Short-term pelvic fracture outcomes in adolescents differ from children and adults in the National Trauma Data Bank

Background Pediatric pelvic fractures are associated with high-energy trauma and injury to other systems, leading to an increased incidence of complication and mortality. Previous studies analyzed the pediatric population as a whole, including both children and adolescents. The purpose of this study was to examine whether adolescents with pelvic fracture have different complication and mortality rates compared to younger children and adults. Methods Using the National Trauma Data Bank, 37,784 patients below the age of 55 years with pelvic fractures were identified and divided into children (age <13 years), adolescents (age 13–17 years), and adults (age >17 years). Descriptive statistics and bivariate and multivariate analyses were performed. Results Children had an increased odds of death [odds ratio (OR) 2.29, 95 % confidence interval (CI) 1.96–2.67] and complications (OR 1.36, 95 % CI 1.20–1.55), whereas adolescents had a decrease in odds of death (OR 0.89, 95 % CI 0.74–1.06) and complications (OR 0.70, 95 % CI 0.61–0.81) compared to the adult population. Conclusions Adolescents with pelvic fractures exhibit a different physiologic response to the children and adult populations. This emphasizes the need to distinguish these subpopulations in future epidemiological research and treatment planning.

Biological, psychological, social, and environmental changes influence the onset and termination of adolescence [12]. Physiologic changes that occur during adolescence, such as increase in muscle size and bone mass, doubling in heart size and lung vital capacity, and rise in blood pressure, blood volume, and hematocrit (particularly in boys), may lead to a theoretical protective effect against trauma [12]. Adolescents also retain the ability to rapidly heal and recover. Similar to growing children, adolescents have an osteoblast:osteoclast activity ratio [1, which increases their capacity to heal fractures. These physiological advantages, together with the increase in distance between internal organs, weight to surface area ratio, and cardiovascular reserve [4], may make adolescents particularly resilient to traumatic injury.
The purpose of this study is to examine whether adolescents with pelvic fracture have different complication and mortality rates compared to younger children and adults.

Methods
We identified the study population by means of the National Trauma Data Bank (NTDB version 7.1). This NTDB version contained over 2.7 million cases from over 900 US trauma centers between the years 2002 and 2006 [13]. The data were imported and merged into a single dataset from the 13 NTDB files using SAS Ò version 9.2 (SAS Institute, Cary, NC).
The initial NTDB population consisted of over 2.7 million entries. For the purpose of our study, all burn or penetrating injuries were excluded, which reduced the group to 1.7 million cases. If a patient's multiple ICD-9 diagnosis codes contained at least one of the following codes, the patient was considered to have a pelvic fracture: 808.2, 808.3, 808.4, 808.41, 808.42, 808.43, 808.49, 808.5, 808.51, 808.52, 808.53, 808.59, 808.8, 808.9. Acetabular fractures were excluded from the study. Those entries without a pelvic fracture were removed, giving a total of 54,459 cases of pelvic fractures. Finally, all adults aged 55 years and older were removed from the study, yielding a final study population of 37,784.
This final study population was subdivided into our three groups of interest: children (younger than 13 years old), adolescents (aged 13-17 years), and adults (aged 18-54 years). The children and adolescents populations together were identified as the pediatric population. Open fractures constituted 3.4 % of all fractures. Each subgroup was assessed for their odds of death and a severe complication. The main outcomes of interest were mortality and a severe complication. Severe complication was defined as having at least one of the following complications recorded: renal failure, pneumonia, bacteremia, acute respiratory distress syndrome (ARDS), deep vein thrombosis (DVT), or a pulmonary embolism. Prehospital risk factors such as sex, race, age, arrival in shock (systolic blood pressure \90 mmHg), Injury Severity Score (ISS), head injury, and mechanism of injury were also analyzed to determine their association with the main outcomes.

Statistical analysis
All statistical analyses were conducted using SAS Ò . Descriptive statistics were performed on the entire study population (n = 37,784).
To determine associations between risk factors and the main outcomes of interest, bivariate analysis was conducted between each prehospital risk factor and two main outcomes (mortality and severe complication). Risk factors that were not already dichotomous were transformed into two mutually exclusive categories. For example, ISS became C25 and \25. For this and subsequent assessment, open fractures (only 4.2 % of pelvic fractures and highly variable in presentation and complication profile) and unknown mechanisms of injury were excluded, yielding a study group of 24,684.
A subpopulation of severe pelvic injuries was also created to determine if any pediatric subgroup has better outcomes. For this population, an Abbreviated Injury Scale (AIS) score of \3, signifying minor and moderate injuries, and all patients with lower extremity fractures besides pelvic fractures were excluded. This created a study subgroup of 2,487 patients with isolated severe pelvic fractures.
For both the main study population and the subgroup of severe pelvic injury, prehospital risk factors were determined to be significant using the Mantel-Haenszel test. This method produced crude odds ratios (ORs) and 95 % confidence intervals (CIs), and those variables of significance were included for the multivariate analysis.
Logistic regression analyses including those variables of significance were performed to determine the association between the prehospital risk factors and the two main outcomes. Each model included a specific age group (children, adolescent, and adult) as compared to all other ages in order to assess the importance of age and outcome after sustaining a pelvic fracture. The Hosmer-Lemeshow goodness-of-fit test was performed on each model to determine whether or not the observed event rates matched the expected event rates.

Results
Our study 's overall incidence rate of pelvic fractures was  26 cases per 10,000 trauma admissions per year for adults,  4 cases per 10,000 trauma admissions per year for adolescents, and 5 cases per 10,000 trauma admissions per year for children. Severe pelvic fractures had an incidence rate of 3.5 cases per 10,000 per year for adults, 0.47 cases per 10,000 per year for adolescents, and 0.63 cases per 10,000 per year for children.
Descriptive data for the whole pelvic fracture population There were 37,784 individuals that matched the preliminary inclusion and exclusion criteria. Descriptive statistics of demographics, injury mechanism, injury severity, treatment, and complications are presented in Table 1. After sustaining a pelvic fracture, survival was 91.4 % for the total population. When separated by age group, adolescents appeared to have the best survival statistics (93.2 %), while children had the worst (89.8 %). At least one severe complication occurred in 14.9 % of the total population. Adolescents had the lowest percentage of severe complication (11.4 %), children had a slightly higher percentage (13.0 %), and adults had the highest percentage (15.7 %).
Most of the pelvic fractures were closed (95.8 %). Pubic rami fractures consistently had the highest prevalence across all age groups (41.5 %). The most common known mechanism of injury was motor vehicle accident (39.9 %), followed by high-energy fall (9.0 %) and motorcycle accident (7.6 %). Most patients had an ISS \25 (71.1 %). When divided by age, more children had an ISS \25 (81.5 %) than adolescents (69.5 %) or adults (69.4 %). Adolescents also had less reports of hypovolemic shock on arrival (systolic pressures below 90 mmHg) (12.9 %) than children (16.5  Multivariate analysis for the whole pelvic fracture population Variables that were found to be significant in the bivariate analysis were included in the final multiple logistic regression models for death and severe complication as outcomes. Adjusted ORs and 95 % CIs are presented in Table 3. When controlling for mechanism of injury and other prehospital conditions, children had more than twice the odds of death (OR 2.29, 95 % CI 1.96-2.67) and increased odds of severe complication (OR 1.36, 95 % CI 1.20-1.55) relative to the adult group after sustaining a pelvic fracture. Adolescents, on the other hand, had a slight decrease in odds of death (OR 0.89, 95 % CI 0.74-1.06) and lower odds of severe complication (OR 0.70, 95 % CI 0.61-0.81) relative to the adult group after sustaining a pelvic fracture.
Descriptive data for the severe pelvic fracture subpopulation A subgroup analysis was performed for patients determined to have a severe, isolated, closed pelvic injury (AIS C3). Table 4 demonstrates the descriptive statistics for this subpopulation (n = 2,487). Gender and race had no significant impact on outcomes compared to the whole population. Adolescents (n = 255) still appeared to have a lower incidence of mortality (3.1 %) and lower incidence of a severe complication (8.2 %) compared to adults (n = 1,886) and children (n = 346). Adolescents contributed the most to motor vehicle accident (89.4 %), which was the most common mechanism of injury in the severe pelvic fracture group. Similar to the whole population, the severe pelvic injury group demonstrated more children with an ISS\25 (82.1 %). For cases of hypovolemic shock on arrival, adolescents had the fewest reports of systolic pressures below 90 mmHg (6.7 %), while children (15.9 %) had the largest contribution. Children had the  Multivariate analysis for the severe pelvic fracture subpopulation Variables that were found to be significant in the bivariate analysis were included in the final multiple logistic regression models for death and severe complication as outcomes (  (Table 6).

Discussion
In this study, we analyzed data gathered from the NTDB to characterize the risks for mortality and complications associated with pelvic fractures in adolescents as compared to younger children and adults. Compared to adults, adolescents were shown to have decreased mortality and severe complication rates, while children have an increased mortality rate but decreased severe complication rate. In the subanalysis of severe pelvic fractures, the same pattern emerged, although the results for complication rates were not statistically significant.
Previous studies have shown conflicting reports regarding mortality associated with pediatric pelvic fractures. Mortality rates as low as 1.4 % to as high as 25 % have been reported [5,10,11,14,15]. Banerjee et al. [7] reported a 16 % mortality rate in 44 patients related mainly to the associated injuries and not to the pelvic fractures themselves. Another study reported 5 % overall mortality for 722 pediatric pelvic fractures versus 17 % among similar injuries in the adult population [15]. Comparable to previously reported studies, the current study reported 10.2 % survival for children and only 6.8 % for adolescents, compared to 8.5 % in adults with pelvic fractures. The pattern of lower survival in children and higher survival in adolescents compared to adults was also seen in the analysis of severe pelvic injuries (Table 4). This study illustrates that simply comparing pediatric mortality to that of adults may be an inadequate evaluation due to the fact that adolescents and children sustaining pelvic injuries may have different outcomes.
Central nervous system head injury was cited as the most common cause of death in children with pelvic fractures in two recent studies [2,11]. Other causes of death in children with pelvic fractures, reported in these studies, include multiorgan failure and visceral injuries [2,11]. In a study that examined the National Pediatric Trauma Registry (NPTR), hospital type (children's or general) and ISS were the only significant parameters in the multivariate analysis [8]. Hemorrhage from a pelvis fracture-related vascular injury was reported as the cause of death in only 0.3 % of children, compared with 3.4 % of adults [15]. The current study supports these previous studies. Multivariate analysis showed hypovolemic shock, head injury, and ISS C25 to correlate with an increased mortality rate in a patient with any type of pelvic fracture (severe or not). The multivariate model for severe pelvic injuries also demonstrated that, despite their lower ISS, children had a significantly higher risk of mortality (OR 1.90, 95 % CI 1.06-3.41) and adolescents had a significantly lower risk of mortality (OR 0.40, 95 % CI 0.18-0.91) as compared to adults. Whereas many studies report the rate of associated injuries, very few studies report the rate of complications after pediatric pelvic fractures [2]. According to a previous study of 120 children (age \16 years) with pelvic fractures, complications included 13.3 % wound complications, 8.3 % urinary tract infections, 4.2 % septicemias, 1.7 % ARDS, and one case (0.8 %) of pulmonary embolus (PE) [2]. In the current study, there were 346 children (age \13 years) with severe pelvic injuries, with a lower overall complication rate of 9.3 %. There was one case (0.3 %) of ARDS and one case of PE (0.3 %), two cases of renal failure (0.6 %), and no cases of septicemia. The 255 adolescents with severe pelvic injury had an even lower complications rate of 8.2 %, with one ARDS (0.4 %), one septicemia (0.4 %), one renal failure (0.4 %), one DVT, but no PE. Urinary tract infection (UTI) was not considered to be a significant complication in our study. However, in the severe pelvic injury group, there were three children (0.87 %), three adolescents (1.2 %), and 30 adults (1.59 %) with UTI. The most common severe complication was pneumonia: three children (0.87 %), 12 adolescents (4.7 %), and 53 adults (2.8 %).
Previous studies have reported a high percentage of associated injuries in the pediatric pelvic fracture population. In one study surveying 166 cases, there were 38.6 % head traumas, 19.9 % significant chest traumas, and 19.3 % abdominal/visceral trauma [11]. Additionally, an incidence of 11-27 % of associated abdominal injuries [2,10,11,[15][16][17], 27 % associated chest injuries, and up to 50 % of associated head injuries have been previously reported [2,10,18]. The proportion of associated injuries (head, thorax, and abdomen ,respectively) in the current study was 7.5, 4.1, and 4.6 % for children; 18, 7.1, and 8.6 % for adolescents. The proportion of a depressed consciousness level [Glasgow Coma Scale (GCS) score \ 15] in the population of severe pelvic fractures was 19.4 % in children and 31.3 % in adolescents (25 % if adolescents and children are grouped together). This is different to the previously reported 38 % in the pediatric population [2].
Other pertinent findings in this study included a 5-7 times higher incidence of pelvic fractures in adults versus children and adolescents, a higher incidence of pelvic fractures in white non-Hispanic males, the vast majority of the injury mechanism of pelvic fractures being blunt trauma from motor vehicle accidents, and a predominance of nonoperative treatment for these fractures. These findings are largely in accordance with previous reports [2,3,8,11].
The main strengths of the current study is that this is the largest cohort used for an investigation of this sort, and the national representation of academic and non-academic centers of various levels provide far-reaching generalizability. However, this study also has a number of  limitations. While the NTDB includes a very large sample of trauma centers from all over the United States, it is a convenience sample that consists solely of data submitted by participating hospitals, leading to a disproportionate number of larger hospitals with younger and more severely injured patients. The quality of data entered into the NTDB is solely dependent on the participants entering the data; however, the NTDB is continuously working to screen, clean, and standardize the quality of the data entered into the database [13]. The larger population provided by the NTDB may be the reason for a higher mortality seen in our pediatric population, unlike the previous studies with smaller sample sizes. The NTDB also reported more pelvic fractures in children than in adolescents, which contributed to our pediatric population's higher mortality rate. We also used pelvis AIS C3 to indicate a severe pelvic injury, while other studies have used a cutoff of AIS C4 [3].
Additionally, since the NTDB does not code separately for pelvic AIS, we derived our pelvic AIS score by using the lower extremity AIS score and excluding all patients with other lower extremity injuries. Assuming an associated lower extremity fracture rate of about 25 % [11], this may have influenced our descriptive statistics. However, this allowed us to compare similar injuries and better isolate the effect of age. The NTDB also does not provide blood gas values such as base deficit and lactate, and, therefore, our analysis was limited to the use of hypovolemic shock on arrival. There were also weaknesses in the tests of the model fitness. This suggests that there could be interactions between variables that were not included in the model. These variables might not have been collected by the NTDB or may be latent factors that are not identified. Such discrepancies may have altered our final parameter estimates by either inflating or deflating their values. Finally,   In conclusion, the results of this study suggest that the adolescents with pelvic fractures behave as a separate group from the children and adult populations. These results emphasize the need for further epidemiological research on disparate injury response between pediatric subpopulations defined by level of maturity. There should be further research aimed towards designating physiological criteria besides age by which to place pediatric patients in their respective subpopulations. This study highlights the fact that the mortality rate seen in a given pediatric population may not be an accurate measurement, as our lower mortality rate largely stems from the adolescents' increased survival, while the children may actually have a higher mortality rate than the adults.
Conflict of interest All authors contributing to this study have no financial interest that relates in any way to this study.
Author contributions Meir Marmor MD was involved in the study design, data analysis, and writing. Joshua Elson MPH was involved in the data analysis and writing. Christopher Mikhail BS was involved in the writing. Saam Morshed MD PHD was involved in the study design and data analysis. Amir Matityahu MD was involved in the study design and writing.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.