A small but significant association between a history of TBI and cognitive decline in a study of older veteran male twins

By Susan Klein, MD, PhD. This article was initially published in our Concussion Update newsletter; please consider subscribing.

“Veterans who reported having at least one traumatic brain injury or TBI at any time in their life had lower cognitive scores in later life, compared to twin siblings who did not experience a TBI,” observed Dr. Chanti-Ketterl during a recent podcast about a twin study of WWII veterans (published in Neurology), of which she was the first author. Three variables increased the small but significant rate of cognitive decline in those who had sustained a TBI compared to their uninjured co-twins: more than one TBI, a more severe TBI, and having a TBI after the age of 24.  The rate of cognitive decline was faster in identical twins with a history of TBI.

This study of veterans, carried out an average of 34 years after their inclusion in the twin registry, provides another window into environmental factors that predispose older adults to dementia, as well as the pervasive effects of even one TBI significant enough to merit medical assessment. The study findings applied to twins with all levels of TBI, including mild TBI.  This study is exciting because it suggests that an environmental factor, in this case head injury, can be a contributor to cognitive decline. What that means is that prevention and early, effective treatment of TBI can potentially have a significant role in mitigating cognitive decline as we age.   

Using serial assessments with a global cognitive measure (modified Telephone Interview for Cognitive Status, or TICS-m) over 12 years, these investigators used the power of a twin design to isolate the contribution of a TBI on cognitive skills in later life. Participants were 8,662 WWII veterans in the Duke Twins Study of Memory in Aging, drawn from a larger NAS-NRC Twin Registry of 15,924 male veteran twin pairs born between 1917 and 1927.  Researchers defined TBI by participant report of a head injury significant enough to merit medical attention, labeling a severe TBI as one that involved loss of consciousness; they also reported their age when the TBI had occurred.  

The sample was divided into a control cohort of 1,195 twin pairs and the other remaining registry members participating in the Duke Twin Study of Memory in Aging, labeled the non-control cohort. Regarding the control cohort [use of this term is confusing, we know] of twin pairs, the twins were either identical MZ twin pairs who shared all genes, or fraternal DZ twin pairs, who shared half their genes in common.  Only one twin in each control cohort pair had sustained a TBI. Regarding the other remaining registry members or non-control cohort, this group included individuals whose twin was no longer alive/included, and others who did not have medical documentation of their twin status other than recruitment into the original twin registry. 

Random effects linear mixed regression models were used on the non-control group data to develop a good model for accounting for how TBI data points might affect the TICS-m scores over time, controlling for age, medical, and social demographic variables.  Then, the model design was tested on the control sample.  In twin studies, this design prevents “fishing” for the best result, and supports the robustness of the findings.  The authors also assessed the impact of genetic risk for dementia (ApoE e4 carriers vs. non carriers) in a smaller subset of 1,392 twin pairs.

These veterans, tested on average more than 30 years after their initial head injury, scored in the low 30s on the TICS-m testing, administered up to three times over 10 years.  The TICS-m test score ranges between 0 (a perfect score – no delays) and 50 (cognitive dysfunction.  Other validation studies on the TICS-m suggest that a score in the low 30 range may be predictive of cognitive impairment.

 In the matched co-twin control sample, where only one of the pair sustained a TBI, the twin with the TBI scored lower on the TICS-m than his unaffected twin. The authors observe: “The twin with TBI [after age 24 years] is declining 0.07 points faster per year than his co-twin without TBI.  Therefore, in 12 years of follow-up of this study, the co-twin with TBI would have steeper cognitive decline than his co-twin without TBI.”  Identical (MZ) twins were more likely to show this difference because they share the same genetic makeup, so any differences in their performance can be attributed to TBI.  Even though over 12 years, the differences between scores in the set of head-injured subjects and the set of their unaffected co-twins is less than a whole point on the TICS-m, that difference is statistically significant. 

ApoE e4 carriers also had lower TICS-m scores, which declined faster over time than non-carriers in the subset analysis, but being an ApoE e4 carrier did not intensify the effect on cognition of having had a TBI.

In an accompanying editorial, Elser and Schneider note the authors’ acknowledgment of the limitations of their study: TBI definitions and use of a sample of White men, most obviously.  Elser and Schneider add that genetic influences on complex disorders in the family of dementias can vary even in identical MZ twins because of changes in the genome made either by epigenetics (how your behaviors and environment can cause changes that affect the way your genes work) or other genetic variations (such as point mutations or mitochondrial sequence variations). They caution that such changes would weaken the study’s conclusion that TBI alone led to cognitive decline.  

However, the editorial writers agree with the authors and with us that the impact of TBI is important and long-lasting.  The power of good observation over many years can help us see how improved management of head injury can mitigate potential long-term cognitive impacts. We owe our veterans thanks for sharing this information with us.