Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
10 OTHER HEALTH OUTCOMES Traumatic brain injury (TBI) is an important cause of death worldwide and contributes to a considerable number of deaths and disability in the United States. The Centers for Disease Control and Protection has estimated that about 50,000 of the 1.4 million people who sustain a TBI in the United States will die as a direct consequence (NCIP, 2008). This section examines mortality and premature death in people who have a TBI. MORTALITY AND TRAUMATIC BRAIN INJURY PRIMARY STUDIES Mortality and Traumatic Brain Injury in Military Populations That head injuries reduce life expectancy has been posited since World War I. Medical records of Bavarian World War I veterans who were patients at a head-injury center were analyzed in 1964â1966 (Walker et al., 1971) (see Chapter 5 for a description of the military cohorts). The head-injury center was established in 1916, and many of the 5,500 men who sustained head injuries in World War I were followed for up to 50 years (Credner, 1930). About 1,000 records were randomly selected, and death certificates were sought from social-welfare offices in Bavaria and West Germany. Vital statistics were obtained for 647 cases and matched with those of 616 uninjured Bavarian World War I decorated veterans, who were able to be traced because they received pensions. Veterans were excluded if they were born before 1880 or died before the age of 35 years, if death dates were unknown, or if records were incomplete. Both penetrating TBI and nonpenetrating TBI were identified; posttraumatic epilepsy was also assessed. Compared with the general population of German men who were at least 35 years old in 1925, those with head injuries had 1.8% more deaths than expected. In 1965, 73% of the veterans with TBI had survived from the age of 35 years to the age of 65 years compared with 80% of those without TBI. Life expectancy of veterans with TBI was about 4 years shorter than that of those without TBI. Posttraumatic epilepsy had the greatest effect on life expectancy after the age of 50 years. Weiss et al. (1982) used data from the same cohort through 1972 to assess life expectancy and correlate it with the severity of injury. By 1972, 77% of TBI veterans (497 of 647) and 78% of control veterans (483 of 616) had died. When TBI was categorized as shallow (nonpenetrating or penetrating less 3 cm; 314 veterans) or deep (penetration greater than 3 cm, perforating wounds, or wounds of ventricle or brainstem; 283 veterans), those with deep wounds 333
334 GULF WAR AND HEALTH had slighter higher mortality between the ages of 45 and 60 years, but the difference was not significant (p = 0.6). Similar results were seen when cases were divided by severity of coma (p > 0.6). Confirming the results of Walker et al., Weiss and colleagues (1982) found that mortality was nonsignificantly (p = 0.08) higher in older TBI veterans with posttraumatic epilepsy than in older TBI veterans without posttraumatic epilepsy; moreover, posttraumatic epilepsy had a marked effect in reducing life expectancy (p = 0.01) in veterans with TBI compared with control veterans. Men with TBI were more likely to die of cerebrovascular disease than the controls (19% vs 12%; p = 0.01), particularly men younger than 60 years old (p = 0.015), but that difference did not correlate with the severity of the TBI. Corkin et al. (1984) continued work begun by Teuber and colleagues in 1948 on a cohort of 190 World War II veterans with penetrating TBI. Corkin et al. (1984) compared those veterans with a control group of 106 veterans who had peripheral nerve injury, matched for age at injury, years of formal education, and preinjury intelligence-test scores. As of 1983, 28.4% of the veterans with penetrating TBI and 17.0% of those with peripheral nerve injury had diedâa statistically significant difference (p = 0.03). However, when veterans with head injuries were categorized by whether they had posttraumatic epilepsy, only those with epilepsy had significantly higher mortality than the controls (p = 0.0002). Such factors as the site of the brain lesion, age at injury, and preinjury and postinjury intelligence scores did not affect survival, although veterans with more education lived longer. Rish and colleagues (1983) followed 1,127 male Vietnam veterans who had penetrating head injuries for 15 years. Over the 15-year period, 90 deaths (8%) had occurred: 46 in the first year after injury, 9 during the second year, and then 1â4 per year. Most deaths occurred early in the first year after trauma and were the result of the injury or coma sequelae. After 3 years, compared mortality in the head-injured Vietnam veteran population approached the norm, according to actuarial projections for North American men 21â35 years old. Length of coma was the best predictor of long-term outcome, and posttraumatic epilepsy was not a significant factor in mortality in the first 15 years, although each seizure event carried its own inherent risk. Mortality and Traumatic Brain Injury in Civilian Populations Mortality in TBI patients can be studied from the time of injury, from the time of discharge from inpatient acute-care hospitals, or from the time of admission into or discharge from inpatient rehabilitation. Rates in cohorts of patients at those different points of entry into a study will be different. For example, rates in patients from the time of injury will be greatest because they include early deaths. In contrast, survivors of the acute phase who are studied during rehabilitation are likely to have lower mortality. Mortality in Patients from Time of Injury or Admission into Acute-Care Hospitals Brown et al. (2004) carried out a study to determine whether mortality from TBI was affected by the severity of the injury. Their population-based retrospective cohort study identified all Olmsted County, Minnesota, residents who had a diagnosis indicative of a potential TBI, and they reviewed the medical records from the Rochester Epidemiology Project for 1985â 1999 (see Chapter 5). The review confirmed 1,448 cases of TBIâ164 (11%) moderate to severe and 1,284 (80%) mild. There were 68 deaths in the moderate-to-severe TBI group. The Kaplanâ
OTHER HEALTH OUTCOMES 335 Meier estimated 30-day case-fatality rate was 29.3% (95% CI, 22.0â35.9). The 68 deaths, compared with mortality in the 1990 Minnesota white population, were significantly more than the 12.8 expected (relative risk [RR], 5.29; 95% confidence interval [CI], 4.11â6.17); but the 14 deaths in those with moderate or severe TBI who survived for the first 6 months after injury did not differ significantly from the 12.7 expected (RR, 1.10; 95% CI, 0.60â1.85; p = 0.72). There were 78 deaths in the mild-TBI group, and the KaplanâMeier estimated 30-day case-fatality rate was 0.2% (95% CI, 0.0â0.4). Over the full followup period, the 78 deaths were significantly greater than the 58.8 expected (RR, 1.33; 95% CI, 1.05â1.65; p = 0.012); but the 69 deaths in the people with mild cases who survived 6 months after injury were not significantly different from the 58.6 expected (RR, 1.18; 95% CI, 0.92â1.49; p = 0.173). Mortality in Patients Discharged from Acute-Care Hospitals Selassie et al. (2005) studied a representative sample of 3,679 TBI patients within a year of their discharge from any of 62 acute-care hospitals in South Carolina in 1999â2000 to document mortality within 15 months of discharge. Patients were stratified on the basis of TBI severity and hospital size. Of the sample, 3,371 (91.6%) were alive, and 308 (8.4%) had died within about 15 months of their discharge. Deaths were confirmed by using the Social Security Death Index (SSDI) and were classified as TBI-related or not. Of the 308 deaths, 17% were classified as TBI-related, and 63% of them occurred within the first 3 months after discharge, compared with 47% of the non-injury-related deaths. Findings indicate that the older the person, the higher the likelihood of early death. Males were more likely than females to die after TBI hospitalization if they were younger than 35 years old or older than 54 years old, but mortality was higher in females in the age range of 35â54 years. The authors also report that the severity of the TBI influenced mortality, as did the place of treatment. Patients who were treated in hospitals with trauma centers were less likely to die within 15 months after hospital discharge than patients treated in hospitals without trauma centers. Mortality in Patients Admitted into or Discharged from Rehabilitation Centers In a retrospective cohort study, Baguley et al. (2000) assessed mortality in 476 people who had sustained severe TBI over a 10-year period compared with an age- and sex-matched sample of the general Australian population. Patients with TBI, resulting primarily from closed head trauma due to motor-vehicle collisions, were admitted into a rehabilitation hospital 1986â 1996. Twenty-seven of the 476 patients with TBI had died by August 1997, for a mortality rate of 5.7%. The median interval between injury and death was 17 months (range, 45 daysâ108 months after injury). Mortality in the TBI group was significantly associated with a lower level of functional independence on the basis of the Functional Assessment Measure (p < 0.001) slightly significantly with male sex (p < 0.078), with greater age (38 vs 32 years of age; p < 0.055), and with a premorbid psychiatric history (p < 0.064), but not significantly with a history of premorbid substance abuse (p < 0.308). In the general-population sample, the mortality rate was 1.5%. A Fisherâs exact test indicated that significantly more people (p < 0.001) with TBI died (compared with the general population), with most deaths occurring in the first 12 months after injury. The leading causes of deaths were cardiorespiratory events (8 of 27) and infection (6 of 27). The study is limited by the short followup period of about 5 years, the use of data from only one state in Australia, and the lack of functional assessment of and preinjury data on 52% of the deceased and 22% of the living TBI patients.
336 GULF WAR AND HEALTH Harrison-Felix et al. (2004) studied 2,178 patients with moderate to severe TBI identified in the Traumatic Brain Injury Model Systems (TBIMS) national database, which covers 15 TBIMS rehabilitation centers. Most of the TBIs resulted from motor-vehicle crashes (62%) or acts of violence (20%). The study participants had a mean age of 37 years, 76% were men, 60% were white, and 37% had severe TBI on the basis of 24-hour postinjury Glasgow Coma Scale (GCS) scores. Using vital-status information obtained from the SSDI the authors identified 161 deaths in 2,178 people with TBI who were followed for 17 daysâ12.8 years after injury for a mortality rate of 7.4%. Compared with age-, sex-, and race-specific mortality rates for the general population, the authors found that individuals with TBI were at twice greater risk of death than those in the general population (95% CI, 1.69â2.31). The median interval between injury and death was 2 years; 38 of the 161 deaths occurred between rehabilitation discharge and 1 year after injury. Life expectancy was reduced by 5â9 years (average, 7 years), depending on age at injury, race, and sex. Long-term survival of severely injured TBI patients was also studied by Ratcliff and colleagues (2005), who reviewed medical records of 640 TBI patients discharged from a rehabilitation hospital in Pennsylvania during 1974â1984 and during 1988â1989. Most of the injuries resulted from motor-vehicle crashes (66.6%), and falls were next most common cause (16.4%). There were 464 males and 176 females in the TBI population, and their mean age was 37 years. Vital status as of 1997 was determined by using the SSDI. There were 126 deaths, for a standardized mortality ratio (SMR), relative to the expected mortality in the Pennsylvania population, of 2.78 (p = 0.0001). An increased risk of early death was associated with a known history of alcohol abuse (SMR, 6.10; p = 0.0001), substance abuse (SMR, 8.00; p = 0.0264), and other personal or social problems (SMR, 7.03; p = 0.0001). A criminal record or history of psychiatric disorder and years of followup did not modify mortality. The studies by Brown et al. (2004) and Ratcliff et al. (2005) indicate that functional status at time of discharge from rehabilitation is a leading indicator of later survival time, although severity of injury and other demographic variables, such as age at time of injury, are not. Shavelle and Strauss (2000) published two mortality studies of people with TBI who received services from the California Department of Developmental Services and survived at least 12 months after injury. The department serves the severely disabled, including those with developmental disabilities and those with long-term cognitive deficits. In nonambulatory patients, observed mortality exceeded expected mortality by a factor of 16.6 in 15- to 29-year- olds and 7.3 in 30- to 44-year-olds. In patients who were more ambulatory, the SMR was 2.5 for 15- to 29-year-olds who could walk with support or could walk unsteadily for at least 10 feet and 2.7 for those who could walk well alone for at least 20 feet. SECONDARY STUDIES The committee identified 12 secondary studies, as discussed below. Several studies used cohort designs to examine mortality in people who have TBI. The people studied have been only patients with TBI; that is, there have been no external comparison groups. The best of the studies have followed patients from the time of TBI. Others have examined subsets of the universe of patients with TBI, such as those admitted into rehabilitation hospitals, in which case the more selective nature of the population (survivors of the acute period who enter rehabilitation) may
OTHER HEALTH OUTCOMES 337 lead to death rates different from that in the whole universe of TBI patients who survive the acute period. Six population-based studies were identified; most found consistent results regarding TBI and increased mortality (Harris et al., 2003; Hukkelhoven et al., 2003; Engberg and Teasdale, 2004; Lu et al., 2005; Flaada et al., 2007; Winqvist et al., 2007;). Using data from the Rochester Epidemiology Project (see Chapter 5), Flaada and colleagues (2007) compared the observed number of deaths after TBI with the expected number of deaths. Mortality within 6 months increased with age in both moderate and severe TBI cases. Mortality in moderate and severe cases was 40 times that in mild cases independently of age. Among those surviving at least 6 months, 10-year mortality differed from expected only in the adult cases. Lu et al. (2005) studied mortality associated with TBI in 1984â1996 to determine whether mortality had been decreasing. The study population consisted of 1,839 severely head- injured patients, whose data were extracted retrospectively from the Traumatic Brain Injury Data Bank (635 patients), the Medical College of Virginia (382), and clinical-trial databases in the United States (822). People with penetrating head injuries and treatment groups in the clinical- trial databases were excluded. Of the 1,839 patients with severe TBI, 526 died, for a mortality rate of 28.6%. Over the period 1984â1996, mortality from severe TBI fell from 39% to 27%. After adjustment for a variety of factorsâincluding age, admission motor score, and pupillary responseâthe difference remained significant (p < 0.05). Engberg and Teasdale (2004) conducted a population-based study of 389 patients identified in the national hospital register in Denmark who had cranial fractures or traumatic cerebral lesions. The mortality rate was assessed 15 years or more after injury. The acute and subacute mortality rate was 27% in those who sustained cerebral lesions and 4% in those who sustained cranial fracture. Harris et al. (2003) assessed mortality after head injury in 13,908 people identified through the New York State Trauma Registry from January 1, 1994, to December 31, 1995. The overall mortality rate in all age groups was 14.7% (range, 5.6â30.2%). The authors found an increased mortality rate with increased age: 10.9% at ages 0â30 years, 12.4% at ages 31â50 years, and 21.3% at age above 50 years. Winqvist et al. (2007) assessed mortality related to TBI by using the Northern Finland Birth Cohorts. One cohort is made up of people born in 1966 in two provinces in Finland and followed over a 34-year period. The authors identified 457 subjects who sustained their first ever TBI during 1966â2000, 78.1% of whom had sustained mild TBIs, including concussions and skull fractures. Nearly 10% had sustained moderate to severe TBIs, including brain contusions, intracranial hematomas, and diffuse traumatic axonal injuries. The authors found that the annual mortality in people with TBI was 14 per 100,000. Mortality related to TBI accounted for 12% of the total mortality in this group. Hukkelhoven et al. (2003) analyzed individual patient data from four prospective studies (three multicenter randomized clinical trials and one prospective series with closed TBI) to assess mortality and outcome related to TBI. The mean mortality rate ranged from 23% to 40% in the four studies; the mean proportion with an unfavorable outcome ranged from 43% to 60%.
338 GULF WAR AND HEALTH The authors found that mortality increased with age from 21% in those 35 years old and younger to 72% in those 65 years old and older. Five secondary studies assessed mortality related to TBI in hospitalized patients (Miller and Jennett, 1968; Fearnside et al., 1993; Lai et al., 1998; Gomez et al., 2000; Pentland et al., 2005). Many of the study populations consisted of consecutive series of patients admitted to hospitals with TBI. The studies found consistent results regarding TBI and increased mortality. Fearnside et al. (1993) conducted a prospective study of 315 consecutive patients with severe TBI admitted to Westmead Hospital in Sydney, Australia. Severe TBI was defined as a GCS of no more than 8 within 6 hours after injury or deterioration to this level within 48 hours after injury. The authors found that GCS was inversely related to mortality. The correlation coefficient between mortality and GCS was 0.418 (t = 8.14; p < 0.005). Gomez and colleagues (2000) studied outcomes in a cohort of 810 patients with severe closed head injuries who were consecutively admitted to a hospital in Spain in 1987â1996. Severe head injury was defined as a GCS at admission of no more than 8 or deterioration to this level 48 hours after injury. At 6 months after injury, the overall mortality rate was 50.3%. In general, older patients had worse outcomes: the mortality rate reached 77% in the subgroup of 142 patients over 55 years old, but it was 37.5% in patients 46â55 years old. Miller and Jennett (1968) evaluated mortality in a consecutive series of 400 patients with depressed fracture of the skull who were treated in the neurosurgical division of the Institute of Neurological Sciences in Glasgow during 1956â1967. The authors found that the cases were associated with a significant increase in mortality (p < 0.005). Lai et al. (1998) conducted a retrospective study to evaluate long-term outcomes related to severe TBI in 70 patients admitted to a surgical intensive-care unit over a 29-month period. At 1 year after injury, the overall mortality was 50%. Pentland (Pentland et al., 2005) assessed mortality in a cohort of 1,871 patients with mild, moderate, and severe TBI admitted to a regional head-injury unit in Scotland between 1981 and mid-2002. Of the 1,871, 93 had severe TBI, 205 moderate, and 1,573 mild. Fifty-seven patients died during the initial admission into the unit (42 with severe TBI, 8 with moderate, and 7 with mild). During subsequent years, 340 patients died (six had severe TBI, 33 moderate, and 301 mild). Studies in the Sports-Injury Literature To study the influence of intense physical training on life expectancy, Bianco et al. (2007) examined male athletes born in 1860â1930 who had been inducted into various halls of fame: baseball (154), ice hockey (130), tennis (83), football (81), boxing (81), track and field (59), basketball (58), swimming (37), and wrestling (32). Because boxing is characterized by repetitive blows to the head, numbers of bouts and rounds were scrutinized. Median life expectancy of all the samples was 76.0 years; boxers had the lowest median life expectancy (73.0 years), and no differences were found by number of rounds or bouts.
OTHER HEALTH OUTCOMES 339 SUMMARY AND CONCLUSIONS There is clear evidence of increased mortality in the acute phase after TBI and for some time afterward in both military and civilian populations with moderate to severe TBI. In the military literature, posttraumatic epilepsy in patients who initially survive penetrating head injuries is associated with an increased risk of death and about a 5-year decrease in life expectancy. In the civilian literature, studies in Minnesota suggest that although there is clear evidence of increased mortality in the first 6 months after injury, there is no evidence of increased mortality in patients with TBI beyond 6 months, regardless of severity. Studies of the subset of more severely injured patients who survive initial hospitalization and require inpatient rehabilitation demonstrate a worse prognosis, consistent with the greater degree of residual compromise: mortality some 2â7 times higher than in age- and sex-matched comparison populations. The committee has reviewed 10 primary studies and 12 secondary studies of TBI and mortality and has found consistent results. The committee concludes, on the basis of its evaluation, that there is sufficient evidence of a causal relationship between penetrating TBI and premature mortality in survivors of the acute injury. The committee concludes, on the basis of its evaluation, that there is sufficient evidence of an association between moderate or severe TBI and premature mortality in the subset of patients who are admitted into or discharged from rehabilitation centers or receive disability services. The committee concludes, on the basis of its evaluation, that there is inadequate/insufficient evidence to determine whether an association exists between surviving 6 months or more after sustaining a mild, moderate, or severe TBI and premature mortality.
340 TABLE 10.1 TBI and Mortality Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations Baguley et Cohort Patients with TBI Severe; 97% Mortality by 476 patients, mean None Missing FAM, al., 2000 composed of admitted to Brain closed, August 1997 duration of followup preinjury clinical case Injury 3% penetrating (mean, 5 years 64 mo; 97% closed information on series Rehabilitation after trauma; head injury, 3% substance abuse, Service, Westmead range, 8 moâ11 penetrating head psychiatric Hospital, New years after injury; 62% MVC, history from South Wales, trauma); 21% falls or hit by patients admitted Australia, 1986â ascertained by object, 12% assault, before 1990 on 1996; cases had New South Wales 4% sports-related 52% of the survived through vital-statistics deceased, 22% admission into search 27 of 476 (5.7%; 95% of the living; no rehabilitation CI, 0.037â0.083) dead multivariate facility; (median, 17 mo after analysis comparison group: trauma; range, 45 dayâ expected mortality 9 years 2 mo after in age- and sex- trauma); expected matched Australian mortality rate, 1.5% population in 1997 (CI, 0.006â0.03) (p<0.001 by Fisherâs exact test) Contributing factors: low FAM on discharge (p < 0.001), being male (p = 0.078), greater age (p = 0.055), prior psychiatric morbidity (p = 0.064), but not prior substance abuse (p = 0.308) by x2 or t test Cause of death: cardiorespiratory arrest
Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations (30%), infection (22%) Brown et Population- Any Olmsted Documented Vital status Age 35.3 years for Age, sex for Unique database al., 2004 based County, MN, concussion with through 2002 moderateâsevere, 26.8 mortality on medical care retrospective resident with LOC; PTA; from medical years for mild; mean analysis; of countyâs cohort from medically attended neurologic records, state followup, 7.4 years age, sex, year entire Rochester TBI, 1985â1999 (N signs of brain death tapes of TBI with population; Epidemiology = 45,831); random injury and/or Mortality in moderateâ Cox cohort not Project 15.7% sample of intracerebral, severe: proportional- generalizable TBI patients (N = subdural, or 68 deaths in 164 cases; hazards beyond Olmsted 7,175) reviewed; epidural overall risk of death model for Countyâfew 1,448 met inclusion hematoma; increased compared comparison minority-group criteria; cerebral with expected, RR, of moderateâ members (96% comparison group: hemorrhage or 5.29 (95% CI, 4.11â severe vs white), all care age- and sex- contusion; brain 6.71) by long-rank mild in only 2 specific 1990 white stem injury; statistic; institutions Minnesotans penetrating 30-day CFR, 29.3% by head injury; KaplanâMeier, risk skull fracture; increased compared or with expected, RR, postconcussive 5.29 (95% CI, 4.11â syndrome 6.71); 14 deaths in those Moderate or surviving 6 mo, no severe (11%): increase in risk, RR skull fracture, 1.10 (95% CI, 0.60â intracranial 1.85) hematoma, brain contusion, Mortality in mild: penetrating 78 deaths in 1,284 head injury, cases; overall risk of brain stem death increased injury, or compared with severe expected, RR, 1.33 complications (95% CI, 1.05â1.65); 9 (neurosurgery, deaths in first 6 mo CNS infection, (CFR, 0.2%), no 341
342 Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations subarachnoid difference from hemorrhage, expected hydrocephaly, 69 deaths in those CSF leak) surviving 6 mo; risk of death not increased Mild (89%): compared with LOC, PTA, expected, RR, 1.18 postconcussive (95% CI, 0.92â1.49) symptoms, focal Comparison of neurologic moderateâsevere with signs mild: risk of death increased in first 6 mo, RH, 5.18 (95% CI, 3.65â7.30) by Cox proportional-hazards model; no difference 6 mo, RH, 1.04 (95% CI, 0.57â1.88) Corkin et Prospective All WWII veterans Penetrating, at Mortality to Mortality: 54 of 190 Cox Vital status al., 1984 cohort (World with penetrating least 3 years 5/1/1983 as (28.4%) with proportional could not be War II head injury from after trauma function of penetrating head injury regression determined on veterans Teuber series (n = various factors dead vs 18 of 106 adjusted for only one subject assembled at 190); excludes few (17.0%), significant age at injury, (treated as alive); NYU by with difference by Kaplanâ years of no cause-of- Teuber in nonpenetrating Meier (p = 0.03); those education, death data 1948) head injury; 106 with PT epilepsy (n = difference in collected WWII controls 82) more likely to be AGCT with peripheral dead than those (preinjury vs nerve injuries without (n = 91) or 10 years after matched for age, controls (p = 0.0002); trauma) education, PT epilepsy (p = preinjury AGCT 0.003), lower (85% of controls in education (p = 0.02) Teuber series with associated with death such injuries) by Cox
Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations Harrison- Retrospective 2,178 TBI patients Age 37.4 years, Mortality from Mortality: 123 of Age, sex, Maximum Felix et al., cohort from 15 16 years old 76% male, 60% SSA Death Index 2,140; median, 2 race in followup only 13 2004 TBIMS completing white through 2001 years; overall, SMR, determining years, average centers inpatient 2.00 (95% CI, 1.69â SMRs from 3.1 years from 1 rehabilitation in Cause of injury: 2.31); <1 year after federal year after 1988â2000; sample MVC, 62%; trauma, 38 deaths; 1 statistics for trauma; 38% loss is 2,140 who violence, 20%; year after trauma, 123 2000; to followup; two survived >1 year falls, 16%; deaths, SMR, 1.95 Cox of 17 centers did after trauma; other, 2% (95% CI, 1.61â2.29); proportional not participate, comparison group: life expectancy, hazards for so sample less US age- and sex- Severity: 37% average reduction, 7 those representative specific mortality severe (24-h years, depending on surviving >1 in 1994 max GCS 8) age at injury, sex, race, year with range 5â9 years ALOS: 21 days acute care, 30 Risk factors: higher days acute age, unemployment at rehabilitation time of injury, higher DRS score at discharge Lewin et Retrospective 7,000 consecutive Severe in large Vital status; for Overall mortality, 178 Age, Only 2% loss to al., 1979 cohort head injured part closed, but 178 (consecutive of 469 (38%) maximum followup; patients admitted complicated by series), 28 central neural developed model (same into John Radcliffe compression or (selected series) Life expectancy for disability for predicting population Infirmary, Oxford, penetration who died, cause four neurophysical- score, long-term studied in 10â24 years earlier (traumatic or of death; for 331 disability patternsâ maximum outcome on basis book by AH (1955â1969); of surgical for survivors, âdecerebrate mental of age at injury, Robert, 1979, these, 479 amnesic internal neurologic dementiaâ: most <1 disability worst category with same or unconscious >1 decompression) examination (all), year, one >10 years; score, of mental and results, but week; additional for 77 and 14, test of cognitive âathetoid duration of neurophysical also mentioned selected series: 64 respectively, of function (217) pseudobulbarâ: PTA for disability, length suicide as cases unconscious 331 survivors reduced only by model of PT amnesia in cause of >1 mo admitted to epilepsy, drowning, selected series increased this or other facility inhalation of food, deaths) 3â25 year earlier suicide; 343
344 Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations (including 24 from âbrain-stem (Cause-of- first set); causes of cerebellarâ or âminor death death in 78 patients hemipareticâ: comparison) discharged from reduction of <5 years initial hospitalization Cause of death among alive were those discharged compared with compared with general causes of death in public: general population meningitis, p < 0.001; of England, Wales epilepsy, p < 0.001; in 1960 (not age- drowning, p < 0.001; or sex-adjusted) respiratory, p < 0.005 Neurologic outcomes at 10 years (consecutive series): 11 (4%) totally disabled; 66 (14%) severely disabled, precluding normal social, occupational life; 214 (46%) recovered; 178 (38%) dead Hospitalization: continuing need discussed but not quantified Ratcliff et Retrospective 640 patients 14 Head injury Mortality through Overall mortality: 128 Age at injury, Subjects outside al., 2005 cohort years old with identified by 1997 (19.7%) deaths; SMR, sex, range of interest moderate to severe ICD-8 and -9 2.78; p < 0.0001 by education, for age at time of TBIs discharged 8â codes 800â801- Poisson regression marital status, injury: 24 years after 9, 803â804.9, race, cause of <18 years, 19%; trauma from 850â854.9, Any preinjury social or injury, 60 years, 13%
Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations Pittsburgh, PA, excluding behavioral problem: severity of rehabilitation comorbid spine SMR, 5.82, p < 0.0001 injury Followup, 8â24 center, 1974â1984, injury year after 1988, 1989; Alcohol abuse: SMR, trauma; excluded comparison group Cause of injury: 6.10; p < 0.0001 1985â1987 to from Pennsylvania MVC, 66%; keep sample size vital-statistics violence, 2%; Substance abuse: smaller, tables falls, 16%; SMR, 8.00; p < 0.0001 manageable; other, 15% 6.5% could not Other personal or be traced Moderate to social problems: SMR, (assumed alive); severe cases 7.03; p < 0.0001 univariate (range, 4â54) analysis of retained, Functional limitations numerous severity based at discharge (seven variables, but on ICD at items with three final multivariate discharge as levels): model contained converted into Bathing, p = 0.01; only preinjury ISS with range grooming, p = 0.002; behavioral 0â75 dressing, p = 0.011; problems, eating, p = 0.003; bed- grooming or to-chair, p = 0.035; eating problems; toilet use, p = 0.017; importance of walking across room, preinjury factors p = 0.019; summation suggests that a partitioned into four property of levels, p = 0.008 people experiencing Years after discharge, TBI, rather than severity of injury not TBI itself, may significant; final increase stepwise regression mortality model if no preinjury behavioral problem or functional limitation at discharge, SMR, 1.69 345
346 Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations Rish et al., Prospective 1,127 male Penetrating Mortality 15 years Overall mortality: 90 Age and sex Exclusively 1983 cohort Vietnam veterans cerebrocranial after trauma of 1127 (8%), 46 in penetrating (registry alive 1 week after wounds first year after trauma, injuries, whose established trauma; 32 in first 3 mo, 16 in consequences 1976â1980 by comparison group: first month; compared may differ from MFUA, WF age- and sex- with North American those of Caveness) matched from males, mortality concussive North American increased up to 13 injuries actuarial data years after trauma (American Council (primarily 1â2 years of Life Insurance) after trauma), near actuarial rates at 14â15 years after trauma Cause of death: after second year, same as general population plus continuing losses due to coma sequelae, seizures and brain abscesses; coma (initial level of consciousness and duration) most predictive of mortality, not PT seizures 26 of 1,050 (2.6%) deaths among those who were discharged to self-care vs 67 of 80 (84%) of those who required continued hospitalization Selassie et 3,679 patients 15 AIS scores of Mortality <15 mo Mortality <15 mo of Age, sex, Some subjects
Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations al., 2005 years old with TBIs severity after discharge discharge: 308 deaths; race for <18, >60 years discharged from 62 converted from from acute care median, 93 days; SMRs based old acute-care ICD-9-CM range, 1â453 days; on US nonfederal codes at survival curves differ population; Focus on only 15 hospitals in South discharge; mild, by severity, p < 0.0001 Cox mo after Carolina in 1999â AIS 2); proportional- discharge 2001, with moderate, AIS Overall SMR, 7.1 hazards selection from 3; severe, AIS (95% CI, 6.3â7.9); model 1,544 (42%) 6,583 eligible 4 cancer (n = 31), SMR, refused or not stratified on 3.1 (95% CI, 2.1â4.2); located severity, hospital heart disease (n = 50), size; comparison SMR, 3.7 (95% CI, Death group: rates, causes 2.8â4.8); unintentional certificates of death in US injury (n = 61), SMR, obtained for 94% population in 2000 36.3 (95% CI, 27.8â of known deaths 46.0); cerebrovascular 74% of injury- disease (n = 18), SMR, related deaths 11.7 (95% CI, 8.2â related to 15.9) original TBI Risk of death Did not find associated with age, excess deaths number of associated with comorbidities, AIS seizures, 4, Medicare, care in respiratory nontrauma center infections, choking and suffocation, suicide Shavelle et Retrospective 2,629 people with TBI by ICD-9 Mortality as Mortality ratio: Stratified by Patients with al., 2000 cohort TBI >15 years old, codes 800â804, recorded in state overall, 277%; ambulation severe in 1988â1997, 850â854 vital statistics nonambulatory status disabilities only, receiving services patients, 660%; not analogous to from California partially ambulatory, incident cohort Department of 196%; ambulatory, Developmental 180% 347
348 Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations Services (implying severe disability) and survived 1 year; comparator: 1990s US life tables by sex Walker et Cohort 1,000 Bavarian Mixed severity, Mortality to 1965 5-year bands of age- PT epilepsy; 50 years of al., 1971 head-injured men type by life-table specific life bracketing followup; from WWI (nonpenetrating analysis; expectancies estimates statistics rather randomly selected slightly >50%) epilepsy at âsome calculated for >35 derived by primitive; biases from among 5,500 time after injuryâ years; 73% of cases, assuming that likely in cases in head- (first event for 80% of controls alive those with selection of injury center in most within year at age 65 years; across unknown study population 1916â1927 with of injury, but all age bands, life vital status (for example, âsufficient persisted for expectancy was were all alive representativenes information for most); broad increasingly lower for or were s of cases at analysisâ of nature classifications of control veterans, head random center of all head of injury; 1,000 cause of death injured without sample of injured and of unwounded epilepsy, head injured population those with Bavarian WWI with epilepsy in sufficient veterans on comparison with information of pension lists for general population; all cases; receiving medal; aside from sequelae of controls all all born 1880â injuries, no cause of received 1900; final, 555 death stood out for medals); vital cases, 563 controls head injured status of 400 of 1,000 not attainable, but same number found for controls; for both groups, date, cause of death found for 56%, but vital status unknown for
Health Outcomes or Outcome Comments or Reference Study Design Population Type of TBI Measures Results Adjustments Limitations about one-fourth Weiss et Cohort 1,010 head-injured Mortality to 1972 Mortality: overall, 497 None If this sample al., 1982 Bavarian men from by life-table of 647 with TBI vs was first defined WWI; analysis; cause of 483 of 616 controls; in or before 1,000 unwounded death ages 35â70 years, 1930, it has Bavarian WWI brain-injured vs proved to be veterans; final, 647 control, effectively a cases, 616 controls no difference; wound prospective 3 cm vs 0â3 cm, ns cohort with 60 increase; years of coma 1 day vs <1day, followup no difference; PT seizures vs no, increase maximal at ages 50â65 years; PT seizures vs control, increase (p = 0.01) Cause of death: TBI vs control, cerebrovascular (p = 0.01), <60 years (p = 0.015), 60 years (p = 0.04), not related to three measures of severity; cardiovascularârenal, no difference NOTE: AGCT = Army General Classification Test, AIS = Abbreviated Injury Scale, ALOS = average length of stay, CFR = case-fatality rate, CI = confidence interval, CNS = central nervous system, CSF = cerebrospinal fluid, CT = computed tomography, FAM = Functional Assessment Measure, GCS = Glasgow Coma Scale, GOS = Glasgow Outcome Score, ICD = International Classification of Diseases, LOC = loss of consciousness, MFUA = Medical Follow-Up Agency, MVC = motor-vehicle crash, NYU = New York University, PT = posttrauma, PTA = posttraumatic amnesia, RH = relative hazards, RR = relative risk, SMR = standardized mortality ratio, SSA = Social Security Administration, TBI = traumatic brain injury. WWI = World War I. 349
350 GULF WAR AND HEALTH BRAIN TUMORS AND TRAUMATIC BRAIN INJURY Brain tumors are growths of abnormal cells in the tissues of the brain and can be benign (noncancerous) or malignant (cancerous). The estimated number of new cases of brain and other nervous system tumors in the United States in 2008 is 21,810, and an estimated 13,070 related deaths are expected (NCI, 2008). Gliomas, which are primary brain tumors, form in the glial cells of the brain or spinal cord and can spread throughout the nervous system. They can be benign or malignant. There are several types of gliomas, such as astrocytomas, ependymomas, and oligodendrogliomas. The committee examined several studies of brain tumors and TBI, and they are discussed in this section. PRIMARY STUDIES The committee reviewed 14 primary studies that examined the association between TBI and brain tumors. Most relied on self-reports of TBI to assess exposure. Three of the 14 studiesâin Olmsted County, Minnesota (Annegers et al., 1979), Denmark (Inskip et al., 1998), and Sweden (Nygren et al., 2001)âused medical records to ascertain TBI and therefore are substantially less prone to recall bias. In a retrospective cohort study, Annegers et al. (1979) followed 2,953 patients in Olmsted County, Minnesota, who had a diagnosis of TBI during 1935â1974 to determine the occurrence of later brain tumors. Patients were selected from among 3,587 head-injured people in the population of Olmsted County who had survived the initial injury and had no history of pre- existing brain tumor. TBI was defined as brain involvement manifested by loss of consciousness (LOC), amnesia, or skull facture. Four brain tumors were observed in those with TBI, compared with an expected 4.13 (RR, 1.0; 95% CI, 0.3â2.6). There were one astrocytoma (SMR, 0.7) and three meningiomas (RR, 1.6; 95% CI, 0.3â4.7). The tumors were not associated with TBI severity and were diagnosed 5â16 years after the injury. In a study of the incidence of brain tumors after TBI in Denmark (Inskip et al., 1998), nationwide registry of hospital discharges in 1977â1992 was used to identify 228,005 people who had been hospitalized for TBI on the basis of ICD-8 codes for fractured skull, concussion, or cerebral laceration or contusion. This registry was linked with the Danish Death Certificate File and the Danish Cancer Registry for 1977â1993 to determine the incidence of cancer. The incidence was compared with that in the Danish population to obtain standardized incidence ratios (SIRs), which were adjusted for sex, age, and calendar year. Concussion and fractured skull were the most common injuries and typically resulted from traffic accidents, falls, and sports-related activities. There were 299 tumors of the brain and nervous system in Denmark during the 7-year period, of which 261 were intracranial tumors, including 113 gliomas, 36 meningiomas, 12 neurilemmomas, 8 medulloblastomas, and 16 vascular tumors. The SIR for any intracranial tumor associated with TBI was 1.36 (95% CI, 1.20â1.53), and most tumors occurred in the first year after the injury (SIR, 3.38; 95% CI, 2.59â4.34), regardless of tumor type. The SIR dropped closer to 1.0 in later years (SIR, 1.15; 95% CI, 0.99â1.32). Sixty-two intracranial tumors were diagnosed in the first year of followup; 43 of these were diagnosed during the first 6
OTHER HEALTH OUTCOMES 351 months. The authors suggest that it is unlikely that the tumors grew quickly enough to be caused by the injuries and were more likely to have been undiagnosed tumors that were present when the injuries occurred and were detected in evaluation of the TBIs. A similar retrospective study was conducted by Nygren et al. (2001) in Sweden. Records of patients hospitalized for TBI during 1965â1994 were linked with the Swedish Cancer Register, Cause of Death Register, and Emigration Register. Some 311,006 patients (192,090 men and 118,916 women) were followed for 3,225,317 person-years. TBI was defined as an ICD-7 or ICD-8 discharge code for skull trauma. There were 400 brain tumors during the period, but 119 that occurred in the first year after injury were excluded, and this left 281 cases of brain cancer (55 benign meningiomas; 161 primary brain tumors, including astrocytomas, glioblastomas, and gliomas; and 65 other brain tumors). When the patients were compared with the general Swedish population, the SIR was 1.0 (95% CI, 0.9â1.2) and did not vary substantially by tumor type, by age group, by severity, by sex, or by time since TBI. Those three studies found that the incidence of brain tumors after severe head trauma was no different from the incidence in the general population if cases diagnosed in the first year after trauma were excluded. The other 11 studies used self-reports or self-reports of physiciansâ diagnoses to ascertain exposure. That creates some methodologic limitations because of the risk of recall bias, which would tend to overestimate risk; it does allow exploration of milder TBI than the large, registry-based studies. These are listed below in order of first authorâs names. In a caseâcontrol study to identify environmental causes of brain tumors, Burch et al. (1987) assessed exposures in 215 people who lived in Toronto and southern Ontario and had brain tumors diagnosed in 1977â1981 and assessed in 1979â1982. Spongioblastomas, ependymomas, meningiomas, neuroepitheliomas, pituitary adenomas, neurilemmomas were excluded from the study. All subjects were individually matched to hospital-based controls by sex, area of residence, marital status, year of birth, date of diagnosis in the case of living patients, and date of death; controls were selected from the hospital nearest the residence of the subjects to reduce referral bias. None of the controls was admitted to the hospital for any type of cancer. Odds ratios (ORs) were calculated by using conditional and unconditional linear logistic regression models. Accidents and injuries were classified as having head involvement or not, and degree of injury was determined by whether medical attention was sought. Of those with brain tumors, 103 reported TBI compared with 41 controls, for an OR of 2.51 (p < 0.0001). When the subset of TBI that required medical attention was considered, the OR, 1.20, was not statistically significant (p = 0.65). The authors suggested that there was a potential for substantial recall bias in the reporting of TBI and that the risk disappeared when assessment was restricted to injuries that required medical attention. The association between epilepsy, TBI, and brain tumors was assessed in a caseâcontrol study by Carpenter et al. (1987). Medical histories of about 66,000 workers employed in 1943â 1979 at the Y12 nuclear facility at Oak Ridge, Tennesssee, or at the Oak Ridge National Laboratory were reviewed. Exposure was assessed by reviewing reports of TBI recorded in pre- employment medical histories or during employment that were in facility medical records. There were 82 fatal cases of primary malignant brain tumor (67 and 15 in white men and women, respectively) on the basis of review of death certificates. The authors compared the records of each of those cases with records of four controls matched for race, sex, facility, year of birth, and year of hire. Two (2.4%) of the 82 subjects reported having had a TBI compared with 9 (2.7%) of the 328 controls. The OR for brain cancer in those with a history of TBI compared with those
352 GULF WAR AND HEALTH without such injury was 0.9 (95% CI, 0.2â4.2); when the tumors were restricted to those of glial origin, the OR was 1.4 (95% CI, 0.3â7.2). The authors note that although the reports were documented, all the reported TBIs were self-reported and therefore subject to recall bias. Nonetheless, the recording of TBI at a time closer to the time of injury and for reasons unrelated to the study lessens the risk of recall bias. Hochberg et al. (1984) studied the effect of TBI on the risk of glioblastoma in a caseâ control study of 125 patients with tumors and their self-chosen âbest-friendâ controls who were of the same sex and within 5 years of age. The subjects were derived from 231 patients, 15â81 years old, who had histologically confirmed glioblastomas in three hospitals in Boston, one in Providence, and one in Baltimore. The history of TBI was initially obtained by self-report but confirmed by interview. TBI was classified as severe (skull fracture or concussion followed by a complication, such as coma, intracranial hemorrhage, epilepsy, shock, or long-lasting impairment of memory, hearing, or vision) or mild (well-described concussion, brief loss of consciousness without any complications, or other inadequately described head trauma); other mild or poorly described head injuries were excluded. The OR for TBI was 2.1 (95% CI, 1.1â 4.0); for mild TBI, 1.5 (95% CI, 0.7â3.3); and for severe TBI, 3.8 (95% CI, 1.3â11.0). If the TBI was sustained when the person was over 15 years old, there was an increased risk of glioblastoma, particularly in those over 50 years old at the time of the study and with severe TBI (OR, 5.1; 95% CI, 0.7â35.6). The age-adjusted OR for all ages was 1.4 (95% CI, 0.4â4.7) for mild trauma received after the age of 15 years and 10.6 (95% CI, 2.1â53.3) for severe trauma received at any ageâa statistically significant trend (p = 0.03). Hu et al. (1998) conducted a hospital-based caseâcontrol study of risk factors, including history of TBI, for astrocytoma and other glioma in residents of Heilongjiang province in northeastern China. Subjects were those with histologically confirmed gliomas (139 astrocytomas and 79 other brain gliomas) who presented for surgery in 1989â1995 at any of six major hospitals in the province. Two controls for each case (436 people) were selected from the same hospitals and were chosen from among those admitted for nonneoplastic or nonneurologic disease and matched for sex, age (within a 5-year interval), and area of residence. All study participants were asked about history of TBI that required medical attention and about diet, socioeconomic status, occupation, smoking, and other indicators of health status. A history of TBI was associated with increased odds of glioma (adjusted OR, 5.09; 95% CI, 2.51â10.31). Adjustments were made for income, education, number of years of drinking alcohol, occupational exposure, and consumption of vegetables and fruit. A hospital-based caseâcontrol study in Rio de Janeiro, Brazil (Monteiro et al., 2006), found increased odds of brain tumors in adults who had experienced a TBI. The 240 subjects were 30â65 years old and had been hospitalized in the Brazilian national health system hospitals and had a diagnosis of primary brain neoplasms in 1999â2002. The 268 controls were age- and sex-matched inpatients in the same geographic region who had diagnoses of diseases other than brain tumor. Assessment of TBI was based on self-reports and had to have occurred at least a year before the diagnosis of the brain tumor (cases) or hospitalization (controls). TBI severity was based on whether hospitalization, LOC, or amnesia had occurred after the injury and on the number of injuries. Brain cancer was associated with prior TBI (OR, 1.49; 95% CI, 1.03â2.15), and the OR was adjusted for age, sex, schooling, epilepsy, and alcohol consumption. However, when analyzed by histologic type, clinical markers of severity (such as LOC and amnesia), numbers of episodes of TBI, and time since TBI, only having had more than one episode of TBI
OTHER HEALTH OUTCOMES 353 was associated with brain tumors (OR, 3.14; 94% CI, 1.50â6.61; p for trend, 0.004) and TBI 10â 19 years before the diagnosis of brain tumor (OR, 1.31; 95% CI, 1.06â1.64). A doseâresponse relationship was observed according to the number of TBIs, particularly for meningioma (OR, 1.63; 95% CI, 0.96â2.75). Phillips et al. (2002) conducted a caseâcontrol study in western Washington state in 1995â1998 to identify the risk of intracranial meningioma after TBI. For each of 200 people with intracranial meningioma (143 women and 57 men), they age- and sex-matched two controls. Cases were identified from the population-based Cancer Surveillance System, which included about 3 million residents. Each study participant was asked about a history of TBI, when it had occurred (by 10-year intervals), and whether it had resulted in LOC, hospitalization, or a visit to an emergency room. Subjects had increased odds of TBI (OR, 1.93; 95% CI, 1.28â2.62). The OR for intracranial meningioma in people with mild TBI was 3.23 (95% CI, 1.82â5.71); in people with severe TBI, 1.27 (95% CI, 0.82â1.98); in people with two or more TBIs, 2.75 (95% CI, 1.48â5.03); and in people who had TBI 10â19 years previously, 4.33 (95% CI, 2.06â9.10). Preston-Martin and colleagues (1980, 1983) conducted a population-based caseâcontrol study of intracranial meningioma in women and men in Los Angeles. All microscopically confirmed cases of meningioma were identified in the Los Angeles County Cancer Surveillance Program, and each subject was matched to a control residing in the same neighborhood by sex, race, and year of birth. Women who received a diagnosis in 1972â1975 completed questionnaires regarding risk factors; questionnaires were returned by 189 subjects and 185 controls. A history of medically treated TBI was a risk factor for meningioma (OR, 2.0; 95% CI, 1.2â3.5) and was independent of having had head or neck radiography. In another caseâcontrol study that focused on men with meningioma, 120 subjects who received a diagnosis in 1972â 1979 were matched with 105 neighbor controls. TBI was associated with increased odds of meningioma if the person had ever participated in boxing as a sport (OR, 2.0; 95% CI, 1.1â3.2) or had a serious TBI that resulted in LOC or a permanent scar (OR, 1.9; 95% CI, 1.1â3.2). The authors noted that many of the men with TBI had not received medical treatment. In a later caseâ control study of the association between serious TBIâdefined as resulting in a medical visit, LOC, or dizzinessâand brain tumor in men, Preston-Martin et al. (1989) investigated cases of glioma and meningioma first diagnosed in 1980â1984 in Los Angeles County and identified in the Cancer Surveillance Program. The 272 subjects (202 with glioma and 70 with meningioma) were interviewed and were compared with 272 neighbor controls. To be included, serious TBI had to have occurred 2 years or more before the brain-cancer diagnosis. The OR for having had a serious TBI 20 years or more before was 0.8 (95% CI, 0.5â1.3) in those with glioma and 2.3 (95% CI, 1.1â5.4) in those with meningioma. Furthermore, the odds of meningioma, but not glioma, increased with the number of serious TBIs: with one TBI, 1.3 (95% CI, 0.6â2.9), with two TBIs, 2.1 (95% CI, 0.8â5.9), and with three or more TBIs, 6.2 (95% CI, 1.2â31.7). Expanding the caseâcontrol approach internationally, Preston-Martin et al. (1998) used a standardized questionnaire to investigate risk factors for glioma and meningioma in six countries (two centers in Australia, one in France, one in Germany, two in Canada, one in Sweden, and one in the United States). Glioma and meningioma cases were in 729 men and 779 women who received diagnoses of either glioma or meningioma in 1984â1992 and were matched to controls by sex, age, and education. Participants were asked about medically treated TBIs, which were classified as serious if they caused LOC or amnesia or required hospitalization. Injuries were also classified as to whether they occurred 5â14 years, 15â24 years, or more than 25 years before
354 GULF WAR AND HEALTH the brain-cancer diagnosis. ORs for glioma or meningioma were not significantly increased in men who had had any TBI 5 years or more before diagnosis (OR, 1.18; 95% CI, 0.94â1.48 and OR, 1.49; 95% CI, 0.86â2.57, respectively) or who had had a severe TBI 5 years or more before (OR, 1.13; 95% CI, 0.87â1.48 and OR, 1.15; 95% CI, 0.57â2.34, respectively). There was no significant increase in the risk of either brain tumor in women regardless of TBI severity. Men had a slightly increased OR for glioma if they had sustained more than one TBI 5 years or more before (OR, 1.52; 95% CI, 1.00â2.32), but not if they had more than one TBI regardless of timing (OR, 1.67; 95% CI, 0.56â4.98). In men who had sustained their TBI 15â24 years before diagnosis, there was a statistically significant increase in the risk of meningioma (OR, 5.35; 95% CI, 1.72â16.62), but the increase was not seen in connection with other latent periods or in women. In a study of primary brain tumors in residents of the Rein-Neckar-Odenwald area of Germany, Schlehofer et al. (1992) identified 226 cases diagnosed in two neurosurgical hospitals in January 1987âDecember 1988, of which 115 were histologically confirmed gliomas, 81 were meningiomas, and 30 were acoustic neuromas. The 99 men and 127 women, 25â75 years old, were interviewed during their hospital stay, as were 418 age- and sex-matched controls from the same residential areas as the cases. TBI that had occurred more than 5 years before and required a visit to a doctor were reported by 46 (20%) of the subjects and 113 (27%) of the controls (OR, 0.71; 95% CI, 0.5â1.1 for any brain tumor; OR, 0.70; 95% CI, 0.4â1.2 for gliomas; and OR, 0.52; 95% CI, 0.3â1.0 for meningiomas, adjusted for age and sex). The authors reported that there was no effect of having multiple TBIs or of varied latent periods, but the data to support these statements are not provided. These 11 studies had mixed results. Eight found evidence of associations between history of TBI and later brain tumors, and four did not. The results of the four Preston-Martin studies suggest that the odds of meningioma are increased in people who have had a TBI, especially those with relatively remote histories (15 years or more before). That was also found by Phillips et al. and in a study with more heterogeneous histologic subtypes by Monteiro. Nonetheless, the findings of those studies are less compelling than the findings of the large population-based studies in Minnesota, Denmark, and Sweden primarily because of the potential for overascertainment of exposure among cases due to self-reporting of TBI. Nonetheless, it is notable that some well-conducted studies yielded a relatively specific association between TBI and risk of later meningioma as opposed to other tumor types and that some studies yielded a finding of a latent period of 10 years of more. SECONDARY STUDIES The committee identified two secondary studies that evaluated the relationship between TBI and brain tumors (Choi et al., 1970; Zampieri et al., 1994). In a retrospective study (Choi et al., 1970) of patients with brain tumors in four University of Minnesotaâaffiliated hospitals in Hennepin County, Minnesota, there were 126 cases of histologically verified tumors diagnosed in 1963â1964. TBI was defined as a fractured skull, unconsciousness, or bleeding from the head that led to hospitalization or surgery. Controls were admitted to the hospitals for any condition other than tumors and were excluded if they had any neurologic, psychiatric, ophthalmologic, or lymphatic disorder; they were matched to cases by hospital of admission, sex, age, race, geographic area of residence, and locale of residence.
OTHER HEALTH OUTCOMES 355 No significant differences were seen in the frequency of TBI between the brain-cancer groups and their matched controls (OR, 0.83 for all verified central nervous system tumors; OR, 1.34 for all gliomas). Zampieri et al. (1994) conducted a caseâcontrol study to assess risk factors related to brain tumors in 195 patients who presented with confirmed cerebral glioma in four neurosurgical departments in Italy. Controls were matched to cases on age, sex, date of hospitalization, and residence. A structured questionnaire was administered to assess education, occupation, environmental exposures, medical history, and history of TBI. TBI was classified as mild if LOC was brief and severe if LOC lasted for over 1 hour, and there were any related neurologic deficits, epilepsy, cranial fracture, or any neurologic procedure. The authors found no statistically significant association between malignant astrocytomas and history of TBI (OR, 0.5; 95% CI, 0.2â1.3). The study was limited in that there was a potential for substantial recall bias. SUMMARY AND CONCLUSIONS The committee reviewed 14 primary studies and two secondary studies of TBI and brain tumors and found mixed results. The three large population-based registry studies in Minnesota, Denmark, and Sweden found no association between TBI and risk of brain tumors, although the Danish study almost reached statistical significance. The Minnesota study was able to ascertain exposure up to 44 years earlier, and there were more than 3.2 million person-years of followup in the Swedish study. However, there is evidence from some of the other studies that there may be a weak but significant association between TBI and meningioma and that risk of brain tumors may be increased 10 years or more after TBI; this suggests the possibility of a long latent period before clinical presentation. The committee therefore does not believe that the possibility of an association between TBI and risk of later brain tumors is a closed question. It believes that longer-term followup, especially in large registry-based studies, is warranted to understand where there is measurable risk and, if it is increased, when and with what types of tumors it is most likely to be observable. For now, however, the committee concludes that the inconsistent results of the studies are most supportive of a classification of inadequate/insufficient evidence to determine whether an association exists. The committee concludes, on the basis of its evaluation, that there is inadequate/insufficient evidence to determine whether an association exists between moderate or severe TBI and subsequent development of a brain tumor.
356 TABLE 10.2 TBI and Brain Tumors Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations Annegers et Double All TBI in Head injury with Brain tumors Four brain tumors observed None TBI that did not al., 1979 cohort Olmsted County, LOC, PTA, or skull (three astrocytomas, one reach medical care MN, 1935â1974 fracture meningioma); RR uncounted survived initial (observed/expected) not trauma; no significant overall or for two Expected numbers known pre- tumor types of tumors not existing tumor; adjusted for age or comparator data sex to match study from previous population incidence study of brain tumors in Olmsted County Burch et Caseâ All brain tumors Accidents, injuries Brain tumors 215 matched pairs analyzed; Matching on Excluded al., 1987 control in Toronto and that involved head more cases than controls basis of sex, area spongioblastomas, southern Ontario (not further reported injuries involving of residence, ependymomas, diagnosed in specified) head (RR, 2.51; p 0.0001), marital status, Â±5 meningiomas, 1977â1981 and but difference not years of birth, neuroepitheliomas, still resident in significant if head injury date of pituitary adenomas, 1979â1982; of required medical attention diagnosis, date neurilemmomas. 328 eligible, 247 (RR, 1.2; p = 0.65) of death (if death (75%) occurred) Recall bias, participated; nonparticipation comparator, bias noted by matched hospital authors controls; of 410 controls asked to participate, 228 (56%) interviewed Carpenter Nested Workers at two Head injury, self- Fatal primary 82 primary brain Matched by race, Misclassification of et al., 1987 caseâcontrol nuclear facilities reported on malignancy of malignancies: OR, 0.9 (95% sex, work site, exposures in Oak Ridge, occupational- brain CI, 0.2â4.2); for tumors of year of birth, TN, in 1943â medicine pre- glial origin: OR, 1.4 (95% year of hire Outcome assessed 1977; cases employment CI, 0.3â7.2) by death certificate,
Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations determined by assessments excluding those death certificate; who had not died four controls per of primary brain case malignancy Hochberg Caseâ Cases with Severe: resulted in Glioblasatoma, Unmatched analysis on 160 Stratification by Participation bias, et al., 1984 control glioblastomas skull fracture or histologically cases and 128 controls: age; RR adjusted recall bias from three concussion, confirmed overall RR, 2.1 (95% CI, (unknown for Boston, one followed by 1.1â4.0); severe RR, 3.8 what) Providence, one complications (95% CI, 1.3â11.0); mild Baltimore (such as coma, RR, 1.5 (95% CI, 0.7â3.3) hospitals; 15 intracranial years old; 160 of hemorrhage, Risk increased with age: 231 (69%) of epilepsy, shock, or RR, 10.6 (95% CI, 2.1â eligibles long-lasting 53.3) for 15 years old at participated; impairment of time of TBI 125 friend memory, hearing, controls matched or vision; mild: by 5-year age well-described group concussion or brief LOC without other complications Hu et al., Caseâ Cases from six History of head Histologically 34 of 218 cases vs 10 of 416 Matching on age Alcohol and skull 1998 control major hospitals trauma by self- confirmed controls reported head (Â±5 years), sex, x-rays also found in Heilongjiang report primary trauma; adjusted OR, 4.85 area of residence as risk factors Province, China, gliomas (95% CI, 2.52â9.44) in 1989â1995; requiring controls from surgery same hospitals with nonneurologic and nonneoplastic disease Inskip et Double All Danish Concussion, Intracranial Overall SIR, 1.36 (95% CI, None al., 1998 cohort; residents fractured skull, or tumors of CNS 1.20â1.53); 1 year PT 357
358 Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations Danish hospitalized with other head injury SIR, 1.15 (95% CI, 0.99â population TBI, 1977â1992 1.32); no difference by cell with TBI (n = 228,955); type compared comparator, with Danish Danish population population without TBI without history of TBI Monteiro et Hospital- 231 patients 30â Head injury >1year New diagnosis Association with prior head Age, sex, Only 80% of cases al., 2006 based caseâ 65 years old before diagnosis of of primary injury: adjusted OR, 1.49 education, confirmed control newly diagnosed brain neoplasm brain (95% CI, 1.03â2.15) epilepsy, alcohol histopathologically, with primary (cases) or neoplasm, consumption but brain tumors in hospitalization including By histologic type: glioma nonhistopathologic 1999â2002, (controls) by self- cerebral (n = 31), OR, 1.30 (95% CI, findings most admitted into 10 report; meningiomas, 0.71â2.35); meningioma (n suggestive; hospitals in Rio hospitalization, brain cancer, = 38), OR, 1.63 (95% CI, participation rate de Janeiro, amnesia, LOC used cranial nerve 0.96â2.75); 94% for cases and Brazil; as indicators of tumors, benign other with histopathology (n 90% for controls; 261 controls trauma severity and = 15), OR, 1.07 (95% CI, reason for matched by age, unspecified 0.52â2.21); other without hospitalization for sex, region of brain tumors histopathology (n = 23), 37.4% of controls residence from OR, 1.92 (95% CI, 0.99â was trauma; recall inpatients for 3.73) bias cannot be conditions other ruled out; than brain cancer As function of severity: information on hospitalized (n = 15), OR, head injury based 0.78 (95% CI, 0.37â1.64); on self-reports lost consciousness (n = 22), OR, 1.03 (95% CI, 0.55â 1.94); amnesia (n = 5), OR, 1.48 (95% CI, 0.38â5.83); any of these (n = 31), OR, 0.93 (95% CI, 0.54â1.60) As function of number of head injuries: 1 (n = 74),
Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations OR, 1.29 (95% CI, 0.85â 1.96); >1 (n = 28), OR, 3.14 (95% CI, 1.50â6.61; (p trend = 0.004) As function of years since head injury: 1â9 (n = 19), OR, 1.18 (95% CI, 0.73â 1.89); 10â19 (n = 27), OR, 1.31 (95% CI, 1.06â1.64); 20â29 (n = 23), OR, 1.07 (95% CI, 0.91â1.27); 30â39 (n = 20), OR, 1.09 (95% CI, 0.94â1.26); 40 (n = 18), OR, 1.09 (95% CI, 0.96â 1.24) Nygren et Retrospec- 311,006 patients Skull trauma that Primary brain 281 cases of brain tumors Stratification by Record-linkage al., 2001 tive hospitalized for survived tumors (55 meningiomas, age at injury, design permits population- TBI in 1965â hospitalization occurring >1 161 primary brain tumors, sex, years after assembly of large based cohort 1994 from (ICD-7 801, 853â year after 65 others) observed in TBI trauma, severity sample, but limited Swedish 855; ICD-8 801, trauma subjects (SIR, 1.0; 95% CI, of injury information Inpatient 850â854; ICD-9 through 1995 0.9â1.2); no relationship for available on other Register (of 801, 850â854); found by individual types of brain risk factors; discharges) considered in three linkage with tumor or severity radiation only without current severity groups: Swedish Suggestion of increase in likely confounder cancer vs age-, concussion, severe Cancer group 30â44 years old at for brain tumors, sex- and year- without Register, time of TBI: overall, SIR, but no apparent specific neurosurgery, Cause of 1.3 (95% CI, 1.0â1.7); problem in these incidence rates severe with Death benign meningiomas, SIR, negative findings; for Swedish neurosurgery Register, 1.0 (95% CI, 0.5â1.8); design adopted population Emigration primary brain tumors, SIR, because of question Register 1.4 (95% CI, 1.0â1.8); of reliability of other, SIR, 1.7 (95% CI, exposure recall in 0.8â3.2) caseâcontrol studies of brain- No suggestion of trend with tumor patients; 359
360 Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations time since trauma (p = 0.69) completeness of or increasing age (p = 0.25) ascertainment of meningiomas in registry of malignant diagnoses unknown Phillips et Population- 200 cases newly History of head Newly 99 cases, 142 controls with Age at diagnosis, Participation 84% al., 2002 based caseâ diagnosed in trauma by self- diagnosed any head trauma: OR, 1.83 sex, skull in cases, 55% control January 1995â report; considered meningiomas (95% CI, 1.28â2.62); mild, radiography, CT random-digit June 1998, 18 âseriousâ if LOC, (intracranial); OR, 3.23 (95% CI, 1.82â scanning of dialing controls, years old, went to ED, or exposures 5.71); severe, OR, 1.27 head; race, 67% in Medicare histologic hospitalized before (95% CI, 0.82â1.98); single, education left out controls confirmation by diagnosis (case OR, 1.51 (95% CI, 0.99â of model when Cancer applied to two 2.29); multiple, OR, 2.75 shown to have Lack of blinding of Surveillance controls) (95% CI, 1.48â5.08) had no effect interviewers to System at Fred gathered by in- case or control Hutchinson; 400 person Time before diagnosis: status might controls, two per interviews <10 years, OR, 1.39 (95% increase potential case matched on CI, 0.72â2.68;) 10â19 years, for recall bias age Â± 5 years, OR, 4.33 (95% CI, 1.28â sex by RDD or 2.62); 20 years, OR, 1.59 Cases arising less Medicare (95% CI, 1.09â2.31) than 1 year after eligibility lists; trauma not all English- excluded, so tumor speaking might have been residents of three cause of injuries or counties in found incidentally western during workup for Washington state TBI with telephone Conditional logistic analysis with information on medical, dental exposures to radiation
Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations Doseâresponse relationship for number of head traumas but not expected direction with âseverityâ of head injury (as defined) Preston- Caseâ Cases, women Head injury >2 Meningioma, 185 matched pairs analyzed: Matched by sex, 189 of 218 (87%) Martin et control 65 years old years before histologically OR for head injury treated race or ethnicity, eligible cases al., 1980 with intracranial interview that was confirmed medically, 2.0 (95% CI, year of birth (Â±5 interviewed; meningiomas medically treated 1.2â3.5) years); by interviewers not identified by history selecting blinded to caseâ through cancer controls from control status registry living in neighborhood, Los Angeles also matched by Differential recall County; one socioeconomic bias matched control status; per case from multivariate neighborhood logistic regression Preston- Caseâ Cases, men 65 Head injury >2 Meningiomas, 105 matched pairs analyzed Matched by sex, One-sided tests of Martin et control years old with years before histologically with exact binomial test: race or ethnicity, significance; al., 1983 intracranial diagnosis by confirmed serious head injury not year of birth (Â±5 differential recall meningiomas history; severe related to boxing, OR, 1.9 (p years); by bias identified head injury defined = 0.01); boxed as sport, OR, selecting through cancer as LOC or 2.0 (p = 0.03); either boxed controls from registry living in permanent scar or had severe head injury neighborhood, Los Angeles unrelated to boxing, OR, 1.8 also matched by County; one (95% CI, 1.1â3.2) SES; matched control multivariate per case from logistic neighborhood regression Preston- Caseâ Cases, men 25â Serious head injury Gliomas and 272 matched pairs (202 Matched by sex, 277 of 478 (58%) Martin et control 69 years old with >2 years before meningiomas, glioma, 70 meningiomas) race or ethnicity, eligible cases 361
362 Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations al., 1989 glioma or diagnosis of case histologically analyzed with exact year of birth (Â±5 interviewed; meningioma that resulted in confirmed binomial test years); by differential recall identified LOC, dizziness, or selecting bias less likely with through cancer medical For history of serious head controls from different findings registry, consultation trauma 20 years before neighborhood, for meningioma diagnosed in diagnosis: glioma, also matched by and glioma 1980â1984 in OR 0.8 (95% CI, 0.5â1.3); SES; Los Angeles meningioma, OR 2.1 (95% multivariate County; one CI, 1.1â5.4) logistic matched regression neighborhood For meningiomas only, control per case number of serious head injuries, p for trend = 0.01 Preston- Caseâ Cases from eight Medically treated Gliomas and 297 gliomas, 59 Individual and Subject to recall Martin et control centers in six head injuries; meningiomas meningiomas frequency bias; different al., 1998 countries subgroup of serious matching by age methods used for (Adelaide, TBI: medically Glioma: any TBI, males, and sex; some matching at Melbourne, treated injuries that OR, 1.18 (95% CI, 0.94â centers matched different centers Australia; resulted in LOC, 1.48), females, OR, 1.03 on race or Grenoble, PTA, or (95% CI, 0.42â2.55); any geographic France; hospitalization; serious TBI, males, OR, region; ORs Heidelberg, also recorded 1.13 (95% CI, 0.87â1.48), computed by Germany; participation in females, OR, 1.07 (95% CI, maximal- Toronto, sports (differed by 0.74â1.56) likelihood Winnipeg, region) that could estimates by Canada; result in TBI; Meningioma: any TBI, using both Stockholm, proxy respondents males, OR, 1.49 (95% CI, conditional, Sweden; Los could be used if 0.86â2.57), females, OR, unconditional Angeles, US; case or control 0.83 (95% CI, 0.54â1.28); logistic men, women 20 unavailable any serious TBI, males, OR, regression years old with 1.15 (95% CI, 0.57â2.34), diagnosed females, OR, 0.79 (95% CI, glioma or 0.45â1.39) meningioma Borderline increase in risk
Health Outcomes or Study Outcome Comments or Reference Design Population Type of TBI Measures Results Adjustments Limitations for >1 TBI in men with glioma (OR, 1.52; 95% CI, .00â2.32) but not seen in women or men with meningioma No correlation with sports participation Risk of meningioma in men higher 15-24 years after trauma (OR, 5.35; 95% CI, 1.72â16.62) Schlehofer Population- 226 cases in Self-reported Primary brain For all tumor types: 46 of Age-, sex- 418 of 521 (72%) et al., 1992. based caseâ Rhein-Neckar- history of head tumors (ICD-9 226 (20%) vs 113 of 418 matching for potential controls control Odenwald area injury requiring 191, 191.1, (27%); RR, 0.71 (95% CI, controls participated; self- of Germany with medical attention; 192.0), 0.5â1.1) reports of head primary brain obtained by restricted to trauma; no tumors interview gliomas (115), For gliomas: 27 cases vs 66 comparisons by diagnosed 1987â meningiomas controls; RR, 0.70 (95% CI, severity or number 1988; controls, (81), acoustic 0.4â1.2) of injuries 418 randomly neuromas (30) selected from For meningiomas: 13 cases residential vs 39 controls; RR, 0.52 registers (95% CI, 0.3â1.0) NOTE: CI = confidence interval, CNS = central nervous system, CT = computed tomography, ED = emergency department, ICD = International Classification of Diseases, LOC = loss of consciousness, OR = odds ratio, PT = posttrauma, PTA = posttraumatic amnesia, RDD = random-digit dialing, RR = relative risk, SES = socioeconomic status, SIR = standardized incidence ratio, TBI = traumatic brain injury. 363
364 GULF WAR AND HEALTH REFERENCES Annegers, J. F., E. R. Laws, Jr., L. T. Kurland, and J. D. Grabow. 1979. Head trauma and subsequent brain tumors. Neurosurgery 4(3):203â206. Baguley, I., S. Slewa-Younan, R. Lazarus, and A. Green. 2000. Long-term mortality trends in patients with traumatic brain injury. Brain Injury 14(6):505â512. Bianco, M., C. Fabbricatore, N. Sanna, C. Fabiano, V. Palmieri, and P. Zeppilli. 2007. Elite athletes: Is survival shortened in boxers? International Journal of Sports Medicine 28(8):697â702. Brown, A. W., C. L. Leibson, J. F. Malec, P. K. Perkins, N. N. Diehl, and D. R. Larson. 2004. Long-term survival after traumatic brain injury: A population-based analysis. Neurorehabilitation 19(1):37â43. Burch, J. D., K. J. Craib, B. C. Choi, A. B. Miller, H. A. Risch, and G. R. Howe. 1987. An exploratory case-control study of brain tumors in adults. Journal of the National Cancer Institute 78(4):601â609. Carpenter, A. V., W. D. Flanders, E. L. Frome, P. Cole, and S. A. Fry. 1987. Brain cancer and nonoccupational risk factors: A case-control study among workers at two nuclear facilities. American Journal of Public Health 77(9):1180â1182. Choi, N. W., L. M. Schuman, and W. H. Gullen. 1970. Epidemiology of primary central nervous system neoplasms. II. Case-control study. American Journal of Epidemiology 91(5):467â485. Corkin, S., E. V. Sullivan, and F. A. Carr. 1984. Prognostic factors for life expectancy after penetrating head injury. Archives of Neurology 41(9):975â977. Credner, L. 1930. Klinische und soziale auswirkungen von hirnsschÃ¤digungen. Z. Gesamte Neurologie et Psychiatrie 126:721â757. Engberg, A. W., and T. Teasdale. 2004. A population-based study of survival and discharge status for survivors after head injury. Acta Neurologica Scandinavica 110(5):281â290. Fearnside, M. R., R. J. Cook, P. McDougall, and R. J. McNeil. 1993. The Westmead Head Injury Project outcome in severe head injury. A comparative analysis of pre-hospital, clinical and CT variables. British Journal of Neurosurgery 7(3):267â279. Flaada, J. T., C. L. Leibson, J. N. Mandrekar, N. Diehl, P. K. Perkins, A. W. Brown, and J. F. Malec. 2007. Relative risk of mortality after traumatic brain injury: A population-based study of the role of age and injury severity. Journal of Neurotrauma 24(3):435â445. Gomez, P. A., R. D. Lobato, G. R. Boto, A. De la Lama, P. J. Gonzalez, and J. de la Cruz. 2000. Age and outcome after severe head injury. Acta Neurochirurgica 142(4):373â380. Harris, C., S. DiRusso, T. Sullivan, and D. L. Benzil. 2003. Mortality risk after head injury increases at 30 years. Journal of the American College of Surgeons 197(5):711â716. Harrison-Felix, C., G. Whiteneck, M. DeVivo, F. M. Hammond, and A. Jha. 2004. Mortality following rehabilitation in the traumatic brain injury model systems of care. Neurorehabilitation 19(1):45â54. Hochberg, F., P. Toniolo, and P. Cole. 1984. Head trauma and seizures as risk factors of glioblastoma. Neurology 34(11):1511â1514.
OTHER HEALTH OUTCOMES 365 Hu, J., K. C. Johnson, Y. Mao, L. Guo, X. Zhao, X. Jia, D. Bi, G. Huang, and R. Liu. 1998. Risk factors for glioma in adults: A case-control study in northeast china. Cancer Detection and Prevention 22(2):100â108. Hukkelhoven, C. W. P. M., E. W. Steyerberg, A. J. J. Rampen, E. Farace, J. D. F. Habbema, L. F. Marshall, G. D. Murray, and A. I. R. Maas. 2003. Patient age and outcome following severe traumatic brain injury: An analysis of 5600 patients. Journal of Neurosurgery 99(4):666â673. Inskip, P. D., L. Mellemkjaer, G. Gridley, and J. H. Olsen. 1998. Incidence of intracranial tumors following hospitalization for head injuries (Denmark). Cancer Causes and Control 9(1):109â 116. Lai, Y. C., F. G. Chen, M. H. Goh, and K. F. Koh. 1998. Predictors of long-term outcome in severe head injury. Annals of the Academy of Medicine, Singapore 27(3):326â331. Lewin, W., T. F. Marshall, and A. H. Roberts. 1979. Long-term outcome after severe head injury. British Medical Journal 2(6204):1533â1538. Lu, J., A. Marmarou, S. Choi, A. Maas, G. Murray, and E. W. Steyerberg. 2005. Mortality from traumatic brain injury. Acta Neurochirurgica Suppl 95:281â285. Miller, J. D., and W. B. Jennett. 1968. Complications of depressed skull fracture. Lancet 2(7576):991â995. Monteiro, G. T. R., R. A. Pereira, R. J. Koifman, and S. Koifman. 2006. Head injury and brain tumours in adults: A case-control study in Rio de Janeiro, Brazil. European Journal of Cancer 42(7):917â921. NCI (National Cancer Institute). 2008. Brain Tumor. http://www.cancer.gov/cancertopics/types/brain/ (accessed September 4, 2008). NCIP (National Center for Injury Prevention and Control). 2008. Traumatic Brain Injury. http://www.cdc.gov/ncipc/tbi/TBI.htm (accessed July 31, 2008). Nygren, C., J. Adami, W. Ye, R. Bellocco, J. L. af Geijerstam, J. Borg, and O. Nyren. 2001. Primary brain tumors following traumatic brain injuryâa population-based cohort study in Sweden. Cancer Causes and Control 12(8):733â737. Pentland, B., L. S. Hutton, and P. A. Jones. 2005. Late mortality after head injury. Journal of Neurology, Neurosurgery and Psychiatry 76(3):395â400. Phillips, L. E., T. D. Koepsell, G. van Belle, W. A. Kukull, J. A. Gehrels, and W. T. Longstreth, Jr. 2002. History of head trauma and risk of intracranial meningioma: Population-based case- control study. Neurology 58(12):1849â1852. Preston-Martin, S., W. Mack, and B. E. Henderson. 1989. Risk factors for gliomas and meningiomas in males in Los Angeles County. Cancer Research 49(21):6137â6143. Preston-Martin, S., A. Paganini-Hill, B. E. Henderson, M. C. Pike, and C. Wood. 1980. Case- control study of intracranial meningiomas in women in Los Angeles County, California. Journal of the National Cancer Institute 65(1):67â73. Preston-Martin, S., J. M. Pogoda, B. Schlehofer, M. Blettner, G. R. Howe, P. Ryan, F. Menegoz, G. G. Giles, Y. Rodvall, N. W. Choi, J. Little, and A. Arslan. 1998. An international case- control study of adult glioma and meningioma: The role of head trauma. International Journal of Epidemiology 27(4):579â586.
366 GULF WAR AND HEALTH Preston-Martin, S., M. C. Yu, B. E. Henderson, and C. Roberts. 1983. Risk factors for meningiomas in men in Los Angeles County. Journal of the National Cancer Institute 70(5):863â866. Ratcliff, G., A. Colantonio, M. Escobar, S. Chase, and L. Vernich. 2005. Long-term survival following traumatic brain injury. Disability and Rehabilitation 27(6):305â314. Rish, B. L., J. D. Dillon, and G. H. Weiss. 1983. Mortality following penetrating craniocerebral injuries. An analysis of the deaths in the Vietnam Head Injury Registry population. Journal of Neurosurgery 59(5):775â780. Schlehofer, B., M. Blettner, N. Becker, C. Martinsohn, and J. Wahrendorf. 1992. Medical risk factors and the development of brain tumors. Cancer 69(10):2541â2547. Selassie, A. W., M. L. McCarthy, P. L. Ferguson, J. Tian, and J. A. Langlois. 2005. Risk of post- hospitalization mortality among persons with traumatic brain injury, South Carolina 1999- 2001. Journal of Head Trauma Rehabilitation 20(3):257â269. Shavelle, R., and D. Strauss. 2000. Comparative mortality of adults with traumatic brain injury in California, 1988â97. Journal of Insurance Medicine (Seattle) 32(3):163â166. Walker, A. E., H. K. Leuchs, H. Lechtape-Gruter, W. F. Caveness, and C. Kretschman. 1971. Life expectancy of head injured men with and without epilepsy. Archives of Neurology 24(2):95â100. Weiss, G. H., W. F. Caveness, H. Einsiedel-Lechtape, and M. L. McNeel. 1982. Life expectancy and causes of death in a group of head-injured veterans of World War I. Archives of Neurology 39(12):741â743. Winqvist, S., M. Lehtilahti, J. Jokelainen, H. Luukinen, and M. Hillbom. 2007. Traumatic brain injuries in children and young adults: A birth cohort study from northern Finland. Neuroepidemiology 29(1-2):136â142. Zampieri, P., F. Meneghini, F. Grigoletto, M. Gerosa, C. Licata, L. Casentini, P. L. Longatti, A. Padoan, and S. Mingrino. 1994. Risk factors for cerebral glioma in adults: A case-control study in an Italian population. Journal of Neuro-Oncology 19(1):61â67.