Tr a u m at i c injuries are major contributors to death

Transcription

1 J Neurosurg Spine 12: , 2010 Epidemiological trends of spine and spinal cord injuries in the largest Canadian adult trauma center from 1986 to 2006 Clinical article Fa r h a d Pi r o u z m a n d, M.D., M.Sc., F.R.C.S.C. Division of Neurosurgery, Sunnybrook Health Sciences Centre, University of Toronto, Ontario, Canada Object. In this study the author documents the epidemiology of spine and spinal cord injuries (SCIs) over 2 decades at the largest Level I adult trauma center in Canada. He describes the current state of spine injuries (SIs), their changing patterns over the years, and the relative distribution of different demographic factors in a defined group of trauma patients. Methods. Data on all trauma patients admitted to Sunnybrook Health Sciences Centre between 1986 and 2006 were collected from the Sunnybrook Trauma Registry Database. Aggregate data on SIs and SCIs, including demographic information, etiology, severity of injuries (injury severity score [ISS]), and associated injuries, were recorded. The data were analyzed in a main category of spinal fracture and/or dislocation with or without SI and in two subgroups of patients with SIs, one encompassing all forms of SCIs and the other including only complete SCIs (CSCIs). Collected data were evaluated using univariate techniques to depict the trend of variables over the years. The number of deaths per year and the length of stay (LOS) were used as crude measures of outcome. Several multivariate analysis techniques, including Poisson regression, were used to model the frequency of death and LOS as functions of various trauma variables. Results. There were 12,192 trauma patients in the study period with 23.2% having SIs, 5.4% having SCIs, and 3% having CSCIs. The SCIs constituted 23.3% of all SIs. The respective characteristics of the SI, SCI, and CSCI groups were as follows: median age 36, 33, and 30 years; median LOS 18, 27, and 29 days; median ISS 29, 30, and 34; female sex ratio 34, 24, and 23%; and case fatality rate 16.7, 16.6, and 21%. Seventy-nine percent of patients had associated head injuries; conversely, 24% of patients with head injuries had SIs. The mean admission age of patients increased by ~ 10 years over the study period, from the early 30s to the early 40s. The relative incidence of SIs remained stable at ~ 23%, but the incidence of SCIs decreased ~ 40% over time to 4.5%. Motor vehicle accidents remained the principal etiology of trauma, although falling and violence became more frequent contributors of SIs. The average annual ISS remained stable over time, but the LOS was reduced by 50% in both the SI and SCI groups. Age, ISS, and SCIs were associated with a longer LOS. The case fatality rate remained relatively unchanged over time. Poisson analysis suggested that the presence of an SCI does not change the case fatality rate. Conclusions. Data in this analysis will provide useful information to guide future studies on changing SI patterns, possible etiologies, and efficient resource allocation for the management of these diseases. (DOI: / SPINE09435) Ke y Wo r d s spinal fracture spinal cord injury epidemiology Tr a u m at i c injuries are major contributors to death and disability in the population. The presence of a spinal fracture or dislocation with or without SCI requires different therapeutic interventions and involves distinct potential long-term consequences. One of the first steps in preventing injury is the collection and analysis of data to help define the problem and identify possible risk factors in various populations. 5,11 A limited number of contemporary studies have described rates and patterns of SIs in a large population. Most of these studies have been restricted to patients with Abbreviations used in this paper: ISS = Injury Severity Score; LOS = length of stay; SCI = spinal cord injury; SI = spinal injury. SCIs. In Canada, the annual incidence of SCIs resulting in permanent paralysis or neurological deficit is ~ 35 cases per million persons. 18 The same estimate in the US is ~ 40 cases per million individuals. 1,15 In a populationbased study from Scotland ~ 5.3% of head injury patients had associated cervical spine injuries. 7 In another study the prevalence of thoracolumbar spine injuries was estimated at 6.3%. 9 In the present study an epidemiological analysis of all SIs was conducted at a large trauma center for a period extending over ~ 2 decades to determine the incidence of This article contains some figures that are displayed in color on line but in black and white in the print edition. J Neurosurg: Spine / Volume 12 / February

2 F. Pirouzmand traumatic SIs; to describe the occurrence of such injuries by year, age, sex, duration of hospital stay, external cause of injury, severity of injury, and presence of SCI; and to identify subgroups at high risk. This surveillance analysis could provide direction for etiological research, resource allocation, and prevention efforts in this common group of trauma patients. Methods Study Design and Patient Population This is a retrospective evaluation of prospectively collected data from a consecutive series of trauma patients who presented to a major Level I trauma center over a period of 2 decades. Data were recorded in the Sunnybrook Trauma Database. Hospital records were reviewed for all trauma patients admitted to our institution between 1986 and Inclusion Criteria and Definitions of Trauma, SI, and SCI The ISS scale is an international scoring system that calculates the severity of one s injuries and predicts death. The higher the ISS, the more severe is the trauma. Adult patients with major trauma and an ISS > 12, who had an appropriate External Cause of Injury Code, and who met one of the following criteria were included in this study: admitted to the hospital, treated in the emergency department (not admitted), or died in the emergency department after treatment was initiated (not admitted). Spine injury was defined as any spinal fracture and/or dislocation with or without SCI. Within the SI group were two subgroups: all forms of SCIs, SCI group ; and CSCIs, CSCI group. Source of Data Sunnybrook Health Sciences Centre is the regional trauma center serving metropolitan Toronto and surrounding areas within the province of Ontario. This hospital is the leading academic adult trauma center in Canada with extensive experience, infrastructure, and appropriate government funding to maintain a high level of care for patients with significant multitrauma. For many years this center has been maintaining a comprehensive registry data set on various aspects of trauma patients, including detailed demographic information, injury severity data, prehospital and hospital care, and patient outcomes. A standardized trauma assessment form is used in data entry. All injuries are identified and coded. This comprehensive data set has been completely registered and electronically maintained in a coded computerized database since Therefore, that year was used as the start date for the present study. All trauma diagnoses with ISSs > 12 along with associated spinal fracture and/or dislocation were included in the analysis. There were no exclusion criteria. The subgroup of patients with spinal fracture and/or dislocation with SCI was included, but the patients with cord injury without obvious spinal column injury were not part of this study. Aggregate data on all the trauma patients were collected. Patients were divided into 3 age-group categories (< 45, 46 65, and > 65 years) to calculate age-specific incidence. The admission date was divided into three 7-year injury brackets ( , , and ) to assess any changes in trauma characteristics in relationship to time. This time division was arbitrary. The ISS was also divided into low (< 25) and high ( 25) categories for the purpose of outcome analysis. To explore the effectiveness of care in trauma patients over the study period, two crude measures of outcome were selected. The first, LOS, was used as a surrogate to predict the changing efficacy of treatment. The second factor was the case fatality rate. To explore any changing pattern in case fatality, the age group, severity of injuries, associated SCI, and admission time categories were analyzed as independent variables over the study period. Statistical Analysis Univariate analyses of age, sex, etiology, level and extent of injury, ISS, associated head injuries, SCIs, time from injury to hospital arrival, LOS, and death rate were conducted. Most of these factors were plotted against study periods to assess the temporal trend of each variable. The number of deaths per year and LOS were used as crude measures of outcome assessment. Multivariate analysis techniques were applied to model the frequency of death and LOS as a function of the time of diagnosis and other trauma variables. We used a log-linear model assuming a Poisson distribution to model the case fatality rate as a function of the admission time period, age group, ISS category, and presence of SCI for all trauma patients. Multiple regression analysis with a backward elimination technique was performed to demonstrate any association (or prediction) between LOS and other independent trauma variables (that is, sex, age, admission date group, ISS, SCI, and mechanism). Commercially available software (SPSS 13, SPSS, Inc.) was used for the analysis. Results Demographic Variables Table 1 summarizes the distribution of sex, age, ISS, spinal injury level, SCI, time from injury to hospital arrival, LOS, and death rate among 12,192 patients during the study period. The average annual percentage of spinal trauma was 23% of all traumas. Sixty-six percent of the patients with SIs were male. Spinal injuries were distributed along the spine as follows: cervical 29%, thoracic 21%, and lumbosacral 50%. Multiple spinal levels were involved in 20% of patients. Among the admitted patients, 79% had associated head injuries; conversely, 24% of patients with head injuries had SIs. Patient Age. The age distribution in the SI group had a peak at early adulthood (18 20 years) with a steady decline as age increased (Fig. 1 upper). The median ages in the SI, SCI, and CSCI groups were 36, 33, and 30 years, 132 J Neurosurg: Spine / Volume 12 / February 2010

3 Epidemiological trends of spine and spinal cord injuries TABLE 1: Summary of characteristics in patients with SIs between 1986 and 2006* Parameter Value no. of cases (relative %) (10.9) (31.4) (46.9) frequency (% relative to all trauma cases) SI 23.2 SCI 5.4 CSCI 3 median patient age in yrs (range) SI 36 (16 96) SCI 33 (16 92) CSCI 30 (16 87) female sex (%) SI 34 SCI 24 CSCI 23 median ISS (range) SI 29 (16 75) SCI 30 (16 75) CSCI 34 (17 75) etiology for SI (%) MVA 65.5 falling 12.9 recreational 4.8 violence 10.3 fire 2.2 other 4.1 etiology for SCI (%) MVA 57 falling 21.2 recreational 9.5 violence 8.2 fire other 4.1 level (%) cervical 29 thoracic 21 lumbo-sacral 50 multiple levels 20 median time to hospital in hrs (range) SI 3.3 ( ) SCI 3.2 ( ) CSCI 3.3 ( ) respectively. Overall, there was an ~ 10-year increase in the age of patients presenting with SI over the 2 decades of the study, that is, from the early 30s to the early 40s (Fig. 1 lower). J Neurosurg: Spine / Volume 12 / February 2010 (continued) TABLE 1: Summary of characteristics in patients with SIs between 1986 and 2006* (continued) Parameter Value median LOS in days (range) SI in (0 517) SI in (0 124) SI for all years 18 (0 778) SCI for all years 27 (0 648) CSCI for all years 29 (0 648) average case fatality rate SI 16.7 SCI 16.6 CSCI 21 * MVA = motor vehicle accident. Frequency. The frequency of trauma patients with SI and their relative annual frequency among all trauma patients are shown in Fig. 2. On average, the SI group made up ~ 23% of all trauma patients, and this rate remained relatively stable during the study period. The annual relative frequency of SCIs in all trauma patients is shown in Fig. 3. This frequency decreased from ~ 7.5 to 4.5% over the 2 decades. Etiology. The causes of traumatic SIs are summarized in Table 1, and the relative frequency of each etiological mechanism in each year of the study and their overall trends are shown in Fig. 4. Motor vehicle accidents accounted for the greatest number of SIs, ~ 66% of all injuries. Falling was the second most common etiology, closely followed by violence. Both of these causes increased in relative frequency in the more recent years. Falling became a more likely predisposing factor for SCIs as compared with SIs. Age is known to be associated with different trauma etiologies; therefore, the mean age was calculated for all trauma etiology groups (Table 2). Falling was associated with the oldest mean age, and violence with the youngest. Statistical analysis revealed a strong association between age and almost all of the defined etiology groups in the study (statistics not shown). Transfer Time. The efficiency in transferring patients with SIs from the injury scene to the tertiary care center trend is shown in Fig. 5. The median time was 3.3 hours for patients with SIs and 3.2 hours for those with SCIs. Over the 2 decades of the study there was an ~ 1 hour reduction in transfer time in both the SI and SCI subgroups, from 3.5 to 2.5 hours. Injury Severity Score. The median ISS for all patients with SIs was 29 for the study period. The median scores were 30 in the SCI group and 34 in the CSCI group. There was no significant change in the ISS in either the SI or SCI group during the study period (Fig. 6). Patient Outcomes Length of Stay. The median LOS in all patients with 133

4 F. Pirouzmand Fig. 2. Upper: Graph of the frequency of trauma patients with SIs from 1986 to Lower: Relative frequency of all SIs from 1986 to Fig. 1. Upper: Bar graph showing the distribution of age in all spinal fractures and/or dislocations from 1986 to Lower: Graph showing the mean age trend for patients with SIs, SCIs, or CSCIs from 1986 to Patients are generally becoming a decade older over this period. Diamonds indicate the mean ages of patients with SIs; squares, CSCIs; and triangles, SCIs. SIs changed from 24 to 14 days over the study period. In patients with SCIs, the median stay was 27 days; 29 days for those with CSCIs. In Fig. 7 the changing pattern of the LOS over time has been plotted, revealing an ~ 40% overall reduction during the study period. For unknown reasons there was a delayed peak in the LOS in all patient groups in Multiple regression analysis with a backward elimination technique was used to show any association (or prediction) between LOS and other independent trauma variables (Table 3). This analysis revealed a significant association between LOS and ISS, and SCI and age. Sex was not a significant contributing factor. Case Fatality Rate. The average case fatality rate for the entire study period was 16.7% for all trauma patients with SIs. There was no major change in this rate over 2 decades, as is evident by the line projection of the case fatality rate (Fig. 8). The average rate was 16.6% for patients with SCIs and 21% for patients with CSCIs. Note that there was a significant range for patients with SCIs in contrast to the more uniform case fatality rate in all patients with SIs. The reason for this difference is not clear based on our data. The results of case fatality rate from the Poisson regression model are shown in Table 4. Younger age groups were ~ 2 3 times less likely to die of similar injuries. The death rate declined minimally during the three consecutive time categories of this study ( , , and ), although this decrease was not evident in a simple linear projection of the case fatality rate in Fig. 8. The most significant effect was evident in the ISS category, with a higher ISS category increasing the risk of dying ~ 3.5-fold. Contrary to popular belief, however, the presence of an SCI is not associated with a higher mortality rate. Discussion The purpose of this study was to perform a long-term assessment of patients with spinal trauma in a tertiary trauma center utilizing a well-maintained prospective database. It was aimed at providing two types of information. The descriptive part documents the information on multiple characteristics of spinal trauma and the degree and pattern of associated injuries. The long-term temporal trend of various important biological or clinical variables 134 J Neurosurg: Spine / Volume 12 / February 2010

5 Epidemiological trends of spine and spinal cord injuries Fig. 3. Relative frequency of patients with SCIs from 1986 to Note the gradual decline. was documented as well. The analytical part explores the relationship among various potential predictor variables and crude outcome measures for trauma (for example, LOS and case fatality rate). This study is unique in both the breadth of aggregated information and the period of data collection over 2 decades. There have been many epidemiological studies on specific aspects of spinal injuries, mainly SCIs, 2,3,10,16,17,21 but there are very limited epidemiological data on SIs overall. Another fundamental assumption in the literature is that the devastating nature of SCIs which is true is somehow associated with high(er) mortality rates. This assertion was not proven in the current study. There has been a well-established comprehensive national database on SCI in the US since ,22 Using data from model SCI systems throughout the US, there have been many studies on the demographics and trends in the injury data of persons with SCI. 10,16 Similar to findings in the present study, the average age and the relative frequency of violence and falling as causes have been shown to be increasing. 16 There have also been studies on the racial and ethnic differences in outcomes following SCI, focused primarily on subjective outcomes in areas in which racial and ethnic minorities have traditionally been disadvantaged in terms of employment and financial status. 13 In Canada, several studies have shown the demographics and short-term trends in patients with SCIs. 8,19,20,24 The authors of one population-based analysis attempted to calculate the rates of and other demographic data on traumatic SCI in the province of Ontario over a 5-year period. Two thousand three hundred eighty-five hospital Fig. 4. Relative frequency of different mechanisms of trauma in patients with SIs for each year during 1986 to J Neurosurg: Spine / Volume 12 / February

6 F. Pirouzmand Fig. 5. The median elapsed time (hours) from the trauma scene to the hospital from 1986 to 2006 for both SIs (dashed line) and SCIs (continuous line). admissions were included in the study. Annual age-standardized rates declined from a maximum of 46.2 to 37.2 hospitalizations per 1 million persons. The rates of affected males declined over the study period, whereas the rates of affected females remained stable. Leading causes of injury included unintentional falls (43.2%) and transport injuries (42.8%). 19 Tator et al. 24 have compared two groups of injured patients in the same geographic area: one between 1947 and 1973, and the second between 1974 and Their results showed that those in the more recently injured group were younger, arrived at the hospital sooner, had less severe cord injuries, and had higher frequencies of motor vehicle, sports, and recreational accidents but fewer work-related injuries. Improvements in critical care medicine for SCIs may explain the marked decline in the short-term mortality rate. And although there have certainly been improvements in long-term rehabilitative care, their effect in enhancing the lifespan of persons with SCI seems overstated. 6,14,23 In one study DeVivo et al. 6 have shown that mortality rates after the first anniversary of injury which had been declining from 1973 to 1992 actually increased 33% in persons injured between 1993 and 1998, relative to those injured between 1988 and To determine whether there have been improvements in survival after SCI over time, Strauss et al. 23 have studied model SCI systems and hospital SCI units across the US from 1973 to They found that over the past 3 decades there has been a 40% decline in the mortality rate during the critical first 2 years after injury. However, the decline in death over time after that 2-year period was small and not statistically significant. Summary and Evaluation of Study Findings Admissions to our institution for traumatic SI have tripled over the past 21 years this while the main catchment area population increased by ~ 9.1% from 1991 to The definition of trauma for the purposes of referral and admission has not changed in this period. Possible explanations for the rise are more trauma patients per population, increased hospital resources coinciding with more centralized care, and referral from other hospitals or TABLE 2: Distribution of causal factors in relation to age Etiology No. of Patients Mean Age 95% CI for Mean Lower Bound Upper Bound MVA fall violance recreational fire other J Neurosurg: Spine / Volume 12 / February 2010

7 Epidemiological trends of spine and spinal cord injuries Fig. 6. The median ISS in patients with SIs (dashed line) and those with SCIs (continuous line) for each year from 1986 to 2006, and their trends over the study period. Note the stability over time. Fig. 7. The median LOS for patients with SIs (dashed line) and those with SCIs (continuous line) and their general declining trend from 1986 to J Neurosurg: Spine / Volume 12 / February

8 F. Pirouzmand TABLE 3: Relationship of various independent variables in predicting LOS in a multivariate analysis* Independent Variable LOS (standardized coefficient) age 0.04 sex NS ISS 0.15 SCI 0.1 * Using a backward elimination technique. Only significant values with p < 0.05 are shown in the table. Abbreviation: NS = not significant. changing criteria for trauma admission. Other possibilities need further investigation. Regardless of the cause, a rise in admissions increases workloads at the hospital. Patients with traumatic SI are becoming almost a decade older according to the data in this study. There are many explanations for this change, including better education for trauma prevention in the younger age group and thus a reduction in preventable injuries. The other hypothetical possibility is the cohort effect for a higher-risk group of patients that followed them during the study period. As this group has become older, they could have shifted the median age of patients with major trauma. There have been clear temporal changes in the causation of trauma over time including an increase in the proportion of falls. In turn, falling is associated with an older age group but an increased age itself could easily explain this association. Traumatic SCI is decreasing in frequency. The present analysis is not a population-based study and only includes patients with SCI associated with spinal fracture and/or dislocation; therefore, this decrease doesn t necessarily mean a reduced incidence in the population. However, this reduction could be explained by better preventive measures for SCI, a temporal shift in causation (thus reducing the chances of SCI), or a spatial shift of SCI patients to another center given the isolated nature of some of these patients. The causation pattern has been changing for all SIs and SCIs. Over time, both falling and violence have become more frequent contributing factors for SIs. One possible explanation is that with the increasing age of a trauma patient, falling becomes a more likely causation factor. This study doesn t provide any explanation for the change in the relative importance of violence. Falling clearly has a more consistent role in leading to SCI (21% of patients) than SI (13% of patients). Motor vehicle accident remains the dominant culprit for both SI and SCI. The average ISS did not change over the study period, but the LOS decreased significantly by ~ 40%. Although one might like to believe that this latter finding is solely based on improvements in patient medical care, there are other explanations such as faster disposition of patients to rehabilitation centers, nursing homes, original referring hospitals, and so forth. It is also possible that the definition of ISS has been applied more rigorously over time, and thus creating a selection bias toward attributing the maximum possible ISS to patients. Government funding for trauma patients is inherently intertwined with the severity of the trauma. Therefore, it is possible that the incentive for special funding per trauma case helps boost the ISS per patient. Fig. 8. The relative frequency of case fatality in all patients with SIs (dashed line), SCIs (continuous line), or CSCIs (dotted line) from 1986 to This rate has remained relatively stable over time. 138 J Neurosurg: Spine / Volume 12 / February 2010

9 Epidemiological trends of spine and spinal cord injuries TABLE 4: Relative risk of dying over 2 decades as a result of trauma* Relative Risk Independent Variable All Patients Significance Upper Bound 95% CI The case fatality rate minimally changed in this study period, and the ISS remained stable. It is difficult to conceive that the LOS as a surrogate of outcome significantly decreased but the fatality rate did not improve. The age of trauma patients steadily increased by ~ 10 years over the study period. Age is a major determinant of the case fatality rate as shown on Poisson analysis. Therefore, in reality, case fatality did improve if one considers the effect of age. But even with this consideration, case fatality remains relatively high. This higher fatality rate may imply that despite medical advances, there has been limited efficiency of improved care to prevent death in the sickest patients. Another possibility is the disproportionate presentation of sicker patients with higher ISSs over time, but this factor could not be corroborated by the data analysis. One of the surprising findings in this study was that SCI is not independently predictive of the in-hospital death rate in the acute setting perhaps because of more comprehensive and coordinated care starting at the injury scene. Greater awareness in the medical community may lead to better immediate and ongoing care. Improved intensive care unit care (for example, tracheotomy, deep brain thrombosis prophylaxis, or urosepsis prevention) has been shown to improve the outcome of patients with SCI. It is important to note that this protective effect was evident only during the acute hospital stay and may not reflect a long-term benefit. Authors of some recent studies have suggested that there is an increased mortality rate after 2 years from the traumatic SCI despite better survival in acute care settings. It is also possible that SCIs have been arbitrarily given disproportionately high scoring values on the injury severity scale, affecting the ultimate J Neurosurg: Spine / Volume 12 / February 2010 Lower Bound age group <45 yrs 2.9 < yrs 2.3 < >65 yrs 1 study period < < ISS category < < SCI absent 1.15 < present 1 * Results are adjusted for age groups, ISS category, cord injury, and admission date period by using Poisson analysis. ISS. This higher ISS may not translate into predisposing to a higher death rate when compared with other organ injuries with similar weight on the injury severity scale. Limitations of Study This study is based on administratively oriented and collected data. The parameters and detailed information may not include all the important variables. The possibility of missing data is very low given the vigorous prospectively arranged, well-funded, and quality-tested data collection at this hospital. The study only provides the existing distribution of variables at one hospital. The trauma patients are not necessarily drawn from a fixed and defined population over time, making generalization difficult. The outcome measures are also crude, adding to the shortcomings of this administrative data, for example, the functional outcome measures cannot be linked with the trauma-related variables. Conclusions Spinal injuries and SCIs occur with an annual frequency of 23 and 5%, respectively. The case fatality rate after admission for an SI or SCI is ~ 17% and has remained relatively unchanged over 2 decades. The LOS has reduced ~ 40% for all trauma patients over 2 decades, from a median of 24 to 14 days. Contrary to popular belief, the presence of SCI is not associated with an increase in fatality on multivariate analysis. Disclosure The author reports no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. References 1. Bracken MB, Freeman DH Jr, Hellenbrand K: Incidence of acute traumatic hospitalized spinal cord injury in the United States, Am J Epidemiol 113: , Carreon LY, Glassman SD, Campbell MJ: Pediatric spine fractures: a review of 137 hospital admissions. J Spinal Disord Tech 17: , Damadi AA, Saxe AW, Fath JJ, Apelgren KN: Cervical spine fractures in patients 65 years or older: a 3-year experience at a level I trauma center. J Trauma 64: , DeVivo MJ, Go BK, Jackson AB: Overview of the national spinal cord injury statistical center database. J Spinal Cord Med 25: , DeVivo MJ, Krause JS, Lammertse DP: Recent trends in mortality and causes of death among persons with spinal cord injury. Arch Phys Med Rehabil 80: , DeVivo MJ, Stover SL, Black KJ: Prognostic factors for 12- year survival after spinal cord injury. Arch Phys Med Rehabil 73: , Drainer EK, Graham CA, Munro PT: Blunt cervical spine injuries in Scotland Injury 34: , Dryden DM, Saunders LD, Rowe BH, May LA, Yiannakoulias N, Svenson LW, et al: The epidemiology of traumatic spinal cord injury in Alberta, Canada. Can J Neurol Sci 30: , Holmes JF, Miller PQ, Panacek EA, Lin S, Horne NS, Mower WR: Epidemiology of thoracolumbar spine injury in blunt trauma. Acad Emerg Med 8: ,

10 F. Pirouzmand 10. Jackson AB, Dijkers M, DeVivo MJ, Poczatek RB: A demographic profile of new traumatic spinal cord injuries: change and stability over 30 years. Arch Phys Med Rehabil 85: , Kirshblum SC, O Connor KC: Predicting neurologic recovery in traumatic cervical spinal cord injury. Arch Phys Med Rehabil 79: , Kondziolka D, Schwartz ML, Walters BC, McNeill I: The Sunnybrook Neurotrauma Assessment Record: improving trauma data collection. J Trauma 29: , Krause JS, Broderick L: Outcomes after spinal cord injury: comparisons as a function of gender and race and ethnicity. Arch Phys Med Rehabil 85: , Krause JS, Sternberg M, Lottes S, Maides J: Mortality after spinal cord injury: an 11-year prospective study. Arch Phys Med Rehabil 78: , Lasfargues JE, Custis D, Morrone F, Carswell J, Nguyen T: A model for estimating spinal cord injury prevalence in the United States. Paraplegia 33:62 68, Nobunaga AI, Go BK, Karunas RB: Recent demographic and injury trends in people served by the Model Spinal Cord Injury Care Systems. Arch Phys Med Rehabil 80: , Piatt JH Jr: Detected and overlooked cervical spine injury in comatose victims of trauma: report from the Pennsylvania Trauma Outcomes Study. J Neurosurg Spine 5: , Pickett GE, Campos-Benitez M, Keller JL, Duggal N: Epidemiology of traumatic spinal cord injury in Canada. Spine (Phila Pa 1976) 31: , Pickett W, Simpson K, Walker J, Briston RJ: Traumatic spinal cord injury in Ontario, Canada. J Trauma 55: , Prasad VS, Schwartz A, Bhutani R, Sharkey PW, Schwartz ML: Characteristics of injuries to the cervical spine and spinal cord in polytrauma patient population: experience from a regional trauma unit. Spinal Cord 37: , Sekhon LH, Fehlings MG: Epidemiology, demographics, and pathophysiology of acute spinal cord injury. Spine (Phila Pa 1976) 26 (24 Suppl):S2 S12, Stover SL, DeVivo MJ, Go BK: History, implementation, and current status of the National Spinal Cord Injury Database. Arch Phys Med Rehabil 80: , Strauss DJ, DeVivo MJ, Paculdo DR, Shavelle RM: Trends in life expectancy after spinal cord injury. Arch Phys Med Rehabil 87: , Tator CH, Duncan EG, Edmonds VE, Lapczak LI, Andrews DF: Changes in epidemiology of acute spinal cord injury from 1947 to Surg Neurol 40: , 1993 Manuscript submitted May 24, Accepted September 14, Address correspondence to: Farhad Pirouzmand, M.D., M.Sc., F.R.C.S.C., Division of Neurosurgery, Sunnybrook Health Sciences Centre, University of Toronto, Suite A131, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada. Farhad.Pirouzmand@ sunnybrook.ca. 140 J Neurosurg: Spine / Volume 12 / February 2010