Frailty and CKD

Frailty is a process theorized to be, at least in part, attributable to chronic inflammation marked by hormonal changes, anorexia, decreased muscle strength, and depression. The phenotypic definition operationalized by Fried et al. (1) includes any three or more of the following: weight loss, weakness, poor endurance and energy, slowness, and low physical activity level. By these criterion, 7% of the initial cohort of 5317 participants and 12% of the black cohort from the Cardiovascular Health Study (CHS) were frail. The 4-year incidence of frailty among these participants was similar to study prevalence and survival was lower among frail individuals. However, frailty by these criteria is more prevalent among patients with CKD and has also been independently linked to lower survival in several cohort studies (2–4).

Despite the overlap of clinical manifestations, frailty is highly prevalent among patient with CKD. Studies across many different cohorts have examined the extent to which kidney dysfunction independently increases the risk of frailty, and the level of kidney function at which the prevalence of frailty substantially rises has also been examined. Wilhelm-Leen et al. (5) estimated kidney function using the Mayo quadratic equation and found that 21% of participants in the Third National Health and Nutrition Examination survey with eGFR<45 ml/min per 1.73 m² were frail according to modified frailty criteria. Two previous CHS studies have also examined the association of kidney function decline and frailty prevalence. In one of those studies, a serum creatinine ≥1.5 mg/dl in men and 1.3 mg/dl in women was associated with a nearly 80% higher odds of prevalent frailty (2). In the second study, an eGFR by cystatin C measurement (eGFR_cys) of 59–45 ml/min per 1.73 m² was associated with a near doubling of the risk of prevalent frailty (odds ratio, 1.75; 95% confidence interval [95% CI], 1.13 to 2.71) and more than double the risk of prevalent frailty (odds ratio, 2.68; 95% CI, 1.58 to 4.54) among individuals with eGFR_cys 44–15 ml/min per 1.73 m² (6). Similarly, Roshanravan et al. examined a cohort of 336 patients referred for nephrology care and estimated kidney function using the CKD Epidemiology Collaboration equation with cystatin C, age, sex, and race. These investigators found that the prevalence of frailty was 2.1- and 2.8-fold higher for participants with an eGFR <30 and 30–44 ml/min per 1.73 m², respectively, compared with an eGFR >60 ml/min per 1.73 m² (4). Among individual with ESKD, frailty prevalence has been documented to be nearly 80% (7). These studies highlight how widespread frailty is among patients with CKD. However, whether CKD itself contributes to the incidence of frailty over time remained an underexplored question.

In this issue of CJASN, Guerville et al. (8) examined the association of kidney function decline with incident frailty and whether kidney function decline would be more predictive of incident frailty over time than baseline kidney function measures. The investigators evaluated participants from the Multidomain Alzheimer’s Preventive Trial (MAPT), a randomized study designed to assess the efficacy of supplementation with omega-3 fatty acid, a multidomain intervention (consisting of nutritional counseling, physical exercise, and cognitive stimulation), or a combination of the two interventions on the change of cognitive function in participants aged ≥70 years for a period of 3 years (9). Participants who completed the 3-year intervention study were followed for an additional 2 years. Frailty was assessed by Fried criteria at study initiation, at 6 and 12 months, and then annually for the remaining 4 years. Kidney function was assessed at baseline, 6, 12, and 24 months by CKD Epidemiology Collaboration equation, using serum creatinine, age, and sex. Among study participants with more than one eGFR measure, eGFR slope over time was calculated. Individuals whose eGFR slope was in the lowest quartile (>4.1 ml/min per 1.73 m² per year) were categorized into the fast decline group for comparisons. The incidence of frailty between 24 and 60 months was determined among participants who were not frail at baseline and 24 months. Kaplan–Meier analyses and Cox models were used to examine the associations. Study analyses were adjusted with the inclusion of body mass index, markers of inflammation, and history of diabetes, hypertension, heart failure, stroke, ischemic heart disease, and death. In addition, the authors performed an impressive series of sensitivity analyses.

Of the 1679 participants in the original cohort, 833 (50%) were not frail at baseline and at 24 months and had complete frailty data and at least two measures of eGFR within the first 24 months of the study period. The cohort mean follow-up time was 53±10 months and the analytic cohort at 5 years dropped to 538 individuals. The median baseline to 24-month eGFR slope was −0.5 (interquartile range, −4.1 to 3.8) ml/min per 1.73 m² per year. Between 24 and 60 months, 11% of participants became frail. In primary analysis, a fast eGFR decline was significantly associated with a higher hazard of incident frailty in Cox adjusted modeling (hazard ratio, 1.66; 95% CI, 1.02 to 2.69). In the sensitivity analysis the strength of the association remained relatively unchanged. When analyses were adjusted for mean eGFR over the first 24 months in lieu of baseline eGFR, eGFR decline remained associated with frailty incidence (hazard ratio, 1.55; 95% CI, 1.01 to 2.40). Neither adding death to the model as a confounder nor change with eGFR expressed as a percentage of baseline eGFR significantly alter associations. Interestingly, baseline eGFR either as a continuous or as a binary variable (eGFR<60 ml/min per 1.73 m²) was not associated with frailty incidence in adjusted Cox modeling.

Perhaps the most striking contribution of this work is the finding that fast eGFR loss was associated with subsequent frailty despite relatively preserved eGFR among MAPT participants. Adding to the novelty of this study was the use of the Fried frailty criterion, which has been widely validated across many cohorts. The ability to examine eGFR change over time utilizing multiple time points for slope assessments was an added advantage. Sensitivity analysis utilizing multiple eGFR cutoff variations contributed to the robustness of this study. Some may argue that the use of serum creatinine to estimate kidney function could complicate the ability to have any firm conclusions on the association of kidney function with frailty. However, a previous study showed similar associations using eGFR calculated using serum cystatin C measures (6). Among 3459 CHS participants, frailty incidence in addition to prevalence has been studied. Among the CHS participants who were not frail at study inception, lower eGFR_cys was associated with a higher risk of incident frailty (6). In that study, the association of eGFR_cys with incident frailty was graded with the strongest association at an eGFR range of 15–44 ml/min per 1.73 m² (incidence rate ratio, 2.08; 95% CI, 1.09 to 3.98). Although these studies may not be directly comparable as the MAPT cohort had a relatively preserved eGFR whereas the CHS cohort had more advanced CKD, both studies do suggest that a lower eGFR contributes to subsequent frailty.

The primary limitation of the study by Guerville et al. (8) was that, despite adjustment for multiple known contributors to both frailty and kidney function, residual confounding may still exist. The severity of comorbidity burden over time was not accounted for leaving the possibility that, worsening heart failure or diabetes for instance, could partly explain study findings. C-reactive protein was the only marker of inflammation used to adjust for the burden of chronic inflammation. Other markers of inflammation have been shown to mediate associations with frailty and CKD (10). Proteinuria and other CKD-related complications, including anemia, acidosis, and disorders of vitamin D handling, which may contribute to muscle weakness and decreased functional status—key components of frailty—were also not included in study analysis. The MAPT trial was not designed to collect such detailed information and thus the data were not available to study investigators. Nevertheless, the study findings do have a few clinical implications.

From the clinical perspective, evaluating the rate of eGFR decline may be a more important indicator of frailty risk than either a single eGFR measure or CKD stage alone. The finding that eGFR slope is an important indicator of frailty incidence stresses the importance of routine screening of all patients with early CKD. Although studies to evaluate whether the findings of Guerville et al. (8) are consistent in a more heterogeneous cohort of patients are needed, the high prevalence of frailty even in nonelderly patients with CKD may partially be explained.

In summary, rapid decline of eGFR in a cohort of elderly individuals with relatively preserved baseline kidney function was associated with incident frailty. Given that baseline eGFR was not associated with incident frailty but rapid decline was, is it plausible that kidney disease causes frailty or is a rapid decline in eGFR another bioassay for frailty? Study results should be confirmed in a more heterogeneous patient population. Findings raise important screening implications for the CKD patient population at large.

Disclosures

Dr. Delgado has nothing to disclose.

Funding

Dr. Delgado’s work is supported by the Department of Veterans Affairs, Clinical Science Research and Development Program under Career Development Award 1IK2CX000527-01A2.