Patients receiving haemodialysis for chronic kidney disease (CKD) who are frail—however that is defined—have been found to be at greater risk of suffering cognitive impairment while receiving the treatment. “To our knowledge,” say the authors behind a new study, led by Benjamin M Anderson (University of Birmingham, Birmingham, UK), “these data are the first to suggest such an interaction between cognition and frailty”.
Anderson and his colleagues, including corresponding author Adnan Sharif (University of Birmingham, Birmingham, UK) note in their introduction the frequency of cognitive impairment among haemodialysis patients, something which they state “worsens in both incident and prevalent haemodialysis populations”. Though they point out that there are many risk factors for cognitive impairment that are consistent across both the general population and haemodialysis patient population, one that is significantly higher among the latter is frailty.
Defining frailty as a “syndrome of increased vulnerability to poor resolution of homeostasis after a stressor event” that is associated with “mortality, hospitalisation and disability,” they highlight that estimated frailty rates for the haemodialysis population range “from 26% to 63%”. They note that there are several possible measures of frailty, including the Frailty Phenotype (FP), Frailty Index (FI), Montreal Cognitive Assessment (MoCA), Edmonton Frail Scale (EFS and Clinical Frailty Scale (CFS), but also point out that previous work utilising the data from the Frailty Intervention Trial in End-Stage patientS on haemodialysiS (FITNESS) cohort, which is used for Anderson et al’s study, “found agreement upon frailty status between these tools is poor”. Nevertheless, “they are all associated with greater mortality”, while a systematic review and meta-analysis led by Marcus Kiiti Borges (University of São Paulo, São Paulo, Brazil) in 2019 found an association between frailty and cognitive impairment in the general population.
To explore that association in the context of haemodialysis patients, Anderson et al drew on the data from FITNESS, which was a “two-stage study that follows a cohort multiple randomised controlled trial design”. The first of those stages is an assessment and long-term follow up with haemodialysis patients whose frailty and bio-clinical status were measured at baseline.
The study enrolled 448 patients from one nephrology centre in Birmingham, England, and were identified using electronic patient records and encompassed any over 18 years, receiving haemodialysis for three months or more, and capable of giving written consent. The two outcomes measured were mortality and hospitalisation. Cognitive impairment was defined as “MoCA scores of <26, or <21 in dexterity impairment, <18 in visual impairment”.
From the cohort, 346 (77.2%) were identified as experiencing cognitive impairment. Increases in frailty “by all definitions” was associated with MoCA-measured impairment. The study authors state that the interaction between MoCA score and increased frailty was present when measuring FI (p=0.002) and CFS (p=0.005), and that “admissions were highest when both MoCA and frailty scores were high, and when both scores were low”.
There was not, however, any direct association found between impaired cognition and mortality (HR 0.99, 95% C.I. 0.95-1.03, p=0.41) or hospitalisation (IRR 1.01, 95% C.I. 0.99-1.04, p=0.39) on multivariable analyses. Univariable analyses did yield an association, as did multivariable analyses for the FP, FI and CFS measures of frailty. Anderson and colleagues also state: “No significant interactions between frailty and cognition were identified upon mortality.” Anderson et al suggest this result, which is different from those of other haemodialysis cohorts, may be a result of the various different definitions of cognitive impairment. It may also suggest that their study, which adjusted for age where others did not, may highlight that others were influenced by the age of the patients enrolled more than by their cognitive impairment status. They suggest that the greater “granularity” of their data increases the
The authors also note that “discordance between frailty and cognition” was the strongest predictor of hospital admission in their results, meaning “severe frailty with no cognitive impairment or vice versa”. They were fewest, meanwhile, “in the absence of frailty or cognitive impairment” as well as “perhaps surprisingly in the setting of severe frailty with severe cognitive impairment”. Anderson and colleagues put forward the suggestion that the latter result may be because those experiencing severe frailty and cognitive impairment may be subject to more assistance from caregivers than those experiencing just one of the two.
Reviewing their study, the authors make the case that its inclusion of “diversity of demographics, comorbidities and socio-economic backgrounds” give its results credibility. They do pick out several limitations, including its cross-sectional nature—something they say may fail to reflect the “year-to-year variability” of frailty, which they describe as a “dynamic state”. The study only examined maintenance dialysis patients rather than incident ones, and “small variations in application of frailty tool criteria from other studies” may have had an influence on the results. As a result, the authors caution against overestimating the immediate applicability of the findings to clinical practice.