Subjective and objective cognitive function among older adults with a history of traumatic brain injury: A population-based cohort study

BACKGROUND Traumatic brain injury (TBI) is extremely common across the lifespan and is an established risk factor for dementia. The cognitive profile of the large and growing population of older adults with prior TBI who do not have a diagnosis of dementia, however, has not been well described. Our aim was to describe the cognitive profile associated with prior TBI exposure among community-dwelling older adults without dementia-an understudied but potentially vulnerable population. METHODS AND FINDINGS In this population-based cohort study, we studied 984 community-dwelling older adults (age 51 y and older and their spouses) without dementia who had been randomly selected from respondents to the 2014 wave of the Health and Retirement Study to participate in a comprehensive TBI survey and who either reported no prior TBI (n = 737) or prior symptomatic TBI resulting in treatment in a hospital (n = 247). Mean time since first TBI was 38 ± 19 y. Outcomes assessed included measures of global cognitive function, verbal episodic memory, semantic fluency, and calculation as well as a measure of subjective memory ("How would you rate your memory at the present time?"). We compared outcomes between the two TBI groups using regression models adjusting for demographics, medical comorbidities, and depression. Sensitivity analyses were performed stratified by TBI severity (no TBI, TBI without loss of consciousness [LOC], and TBI with LOC). Respondents with TBI were younger (mean age 64 ± 10 y versus 68 ± 11 y), were less likely to be female, and had higher prevalence of medical comorbidities and depression than respondents without TBI. Respondents with TBI did not perform significantly differently from respondents without TBI on any measure of objective cognitive function in either raw or adjusted models (fully adjusted: global cognitive function score 15.4 versus 15.2, p = 0.68; verbal episodic memory score 4.4 versus 4.3, p = 0.79; semantic fluency score 15.7 versus 14.0, p = 0.21; calculation impairment 22% versus 26%, risk ratio [RR] [95% CI] = 0.86 [0.67-1.11], p = 0.24). Sensitivity analyses stratified by TBI severity produced similar results. TBI was associated with significantly increased risk for subjective memory impairment in models adjusted for demographics and medical comorbidities (29% versus 24%; RR [95% CI]: 1.26 [1.02-1.57], p = 0.036). After further adjustment for active depression, however, risk for subjective memory impairment was no longer significant (RR [95% CI]: 1.18 [0.95-1.47], p = 0.13). Sensitivity analyses revealed that risk of subjective memory impairment was increased only among respondents with TBI with LOC and not among those with TBI without LOC. Furthermore, the risk of subjective memory impairment was significantly greater among those with TBI with LOC versus those without TBI even after adjustment for depression (RR [95% CI]: partially adjusted, 1.38 [1.09-1.74], p = 0.008; fully adjusted, 1.28 [1.01-1.61], p = 0.039). CONCLUSIONS In this population-based study of community-dwelling older adults without dementia, those with prior TBI with LOC were more likely to report subjective memory impairment compared to those without TBI even after adjustment for demographics, medical comorbidities, and active depression. Lack of greater objective cognitive impairment among those with versus without TBI may be due to poor sensitivity of the cognitive battery or survival bias, or may suggest that post-TBI cognitive impairment primarily affects executive function and processing speed, which were not rigorously assessed in this study. Our findings show that among community-dwelling non-demented older adults, history of TBI is common but may not preferentially impact cognitive domains of episodic memory, attention, working memory, verbal semantic fluency, or calculation.