Year: 2024
Type: article
Study design: Observational Study
Abstract: To the Editor, Hemodialysis (HD) patients have high rates of cognitive impairment and features of cerebral ischemic injury on brain imaging [1, 2]. The pathophysiology underlying these changes in the HD population is likely multifactorial, however, recent studies have implicated the HD process as a risk factor for cerebral injury and cognitive decline. Specifically, hypotension induced during HD treatments may lead to cerebral hypoperfusion in ESKD patients, who often have vascular disease and reduced cerebrovascular reserve [3–7]. This hypoperfusion may lead to brain injury and subsequent cognitive decline. In this study we assessed cognitive function and intradialytic cerebral oxygen saturation in an HD cohort. We hypothesize that patients who experience a greater decline in cerebral oxygen saturation during HD will have worsening cognitive function. This can help clarify the role of HD and intradialytic cerebral perfusion cognitive impairment. We recruited prevalent in-center HD participants ≥50 years of age who had initiated dialysis within the last 2 years. We measured cerebral oxygen saturation during HD sessions at baseline and 3, 6, 9 and 12 months using near-infrared spectroscopy with the SenSmart X-100 Universal Oximetry System (Nonin Medical, Plymouth, MN, USA). The decline in cerebral perfusion was calculated as the percentage decline in cerebral oxygen saturation area under the curve (AUC) (details in the supplementary materials). At baseline and 12 months, participants completed the National Institutes of Health (NIH) Toolbox Cognition Battery that evaluates domains of language, attention, processing speed, executive function, working memory and episodic memory and has three composite scores (see Table 1). We excluded domain-specific cognitive tests if the participant was unable to complete that test. Participants also completed the NIH PROMIS Cognitive Function computer adaptive testing survey version 2.0. We used a paired t-test to assess the change in cognition between baseline and 1 year and linear regression models to evaluate the relationship between the decline in cerebral perfusion and cognitive changes. Two participants with a negative decline in AUC were removed from the linear regression analysis, as it was unclear if a negative AUC represented an increase in cerebral oxygen saturation or is confounded by some interference. P-values <.05 were considered significant for all statistical testing. R version 4.3.3 (R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical testing. Detailed methods are available in the supplementary materials. Cognitive scores at baseline and the change in score associated with intradialytic cerebral oxygen saturation. There are seven assessments in the NIH Toolbox and three composite scores of fluid (memory, executive function, attention, processing speed), crystallized (language) and total cognition. The mean baseline score is a fully adjusted t-score and accounts for age, race, ethnicity and education levels. P-values are for comparison with the general population normative score, which is 50 (SD 10) for all measures. A lower score is worse for all tests. The β estimate is for the linear regression between the change in the uncorrected standard cognitive score and the decrease in the AUC. n refers to the number of participants included in the linear regression analysis for the β-estimate. The number is reduced, as one participant did not complete the picture sequence memory test and one did not complete the oral reading test. This decreased the n for the composite scores of fluid and crystallized by one each and the total score by two. SD: standard deviation. Cognitive scores at baseline and the change in score associated with intradialytic cerebral oxygen saturation. There are seven assessments in the NIH Toolbox and three composite scores of fluid (memory, executive function, attention, processing speed), crystallized (language) and total cognition. The mean baseline score is a fully adjusted t-score and accounts for age, race, ethnicity and education levels. P-values are for comparison with the general population normative score, which is 50 (SD 10) for all measures. A lower score is worse for all tests. The β estimate is for the linear regression between the change in the uncorrected standard cognitive score and the decrease in the AUC. n refers to the number of participants included in the linear regression analysis for the β-estimate. The number is reduced, as one participant did not complete the picture sequence memory test and one did not complete the oral reading test. This decreased the n for the composite scores of fluid and crystallized by one each and the total score by two. SD: standard deviation. At baseline, 43 participants completed the NIH Toolbox Cognition Battery and 26 completed it at 12 months. Two participants had partial completion, leaving 24 with a full complement of scores at baseline and 12 months. The mean age was 64 years, 73% were male, 42% Black and 46% were White non-Hispanic, 65% had diabetes mellitus and 81% had hypertension (see Supplementary Table S1). The mean decrease in cerebral oxygen saturation AUC during the HD sessions was 2.6% (standard deviation 1.3) for 24 participants after excluding the two negative AUC declines. Table 1 shows the number of participants included in each linear regression analysis, which ranged from 22 to 24. Tests of executive function, attention, processing speed and total cognition were significantly lower in the HD cohort than general population normative scores, but participant self-reported cognition (PROMIS survey) was similar to that of the general population (see Table 1). The only change in cognitive performance scores from baseline to 12 months was a decrease in the List Sorting Test for working memory (−3.4, P = .04) and in the PROMIS score (−3.8, P = .03) (see Supplementary Table S2). A greater decrease in cerebral oxygen saturation during HD was significantly associated with a decrease in the pattern comparison (processing speed), fluid cognition and total cognition scores. See Table 1 for the results of the linear regression and Fig. 1 for the graphical relationship between the decrease in cerebral oxygen saturation during HD and the change in 12-month total cognition score. Relationship between the decline in cerebral oxygen saturation during HD and the change in the 12-month total cognition score. A greater decrease in the AUC of cerebral oxygen saturation was associated with a decrease in the total cognition score over a 12-month period. In this longitudinal HD cohort study, we found that a greater decrease in cerebral oxygen saturation during HD was associated with worsening cognition. This relationship was notable for processing speed, fluid and total cognition scores. Also, executive function and memory showed similar trends of worsening scores associated with a greater decrease in cerebral oxygen saturation. Slow processing speed, executive function and memory are known areas of deficit in the HD population [1, 8]. The significant association observed between reduced cerebral oxygen saturation and a decrease in composite cognition scores, but not all domain- specific cognitive scores, may indicate that intradialytic cerebral perfusion affects some domains more than others but together significantly impact overall cognitive function in HD patients. The fluid cognitive score is made up of cognitive domains that are thought to be more susceptible to neurobiological injury in contrast to the crystallized (e.g. language) domains [9]. Our results are consistent with prior work that has demonstrated a decrease in intradialytic cerebral perfusion, as measured with cerebral oxygen saturation and cerebral blood flow velocity, is associated with worsening global cognition [10]. Despite the relationship between cerebral hypoperfusion and cognitive decline, only the List Sorting Test of the working memory score in the Toolbox and the patient-reported PROMIS cognition score had an overall decrease over 1 year in our cohort. This may be due to floor effects, as there were significant baseline deficits in the fluid cognitive assessments (those more susceptible to decline), but of those, the List Sorting Test scores were the highest, which may have allowed for the detection of change over the 1 year in that assessment. The decrease in the PROMIS score may indicate patients do become aware of impairments over time. The main limitation was our small HD cohort size. However, our cohort is largely reflective of the US HD population, thus our results are likely generalizable to the US HD population. Further, we did not specifically measure cerebral perfusion or blood flow but used cerebral oxygen saturation as our surrogate marker. The advantage of this method is that it allowed us to monitor the whole HD session (rather than intermittent monitoring that may miss episodic hypoperfusion), was well tolerated by patients and was easily completed in a community HD unit during a routine HD session. In conclusion, using comprehensive measures of cognitive performance outcomes, we demonstrated that the decrease in cerebral oxygen saturation experienced during HD is associated with worsening cognition. Our results reinforce concerns that intradialytic cerebral hypoperfusion contributes to the cognitive decline in persons undergoing HD. Future investigations should seek to clarify the mechanism of this association, focusing on vascular changes that may be contributing. Funding was received from the National Institute of Diabetes and Digestive and Kidney Diseases (5K23DK113119 to D.F.W.) and the National Center for Advancing Translational Sciences (TL1TR001437 to W.T.R.). The funders had no role in the data collection, data analysis and reporting of this study. D.F.W., B.D.S. and W.T.R. were responsible for the project idea and design and data collection and processing. D.F.W., B.D.S., W.T.R. and A.M.M. were responsible for the analysis and interpretation of data and wrote and edited the manuscript. The authors declare no conflicts of interest.
DOI: 10.1093/ndt/gfae101
Language: English
Subfield: Nephrology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W4396560950