American Journal of Epidemiology
Copyright © 1999 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 149, No. 1
Printed in U.S.A.
Traumatic Brain Injury and Time to Onset of Alzheimer's Disease: A
Population-based Study
Peter N. Nemetz,12 Cynthia Leibson,1 James M. Naessens,1 Mary Beard,1 Emre Kokmen,3 John F. Annegers,14
and Leonard T. Kurland1
Controversy continues as to whether traumatic brain injury is a risk factor for Alzheimer's disease. The authors
examined a related hypothesis that among persons with traumatic brain injury who develop Alzheimer's disease,
time to onset of the disease is reduced. They used data on all documented episodes of traumatic brain injury
that occurred from 1935 to 1984 among Olmsted County, Minnesota, residents. Community-based medical
records were used to follow traumatic brain injury cases who were aged 40 years or older at last contact prior to
June 1, 1988, for Alzheimer's disease until last contact, death, or June 1, 1988. The test of the hypothesis was
restricted to those cases who developed Alzheimer's disease. The expected time to onset of Alzheimer's disease
was derived from a life table constructed by using age-of-onset distributions within sex groups for a previously
identified cohort of Rochester, Minnesota, Alzheimer's disease incidence cases without a history of head trauma.
The authors found that of the 1,283 traumatic brain injury cases followed, 31 developed Alzheimer's disease, a
number similar to that expected (standardized incidence ratio = 1.2, 95% confidence interval 0.8-1.7). However,
the observed time from traumatic brain injury to Alzheimer's disease was less than the expected time to onset
of Alzheimer's disease (median = 10 vs. 18 years, p = 0.015). The results suggest that traumatic brain injury
reduces the time to onset of Alzheimer's disease among persons at risk of developing the disease. Am J
Epidemiol 1999; 149:32-40.
Alzheimer's disease; brain injuries; cohort studies; head injuries; incidence; medical records; risk factors;
traumatology
of this etiologic factor has been mixed (4, 5, 8-27), as
illustrated in figure 1. Although most of these studies
reported odds ratios of more than 1, very few of these
odds ratios were statistically significant. A metaanalysis of seven case-control studies supported an
association between head injury and Alzheimer's disease for males only (26). However, in a populationbased cohort study of 821 persons with traumatic brain
injury whose medical records were followed for subsequent evidence of Alzheimer's disease, the overall
incidence of Alzheimer's disease was not greater than
that expected based on rates for the general population
(25).
While the controversial issue of prior traumatic brain
injury and risk for Alzheimer's disease may not be
resolved satisfactorily for some time, a related hypothesis has been advanced concerning traumatic brain injury
and the timing of onset to Alzheimer's disease (28-30).
This hypothesis suggests that for those who for some as
yet unspecified reason are at risk of developing
Alzheimer's disease, a head injury could hasten the
onset of dementia. In this paper, we use the populationbased resources of the Rochester Epidemiology Project
to examine the hypothesis that among persons with traumatic brain injury who develop Alzheimer's disease, the
The number of Alzheimer's disease cases in the
United States has been estimated at 1.1-3.4 million
during the early 1980s, 2-4 million during the early
1990s, and more than 15.4 million by the year 2050
(1-3). Although research has focused on a broad range
of potential risk factors or pathogenetic mechanisms
including, among others, family history, specific
genetic factors, autoimmune disorders, slow viruses,
aluminum toxicity, and prior thyroid disease, the causes of this devastating and frequently occurring disease
remain largely unknown (4-9).
One of the most controversial suggested risk factors
for Alzheimer's disease is prior traumatic brain injury.
Evidence in the medical literature of the potential role
Received for publication November 7, 1997, and accepted for
publication May 5, 1998.
Abbreviations: Cl, confidence interval; SIR, standardized incidence ratio.
1
Department of Health Sciences Research, Mayo Clinic,
Rochester, MN.
2
Policy Analysis Division, Faculty of Commerce and Business
Administration, University of British Columbia, Vancouver, British
Columbia, Canada.
3
Department of Neurology, Mayo Clinic, Rochester, MN.
4
Houston School of Public Health, University of Texas, Houston,
TX.
32
Brain Injury and Alzheimer's Disease
33
Odds ratio
0 .1
1
10
Molgaardetal. (18)*
Heyman et al. (16)
100
•
French et al. (4) f
Mayeux et al. (17)
Graves et al. (5)
Mortimer etal. (26) t/
Statistically
significant
•
Chandra etal. (13)
Amaducci et al. (11) f
French et al. (4) §
•
Shalat et al. (22)
Paschalis et al. (27)
Amaducci et al. (11) f
Not
statistically
significant
•
van Duijn et al. (24)
Broeetal. (12)
Chandra etal. (14)
Ferini-Strambi et al. (15)
•
Soininen & Heinonen (23) #
W
95 percent confidence interval - upper limit
H Odds ratio
W' 95 percent confidence interval - lower limit
FIGURE 1. Odds ratios and 95% confidence intervals reported in previous investigations of the association between Alzheimer's disease and
head injury. *, "mild intellectual impairment"; t , hospital controls; t, meta-analysis—only relative risk reported; §, population controls; % medium to severe trauma; #, mild trauma. Four studies (11,13,16, 23) did not provide confidence intervals for their estimates but did provide p values for test of significance.
observed time to onset is less than the expected time to
onset of Alzheimer's disease.
MATERIALS AND METHODS
The Rochester Epidemiology Project is particularly
suitable for use in addressing the issue of traumatic
brain injury and the timing of subsequent Alzheimer's
disease because of its virtually complete longitudinal
coverage of the relatively stable, geographically
defined population of Rochester, Minnesota, and surrounding Olmsted County (total 1990 population,
106,470) (31). Established in 1966 with funding from
the National Institutes of Health (grant AR30582), the
Rochester Epidemiology Project affords access to
complete medical records from essentially all sources
of medical care available to and used by the local population. These data are drawn from the Mayo Clinic
unit record system, in existence since 1907, and are
supplemented by the records of subsequent non-Mayo
providers. These records, together with indexes of virAm J Epidemiol
Vol. 149, No. 1, 1999
tually all clinical and histologic diagnoses and all surgical procedures for the local population, make population-based epidemiologic
studies possible.
Population-based studies alleviate problems of selective survival and referral and of recall bias that hamper
most other studies of Alzheimer's disease etiology.
A previous population-based study (32) used the
resources of the Rochester Epidemiology Project to
identify all documented episodes of head trauma with
presumed brain involvement among residents of
Olmsted County, Minnesota. Brain involvement was
defined as injury with loss of consciousness, posttraumatic amnesia, or neurologic signs of brain injury.
Cases of head trauma with skull fracture were included, even if alterations in consciousness could not be
documented. A subsequent study by Williams et al.
(25) was limited to 821 traumatic brain injury cases
who were aged 40 years or older at last medical contact. The medical records of these persons were
reviewed for neurologic sequelae, including
Alzheimer's disease. Standardized incidence ratios
34
Nemetz et al.
were used to investigate whether the number of
Alzheimer's disease cases was higher than expected;
the expected number was based on published incidence rates of Alzheimer's disease for the community
of Rochester, Minnesota (33). The investigation by
Williams et al. was based on episodes of traumatic
brain injury during 1935-1974 (25).
The Olmsted County traumatic brain injury cohort
has subsequently been updated through 1984. The present study used the updated cohort to address the
hypothesis that among persons with traumatic brain
injury who developed Alzheimer's disease, the
observed time to onset was less than the expected time
to onset of the disease. Our study also replicated the
Williams et al. (25) analysis of the standardized incidence ratios on the updated cohort. All Olmsted
County residents with traumatic brain injury in
1935-1984 who were aged 40 years or older at last
medical contact prior to June 1, 1988, were followed in
their medical records for evidence of Alzheimer's disease until last contact, death, or June 1, 1988.
Alzheimer's disease was defined by using the same
criteria used to determine the incidence of this disorder
in Rochester, Minnesota, between 1960 and 1984
(table 1) (34). For both the Rochester Alzheimer's disease incidence cohort and the Alzheimer's disease
cases in the traumatic brain injury cohort, the diagnosis of Alzheimer's disease was confirmed by the study
neurologist (E. K.), who used the best available information to review the records in a standardized manner
across calendar time to establish a year of Alzheimer's
disease onset (35). This date was usually the first mention in the medical record of cognitive or behavioral
difficulties that subsequently were determined to be
due to Alzheimer's disease.
To replicate the standardized incidence ratio analysis
by Williams et al. (25) in their investigation of
Alzheimer's disease incidence in the 1935-1974 traumatic brain injury cohort, we adopted the following procedure: we divided the observed number of Alzheimer's
disease cases with traumatic brain injury in the updated
cohort by the expected number of cases, based on ageand sex-specific incidence rates of Alzheimer's disease
observed for the Rochester, Minnesota, population in
1960-1984 (34). We expanded on the earlier investigation by calculating standardized incidence ratios separately for men and women, for early (age <75 years) and
late (age >75 years) Alzheimer's disease onset, and for
short (<10 years) and long (>10 years) periods of follow-up after traumatic brain injury. Ninety-five percent
confidence intervals were calculated, assuming a
Poisson distribution (36).
Our principal hypothesis was that among persons
with traumatic brain injury who developed
TABLE 1. Diagnostic criteria for dementia and Alzheimer's
disease*
Dementia
Documented evidence:
Previously normal intellectual and social function
Progressive decline of intellectual/cognitive/social function
that cannot be reversed with medical or psychiatric
treatment
Memory impairment
Dementia sufficient to impair age-/education-/occupationappropriate lifestyle adjustment
In addition, documented evidence of at least two of the
following:
Disorientation
Personality/behavioral problems
Dyscalculia
Aphasia or apraxia or agnosia
Impairment of judgment/abstract thinking
Alzheimer's disease
Clinical diagnosis (probable Alzheimer's disease):
Presence of dementia (as defined above)
Insidious onset of symptoms of dementia
Gradual progression and irreversible course
Other potential causes of dementia either not present or
occurring definitely after the onset of symptoms of
dementia
Neuropathologic diagnosis (definite Alzheimer's disease):
Presence of clinical Alzheimer's disease (as defined
above)
Presence of abundant neuritic (senile) plaques and/or
neurofibrillary tangles in neocortical regions
excluding the hippocampus and subiculum
* Adapted from Kokmen et al. (34).
Alzheimer's disease, time to onset of the disease was
shorter than expected. Observed time to onset was calculated as the date of Alzheimer's disease onset minus
the date of the first documented traumatic brain injury.
Persons who developed Alzheimer's disease prior to
traumatic brain injury were excluded from the analysis. Expected time to onset was estimated by using a
life table approach. This approach, frequently used in
demography, is initiated by identifying all persons in a
population who experience a particular event (e.g.,
death). Using distributions of the ages at which that
event occurred, a life table is constructed that enables
calculation of the length of survival free of that event
for every age (e.g., remaining life expectancy at age 40
years, at age 45 years, at age 50 years). In our study,
the event of interest was Alzheimer's disease rather
than death. We took advantage of data from the previously identified cohort consisting of all Rochester
Alzheimer's disease incidence cases in 1965-1984,
after excluding 151 cases previously identified as having a history of head trauma. By using age at
Alzheimer's disease onset for each of the 689
Rochester Alzheimer's disease incidence cases without
Am J Epidemiol Vol. 149, No. 1, 1999
Brain Injury and Alzheimer's Disease
a history of head trauma, we constructed a life table to
calculate the length of survival free of Alzheimer's disease for each age. For all traumatic brain injury cases
who developed Alzheimer's disease, we compared
their observed length of survival from traumatic brain
injury to Alzheimer's disease with their expected
length of survival free of Alzheimer's disease (as
obtained from the life table estimate, specific for the
age at which that traumatic brain injury case experienced traumatic brain injury).
The analysis excluded persons for whom both traumatic brain injury and Alzheimer's disease onset
occurred in the same year, because it was not possible
to determine whether Alzheimer's disease onset preceded traumatic brain injury. By excluding these persons, we drastically reduced the likelihood of including any Alzheimer's disease events that occurred in the
first 6 months following traumatic brain injury.
Therefore, the first 6 months of follow-up were
ignored in the analysis of time to onset. Observed time
to Alzheimer's disease onset was compared graphically with expected time to Alzheimer's disease onset.
The significance of observed versus expected time to
onset of Alzheimer's disease was tested by using a
one-sample log-rank test (37). Separate analyses were
also performed for men and women, for traumatic
brain injury before age 65 years and at age 65 years or
older, and for duration of follow-up after traumatic
brain injury of less than 10 years and of 10 years or
longer.
RESULTS
The 1935-1984 Olmsted County, Minnesota, traumatic brain injury cohort consisted of 5,023 persons.
The 1,283 who were aged 40 years or older as of last
medical contact prior to June 1, 1988, contributed a
total of 17,279 person-years of follow-up, during
which 31 persons were identified as subsequently
developing Alzheimer's disease. Because traumatic
brain injury was more frequent among men than
women (65 percent of the 1,283 traumatic brain injury
cases were men), the proportion of men among the 31
traumatic brain injury cases who developed
Alzheimer's disease (55 percent) was higher than the
proportion of men in the 1960-1984 Rochester
Alzheimer's disease incidence cohort (28 percent). For
the 31 traumatic brain injury cases, the calendar year
of Alzheimer's disease onset ranged from 1954 to
1986; the median age at onset was 76 years for men
and 80 years for women. Although the median age of
the Rochester Alzheimer's disease incidence cohort in
the first quinquennium (1960-1964) was also 76 years
for men and 80 years for women, these ages had
increased to 80 and 83 years, respectively, by
Am J Epidemiol Vol. 149, No. 1, 1999
35
1980-1984, due in part to a shift in the age distribution
of the underlying population (34).
For the 31 persons with traumatic brain injury who
developed Alzheimer's disease, observed time from
injury to Alzheimer's disease onset was compared with
expected time to onset of Alzheimer's disease. As
shown in figure 2, the observed time to onset was
shorter than expected (median 10 vs. 18 years; onesample log-rank test, p = 0.015). The observed time to
onset was significantly shorter than the expected time
to onset of Alzheimer's disease regardless of whether
onset occurred less than 10 years after traumatic brain
injury (n = 15, p = 0.004) or 10 years or more after
traumatic brain injury (n = 16, p < 0.001). When stratified by age at time of brain injury, the observed time
from injury to Alzheimer's disease onset was significantly shorter than the expected time to onset for the
16 cases whose injuries occurred before age 65 years
(p < 0.001) but was not different from the expected
time to onset of Alzheimer's disease for the 15 cases
whose injuries occurred after age 65 years (p = 0.867).
Of those with traumatic brain injury who developed
Alzheimer's disease, the median age at traumatic brain
injury was 66 years for the 14 women compared with
58 years for the 17 men (rank sum test, p = 0.1).
Among these men, a significant reduction in time to
Alzheimer's disease onset was apparent only for the 11
cases whose traumatic brain injuries occurred before
age 65 years (p - 0.01). The observed time to onset of
Alzheimer's disease was not different from the expected time to onset (p = 0.20) for the six men whose
injuries occurred after age 65 years, although the relatively small numbers contributed to low statistical
power. For the 14 women, the observed time to onset
was significantly shorter than the expected time to
onset, irrespective of age at time of injury (p < 0.001
for age <65 years, p = 0.03 for age >65 years).
We were concerned that the significant reduction in
time to onset was due to an outlier, that is, a single
female whose age at Alzheimer's disease onset was
less than 45 years. This person was excluded, and the
analyses were rerun; the differences between the
observed and the expected time to onset remained significant (p < 0.001 for women, p < 0.001 for age at
head trauma of <65 years).
Age- and sex-specific Alzheimer's disease incidence
rates for the 1,283 persons with traumatic brain injury
were compared with 1960-1984 Alzheimer's disease
incidence rates for the community of Rochester,
Minnesota (table 2). The number of cases of
Alzheimer's disease among persons with traumatic
brain injury (n = 31) was not higher than expected
(standardized incidence ratio (SIR) for both sexes
combined = 1.2, 95 percent confidence interval (CI)
36
Nemetz et al.
eo
es
c
I
a
o
t
3
0
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30 32 34 36
Time to onset of Alzheimer's disease (years)
FIGURE 2. Observed versus expected percentage of survival free of Alzheimer's disease (AD) following traumatic brain injury (TBI) among
the 31 members of the 1935-1984 Olmsted County, Minnesota, traumatic brain injury incidence cohort who developed Alzheimer's disease prior
to June 1,1988. Expected length of survival was based on life table estimates derived from age- and sex-specific Alzheimer's disease incidence
rates for the 689 1965-1984 Rochester, Minnesota, Alzheimer's disease incidence cases with no mention of prior head trauma.
0.8-1.7), a result consistent with the findings of
Williams et al. (25) in their analysis of Alzheimer's
disease incidence in the 1935-1974 traumatic brain
injury cohort (SIR = 1.0, 95 percent CI 0.6-1.5). The
findings were similar when stratified by sex (SIR for
men =1.4, 95 percent CI 0.8-2.3; SIR for women =
1.0, 95 percent CI 0.5-1.6). However, the analysis
stratified by age at onset of Alzheimer's disease
revealed that the 11 traumatic brain injury cases who
experienced the onset of Alzheimer's disease before
age 75 years were more than twice the number expected (SIR = 2.2, 95 percent CI 1.1-4.0). The standard-
ized incidence ratio for Alzheimer's disease onset at
age 75 years or older was 0.9 (95 percent CI 0.6-1.5).
The number of observed cases did not differ from that
expected when stratified by years of observation (SIR
for <10 years of follow-up = 1.2, 95 percent CI
0.7-1.9; SIR for >10 years of follow-up = 0.9, 95 percent CI 0.5-1.5).
DISCUSSION
This study took advantage of the fact that all documented cases of traumatic brain injury that occurred in
TABLE 2. Incidence rates* of Alzheimer's disease among Olmsted County, Minnesota, residents with traumatic brain injury in
1935-1984 who were aged >40 years at last medical contact and among Rochester, Minnesota, residents in 1960-1984
Incidence in Rochester
Incidence in the traumatic brain injury cohort
Men
(years)
40-44
45-54
55-64
65-74
75-84
>85
Women
Women
Men
Total
No.
Rate
No.
Rate
No.
Rate
No.
Rate
No.
0
0
1
6
8
2
0.00
0.00
38.62
352.96
999.41
1,119.48
1
0
1
2
8
2
62.83
0.00
96.41
231.67
1,321.75
836.18
1
0
18.74
0.00
55.15
312.11
1,138.20
957.31
0
1
13
62
0.00
1.80
30.50
229.40
960.49
1,704.06
2
11
96
336
243
2
8
16
4
132
60
1
Total
Rate
No.
Rate
0.20
3.19
19.76
211.86
1,110.05
2,262.57
1
3
24
158
468
303
0.11
2.54
24.42
218.41
1,063.35
2,124.68
Per 100,000 person-years.
Am J Epidemiol
Vol. 149, No. 1, 1999
Brain Injury and Alzheimer's Disease
1935-1984 among residents of Rochester, Minnesota,
had already been identified. Cases who were aged 40
years or older at last medical contact prior to June 1,
1988, were followed in community-based medical
records for evidence of Alzheimer's disease. The finding that the number of observed Alzheimer's disease
cases was not significantly higher than expected was
consistent with findings from previous Rochester
Epidemiology Project studies (14, 25), which concluded that traumatic brain injury was not a significant
independent risk factor for Alzheimer's disease. The
present study expanded on previous Rochester
Epidemiology Project investigations to ask whether,
for persons with traumatic brain injury who developed
Alzheimer's disease, the time from traumatic brain
injury to Alzheimer's disease onset was significantly
shorter than the expected time to onset. Our results
support that hypothesis.
In addition, when we stratified the analysis of standardized incidence ratios by age at onset of
Alzheimer's disease, the number of traumatic brain
injury cases who had developed early-onset
Alzheimer's disease was more than twice that expected. The increased risk for early-onset Alzheimer's disease reported here is consistent with the significant
association between a history of head trauma and the
onset of Alzheimer's disease before age 70 years for
males found by Mortimer et al. (26) in their metaanalysis of seven case-control studies. However, in
contrast to the results reported here, the meta-analysis
also revealed a significant association between a history of head trauma and the onset of Alzheimer's disease
at age 70 years or older (26).
A community-based prospective study by Schofield
et al. (30) investigated time to Alzheimer's disease
onset in a manner somewhat different from that used
here. They identified 271 elderly persons free of
dementia at baseline and followed them for up to 5
years for evidence of dementia. A self-reported history
of head trauma at baseline was available from two
sources, a physician's query about head trauma with
loss of consciousness and a trained interviewer's query
about head trauma with loss of consciousness or with
amnesia. The limitations of self-report were evidenced
by the finding that of those who responded yes to the
former query, 26 percent responded no to the latter
query. Cox proportional hazards analysis revealed a
significantly increased hazard for persons who
responded yes to the physician's query. The relative
hazard was not significantly increased for those who
responded yes to the interviewer's query, but there was
a significant difference between those who responded
no and those who reported a loss of consciousness of
more than 5 minutes (30).
Am J Epidemiol Vol. 149, No. 1, 1999
37
A few studies have compared age at onset for
Alzheimer's disease cases with and without mention of
head trauma; these studies have yielded conflicting
results (26, 28-29). A study by Rasmusson et al. (38)
of 68 Alzheimer's disease cases found no effect of
head injury on age at onset. However, a study by
Sullivan et al. (29) of 17 Alzheimer's disease cases
reported a younger mean age at onset for cases with a
history of head trauma than for cases without head
trauma. Gedye et al. (28) studied 148 referral patients
with confirmed probable Alzheimer's disease and also
found that the mean age of Alzheimer's disease onset
was significantly younger for Alzheimer's disease
cases with a history of head injury than for cases with
no history of head injury. However, persons with a history of head injury after age 65 years were excluded
from the exposed group in this comparison. Since no
comparable adjustment was made or could have been
made in the unexposed group, the average age of onset
among those with a history of head injury may have
been biased downward, thus contributing to the
observed statistical difference. In all of these studies, a
history of head trauma was obtained by proxy report.
In addition to the potential for recall bias with proxy
or self-report, most previous investigations of the association between traumatic brain injury and time to
Alzheimer's disease onset or age at Alzheimer's disease onset were also subject to problems of referral
bias and selective survival. By using the Rochester
Epidemiology Project resources, we were able to
address a number of these issues. Advantages included
the following: 1) identification of all documented
episodes of traumatic brain injury in a geographically
defined population, 2) comparison with populationbased age- and sex-specific Alzheimer's disease incidence rates for unaffected persons, and 3) a historic
cohort study design with sufficient follow-up from
first documented traumatic brain injury to last medical
contact, death, or end of study period to accumulate a
relatively large number of Alzheimer's disease cases.
Thus, the problems of insufficient power and the
potential for selection bias, recall bias, and/or referral
bias were minimized in the present study.
Our study is not without limitations. Confirmation
of a diagnosis of dementia, classification as
Alzheimer's disease, and assignment of year of
Alzheimer's disease onset were based on retrospective
record review. Not all persons underwent neuropsychological or imaging tests, and some were never evaluated by either a neurologist or a psychiatrist. The
progress of all medical conditions and symptoms of
dementia over time was assessed with the longitudinal
information available from the complete communitybased medical records. Cases were often followed
38
Nemetz et al.
until death. Information was accumulated over many
years from multiple sources, including autopsy, to
reconstruct a clear clinical picture of the dementing illness. Postmortem findings took precedence over clinical evaluation results. In a previous Rochester study
(33) of 177 patients with a clinical diagnosis of
Alzheimer's disease, only 42 underwent microscopic
examination of brain tissue; however, the clinical diagnosis was confirmed by histopathologic examination
in 40 of the 42.
This study would miss those Alzheimer's disease
cases who never came to the attention of medical personnel or for whom Alzheimer's disease or dementialike diagnoses were not noted in the patient record.
However, the likelihood of this occurring is low in light
of the comprehensive medical coverage of the target
population. For example, it has been demonstrated that
between 1989 and 1991, more than 96 percent of elderly female Olmsted County residents were seen at least
once by medical care providers, whose diagnoses are
systematically included in the centralized records linkage file (39). Even if several cases were missed by
using this system, the analysis of time to onset would
be biased only if ascertainment was selective against
persons who had traumatic brain injury and experienced late onset of Alzheimer's disease.
Our results are potentially vulnerable to observation
and/or detection bias, as it could be hypothesized that
clinicians might observe traumatic brain injury
patients more closely in subsequent years, thereby
leading to an earlier estimated date of onset of
Alzheimer's disease. If this is the case, a shorter time
period between Alzheimer's disease onset and the clinical diagnosis of Alzheimer's disease for the traumatic
brain injury cases would be expected. To test for this
problem, we compared the time from Alzheimer's disease onset to Alzheimer's disease diagnosis for the 631
persons with no prior head injury for whom dates of
both Alzheimer's disease onset and diagnosis were
available in the 1965-1984 Rochester Alzheimer's disease incidence cohort and the 31 Alzheimer's disease
cases in the traumatic brain injury cohort. The median
time lag between estimated onset and diagnosis for the
Rochester Alzheimer's disease incidence cases without
prior head injury was 1.47 years compared with 2.06
years for the 31 Alzheimer's disease cases with traumatic brain injury (rank sum test, p = 0.08), essentially minimizing if not eliminating the possibility of
observation bias. An alternative possible explanation is
that preclinical features of Alzheimer's disease could
have predisposed some patients to falls or other causes of head injury. To determine whether this type of
explanation is possible, a neurologist must estimate
precisely the date of onset of the disease.
The finding that time to onset of Alzheimer's disease
among persons with traumatic brain injury who developed Alzheimer's disease was shorter than the expected time to onset could also have occurred if there were
preferential loss to follow-up at older ages in the traumatic brain injury cohort relative to the Rochester population. We investigated that possibility by calculating
the proportion of all person-years of follow-up for
those aged 75 years or older. The proportions were 12
percent for men and 21 percent for women in the traumatic brain injury cohort and 12 percent for men and
23 percent for women in the Rochester population.
Therefore, it does not appear as if preferential loss to
follow-up at older ages in the traumatic brain injury
cohort was a problem.
A number of recognized Alzheimer's disease risk
factors were not considered in the analysis, for example, apolipoprotein E genotype. The extent to which
the prevalence of these other risk factors differed
between the traumatic brain injury cohort and the
Rochester population is not known.
The findings presented here may help clarify some
apparent contradictions arising from earlier research.
Our study found that the overall risk of Alzheimer's
disease for persons with traumatic brain injury was not
greater than expected but that the risk of early-onset
Alzheimer's disease was more than twice that expected. Among traumatic brain injury cases who developed
Alzheimer's disease, the observed time to onset was
shorter than the expected time to onset, and the effect
of traumatic brain injury on time to onset was greatest
when the injury occurred before age 65 years. These
findings are consistent with the interpretation that traumatic brain injury interacts with other risk factors to
hasten the onset of Alzheimer's disease in persons susceptible to the disease but that the effect of traumatic
brain injury on time to onset may be diluted by the
exponential increase in the risk of Alzheimer's disease
that occurs with advancing age.
If there is indeed a reduced time to onset of
Alzheimer's disease among Alzheimer's diseasesusceptible persons following traumatic brain injury,
then the question remains as to the causative factors
involved. One central issue in the study of Alzheimer's
disease is the role of P-amyloid, which forms the main
component of senile plaques (40-^42). Deposition of (3amyloid occurs in most persons during the process of
aging (3), and Selkoe states that "the distinction
between Alzheimer's disease and ageing of the brain is
essentially quantitative rather than qualitative" (43, p.
433). Several papers (10, 17, 44-50) have reported an
association between traumatic brain injury and subsequent deposition of P-amyloid. It has also been reported that synthesis of apolipoprotein E, a plasma protein
Am J Epidemiol Vol. 149, No. 1, 1999
Brain Injury and Alzheimer's Disease
involved in the transport of cholesterol, a vital component in cell structure and repair (51-55), increases
markedly after injury and that the apolipoprotein E *e4
allele, a recognized independent risk factor for
Alzheimer's disease, binds closely to the P-amyloid
type 4 peptide (54, 56-62). It has been posited that pamyloid is transported into brain cells by the
apolipoprotein E *e4 allele (55). Several recent studies
lend credence to a possible interaction between the
presence of an apolipoprotein E *e4 allele and traumatic brain injury (63-66), although a more recent study
did not find evidence of an interaction (67). Although
we remain agnostic with respect to the question of Pamyloid causality and Alzheimer's disease, the linkages
between traumatic brain injury, subsequent P-amyloid
deposition, and posited direct or indirect neurotoxicity
are sufficiently compelling to require further study.
Our research suggests that Alzheimer's cases with
prior traumatic brain injury exhibit a reduced time to
onset of Alzheimer's disease relative to those cases
without prior head injury. These conclusions are
strengthened because we used the population-based
resources of the Rochester Epidemiology Project,
thereby significantly reducing or eliminating problems
of selective survival and of referral and recall bias.
This study, as well as the discrepant findings in the
medical literature, supports the argument that the etiology of Alzheimer's disease is multifactorial. Biologic
mechanisms consistent with the results of this study
have been proposed in the literature, therefore suggesting that further research in this area is warranted.
REFERENCES
1. Evans DA. Estimated prevalence of Alzheimer's disease in the
United States. Milbank Q 1990;68:267-89.
2. Hay JW, Ernst RL. The economic cost of Alzheimer's disease.
Am J Public Health 1987;77:1169-75.
3. Yankner B A, Mesulam MM. Seminars in medicine of the Beth
Israel Hospital, Boston. p-Amyloid and the pathogenesis of
Alzheimer's disease. N Engl J Med 1991;325:1849-57.
4. French LR, Schuman LM, Mortimer JA, et al. A case-control
study of dementia of the Alzheimer type. Am J Epidemiol
1985;121:414-21.
5. Graves AB, White E, Koepsell TD, et al. The association
between head trauma and Alzheimer's disease. Am J
Epidemiol 1990;131:491-501.
6. Henderson AS. The risk factors for Alzheimer's disease: a
review and a hypothesis. Acta Psychiatr Scand 1988;78:
257-75.
7. Kokmen E, Beard CM, Chandra V, et al. Clinical risk factors
for Alzheimer's disease: a population-based case-control
study. Neurology 1991 ;41:1393—7.
8. Rocca WA. The etiology of Alzheimer's disease: epidemiologic contributions with emphasis on the genetic hypothesis. J
Neural Transm 1987(suppl);24:3-12.
9. Rocca WA, Amaducci L. The etiology of Alzheimer's disease:
an epidemiologic point of view. Neurobiol Aging 1989; 10:
440-1.
Am J Epidemiol
Vol. 149, No. 1, 1999
39
10. Amaducci L, Lippi A. Risk factors and genetic background for
Alzheimer's disease. Acta Neurol Scand Suppl 1988; 116:
13-18.
11. Amaducci LA, Fratiglioni L, Rocca WA, et al. Risk factors for
clinically diagnosed Alzheimer's disease: a case-control study
of an Italian population. Neurology 1986;36:922-31.
12. Broe GA, Henderson AS, Creasey H, et al. A case-control
study of Alzheimer's disease in Australia. Neurology 1990;40:
1698-707.
13. Chandra V, Philipose V, Bell PA, et al. Case-control study of
late onset "probable Alzheimer's disease." Neurology 1987;
37:1295-300.
14. Chandra V, Kokmen E, Schoenberg BS, et al. Head trauma
with loss of consciousness as a risk factor for Alzheimer's disease. Neurology 1989;39:1576-8.
15. Ferini-Strambi L, Smirne S, Pinto P, et al. Clinical and epidemiological aspects of Alzheimer's disease with presenile
onset: a case control study. Neuroepidemiology 1990;9:39-49.
16. Heyman A, Wilkinson WE, Stafford JA, et al. Alzheimer's disease: a study of epidemiological aspects. Ann Neurol
1984;15:335^».
17. Mayeux R, Ottman R, Tang M, et al. Genetic susceptibility and
head injury as risk factors for Alzheimer's disease among community-dwelling elderly persons and their first-degree relatives. Ann Neurol 1993;33:494-501.
18. Molgaard CA, Stanford EP, Morton DJ, et al. Epidemiology of
head trauma and neurocognitive impairment in a multi-ethnic
population. Neuroepidemiology 1990;9:233^2.
19. Mortimer JA, French LR, Hutton JT, et al. Head injury as a risk
factor for Alzheimer's disease. Neurology 1985;35:264—7.
20. Rocca WA, Amaducci LA, Schoenberg BS. Epidemiology of
clinically diagnosed Alzheimer's disease. Ann Neurol
1986; 19:415-24.
21. Schoenberg BS. Epidemiology of dementia. Neurol Clin
1986;4:447-57.
22. Shalat SL, Seltzer B, Pidcock C, et al. Risk factors for
Alzheimer's disease: a case-control study. Neurology 1987;37:
1630-3.
23. Soininen H, Heinonen OP. Clinical and etiological aspects of
senile dementia. Eur Neurol 1982;21:401-10.
24. van Duijn CM, Tanja TA, Haaxma R, et al. Head trauma and
the risk of Alzheimer's disease. Am J Epidemiol 1992;
135:775-82.
25. 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—7.
26. 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. Int J Epidemiol 1991;20(suppl):
S28-35.
27. Paschalis C, Polychronopoulos P, Lekka NP, et al. The role of
head injury, surgical anaesthesia and family history as aetiological factors in dementia of Alzheimer type. Dementia
1990; 1:52-5.
28. Gedye A, Beattie BL, Tuokko H, et al. Severe head injury hastens age of onset of Alzheimer's disease. J Am Geriatr Soc
1989;37:970-3.
29. Sullivan P, Petitti D, Barbaccia J. Head trauma and age of
onset of dementia of the Alzheimer type. (Letter). JAMA
1987;257:2289-90.
30. Schofield PW, Tang M, Marder K, et al. Alzheimer's disease
after remote head injury: an incidence study. J Neurol
Neurosurg Psychiatry 1997;62:119-24.
31. Melton LJ III. History of the Rochester Epidemiology Project.
Mayo Clin Proc 1996;71:266-74.
32. Annegers JF, Grabow JD, Kurland LT, et al. The incidence,
causes, and secular trends of head trauma in Olmsted County,
Minnesota, 1935-1974. Neurology 1980;30:912-19.
33. Schoenberg BS, Kokmen E, Okazaki H. Alzheimer's disease
and other dementing illnesses in a defined United States population: incidence rates and clinical features. Ann Neurol
l987;22:724-9.
34. Kokmen E, Beard CM, O'Brien PC, et al. Is the incidence of
40
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
Nemetz et al.
dementing illness changing? A 25-year time trend study in
Rochester, Minnesota (1960-1984). Neurology 1993;43:
1887-92.
Kokmen E, Beard CM. Epidemiologic studies of dementia in
Rochester, Minnesota. In: Iqbal K, McLachlan DRC, Winblad
B, et al, eds. Alzheimer's disease: basic mechanisms, diagnosis and therapeutic strategies. New York, NY: John Wiley &
Sons Ltd, 1991:343-51.
Schoenberg BS. Calculating confidence intervals for rates and
ratios: simplified method utilizing tabular values based on the
Poisson distribution. Neuroepidemiology 1983;2:257-65.
Harrington DP, Fleming TR. A class of rank test procedures for
censored survival data. Biometrika 1982;69:553-66.
Rasmusson DX, Brandt J, Martin DB, et al. Head injury as a
risk factor in Alzheimer's disease. Brain Inj 1995;9:213-19.
Gabriel SE, O'Fallon WM, Kurland LT, et al. Risk of connective-tissue diseases and other disorders after breast implantation. N Engl J Med 1994;330:1697-702.
Special issue: is P-amyloid neurotoxic? Neurobiol Aging
1992; 13:535-625.
Price DL, Borchelt DR, Walker LC, et al. Toxicity of synthetic A (3 peptides and modeling of Alzheimer's disease.
Neurobiol Aging 1992;13:623-5.
Yankner BA. Amyloid and Alzheimer's disease—cause or
effect? Neurobiol Aging 1992; 10:471-2.
Selkoe DJ. Alzheimer's disease. In the beginning . . . . Nature
1991;354:432-3.
Clinton J, Ambler MW, Roberts GW. Post-traumatic
Alzheimer's disease: preponderance of a single plaque type.
Neuropathol Appl Neurobiol 1991;17:69-74.
Griffin WS, Sheng JG, Gentleman SM, et al. Microglial interleukin-1 alpha expression in human head injury: correlations
with neuronal and neuritic beta-amyloid precursor protein
expression. Neurosci Lett 1994;176:133—6.
McKenzie JE, Gentleman SM, Roberts GW, et al. Increased
numbers of p APP-immunoreactive neurones in the entorhinal
cortex after head injury. Neuroreport 1994;6:161^l-.
Roberts GW, Allsop D, Bruton C. The occult aftermath of boxing. J Neurol Neurosurg Psychiatry 1990;53:373-8.
Roberts GW, Gentleman SM, Lynch A, et al. P A4 amyloid
protein deposition in brain after head trauma. Lancet
1991;338:1422-3.
Roberts GW, Gentleman SM, Lynch A, et al. Beta amyloid
protein deposition in the brain after severe head injury: implications for the pathogenesis of Alzheimer's disease. J Neurol
Neurosurg Psychiatry 1994;57:419-25.
Weidemann A, Konig G, Bunke D, et al. Identification, biogenesis, and localization of precursors of Alzheimer's disease
A4 amyloid protein. Cell 1989;57:115-26.
Mahley RW. Apolipoprotein E: cholesterol transport protein
with expanding role in cell biology. Science 1988;240:
622-30.
52. Namba Y, Tomonaga M, Kawasaki H, et al. Apolipoprotein E
immunoreactivity in cerebral amyloid deposits and neurofibrillary tangles in Alzheimer's disease and kuru plaque amyloid
in Creutzfeldt-Jakob disease. Brain Res 1991;541:163-6.
53. Noguchi S, Murakami K, Yamada N. Apolipoprotein E genotype and Alzheimer's disease. (Letter). Lancet 1993;342:737.
54. Strittmatter WJ, Saunders AM, Schmechel D, et al.
Apolipoprotein E: high-avidity binding to p-amyloid and
increased frequency of type 4 allele in late-onset familial
Alzheimer disease. Proc Natl Acad Sci U S A 1993;90:
1977-81.
55. Wisniewski T, Frangione B. Apolipoprotein E: a pathological
chaperone protein in patients with cerebral and systemic amyloid. Neurosci Lett 1992; 135:235-8.
56. Roses AD, Pericak-Vance MA, Saunders AM, et al. Complex
genetic disease: can genetic strategies in Alzheimer's disease
and new genetic mechanisms be applied to epilepsy? Epilepsia
1994;35(suppl):S20-8.
57. Strittmatter WJ, Weisgraber KH, Huang DY, et al. Binding of
human apolipoprotein E to synthetic amyloid P peptide: isoform-specific effects and implications for late-onset Alzheimer
disease. Proc Natl Acad Sci U S A 1993,90:8098-102.
58. Saunders AM, Strittmatter WJ, Schmechel D, et al. Association
of apolipoprotein E allele e4 with late-onset familial and sporadic Alzheimer's disease. Neurology 1993;43:1467-72.
59. Corder EH, Saunders AM, Strittmatter WJ, et al. Gene dose of
apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 1993;261:921-3.
60. Mayeux R, Stern Y, Ottman R, et al. The apolipoprotein e4
allele in patients with Alzheimer's disease. Ann Neurol
1993;34:752^t.
61. Poirier J, Davignon J, Bouthillier D, et al. Apolipoprotein E
polymorphism
and
Alzheimer's
disease.
Lancet
1993;342:697-9.
62. Saunders AM, Schmader K, Breitner JC, et al. Apolipoprotein
E e4 allele distributions in late-onset Alzheimer's disease and
in other amyloid-forming diseases. Lancet 1993;342:710-11.
63. Mayeux R, Ottman R, Maestre G, et al. Synergistic effects of
traumatic head injury and apolipoprotein-e4 in patients with
Alzheimer's disease. Neurology 1995;45:555-7.
64. Nicoll JA, Roberts GW, Graham DI. Apolipoprotein E e4
allele is associated with deposition of amyloid p-protein following head injury. Nat Med 1995; 1:135-7.
65. Nicoll JA, Roberts GW, Graham DI. Amyloid beta-protein,
APOE genotype and head injury. Ann N Y Acad Sci
1996:271-5.
66. Jordan BD, Relkin NR, Ravdin LD, et al. Apolipoprotein E e4
associated with chronic traumatic brain injury in boxing.
JAMA 1997;278:136-40.
67. 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; 146:373-84.
Am J Epidemiol Vol. 149, No. 1, 1999