Acquired brain injuries are any injury to the brain occurring after birth, and are one of the leading causes of death and disability in adults;1,2 with stroke and traumatic brain injury (TBI) being the two most common causes of acquired brain injury in developed countries. Despite differing etiologies, the consequences of stroke and TBI share considerable overlap in regards to the types of disabilities which might result and rehabilitation services engaged.3 As such it is important to examine differences in who is at most risk of experiencing stroke and TBI via incidence profiles, as this will inform planning of evidence-based healthcare and allow evaluation of the impact of preventative/management strategies. Below we review the definition of stroke and TBI, respectively; followed by a brief examination of epidemiological findings.

Stroke

Stroke is defined by the World Health Organization (WHO) as “rapidly developing signs of focal (or global) disturbance of cerebral function, lasting longer than 24 hours (unless interrupted by death) with no apparent non-vascular cause”.4 Strokes can be either ischaemic (IS, occlusion of a blood vessel) or haemorrhagic (HS, rupture of a blood vessel) with both prognosis and treatment differing according to the specific nature of stroke identified.5 IS comprises 80–85% of all strokes.4

There is variability in reported age-adjusted incidence rates of stroke across studies. It is thought that this may reflect true differences between the populations being studied, methodological differences and/or differences within countries (eg, higher incidence in rural areas).6,7

Despite national variations, a systematic review of worldwide stroke incidence6 revealed distinct trends in stroke incidence according to country income levels, with >100% increase in low- to middle-income countries and a 42% decrease in high-income countries since 1970. However, similar age-adjusted incidence of IS were reported for high-income and low- to middle-income countries in the last decade. In contrast, the pooled proportional frequencies of IS in high-income countries were higher (83%) in 2000–2008 compared to low- to middle-income countries (67%), who had higher HS incidence.

Case-fatality rates (21–30 days) following IS have declined in high-income countries from 10–32% (1990–1999) to 13–23% (2000–2008).6 Case-fatality rates for low- to middle-income countries for IS have only recently been published and suggest higher case-fatality in these countries (range 17.8–23.2%), possibly reflecting differences in acute care.5 Years of life lost due to premature mortality (YLLs) attributed to stroke (80–85% IS) vary widely but global estimates are 9.5% and 9.9% of total deaths for low-income countries and high-income countries respectively, making it the second leading cause of death after ischaemic heart disease.8 In addition to mortality, stroke is also a leading cause of disability.9,10

In terms of demographic factors linked to stroke incidence that might be used to assist in targeting preventive efforts, age (increased age linked to greater risk),11–13 gender (with male gender having higher risk)14–16 and ethnicity (with minority ethnicity having higher risk)17–19 have received the greatest attention.

Traumatic brain injury

TBI is defined by the WHO as an “acute brain injury resulting from mechanical energy to the head from external physical forces”. Clinical identification of TBI requires the presence of at least one of the following: (1) confusion or disorientation; (2) loss of consciousness; (3) post-traumatic amnesia; or (4) other neurological abnormalities (eg, focal neurological signs, seizure, intracranial lesion).20 TBI severity is classified as mild, moderate or severe, based on Glasgow Coma Scale (GCS) scores and/or duration of post-traumatic amnesia.20 According to the literature, irrespective of age, 70–90% of TBIs are mild with 5–20% being moderate and severe.21

In developed countries (eg, Australia, UK), incidence of TBI is approximately 200–300 people per 100,000 population annually.22 However, a recent WHO systematic review23 suggests the annual incidence of mild TBI is probably >600/100,000.23 This discrepancy reflects that, in contrast to stroke, in Europe and North America only about 25% of persons who experience a TBI are admitted to hospital.24,25

In contrast to stroke, TBI has a high incidence in those under five years of age and those aged <35 years resulting in a long-term impact on employment and disability,26 with DALYs lost due to TBI is significant.27 The economic cost of TBI varies significantly, due to both individual characteristics and the nature of the injury.28 As noted by Maas et al29 TBI-related costs in Europe for 2017 are estimated at US$49.7 billion, which comprises 41% direct costs and 59% indirect costs; while in the US aggregated direct and indirect cost estimates range from US$60.4 billion in 2000 to US$221 billion in 2009.

As for stroke, age, male gender and ethnic minority status are linked to increased incidence of TBI. Indeed, the literature suggests a peak in incidence between 15 and 24 years of age, with smaller peaks in infancy (<5 years of age) and in older age (65+ years of age).30 Furthermore, males are reported to have approximately twice the risk of TBI compared to females (95% CI 1.6–2.8), though this varies with age.31,32

Ethnic minority groups also have greater risk of TBI and have higher post-TBI mortality rates.33,34 These ethnic disparities remain even when socioeconomic status (SES) is controlled for.35 Studies indicate that lower SES is associated with increased risk of TBI,36 and poorer TBI outcomes.35 An added factor, alcohol use, also increases risk of TBI related to falls, motor vehicle accidents and assault.37 These later factors have not been examined in regards to incidence due to lack of population data that could be used as a denominator.

Summary

Accurate and representative population-based data are crucial to determining the true incidence of TBI and stroke; for planning evidence-based healthcare and for developing and evaluating the impact of preventative and management strategies. Despite stroke and TBI being the two main causes of acquired brain injury there are differences in their epidemiology.38 While there is a high level of epidemiological knowledge about stroke generally, there is less knowledge in relation to TBI, with few population-based studies available and issues related to capture of cases with mild TBI being particularly relevant.

The need to conduct and contrast high-quality epidemiological studies separately for stroke and TBI is suggested by different demographic profiles of the two. As noted above, while certain age groups, male gender and minority ethnicity increase risk for both stroke and TBI, their profiles differ. Further information is required to understand how these demographic profiles might differ, and therefore inform different targets for preventive efforts.

The aim of the present study is to examine differences in incidence profiles of stroke and TBI in relation to age and ethnicity. New Zealand is in a unique position to contrast these two main forms of acquired brain injury with two population-based epidemiological studies conducted in New Zealand using comparable methodologies. Both studies meet the gold standard requirements for incidence studies. These are the Brain Injury Outcomes New Zealand In the Community (BIONIC) and the fourth round of the Auckland Regional Outcomes of Stroke Studies (ARCOS-IV). While the findings of these studies have been published individually elsewhere,39–43 they have not been previously contrasted.

Methods

Table 1 summarises the basic methodologies of the prospective BIONIC and ARCOS-IV studies, with greater detail on each provided in the text which follows. Both studies captured cases across one year, with follow-up of those who provided consent for up to 12 months post-ascertainment. For both BIONIC and ARCOS-IV, demographic information was captured for all eligible participants in order that incidence could be calculated by age, gender and ethnicity.

As can be seen in Table 1, methodologies for the two samples diverged due to the nature of the samples with a much wider range of sources for cases required to identify TBI given that many do not attend hospital.25,44

Also of note is that two different regions of New Zealand were used. For stroke, the Auckland region defined the base population. The data used here are from the fourth ARCOS study, with all three prior studies also being conducted within the Auckland region in order to allow examination of change over time. Thus, Auckland was used in order to remain constant across the ARCOS studies, allowing for examination of trends and patterns in incidence and outcomes over time. While consideration was given to using Auckland as the base population for the BIONIC study, given the expected incidence of TBI is much greater than that for stroke, it was not deemed financially or logistically feasible. As such the Waikato region, whose main city, Hamilton, is located approximately 90 minutes’ drive south of Auckland was selected. It was also noted at the time that the demographic characteristics of the Waikato region reflect those of the New Zealand Census in regards to gender and ethnicity, which meant that the findings could be more easily extrapolated to the wider New Zealand population.42

Stroke

Stroke was defined according to the World Health Organization.45 This definition includes primary intracerebral haemorrhage (PICH) and subarachnoid haemorrhage (SAH), but excludes cases of ‘silent stroke’ detected by neuroimaging, without appropriate clinical signs and symptoms.

Complete case ascertainment was assured by multiple overlapping sources of information (Table 1) on all new hospitalised or non-hospitalised cases (‘hot-pursuit’ method46).

Regular checks of private hospitals, rest homes and community health services (general medical practice, rehabilitation centres, outpatient clinics) were made to capture non-hospitalised cases. New Zealand Health Information Systems (NZHIS) data from the New Zealand Ministry of Health of all fatal and non-fatal stroke cases in the study population (‘cold-pursuit’ methods) were also examined. A diagnostic review committee comprising four stroke neurologists met fortnightly to confirm the diagnosis of stroke.

The study was approved by the Northern X Regional Ethics Committee and the Auckland University of Technology Ethics Committee.

Table 1: General characteristics of the Bionic and Arcos-IV studies.

‍

Note: CT = Computed Tomography, PHO= Public Healthcare Organisation; MRI= Magnetic Resonance Imaging; ACC = Accident Compensation Corporation, a no-fault government-funded body responsible for healthcare provision after any injury.

aThe Waikato region includes the city of Hamilton (129,249 urban residents) and surrounding rural area (Waikato district 43,959 residents).

TBI

The methods of the BIONIC study are provided in detail elsewhere.40,42 TBI was defined according to World Health Organization criteria.47 TBI was operationally defined as the presence of one or more of: (1) confusion or disorientation; (2) loss of consciousness; (3) post-traumatic amnesia; (4) other neurological abnormalities (eg, focal signs, seizure20). These symptoms needed to be primarily related to the TBI and not other causes (eg, substance use, other injury, psychological/medical conditions). Each TBI diagnosis was confirmed by medical record (or clinical details) review for each patient by a diagnostic adjudication group, including study neurologists. Less than 40 individuals in the sample had CT scans (as compared to 97% in stroke). Standard definitions of TBI severity were applied using the Glasgow Coma Scale (GCS) and/or Westmead Post-traumatic Amnesia (PTA) scale.48 Classification of mechanisms of TBI was guided by the ICD-10 and included injuries due to falls, motor vehicle accidents, assaults, exposure to external forces (eg, being hit/struck by another person/animal/inanimate object) and other causes (including unspecified causes).

Case ascertainment included both prospective and retrospective surveillance systems to ensure registration of all events in residents (defined as having lived in the study area for a minimum of six months). It included all ages and all severities over the 12-month period. Multiple overlapping sources of information on all new hospitalised and non-hospitalised cases were used (see Table 1 and Theadom et al42 for details). The study was approved by the Northern Y Regional Ethics Committee of New Zealand and the Auckland University of Technology Ethics Committee.

Statistical methodology

For both studies, multiple self-identified ethnicities were recorded, and so prioritised ethnicity corresponding to the New Zealand Census definitions were used for Māori, Pasifika, Asian/other and New Zealand Europeans/Pākehā. Age, sex and ethnic structure of the corresponding census data (Auckland for ARCOS-IV, and Hamilton and Waikato for BIONIC) were used as the denominator in calculating incidence. Crude annual age-, sex- and ethnic-specific stroke and TBI incidences per 100,000 people were calculated assuming a Poisson distribution. Standardised rates were calculated using the direct method and age-standardised to World Health Organization ‘World’ standard population.49

Results

Overall findings

Table 2 presents an overview of the findings from the two studies. As seen in the table, for both samples over half of incident cases were male and less than half of cases consented to follow-up. Males and females shared a similar level of risk for stroke, but this was not true of TBI where males had twice the risk of mild TBI and a three-fold risk of moderate or severe TBI when compared to females. Of note a much larger proportion of cases with TBI (35.6%) were identified through non-medical sources as compared to those with stroke (3%). The mean age of those with TBI (mean = 28.1 years) was also considerably younger than for those with stroke (mean = 69.24 years).

The incidence of TBI was more than five times greater overall than that of stroke. Incidence of TBI in Māori was not only 1.5 times higher than that of TBI incidence overall, but was 10 times the risk of stroke in Māori. Indeed, Māori had the highest incidence rate for TBI while New Zealand European/Pākehā had the highest incident rate for stroke.

Table 2: General characteristics and overall incidence rates and by ethnicity for NIONIC and ARCOS-IV studies.  

‍

Note: Pākehā = New Zealand European.

aNote BIONIC was completed prior to ARCOS-IV and as such a previous version of the census was used as the denominator. The newer versions of the census provide additional breakdown of the population into ethnicity, which includes Pacific Island peoples as a separate group. Tables 3 and 4 provide the population denominator, n and incidence (with 95% confidence intervals) by gender and ethnicity within each age range for TBI and stroke.

Figure 1: The incidence of TBI and stroke by age and gender.

Age-related incidence by gender and ethnicity

Figures 1 and 2 present ethnicity for stroke and TBI by gender and by ethnicity, respectively. It should be noted that differing age ranges in the figures reflect the age limits for the two studies (eg, ACROS-IV lower age limit for inclusion was 15 years) as well as the census data age ranges available.

As can be seen in Figure 1 the patterns of incidence by age differ quite extensively. For TBI incidence peaks in those aged <5 years, and then again in those aged 15–35 with males having the highest incidence across the lifespan. In contrast, incidence of stroke was lowest in those aged under 65 years and then increases over time to peak in those aged 85 year and older. Stroke incidence for males was slightly higher than that of females until age 85+ where female incidence was higher.

When incidence is examined by age and ethnicity (see Figure 2) it becomes evident that while New Zealand European/Pākehā have the highest incidence of TBI when under age five, this reduces noticeably from age 15. In contrast, Māori have the highest incidence from age five years onwards, and remains at a high level.

In contrast, for stroke Pacific Island people and Māori have higher incidence rates than New Zealand European/Pākehā until the 75–84-year age range. From the age of 85 years Europeans have the highest stroke incidence. It should be noted (see Table 4) that this pattern occurs in the context of reduced sample size over 85 due to smaller population size in older age groups of non-Europeans compared to Europeans. For example, in Māori over age 85 years, the incidence reflects only two cases from 243 in the population, as compared to 279 cases from a population of 16,776 for Europeans.

Figure 2: Incidence for TBI and stroke by age and ethnicity.

Table 3: TBI incidence rates (per 100,000 person-years) by gender and ethnicity in Hamilton/Waikato District, New Zealand 2010-2011.

‍

Table 3: TBI incidence rates (per 100,000 person-years) by gender and ethnicity in Hamilton/Waikato District, New Zealand 2010-2011 (continued).

‍Table 4: Stroke incidence rates (per 100,000 person-years) by gender and by ethnicity in Auckland, New Zealand 2011-2012.

‍

*Standardised to the WHO World Standard Population.49

Discussion

First, while TBI and stroke, as the two most common causes of acquired brain injury, share a number of characteristics, TBI incidence is much higher than that of stroke in New Zealand. While the number of TBI cases identified was smaller than that for stroke, the smaller size of the population from which these were drawn (the denominator) means a higher incidence. The overall incidence rate for TBI (790 per 100,000 overall, 749 per 100,000 for mild TBI) was over five times that for stroke. In addition, the incidence of TBI reported here is much higher than previous estimates in most other studies with comparable methodologies from the US, Europe and Asia-Pacific region (range from 55–600/100,000 people per year).50–56 A recent comprehensive World Health Organization systematic review of mild TBI epidemiology56 showed an annual incidence of in the range of 100–300/100,000 population. However, because mild TBI is usually not treated at hospitals, they suggested that the true population-based incidence of mild TBI is probably above 600/100,000.56 The BIONIC study is the first population-based study to confirm this suggestion, and to show that 36% of cases did not seek medical attention and that 95% of all TBI cases are mild.40 In the present sample the most common cause of TBI was falls (37.7%), followed by exposure to a mechanical force (ie, being hit by an object; 21%), vehicular accidents (20.2%) and assaults (16.7%). This is similar to a report on the number of TBI related hospitalisations, deaths and emergency department visits caused by an external force in the US (2002 to 2006) where falls were the most common mechanism, followed by being struck/by, motor vehicle traffic and assault.57 This suggests that medical versus population-based samples do not differ in mechanism of injury.

In the present sample 52% of strokes occurred in females, with males and females having similar crude incidence rates. Standardised stroke incidence was slightly higher in males than females (129 and 110 per 100,000, respectively). This later finding presents the same pattern of gender differences (though to a lesser extent) as standardised global data; where in 2013 global ischaemic stroke (IS) and haemorrhagic stroke (HS) incidence (per 100 000) in men (IS 132.77 [95% UI, 125.34–142.77]; HS 64.89 [95% UI 59.82-68.85]) exceeded those of women (IS 98.85 [95%UI, 92.11–106.62]; HS 45.48 [95% UI, 42.43–48.53]).58

While examining overall incidence of TBI and stroke suggests that TBI is likely to result in greater burden and greater need for medical care due to higher numbers, it could be reduced through targeted prevention efforts. For example, while males and females shared a similar level of risk for stroke, this was not true of TBI where males had twice the risk of mild TBI and a three-fold risk of moderate or severe TBI when compared to females. This suggests the need for prevention strategies developed to target on males. It has been suggested that males are at heightened risk of TBI due to risk-taking behaviours such as engaging in excess alcohol consumption, playing in high-impact sports and driving more aggressively. For example, in one study of severe TBI59 the data suggest that younger people acquired severe TBI from riskier behaviours compared to older individuals; and being male and being alcohol intoxicated also contributed significantly to risk-taking behaviour. Such risk-taking behaviours could become a target for TBI prevention.

Examining incidence across age ranges showed that the mean age of those with TBI was also considerably younger than for those with stroke, resulting in a longer time living with disability. This suggests that stroke prevention and education should be targeted at reducing risk through addressing modifiable risk factors such as diet and exercise before individuals reach this older age range. A limitation to this suggestion appears if one also considers ethnicity, as those who self-identified as Māori and Pacific Island ethnicities were at highest risk of young stroke. This indicates the potential of a different stroke prevention strategy for this group, targeting risk behaviours in a much younger age range; including risk factors/behaviours identified as more prevalent in Māori and Pacific Island communities in New Zealand, which include increased rates of atrial fibrillation,60 smoking61 and obesity.62

One of the other risk factors for stroke not mentioned as yet is the presence of ‘mini stroke’ or transient ischaemic attack (TIA). TIA is an acute focal cerebrovascular event with symptoms lasting less than 24 hours.63 The presence of TIA allows for identification of specific individuals seeking medical attention who should then be the particular focus of stroke preventive efforts. Unfortunately, no such ‘early warning’ exists for TBI. Prevention of stroke is also linked to other medical indicators of increased risk for cerebrovascular disease (eg, high blood pressure, high cholesterol), which can be identified and monitored by a health professional. No such indicators exist for TBI. However, in examining the demographic profile of TBI we do know that some age, gender and ethnic populations are at greater risk. For example, those in older age ranges (>65 years) are at increased risk of TBI, and in particular the literature reflects an increased risk of TBI as a result of falls.40 Thus reducing the risk of falls in older age ranges through preventive strategies such as increasing mobility, flexibility and balance could be one means of preventing such injuries and reducing incidence.64

For TBI, incidence also peaked in those under five years of age (and particularly for New Zealand Europeans/Pākehā) and then again from 15 to 35 (with a particularly high and extended peak for Māori). Taken together the above suggests that the greatest reduction in TBI incidence is likely to occur through targeting prevention strategies in those who are male, Māori and ages 15 to 35 years. It should be noted here that given less than 36% of TBI cases were located through medical sources, peaks in incidence and differences in incidence between demographic groups cannot be attributed solely to differences in medical help-seeking behaviours. It is also important to ensure that TBI and stroke services are culturally responsive and accessible for all groups.

It should further be noted that given the young age at which TBI incidence peaks occur, the burden (eg, direct and indirect costs, DALYs, etc) of this condition across the lifespan are likely to be much greater than those of stroke. As such, targeting TBI for prevention should be a priority.

Of note a much larger proportion of cases with TBI (35.6%) were identified through non-medical sources as compared to those with stroke (3%). This suggests a potential area for community education, as the literature shows that even minimal intervention early following mild TBI (ie, provision of an information booklet) can reduce long-term negative outcomes and risk for recurrent injury.65 Further refining of TBI preventive strategies could potentially result from using information on mechanisms/causes of injury including the collection of more detailed injury information to identify new opportunities for prevention.

In addition, in contrasting the incidence rates produced by the two studies, methodological issues which impact the ability to conduct such studies also became evident. For example, while only 3% of stroke cases were not identified through medical sources, over 35% of those with TBI did not seek medical attention. This has implications for both research and clinical practice. In regards to research, funding of studies to identify TBI cases cannot rely upon hospital admission and discharge records as is the case for stroke. As can be seen in Table 1, not only were TBI cases more common than stroke cases, a much wider variety of overlapping sources for case identification were used to identify TBI cases, including outpatient clinics, rest homes, sports centres, ambulance service, prison, schools and self-referrals to the study following publicising within the media (eg, local newspapers, radio and television). The amount of time required to conduct regular searches of all such sources of cases makes conducting community-based TBI studies of this nature much more costly. In addition, despite TBI having such a high incidence it should be noted that this likely remains an underestimate. Indeed, additional sources of potential cases were not specifically accessed in BIONIC, including police records (eg, domestic and non-domestic violence), domestic refuge shelters and child protective service records.

Conclusion

Though as acquired brain injuries, TBI and stroke are often subject to similar outcomes and therefore rehabilitation interventions, differences in their incidence suggest the need to target efforts at prevention very differently for the two groups.

TBI incidence is over five times that for stroke and was also much higher than previous estimates of other countries; though the commonness of various mechanisms of injury mirrored those in other countries. In contrast, stroke incidence was similar to that reported elsewhere. Differences in the demographic profiles of stroke and TBI suggest different preventive strategies. For example, while males and females shared a similar level of risk for stroke, males had twice the relative risk of mild TBI and a three-fold risk of moderate or severe TBI compared to females; suggesting prevention targeted at males. Similarly, given the mean age of those with stroke was higher than that of TBI, stroke prevention and education should target modifiable risk factors (eg, diet and exercise) before individuals reach this older age range. However, if ethnicity is also considered, prevention strategies for Māori and Pacific Island peoples, who are at highest risk of young stroke, should be targeted a much younger age group.

‍