Risk of hospitalization with neurodegenerative disease after moderate-to-severe traumatic brain injury in the working-age population: A retrospective cohort study using the Finnish national health registries
In a retrospective cohort study using the Finnish national health registries, Rahul Raj and colleagues investigate the association between traumatic brain injury and dementia, Parkinson disease, and amyotrophic lateral sclerosis.
In a retrospective cohort study using the Finnish national health registries, Rahul Raj and colleagues investigate the association between traumatic brain injury and dementia, Parkinson disease, and amyotrophic lateral sclerosis.
Traumatic brain injury (TBI) is a globally increasing healthcare problem, affecting persons of all ages . Following the early phases of TBI, patients face a significant risk of long-term disability and neurological morbidity . Previous epidemiological studies have found an association between history of TBI and risk for future neurodegenerative disease (NDD) (a concept including dementia, Parkinson disease [PD], and amyotrophic lateral sclerosis [ALS]), but the results have been conflicting and few studies have focused on the working-age population [3–7]. Gardner et al. showed that persons under 65 years with a history of mild TBI did not have an increased risk for dementia compared to non-TBI controls . The development of NDD is supposedly most deleterious in the working-age population, as this not only causes significant morbidity but also has major socioeconomic consequences. Yet, to our knowledge, no previous studies have specifically looked at the overall risk for developing NDD in working-aged persons hospitalized due to TBI.
Our aim is to contrast the risk of NDD in working-aged persons hospitalized due to moderate-to-severe TBI to that of persons hospitalized due to mild TBI. Persons with a history of mild TBI and moderate-to-severe TBI are similar in regard to TBI-specific risk factors, such as alcohol use, which is why individuals with mild TBI serve as a suitable control group . Further, since data on the possible effects of socioeconomic factors on the association between TBI and NDD are lacking, we adjust for education and socioeconomic group . We hypothesized that working-aged persons with a history of moderate-to-severe TBI would have an increased risk for future NDD compared to persons with a history of mild TBI, after adjusting for covariates.
The National Institute for Health and Welfare (THL/1326/5.05.00/2015) approved of the study, in accordance with Finnish national legislation. Statistics Finland (Dnro TK-53-1179-16) and the Population Register Centre (Dnro 1873/410/16) granted us access to their databases. The Finnish Office of the Data Protection Ombudsman (Dnro 2794/402/2015) approved the data collection and combining of data registries. The study was conducted according to the Declaration of Helsinki of the World Medical Association.
Study design and setting
We used the nationwide Finnish Care Register for Health Care to identify persons treated in a public hospital due to TBI in Finland during 1987–2014. The Care Register for Health Care (a continuation of the previous Hospital Discharge Register) was established by the National Institute for Health and Welfare and contains data on persons discharged from every public hospital in Finland from 1969 onwards. The Finnish healthcare system is tax-funded by local municipalities and by the state. In practice, acute care of TBIs is provided solely by public nonprofit healthcare providers and not by private institutions. Thus, the Care Register for Health Care comprehensively includes persons hospitalized due to TBI. The same register can also be used to identify persons hospitalized with NDD. Previous studies have verified the diagnostic accuracy of the registers [11,12].
Study population definition and exposure assessment
The study population consisted of working-aged persons (18–65 years) hospitalized due to moderate-to-severe TBI or mild TBI between 1 January 1987 and 31 December 2014. Mild TBI was defined as a discharge diagnosis indicating no traumatic intracranial lesion (ICD-9 850; ICD-10 S06.0) according to US Centers for Disease Control and Prevention (CDC) criteria, with the exception of isolated skull fractures . Moderate-to-severe TBI was defined as a discharge diagnosis indicating traumatic intracranial lesion (ICD-9 851–854; ICD-10 S06.1–S06.9).
To diminish the likelihood of persons in the mild TBI group having significant traumatic intracranial lesions, we included only those hospitalized for less than 1 day. To diminish the likelihood of persons in the moderate-to-severe TBI group having no significant traumatic intracranial lesion, we included only persons hospitalized for 3 days or longer. In the case of several hospitalizations due to moderate-to-severe TBI or mild TBI, we used the first date of the most severe head injury (i.e., if the person first had a mild TBI and later a moderate-to-severe TBI, the latter was used).
Follow-up and outcome assessment
The study population was prospectively followed up from time of TBI until diagnosis of NDD, death, emigration, or end of follow-up on 31 December 2014. We identified persons hospitalized, for any reason, who were diagnosed with a new NDD diagnosis of dementia (ICD-9 290, 331, 797; ICD-10 G30, F00, F01, F02, F03), PD (ICD-9 332; ICD-10 G20), or ALS (ICD-9 335.2; ICD-10 G12.2) from the Care Register for Health Care. The admission date for the hospitalization including the new NDD diagnosis was defined as the date of diagnosis. We excluded persons with a NDD diagnosis prior to the TBI and persons permanently living outside of Finland. We further excluded persons who received a NDD diagnosis or died within 1 year of the TBI to diminish the possibility of reverse causality. Fig 1 shows a flow chart of the selection and follow-up protocol.
Definition of covariates
Statistics Finland classification of socioeconomic groups is based on the United Nation’s recommendations for the 1990 population censuses. The education classification system is based upon the International Standard Classification of Education 1997 and 2011 classifications. Data on socioeconomic group and highest level of education were obtained for the year closest to the end of follow-up. The classification systems are presented in S1 Text.
Data on mortality and date of emigration were obtained from the Population Register Centre of Finland. Information on date of birth, sex, hospitalization dates, and diagnoses came from the Care Register for Health Care. Data on socioeconomic group and education were obtained from Statistics Finland. We used the unique identification number assigned to all Finnish citizens to identify individuals and combine data from the registers.
Descriptive characteristics of the cohort are presented either as categorical data (N [percent] and compared using a two-sided χ2 test) or as continuous data (mean [standard deviation] and compared using a t test). We calculated the unadjusted rates of NDD per 100,000 person-years. This was done overall and in prespecified age groups, given the increase in risk of NDD with age. Data are presented as NDD rate per 100,000 person-years with 95% confidence intervals.
We used Cox proportional hazards models in Stata (version 14, StataCorp, College Station, TX) to estimate covariate-adjusted hazard ratios (HRs) with 95% confidence intervals. Date of NDD was set as date of study exit. Persons dying before the end of follow-up or moving outside of Finland before the end of follow-up were censored at the time of death or emigration. Date of study entry was the hospital admission date for the hospitalization due to TBI. Age (continuous variable) was the underlying time parameter in all analyses. Socioeconomic group and level of education were treated as categorical variables, using the group with the most individuals as the reference category.
In the primary analysis, we used NDD as a composite outcome variable. In the sensitivity analyses, we separately assessed the risk for dementia, PD, and ALS. Subgroup analyses were conducted by sex, by prespecified age groups (18 to 40 years, 41 to 50 years, 51 to 60 years, and 61 to 65 years), and by equally sized hospital length of stay quartiles (3–5 days, 6–10 days, 11–24 days, ≥25 days). Hospital length of stay served as a surrogate marker of TBI severity .
We repeated the analysis using the Cox proportional hazards model in a matched sample within the cohort. We matched persons with a history of moderate-to-severe TBI and mild TBI in a 1:1 fashion based on age, sex, education, and socioeconomic group using the “ccmatch” function in Stata. If there were multiple mild TBI cases that matched a moderate-to-severe TBI case with respect to these variables, the “ccmatch” function included all of them.
The results are presented as HRs with 95% confidence intervals. We used the group of persons with a history of mild TBI as the reference group in all analyses. p-Values < 0.05 were considered statistically significant. We derived log–log plots of survival curves of TBI to verify that the proportional hazards assumption was not violated.
The reporting of this study is in accordance to the REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement (S1 RECORD Checklist) .
The preplanned statistical plan is presented in S2 Text. One additional analysis assessing how the risk of dementia behaves as a function of time (in the matched cohort) when including persons diagnosed with dementia within the first year after TBI was performed in response to a reviewer comment.
A total of 19,936 persons with a history of moderate-to-severe TBI and 20,703 persons with a history of mild TBI were identified (Fig 1). Baseline characteristics of the two groups are shown in Table 1. Persons in the moderate-to-severe TBI group were on average 7 years older at the time of injury than persons in the mild TBI group (46 versus 39 years). The male to female ratio was higher in the moderate-to-severe TBI group than in the mild TBI group, although men predominated in both groups. There were no major differences in educational level between the groups. Nearly half of all persons had an upper-secondary level of education, while the higher education levels contained between 4% and 9% of individuals. Twenty-nine percent of persons in the moderate-to-severe TBI group died during follow-up, compared to 12% in the mild TBI group.
Unadjusted risk for neurodegenerative disease
Overall time at risk was 453,079 person-years (mean 11 years [SD 8], median 10 years [IQR 4–17], per person). During the follow-up, 696 (3.5%) persons with a history of moderate-to-severe TBI developed NDD compared to 326 (1.6%) of those with a history of mild TBI (p < 0.001). Dementia was the most frequent NDD, followed by PD and ALS, in both groups. Persons in the moderate-to-severe TBI group were on average 4 years younger than persons in the mild TBI group at the time of NDD diagnosis (67 versus 71 years). Significantly more persons in the moderate-to-severe TBI group were diagnosed with NDD before the age of 65 years compared to the mild TBI group (40% versus 26% of all NDD cases in the respective groups, p < 0.001). Persons with a history of moderate-to-severe TBI who went on to develop NDD tended to have a longer hospital length of stay than those who did not develop NDD, suggesting higher TBI severity.
There were no significant differences between persons in the moderate-to-severe TBI group and persons in the mild TBI group who developed NDD in terms of age, socioeconomic group, or level of education (S1 Table). The male to female ratio was higher among individuals in the moderate-to-severe TBI group who developed NDD than among individuals in the mild TBI group who developed NDD.
The unadjusted rate of NDD was 331 per 100,000 person-years in the moderate-to-severe TBI group (318 per 100,000 in men and 373 per 100,000 in women) and 134 per 100,000 person-years in the mild-TBI group (115 per 100,000 in men and 162 per 100,000 in women). The unadjusted rates for all three NDD subtypes were notably higher in the moderate-to-severe TBI group than in the mild TBI group (Table 2). The incidence of NDD increased with age. In the two youngest age groups (18–40 and 41–50 years), the rate of NDD was three to five times higher in persons with moderate-to-severe TBI compared to mild TBI, whereas in the two older age groups (51–60 and 61–65 years), the incidence was approximately one and a half times higher in the moderate-to-severe TBI group.
Adjusted risk of neurodegenerative disease
In the primary analysis, adjusting for age, sex, level of education, and socioeconomic group, moderate-to-severe TBI was associated with an increased risk of NDD, with a HR of 1.8 (95% CI 1.6–2.1, p < 0.001) compared to mild TBI (Fig 2). When analyzing all persons, female sex was associated with a decreased risk of NDD (HR 0.8, 95% CI 0.7–0.9, p < 0.001).
The sensitivity (for NDD subtypes) and subgroup analyses (by sex, age group, and hospital length of stay) are shown in Table 2. In the sensitivity analyses, moderate-to-severe TBI was associated with an increased risk for dementia compared to mild TBI (HR 1.9) but not for PD or ALS. When analyzing women separately, moderate-to-severe TBI was associated with an increased risk for NDD compared to mild TBI, with a HR of 1.9. When analyzing men separately, moderate-to-severe TBI was associated with an increased risk for NDD with a HR of 1.7 (Fig 3). The relative risk of NDD among those with a history of moderate-to-severe TBI was highest among those aged 41 to 50 years and 51 to 60 years of age at baseline (HR 2.7 and 2.0, respectively), though the differences between age groups were not significant (overlapping 95% CIs). Increasing TBI severity, as reflected by duration of hospitalization, was associated with an increased risk for NDD (HR 1.3 to 2.1) in a dose–response pattern.
We identified a total of 25,747 exposure–control matched persons within the follow-up cohort, of which 13,470 were persons with a history of mild TBI, and 12,277 persons with a history of moderate-to-severe TBI. There were no major differences in age, sex, level of education, or socioeconomic group distribution between the matched groups (S2 Table). In the matched sample, 3.2% of persons with a history of moderate-to-severe TBI developed NDD compared to 2.3% of persons with a history of mild TBI. In the Cox proportional hazards model, TBI was associated with an increased risk for NDD, with a HR of 1.8 (95% CI 1.6–2.1, p < 0.001), providing additional support for the results from the primary analysis.
A reviewer requested additional analysis to show how the risk of dementia behaved as a function of time (in the matched sample cohort). This analysis did not exclude persons diagnosed with dementia within the first year after the TBI. The risk for dementia was continually higher for persons with a history of moderate-to-severe TBI compared those with a history of mild TBI (S1 Fig). Furthermore, the lines diverge with time, providing further support for the association between moderate-to-severe TBI and dementia.
In this nationwide study in Finland investigating the association between TBI and NDD, we found that persons with a history of moderate-to-severe TBI had an 80% increased probability of future NDD compared to persons with a history of mild TBI. The matched sample analysis strengthened our results. The risk of future NDD increased with TBI severity (length of hospital stay) in a dose–response pattern. However, when the three NDD subtypes were analyzed separately, TBI was associated with increased risk only for dementia (90% increased probability).
Similar to our findings, three previous meta-analyses found TBI to be associated with an increased risk for dementia [3–5]. We further found that the associated risk between NDD and dementia increased with TBI severity. Thus, the evidence for the association between TBI and dementia seems to be compelling. Conversely, the association of TBI with PD and ALS is not as clear. Some studies suggest a significant association, and some no association [6,16–19]. In the present study, no association between moderate-to-severe TBI and PD or ALS could be established. The absolute numbers of persons developing PD and ALS were limited. Thus, the negative finding might be the consequence of a type II error. Yet, considering that this was an almost 30-year-long nationwide follow-up study including all persons hospitalized for TBI in Finland, ALS and PD do not seem to be a significant long-term neurological problem in individuals with a history of moderate-to-severe TBI. It should be noted that we were only able to identify persons diagnosed with ALS and PD who had been hospitalized. Thus, it is possible that we missed persons diagnosed with ALS or PD in the primary care setting or by private sector specialists.
In the present study, we could not study the risk of dementia in persons without a history of mild or moderate-to-severe TBI. However, based on previous studies, 14,500 persons are diagnosed with dementia annually in Finland. . For a Finnish adult population (18 years or older) of 4,385,426, this translates to an incidence of 331 per 100,000 person-years. In comparison, the unadjusted incidence of dementia was 293 per 100,000 person-years for persons with a history of moderate-to-severe TBI and 114 per 100,000 person-years for persons with a history of mild TBI. Thus, working-aged persons with a moderate-to-severe injury at a young age have a similar incidence of dementia as the general population, where most cases are among elderly individuals . The incidence of dementia in persons with a history of mild TBI is somewhat lower than the general incidence, most probably because our cohort included persons under 65 years of age at the time of injury.
A major limitation for comparing previous studies of the association between TBI and NDD is the large variation in TBI and NDD definitions . Some studies rely upon self-reported diagnoses for TBI (and thus suffer from recall bias), while others use ICD diagnoses (eight, ninth, or tenth revisions). For NDD diagnosis, some studies use DSM criteria, ICD diagnoses, or NINCDS-ADRDA diagnostic criteria. Thus, not surprisingly, with such a wide spectrum of the basic definitions, the results have been conflicting. In this study, we defined mild TBI according to the CDC criteria . The same definition was used by Gardner et al. , who showed that persons under the age of 65 years with mild TBI do not have an increased risk for dementia. Thus, as we specifically investigated persons under 65 years of age, the mild TBI population formed a suitable control group for individuals with moderate-to-severe TBI. Persons treated for mild and moderate-to-severe TBI are likely to be similar in TBI-specific risk factors, such as age, gender, alcohol use, and socioeconomic factors . Yet, the included mild TBI population might differ somewhat from the most common form of mild TBI, when hospital admission is not required. Identifying a control group of individuals with a history mild TBI without hospitalization is, however, impossible in large epidemiological studies such as the present study.
The role of sex differences in risk of developing dementia after TBI has been widely discussed. Both Mortimer et al.  and Fleminger et al.  found men, but not women, with a history of TBI to have an increased risk of dementia after TBI. Speculated theories include estrogen- and progesterone-induced neuroprotection [23,24]. In our study, both men and women with a history of moderate-to-severe TBI had an increased risk for dementia compared to those with a history of mild TBI. Yet, female sex was associated with a reduced risk for dementia in comparison to male sex, both in the mild TBI and moderate-to-severe TBI groups. Population studies generally do not find differences in dementia incidence between men and women [25,26]. Our study cannot establish any causation between sex, TBI, and dementia, although our results imply that sex may play a role in TBI-related dementia. The underlying causes of TBI may be different in men and women, resulting in a differential capacity to recover from brain injury and hence differential risk of dementia.
Poor socioeconomic factors (such as low education level and socioeconomic group) increase the risk for sustaining a TBI . However, they also serve as a risk factor future NDD . Thus, the association between TBI and NDD in previous studies may have reflected the underlying association of these socioeconomic factors, rather than of TBI itself, with NDD. Yet, we found that even after adjusting for socioeconomic group and level of education, TBI was significantly associated with an increased risk for NDD.
Strengths and limitations
This is, to our knowledge, the first nationwide study on the subject (including over 40,000 persons). The nationwide coverage and the high data quality of the registries strengthen the study’s generalizability [11,12]. Only a few previous studies match the present one in size. One of the larger studies (with a maximum follow-up time of 5 to 7 years) was by Gardner et al. , in which the authors identified approximately 50,000 persons with a history of TBI and found results like ours, i.e., moderate-to-severe TBI increased the risk for dementia, with a HR of 1.7 (compared to 1.9 in our study) . Yet, to date, the present study has one of the longest follow-up times (mean time at risk 11 years, or 453,079 person-years), something that is essential when investigating long-term neurological morbidity after TBI. By using persons with a hospitalization due to mild TBI as controls, we diminished the likelihood of detecting an effect that is not present (i.e., type I error), as it is unlikely that persons with a history of mild TBI have an increased risk of dementia compared to persons with a history of non-brain trauma . Yet, there are studies suggesting that mild TBI itself might increase the risk for NDD, and therefore it is possible that our results underestimate the effect of moderate-to-severe TBI in the development of NDD . On the other hand, a recent systematic review found no association between mild TBI and dementia .
Despite the high quality of the registries used, all register-based studies include diagnostic inaccuracies, coding errors, and other confounding factors that cannot be controlled for. First, as the Care Register for Health Care includes only hospitalized persons; it is possible that we missed persons being diagnosed with NDD in the outpatient setting (e.g., milder forms of PD and dementia). Second, although the median follow-up time was 10 years, persons hospitalized for TBI during the more recent years inevitably had shorter follow-up times and may not yet have been diagnosed with a NDD. Third, a notably higher proportion of persons in the moderate-to-severe TBI group died during the follow-up period compared to the mild TBI group (29% versus 12%), decreasing the moderate-to-mild TBI cases’ exposure time. Thus, it is likely that the risk of NDD in persons with a history of moderate-to-severe TBI is even higher than presented in this study. Furthermore, the diagnosis of NDD, especially dementia, is prone to error if it occurs too soon after TBI. Such diagnosis may be a residual effect of delirium, medication, or other complications following TBI. To avoid the possibility of reverse causality, we recorded NDD diagnoses starting 1 year following the TBI. Yet, even after including persons diagnosed with dementia within the first year after the TBI, moderate-to-severe TBI was associated with a significantly higher risk for dementia than mild TBI, strengthening the association (S1 Fig).
As in many register-based studies, we used ICD-9 and ICD-10 discharge diagnoses to identify persons with a history of moderate-to-severe TBI and mild TBI . For mild TBI we used diagnoses indicating no structural intracranial injury, and for moderate-to-severe TBI, we used diagnoses indicating an objective intracranial injury . We further excluded persons with mild TBI hospitalized for longer than 1 day, as these may have had an undiagnosed intracranial injury, and persons with moderate-to-severe TBI hospitalized for shorter than 3 days, as these may represent cases of either rapid death or a milder form of intracranial injury not requiring hospitalization. It is possible that in the mild TBI group there were persons with clinically silent diffuse axonal injuries that passed undetected. How such traumatic microlesions affect the risk of future NDD is unknown.
The most evident limitation, which is shared by most large-scale epidemiological studies, is that the study setup does not allow to us analyze any causative factors. It has been hypothesized that TBI does not itself cause NDD but rather accelerates an underlying process of developing NDD in persons with predisposing factors . Such predisposing factors may include comorbidities; genetic variations, such as APOE ε4 allele expression and neprilysin polymorphism; and lifestyle factors, such as cognitive reserve, physical activity, obesity, alcohol, and smoking . For example, a substantial proportion of TBIs in Finland relate to alcohol use . Alcohol-related TBIs are much more common in men than in women, especially among less educated people, who also have a higher baseline risk for NDD, which may confound our results [35,36]. Furthermore, comorbidities, such as hypertension , stroke , and diabetes , have been found to significantly associate with risk of NDD, especially with dementia. Therefore, differences in comorbidities between the mild and moderate-to-severe TBI groups may potentially have affected our results. The aspect of physical activity after TBI is interesting. Decreased physical activity is probably more likely to happen after moderate-to-severe TBI than after mild TBI. Decreased physical activity is associated with an increased risk for dementia [40,41]. Thus, increasing physical activity in persons after moderate-to-severe TBI, in combination with aggressive treatment of cardiovascular comorbidities, might decrease the risk of subsequent dementia. Yet, further studies investigating the causative relationship between TBI, other environmental risk factors, and genetics are needed.
Our results suggest that in working-aged persons, moderate-to-severe TBI increases the risk for developing NDD later in life. In our study, the risk seemed to increase with TBI severity in a dose–response pattern, and the risk was higher in men. With regard to the NDD subtypes, moderate-to-severe TBI was associated with increased risk for dementia but not for PD and ALS. The effect of covariates, such as comorbidities, lifestyle factors, and genetic factors, should be accounted for in future etiological studies, as well as studies to improve diagnostics and prevention of dementia after TBI.
1. Roozenbeek B, Maas AIR, Menon DK. Changing patterns in the epidemiology of traumatic brain injury. Nat Rev Neurol. 2013;9:231–6. doi: 10.1038/nrneurol.2013.22 23443846
2. Selassie AW, Zaloshnja E, Langlois JA, Miller T, Jones P, Steiner C. Incidence of long-term disability following traumatic brain injury hospitalization, United States, 2003. J Head Trauma Rehabil. 2008;23:123–31. doi: 10.1097/01.HTR.0000314531.30401.39 18362766
3. Fleminger S, Oliver D, Lovestone S, Rabe-Hesketh S, Giora A. Head injury as a risk factor for Alzheimer’s disease: the evidence 10 years on; a partial replication. J Neurol Neurosurg Psychiatry. 2003;74:857–62. doi: 10.1136/jnnp.74.7.857 12810767
4. Mortimer JA, Van Duijn CM, Chandra L, Fratiglioni L. Head trauma as a risk factor for Alzheimer’s disease: a collaborative reanalysis of case control studies. Int J Epidemiol. 1991;20(Suppl 2):S28–35. doi: 10.1093/ije/20.Supplement_2.S28
5. Li Y, Li Y, Li X, Zhang S, Zhao J, Zhu X, et al. Head injury as a risk factor for dementia and Alzheimer’s disease: a systematic review and meta-analysis of 32 observational studies. PLoS ONE. 2017;12:e0169650. doi: 10.1371/journal.pone.0169650 28068405
6. Armon C, Nelson LM. Is head trauma a risk factor for amyotrophic lateral sclerosis? An evidence based review. Amyotroph Lateral Scler. 2012;13:351–6. doi: 10.3109/17482968.2012.660954 22424129
7. Jafari S, Etminan M, Aminzadeh F, Samii A. Head injury and risk of Parkinson disease: a systematic review and meta-analysis. Mov Disord. 2013;28:1222–9. doi: 10.1002/mds.25458 23609436
8. Gardner RC, Burke JF, Nettiksimmons J, Kaup A, Barnes DE, Yaffe K. Dementia risk after traumatic brain injury vs nonbrain trauma. JAMA Neurol. 2014;71:1490. doi: 10.1001/jamaneurol.2014.2668 25347255
9. Nordstrom A, Edin BB, Lindstrom S, Nordstrom P. Cognitive function and other risk factors for mild traumatic brain injury in young men: nationwide cohort study. BMJ. 2013;346:f723. doi: 10.1136/bmj.f723 23482939
10. Vuoksimaa E, Rinne J, Lindgren N, Heikkilä K, Koskenvuo M, Kaprio J. Middle age self-report risk score predicts cognitive functioning and dementia in 20–40 years. Alzheimers Dement (Amst). 2016;4:118–25. Available from: http://www.dadm.alzdem.com/article/S2352-8729(16)30043-4/pdf. Accessed 17 May 2017.
11. Sund R. Quality of the Finnish Hospital Discharge Register: a systematic review. Scand J Public Health. 2012;40:505–15. doi: 10.1177/1403494812456637 22899561
12. Keskimäki I, Aro S. Accuracy of data on diagnoses, procedures and accidents in the Finnish Hospital Discharge Register. Int J Heal Sci. 1991;2:15–21.
13. Centers for Disease Control and Prevention. Report to Congress on mild traumatic brain injury in the United States: steps to prevent a serious public health problem. Atlanta: Centers for Disease Control and Prevention; 2003.
14. Newgard CD, Fleischman R, Choo E, Ma OJ, Hedges JR, McConnel KJ. Validation of length of hospital stay as a surrogate measure for injury severity and resource use among injury survivors. Acad Emerg Med. 2016;17:142–50.
15. Benchimol EI, Smeeth L, Guttmann A, Harron K, Moher D, Petersen I, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLOS Med. 2015;12:e1001885. doi: 10.1371/journal.pmed.1001885 26440803
16. Kenborg L, Rugbjerg K, Lee P-C, Ravnskjær L, Christensen J, Ritz B, et al. Head injury and risk for Parkinson disease: results from a Danish case-control study. Neurology. 2015;84:1098–103. doi: 10.1212/WNL.0000000000001362 25681453
17. Rugbjerg K, Ritz B, Korbo L, Martinussen N, Olsen J. Risk of Parkinson’s disease after hospital contact for head injury: population based case-control study. BMJ. 2008;337:a2494. doi: 10.1136/bmj.a2494 19074944
18. Gardner RC, Burke JF, Nettiksimmons J, Goldman S, Tanner CM, Yaffe K. Traumatic brain injury in later life increases risk for Parkinson disease. Ann Neurol. 2015;77:987–95. doi: 10.1002/ana.24396 25726936
19. Peters TL, Fang F, Weibull CE, Sandler DP, Kamel F, Ye W. Severe head injury and amyotrophic lateral sclerosis. Amyotroph Lateral Scler Frontotemporal Degener. 2013;14:267–72. doi: 10.3109/21678421.2012.754043 23286749
20. Prince M, Ali G-C, Guerchet M, Prina AM, Albanese E, Wu Y-T. Recent global trends in the prevalence and incidence of dementia, and survival with dementia. Alzheimers Res Ther. 2016;8:23. doi: 10.1186/s13195-016-0188-8 27473681
21. Rossor MN, Fox NC, Mummery CJ, Schott JM, Warren JD. The diagnosis of young-onset dementia. Lancet Neurol. 2010;9:793–806. doi: 10.1016/S1474-4422(10)70159-9 20650401
22. Puljula J, Vaaramo K, Tetri S, Juvela S, Hillbom M. Risk for all-cause and traumatic death in head trauma subjects: a prospective population-based case-control follow-up study. Ann Surg. 2016;263:1235–9. doi: 10.1097/SLA.0000000000001392 26334638
23. Stein DG. Brain damage, sex hormones and recovery: a new role for progesterone and estrogen? Trends Neurosci. 2001;24:386–91. doi: 10.1016/S0166-2236(00)01821-X 11410269
24. Roof RL, Hall ED. Gender differences in acute CNS trauma and stroke: neuroprotective effects of estrogen and progesterone. J Neurotrauma. 2000;17:367–88. doi: 10.1089/neu.2000.17.367 10833057
25. Fratiglioni L, Launer LJ, Andersen K, Breteler MM, Copeland JR, Dartigues JF, et al. Incidence of dementia and major subtypes in Europe: a collaborative study of population-based cohorts. Neurologic Diseases in the Elderly Research Group. Neurology. 2000;54:S10–5. 10854355
26. Jorm AF, Jolley D. The incidence of dementia: a meta-analysis. Neurology. 1998;51:728–33. doi: 10.1212/WNL.51.3.728 9748017
27. Stern Y. Cognitive reserve in ageing and Alzheimer’s disease. Lancet Neurol. 2012;11:1006–12. doi: 10.1016/S1474-4422(12)70191-6 23079557
28. Gavett BE, Stern RA, Cantu RC, Nowinski CJ, McKee AC, Mortimer J, et al. Mild traumatic brain injury: a risk factor for neurodegeneration. Alzheimers Res Ther. 2010;2:18. doi: 10.1186/alzrt42 20587081
29. Godbolt AK, Cancelliere C, Hincapié CA, Marras C, Boyle E, Kristman VL, et al. Systematic review of the risk of dementia and chronic cognitive impairment after mild traumatic brain injury: results of the international collaboration on mild traumatic brain injury prognosis. Arch Phys Med Rehabil. 2014;95(3 Suppl):S245–56. doi: 10.1016/j.apmr.2013.06.036 24581910
30. Tagliaferri F, Compagnone C, Korsic M, Servadei F, Kraus J. A systematic review of brain injury epidemiology in Europe. Acta Neurochir (Wien). 2006;148:255–68.
31. Koskinen S, Alaranta H. Traumatic brain injury in Finland 1991–2005: a nationwide register study of hospitalized and fatal TBI. Brain Inj. 2008;22:205–14. doi: 10.1080/02699050801938975 18297592
32. Washington PM, Villapol S, Burns MP. Polypathology and dementia after brain trauma: does brain injury trigger distinct neurodegenerative diseases, or should they be classified together as traumatic encephalopathy? Exp Neurol. 2016;275(Pt 3):381–8. doi: 10.1016/j.expneurol.2015.06.015 26091850
33. Ballard C, Gauthier S, Corbett A, Brayne C, Aarsland D, Jones E. Alzheimer’s disease. Lancet. 2011;377:1019–31. doi: 10.1016/S0140-6736(10)61349-9 21371747
34. Raj R, Skrifvars MB, Kivisaari R, Hernesniemi J, Lappalainen J, Siironen J. Acute alcohol intoxication and long-term outcome in patients with traumatic brain injury. J Neurotrauma. 2015;32:95–100. doi: 10.1089/neu.2014.3488 25010885
35. Kivipelto M, Ngandu T, Laatikainen T, Winblad B, Soininen H, Tuomilehto J. Risk score for the prediction of dementia risk in 20 years among middle aged people: a longitudinal, population-based study. Lancet Neurol. 2006;5:735–41. doi: 10.1016/S1474-4422(06)70537-3 16914401
36. Helakorpi S, Mäkelä P, Uutela A. Alcohol consumption before and after a significant reduction of alcohol prices in 2004 in Finland: were the effects different across population subgroups? Alcohol Alcohol. 45:286–92. doi: 10.1093/alcalc/agq007 20164044
37. Qiu C, Winblad B, Fratiglioni L. The age-dependent relation of blood pressure to cognitive function and dementia. Lancet Neurol. 2005;4:487–99. doi: 10.1016/S1474-4422(05)70141-1 16033691
38. Savva GM, Stephan BCM, Alzheimer’s Society Vascular Dementia Systematic Review Group. Epidemiological studies of the effect of stroke on incident dementia: a systematic review. Stroke. 2010;41:e41–6. doi: 10.1161/STROKEAHA.109.559880 19910553
39. Lu F-P, Lin K-P, Kuo H-K. Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis. PLoS ONE. 2009;4:e4144. doi: 10.1371/journal.pone.0004144 19127292
40. Hamer M, Chida Y. Physical activity and risk of neurodegenerative disease: a systematic review of prospective evidence. Psychol Med. 2009;39:3–11. doi: 10.1017/S0033291708003681 18570697
41. Fratiglioni L, Paillard-Borg S, Winblad B. An active and socially integrated lifestyle in late life might protect against dementia. Lancet Neurol. 2004;3:343–53. doi: 10.1016/S1474-4422(04)00767-7 15157849