Abstract Rationale Delirium commonly complicates critical illness and contributes to disability, mortality, and cognitive decline. Multi-organ failure and treatments including sedation confound delirium detection and staging. A blood biomarker algorithm to grade neuronal injury during critical illness might improve diagnosis or prognostication, and guide future interventional trials. We hypothesized that total tau protein - a microtubule-binding protein implicated in neurodegeneration - and neurofilament light chain (NFL) - a neuron-specific cytoskeletal protein - would associate with delirium duration, survival, and subjective cognitive difficulty. Methods We performed a multicenter cohort study of critically ill adults with cardiopulmonary failure at high risk for ARDS, pneumonia, and/or sepsis. Participants were enrolled within 48h of qualifying organ failure, defined as shock requiring vasopressors or respiratory failure requiring invasive or non-invasive mechanical ventilation or supplemental oxygen ≥ 6 liters per minute flow, and underwent plasma sampling. Clinical data including SOFA score and daily RASS and CAM ICU scores were collected for 7 days. Delirium and coma free days (dcfd) within 8 days were calculated between days 0 and 7. We assessed survival at 28 days and surveyed using the WHO Disability Assessment Scale 2.0 (WHODAS) at 3 months, focusing on the ‘difficulty concentrating’ domain to measure subjective cognition. We measured tau and NFL using electrochemiluminescence and used logistic (mortality, WHODAS dichotomized at moderate or worse) or ordinal (dcfd) regression of log-normalized values to test association with outcomes adjusting for age, sex, and SOFA score at enrollment. Results Day 0 plasma was available for 402 of the first 500 participants. Delirium assessments were missing for 65; 115 survivors completed WHODAS. Admission plasma tau strongly associated with mortality (OR 1.39, 95% CI 1.09, 1.78 per log, p = 0.007) and with fewer adjusted dcfd (coefficient -0.42, 95% CI -0.64, -0.21, p 0.001). NFL similarly associated with adjusted mortality (1.43 per log, 95% CI 1.11, 1.84, p = 0.005) and fewer adjusted dcfd (-0.30, 95% CI -0.51, -0.09, p = 0.004). Difficulty concentrating at 3 months associated with plasma tau (p = 0.04). Discussion Plasma tau and NFL, markers of neural injury in dementia and brain injury, identify critically ill patients at high risk for delirium and death. Admission tau may predict future subjective cognitive complaints. These markers may identify a high risk population for future trials to decrease post-ICU cognitive impairment and may represent potentially modifiable intermediate phenotypes in model systems testing preventative interventions. Ongoing work will focus on markers’ predictive ability for long term cognitive function. This abstract is funded by: NIH HL168419, NIH HL168478, NIH HL168416, HL168145, HL168415, NIH HL168412, NIH HL168308
Meyer et al. (Fri,) studied this question.