A96-09 Characterizing Cognitive Resilience in Survivors of Critical Illness

Rationale Cognitive resilience, defined as better-than-expected cognitive function after an acute stressor, is poorly understood among ICU survivors. We identified survivors with varying levels of cognitive resilience and characterized differences in biological, clinical, and social factors. Methods We included participants ≥18 years old with respiratory failure and/or shock from the BRAIN-ICU and MIND-ICU prospective cohort studies that enrolled at academic, community, and Veterans Affairs (VA) hospitals. Using proportional odds logistic regression, we created a predictive model for global cognition scores, as measured by the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), using variables associated with cognitive impairment including age, sex, BMI, comorbidities, education, and pre-ICU cognition measured by Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE). For each participant, we predicted the “expected” global cognition score at 3- and 12-month follow-up. To account for competing risks of mortality and attrition, actual RBANS scores (with tests administered by study personnel) were scored as a composite outcome with participants who died being assigned the worst possible RBANS score minus 1 and those lost to follow-up being assigned the worst possible score. Participants with actual RBANS scores within ± 7.5 points (0.5 SD - minimal clinically important difference) of their predicted score were categorized as “expected resilience”. Participants with actual RBANS scores more than 7.5 points above their predicted score were categorized as “higher resilience” (i.e., better cognition) while those with actual scores more than 7.5 points below predicted were categorized as “lower resilience” (i.e., worse cognition). We compared pre-specified biological, clinical, and social factors between groups using the Kruskal-Wallis test. Results We included 781 participants who were a median [IQR] of 62 [52-71] years old and 41% female with an enrollment SOFA of 9 [7-11]. At 3-month follow-up, 265 (34%) participants demonstrated higher resilience, 253 (32%) demonstrated lower resilience, and 263 (34%) demonstrated expected resilience. At 3 months, participants with higher resilience had lower mean IL-6 levels early in critical illness, fewer days of severe sepsis, delirium, and coma, and were more likely to be treated at a VA hospital than participants with lower or expected cognitive resilience. At 12 months, these factors remained significant except for IL-6. Other predictors, including rurality and socioeconomic deprivation, were not associated with differential resilience. Conclusions Lower IL-6 levels, fewer days of severe sepsis, delirium, and coma, and treatment at a VA hospital were associated with higher cognitive resilience in ICU survivors. This abstract is funded by: Department of Veterans Affairs (IK2RX004799)