Severe neurotrauma in Switzerland : have short-term outcomes improved ?

Neurotrauma has a high incidence in high-income countries (790 per 100,000 population per year) and can be considered a silent epidemic. Severe traumatic brain injury (TBI) is a major burden for societies and is associated with high costs for both immediate and long-term care. Population-based studies including patients with severe TBI are rare. A recent cohort study in Switzerland observed an incidence of 11 / 100,000 population / year. Mortality rate at 14 days post-injury was 30% in Switzerland and was associated with the severity of the injury and the age of the injured person. Thirty-five percent of patients were >65 years; in this subpopulation the incidence (22/100,000/ year) and death rate (41%) were higher; this high proportion of elderly patients in this setting is new. A decrease in disability in the first year after TBI was observed in large multicentre cohort studies including the Swiss cohort study. There is some evidence that the speed of decrease of disability over time is associated with intensive neurorehabilitation. In conclusion, short-term outcome may have improved for younger patients over recent years, but this improvement may be masked by the higher proportion of elderly patients with less favourable outcomes. Additionally, we propose that clinical pathways from the prehospital period to rehabilitation could be improved, and in turn allow a higher level of positive outcomes not only in young but also in elderly patients.


Introduction
Neurotrauma, in particular severe traumatic brain injury (TBI) -which makes up about 5% of all neurotrauma -is a worldwide health concern and a great burden for a society.Costs of care are reported to be substantial (about $250,000 Australian [180,000 CHF] per patient with severe TBI) [1].Neurotrauma has been considered a "silent epidemic", with an incidence of 790 per 100,000 individuals per year [2] in New Zealand, for example."Silent" because our society seems to accept a high prevalence of neurotrauma as an inevitable fact, which may explain why brain research is underfunded [3].Interestingly, newspapers and television regularly report on spectacular accidents involving young athletes that result in severe TBI, but fail to integrate those cases into existing scientific evidence yielded by epidemiological studies.The epidemiological evidence just mentioned indicates a considerable increase in severe TBI in elderly individuals that results from falls [4].In Switzerland, research on neurotrauma is rare.Among the few studies that have investigated neurotrauma, single centre observations are the most common.In the last 10 years a national research network -entitled "Patient-relevant Endpoints after Brain Injury from Traumatic Accidents" (PEBITA) -has been conducting research in the domain of severe TBI.The goal of this research network is to increase knowledge about severe TBI in Switzerland, and make data transparent and available for the public by reporting and standardizing patient recovery from the prehospital setting to the post-rehabilitation period.The PEBITA network intends to increase the knowledge and perception of the neurotrauma epidemic.In view of the limited high quality evidence of effective interventions for patients with severe TBI, the research network decided to focus on an epidemiological approach, which can identify specific population-based risk factors for Switzerland and thus other high-income countries.The present review aims at summarising incidence and short-term outcome in adult patients suffering from severe TBI in Switzerland and other high-income countries.Furthermore, it will identify potential factors that may improve the outcome after a severe TBI.Short-term outcomes include mortality and functional disability (0.5 to 12 months), long-term outcomes include neuropsychological deficits, quality-of-life and burden on the family (3 months to many years).

Incidence
Reported population-based estimates of the incidence may vary considerably because of differences in inclusion criteria across studies (table 1).First, the estimated incidence is higher if data collection starts in the prehospital period as opposed to after hospital admission.Second, estimated incidence depends on the inclusion criteria for severe TBI: If at hospital admission a Glasgow Coma Score (GCS) <9 is used as inclusion criterion, the incidence will be lower compared with an inclusion based on the abbreviated injury score of the head (HAIS) >3.The estimated incidence will also be different if the International Classification of Diseases (ICD) code is used as inclusion criterion.The ICD is the diagnostic tool commonly used at the end of the acute hospital stay; consequently, it may be more useful for retrospective studies or studies starting after the acute hospital stay.Third, there is considerable inter-rater variability for the GCS at hospital admission, which may bias the estimates of incidence [5].Thus, results of all populationbased investigations, including the ones involving individuals with a TBI, have to be interpreted with caution.Most population-based studies suffer from underestimation of the "true" incidence.This underestimation is related to the absence of data for patients without contact with healthcare providers.Population-based estimates have rarely been performed in high-income countries with a low rate of penetrating TBI [6][7][8][9][10], and among the few investigations incidences have varied substantially (table 1).This variation could be partly related to the different inclusion criteria.Two populationbased studies used the same inclusion criterion: HAIS >3 [7,9].A recent investigation in Switzerland reported an incidence of 11 per 100,000 individuals per year [6].A French study in the rural region of Aquitaine estimated the incidence of severe TBI to be 17 per 100,000 individuals per year in 1996 [7].The distribution of HAIS was similar in both studies (HAIS 4: 41.1%, HAIS 5: 58.9%), including the ratio of multiple trauma (32.2%).The lower incidence in the Swiss population may be related to increased traffic safety and/or to different risk behaviours of the different populations.A population-based study in Norway showed the incidence to be 4 to 5 per 100,000 individuals per year using the ICD-10 codes, and GCS <9 [6].If only patients with a GCS <9 on scene were taken into account for the Swiss TBI population, the incidences in either study would be similar (5 per 100,000 individuals per year in Switzerland, 4 to 5 per 100,000 individuals per year in Norway).Thus, incidences in both countries should have been similar if identical inclusion criteria had been chosen.

Short-term mortality
Mortality is highest in the first few days after severe TBI.Main risk factors are the severity of the TBI and the patient's age (i.e., the greater the severity and the older the patient, the greater the mortality).The severity of additional injuries to other body parts associated with hypovolaemic shock and arterial hypotension may further contribute to mortality [11].It is possible that improvement over time in care of patients with severe TBI may have led to the lower reported mortality rates in newer studies.However, in a recent metaregression, no such trend could be observed since 1990 [12].Most large multicentre cohort studies of patients with severe TBI in high-income countries observed mortality rates between 30% and 45% (tables 2 and 3); [2,[6][7][8][9][10][13][14][15][16][17][18][19].Two studies that included only an urban population with an unknown injury severity score (ISS) reported higher death rates (47% and 49%) and one study in Australia with a lower TBI severity reported a death rate of 18%.Interestingly, the Swiss population-based study, which included the oldest population (median age of 55 years), did not report the highest mortality of all the studies.This may speak for the acceptability of the overall quality of care for patients with TBI in Switzerland.Two population-based studies that used the same inclusion criteria (i.e., HAIS >3) [7,9], reported similar death rates per HAIS category (HAIS 4: 7.7% in Aquitaine, 10.4% in Switzerland; HAIS 5: 46.0% in Aquitaine, 40.9% in Switzerland).Considering that age is one of the most important risk factors for poor post-TBI outcome [20], the similar mortality rates in the Swiss study in 2007-11 (with older patients) and the French study in 1996 (with comparably younger patients) may suggest that care and general safety measures have improved over recent years in Europe, thus potentially counterbalancing the effects of age on mortality.In other words, if mortality rates are similar in the two investigations with similar injuries but with two different age distributions, the one with the older population may have been carried out in a region with superior care.The population-based study in Norway was carried out in a similar time-period to the Swiss study.Similarities between the studies included cause distribution, age distribution, death rate (29%) and an increased early mortality rate (within 48 hours following injury).These studies sup- port previously reported evidence that mortality is linked to the severity of the injury and the age of the injured [21].
The time-point when mortality rates were measured and reported differed across studies.Calls for consensus on the definition of the time-point of the mortality assessment in neurotrauma research are justified.We propose that this time-point should be around 14 days postinjury, because at this time most patients with severe TBI are still inpatients.

Short-term disability outcome
Most often disability in patients with neurotrauma is measured with the Glasgow Outcome Scale (GOS) or, more recently, with the Glasgow Outcome Scale extended (GOSE).The scale includes the items death, vegetative state, severe disability, moderate disability and good recovery.The GOSE subclassifies the last three items into lower and upper.The use of the scale differs across studies in that some included patients after acute hospital admission and therefore included the initial high mortality group, whereas some included patients after in-hospital rehabilitation admission, and others included patients even later and without the initial high mortality.Either approach has limitations.Inclusion after acute hospital admission is usually accompanied by a high loss of follow-up because of the high incidence of refusal or unavailability of patients at follow-

Short-term outcome of elderly traumatic brain injury patients
The definition of elderly or geriatric patients lacks homogeneity.There is a certain consensus that patients with neurotrauma who are aged >65 years are considered elderly patients [24,25].Using this definition, we observed a high proportion of elderly patients with severe TBI (35%) in Switzerland.In comparison with patients ≤65 years, pa-tients >65 years showed a higher incidence (22/100,000/ year), a higher GCS at hospital admission (8) [3][4][5][6][7][8][9][10][11][12][13][14], but similar abnormal pupil reaction and similar HAIS compared with patients ≤65 years (i.e., comparable TBI severity) [9].Death rate at 14 days was higher in older patients (41%), and consciousness at 14 days assessed with the GCS was similar across the two age groups.Normed GOSE data in older Swiss patients are presently not available; however, it has been suggested that GOSE scores are lower for older than younger patients.Severe TBI has rarely been investigated in older patients, and the few studies that did this, have reported incomplete data (table 5) [8-10, 25, 26].Nevertheless, those few studies all reported: (a) a high incidence rate; (b) falls as the main trauma mechanism; and (c) a high death rate in older patients with neurotrauma.

Implications for further quality improvement programmes in Switzerland
There is a debate about the decreasing incidence of severe TBI in high-income countries [12].An argument in favour of a decrease in incidence may be improved road safety and/or other safety-related interventions.An argument against a decrease in incidence may be the increasing age of the population that is associated with more falls.The Swiss Trauma Register, which will include similar standardised data, may be able to answer this question in the near future.We suspect that the incidence of severe TBI will increase in relation to the demographic increase in the eld-  erly population, a trend that has previously been reported in Canada [27].Mortality after severe trauma and severe TBI is associated with trauma centre patient volume: the higher the volume, the lower the mortality [28].US trauma centres were considered "high volume" if they had more than 160 patients with moderate and severe TBI (expected mortality >15%) per year [29].A high-volume effect has also been observed in Taiwan, although with lower annual numbers (>33 patients with TBI needing surgery) [30].Based on these international data, some Swiss trauma centres would not be considered high-volume trauma centres and may thus have less favourable outcomes.Potentially, the increase in mortality in low volume centres is not related to healthcare per se, but to structural problems: Many patients with neurotrauma are admitted outside usual business hours and days, and thus require twenty-four/seven coverage with highly competent medical staff at night and during weekends, which may be difficult to realise in low-volume trauma centres.However, reduced staffing in trauma centres impacts upon outcome [31].
There is evidence that the speed of functional recovery depends on early and intensive neurorehabilitation [26,32,33].Discharge after inpatient and/or outpatient neurorehabilitation should be based on objective discharge criteria including cognitive functioning [34].However, at least in US rehabilitation centres, huge variability in risk-adjusted centre-specific disability outcomes can be observed by use of the Disability Rating Scale [35].Normally, high centrespecific outcome variability is associated with care quality differences, which can be improved.However, on the basis of the large number of rehabilitation centres in Switzerland where severe TBI patients are actually treated, and the fact that the time-span between discharge from acute care to rehabilitation admission is different across centres, we suspect that centre-specific outcome variability may be present in Switzerland as well.Further scientific investigations are clearly indicated to define the optimal post-acute care pathway for neurotrauma in Switzerland.It is evident that populations in high-income countries -including Switzerland -are ageing.Older age in turn is associated with an increased risk of falls from low heights [36] and thus a higher probability of neurotrauma.These older populations with severe TBI show comorbidities with illnesses that need permanent drugs including drugs with increased risk of bleeding (e.g, phenprocoumon, warfarin).In neurotrauma anticoagulation aggravates the initial lesions as a result of space-occupying intracranial bleeding.In elderly TBI patients with brain atrophy and low energy impact, however, there could be a time window for surgical and drug interventions before space-occupying haematomas occur.This hypothesis is supported by our observation that initial GCS scores in elderly patients were quite high.
In other words, if elderly patients with previously greater brain atrophy are exposed to a TBI, the bleeding will not immediately occupy space (reflected by higher GCS at scene or at hospital admission) compared with younger patients (without premorbid brain atrophy).As a result, this observed unique time window among elderly patients should be used for interventions before major cerebral injury occurs.
The prehospital and early in-hospital pathway for elderly patients with TBI should include diagnostic measures that allow optimal use of this unique opportunity.Furthermore, and related to the prolonged recovery process in elderly patients, specific in-hospital care (for instance, in intermediate care units) and adapted rehabilitation programmes should be investigated.Interdisciplinary, standardised pathways including healthcare professionals in acute care, as well as the combined work of rehabilitation specialists for geriatrics, neurology and neuropsychology may be the most promising future for this specific population group.Based on our data, we suspect that elderly patients receive different rehabilitation compared with younger patients (either no rehabilitation or qualitatively relatively lower rehabilitation).A general behavioural change in healthcare providers towards elderly trauma patients may be needed to counter lower care intensity, which in turn may so far have been associated with an increased death rate [37].
Because the incidence of severe TBI in elderly patients is increasing, and is probably even increasing at a faster rate than the ageing population itself [36], effective preventive measures are required.Given that falls at home and anticoagulation or antiaggregation are causal factors in a high number of these patients, fall prevention (e.g., strength and balance training [38]) and possibly modifications of the home [39], combined with drug adaptation [40] could be possible approaches to reducing and possibly minimising the occurrence of TBI in the elderly.However, further research is required to confirm these potential preventive strategies.

Conclusions
Severe neurotrauma is still a challenge, even in high-income countries.The probability of unfavourable short-term outcome after severe neurotrauma continues to be high and is, at least partially, related to an increased proportion of elderly patients.Healthcare systems need to provide high standard emergency and trauma care followed by comprehensive critical care and seamless transitions to timely and aggressive (neuro-)rehabilitation, in order to be able to send patients home with the best possible functional recovery.Procedural and structural improvements can be achieved with a system that includes interdisciplinary teams: All elements of care have to interact closely (i.e., efficient clinical pathways) to provide patients with the best possible outcomes.Based on our data from -and experience with -the Swiss cohort with severe TBI, quality of care in Switzerland can be further improved.

Table 1 :
Recent population-based studies of severe traumatic brain injury in adults and in high-income countries (last 20 years).
CT = computed tomography; GCS = Glasgow Coma Score; HAIS = abbreviated Head Injury Score; ICD = International Classification of Diseases; ICP = intracranial pressure; ISS = Injury Severity Score Review article: Current opinion Swiss Med Wkly.2015;145:w14177 Swiss Medical Weekly • PDF of the online version • www.smw.ch

Table 2 :
Large, multicentre cohort studies in high-income countries reporting on short-term mortality in adults with severe traumatic brain injury (Part A).

Table 3 :
Large, multicentre cohort studies in high-income countries reporting on short-term mortality in adults with severe traumatic brain injury (Part B). -based studies of neurotrauma.This refusal and unavailability after acute hospital discharge is often based on alcoholism, depression and low socioeconomic status[22].The resulting high attrition rate may skew the data distribution on the items severe disability, moderate disability and good recovery.The late inclusion of TBI survivors facilitates an improved follow-up; however, without the analysis of the initial distribution of surviving and not surviving patients, results may be skewed as well.
GCS = GlagowComa Score; IQR = interquartile range; ISS = Injury Severity Score; N = no; NA = not available; SD = standard deviation; Y = yes ° including prehospital period Review article: Current opinion Swiss Med Wkly.2015;145:w14177 Swiss Medical Weekly • PDF of the online version • www.smw.ch up in population

Table 4 :
Large, multicentre cohort studies in high-income countries reporting on short-term disability of adults with severe traumatic brain injury.

Table 5 :
Large, multicentre cohort studies in high-income countries reporting on short-term mortality in elderly adults with traumatic brain injury.
GCS = GlasgowComa Score; HAIS = abbreviated Head Injury Score; ICD = International Classification of Diseases; ISS = Injury Severity Score; NA = not available; SD = standard deviation; TBI = traumatic brain injury Review article: Current opinion Swiss Med Wkly.2015;145:w14177 Swiss Medical Weekly • PDF of the online version • www.smw.ch