Cutaneous impact location: a new tool to predict intracranial lesion among the elderly with mild traumatic brain injury?

Mild traumatic brain injury (mTBI) is the leading cause of arrivals to emergency department (ED) due to trauma in the 65-year-old population and over [1, 2]. Lately, elderly visits to ED for mTBI have increased disproportionately, mTBI in the elderly is associated with an increase in morbidity and mortality. This is a frequent reason for hospitalization and it is associated with an alteration in functional and cognitive capacities [2, 3]. Current international guidelines are consistent with the large indication of non-contrast head computed tomography scan (head CT-scan) after mTBI in patients over 65 years old, even without initial consciousness loss [4, 5]. Head CT-scan is also recommended in all patients under antiplatelets or anticoagulants after mTBI. Former cohorts from which most international guidelines were derived showed that the intracranial lesions prevalence after mTBI in the elderly was higher than in younger subjects. Regarding to this population, intracranial lesion increased until 30% [6]. These guidelines are also discussed with specific elderly epidemiological and physiological features. Older patients showing less specific sign of intracranial lesion following head trauma compared to younger patients, they also present less of signs of intracranial hypertension. As well, the Glasgow score is less sensitive and the event taking place during the mTBI is often missing [7, 8].

Recent studies conducted in ED suggested a low intracranial lesion prevalence reaching 2.2% and a rate of neurosurgery lower than 1% [9]. These data suggest that the brain CT-scan indication following mTBI in this population could be more targeted. In addition to the cost issue, head CT-scan overuse could also have an impact on the patient in terms of radiation-induced neoplasia and cataracts [10]. A prospective study containing detailed injury history and clinical findings at the ED is required to better characterize prevalence and potential risk factors of intracranial lesions.

Our objectives were to assess the prevalence of mTBI-related intracranial lesions in subjects over 65 years old admitted to the ED and to identify risk factors for intracranial lesions by recording precise mTBI anamnesis as well as clinical findings.

Between January 2019 and June 2019, 365 patients were included. The average age of the entire cohort was 86.5(±8.5) years old, ranging from 65 to 104, and 244 patients (66.8%) were female. Overall, 141 patients (38.6%) were receiving antiplatelet and 128 (35.1%) anticoagulant medication. One hundred and twenty-six patients (34.5%) presented cognitive impairment prior to ED visit.

The most common mTBI mechanism was ground-level fall (331/365, 91.8%). Among the 7 patients (1.9%) who had a road accident, 3 were drivers or passengers of a motor vehicle, 3 were cyclists and one was a pedestrian hit by a car. The mean blood alcohol level of the 7 inebriated patients was 2.2 g/l (SD = 1.1).

Furthers characteristic of the study population with univariate analysis for intracranial lesions are presented in Table 1.

Median time between mTBI and ED arrival was 120 min (IQR = 90 to 180) and 142 min (IQR = 112 to 180) between ED arrival and the performance of head CT-scan. Time of the mTBI event was unknown for 117 patients (32.1%). The history of the injury and the clinical findings at ED are displayed in Table 2 with univariate analysis according to the presence of intracranial lesions.

Of the 365 patients included, 26 (7.1%) had intracranial lesion. Compared with cutaneous frontal impact, the relative risk of intracranial lesion was 2.54 (95% CI 1.20 to 5.42) for patients with temporoparietal or occipital impact and 0.12 (95% CI 0.01 to 0.93) for patients with cutaneous facial impact or no cutaneous impact. Relative risk of intracranial lesions according to the cutaneous impact location are presented on Fig. 1. There was no statistical increase in risk of intracranial lesion in patients receiving antiplatelets (RR = 1.43; 95% CI 0.68 to 2.99) or anticoagulants (RR = 0.98; 95% CI 0.45 to 2.14). The univariate analysis of relative risk for intracranial lesions are displayed on Table 3.

Among the 26 patients with intracranial lesions, 3 (11.5%) had a focal deficit: one presented aphasia (3.8%), one homonymous hemianopsia (3.8%), and one hemiparesis (3.8%). Thirteen patients (50%) showed a subdural hematoma, 9 (34.6%) had subarachnoid hemorrhage, and 4 (15.4%) an intraparenchymal hematoma. No patient required neurosurgical intervention. Eighteen patients (69.2%) with intracranial lesion were hospitalized and 2 (7.7%) were transferred to the intensive care unit. Warfarin effects were reversed by prothrombin complex concentrate and vitamin K in 5 patients (19.2%). One patient treated with antivitamin K had an extensive intraparenchymal hematoma despite reversion and died 2 days after admission.

Overall, 106 patients (29.0%) were hospitalized and 141 (38.6%) had an isolated mTBI.

Patients’ characteristics, injury history, clinical findings and outcomes according to cutaneous impact location according to the cutaneous impact location are presented in the Table 4.

Our mTBI-related intracranial lesion prevalence of 7.2% was consistent with recent studies performed in the elderly [9, 13]. This lower prevalence found in EDs recent studies may have several causes. Firstly, current guidelines recommend a head CT-scan in most patients over 65 years old after mTBI. Thus, this large indication may decrease intracranial lesion prevalence among this population. Secondly, according to the mTBI definition of 1993, injury mechanisms were not only direct head strikes but also acceleration/deceleration movements without any direct external trauma of the head [14]. In accordance with literature our findings showed that ground-level fall was the most common cause of mTBI [1, 13]. Therefore, the amount of mTBI kinetics involved in ground level fall in the elderly is less than in younger populations (ie: falls from over one meter or motor vehicle accidents). Thus, ground-level fall, especially without head impact, may not have enough kinetics to induce intracranial lesion in the elderly.

Compared with cutaneous frontal impact, temporoparietal and occipital impact had an intracranial lesion relative risk of 2.54 (95% CI 1.19 to 5.42). To our knowledge, the cutaneous impact location in the context of mTBI in elderly has not been assessed. Several hypotheses could explain these results. The musculature reduction of the trunk and neck, to which ageing of the individual contributes, could increase the force of temporoparietal or occipital impact in the event of ground-level fall [4, 15]. Conversely, in case of frontal impact, some of this kinetic energy may be reduced by postural adaptation reflexes of the upper limbs. In addition, frontal impact may have less risk of intracranial lesion due to the fact the frontal bone is thicker than temporal, parietal and occipital bones, [16]. Moreover, in our study and according to literature, 50% of intracranial injuries were subdural hemorrhages [4, 13]. The pathophysiology of this hemorrhage as Miller JD et al. reminds us, involves a direct trauma with low kinetic energy able to affect the venous network of the dura mater and the arachnoid [17]. Furthermore, several hypotheses involving the vulnerability of vascular tissue and age-related white matter alterations have been put forward to explain the susceptibility of old patients to hemorrhage subsequent to a direct trauma with low kinetic energy [18, 19]. The cutaneous impact location in a context of mild traumatic brain injury has already been studied among children. In a recent large study of 3866 children younger than 17 years old, Burns EC et al., [20] showed that temporal/parietal and occipital impact location had significantly greater odds of intracranial lesions than other impact locations. Furthermore, these odds were greatest in children aged from 0 to 6 months. The mechanisms involved in youngest children and frailty elderly might be the same: despite the low kinetic energy, the absence of postural adaptation reflexes of the upper limbs may trigger intracranial lesions especially in case of temporal/parietal and occipital impact. Taken together, these findings could explain why patients with cutaneous temporoparietal or occipital impact had a higher risk of intracranial lesion than those with cutaneous frontal impact. Thus, to better apprehend the relevance of head CT-scan in the context of ground-level fall with mTBI among the elderly, risk groups of intracranial lesion may be identified regarding the cutaneous impact location as follow: low risk groups (with facial or no cutaneous impact), medium risk (cutaneous frontal impact) and high risk (cutaneous temporoparietal and occipital impact). This objective sign may help us to better assess the risk of intracranial lesions and better target head CT-scan indication.

In spite of the prospective nature of our study, it was not possible to determine the fall cause in more than one third of the cases. Timler et al. showed in their retrospective study that the mTBI mechanism was not identified in 23.6% of the cases [13]. Furthermore, in our study, symptoms after mTBI were unknown in 13.5% of the cases. Not only the symptom sensitivity for intracranial lesion diagnosis are low, but the occurrence of those are also often unknown by the physician in charge of the patient [7, 13, 21]. These results strengthen the hypothesis of Papa et al. suggesting that the term mTBI is sometimes misused in the elderly [7]. However, clinical findings as headaches (RR = 5.49; 95% CI 2.70–11.17), vomiting and focal neurological signs (RR = 15.74; 95% CI 10.59–23.37) were significantly associated with the development of intracranial lesions. Regarding the Glasgow score, a score of 14 was significantly associated with intracranial lesion (RR = 3.34; 95% CI 1.51–7.39). Literature shows divergent findings; on the one hand, some studies discredited Glasgow score in elderly by showing less sensitive than in younger patients to detect intracranial lesion [22]. On the other hand, some studies suggested that a Glasgow score limit of 14 instead of 13 may improve sensitivity to predict poor outcome in elderly [23].

Ongoing treatments with antiplatelets or anticoagulants were not associated with a significant increase in the risk of intracranial lesion. Recent studies suggest that the use of antiplatelet alone does not increase the risk of mTBI-related intracranial lesion [24]. About anticoagulant, especially in ground-level fall context, recent studies suggest an increasing risk of intracranial lesion but not a worsen morbidity or mortality [25, 26]. These results should be analyzed carefully, despite the absence of multivariate analysis performed in our study due to the small sample size of it. Moreover, even if ICL prevalence was similar we still notice a higher prevalence in the Aspirin group with 9.8% (11/112) and 8.3% in the Warfarin group with 8.3% (5/60) than the overall population (7.2%, 26/365).

Contrary to our initial hypothesis, patients’ frailty assessed by the Clinical Frailty Scale was not associated with a significant increase in the risk of developing intracranial lesion. However, the frailty measured by this scale was high in our population (mean 4.6/7; SD 1.6). This finding is consistent with literature suggesting that frailty is associated with an high incidence of mTBI [15]. Nonetheless, getting up with assistance was associated with great risk of intracranial lesion (RR = 2.66; 95% IC 1.09–6.47). This may reflect declined functional abilities in these patients who might have impaired postural adaptive reflexes.

Additionally, we identified the same risk factors for intracranial lesion classically identified in older patients such as female (RR = 2.76; 95% CI 0.98–7.83) [21]. The age population average was high. Due to this Fournier et al. suggest increasing the age limit from 65 to 75 years-old. The fact of adjusting the age limit to 75 year-old in the Canadian head CT rule may reduce head CT-scans performed of 25% without any intracranial lesion missed [27]. In their retrospective cohort, Riccardi et al. even suggested to increase this limit to 80 years old [9].

Hospitalization rate in our cohort was low (29%). This is consistent with literature and reinforces recent hypotheses of potentially avoidable and/or inappropriate use of emergency services, particularly for the NH residents counting for 39.5% of our cohort [13]. The following term: ‘potentially avoidable and/or inappropriate use of emergency services’ has no consensual definition, but may represent almost half of the transfers to the ED from nursing homes [28]. Better target intracranial lesions risk after mTBI among this population may reduce the need for ED visits. Our study further strengthens this theory since 38.7% of the patients admitted to the ED had an isolated mTBI, without wounds requiring suture or any medical and/or trauma emergency care.

The strength of our prospective study relies on the anamnesis accuracy and clinical findings in the ED since it was a real-life observational study. This prospective inclusion is a true reflection of the semiological survey conducted by ED practitioners. The main limitation of our study was the population’s size. Its main objective was to establish the prevalence of intracranial lesion and it may lack power to identify certain risk factors. Furthermore, the number of cases was not sufficient to conduct a multivariate analysis. In addition, there was no patient follow-up while according to literature, there might be an increased risk of delayed hemorrhage especially in patients treated with anticoagulants [29]. Moreover, the readmission rate to hospital within 1 month after an ED visit for mTBI-related intracranial lesions might be high, particularly for the elderly with fall-related mTBI [30, 31].

To sum up, the prevalence of mTBI-related intracranial lesion in elderly patients admitted to the ED was similar to younger patients. This finding might be partly explained by the low kinetic energy and the pathophysiology of hemorrhage in the context of ground-level fall. These results suggest that the cutaneous impact location may help to identify risk for mTBI-related intracranial lesions in older patients. A prospective, multicenter ED study would be useful to confirm these potential risk factors.