, Yeungnam Uni- Support: This research was supported by a grant of the Research Driven Hos-
versity Hospital, Daegu, South Korea. pital R&D project, funded by the CHA Bundang Medical Center (grant number:
§
Department of Internal Medicine, Division of Nephrology, Kyungpook Na- BDCHA R&D 2015-35).
tional University Hospital, Daegu, South Korea. Address correspondence to Jun Chul Kim, MD, PhD, 12, Sinsi-ro 10-gil, Gumi-
{
Department of Internal Medicine, Division of Nephrology, Soonchunhyang si, Gyeongsangbuk-do 39295, South Korea. E-mail: [email protected]
University Gumi Hospital, Gumi, South Korea. Ó 2016 The Authors. Published by Elsevier Inc. on behalf of the National
**
Department of Internal Medicine, Division of Nephrology, Daegu Fatima Kidney Foundation, Inc. This is an open access article under the CC BY-NC-
Hospital, Daegu, South Korea. ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
††
Department of Preventive Medicine, School of Medicine Gyeongsang Na- 1051-2276/$36.00
tional University, Jinju, South Korea. http://dx.doi.org/10.1053/j.jrn.2016.11.003

106 Journal of Renal Nutrition, Vol 27, No 2 (March), 2017: pp 106-112

 FRAILTY IN DIALYSIS PATIENTS 107

people older than 65 years.9,11 In addition, chronic kidney Cardiac diseases were defined when the patient had any med-
disease (CKD) per se accelerates the aging process at the ical history of angina pectoris, acute myocardial infarction,
cell, tissue, and organ level via protein energy wasting, congestive heart failure, positive cardiac exercise test results,
various uremic toxins, inflammation, and oxidative or interventions such as percutaneous transluminal coronary
stresses.12-14 These combined effects of chronological and angioplasty or coronary artery bypass surgery. Cerebrovascu-
pathological aging could explain why the frailty lar diseases were defined when the participant had any history
phenotype is much more common in the CKD of transient ischemic attack, cerebrovascular infarction, or
population regardless of dialysis therapy use and age, hemorrhage. The blood test results are the averages of the
compared with the general population with no monthly measurements obtained for the 3 months immedi-
impairment of kidney function.15-17 ately prior to obtaining other study measurements to capture
Individuals with ESRD requiring dialysis still have high a reliable and consistent picture of each participant’s condi-
hospitalization and mortality rates despite continuous and tion. The blood chemistry data included hemoglobin, serum
vigorous efforts to reduce known related risk factors such albumin, blood urea nitrogen (BUN), creatinine, potassium,
as anemia, hypertension, inflammation, hyperparathyroid- total cholesterol, calcium, phosphate, intact parathyroid hor-
ism, etc.18-20 For this reason, it is important to identify mone (iPTH), total iron binding capacity (TIBC), and high-
other potential factors that may lead to the many adverse sensitive C-reactive protein.
clinical outcomes in this population. This study was approved by the IRB of six medical
Several studies have examined the frailty phenotype and its school-affiliated hospitals (Keimyung University School
clinical implications for ESRD patients on dialysis.15,16,21-23 of Medicine, Yeungnam University Hospital, Kyungpook
However, most of those studies focused on patients who national University Hospital, Soonchunhyang University
were starting maintenance dialysis therapy, and data on Gumi hospital, Daegu Fatima Hospital, and CHA Gumi
peritoneal dialysis patients are limited. The aim of our Medical Center at CHA University). All of the study sub-
study was to investigate the prevalence of the frailty jects gave informed written consent.
phenotype and its relation to clinical outcomes such as
hospitalization and mortality in maintenance hemodialysis Definition of Frailty
(MHD) and chronic peritoneal dialysis (CPD) patients. Our study adopted a definition of frailty with compo-
nents that were identical or similar to those used in the
Methods Dialysis Morbidity and Mortality Study (DMMS) Wave
Subjects and Design 2,15 which were modified from the frailty definitions of
The inclusion criteria for MHD and CPD patients were the Cardiovascular Health Study by Fried et al.1 and the
as follows: (1) age $ 20 years, (2) dialysis duration Women’s Health Initiative Observational Study by Woods
$ 6 months, (3) no hospitalizations during the previous et al.7 A trained interviewer asked the study participants
3 months, except for vascular access repair, (4) able to questions pertaining to the frailty components of slowness,
ambulate with or without assistive devices, (5) sufficient weakness, exhaustion, shrinking, and physical inactivity us-
cognitive function to communicate with the interviewer ing the RAND 36-item Short Form (SF 36). We used a
to complete the questionnaires without help from others, Korean version of the Kidney Disease Quality of Life SF
and (6) willingness to give informed consent. The exclu- 36TM, which was linguistically validated.25 The presence
sion criteria included the following: (1) any acute infectious or absence of slowness and weakness was determined using
or other inflammatory illnesses, (2) active cancer except of the Physical Function (PF) Scale of the SF 36, which
basal cell carcinoma, and (3) current severe heart or lung consists of 10 items pertaining to physical activities usually
failure with unstable vital signs. performed during a typical day. The limitations of each
The study participants were recruited from 27 dialysis physical activity were classified into 3 categories: ‘‘limited
centers in the Daegu/Kyungsangbuk-do area of South Ko- a lot,’’ ‘‘limited a little,’’ and ‘‘not limited at all,’’ and each
rea. Eligible patients who met the inclusion criteria were response yielded a score of 0, 50 or 100, respectively. The
recruited from July 2012 to December 2012. We followed final score was determined by summing the scores for the
them until December 31, 2014, to evaluate the associations 10 physical activity items and dividing the total by 10. A
of frailty status with hospitalization and mortality. score lower than 75 on the PF scale of the SF 36 was consid-
To minimize the likelihood of the study subjects misunder- ered to indicate slowness and weakness, thus counting for 2
standing the interview questions, well-trained interviewers points. Exhaustion was measured with vitality scale, con-
directly administered the questionnaire, which consisted of sisting of four questions about how the respondent feels
the Kidney Disease Quality of Life questionnaire, Korean and how things have been during the previous 4 weeks as
version,24 and the physical activity pattern survey at enroll- follows: ‘‘Did you feel full of pep?’’, ‘‘Did you have a lot
ment and helped the participants complete it correctly. De- of energy?’’, ‘‘Did you feel worn out?’’, and ‘‘Did you feel
mographic data and hospitalization and death records for tired?’’ The average score for these 4 questions was calcu-
the study period were obtained via medical record review. lated; exhaustion was indicated if the score was lower
108 LEE ET AL

than 55, and the patient then received 1 point. Shrinking

was defined as the unintentional loss of more than 4.5 kg
(10 pounds) or 5% of the baseline body weight (BW) for
the past year; in such cases 1 point was added to the patient’s
score. Low regular physical activity of less than once per
week during leisure time for the previous 3 months was
considered physical inactivity and was scored as 1 point.
The points for each frailty component were totalled, and
the possible scores ranged from 0 to 5 point(s). Participants
with 3 or more points were considered ‘‘frail’’ and
those with 1 or 2 points were considered ‘‘prefrail.’’ Patients
with 0 points were described as ‘‘not frail.’’
Definition of Disability
Upon enrollment, we asked the participants four ques-
tions regarding activities of daily living (ADLs), including
whether they currently needed help from another person
to have a meal, dress/undress, get out of bed/lie in bed
and take a bath/shower. The respondents indicated the de-
gree of their need for help to complete these physical activ-
ities with one of three responses: no need for help, some
need for help and help is needed for the entire process.
Disability was defined as the inability to perform at least Figure 1. Study flow diagram. CPD, chronic peritoneal dial-
one of the 4 domains of ADLs with no help.7 ysis; MHD, maintenance hemodialysis.
Statistical Analysis
The baseline characteristics of the patients in this study groups, respectively (Table 1, Fig. 2). Figure 2 shows the
were compared with the frailty status data using the un- increasing tendency of the prevalence of frailty as the pa-
paired t test or ANOVA (continuous variables) and c2 anal- tients became older and the decreasing tendency of the
ysis (categorical variables). Multivariable logistic regression prevalence of prefrailty with age. In all age groups, a poor
analysis was used to determine the patient characteristics Vitality (VT) score (scored ,55) was the most common
that were associated with frailty at baseline. A survival anal- factor among the frailty components, followed in order
ysis was conducted using a Cox proportional hazards by inactivity, PF score ,75 and shrinking (Fig. 2).
regression model. The models were adjusted for the same Maintenance dialysis patients who were female, older,
set of predictor variables, including age, sex, comorbidities, had a higher BMI, were not employed, had a lower educa-
disability, dialysis modality, unemployment, body mass in- tional level, were disabled and with comorbid conditions,
dex (BMI), educational levels, serum albumin, iPTH, and such as diabetes mellitus and cardiac and cerebrovascular
hemoglobin. Two-tailed P values , .05 were considered disease, were significantly more likely to be frail, but dialysis
statistically significant. All of the analyses were completed modality and duration were not related to frailty (Table 1).
using SPSS software, for Windows (version 22; SPSS, Chi- Some nutritional markers in blood testing, such as BUN,
cago, Ill., USA). serum creatinine, and TIBC, showed significantly negative
relationships with frailty status; however, others, such as
Results serum albumin, total cholesterol, and high-sensitive C-
Eligible patients (n 5 1,658: 1,255 MHD and 403 CPD reactive protein, failed to show any significant relation to
patients) were recruited from all prevalent maintenance frailty. Lower serum levels of potassium and phosphorus
dialysis patients (n 5 2,737: 1,994 MHD and 743 CPD pa- were significantly associated with frailty (Table 1).
tients) at 27 dialysis centers to participate in this study After adjustment for all related covariates (age, sex, un-
(Fig. 1). The median follow-up period was 17.1 months. employment, disability, BMI, education level, dialysis mo-
The baseline demographic and blood chemistry character- dality, comorbidity, hemoglobin, phosphorous, serum
istics are described in Table 1. The mean age was albumin, potassium, TIBC, iPTH, BUN, and creatinine),
55.9 6 12.9 years, 25.6% were aged 65 year old and older, the variables that showed significant associations with frailty
55.7% were male, and 39.4% had diabetes. The mean time were age, comorbid conditions, disability, unemployment,
on dialysis was 5.2 6 4.5 years. Disability was observed in higher BMI, and lower education level (Table 2).
19.1% of the patients. Of the total group of patients, 608 (37%) were hospital-
Of the total study subjects, 577 (34.8%) and 757 (45.7%) ized at least once, 87 died, and 66 underwent kidney trans-
patients met the study definitions of frail and prefrail plantation during the 30-month follow-up period. The
 FRAILTY IN DIALYSIS PATIENTS 109

Table 1. Baseline Characteristics of Study Population According to Frailty Status

Characteristics Overall (n 5 1,658) Not Frail (n 5 324) Prefrail (n 5 757) Frail (n 5 577) P Value

Male, n (%) 923 (55.7) 205 (63.3) 446 (58.9) 272 (47.1) ,.001
Age, y 55.9 6 12.9 53.8 6 11.6 53.1 6 12.1 60.8 6 13.2* ,.001
MHD, n (%) 1,255 (75.7) 251 (77.5) 581 (76.8) 423 (73.3) .247
Vintage, y 5.2 6 4.5 5.4 6 4.6 5.2 6 4.4 5.1 6 4.5 .744
BMI, kg/m2 22.4 6 3.2 22.1 6 2.9 22.2 6 3.1 22.7 6 3.6*† .017
Employed, n (%) 440 (26.5) 103 (31.8) 263 (31.8) 74 (12.8) ,.001
Educational level, n (%) ,.001
6th grade 362 (21.9) 42 (13.0) 126 (16.7) 194 (33.6)
7th-12th grade 947 (57.2) 187 (57.9) 456 (60.3) 304 (52.7)
.12th grade 347 (21.0) 94 (29.1) 174 (23.0) 79 (13.7)
Diabetes, n (%) 654 (39.4) 105 (32.4) 255 (33.7) 294 (51.0) ,.001
Cardiac, n (%) 255 (15.4) 37 (11.4) 92 (12.2) 126 (21.8) ,.001
Cerebrovascular, n (%) 145 (8.7) 24 (7.4) 44 (5.8) 77 (13.3) ,.001
Disability, n (%) 317 (19.1) 23 (7.1) 70 (9.2) 224 (38.8) ,.001
Hemoglobin, g/dL 10.4 6 1.0 10.4 6 0.9 10.4 6 1.0 10.3 6 1.1 .297
Albumin, mg/dL 3.7 6 0.7 3.6 6 0.8 3.8 6 0.7* 3.7 6 0.7† .005
BUN, mg/dL 58.6 6 16.3 59.6 6 17.2 59.9 6 15.5 56.3 6 16.5*† ,.001
Creatinine, mg/dL 10.2 6 3.1 10.5 6 3.2 10.7 6 3.2 9.4 6 2.9*† ,.001
Potassium, mEq/L 5.0 6 0.8 5.0 6 0.8 5.1 6 0.8 4.9 6 0.8*† ,.001
T. cholesterol, ug/dL 154.3 6 37.3 152.0 6 36.9 154.8 6 37.3 154.9 6 37.2 .476
Calcium, mg/dL 8.6 6 1.0 8.6 6 1.0 8.7 6 1.0 8.6 6 1.0† .070
Phosphate, mg/dL 5.2 6 1.4 5.2 6 1.4 5.4 6 1.4* 5.0 6 1.0*† ,.001
iPTH, pg/dL 278.9 6 344.5 252.2 6 248.0 297.4 6 345.1 269.3 6 386.6 .107
TIBC, ug/dL 228.9 6 48.8 228.6 6 49.9 234.5 6 49.9 221.6 6 45.9*† ,.001
hsCRP, mg/dL 0.7 6 1.7 0.6 6 1.7 0.6 6 1.6 0.8 6 1.7 .133
BMI, body mass Index; BUN, blood urea nitrogen; hsCRP, high-sensitive C-reactive protein; iPTH, intact parathyroid hormone; MHD, main-
tenance hemodialysis; T. cholesterol, total cholesterol; TIBC, transferrin iron binding capacity.
Post hoc analysis was done by LSD test.
*P , .05 versus not frail patients.
†P , .05 versus prefrail patients.

hospitalization rate was 24.4% for the nonfrail, 33.0% for prefrail and frail patients were 1.4 (95% confidence interval
the prefrail, and 48.4% for frail patients (P , .001). The [CI]: 1.07-1.78) and 2.4 (95% CI: 1.87-3.08) times more
causes of hospitalization included infectious diseases likely, respectively, to be hospitalized over the 30-month
(29.3%), cardiovascular diseases (12.0%), gastrointestinal study period (Table S2). Moreover, the proportion of
bleedings (11.4%), cerebrovascular disease (4.0%), and patients with two or more hospitalizations was significantly
others (Table S1). The univariate analysis showed that the higher (25.3%) among the frail patients than among the

Table 2. Predictors of Frailty Analyzed by Multiple Logistic

Regression
Characteristics OR (95% CI) P

Age (per y) 1.03 (1.01-1.04) ,.001

Female gender 1.23 (0.92-1.63) .157
Diabetes 1.44 (1.11-1.87) ,.001
Cardiac 1.43 (1.01-1.98) .029
Cerebrovascular 1.56 (1.04-2.35) .031
Disability 5.60 (4.12-7.62) ,.001
Dialysis modality (PD) 1.11 (0.77-1.61) .571
Unemployed 1.89 (1.36-2.62) ,.001
BMI (kg/m2) 1.06 (1.02-1.10) .002
Education level
#6th grade 1.00 (referent)
7th-12th grade 0.67 (0.49-0.91) .012
.12th grade 0.53 (0.35-0.82) .010
Albumin (g/dL) 1.23 (0.98-1.55) .069
Figure 2. The prevalence of frailty and its components by BMI, body mass index; CI, confidence interval; OR, odds ratio; PD,
age. *P , .05. PF, physical performance; VT, vitality. peritoneal dialysis.
110 LEE ET AL

Figure 3. (A) Kaplan–Meier estimates of hospitalization-free survival probability of chronic dialysis patients in relation to frailty
status and (B) Kaplan–Meier estimates of cumulative survival probability of chronic dialysis patients in relation to frailty status.

nonfrail (9.6%) and prefrail (12.5%) patients (P , .001, cerebrovascular diseases (12.6%), in that order (Table S1).
Table S3). Figure 3 shows the lower hospitalization-free The univariate analysis indicated that the frail patients
survival and cumulative survival rates in the frail group were three times more likely to die during the follow-up
compared with the prefrail or nonfrail groups (P , .001). period than the patients in the other groups (HR, 3.05;
After adjustment for age, sex, comorbidities, dialysis mo- 95% CI: 1.55-6.00; Table S2). This significant relationship
dality, disability, serum albumin and creatinine, and the persisted even after adjustments for multiple other potential
other factors shown in Table 3, frailty remained signifi- risk factors of mortality (HR, 2.37; 95% CI: 1.11-5.02;
cantly associated with hospitalization (adjusted hazard ratio Table 3). Prefrailty did not show any significant associations
[HR], 1.80; 95% CI: 1.38-2.36; Table 3). However, prefrail with mortality in either the univariate or multivariate anal-
status was not related to hospitalization after those adjust- ysis (Table 3).
ments (Table 3).
The mortality rates were 3.1% for nonfrail, 3.2% for Discussion
prefrail, and 9.2% for frail patients (P , .001) during the This study indicates that the overall prevalence of the
study period. The three most common causes of death frailty phenotype in patients undergoing maintenance dial-
were infection-related (24.1%), cardiovascular (23%), and ysis therapy was substantially high (34.8%) compared with

Table 3. Multivariable Analysis of the Association of Frailty With Mortality and Hospitalization
Mortality Hospitalization
Characteristics Hazard Ratio (95% CI) P Value Hazard Ratio (95% CI) P Value

Status of frailty
Prefrail 1.01 (0.48-2.12) .980 1.29 (1.00-1.67) .050
Frail 2.08 (1.04-4.16) .039 1.83 (1.41-2.37) ,.001
Age (y) 1.03 (1.01-1.05) .002 1.01 (1.00-1.02) .012
Female gender 0.56 (0.36-0.88) .012 0.77 (0.64-0.93) .006
Diabetes 2.07 (1.30-3.29) .002 1.52 (1.59-2.31) ,.001
Cardiac 2.22 (0.42-1.13) .137 1.43 (1.17-1.75) .001
Dialysis modality (PD) 1.97 (1.18-3.28) .010 1.92 (1.59-2.31) ,.001
Unemployed 1.83 (0.97-3.69) .091 1.41 (1.13-1.77) .003
BMI (kg/m2) 0.91 (0.85-0.97) .006 0.99 (0.97-1.02) .563
iPTH (pg/dL) 1.00 (1.00-1.01) .012 1.00 (1.00-1.01) .002
Hemoglobin (g/dL) 0.76 (0.62-0.92) .006 0.89 (0.83-0.97) .006
Education level
#6th grade — — 1.00 (referent)
7th-12th grade — — 0.92 (0.75-1.15) .469
.12th grade — — 0.60 (0.43-0.81) .001
BMI, body mass Index; CI, confidence interval; iPTH, intact parathyroid hormone; PD, peritoneal dialysis.
 FRAILTY IN DIALYSIS PATIENTS 111

that of community-dwelling older adults (6.9% in the Car- tency seems to result from differences in the characteristics
diovascular Health Study and 16.3% in the women’s Health of study populations, the complexity of pathophysiology
Initiative Observational Study),1,7 although our study depending on the CKD stage, the collected/analyzed
population was relatively young (74.4% were younger data, the definition of frailty, and other factors.
than 65 years old) and in stable physical condition The adverse effects of frailty on clinical outcomes have
according to the inclusion criteria described above. been widely examined in people without or with CKD
Even in predialysis stage, CKD patients are more across all stages. Large-scale prospective frailty studies
likely to be frail than people with normal kidney func- have revealed high HRs of disability (HR, 1.79-3.15),
tion. Shilpak et al.17 reported a higher prevalence of hip fracture (HR, 1.57), hospitalization (HR, 1.27-1.95),
frailty in CKD patients than in those without renal and mortality (HR, 1.63-2.24) in elderly community-
dysfunction (15% vs. 6%; P , .001) and an increasing dwelling people.1,7 Frailty was also significantly associated
trend as kidney function decreased. Consistent findings with high risks of death and dialysis initiation (HR, 2.0-
have been observed in populations with CKD stages 2.5) in predialysis CKD patients.26,28 More falls or
1-426,27 and in the Third National Health and fractures (HR, 1.60),22 earlier or more frequent hospitaliza-
Nutrition Evaluation Survey.28 tion (HR, 1.26-1.56), and higher mortality (HR, 1.22-
Much higher prevalence of frailty has been reported in 2.60) were also significantly associated with the frailty
CKD patients on dialysis than in the predialysis CKD pop- phenotype in incident and prevalent dialysis pa-
ulation. Johansen et al.15 reported that the frailty prevalence tients.15,16,21,23 In our study, the frailty phenotype was
of incident hemodialysis and peritoneal dialysis patients was also associated with higher risks of hospitalization (HR,
67.7% in the DMMS Wave 2, and it was 73.3% for the 1.80; 95% CI: 1.38-2.36; Table 3) and mortality (HR,
Comprehensive Dialysis Study participants.21 Our study 2.37; 95% CI: 1.11-5.02; Table 3) in the multivariate anal-
reported a frailty prevalence among CKD patients on dial- ysis. These associations of adverse clinical outcomes with
ysis (34.8%) that was higher than that of the predialysis frailty described in our study are consistent with and com-
CKD patients and much lower than that of the in incident parable to the results of other studies.
dialysis patients from the DMMS Wave 2 and Comprehen- Our study had several strengths. First, it included large
sive Dialysis Study. Other studies including subjects on number of participants (n 5 1,658) recruited from 27 dial-
prevalent MHD showed frailty prevalence from 13.8% to ysis centers for the specific subgroup of prevalent dialysis
41.8%.16,29,30 patients, that is, ambulatory ESRD patients with no recent
The reasons CKD patients are more likely to be frail are admission history. Second, a relatively large proportion of
thought to be: (1) a trend toward a rapidly increase in the CPD patients (n 5 403) was also included. Third, well-
number of elderly CKD patients,9,11,14 (2) a ‘‘premature’’ trained study assistants interviewed all of the study partici-
or ‘‘accelerating’’ aging process caused by CKD pants directly to minimize misunderstandings when
itself,12-14,31 and (3) the combination of aging and completing the questionnaire.
CKD.14 The differences in the frailty prevalence among There are also several limitations. First, we did not use a
CKD patients could be caused by the different definitions physical performance–based definition of frailty; rather, we
of frailty that each study adopted.32-34 We adopted the adopted self-reported (perceived) constructs to define
modified Fried frailty criteria based on self-report, which frailty. It has been reported that perceived frailty reports
can be easily implemented in routine clinical practice and tend to overestimate the frailty prevalence, compared
has been validated in frailty studies of CKD patients.32-34 with measured frailty.32-34 However, considering the
This modified Fried criteria has proven to be predictive prevalence rate and significant associations with adverse
of adverse outcomes such as hip fracture, disability, health outcomes shown in this study, the modified
hospitalization, and mortality in large population–based criteria for the frailty construct that we used seemed to
studies.7 However, several reports have shown that this be quite reliable and more easily applicable than the
tool tended to overestimate the frailty prevalence compared classic criteria, especially for this type of large-scale study.
with the use of objective criteria.32 Second, no significant clinical implications of prefrailty
The factors shown to be significantly related to the frailty were found here, whereas other studies have reported a sig-
phenotype in this study (older age, comorbidities, disability, nificant relationship between prefrailty or intermediate
unemployment, higher BMI and lower education level; frailty and an increased mortality risk in people with
Table 2) were found also in other frailty studies of the gen- CKD on dialysis.15,16,23 We suggest that the follow-up
eral population1 and CKD patients receiving dialysis,15,21,29 period of future studies be sufficiently long to prove the
but differences also exist. Johansen et al.15 showed that fe- clinical significance of prefrailty in the CKD population.
male gender and dialysis modality (MHD) were significant
predictors of frailty, but we did not. Some studies reported Practical Application
that a low serum albumin level had a positive relationship The findings show a high prevalence of frailty in ambu-
with frailty,15,30 whereas others did not.21 This inconsis- latory chronic dialysis patients without a recent admission
112 LEE ET AL

history. The frail phenotype is significantly associated with 18. Kalantar-Zadeh K, Block G, Humphreys MH, et al. Reverse epidemi-
hospitalization and mortality. Therefore, we should pay ology of cardiovascular risk factors in maintenance dialysis patients. Kidney Int.
2003;63:793-808.
more attention to the frailty status of patients, even those 19. Longenecker JC, Coresh J, Powe NR, et al. Traditional cardiovas-
who appear to be in a good condition, to improve their cular disease risk factors in dialysis patients compared with the
morbidity and mortality. general population: the CHOICE Study. J Am Soc Nephrol. 2002;13:
1918-1927.
Supplementary Data 20. Beto JA, Bansal VK, Gohlke NP, et al. Using the hemodialysis prog-
nostic nutrition index and urea reduction ratio to predict morbidity and mor-
Supplementary data related to this article can be found at tality: a pilot study of the 1995 council on renal nutrition national research
http://dx.doi.org/10.1053/j.jrn.2016.11.003. question. J Ren Nutr. 1998;8:21-24.
21. Bao Y, Dalrymple L, Chertow GM, et al. Frailty, dialysis initia-
References tion, and mortality in end-stage renal disease. Arch Intern Med.
1. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence 2012;172:1071-1077.
for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146-M156. 22. Delgado C, Shieh S, Grimes B, et al. Association of self-reported frailty
2. Lipsitz LA, Goldberger AL. Loss of ’complexity’ and aging. Potential appli- with falls and fractures among patients new to dialysis. Am J Nephrol.
cations of fractals and chaos theory to senescence. JAMA. 1992;267:1806-1809. 2015;42:134-140.
3. Buchner DM, Wagner EH. Preventing frail health. Clin Geriatr Med. 23. Alfaadhel TA, Soroka SD, Kiberd BA, et al. Frailty and mortality in
1992;8:1-17. dialysis: evaluation of a clinical frailty scale. Clin J Am Soc Nephrol.
4. Bortz WM 2nd. The physics of frailty. J Am Geriatr Soc. 1993;41:1004-1008. 2015;10:832-840.
5. Campbell AJ, Buchner DM. Unstable disability and the fluctuations of 24. Park HJ, Kim S, Yong JS, et al. Reliability and validity of the Korean
frailty. Age Ageing. 1997;26:315-318. version of Kidney Disease Quality of Life instrument (KDQOL-SF). Tohoku
6. Hamerman D. Toward an understanding of frailty. Ann Intern Med. J Exp Med. 2007;211:321-329.
1999;130:945-950. 25. Park KS, Cho MH, Ha IS, et al. Validity and reliability of the Korean
7. Woods NF, LaCroix AZ, Gray SL, et al. Frailty: emergence and conse- version of the pediatric quality of life ESRD module. Health Qual Life Out-
quences in women aged 65 and older in the Women’s Health Initiative Obser- comes. 2012;10:59.
vational Study. J Am Geriatr Soc. 2005;53:1321-1330. 26. Roshanravan B, Khatri M, Robinson-Cohen C, et al. A prospective
8. Lam M, Jassal SV. The concept of frailty in geriatric chronic kidney dis- study of frailty in nephrology-referred patients with CKD. Am J Kidney Dis.
ease (CKD) patients. Blood Purif. 2015;39:50-54. 2012;60:912-921.
9. Collins AJ, Foley RN, Chavers B, et al. US renal data system 2013 27. Dalrymple LS, Katz R, Rifkin DE, et al. Kidney function and prevalent
annual data report. Am J Kidney Dis. 2014;63:A7. and incident frailty. Clin J Am Soc Nephrol. 2013;8:2091-2099.
10. Saran R, Li Y, Robinson B, et al. US renal data system 2014 annual data 28. Wilhelm-Leen ER, Hall YN, K Tamura M, et al. Frailty and chronic
report: epidemiology of kidney disease in the United States. Am J Kidney Dis. kidney disease: the Third National Health and Nutrition Evaluation Survey.
2015;66:S1-S305. Svii. Am J Med. 2009;122:664-671.
11. Jin DC. Major changes and improvements of dialysis therapy in Korea: 29. Johansen KL, Dalrymple LS, Delgado C, et al. Association between
review of end-stage renal disease registry. Korean J Intern Med. 2015;30:17-22. body composition and frailty among prevalent hemodialysis patients: a US
12. Stenvinkel P, Larsson TE. Chronic kidney disease: a clinical model of renal data system special study. J Am Soc Nephrol. 2014;25:381-389.
premature aging. Am J Kidney Dis. 2013;62:339-351. 30. Kutner NG, Zhang R, Huang Y, et al. Risk factors for frailty in a
13. Kooman JP, Broers NJ, Usvyat L, et al. Out of control: accelerated ag- large prevalent cohort of hemodialysis patients. Am J Med Sci.
ing in uremia. Nephrol Dial Transplant. 2013;28:48-54. 2014;348:277-282.
14. Kim JC, Kalantar-Zadeh K, Kopple JD. Frailty and protein-energy 31. Walker SR, Wagner M, Tangri N. Chronic kidney disease, frailty, and
wasting in elderly patients with end stage kidney disease. J Am Soc Nephrol. unsuccessful aging: a review. J Ren Nutr. 2014;24:364-370.
2013;24:337-351. 32. Painter P, Kuskowski M. A closer look at frailty in ESRD: getting the
15. Johansen KL, Chertow GM, Jin C, et al. Significance of frailty among measure right. Hemodial Int. 2013;17:41-49.
dialysis patients. J Am Soc Nephrol. 2007;18:2960-2967. 33. Johansen KL, Dalrymple LS, Delgado C, et al. Comparison of self-
16. McAdams-DeMarco MA, Law A, Salter ML, et al. Frailty as a novel report-based and physical performance-based frailty definitions among
predictor of mortality and hospitalization in individuals of all ages undergoing patients receiving maintenance hemodialysis. Am J Kidney Dis.
hemodialysis. J Am Geriatr Soc. 2013;61:896-901. 2014;64:600-607.
17. Shlipak MG, Stehman-Breen C, Fried LF, et al. The presence of frailty 34. Salter ML, Gupta N, Massie AB, et al. Perceived frailty and measured
in elderly persons with chronic renal insufficiency. Am J Kidney Dis. frailty among adults undergoing hemodialysis: a cross-sectional analysis. BMC
2004;43:861-867. Geriatr. 2015;15:52.