Validity of the NINDS traumatic encephalopathy syndrome criteria for predicting chronic traumatic encephalopathy

Importance: Accurate prediction of chronic traumatic encephalopathy (CTE) remains challenging in life. Objective: To assess the reliability and validity of the NINDS traumatic encephalopathy syndrome (TES) criteria to predict CTE pathology in life. Design: Clinicopathological Diagnostic/Prognostic Study Setting: Six brain banks with varied recruitment criteria Participants: Brain donors were selected across 6 brain banks (15+ donors each), 5 age groups spanning ages 20 to 80+ (25+ donors each) and 9 repetitive head impact (RHI)/traumatic brain injury (TBI) groups (15+ donors each): (1) college or professional American football; (2) less than college football; (3) college or professional contact sports, non-football; (4) less than college contact sports, non-football; (5) military combat, no contact sports; (6) military combat and contact sports; (7) concussion with loss of consciousness, no RHI; (8) moderate to severe TBI, no RHI; (9) no RHI/TBI. Exposures: Blinded to neuropathological information, clinicians reviewed prospective study and medical records and conducted informant interviews, and an expert panel adjudicated TES diagnoses, including provisional levels of certainty for CTE pathology (suggestive/possible/probable). TES diagnoses were a priori dichotomized: TES with possible/probable CTE (CTEpos/prob) vs. no TES/TES with suggestive CTE (CTEsug). Main Outcomes and Measures: Blinded to clinical information, neuropathologists applied NINDS/NIBIB CTE neuropathological criteria and staging (I-IV). CTE diagnoses were a priori dichotomized: stages II-IV vs. no CTE/stage I. Results: Among 193 brain donors [men:153 (79.3%), mean age:66.4 (SD:22.0)], 57 (29.5%) donors met clinical criteria for CTEpos/prob and 42 (21.8%) donors met neuropathological criteria for CTE stages II-IV. There was high agreement between panelists for CTEpos/prob vs. no TES/CTEsug (ICC:0.95, 95%CI:0.88-0.97). CTEpos/prob sensitivity, specificity, positive likelihood ratio (LR) and negative LR for CTE stages II-IV were: 0.77 (95%CI:0.64-0.89), 0.84 (95%CI:0.78-0.90), 4.8 (95%CI:3.02-7.61), 0.28 (95%CI:0.15-0.50); age[&ge;]50:0.90 (95%CI:0.80-1), 0.90 (95%CI:0.85-0.96), 9.2 (95%CI:4.9-17.27), 0.11 (95%CI:0.04-0.33). All younger false positives (age<50; n=13) had a mental health, substance use and/or pain disorder. All older false positives (age[&ge;]50; n=11) had non-CTE neurodegenerative and vascular pathologies. Among 10 false negatives, 8 had stage II CTE. Conclusions and Relevance: The NINDS TES criteria demonstrated good reliability, sensitivity and specificity, and provided moderate to large evidence to both rule out and rule in CTE pathology, particularly above age 50.