RESEARCH ARTICLE
Traumatic Brain Injury and Young Onset Dementia: A Nationwide Cohort Study Peter Nordstr€ om, PhD,1 Karl Micha€ elsson, PhD,2 Yngve Gustafson, PhD,1 and Anna Nordstr€ om, PhD3 Objective: To investigate the association between traumatic brain injuries (TBIs) and the risk of young onset dementia (YOD), that is, dementia before 65 years of age. Methods: The study cohort comprised 811,622 Swedish men (mean age 5 18 years) conscripted for military service between 1969 and 1986. TBIs, dementia, and covariates were extracted from national registers. Time-dependent exposures using Cox proportional hazard regression models were evaluated. Results: During a median follow-up period of 33 years, there were 45,249 men with at least 1 TBI in the cohort. After adjustment for covariates, 1 mild TBI (hazard ratio [HR] 5 1.0, 95% confidence interval [CI] 5 0.5–2.0), at least 2 mild TBIs (HR 5 2.5, 95% CI 5 0.8–8.1), or 1 severe TBI (HR 5 0.7, 95% CI 5 0.1–5.2) were not associated with Alzheimer dementia (AD). Other types of dementia were strongly associated with the risk of 1 mild TBI (HR 5 3.8, 95% CI 5 2.8–5.2), at least 2 mild TBIs (HR 5 10.4, 95% CI 5 6.3–17.2), and 1 severe TBI (HR 5 11.4, 95% CI 5 7.4–17.5) in age-adjusted analysis. However, these associations were largely attenuated after adjustment for covariates (1 mild TBI: HR 5 1.7; at least 2 mild TBIs: HR 5 1.7; 1 severe TBI: HR 5 2.6; p < 0.05 for all). Interpretation: In the present study, we found strong associations between YOD of non-AD forms and TBIs of different severity. These associations were, however, markedly attenuated after multivariate adjustment. ANN NEUROL 2014;75:374–381
T
he estimated annual global incidence of traumatic brain injury (TBI) requiring medical attention or resulting in hospitalization or death is >10 million.1,2 Although changes in cognitive function shortly after TBI can be linked to the physiological consequences of the injury, longer-term consequences of TBI frequently persist.3–5 Increased risk of dementia is among the most feared late onset outcomes, which after a TBI may occur at an earlier age than expected.6 The estimated population at risk of dementia attributable to TBI is 5 to 15%, and this injury has been suggested to be the best established risk factor for dementia.7 However, the association between TBI and the risk of dementia has been debated. Two meta-analyses of case–control studies suggested that previous TBI resulting in loss of consciousness was associated with a doubled risk of Alzheimer dementia (AD) in men, but not in women.8,9 A case–control study of former Navy and
military veterans hospitalized for TBI during World War II with a follow-up period of 50 years found a graded association between TBI severity and the risk of AD and other forms of dementia.10 In contrast, 4 cohort studies suggested that head trauma did not increase the risk of dementia.11–14 These conflicting results may have been influenced by insufficient numbers of dementia cases in study populations; short follow-up periods; and especially profound lack of assessment of covariates influencing the risks of TBI and dementia, such as low socioeconomic status, alcohol intoxication, physical fitness, blood pressure, and low premorbid cognitive function.15–19 To the best of our knowledge, no previous population-based study has investigated the association between TBI and the risk of young onset dementia (YOD) exclusively, that is, dementia before 65 years of age, which has devastating consequences for individuals, their families, and society.20–22
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24101 Received Sep 10, 2013, and in revised form Dec 26, 2013. Accepted for publication Dec 31, 2013. Address correspondence to Dr Nordstr€ om, Department of Community Medicine and Rehabilitation, Geriatrics, 90187 Umea˚ University, Umea˚, Sweden. E-mail: peter.nordstrom@germed.umu.se From the 1Department of Community Medicine and Rehabilitation, Geriatric Medicine, Umea˚ University, Umea˚; 2Department of Surgical Sciences, Section of Orthopedics, Uppsala University, Uppsala; and 3Department of Surgical and Perioperative Sciences, Sports Medicine, Umea˚ University, Umea˚, Sweden.
C 2014 American Neurological Association 374 V
Nordstr€ om et al: TBI and Young Onset Dementia
In the present nationwide cohort study, we investigated associations between TBIs of different severity and the risk of YOD in men during a median follow-up period of 33 years.
Subjects and Methods Baseline Examination of the Study Cohort All Swedish men conscripted for mandatory military service at 6 centers between 1969 and 1986 (n 5 825,816) were considered for inclusion in the present study. During these years, about 97% of all men in Sweden participated in conscription tests. The Swedish conscription register has been described previously.23 After the exclusion of subjects with body weight < 40 or > 170kg, body height < 140 or > 215cm, or age < 16 years at conscription, those who were conscripted twice, and subjects for whom register information was insufficient, 811,622 men were included in the study. The baseline date was the date of conscription. During conscription, physicians examined all men and registered all diagnoses according to the International Classification of Disease version 8 (ICD 8). Weight and height were measured using standardized equipment, and blood pressure was measured using mercury sphygmomanometers. Isometric muscle strength for right knee extension was measured with the subject in a seated position with crossed arms. The strength test was repeated 33 or until the value stopped increasing, and the highest value was recorded. Testing equipment was calibrated daily. Four tests were used to assess cognitive function: logical, word recollection, visuospatial, and technical=mathematical tests. These tests have been described and evaluated in detail previously.24,25 For the present study, the 4 cognitive test scores were converted to z scores separately for each test, that is: (the individual test score 2 the mean score for the cohort)=standard deviation for the cohort. The individual z scores were then added and linearly transformed based on the highest test result in the cohort, to give an overall cognitive function score ranging from 1 to 40.
Socioeconomic Information and Diagnoses in Participants and Parents Every Swedish citizen is linked to the national patient register, launched in 1964, and=or other registers based on his or her unique individual social security number. The biological parents of all men in the cohort were also tracked in the Statistics Sweden database using social security numbers. The following diagnoses were identified in the cohort during follow-up and=or in parents using the following ICD 8, 9, and 10 codes: mild TBI (S06.0x, 850), severe TBI (851, 852, 853, and S06.x, excluding S06.0x and S06.1x), AD (F00.x, G30.x, and also 290.x in the parents), vascular dementia (F01.x), alcohol dementia (F10.7A), and dementia of unspecified type (NUD; F03.9) in subjects and parents; and alcohol intoxication (303 and F10.x, excluding F10.7A), drug intoxication (F11.x, 304), depression (F32.x, 311), ischemic stroke (I63.x, 433, 434), and cerebral hemorrhage (I61.x, I62.x, 431) in subjects. Information about sub-
March 2014
jects’ total income per year and highest achieved education level was collected from the Statistics Sweden database 15 years after conscription testing. Education level was classified as the presence=absence of university-level education. Information about parents was linked to subjects in the cohort as a proxy of the heritable components of YOD and TBI. Information about deaths, immigration, and emigration during the study period was obtained through record linkage with the National Cause of Death Register and the Statistics Sweden database.
Statistical Analyses Descriptive data are presented as means and standard deviations throughout the article, if not indicated otherwise. To investigate the association between TBI of different severity (no TBI, only 1 mild TBI, at least 2 mild TBIs but no severe TBI, or at least 1 severe TBI) and the risk of YOD, a Cox regression model was used. Diagnoses occurring during follow-up were included as time-dependent variables, that is, the exposure changed during follow-up. In the first model, the hazard ratios (HRs) were adjusted for age, place, and year of conscription, and in a second model also for the other variables as specified in Table 1. Followup data were calculated until date of YOD, date of emigration, date of death, or December 31, 2011, whichever came first. To test the consistency of the associations between TBI and YOD after adjustment for confounders, we performed nested case–control studies within the total study cohort. Controls were selected using propensity score matching. For men with 1 mild TBI, every case was matched with 1 control, and for men with at least 2 mild TBIs or 1 severe TBI, every case was matched with 9 controls. Propensity scores were estimated using separate logistic regression models for each type of TBI, including the covariates as specified in Table 1. Cox regression was then used to analyze the relation between TBI and YOD, where TBIs were entered as a time-dependent variable, with the same endpoints for follow-up as above. In a complementary analysis, TBIs of different severity and the risk of death were analyzed in propensity score– matched cohorts, where every case was matched against 1 control based on the risk factors specified in Table 1 as described above. The date of TBI was the baseline date in these models both for cases and for corresponding controls. Case–control pairs in which the control died or emigrated before the date of TBI in the corresponding case were excluded. The associations between TBI of different severity and death were then investigated using piecewise Cox regression with endpoints for followup including emigration, death, or December 31, 2011 (end of follow-up). The SPSS software package (version 20.0 for Macintosh; SPSS, Chicago, IL) with the R essential application was used for statistical analyses, with p < 0.05 considered significant.
Results Characterization of TBIs in the Study Cohort The study cohort included 811,622 Swedish men with a mean age of 18 years at the time of conscription. During 375
ANNALS
of Neurology
TABLE 1. Characteristics of the Study Cohort Based on the Occurrence of TBI during Follow-up
Characteristic No.
No TBI
1 Mild TBI
>1 Mild TBI
1 Severe TBI
766,373
34,698
4,569
5,982
32.9 6 11.1
37.4 6 10.6
39.1 6 11.3
Age at time of TBI, yr Baseline/conscription Age, yr
18.4 6 0.8
18.4 6 0.8
18.4 6 0.9
18.5 6 0.9
Weight, kg
68.9 6 10.0
68.4 6 10.1
67.8 6 10.1
67.7 6 9.9
Height, cm
179 6 6
178 6 7
178 6 7
178 6 6
Knee strength, N
543 6 104
540 6 105
537 6 106
532 6 106
Systolic blood pressure, mmHg
128 6 11
128 6 11
127 6 11
128 6 11
Diastolic blood pressure, mmHg
68 6 10
68 6 10
68 6 10
69 6 10
Overall cognitive function score, 1–40 points
21.3 6 5.0
19.8 6 5.1
18.4 6 5.1
19.7 6 5.1
University-level education, %
28.7
17.6
9.8
15.0
Total income, thousand US dollars
27.2 6 16.6
22.5 6 16.2
17.2 6 14.6
18.9 6 15.4
Dementia in father
5.1
4.8
4.5
4.8
Dementia in mother
5.5
5.4
4.8
5.6
TBI in father
3.8
4.8
5.8
4.8
TBI in mother
2.9
3.6
4.3
3.7
Alcohol intoxication
3.9
19.4
41.2
29.1
Drug intoxication
0.9
5.1
12.3
6.8
Depression
13.0
22.2
32.6
26.6
Ischemic stroke
0.7
1.2
1.7
2.3
Nonischemic hemorrhage
0.2
0.6
1.1
3.7
Socioeconomic factors 15 years after baseline
Heritability, %
Diagnoses during follow-up, %
TBI 5 traumatic brain injury.
a total follow-up time of 26.9 million (median 5 33, range 5 0–43) years, 34,698 subjects sustained 1 mild TBI, 4,569 subjects had at least 2 mild TBIs but no severe TBI, and 5,982 subjects had at least 1 severe TBI. For subjects with mild TBIs, the mean duration of hospitalization was 1.5 (range 5 0–369) days. The most common diagnoses associated with severe TBI were traumatic subdural and subarachnoid hemorrhage, and the mean duration of hospitalization was 15.1 (range 5 0–1,972) days. Baseline characteristics of the cohort based on TBI status during follow-up are shown in Table 1. The largest differences between subjects with and without TBI during follow-up included lower overall premorbid cognitive function, educational level, and 376
total income, and higher risk of intoxication and depression in subjects with TBI. YOD Occurring during Follow-up according to TBI Severity During follow-up, 566 (0.07%) subjects were diagnosed with YOD at a median age of 52 (range 5 30–61) years. Diagnoses of AD (n 5 177), vascular dementia (n 5 119), alcohol dementia (n 5 78), and dementia NUD (n 5 192) were distributed among subjects with 1 mild TBI (n 5 100, 0.25%), at least 2 mild TBIs (n 5 23, 0.5%), and at least 1 severe TBI (n 5 25, 0.4%). The Figure shows the cumulative survival of YOD (all causes) by TBI type before and after adjustment for all Volume 75, No. 3
Nordstr€ om et al: TBI and Young Onset Dementia
FIGURE: Cumulative survival with respect to all cases of young onset dementia (YOD) stratified for type of traumatic brain injury (TBI) as estimated by Cox regression. (A) Survival was adjusted for age and year of conscription. (B) Survival was also adjusted for weight, height, knee extension strength, cognitive function and blood pressure at baseline (conscription), total income, university education, intoxication, depression, and cerebrovascular disease during follow-up. Follow-up time was terminated at 40 years. [Color figure can be viewed in the online issue, which is available at www.annalsofneurology.org.]
confounders. Differences in cumulative survival for YOD according to TBI type were markedly reduced after adjustment for all covariates. TBI Severity and the Risk of YOD Associations between TBI type and the rate of YOD were evaluated using Cox regression, with all diagnoses (TBIs, depressions, cerebrovascular disease, intoxications) entered as time-dependent variables. A diagnosis of AD was very rare after any TBI (1 mild TBI, n 5 10; at least 2 mild TBIs, n 5 3; 1 severe TBI, n 5 1). After adjustment for all covariates, the risk of AD was not significant among men with 1 previous mild TBI (HR 5 1.0, 95% confidence interval [CI] 5 0.5–2.0), at least 2 mild TBIs (HR 5 2.5, 95% CI 5 0.8–8.1), or 1 severe TBI (HR 5 0.7, March 2014
95% CI 5 0.1–5.2; Table 2). For other types of YOD the risks were higher in men with 1 mild TBI (HR 5 3.8, 95% CI 5 2.8–5.2), at least 2 mild TBIs (HR 5 10.4, 95% CI 5 6.3–17.2), and 1 severe TBI (HR 5 11.4, 95% CI 5 7.4–17.5), after adjustment for age at baseline and place and year of conscription. These associations were, however, markedly attenuated after adjusting for all covariates (1 mild TBI: HR 5 1.7, 95% CI 5 1.2–2.3; at least 2 mild TBIs: HR 5 1.7, 95% CI 5 1.0–2.9; 1 severe TBI: HR 5 2.6, 95% CI 5 1.6–4.1). Sensitivity Analyses TBIs of different severity showed similar associations with the risk of YOD in propensity score–matched cohorts (see Table 2). Furthermore, adjusting for some of 377
ANNALS
of Neurology
TABLE 2. Associations between TBIs and the Outcome of Dementia
All Cases of Dementia, n 5 566 TBI Type
Alzheimer Dementia, n 5 177
Other Forms of Dementia,a n 5 389
No.
HR
95% CI
No.
HR
95% CI
No.
HR
95% CI
63
3.0
2.3–4.0
10
1.5
0.8–2.8
53
3.8
2.8–5.2
62
1.6
1.2–2.2
10
1.3
0.7–2.5
53
1.7
1.3–2.3
Model adjusted for all confoundersd
59
1.5
1.1–2.0
8
1.0
0.5–2.0
51
1.7
1.2–2.3
Case–control studye
59
1.8
1.2–2.8
8
1.5
0.5–4.1
51
2.1
1.3–3.1
20
8.3
5.3–13.1
3
3.9
1.3–12.3
17
10.4
6.3–17.2
19
2.1
1.3–3.4
3
2.8
0.9–9.1
17
2.0
1.2–3.4
Model adjusted for all confoundersd
18
1.8
1.1–3.0
3
2.5
0.8–8.1
15
1.7
1.0–2.9
Case–control studye
18
2.4
1.4–4.1
3
—
15
2.1
1.2–3.8
25
7.9
5.2–11.9
1
1.0
0.1–7.1
24
11.4
7.4–17.5
25
2.9
1.9–4.5
1
0.8
0.1–5.6
24
3.4
2.2–5.3
Model adjusted for all confoundersd
25
2.3
1.5–3.6
1
0.7
0.1–5.2
24
2.6
1.6–4.1
Case–control studye
25
2.9
1.8–4.6
1
—
24
3.0
1.6–5.3
1 mild TBI, n 5 34,698 Age-adjusted modelb Minimally adjusted model
c
>1 mild TBI, n 5 4,569 Age-adjusted modelb Minimally adjusted model
c
1 severe TBI, n 5 5,982 Age-adjusted modelb Minimally adjusted model
c
a
Vascular dementia, alcohol dementia, dementia of unspecified type. Adjusted for the influence of age, and place and year of conscription. c Adjusted for the variables listed above and overall cognitive function and alcohol intoxication. d Adjusted for the variables listed above and weight, height, knee extension strength, TBI in parents, dementia in parents, income, educational level, systolic blood pressure, drug intoxication, depression, and cerebrovascular disease. e For the case–control study, controls were selected using propensity scores based on all confounders listed above. Every subject with 1 mild TBI was matched with 1 control, and every subject with at least 2 mild TBIs or 1 severe TBI was matched with 9 controls. Hazard ratios are adjusted for propensity scores. No case–control study was performed for Alzheimer dementia in subjects with at least 2 mild TBIs or 1 severe TBI, because the cohort included few such cases. CI 5 confidence interval; HR 5 hazard ratio; TBI 5 traumatic brain injury. b
the covariates might be inappropriate, because for example depression might be an early sign of YOD rather than a risk factor for YOD. In 1 of the statistical models, we therefore only adjusted the hazards for overall premorbid cognitive function, alcohol intoxications, age, place, and year of conscription. This resulted in hazards that were attenuated to an almost similar degree as in the fully adjusted models (see Table 2). We also evaluated the relationship between TBIs of different severity and the risk of death in propensity score–matched cohorts based on all variables in Table 1. One severe TBI was associated with a higher risk of death in the first year after TBI (HR 5 11.4, 95% CI 5 7.6–17.0) compared with 1 mild TBI (HR 5 2.2, 95% CI 5 1.6–3.1) and at least 2 mild TBIs (HR 5 4.0, 95% CI 5 2.2–7.0; Table 3). Finally, we evaluated whether there were any significant interactions for the dif378
ferent covariates and TBIs according to the outcome of YOD of all causes (Table 4). For those with at least 2 mild TBIs, there were significant interactions with respect to height and depression. Accordingly, men with at least 2 mild TBIs and a height of at least 180cm had a higher risk of YOD than those with a lower height (odds ratio [OR] 5 13.2 vs 4.4), and those with no diagnosed depression had a higher risk than those with a diagnosed depression (OR 5 9.5 vs 3.3). After adjusting the probability values according to Bonferroni, none was significant.
Discussion In the present nationwide cohort study, we found only 14 cases of YOD during >30 years of follow-up in 45,249 patients with any type of TBI. Thus, although the HR for the association between TBI and YOD of Volume 75, No. 3
Nordstr€ om et al: TBI and Young Onset Dementia
TABLE 3. Cox Regression Analysis of TBI and the Risk of Death in Propensity Score–Matched Cohorts
Deaths during the First Year of Follow-up
All Deaths during Follow-up TBI Type
Deaths during Follow-up after the First Year
No.
HR
95% CI
No.
HR
95% CI
No.
HR
95% CI
1 mild TBI, 31,755 cases and 31,755 controls
2,760
1.4
1.3–1.5
143
2.2
1.6–3.1
2,617
1.3
1.2–1.4
At least 2 mild TBIs, 4,069 cases and 4,069 controls
637
2.0
1.7–2.3
74
4.0
2.2–7.0
563
1.8
1.5–2.1
1 severe TBI, 5,102 cases and 5,102 controls
998
3.1
2.7–3.6
355
11.4
7.6–17.0
643
2.0
1.7–2.3
Cohorts were matched based on propensity scores originating from all variables listed in Table 1. The baseline date for each case– control pair was the date of TBI in the case. Pairs in which the control died or emigrated before the date of TBI were excluded. Piecewise Cox regression analysis was used to estimate the risk of death within 1 year of the TBI, and >1 year after the TBI. CI 5 confidence interval; HR 5 hazard ratio; TBI 5 traumatic brain injury.
AD type had a modest precision, the absolute risk seems to be low. For other forms of YOD, there were strong associations with previous TBIs in initial age-adjusted analysis. These associations were, however, markedly reduced after adjustment for rather crude risk factors for TBI and YOD, such as cognitive function in young adulthood and alcohol intoxication.
Based on the findings of previous studies, we hypothesized that severe TBI would be associated with a higher risk of YOD than would 1 mild TBI.6,10 Such an association would also make sense from a mechanistic perspective. However, only 1 of about 6,000 men with at least 1 severe TBI was diagnosed with AD during a follow-up period of >30 years. Thus, the absolute risk of
TABLE 4. Logistic Regression Was Used to Test for Interaction Effects for the Different Covariates with Respect to 1 mTBI, >1 mTBI, and at Least 1 Severe TBI (1 TBI) and the Outcome of All Cases of Young Onset Dementia
Covariates
1 mTBI
>1 mTBI
1 TBI
Body height < 180cm
0.87
0.01
0.66
Body weight > 68kg
0.19
0.18
0.49
Knee strength < 540N
0.09
0.93
0.79
Systolic blood pressure > 128mmHg
0.57
0.64
0.61
Overall cognitive function < 21.45 points
0.85
0.45
0.31
Total income < $26,800 US
0.53
0.44
0.52
University education
0.61
0.61
0.11
Dementia in father
0.51
0.35
0.96
Dementia in mother
0.37
0.61
0.99
TBI in father
0.71
0.73
0.99
TBI in mother
0.72
0.14
0.83
Alcohol intoxication
0.61
0.69
0.21
Drug intoxication
0.54
0.60
0.66
Stroke
0.10
0.42
0.25
Depression
0.18
0.02
0.26
Probability values are presented. The associations were adjusted for the influence of age and year of conscription. Continuous variables were analyzed according to the upper and lower half. mTBI 5 mild traumatic brain injury; TBI 5 traumatic brain injury.
March 2014
379
ANNALS
of Neurology
AD early in life must be considered to be extremely low, even after a severe TBI. For other forms of YOD, the observed graded relationship with TBI severity in the initial analysis (adjustment for age and place and year of conscription) supported the results of a previous study10; 1 mild TBI increased the risks of all non-AD forms of YOD 4-fold, whereas multiple mild TBIs and 1 severe TBI increased the risks about 10-fold. However, these high risks were largely explained by premorbid cognitive function and alcohol abuse. Thus, after adjusting for these covariates, and age at baseline, the graded association between TBI severity and the risk of YOD was no longer evident, and the overall associated risk was markedly reduced for TBIs of all severities. Nevertheless, a severe TBI was still associated with a more than doubled risk of YOD of non-AD forms. It seems likely that at least some of this association is explained by residual confounding. A TBI might also increase the risk of YOD by interacting with other known risk factors. In the present study, there were no significant interactions supporting such a hypothesis. We evaluated previous studies investigating associations between TBI and the risk of predominantly late life dementia, including a meta-analysis of 15 case–control studies published before August 2001. This meta-analysis reported that a previous TBI resulting in loss of consciousness was associated with an increased risk of AD in men.8 However, in separate analyses of the results, only 3 of 15 studies showed significant associations between TBI and the risk of AD.26–28 One of 4 cohort studies10,11,29,30 reported a significant association between previous TBI and the risk of AD,10 and another reported a significant association only in a subcohort of younger participants.30 Two of the studies above reporting significant associations adjusted for alcohol use according to interviews or questionnaires,26,30 whereas no study considered other important risk factors for TBI and dementia identified in the present and previous recent studies, including diagnosed alcohol intoxication, premorbid cognitive function, physical fitness, and blood pressure.15–19 In contrast, Lee et al31 showed in a quite recently published study that mild TBIs were associated with 3.33 higher risk of AD after taking into account an impressive number of covariates including many of those above. However, given a mean follow-up time of 1 year between mild TBI and the diagnosis of dementia, the risk of reverse causality seems substantial. There are limitations to the present study that should be considered. The cohort studied included only men. Risk factor profiles for TBI and early dementia are likely to differ in women, in whom the risk of TBI is lower.32 Diagnoses of TBIs of different severity and 380
dementia were retrieved from national registers. Thus, the absent graded associations between TBI severity and YOD could have been influenced by misclassification of TBI and=or dementia. A number of sensitivity analyses were therefore performed to evaluate these diagnoses in the present study. To test the sensitivity of TBIs of different severity, we evaluated the outcome of death, expecting to find a graded association. In propensity score–matched cohorts, the risk of death was 113 higher in the first year after severe TBI, compared with 23 higher after mild TBI; these results clearly indicate the presence of a graded association. We did also find that the time in hospital was on average 103 longer for men with TBIs classified as severe compared to mild. In a recent study, we have also shown that subjects who sustain 1 mild TBI have clearly different characteristics compared to subjects who sustain multiple mild TBIs or 1 severe TBI.19 The diagnosis of YOD has been validated in a recent study, where we found that 75 of 79 ( 95%) cases of YOD were correctly diagnosed in a similar cohort.33 This high percentage is to be expected, given that suspected cases of dementia early in life are usually carefully evaluated in Sweden. Finally, with a higher risk of death after a TBI, mortality might be considered as a competing event for the development of dementia. Nevertheless, competing risk subdistribution hazard models are considered inappropriate in etiologic research.34 The strengths of the present study include the nationwide cohort of men, well characterized at baseline, the main exposure of about 45,000 TBIs of different severity, and finally the use of Swedish national registers covering all citizens living in Sweden resulting in no loss to follow-up, increasing the external validity of the results. In the present study, the risk of YOD of AD type after TBIs of different severity was low during >30 years of follow-up. For other forms of dementia, the associations were largely attenuated but remained significant after adjustment for 2 covariates: cognitive function in young adulthood and alcohol intoxications. Finally, there was no indication of a graded association between TBI severity and the risk of YOD of any subtype after adjustment for confounders. The association between TBIs and YOD would be of interest to study further in other cohorts including also women.
Acknowledgment The study was supported by grants from the Swedish research council, Swedish Dementia Foundation, and Swedish Alzheimer Foundation. The funding bodies did Volume 75, No. 3
Nordstr€ om et al: TBI and Young Onset Dementia
not have any active role in any part of the research performed within the present study.
Authorship P.N. conceived the idea for the study; compiled and analyzed the data, and made initial drafts of tables, figures, and results, with input from all other authors; and led the writing of the article, with contributions from all other authors.
Potential Conflicts of Interest
14.
Dams-O’Connor K, Gibbons LE, Bowen JD, et al. Risk for late-life re-injury, dementia and death among individuals with traumatic brain injury: a population-based study. J Neurol Neurosurg Psychiatry 2013;84:177–182.
15.
Snowdon DA, Kemper SJ, Mortimer JA, et al. Linguistic ability in early life and cognitive function and Alzheimer’s disease in late life. Findings from the Nun Study. JAMA 1996;275:528–532.
16.
Whalley LJ, Deary IJ. Longitudinal cohort study of childhood IQ and survival up to age 76. BMJ 2001;322:819.
17.
Fratiglioni L, Paillard-Borg S, Winblad B. An active and socially integrated lifestyle in late life might protect against dementia. Lancet Neurol 2004;3:343–353.
18.
Nordstrom A, Nordstrom P. Cognitive performance in late adolescence and the subsequent risk of subdural hematoma: an observational study of a prospective nationwide cohort. PLoS Med 2011;8:e1001151.
19.
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.
20.
Rossor MN, Fox NC, Mummery CJ, et al. The diagnosis of youngonset dementia. Lancet Neurol 2010;9:793–806.
Nothing to report.
References 1.
Murray C, Lopez A. Global health statistics: a compendium of incidence, prevalence and mortality estimates for over 200 conditions. 1996.
21.
2.
Hyder AA, Wunderlich CA, Puvanachandra P, et al. The impact of traumatic brain injuries: a global perspective. NeuroRehabilitation 2007;22:341–353.
Atkins ER, Bulsara MK, Panegyres PK. The natural history of earlyonset dementia: the Artemis Project. BMJ Open 2012;2(5).
22.
Alexander MP. Mild traumatic brain injury: pathophysiology, natural history, and clinical management. Neurology 1995;45:1253– 1260.
Svanberg E, Spector A, Stott J. The impact of young onset dementia on the family: a literature review. Int Psychogeriatr 2011;23:356–371.
23.
Ashman TA, Gordon WA, Cantor JB, Hibbard MR. Neurobehavioral consequences of traumatic brain injury. Mt Sinai J Med 2006; 73:999–1005.
Andreasson S, Allebeck P, Romelsjo A. Alcohol and mortality among young men: longitudinal study of Swedish conscripts. Br Med J (Clin Res Ed) 1988;296:1021–1025.
24.
Carlstedt B, Ma˚rdberg B. Construct validity of the Swedish enlistment battery. Scand J Psychol 1993;34:353–362.
25.
Carlstedt B, Gustafsson JE. Construct validation of the Swedish Scholastic Aptitude Test by means of the Swedish Enlistment Battery. Scand J Psychol 2005;46:31–42.
26.
Mortimer JA, French LR, Hutton JT, Schuman LM. Head injury as a risk factor for Alzheimer’s disease. Neurology 1985;35:264–267.
27.
Shively S, Scher AI, Perl DP, Diaz-Arrastia R. Dementia resulting from traumatic brain injury: what is the pathology? Arch Neurol 2012;69:1245–1251.
Graves AB, White E, Koepsell TD, et al. The association between head trauma and Alzheimer’s disease. Am J Epidemiol 1990;131: 491–501.
28.
Fleminger S, Oliver DL, Lovestone S, et al. Head injury as a risk factor for Alzheimer’s disease: the evidence 10 years on; a partial replication. J Neurol Neurosurg Psychiatry 2003;74:857–862.
O’Meara ES, Kukull WA, Sheppard L, et al. Head injury and risk of Alzheimer’s disease by apolipoprotein E genotype. Am J Epidemiol 1997;146:373–384.
29.
Katzman R, Aronson M, Fuld P, et al. Development of dementing illnesses in an 80-year-old volunteer cohort. Ann Neurol 1989;25: 317–324.
30.
Plassman BL, Havlik RJ, Steffens DC, et al. Documented head injury in early adulthood and risk of Alzheimer’s disease and other dementias. Neurology 2000;55:1158–1166.
Schofield PW, Tang M, Marder K, et al. Alzheimer’s disease after remote head injury: an incidence study. J Neurol Neurosurg Psychiatry 1997;62:119–124.
31.
Williams DB, Annegers JF, Kokmen E, et al. Brain injury and neurologic sequelae: a cohort study of dementia, parkinsonism, and amyotrophic lateral sclerosis. Neurology 1991;41:1554–1557.
Lee YK, Hou SW, Lee CC, et al. Increased risk of dementia in patients with mild traumatic brain injury: a nationwide cohort study. PLoS One 2013;8:e62422.
32.
Mehta KM, Ott A, Kalmijn S, et al. Head trauma and risk of dementia and Alzheimer’s disease: the Rotterdam Study. Neurology 1999;53:1959–1962.
Feigin VL, Theadom A, Barker-Collo S, et al. Incidence of traumatic brain injury in New Zealand: a population-based study. Lancet Neurol 2013;12:53–64.
33.
Katzman R, Aronson M, Fuld P, et al. Development of dementing illnesses in an 80-year-old volunteer cohort. Ann Neurol 1989;25: 317–324.
Nordstrom P, Nordstrom A, Eriksson M, et al. Risk factors in late adolescence for young-onset dementia in men: a nationwide cohort study. JAMA Intern Med 2013;173:1612–1618.
34.
Lau B, Cole SR, Gange SJ. Competing risk regression models for epidemiologic data. Am J Epidemiol 2009;170:244–256.
3.
4.
5.
McAllister TW, Flashman LA, McDonald BC, Saykin AJ. Mechanisms of working memory dysfunction after mild and moderate TBI: evidence from functional MRI and neurogenetics. J Neurotrauma 2006;23:1450–1467.
6.
Nemetz PN, Leibson C, Naessens JM, et al. Traumatic brain injury and time to onset of Alzheimer’s disease: a population-based study. Am J Epidemiol 1999;149:32–40.
7.
8.
9.
10.
11.
12.
13.
Mortimer JA, van Duijn CM, Chandra V, et al. Head trauma as a risk factor for Alzheimer’s disease: a collaborative re-analysis of case-control studies.EURODEM Risk Factors Research Group. Int J Epidemiol 1991;20(suppl 2):S28–S35.
March 2014
381