Introduction

Frailty is a state of homeostenosis associated with increased risks of disability, falls, impaired quality of life (QoL), and mortality. It is considered reversible given that it is acknowledged early and managed appropriately (1). Frailty has been mostly studied in older age. However, the etiological factors underlying frailty, e.g., increased inflammation, protein/energy malnutrition, limited physical activity, increased oxidative stress, loss of neurologic input, are not limited to older age. As such, frailty has been increasingly reported in diseases that have the outlined pathophysiological factors (2–5). Chronic kidney disease (CKD) is associated with protein loss and multiple metabolic disorders. It constitutes a catabolic process causing wasting of skeletal muscle predisposing to the development of frailty, via also increasing inflammatory cytokines (6). CKD comes front as a prevalent disorder that increases exponentially by aging and shares some common risk factors of frailty as noted above. Accordingly, CKD seems to constitute a high-risk state for frailty.

CKD is associated with disability, impaired QoL and mortality, some of which are probably contributed by the simultaneous presence of frailty. While there is limited means to counteract the progression of renal disease, frailty has been reported as reversible early in its course. Prefrailty state is reversible more than the overt frailty state as it is prodromal phase of frailty with less affected homeostatic reserves. Hence, recognition of prefrailty/frailty (i.e., frailty severity) and determination of their associates have potential to manage and improve care in CKD.

The presence of frailty in CKD settings has drawn attention recently, beginning from 2015 (7). In the literature, there are a variety of to evaluate frailty (8–10). Hence, frailty has been evaluated with a variety of tools in CKD. FRAIL scale is one of the recently proposed tools to evaluate frailty in clinical practice based on self-assessment via answers of the patient (11). It is performed in < 5 minutes making it one of the most convenient tools in crowded clinical settings and demonstrated its clinical validity in various researches (12–15). A few studies investigated frailty comparatively in different CKD stages with its independent associates (16–21) of which only one used FRAIL scale. Regarding this limited literature, we aimed to examine the prevalence and independent associates of frailty in CKD patients with estimated glomerular filtration rate (eGFR) <60 ml/min/1.73m2. We hypothesized that frailty or prefrailty is prevalent in stage 3 and higher CKD patients, and it increases by advanced renal disease.

Materials and Methods

This is a cross-sectional, single-centre study approved by the Health Sciences University, Haseki Training and Research Hospital Ethics Committee (number: 2019–52) and all participants gave informed consent. We followed STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (22).

Population and setting

We included a total of 148 consecutive patients aged 18–80 from the outpatient clinic and hemodialysis unit of our hospital, provided that they were cognitively intact and gave informed consent. Cognition was assessed by considering the patients’ clinical evaluation and prior diagnoses. Sixty (60) patients were maintenance hemodialysis (HD) patients (at least for 3 months), and 88 were stage 3–4 CKD patients. Thirty-seven (37) patients (42%), were eGFR G3a, 31 patients (35.3%) were eGFR G3b and 20 patients (22.7%) were eGFR G4 patients in stage 3–4 CKD patients. The exclusion criteria were as follows: the presence of advanced systemic disease [malignancy, sepsis, autoimmune diseases, New York Heart Association (NYHA) stage 4 congestive heart failure, cirrhosis, amyloidosis], acute coronary syndrome or cerebrovascular event in the last month, pregnancy/lactation, chronic debilitating neuromuscular or musculoskeletal disease, medication use that may affect bone, muscle or lipid metabolism (i.e., corticosteroids, immunosuppressives, statins), dependency in activities of daily living.

Assessments and measurements

We recorded age, sex, weight, height, body mass index, blood pressure values. Comorbid diseases, [diabetes mellitus (DM), hypertension, ischemic heart disease (IHD), congestive heart failure (CHF), peripheric artery disease (PAD), cerebrovascular event, chronic obstructive pulmonary disease (COPD)], etiology underlying the CKD, and regular medications were noted. Laboratory evaluations included serum glucose(mg/dl), HbA1c (g/dl), Na (mmol/l), K (mmol/l), Mg (mg/dl), Ca (mg/dl), Cl (mmol/l), P (mg/dl), C (mmol/l), urea (mg/dl), creatinine (mg/dl), eGFR (in non-HD participants, via CKD-EPI formula), uric acid (mg/dl), albumin (g/dl), AST (U/l), ALT (U/l), LDL (mg/dl), HDL (mg/dl), total cholesterol (mg/dl), triglycerides (mg/dl), LDH (U/l), CK (U/l), C-reactive protein (mg/dl), pro-BNP (pg/ml), PTH (pg/l), hemoglobin (g/dl), hematocrit (%), thrombocytes (x1000/mm3), ferritin (mg/L), 25 (OH) vitamin D3(ng/ml), lactate(mmol/l), and bicarbonate (mmol/l) levels. All laboratory evaluations were analyzed from blood samples taken after at least 8 hours of fasting in the morning and before the dialysis session in HD patients.

Evaluation of frailty

We used FRAIL scale to evaluate frailty (23). FRAIL scale is composed of 5 questions based on the answers of the individuals. It includes evaluation of fatigue, resistance, ambulation, illnesses, and weight loss. The application of the test is as follows: The first question evaluates the fatigue status of the participant and asks, “How much of the last 4 weeks did you feel tired?” the question is asked; 1=Always, 2=Most of the time, 3=Sometimes, 4=Sometimes, and 5=Never, and 1 point is given if the patient’s response is 1 or 2, all others are given 0 points. With the second question, it was aimed to measure the resistance of the participant and asked, “Do you have difficulty in climbing 10 steps of stairs on your own and without using an assistive device?” the question is asked. If the participant’s answer is yes, 1 point is awarded, if no, 0 points are awarded. With the third question, the participant’s mobility is evaluated and «Do you have difficulty walking a few hundred meters on your own and without using an assistive device?» (i.e., wheelchair, scooter, walker, cane) the question is asked. If the participant’s answer is yes, 1 point is awarded, if no, 0 points are awarded. With the fourth question, the health status of the participant is questioned. “Has a doctor ever told you that you have any of these diseases? (Hypertension, diabetes, cancer (other than minor skin cancer), chronic lung disease, heart attack, congestive heart failure, angina, asthma, arthritis, stroke, kidney disease)” is asked. If the participant has 0–4 diseases, 0 points are given, and if there are 5–11 diseases, 1 point is given. With the fifth question, the participant’s weight loss in the last year is questioned and the percentage of weight change is calculated. If the percent weight change is >5 (representing 5% weight loss), 1 point is given.” Each question is scored by zero or one. Total scores 0 indicate robust status, 1–3 pre-frailty, and 4–5 frailty (11). In this study, we grouped prefrail and frail as frail group and the score=0 individuals as robust (non-frail) group to achieve a comparable number of participants in 2 different frailty groups. This approach has been applied successfully in various researches (24, 25).

Outcome measures

The outcome measures were the prevalences of frailty status in the HD vs stage 3–4 CKD patients comparatively and the independent associates of frailty in these patients.

Statistical Analysis

We used IBM SPSS 26.0 for Windows (IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY: I.B.M. Corp). Normality was assessed by the Kolmogorov Simonov test and visual histograms. Descriptive statistics were given as number and percentage for categoric variables, and median and interquartile range (IQR) for continuous variables as appropriate. Independent two groups were compared as appropriate by ANOVA (with post hoc Tukey analysis) or Kruskal Wallis test (with Mann Whitney U test for post hoc analyses). Pearson analyses or Spearman’s rho analysed correlations between the numerical variables. Binary logistic regression analysis was used to detect independent associates of frailty. Odds ratio (OR) and 95% confidence interval (CI) were given. P-value was set at 0.05 for statistical significance.

Results

Demographic and Baseline Characteristics

We included a total of 148 participants (60 HD patients, 88 stages 3–4 CKD patients). We outlined demographics, disease characteristics, and the underlying CKD etiology in Table 1. There were no differences in age, sex, smoking status, and blood pressure measurements between HD and CKD stage 3–4 groups. Body mass index was significantly lower (24.05 vs. 30.8 kg/m2), while the prevalences of CHF (11.7% vs. 3.4%) and PAD (13.3% vs. 3.4%) were higher in the HD group, p<0.05 for all). Laboratory analysis results of the groups were in accordance with the patient characteristics (Table 2). Urea, creatinine, uric acid, potassium, phosphate, magnesium levels were higher in HD group than in CKD group. LDL-cholesterol, transaminases, hemoglobin, calcium were lower in HD groups than in CKD group.

Table 1 Characteristics of the study population stratified by CKD stage (n=144)
Full size table
Table 2 Laboratory data of the study participants (n=144)
Full size table

When patients were classified as prefrail or frail vs. robust, prefrailty+frailty was more prevalent in the HD group than the CKD stage 3–4 group (71.7% vs. 53.4%, respectively, p = 0.025). The prevalences of frailty and pre-frailty were 3.3% and 68.3% in HD group, respectively, while there were no frail patients in CKD stage 3–4 group, and pre-frailty prevalence was 53.4%. We outlined the prevalence of frailty by patient groups in Table 3 and Figure 1. The prefrailty+frailty prevalences were 23% in patients aged< 50; 60% in 51–64 ages and 71.2% in >=65 years old participants.

Figure 1
figure 1

Prevalence of frailty stratified by age groups (n=144)

Full size image
Table 3 Prevalence of frailty stratified by CKD stages (n=144)
Full size table

Multivariate Regression Analyses

Factors associated with frailty (Table 4)

Table 4 Logistical regression analysis to detect independent factors associated with prefrailty+frailty (n=144)
Full size table

We performed multivariate regression analysis to detect independent associates of frailty. Variables selected for inclusion in the models were based on clinical judgment and known literature predictors of mortality in patients on dialysis. The dependent factor was the presence of frailty/prefrailty; independent variables were patients’ group (HD vs. CKD stage 3–4 group), age, sex, presence of DM, IHD, CHF, peripheral artery disease, COPD, smoking status, hematocrit, albumin, TSH, 25 (OH) vitamin D3. In this analysis, patient group, age, and sex emerged as independent associates of frailty. Compared to CKD stage 3–4 patients, HD patients had a higher risk of suffering from frailty (OR=3.87, 95% CI= 1.06–14.19, p=0.04). Older age (OR=1.09, 95% CI= 1.04–1.13) and female sex (OR=9.13, 95%CI= 2.82–29.46) were associated with higher frailty risk (p<0.001, for both).

Discussion

In this study, including 148 participants (60 HD and 88 stages 3–4 CKD patients), we found that prefrailty/frailty status was quite common (71.6%) in HD patients and in patients with stage 3–4 CKD (53.4%). It should be noted that there was no individual with overt frailty in stage 3–4 CKD group and only 3.3% of the participants were frail in HD group. On the other hand, prefrailty+frailty status was prevalent even in younger HD patients as high as 23.0% in those < 50 years old and increased per increase in age. These results indicate that the presence of frailty in young age in CKD patients is confirming that although these individuals are chronologically young, they are biologically old. We believe that this is one of the major findings of the present study. In addition, the high prevalence of prefrailty in HD and CKD stage 3–4 groups rather than prevalent frailty indicate a great potential to reverse and prevent frailty in these patients which is another important finding of our analyses.

The prevalence of frailty has been mostly studied in patients on chronic maintenance dialysis. In a 2015 study, 324 adults were enrolled at HD initiation and frailty prevalence was assessed by Fried frailty tool (26). The prevalence of frailty (frailty score>=3) and intermediate-frailty (frailty score=2) were reported as 34.0% and 37.7%, respectively. In another 2015 study, frailty was assessed in 390 dialysis patients with the clinical frailty scale (27) reporting 26% of patients as mildly-severely frail and 53% were vulnerable-severely frail. In another study, van Loon et al. included 123 older patients (>= 65 years) on dialysis (28). They applied 6 different frailty tools and reported frailty between 48% (via modified Fried Frailty Index) and 88% (via Geriatric 8), stating that ∼75% of the older dialysis patients were frail. Garcia-Canton et al. investigated frailty prevalence in specifically 277 hemodialysis patients by Edmonton Scale (29), reporting 29.6% frailty and 19.1% vulnerability. In a 2020 study, Yabuuchi et al. involved 37 maintenance HD patients (7) using a modified version of Fried frailty criteria and reported frailty in 22%, prefrailty in 65%. In a 2016 study, Drost et al. included a total of 95 patients with ESRD (receiving hemodialysis, peritoneal dialysis and pre-dialysis care) and studied frailty with two different tools, i.e., frailty index and Fried frailty phenotype (4) declaring 36.8 % and %27.3 prevalences, respectively. The largest study on the prevalence and associations of frailty was performed by Lee et al. examined 1658 maintenance dialysis patients (30). Via a modified Fried frailty tool, they reported 34.8% as frail and 45.7% as prefrail. In a 2020 report, Yuan et al. studied frailty prevalence among 187 patients undergoing maintenance HD, and of note, they used FRAIL scale like our study (31). They reported frailty as 5.9%, prefrailty as 36.3%.

There were only two studies that examined the prevalence of frailty in non-dialysis patients. The first one included 104 older pre-dialysis patients (>=65 years) with an e-GFR <= 25 mL/min/1.73m2 (5) using PRISMA questionnaire and Timed up and Go test (5). They reported a considerably high prevalence of 53.8%. The second one included 105 older males (≥65 years) with non-dialysis CKD (eGFR at stage 3–4 CKD). Frailty was assessed by the Fried frailty criteria with a frailty prevalence, 35.2%; prefrailty, 29.5% (32). These figures suggest a considerably high prevalence of prefrailty+frailty or frailty, being also slightly higher than the present study among the non-dialysis patients (53.4%). These former studies included exclusively older patients and we suggest that their slightly higher prevalence of prefrailty+frailty is likely due to their higher age.

Similar to our study Rodriguez et al.(33, 34) had noted that the prevalence of overt frailty was 0% in patients with severe CKD (mean eGFR = 16) (33). Another study in HD patients, overt frailty was noted 22.0% and 65.0% was prefrail (7). This study has a participant number less than half of our study and assessed frailty by a more complicated tool Fried Criteria. The prevalence of frailty has been noted lowest among 5 other frailty tools (i.e. Strawbridge questionnaire, Edmonton Frail Scale, Groningen Frail Indicator, G8 questionnaire, and Tilburg Frail Indicator) (35) while other tools did not show consistently significant relationships with important dialysis-related complications while FRAIL questionnaire did. These points may explain the difference of their results and ours and point out the strength of this study considering the use of FRAIL criteria. Lastly, sarcopenia, is considered a major component and in some cases precursor of frailty (1) and may be indirectly considered to include those with prefarilty+frailty condition. When taking the studies examining sarcopenia prevalence among CKD patients, sarcopenia prevalence between 5.9%–49.4%, our prevalence of prefrailty+frailty as 71.6% in HD population and 53.4% in stage 3–4 CKD population is compatible with some of those in current literature.

As an overview, all these studies have been published so far, and our study put forward a considerable prevalence of frailty syndrome (prefrailty+frailty) in dialysis and ESRD patients, ranging between 40–80%, with the finding of the foremost largest study as 81.5%. Frailty has not a universally accepted globally applied standard evaluation tool, and different tools were applied in between the studies. As such, some differences in frailty prevalences may be due to the different tools applied to evaluate frailty. On the other hand, there are only three studies evaluating the significance of frailty in non-dialysis CKD settings, including the present study. These studies all indicated prefrailty+frailty prevalence of about 50%, indicating a considerable prevalence of frailty syndrome in different stages of severity in these CKD patients as well.

We analyzed the independent factors associated with frailty+prefrailty. Strikingly, apart from age and female sex, the only other factor independently associated with frailty+prefrailty emerged as the progressed CKD. Only a few reports are examining the multivariate associates of frailty, which are again mostly studied in end-stage renal failure settings. These studies included variable possible independent factors for association with frailty. In the above study conducted by Drost et al. in 2016 included a total of 95 patients with ESRD (receiving hemodialysis, peritoneal dialysis and pre-dialysis care) (4) and female sex and a higher comorbidity burden were associated with higher risk of frailty. In the above-mentioned largest study on frailty in 1658 patients under maintenance dialysis (30), multivariate analysis revealed significant associations of frailty with age, comorbidity, disability, unemployment, higher body mass index, and a lower educational level. In the most recent report performed by Yuan et al. in 2020 studied frailty also by FRAIL scale in 187 patients undergoing maintenance HD (31) and older age, comorbidities, depression, and sleep disorders were associated with frailty risk.

The only study other than the present study that examined multivariate associations of frailty in non-dialysis CKD patients included 105 older males (≥65 years) with eGFR at stage 3–4 CKD (32). Frailty was independently associated with older age, lower body mass index, and eGFR levels. Frailty was more prevalent in patients with eGFR<45 ml/min/1.73 m2 and 45–59 ml/min/1.73 m2 compared to patients with eGFR≥60 ml ml/min/1.73 m2. This finding is in line with our findings which signifies the significant association of severity of the renal disease with the occurrence of frailty. In particular, HD patients had 3.87 times higher risk than stage 3–4 CKD patients for frailty/prefrailty. Therefore, our findings suggest that ESRD may be a unique contributor to the development of frailty. However, the vice versa may also be valid and frailty status may be contributing to the progression in CKD severity. Nevertheless, all these studies point out considerably increasing prevalences of prefrailty/frailty by decreasing eGFR and progression of CKD. The ESRD represents the top level of noxious stimulus that drives elements of frailty, which may be a much more contributor than that of other diseases and laboratory markers in ESRD patients. In line with the literature, older age was also the common factor for higher frailty risk in all these studies, including the present one. Advanced age is a well-known factor related to the emergence of frailty. We found female sex as another independent risk factor for frailty, similar to the study reported by Drost et al. (4). Female sex is also one of the established risk factors in many frailty reports (4, 36, 37).

Our study has some limitations and strengths. While the participant number was comparable or higher than many studies, it was still limited to 148 CKD patients of whom 60 (40.5%) were hemodialysis patients. When we consider the other studies that examined the issue of frailty in CKD patients, there are recent studies including a total of 37 patients (7), 46 patients (35), while there are some studies that includes higher number of participants (38). When we consider these data, we suggest that the number of study participants is comparable with the data in the literature, being higher than some recently published studies. In our study age range was 18–80 which can be considered as a wide range. On the other hand, this is a real life data which included subsequent cases in a practicing nephrology unit. Therefore, the study reflects the current nephrology practice. Hence, we suggest that it provides the data for the current practice in our outpatient clinics without artificially grouping the patients by age. Focusing on particular age groups will also provide important data and it is a way of different study aim and analysis intend. Our aim was to outline the current situation in a standard outpatient nephrology clinics and therefore, we suggest that the age range is in line with our study aim. We suggest that, this point may be considered as one of the strength of our study as it reflects the CKD patients that we encounter in our daily practice. We included CKD patients from a single centre and country, and therefore, our findings lack generalisability to the overall CKD population. This is a cross-sectional study, and hence, one cannot suggest a cause-effect relationship between frailty and CKD. On the other hand, our strengths are, we included CKD patients at different stages, which enabled us to analyse frailty and frailty severity with advancing renal disease. In addition, we performed multivariate analysis to examine independent associates of frailty+prefrailty in CKD settings involving various risk factors for frailty, including known diseases and biomarker associates. These included well-known debilitating diseases such as diabetes, ischemic heart disease, heart failure and some well-known laboratory parameters that may be related to development of frailty including vitamin D level, albumin, hematocrit and TSH levels, in addition to renal disease. Of note, neither of these diseases or laboratory values appeared significant in multivariate analysis while ESRD did.

In conclusion, we found that prefrailty/frailty status was quite common (71.6%) in HD patients and in patients with stage 3–4 CKD (53.4%). Being a maintenance HD patient was identified as another independent risk factor, in addition to age and female gender, which are well known to be independently associated with frailty. Longitudinal studies including a larger number of patients are needed to comment more on the relationship between frailty and CKD and their consequences.