Cerebrospinal Fluid Biomarkers and Cognitive Trajectories in Patients with Alzheimer’s Disease and a History of Traumatic Brain Injury

Trau matic brain injury (TBI) and Alzheimer’s disease (AD) have overlapping mechanisms but it remains unknown if pathophysiological characteristics and cognitive trajectories in AD patients are influenced by TBI history. Here, we studied AD patients (stage MCI or dementia) with TBI history (AD TBI+, n=110), or without (AD TBI, n=110) and compared baseline CSF concentrations of amyloid beta 1-42 (A β 42), phosphorylated tau181 (pTau181), total tau, neurofilament light chain (NfL), synaptosomal associated protein-25 (SNAP25), neurogranin (Ng), neuronal pentraxin-2 (NPTX2) and glutamate receptor-4 (GluR4), as well as differences in cognitive trajectories using linear mixed models. Explorative, analyses were repeated within stratified TBI groups by TBI characteristics (timing, severity, number). We found no differences in baseline CSF biomarker concentrations nor in cognitive trajectories between AD TBI+ and AD TBI-patients. TBI >5 years ago was associated with higher NPTX2 and a tendency for higher SNAP25 concentrations compared to TBI ≤ 5 years ago, suggesting that TBI may be associated with long-term synaptic dysfunction only when occurring before onset or in a pre-clinical disease stage of AD.


Introduction
Traumatic brain injury (TBI) affects about 50 million people globally and is a growing burden on healthcare (Feigin et al., 2013).This not only includes the direct disabling effects of TBI, but also the persistent symptoms and long-term chronic consequences of head trauma (Wilson et al., 2017).In fact, biological changes after TBI have overlap with neurodegenerative diseases, such as Alzheimer's disease (AD), although this connection lacks conclusive clinical evidence.Postmortem neuropathological examination in dementia-free individuals have revealed a higher burden of amyloid-beta (Aβ) plaques and phosphorylated tau (pTau) positive neurofibrillary tangles in individuals with a history of TBI compared to those without TBI (Johnson, Stewart and Smith, 2012).One hypothesis that can explain the link between TBI and AD pathology is that TBI-related axonal injury and progressive axonal degeneration may result in cytoskeletal disruption and axonal bulb formation, thus leading to the onset and accumulation of Aβ plaques and pTau tangles (Graham and Sharp, 2019;Johnson, Stewart and Smith, Abbreviations: AD, Alzheimer's disease; ADC, Amsterdam Dementia Cohort; APOE, Apolipoprotein E; Aβ, Amyloid beta; Aβ42, Amyloid beta 1-42; ANCOVA, Analysis of covariance; CSF, Cerebrospinal fluid; Df, Degrees of freedom; EEG, Electroencephalography; FDR, False Discovery Rate; GluR4, Glutamate receptor-4; IQR, Interquartile range; LOC, Loss of consciousness; MCI, Mild cognitive impairment; MMSE, Mini mental state examination; MRI, Magnetic resonance imaging; NfL, Neurofilament light chain; Ng, Neurogranin; NPTX2, Neuronal pentraxin-2; pTau, Phosphorylated tau; pTau181, Phosphorylated tau 181; SD, Standard deviation; SE, Standard error; SNAP25, Synaptosomal associated protein-25kDa; TBI, Traumatic brain injury; β, Standardized beta.2013).Synaptic dysfunction has also been observed in both experimental AD and TBI models.Shared pathophysiological processes include reduced synaptic density, decreased synaptic plasticity, or disruption of the excitation/inhibition balance (Li, Liang and Fu, 2021).Furthermore, synaptic dysfunction has been associated with increased Aβ and pTau load in AD and TBI models (Jamjoom et al., 2020;Li, Liang and Fu, 2021).Despite the suggested pathophysiological links between AD and TBI, these links are yet to be fully elucidated in human studies.
The investigation of cerebrospinal fluid (CSF) proteins offers an opportunity to detect altered pathophysiological brain processes in humans.Amyloid beta 1-42 (Aβ42), phosphorylated tau 181 (pTau181), and total tau are the core fluid biomarkers to diagnose AD, and may become abnormal years prior to AD symptom onset (Cohen et al., 2019;Jack et al., 2018).Neurofilament light chain (NfL) is a marker reflective of axonal damage and exhibits elevated concentrations in various neurological diseases (Bridel et al., 2019;Shahim et al., 2020), including sports-related mild TBI and moderate to severe TBI (Graham et al., 2021;Neselius et al., 2012).In addition, several pre-and postsynaptic CSF biomarker tests are now available to asses synaptic loss or degeneration, such as neurogranin (Ng), neuronal pentraxin 2 (NPTX2), or synaptosomal-associated protein 25kDa (SNAP25).The impact of previous TBI on these core AD biomarkers, neuroaxonal damage biomarkers and on synaptic biomarkers in patients with AD remains unknown, but it is conceivable that TBI in history affects the expression of these biomarkers in AD.
There is also limited knowledge whether TBI history impacts the clinical course in AD patients.Previous work from our group revealed no distinct baseline clinical features in AD patients with TBI history (van Amerongen et al., 2022).One study reported that TBI history within 10 years of AD-type dementia onset was associated with a more rapid decline in functional abilities (Gilbert et al., 2014), but other studies did not find distinct clinical trajectories (LoBue et al., 2018;Tripodis et al., 2017).
Taking all aforementioned together, one may hypothesize that TBI history within the context of AD is associated with more pronounced or accelerated pathophysiological processes, reflected by altered CSF biomarker concentrations, potentially associated with a faster cognitive decline.Studying this association could yield valuable insights into the underlying pathophysiological mechanisms linking TBI and AD, and aiding in predicting disease course in patients with AD and TBI history.Furthermore, such research may potentially identify new therapeutic targets aimed at preventing or slowing cognitive decline in AD patients with TBI history.Therefore, the objective of this study is to examine whether AD patients with a history of TBI demonstrate distinct pathophysiological characteristics as reflected in the CSF, and whether this associates with rates of cognitive decline.

Sample selection
This is a longitudinal case-control study that included two groups of patients selected from the Amsterdam Dementia Cohort (ADC): (1) AD patients with history of TBI, and (2) AD patients without history of TBI (van der Flier et al., 2014).Individuals were included when there was evidence of biological AD based on their CSF biomarker status, including a baseline syndrome diagnosis of possible or probable AD dementia, or mild cognitive impairment (MCI) according to diagnostic criteria (Albert et al., 2011;McKhann et al., 2011).In addition, only patients with available CSF in the biobank were eligible for this study.Consensus diagnosis was made according to applicable guidelines (Jack et al., 2018;van der Flier et al., 2014), during a multidisciplinary consensus meeting following an extensive cognitive screening day including: neurological examination, neuropsychological testing, brain magnetic resonance imaging (MRI), electroencephalography (EEG), CSF biomarker measurements, and Apolipoprotein E (APOE) genotyping (van der Flier et al., 2014).This study was approved by the Medical Ethical Committee of the Amsterdam UMC, location VUmc and all participants provided written informed consent.

TBI assessment and matching
We searched in the medical history files of 480 eligible patients for information about previous traumatic brain injury/injuries.In addition, we obtained information from a baseline questionnaire including concussion history, stating: "have you ever suffered from a concussion".If answering "yes", the next question asked was: "Have you also lost consciousness"?and "Which year(s) have this been?"If both of these sources did not indicate a history of concussion and/or TBI, the patient was considered TBI negative.TBI positive patients were further classified as with or without reported loss of consciousness (LOC) or as single or multiple TBI (2 or more TBI sustained in history).Also, the year of each TBI was noted.We detected 110 patients that were TBI positive (AD TBI+ ).These patients were subsequently 1:1 matched to 360 TBI negative patients (AD TBI-), based on 'optimal matching' procedure with propensity score distances on four matching criteria: age, sex, baseline mini mental state examination (MMSE) score, and baseline disease stage (MCI or dementia).Matching procedures were performed in RStudio with package "MatchIt" (Ho et al., 2011).Adequate matching was obtained, with all absolute standardized mean matching differences below 0.10.The final cohort thus consisted of 110 AD TBI+ (89 dementia, 21 MCI) and 110 AD TBI-(92 dementia, 18 MCI) patients.
Besides screening for TBI history, all participants were screened for participation in contact and/or collision sports (i.e., soccer, rugby, fighting), based on the questionnaire stating: "Do you participate in sport, or have you ever participated in sport".If answering yes, the next question asked was: "Which sport and when did you participate?".Contact and/or collision sports participation was detected in 28 patients (16 AD TBI-and 12 AD TBI+ ), 19 patients had missing data.

CSF biomarkers
At the time of baseline diagnostic procedures, CSF was collected through lumbar puncture according to the standardized methods (van der Flier et al., 2014).Routine measurements of CSF Aβ42, pTau181, and total tau were performed using Innotest or Elecsys sandwich ELISAs.To align biomarker concentrations, we transformed Elecsys values into Innotest equivalents using a conversion formula that has been established previously (Willemse et al., 2018).AD biomarker positivity was determined by previously set cut points of the ratio between CSF pTau181 and Aβ42 (Willemse et al., 2021).Remaining CSF was aliquoted and stored in the Amsterdam Dementia Biobank (van der Flier et al., 2014).This aliquoted CSF was used to measure additional biomarkers, which was performed in 2022.In-house protopype immunoassays from ADx NeuroSciences were used to measure NfL (ELISA), NPTX2 (synaptic cleft protein, measured with ELISA), SNAP25 (pre-synaptic protein, measured with Quanterix Simoa HD-X), and glutamate receptor AMPA R4 (GluR4; post-synaptic protein, ELISA) (Andreasson et al., 2015;Bolsewig et al., 2022;Das et al., 2022;Das et al., 2023).Neurogranin (Ng; post-synaptic protein) was measured with a commercial ELISA from EuroImmun.Specific assay details can be found in the supplementary materials.NfL and SNAP25 measurements were missing in one individual, Ng was missing in two individuals, NPTX2 was missing in four individuals, and GluR4 was missing in 8 individuals.

Cognition
Global cognition was assessed using MMSE scores.Baseline MMSE was available for all individuals, follow-up MMSE was available for 183 individuals.In addition, four cognitive domains (attention, executive functioning, language, memory) were assessed, where domain scores were calculated from raw scores of neuropsychological tests (Table S1).
All neuropsychological test results (including MMSE) were standardized into test-specific z-scores using mean and standard deviation of baseline test results, which were also applied to longitudinal test results.The testspecific z-scores representing a specific cognitive domain were subsequently averaged into a domain z-score.For each cognitive domain score, a minimum of two test scores was required to provide a reliable outcome.Neuropsychological domain scores were available at baseline (n=219) and follow-up (n=152).In this cohort, the median number of visits was three (range: 1-13), and the mean follow-up duration since baseline AD or MCI diagnosis was 2.2 years with a maximum of 8.1 years.The total number of visits for all participants combined was 764.From these visits, we were able to calculate a total of 469 composite scores for attention, 448 for executive function, 459 for language and 510 for memory for the complete cohort consisting of 220 individuals.Educational level was classified with the Verhage score (range: 1-7) (Rijnen et al., 2020).

Data analyses
Log-transformation (log10) was applied to biomarker concentrations to optimize normal distribution of the residuals in the analyses.Concentrations were compared between AD TBI-and AD TBI+ with independent T-tests.Separate linear mixed models with random intercepts and random slopes were employed to test the effect of TBI history on MMSE and the four cognitive domain scores (standardized Z-scores), at baseline and over time.These models included TBI history (AD TBI-as reference group), follow-up time in years, the interaction between TBI history and years, as well as educational level to correct for potential confounding.To test whether the effect of TBI history on cognitive slopes was dependent on biomarker status, we subsequently repeated the linear mixed models now including an additional three-way interaction: AD TBI-/+ x biomarker concentration x follow-up time in years.Since the AD TBI-and AD TBI+ groups were age, sex, MMSE and disease stagematched, we did not consider these factors as potential covariates in our analyses.False Discovery Rate (FDR) was used to correct for multiple testing in case of significant results in primary analyses.Data is, therefore, presented with uncorrected p-values or FDR-corrected pvalues (FDR-p).
We subsequently conducted exploratory analyses to investigate the impact of several TBI characteristics on biomarker concentrations and cognitive trajectories in AD.To investigate the effect of timing of TBI, we divided the AD TBI+ group into two groups based on the number of years between the latest TBI and AD diagnosis; ≤5 years (TBI≤5yrs, n=33) or >5 years (TBI>5yrs, n=68).The selection of this cutoff point is based on the likelihood that patients would have developed symptoms within the five years preceding their diagnosis with AD in a tertiary care setting, thereby reflecting a TBI occurring in a clinical AD stage.Nine cases were excluded because of missing year of TBI.For TBI severity, we divided the AD TBI+ group into TBI with LOC (n=54) or without LOC (n=36) (n=20 was missing).To investigate the effect of number of TBI, we stratified the AD TBI+ group into single TBI (n=89) or multiple TBI (when ≥2 TBI in history, n=21).We used analyses of covariance (ANCOVA) with Tukey pairwise post-hoc tests to compare biomarker concentration among each of the three groups.Linear mixed models were used to assess differences in cognitive slopes between three TBI subgroups.With these new TBI subgroup divisions, age, sex, and disease stage-matching was lost, therefore we accounted the models for age, sex, and disease stage, and the cognition models additionally for educational level.
Lastly, two sensitivity analyses were performed for the primary analyses.First, to account for potential repetitive head injury exposure, we excluded patients that (previously) participated in contact sports and repeated our statistical analyses.Second, we repeated analyses only within AD dementia patients, to account for the fact that MCI patients may have different disease features than dementia patients.Both sensitivity analyses showed no differences in findings compared to the primary results, therefore these results are not shown (available upon request).

Demographics
The baseline demographic characteristics of our cohort of AD patients with TBI history (n=110) or without TBI history (n=110) (mean age 65.6 years) are detailed in Table 1.Within AD TBI+ , 60 % (n=54) had a TBI with LOC, and 19 % (n=21) indicated to have two or more TBIs in their medical history.Thirty three patients (33 %) suffered from a TBI within 5 years of AD diagnosis.Between the AD TBI-versus AD TBI+ groups, there were no differences in educational level or percentage of APOE-ε4 carriers.AD TBI+ had less follow-up data available than AD TBI- and consequently a shorter mean follow-up duration.Data are shown as frequency (N), with percentage (%), or mean ± standard deviation (SD).P-values for baseline biomarker comparisons were generated using independent T-tests (with log10 transformed concentrations; additional statistical details are displayed in Table S2), while those for baseline cognitive performance were generated using linear mixed models.AD: Alzheimer's disease, TBI: traumatic brain injury, LOC: loss of consciousness, FDR: false discovery rate, MMSE: mini mental state examination, Aβ42: amyloid beta 1-42, pTau181: phosphorylated Tau 181, NfL: neurofilament light, SNAP25: synaptosomal associated protein 25kDa, Ng: Neurogranin, NPTX2: neuronal pentraxin 2, GluR4: glutamate receptor 4.

CSF biomarkers
All raw biomarker concentrations are summarized in Table 1.No differences in any of the baseline log-transformed CSF core AD biomarker concentrations (Aβ42, pTau181, total tau), nor in the presynaptic (SNAP25), synaptic cleft (NPTX2), or post-synaptic (GluR4, Ng) biomarker concentrations were detected between AD TBI-and AD TBI+ patients at baseline.(Table 1, S2)

Cognitive trajectories
There were no differences in baseline cognitive scores (Table 1), nor differences in the rate of cognitive changes over time between AD TBI+ and AD TBI-(Table 2, Figure 1).We found no differences between AD TBI+ and AD TBI-in their associations of biomarker concentrations with cognitive performance at baseline nor with rate of cognitive change over time (p-values of interaction term all >0.05,Table 2).The interaction between AD TBI and NPTX2 on decline in attention over time was attenuated after FDR-correction (FDR-p = 0.225).We did observe associations of baseline biomarker concentrations with rate of cognitive change over time when investigating this over the total cohort (derived from the three-way interaction models: AD TBI-/+ x biomarker concentration x time in years): high baseline CSF NfL was most robustly associated with steeper cognitive slopes over time (significant associations with longitudinal MMSE and all four cognitive domain scores), followed by low baseline NPTX2 (association with longitudinal MMSE score and three cognitive domain scores), and high CSF total tau (association with longitudinal MMSE score and attention domain), whereas Aβ42, pTau181, SNAP25, Ng, and GluR4 did not show any associations with cognitive slopes in any of the cognitive domains over the total cohort (Table S3).

Discussion
In this study, we investigated the impact of TBI history on the pathophysiological characteristics and cognitive trajectories in Alzheimer's disease.Despite observing no differences in baseline concentrations of AD core biomarkers, NfL, or synaptic proteins, as well as no variations in cognitive slopes between AD patients with or without TBI history, minor alterations in synaptic biomarker concentrations of SNAP25 and NPTX2 but not in cognitive slopes were detected in those that sustained a TBI more than 5 years ago.While TBI history did not seem to impact cognitive slopes via altered pathophysiological processes represented by CSF biomarkers, our findings did highlight that high baseline NfL and low NPTX2 concentrations were associated with steeper rates of cognitive decline over time in individuals with AD compared to controls.Taken together, our findings suggested that TBI had limited influence on AD pathophysiology and its associated cognitive decline.However, the CSF biomarkers NfL and NPTX2 have prognostic value in AD.Understanding the prognostic utility of synaptic biomarkers in TBI may pave the way towards their use for disease and therapeutic intervention monitoring.This study thus represents a unique cohort focusing on the interplay of TBI, AD and CSF biomarkers that are reflective of various pathophysiological processes.
Several studies have demonstrated that axonal damage, as reflected by NfL, persists up to 18 years post-TBI, sometimes detected years after the injury (Graham et al., 2021;Neselius et al., 2012;Shahim et al., 2020).Furthermore, there is clinical evidence to suggest that patients with AD have higher CSF or plasma NfL concentrations compared to healthy individuals (Xiong et al., 2021), and is also predictive of the rate of cognitive decline in patients with MCI or AD-dementia (Bos et al., 2019;Mattsson et al., 2016).The latter finding aligns with our results and may reflect the potential of NfL for investigating effects of TBI in AD patients.In our study, however, clear evidence for pronounced neuro-axonal damage or degeneration, as reflected by NfL, was not found in patients with AD and a history of TBI.Established knowledge indicates that core AD biomarkers tend to return to normal concentrations within days to weeks after TBI (Neselius et al., 2012;Zetterberg et al., 2006), which might explain why we did not find differences in the baseline biomarker concentrations between the AD patients with and without TBI.
Synaptic dysfunction is a pathophysiological feature in several neurodegenerative conditions, including AD (Milà-Alomà et al., 2020).In fact, synaptic pathologies have been identified in pre-clinical stages of AD and there is ample clinical evidence for altered concentrations of CSF synaptic biomarkers in AD compared to healthy individuals.Elevation of CSF SNAP25 and Ng, for example, is found to be specific to AD, while studies report reduced concentrations of NPTX2 and GluR4 in patients with AD compared to cognitively healthy controls (Das et al., 2023; Linear mixed models tested the effect of TBI history on changes in MMSE and cognitive scores over time (Model: TBI x Time, AD TBI-as reference group), as well as the interaction between TBI and log-transformed (log10) biomarker concentration (Model: TBI x 'CSF biomarker' x Time).Models included years of education as covariate.Data is presented with standardized beta (β) and standard errors (SE).AD: Alzheimer's disease, TBI: traumatic brain injury, MMSE: mini mental state examination, Aβ42: amyloid beta 1-42, pTau181: phosphorylated Tau 181, NfL: neurofilament light, SNAP25: synaptosomal associated protein 25kDa, Ng: Neurogranin, NPTX2: neuronal pentraxin 2, GluR4: glutamate receptor 4. * FDR-corrected p-value: 0.225 Kivisäkk et al., 2022;Lleó et al., 2019).In the context of TBI, synaptic proteins have been explored as potential biomarkers only within experimental settings.SNAP25, for example, plays a crucial role in synaptic vesicle exocytosis and in modulating Ca 2+ ion channels (Das et al., 2023).Given that excess intracellular Ca 2+ is an identified pathological feature following TBI, this supplies a rationale that TBI may lead to degradation of synaptic vesicles and excess SNAP25 being released into CSF (Zhang et al., 2016).NPTX2 is a protein that plays a crucial role in synaptic plasticity and is linked to excitatory synapses associated with parvalbumin interneurons (Boiten et al., 2020).The loss of these interneurons has also been observed in experimental TBI studies (Frankowski, Kim and Hunt, 2019;Nichols et al., 2018), potentially leading to decreased excitation and, consequently, reduced NPTX2 concentrations.
Here, we found that AD patients with a remote TBI exhibit slightly higher SNAP25, along with slightly higher NPTX2 concentrations, compared to AD patients with a more recent TBI history.While this may imply that synaptic alterations related to TBI may occur in the earlier pre-disease or pre-clinical AD stages, the finding of higher (instead of lower) NPTX2 concentrations is unexpected.This may indicate an upregulation of NPTX2 as a protective mechanism.Nonetheless, as of now, this is the first clinical study to explore CSF synaptic biomarkers in the context of TBI.Longitudinal studies with repeated CSF collection following TBI patients at risk for a future AD diagnosis could further elucidate the synaptic pathophysiology.
In this study, the biomarkers NfL, NPTX2, and total tau were predictive of cognitive decline over time over the whole cohort.This is in line with previous evidence towards the association of these biomarkers with cognitive deterioration in dementias (Mattsson et al., 2016;Nilsson et al., 2024;Sämgård et al., 2010).However, we observed no differences in cognitive trajectories between AD patients with and without TBI history and the relationship between TBI history and cognitive decline was not dependent on specific biomarker concentrations.Additionally, we found no discernible effects of whether TBI occurred recently or further in the past.These findings align with previous studies that reported a lack of association between TBI with LOC and rates of cognitive decline in both AD patients and healthy control subjects nor with the  rate of progression from MCI to AD dementia in individuals with history of TBI (LoBue et al., 2018;Tripodis et al., 2017).One reason for not detecting differences in decline rates within this cohort may be related to the sample set of relatively young AD patients (mean age 65 years).Since early onset AD is associated with a more aggressive disease course than late-onset AD (Jacobs et al., 1994;Tort-Merino et al., 2022), the effect of TBI on the AD trajectory might be too subtle to discern in this cohort.Secondly, although numerous population-wide studies have indicated an elevated risk of dementia in individuals with a lifetime history of TBI (Fann et al., 2018;Nordström and Nordström, 2018;Plassman et al., 2000), most neuropathological studies have not found an increased risk of AD neuropathology in individuals with TBI history (Agrawal et al., 2022;Chosy et al., 2020;Crane et al., 2016;Sugarman et al., 2019).Therefore, the failure to detect differences within this cohort prompts a broader question regarding the true association of TBI with AD-specific neurodegeneration or cognitive decline.While the cohort is unique in its design, it has some limitations.First of all, though all patients had clinically and biomarker-supported AD, the risk of developing AD after TBI was not investigated in this longitudinal observational cohort study, so caution is warranted in drawing conclusions about the role of TBI as risk factor for AD.Secondly, to determine a lifetime history of TBI, studies commonly apply self-report assessments, either with or without a validated TBI assessment tool.While these tools find utility in various clinical and research contexts, they are susceptible to recall bias and the reliability of recalling such information within a population of cognitively impaired elderly individuals is questionable (Weiner et al., 2017).We used a self-report questionnaire, focusing specifically on concussions.Unfortunately, other types of TBI, such as brain contusions or traumatic skull fractures may not have been identified with this questionnaire.To address this limitation, we incorporated hospital-based medical records and information obtained through a medical history assessment conducted by physicians, which were likely to detect these other types of more severe TBI.Nevertheless, it remains possible that not all TBI events were captured with this method.The inclusion of athletics history to identify potential exposure to repetitive head impacts strengthens this study.Head impacts sustained in contact sports may not always be recognized as TBI or may occur on a sub-concussive level, even though they are also associated with long-term neurological consequences (McKee et al., 2023).Another limitation concerns the differences in follow-up visits between the two study groups, resulting in less follow-up data available in AD patients with TBI history.This may limit the identification of true differences in clinical trajectories within the groups.Finally, measures of TBI severity such as the Glasgow coma scale (GCS) scores were not available for patients in this cohort.Availability of such scores in future prospective cohorts would allow better understanding of TBI pathophysiology and disease prognosis.For a more comprehensive understanding of the complex interplay between TBI, axonal degeneration, synaptic dysfunction and AD, future studies in larger, more ethnically and racially diverse cohorts may benefit from focusing on longitudinal trajectories, including repeated CSF sampling, covering the examination of multiple disease stages, including the post-TBI period, preclinical AD and clinical AD phases.

Conclusion
Our study suggests that a history of TBI may not impact cognitive trajectories in AD, nor does it have a robust association with distinct AD pathophysiological characteristics.Our results indicated slight alterations in synaptic functioning when TBI occurred before disease onset or during an early pre-clinical AD phase.This study underscores the complexity of unraveling the interplay between TBI and AD.Further research is needed to provide a comprehensive understanding of the complete clinical and biological longitudinal trajectories spanning from TBI to the stage of AD-related neurodegeneration.

Verification
We confirm that this article is original and has not been published or submitted elsewhere nor will it be submitted elsewhere while under consideration for this journal.If accepted, it will not be published elsewhere in the same form, in English or in any other language.All authors have contributed to the work and have approved the manuscript for submission.We declare that all potential competing interests are mentioned in the manuscript.IMWV: no competing interest During the preparation of this work the authors used ChatGPT 3.5 in order to improve readability.After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Fig. 1 .
Fig. 1.Spaghetti plots for cognitive trajectories between AD TBI-and AD TBI+.The slopes are uncorrected for potential confounder education.Average slope with 95 % confidence interval is superimposed on the graphs.TBI: traumatic brain injury, MMSE: mini mental state examination.

Table 1
Baseline demographic characteristics.

Table 2
Changes in cognition over time.