Abstract: FR-PO884

Implications of Variation in Cognitive Performance on Dialysis: A Pilot Study of an Electronic Cognitive Battery

Session Information

Category: Dialysis

  • 607 Dialysis: Epidemiology, Outcomes, Clinical Trials - Non-Cardiovascular

Authors

  • Harhay, Meera Nair, Drexel University College of Medicine, Philadelphia, Pennsylvania, United States
  • Robinson, Lucy, Drexel University , Philadelphia, Pennsylvania, United States
  • Arif, Hasan, Drexel University College of Medicine, Philadelphia, Pennsylvania, United States
  • Ranganna, Karthik M., Drexel University College of Medicine, Philadelphia, Pennsylvania, United States
  • Schultheis, Maria T, Drexel University , Philadelphia, Pennsylvania, United States
Background

Cognitive impairment (CI) is common among dialysis patients, and some patients may exhibit transient CI during dialysis therapy. However, traditional neurocognitive testing has limited feasibility for use in clinical settings, and no studies have examined whether a decline in cognition during dialysis predicts health outcomes. We examined the association of variation in performance on the electronic Cogstate Brief Battery, consisting of four card-based cognitive tasks called Detection (processing speed), Identification (attention), One Card Learning (visual memory) and One Back (working memory), on the risk of subsequent hospitalization in a cohort of hemodialysis patients.

Methods

We enrolled 28 participants from a single dialysis unit in Philadelphia, PA. Participants completed the 10-minute Cogstate battery twice, once prior to dialysis and again during the last hour of dialysis therapy. For both sessions, participants were defined as “low” scorers on individual tasks if they scored below the cohort median score. We estimated age-adjusted Poisson regression models for the association of pre-dialysis and intradialytic cognitive performance and the number of hospitalizations participants experienced during six months of follow-up.

Results

The average age of the cohort was 58 years (SD 13), 97% were black, 60% were female. Mean dialysis duration was 6 years (range 2-20). After six months of follow-up, 35% of the cohort had at least 1 hospitalization, and 18% of participants (n=5) had >1 hospitalization. In age adjusted Poisson regression models, compared to participants who scored higher than the median both times for the Identification task, those who scored 1) higher pre-dialysis and lower during dialysis and 2) those that scored lower both times had a 9-fold and 7.5-fold increase in the expected number of hospitalizations in six months, respectively (p=0.03 for both). Reasons for hospitalization in low scorers included syncope and mechanical fall.

Conclusion

In this pilot study, we found that consistently low performance on a self-administered electronic cognitive task, or poorer performance during dialysis, signaled hospitalization risk among prevalent hemodialysis patients. Given limitations due to low sample size, future studies should confirm and expand on these findings in larger cohorts.

Funding

  • NIDDK Support