This study analyzed data on 1,995 patients with traumatic spinal cord injuries (TSCI) admitted to hospitals in South Carolina between 1993-2003 to identify factors predicting early mortality. Early mortality was defined as death during the initial hospital admission. The study found through logistic regression that increasing age after 20, male gender, severe systemic injuries, traumatic brain injury, medical comorbidities, poor neurological status at admission, and treatment at a level 1 trauma center were associated with higher odds of early mortality after controlling for other variables. The results suggest timely identification of at-risk patients is important to improve survival in acute care settings, as severe injuries, comorbidities, and brain trauma limit outcomes following TSCI.
1. Objective. To identify predictors of early mortality following traumatic spinal cord injury (TSCI).
Summary of Background Data. Limited information is available on factors associated with early
mortality following TSCI. Ability to identify high risk individuals can help to appropriately treat them, and
reduce mortality.
Methods. Early mortality was defined as death occurring during the initial hospital admission.
Retrospective analysis of 1995 patients with TSCI, admitted to various hospitals of South Carolina from
1993 to 2003, was performed. There were 251 patients with early mortality. Multivariable logistic
regression was used in modeling of early death following TSCI with gender, race, age, Frankel grade,
trauma center, level of injury, injury severity score (ISS), traumatic brain injury (TBI), and medical
comorbidities as covariates.
Results. Increasing age after 20 years (OR: 1.2, P = <0.0001), male gender (OR: 1.6, P = 0.016), severe
(ISS ≥15) systemic injuries (OR: 1.9, P = 0.012), TBI (OR: 3.7, P < 0.0001), 1 or more comorbidities (P <
0.0001), poor neurologic status (P = 0.015), and level 1 trauma center (OR: 1.4, P = 0.026) were
significantly associated with early mortality, after adjusting for other covariates.
Conclusion. Early mortality following TSCI is influenced by multiple factors. Timely recognition of these
factors is crucial for improving survival in the acute care setting. Severe systemic injuries, medical
comorbidities, and TBI continue to be the main limiting factors affecting the outcome. These findings also
suggest the need to allocate resources for trauma prevention, and promote research towards improving
the care of acutely injured patients.
Introduction
Traumatic spinal cord injury (TSCI) is a serious medical disorder in terms of both mortality and morbidity.
The annual incidence of TSCI in the United States, excluding those that die at the scene of accident, is
approximately 11,000 and the prevalence is 253,000.[1]
Though TSCI primarily affects young adults, the
average age of TSCI has increased steadily from 28.7 years in 1979 to 37.6 years in 2000, a reflection of
rise in the median age of the general population. The proportion of persons older than 60 years at injury
has also increased from 4.7% before 1980 to 10.9% after 2000. Since 2000, 79.6% of injuries have
occurred among males, with a slight trend towards decreasing male percentage, compared to 1980 and
earlier.[1]
Frequently TSCI is associated with other injuries. Due to the complexity and multitude of clinical
problems associated with TSCI, early and long-term outcomes are influenced by a host of factors. There
have been numerous studies on long-term morbidity following TSCI.[2–8]
Causes of death or factors
predisposing to mortality following TSCI reported in long-term prospective and retrospective studies, have
included advanced age; cord injury at a higher spinal level; infections (septicemia from infected decubiti,
urinary tract infection, and pneumonitis); pulmonary thromboembolism; medical comorbidities
(cardiovascular disease, diabetes, and pulmonary disease); and suicide. However, there is a paucity of
2. information on early mortality following TSCI. In a review of English language medical literature published
from 1960 to 1978, Krause[9]
reported a mortality rate ranging from 4.4% to 16.7% following admission for
acute TSCI. Current data on early mortality, after TSCI, and an understanding of factors contributing to
the same are lacking. Further, few studies have used statewide population based dataset. This
retrospective study was performed to identify predictors of early mortality following TSCI. In this study, the
term "early mortality" is defined as death occurring within the acute care facility, before discharge or
transfer to an extended care facility. To our knowledge, this is among the very few studies that address
the influence of demographic, clinical, and provider characteristics on early mortality among acutely
injured spinal cord patients. Recognition of factors associated with early mortality can help streamline the
trauma care system, identify high-risk patients and potentially reduce deaths in this patient population.
Variable Definitions
The primary outcome variable for all analyses was early mortality status (yes or no),
defined as death in the acute care setting before discharge or transfer to an
extended care facility. Covariates were selected based on clinical relevance. Race
was dichotomized as white and nonwhite, with the nonwhite category including
African Americans as well as those identified as other than white or black (64
subjects or 3% overall). Age (years) was treated as a continuous variable in all
analyses. Functional state was defined according to Frankel grading system.[12] The
database included a comorbidity scale developed for use with administrative
data.[13] This method identifies 30 conditions known to be significant predictors of in-
hospital mortality and resource use. All 10 ICD-9-CM diagnosis fields were searched
for any of these 30 conditions. For the purposes of our analysis, subjects were
categorized as having 0, 1, 2, or 3 or more comorbid conditions. Level of injury was
categorized into cervical, dorsal, and lumbosacral. For the purposes of analysis,
injury level was coded as missing for nonspecific injuries (n = 151). Overall injury
severity was graded using the Injury Severity Score (ISS).[14] The ISS is an
anatomically-based measure ranging from 1 to 75. Briefly, the injury scoring system
divides the body into 6 regions: head or neck, face, chest, abdominal or pelvic
contents, pelvic girdle or extremities, and external. The Injury Severity Score is then
calculated using the 3 most severely injured body regions. ICDMAP-90 software was
used to translate the ICD-9-CM diagnosis codes into an equivalent ISS. Subjects
were then categorized as mildly, moderately or severely injured based on the
following ISS values: mild = ISS from 1 to 8; moderate = ISS from 9 to 14; and
severe = ISS ≥15. Because of sparse numbers for the mild ISS class (only 2 early
mortality subjects were classified with a mild injury), we combined the mild and
3. moderate injury severity categories to facilitate subsequent data analyses. Subjects
were classified as experiencing traumatic brain injury (TBI), based on case definition
for TBI provided by CDC.[11] Hospitals were classified as either level 1 or level 2
trauma centers based on the SC Department of Health and Environmental Control
Health Regulation Office that evaluates trauma level status of hospitals. For analysis
purposes, trauma facilities designated as 3 and 4 were grouped with level 2.