SPINE TRAUMA EPIDEMIOLOGICAL PROFILE IN A TERTIARY NEUROSURGERY HOSPITAL IN SOUTH BRAZIL PERFIL EPIDEMIOLÓGICO

Objective: The aim of this paper is to analyze the epidemiological profile of a large series of spine trauma victims in the Southernmost state of Brazil. Methods: A retrospective study including spine trauma patients was performed at a tertiary hospital from January 1st, 2013 to December 31st, 2018. The variables analyzed include demographic data, information related to the trauma (etiology, trauma mechanism, type of spine injury, number of vertebrae involved, vertebral segment involved), neurological status at hospital admission (Frankel scale), treatment performed and the outcome (number of days in hospital, neurological outcome, and mortality). Results: A total of 808 patients were included. The mean age was 47.9 (±19.0), and the majority were male and Caucasian. The most frequent etiology was falls from height (N=508; 62.9%) followed by traffic accidents (N=185; 22.9%). The thoracolumbar segment was the spinal segment most frequently affected, occurring in 401 (52.1%) patients, followed by the cervical, thoracic and lumbar segments. The incidence of SCI was 16.7%. Non-operative treatment was indicated in 510 (63.1%) patients. Conclusion: The authors presented the largest epidemiological profile regarding spine trauma in Latin America, analyzing a total of 808 patients, which represents an incidence of 134.6 cases/year. This paper fills a gap in the medical literature regarding the epidemiological profile of this disease in Latin America. Level of evidence II; Prognostic study.


INTRODUCTION
Spine trauma is defined as any traumatic lesion to the spine secondary to an external etiology, 1,2 which may be associated with spinal cord injury (SCI). 3 Spine trauma patients have a high rate of morbidity and when there is associated SCI, most patients have motor, sensitive, and autonomic impairment. [4][5][6] Men are three to four times more likely to be affected than women, corresponding to 76-88% of patients, with a peak age of between 34 and 58 years. [7][8][9] Incidence of spine trauma and SCI varies around the world, with markedly different statistics between developing and non-developed countries. 10,11 In Asia Pacific, Taiwan has the highest rate , with 40.2 patients per million people with spine trauma, 12 followed by China and Saudi Arabia, which have rates of 23.7 and 38 patients per million people, respectively. [13][14][15] The incidence in the United States varies from 25 to 59 patients per million, with a national average of 40 cases per million people. [16][17][18] The incidence is lower in Europe, varying from 10 to 19 cases per million people in countries such as Denmark, Germany, France, Turkey and Sweden. Portugal has the highest rate of spinal trauma cases (58 per million people) when compared to the other European countries. 9,13,[19][20][21][22][23] A systematic review published in 2013 reported the worldwide publications on spine trauma. However, no experience from Latin America was reported. 17 The incidence of spine trauma in different regions of Brazil seems to indicate an approximate incidence of 21 cases per million people. 7,10 The Brazilian data had small samples and lacked essential epidemiological information. The objective of this study is to analyze the epidemiological profile of patients with spine trauma and SCI in Rio Grande do Sul, the Southernmost state of Brazil.

Study design
The authors performed a retrospective study of traumatic spine injury patients admitted to the Neurosurgery Institute of Cristo Redentor Hospital from January 1 st , 2013 to December 31 st , 2018. The study was approved by the Ethics in Research Committee of the hospital where it was conducted (CAAE: 75903717.3.0000.5530) under number 3.744.697. Since it is a retrospective study, for which data was collected without identification of the patients and without any impact on the treatment, the need for the patients to sign an informed consent form was dispensed with.

Eligibility of patients
The inclusion criteria were patients with traumatic spine injury. Patients admitted with spine infection disease (spondylodiscitis), degenerative spine cases, and pathological or osteoporotic fractures, were excluded.

Variables analyzed
The variables analyzed include demographic data (age, sex, race), information about the trauma (etiology, trauma mechanism, type of spine injury, number or vertebrae involved, vertebral segment involved), neurological status at hospital admission, the treatment performed (operative and non-operative) and the outcome (number of days in hospital, neurological outcome, and mortality). Neurological status at hospital admission and discharge was evaluated based on the Frankel scale. The outcome was analyzed based on mortality rate and the Frankel scale at hospital discharge. All variables were collected retrospectively from the hospital's computer records.
The spine segment injured was first classified as cervical, thoracic, thoracolumbar or lumbar. The cervical lesions were further stratified as C1, C2 or Cervical Subaxial Trauma (C3 to C7). Thoracic fractures from T1 to T10 were classified as Thoracic Spine Traumas. Fractures between T11 and L3 were classified as Thoracolumbar Spine Traumas. Fractures L3 to L5 were classified as Lumbar Spine Traumas.
Specific classifications were used to report the fractures according to the vertebrae affected. C1 fractures were reported according to the Jefferson classification as posterior arch, anterior arch, anterior and posterior arch, and lateral mass. 24 Odontoid fractures were classified according to the Anderson and D'Alonzo classification. 25 Fractures involving the articular pars of C2 were stratified based on the Levine and Edwards classifications. 26 CT scans of Subaxial cervical, Thoracic, Thoracolumbar and Lumbar spine fractures were reviewed and classified using the AO Spine classification system. 27,28 Statistical analysis Data were collected using the software program Microsoft Excel 2019. The statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS, version 22.0). The categorical variables were presented in proportion. The numerical variables were submitted to the Kolmogorov-Smirnov Test in order to evaluate their parametricity, and presented as mean and standard deviation if parametric, or median and interquartile range if non-parametric.

Treatment related data
Operative treatment was performed in 298 (36.9%) cases and non-operative treatment in 510 (63.1%) patients. The most common non-operative treatment was the use of an orthosis. The median length of hospital stay was 10 (4-37.75) days. Among the operative patients, the median hospital stay was 20 (11-.2), 9 (+8.14) before surgery and 5 (3-12) days after surgery. Among the non-operative patients, the median hospital stay was 5 (3-20.2) ( Table 1).

Neurological status and outcome
The majority of the cases had normal neurological exam results at hospital admission (83.3%). Complete spinal cord lesion was observed in 51 (6.3%) patients. The cases of incomplete spinal cord injury were classified according to Frankel B (N=9; 1.1%), Frankel C (N=16; 2.0%), and Frankel D (N=26; 3.2%).
The mortality rate was 1.9% (15 patients

DISCUSSION
The authors present the largest epidemiological study of patients with spine trauma performed in Latin America. Additionally, the authors evaluated the incidence and the characteristic of trauma according to different spine segments. The information presented here will fill a gap, because the epidemiological information was based on findings of European and North American countries, which do not necessarily have the same patient profile.

Distribution by age and sex
The present work showed that spine trauma mostly occurred in young males, corroborating previously published articles. [7][8][9]12,29 A study published in the same state found similar distribution by age and sex. 30 A progressive increase in rates of spine trauma was observed in patients aged over 65 years old; from 4.8% to 12.1%. 11 Our study reinforces this finding, as one fifth of the spine trauma patients in our study were older than 65 years.

Trauma etiology
A previous study reports that traffic accidents and violence, especially firearm injury, were the main spine trauma mechanisms in Brazil. 31 Our study showed that fall from height was the main etiological mechanism, particularly from higher than 2 meters, which corroborates data from China, the USA and Spain. 13,15,23,32 Traffic accidents were the second most frequent mechanism in our series, especially car accidents, which were also frequent in many other studies. 21,32 Unlike Scandinavian studies, where there was a high percentage of spine trauma during sports practice, 9,21,22 only 3 of the patients in our study had this trauma mechanism.

Trauma mechanism
Vertebral fracture was the most frequent injury after spine trauma. The incidence of patients with multiple fractures in this series is in accordance with the literature. 5 We demonstrated that 70% of multiple fractures occurred in adjacent vertebrae and 30% in distant vertebrae.

Cervical spine trauma
Cervical spine fractures are commonly classified according to the vertebrae involved, as the mechanism, neurological status, fracture classification and treatment are different. Our study showed that 13.4% of cervical fractures occurred in C1. In general, atlas (C1) fractures represent 2% to 13% of all cervical spine injuries. 33 Fractures of the second cervical vertebra (axis) comprise approximately one-third of cervical spine fractures, 34 a percentage close to the 36% found in the present sample. The peculiar anatomy of the second cervical vertebra is different from that of all other levels, with the odontoid process and an elongated pars interarticularis between the atlantoaxial joint and the C2-C3 joint. For this reason, fracture of the axis is classified differently if it occurs at the odontoid process or the pars interarticularis. C2 fractures involved mainly the odontoid process in 79% of cases and the pars interarticularis in 21% of cases. Fifty percent of cervical fractures involved the subaxial segment, which corroborates the literature as the most common among cervical trauma. 35

Thoracic spine trauma
The thoracic spine is functionally rigid due to coronally oriented facet joints, thin intervertebral discs and the ribcage. It requires great amounts of energy to produce fractures and dislocations. 36 According to the literature, of the fractures involving the thoracic, thoracolumbar or lumbar segments, 25-40% occur in the thoracic segment. 37 The present study found only 20% of fractures occurring in the thoracic segment (T1-T10).

Thoracolumbar spine trauma
The thoracolumbar junction (T10-L2) is uniquely positioned between the rigid thoracic spine and the mobile lumbar spine. This transition from the less mobile thoracic spine, with its associated ribs and sternum, to the more dynamic lumbar spine results in the thoracolumbar region being subjected to significant biomechanical stress. 37 Fractures of the thoracolumbar region are the most common injuries of the vertebral column. 36 Our study showed that fractures between T11-L2 occurred in 52.1% of the whole sample, and 68.7% of the cases when cervical segment was excluded.

Lumbar spine trauma
The lumbar spine is flexible due to the thicker intervertebral discs, sagittal orientation of the facet joints and the absence of the rib cage. The lumbar segment is not frequently involved in spine trauma, representing 10-14% of cases, according to the literature, 37 a range that corroborates the 11.14% found in this sample.

Neurological status
Approximately, 15 to 20% of patients with spinal trauma have associated neurological damage. 38 Based on our study, 17% of patients have neurological impairment. Most of them have complete SCI. Different findings were described in a study published previously, in the same city. 30

Treatment and outcome
Operative therapy was performed in 40% of the patients, which demonstrated the high proportion of severe cases in our sample. However, the mortality rate of 2% was low. A greater number of patients had normal neurological function at hospital discharge or improved neurologic function, when compared with their status at admission.

Strengths and Limitations
The main strengths of this study are the large sample size and the fact that several variables are analyzed, allowing a broad evaluation of different aspects of the pathology, radiology, and patients. This allows an overall understanding of the importance of each variable in the scenario of the spine trauma patient. Some limitations must also be recognized, particularly the fact that this study uses only retrospective data. Despite the retrospective study design, all the variables analyzed were successfully retrieved from the hospital digital system. The results came from experience at a single center, which may represent a selection bias, as it is located within a tertiary hospital. Therefore, minor traumas or fractures may not have been referred to this service, or may have been discharged from the emergency room for treatment at the primary care level.

CONCLUSION
Adult males are the subgroup at higher risk of spine trauma, especially as a result of falls and traffic accidents. The thoracolumbar segment is the one most involved and, in accordance with the other segments, the majority of patients have stable fractures without neurological involvement.