Abstract
Background
Open pelvic fractures are rare injuries, associated with high patient morbidity and mortality. Few studies have investigated the impact of patient demographics, comorbidities, and injury related factors on complication and mortality rates. The purpose of this study was to: (1) identify the overall incidence of complications and mortality after open pelvic fractures, (2) compare patient factors between those who did and did not develop complications, (3) identify perioperative independent risk factors for complications and mortality.
Methods
A query was performed for patients with open pelvic fractures between 2007 and 2017 using the American College of Surgeons National Trauma Data Bank. Patient and injury specific variables were collected and complications were identified using International Classification of Disease Ninth and Tenth edition Codes. Patient demographic and perioperative data was compared using Fisher’s exact test and chi-square test for categorical variables, and Welch’s t-test for continuous variables. Using pooled data from multiple imputations, logistic regressions were used to calculate odds ratios and confidence intervals of independent risk factors for complications.
Results
A total of 19,834 open pelvic fracture cases were identified, with 9622 patients (48.5%) developing at least one complication. Patients who developed complications were older (35.0 vs 38.1 years), and had higher Injury Severity Scores (17.7 vs 26.5), lower Glasgow Coma Scores (14.2 vs 11.7), and a larger proportion presenting with hypotension (21% vs 6.9%). After pooled regression involving 19 factors, these were the strongest independent predictors of inpatient complication and mortality.
Conclusion
We report a mortality rate of 14%, with an inclusive complication rate of 48.5%. Evaluating risk factors for morbidity and mortality for this devastating orthopaedic injury provides knowledge of an inherently sparse population.
Level of Evidence
Level II, Retrospective study.
Keywords: Orthopaedic surgery, Orthopaedic trauma, Epidemiology, Open pelvic fractures, Pelvic fractures, National trauma data bank, NTDB
Abbreviations: RTS, Revised Trauma Score; ISS, Injury Severity Score; ACS-NTDB, American College of Surgeons National Trauma Data Bank; ED, Emergency Department; ICD-9, International Classification of Disease Ninth; ICD-10, International Classification of Disease Tenth; ACS, American College of Surgeons; E-Code, external cause of injury; GCS, Glasgow Coma Scale; AIS, Abbreviated Injury Scale; MCAR, missing completely at random; PE, pulmonary embolism; UTI, urinary tract infection; BP, blood pressure; SBP, systolic blood pressure
1. Introduction
Pelvic fractures are rare injuries, composing 2–8% of all skeletal trauma injuries.1, 2, 3, 4 Open pelvic fractures occur more infrequently, accounting for 2–4% of all pelvic fractures.5,6 High energy mechanisms, such as motor vehicle versus pedestrian, and motorcycle injuries, are needed to disrupt the pelvic ring to an extent to cause an open injury. The high energy imparted causes not only significant bony injury, but also severe soft tissue damage. The literature reports varying mortality rates, ranging from 7% to as high as 58%.7, 8, 9 The timing of mortality in open fractures usually demonstrates a bimodal distribution. Early mortality often occurs secondary to exsanguinating hemorrhage.10,11 Late mortality, on the other hand, manifests on average 17 days the after the injury and is attributed to sepsis and multiorgan failure.12,13 Mortality rates after open pelvic fractures were first described in the 1970’s and 1980’s as near 50%.7,14,15 At the time, this was significantly more than reported for closed pelvic fractures (10.5%).16 Treatment protocols including: damage control surgery, exploratory laparotomy, aggressive transfusion protocols, hemorrhage control, arterial embolization, and fecal diversion have all lead to decreased mortality from these injuries.9, 17, 18, 20
Predictors of overall patient mortality after an open pelvic fracture have been studied in small case series and retrospective reviews. Dente et al.12 identified factors predicting mortality in patients with open pelvic fractures: the Revised Trauma Score, (RTS), Injury Severity Score (ISS), and number of packed red blood cell transfusion requirements. In addition to patient mortality, many complications are associated with these injuries due to the high energy causes of open pelvic fractures.19, 20 Early complications in these patients include thromboembolic disease, pneumonia, infection, genitourinary disability, and multiorgan failure.21 Patient specific risk factors and demographics have been reported inconsistently and in small patient cohorts, which has led to heterogeneous results between studies. Moreover, previous studies were predominantly from single institutions which may limit their generalizability to other hospital settings.
The identification of patient risk factors that lead to complications and mortality after open pelvic fractures has been infrequently described. The rarity of open pelvic fractures makes it difficult to produce adequately powered studies to evaluate the frequency of complications and the associated risk factors. The purpose of this study is to: (1) identify overall incidence of complications and mortality after open pelvic fractures (2) compare patient factors between those who did and did not develop a complication, (3) identify independent risk factors for complications and mortality.
2. Methods
2.1. Registry
A query was performed for patients treated at trauma centers using the 2007 to 2017 American College of Surgeons National Trauma Data Bank (NTDB) Participant Use File/Research Dataset. This database includes trauma patient information collected from emergency departments (ED) and hospitals, with the goal to inform and improve the care of these patients. In our study, patients were identified using International Classification of Disease, Ninth and Tenth Revisions Diagnosis Codes (ICD-9, ICD-10). Information in the registry was de-identified, and therefore, did not require approval from our Institutional Review Board. The NTDB remains the full and exclusive copyrighted property of the American College of Surgeons (ACS). The ACS is not responsible for any claims arising from works based on the original Data, Text, Tables, or Figures.
2.2. Study population
Open pelvic ring fractures were queried by corresponding diagnosis codes from the ICD-9 and ICD-10 (Appendix 1 and 2). This resulted in 22,156 patients with open pelvic ring fractures. Subsequently, patients less than 18 years of age (N = 1770) and those who were dead on arrival to the emergency department (N = 552) were excluded, resulting in our study population (N = 19,834).
2.3. Variable categorization and outcomes assessed
Due to the differing data collection methods for NTDB files before 2016, complications were extracted from a list format by the presence of the complication and the confirmation of its absence. Cases labeled missing or not collected were labeled accordingly. Complications included in Table 1 had less than ten percent of data listed as missing and incidences were determined by valid cases. Other specific complications (Abdominal Fascia Left Open, Base Deficit, Bleeding, Coagulopathy, Coma, Graft Failure, Increased Intracranial Pressure, Wound Disruption) included data points not collected across numerous years, and were combined to determine incidence. Other unspecified complications were labeled as “Other”. Mortality was extracted from the emergency department and hospital discharge information. Mortality was treated as a complication in the determination of All Complications. The compilation of these variables resulted in an endpoint with no missing values.
Table 1.
Incidence of complications and mortality in patients with open pelvic fractures.
| Complication | N | Percent∗ |
|---|---|---|
| Abdominal Compartment Syndrome | 70 | 0.5 |
| Alcohol/Drug Withdrawal | 131 | 0.7 |
| Acute Kidney Injury | 599 | 3.4 |
| Acute Respiratory Distress Syndrome | 579 | 3.2 |
| Cardiac Arrest | 597 | 3.4 |
| Deep and Organ Space Surgical Site Infection | 631 | 4.1 |
| Superficial Surgical Site Infection | 290 | 1.5 |
| Decubitus Ulcer | 362 | 2.0 |
| Deep Vein Thrombosis | 716 | 4.0 |
| Extremity Compartment Syndrome | 185 | 1.0 |
| Myocardial Infarction | 45 | 0.2 |
| Osteomyelitis | 36 | 0.2 |
| Pneumonia | 1025 | 5.8 |
| Pulmonary Embolism | 237 | 1.3 |
| Return to Intensive Care Unit | 218 | 1.1 |
| Return to Operating Room | 315 | 1.8 |
| Sepsis | 235 | 1.3 |
| Catheter Septicemia | 46 | 0.2 |
| Stroke/Cerebrovascular Accident | 86 | 0.4 |
| Unplanned Intubation | 224 | 1.3 |
| Urinary Tract Infection | 351 | 2.0 |
| Other Specified Complications2 | 840 | 4.2 |
| Other Unspecified Complications | 2860 | 14.4 |
| Mortality | 2211 | 11.6 |
| All Complications1 | 9622 | 48.5 |
Incidence based on valid cases.
Some patients experienced more than one complication
Refer to Methods for specifics.
Patients undergoing pelvic angiography were determined by ICD-9 procedure codes (38.86, 88.4, 88.40, 88.47, 88.49, 39.79) or extracted from Angiography variable information. Alcohol and Drug Screen Result values were combined and reported as either positive or negative. Protective Devices were extracted from a list format in a similar manner to complications, with Other compiled from protective clothing (e.g. leather pants), protective non-clothing (e.g., shin guard), eye protection, and unspecified other. Trauma Type was differentiated by classification of external cause of injury (E-Code), into either Blunt, Penetrating, or Other. Fracture Involvement was determined by ICD-9, ICD-10 diagnosis codes used for the determination of study population.
In addition, demographic factors like age, race, gender, and comorbidities were collected. Perioperative factors like Glasgow Coma Scale (GCS), Injury Severity Score (ISS), ACS Verification Level, and ED blood pressure were also assessed. Systolic blood pressure was turned into a dichotomous variable by noting patients with ED values less than 90 mm Hg, previously indicated as a risk factor in the trauma literature.23 The GCS used was recorded in the ED. ISS was calculated by Abbreviated Injury Scale (AIS).
2.4. Statistical analysis and missing data
As noted above, there were several variables with missing data. Variables had varying percentages of missing values ranging from less than one to five percent, with data for Pelvic Angiography missing in approximately 11% of cases. Protective Devices, Fracture Involvement, Alcohol and Drug Screen were non-disjoint or missing large amounts of data. These were presented as comparisons between groups but were not included in the final regression.
Missing value analysis and Little’s MCAR (missing completely at random) test confirmed data was not missing at random. Multiple imputation, with inclusion of all factors in Table 1, ACS verification level, Trauma Type, and Pelvic Angiography, was performed. Number of imputations was set to 20, and linear and logistic regressions were used to determine values for continuous and categorical variables. Pooled logistic regression was then performed including variables with a p-value less than 0.2 in the univariate analysis. This was compared to a regression done with the original data. Improvements were noted in the predictive ability of the model (64.5 vs 67.8% correct). Of note, in the pooled analysis as opposed to the original, gender was determined to be a significant risk factor. Obesity was also approaching significance, which was not seen in the original analysis.
All statistical analyses were performed with SPSS version 26 (IBM Corporation, Armonk, New York). Patient demographics and perioperative data were compared using Fisher’s exact test and chi-square test for categorical variables, and Welch’s t-test for continuous variables. Logistic regressions, with pooled data from multiple imputations, were used to calculate odds ratios (OR) and confidence intervals of independent risk factors for complications. Model was significant (p < 0.001) and demonstrated appropriate goodness of fit. Tests were deemed significant with a p-value less than 0.05 or, where applicable, a Holm-Bonferroni correction was done to determine an adjusted p-value.
3. Results
In our analysis of 19,834 open pelvic fracture cases, 9622 patients (48.5%) developed at least one complication. Of the patients who were admitted to the emergency, 2211 patients (11.6%) expired during their hospital stay (Table 1). The most common defined complication was pneumonia (5.8%), followed by deep and organ space surgical site infection (4.1%), deep vein thrombosis (4.0%), cardiac arrest (3.4%), acute kidney injury (3.4%), and acute respiratory distress syndrome (3.2%).
The average age of patients who developed complications was statistically higher (35.0 vs 38.1 years, p < 0.001), as was the Injury Severity Score (17.7 vs 26.5, p < 0.001). The Glasgow Coma Score was also significantly lower in the patients with complications (14.2 vs 11.7, p < 0.001) (Table 2). Males made up 80.3% of the cases and accounted for 79.4% of complications (p = 0.001). Regarding race, whites made up 52.9% of the complications and 48.8% of the study population, while blacks made up 34.0% of the complications and 39.0% of the study population. There were less patients with substance abuse/dependence and fewer smokers in the complications group. Patients with bleeding disorders and diabetes had higher numbers of patients in the complication group (p < 0.001) (Table 2).
Table 2.
Demographic and comorbidity factor comparison between complication groups.
| Factors | Study Population |
No Complication |
Complication |
|||
|---|---|---|---|---|---|---|
| Mean ± SD | Mean | SD | Mean | SD | p-value | |
| Age, years | 36.5 ± 15.0 | 35.0 | 15.2 | 38.1 | 16.3 | <0.001 |
| Glasgow Coma Scale | 13.0 ± 3.9 | 14.2 | 2.5 | 11.7 | 4.9 | <0.001 |
| Injury Severity Score |
22.0 ± 12.2 |
17.7 |
9.5 |
26.5 |
14.2 |
<0.001 |
| Study Population | No Complication | Complication | ||||
| Demographics |
N (Percent∗) |
N |
Percent∗ |
N |
Percent∗ |
p-value |
| Gender | 0.001 | |||||
| Female | 3898 (19.7) | 1914 | 18.8 | 1984 | 20.6 | |
| Male | 15,867 (80.3) | 8243 | 21.2 | 7624 | 79.4 | |
| Race | ||||||
| White | 9164 (48.8) | 4357 | 44.9 | 4807 | 52.9 | <0.001 |
| Black | 7323 (39.0) | 4238 | 43.7 | 3085 | 34.0 | <0.001 |
| Asian | 217 (1.2) | 96 | 1.0 | 121 | 1.3 | 0.029 |
| Ame. Ind., Pac. Isl. | 156 (0.8) | 70 | 0.7 | 86 | 0.9 | 0.920 |
| Other Race | 1918 (10.2) | 934 | 9.6 | 984 | 10.8 | 0.007 |
| Comorbidities | ||||||
| Alcohol Abuse | 1313 (6.9) | 621 | 6.3 | 692 | 7.4 | 0.004 |
| Substance A/D | 1231 (6.4) | 803 | 8.2 | 428 | 4.6 | <0.001 |
| Bleeding Disorder | 267 (1.4) | 101 | 1.0 | 166 | 1.8 | <0.001 |
| CHF | 115 (0.6) | 48 | 0.5 | 67 | 0.7 | 0.049 |
| COPD | 742 (3.9) | 405 | 4.1 | 337 | 3.6 | 0.061 |
| CVA | 95 (0.5) | 35 | 0.4 | 60 | 0.6 | 0.005 |
| Diabetes | 836 (4.4) | 364 | 3.7 | 472 | 5.1 | <0.001 |
| Hypertension | 2107 (11.0) | 1011 | 10.3 | 1096 | 11.7 | 0.002 |
| Obesity | 1296 (6.9) | 684 | 7.1 | 612 | 6.6 | 0.158 |
| Smoker | 3397 (15.6) | 2159 | 22.1 | 1238 | 13.3 | <0.001 |
| Chronic Steroid Use | 33 (0.2) | 15 | 0.2 | 18 | 0.2 | 0.602 |
SD=Standard Deviation; ∗Frequency based on valid cases; Ame. Ind.=American Indian; Pac. Isl.=Pacific Islander; A/D=Abuse/Dependence; CHF=Congestive Heart Failure; COPD=Chronic Obstructive Pulmonary Disease; CVA=Cerebrovascular Accident.
Of patients who presented to the ED with a systolic blood pressure less than 90 mmHg, 21% developed complications compared to 6.9% who did not (p < 0.001) (Table 3). Patients that tested positive on a drug screen were more likely to be in the group with no complications. Patients who had airbags present during their accident were more likely to be in the no complication group. Patients with other protective devices (e.g., lap belt, shoulder belt, protective clothing, shin guards, unspecified) had increased frequency in the complications group. Patients who experienced blunt trauma and multiple fractures were more likely to fall into the complication group. Hospital length of stay was over 7 days longer for patients with complication on average (6.9 ± 9.5 vs 17.2 ± 22.0 days, p < 0.001).
Table 3.
Perioperative factor comparison between complication groups.
| Factors | Study Population |
No Complication |
Complication |
p-value | ||
|---|---|---|---|---|---|---|
| N (Percent∗) | N | Percent∗ | N | Percent∗ | ||
| ACS Verification Level | <0.001 | |||||
| I | 8536 (44.8) | 4385 | 45.4 | 4151 | 44.3 | |
| II | 3487 (18.3) | 1717 | 17.7 | 1774 | 18.9 | |
| III | 222 (1.2) | 138 | 1.4 | 84 | 0.9 | |
| IV | 14 (0.1) | 11 | 0.1 | 3 | 0.0 | |
| Not Applicable | 6777 (35.6) | 3417 | 35.4 | 3360 | 35.9 | |
| ED BP, <90 mm Hg | 2636 (13.7) | 685 | 6.9 | 1951 | 21.0 | <0.001 |
| Pelvic Angiography | 721 (4.4) | 229 | 3.0 | 492 | 5.7 | <0.001 |
| Alcohol Screen, Pos.2 | 4294 (26.1) | 2226 | 26.3 | 2068 | 25.8 | 0.435 |
| Drug Screen, Pos.2 | 4360 (28.0) | 2399 | 29.6 | 1961 | 26.3 | <0.001 |
| Protective Devices1 2 | ||||||
| Airbag Present | 1959 (16.2) | 1073 | 17.3 | 886 | 15.1 | 0.001 |
| Helmet | 2036 (16.9) | 891 | 14.4 | 1145 | 19.5 | <0.001 |
| Lap Belt | 1173 (9.7) | 512 | 8.3 | 661 | 11.3 | <0.001 |
| Shoulder Belt | 1030 (8.5) | 443 | 7.2 | 587 | 10.0 | <0.001 |
| Other | 1994 (15.6) | 917 | 14.0 | 1077 | 17.2 | <0.001 |
| None | 8041 (66.7) | 4355 | 70.4 | 3686 | 62.8 | <0.001 |
| Trauma Type | <0.001 | |||||
| Blunt | 10,553 (48.5) | 4461 | 44.0 | 5674 | 59.2 | |
| Penetrating | 9342 (43.0) | 5521 | 54.4 | 3748 | 39.1 | |
| Other | 1848 (8.5) | 166 | 1.6 | 163 | 1.7 | |
| Fracture Involvement1 | ||||||
| Pubic | 3065 (17.2) | 1633 | 18.6 | 1432 | 15.8 | <0.001 |
| Ilium | 7085 (38.5) | 4053 | 43.9 | 3032 | 33.0 | <0.001 |
| Ischium | 804 (4.6) | 504 | 5.9 | 300 | 3.3 | <0.001 |
| Sacrum and Coccyx | 2836 (15.9) | 1509 | 17.3 | 1327 | 14.6 | <0.001 |
| Multiple Fractures | 4351 (24.2) | 1549 | 17.6 | 2802 | 30.5 | <0.001 |
| Other Spec./Unsp. |
3376 (18.9) |
1724 |
19.7 |
1652 |
18.2 |
0.010 |
| Study Population | No Complication | Complication | ||||
| Mean ± SD |
Mean |
SD |
Mean |
SD |
p-value |
|
| Hospital LOS, days | 11.5 ± 17.0 | 6.9 | 9.5 | 17.2 | 22.0 | <0.001 |
ACS=American College of Surgeons; ED=Emergency department; BP=Systolic Blood Pressure; Pos.=Tested Positive; Spec./Unsp=Specified/Unspecified; LOS=Length of Stay; SD=Standard Deviation;∗Frequency based on valid cases.
Excluded from regression due to non-disjoint factors.
Excluded from regression due to missing data.
While controlling for the variables in the regression analysis (Table 4), independent risk factors of complications were examined. Alcohol abuse significantly increased the risk of complications (OR = 1.29, 95% CI: 1.14–1.46, p < 0.001), though substance abuse/dependence (OR = 0.70, 95% CI: 0.61–0.80, p < 0.001) and smoking (OR = 0.73, 95% CI: 0.67–0.79, p < 0.001) decreased the likelihood of being in the complication group. The largest contributor to predicting complications was presentation to the ED with a systolic blood pressure less than 90 mmHg (OR = 1.72, 95% CI: 1.55–1.91, p < 0.001). Males were 1.14 times as likely to develop complications compared to females (95% CI: 1.05–1.23, p < 0.001). Increasing age was a significant risk factor for the development of complications (OR = 1.01, 95% CI: 1.01–1.01, p < 0.001). The increase of one point on the GCS makes it 0.90 times as likely to develop a complication (95% CI: 0.89–0.91, p < 0.001). Additionally, an increase in one point of the ISS score, demonstrating increasing severity, corresponded with a 1.05 times increased likelihood of developing a complication. Hospital ACS Trauma Center verification level and trauma type (blunt versus penetrating) failed to have significant effects on the likelihood of complications.
Table 4.
Independent risk factors for complications after an open pelvic ring fracture using pooled data from multiple imputations.
| Risk Factor | OR | 95% CI | p-value |
|---|---|---|---|
| Age, years | 1.01 | 1.01–1.01 | <0.001 |
| Gender, ref to Female | 1.14 | 1.05–1.23 | 0.002 |
| Race | |||
| White | 0.87 | 0.42–1.80 | 0.679 |
| Black | 0.84 | 0.40–1.76 | 0.616 |
| Asian | 0.90 | 0.42–1.91 | 0.764 |
| American Indian, Pacific Islander | 1.00 | 0.43–2.33 | 0.994 |
| Other Race | 1.06 | 0.48–2.37 | 0.870 |
| Comorbidities | |||
| Alcohol Abuse | 1.29 | 1.14–1.46 | <0.001 |
| Substance Abuse/Dependence | 0.70 | 0.61–0.80 | <0.001 |
| Bleeding Disorder | 1.28 | 0.97–1.68 | 0.077 |
| Congestive Heart Failure | 1.08 | 0.68–1.71 | 0.736 |
| Chronic Obstructive Pulmonary Disease | 1.09 | 0.93–1.28 | 0.289 |
| Cerebrovascular Accident | 1.18 | 0.67–2.09 | 0.542 |
| Diabetes | 1.20 | 1.02–1.41 | 0.025 |
| Hypertension | 1.01 | 0.91–1.13 | 0.817 |
| Obesity | 0.83 | 0.72–0.95 | 0.008 |
| Smoker | 0.73 | 0.67–0.79 | <0.001 |
| Glasgow Coma Score | 0.90 | 0.89–0.91 | <0.001 |
| Injury Severity Score | 1.05 | 1.04–1.05 | <0.001 |
| ED Blood Pressure, less than 90 mm Hg | 1.72 | 1.55–1.91 | <0.001 |
| Pelvic Angiography | 1.08 | 0.87–1.34 | 0.473 |
| Trauma Type, ref to Blunt | |||
| Penetrating | 1.00 | 0.92–1.09 | 0.951 |
| Other | 1.10 | 0.86–1.40 | 0.459 |
| ACS Verification Level, ref to I | |||
| II | 1.10 | 1.00–1.21 | 0.039 |
| III | 0.72 | 0.53–0.98 | 0.037 |
| IV | 0.45 | 0.22–0.93 | 0.032 |
| Not Applicable | 1.05 | 0.98–1.13 | 0.166 |
OR=Odds Ratio; CI=Confidence Interval; ref=In reference to noted value; Holm-Bonferroni method used to adjust p-value to <0.003 for significance.
4. Discussion
Open pelvic fractures often present with associated injuries and substantial hemorrhage, leading to significant challenges in their management. In recent decades, advancements in treatment protocols focused on targeting coagulopathy, acidosis, and hypothermia (lethal triad of severe trauma) have contributed to the management of these deadly injuries.13,24, 25, 26 Previous published studies have lacked large cohorts, limiting their generalizability.10,13,27, 28, 29 Furthermore, a large portion of the epidemiological data of pelvic ring fracture morbidity and mortality are largely based on non-United States populations.9,30, 31, 32 Using the National Trauma Data Bank, our study sought to characterize risk factors for acute morbidity and mortality following open pelvic fractures, in a series of patient’s representative of the American population. In addition, this study evaluated the incidence of open pelvic fractures in the United States between 2007 and 2017 and identified prognostic factors associated with in hospital complications and mortality.
To our knowledge this is the largest study of open pelvic fractures in an adult population (N = 19,834). We found an overall mortality rate of 14% in all patients with an open pelvic fracture. After excluding patients who were determined to be dead on arrival to the ED (N = 552), the inpatient mortality rate was 11.6%. The current literature reports incredible variability in mortality rates after open fractures. Studies over the past four decades have shown mortality rates are decreasing over time. This can be partially attributed to the advances in care for trauma patients since the first published studies in the 1970’s. More recently, Mi et al. published a systematic review of open pelvic fractures combining data from 15 studies between 2005 and 2019.33 Their patient population had a mean age 35 years, ISS of 27, and 53-day hospital length of stay and reported a mortality rate of 23.7%.33 Our study examined a similar period of time, but found a lower mean ISS, shorter length of stay and decreased mortality. We believe our study is more indicative of the whole population given their cohort contained only 646 cases.
We found a predictably high rate of complications associated with open pelvic fractures (48.5%). The complications with the highest incidence were Pneumonia (5.8%), Deep and Organ Space Surgical Site Infection (4.1%), and Deep Vein Thrombosis (4.0%). Jones et al.29 analyzed complications in patients with open pelvic fractures and had a similar incidence of complications (46%). They reported incidence of PE (11%), UTI (22%) and wound infection (11%) which was in contrast to our rates of PE (1.3%), UTI (2%) and wound infection (5.6%).29 One reason for the discrepancy in our complication rates could be their study had a higher mean ISS (29 vs 23). Upon further examination of their treatment protocols, 30% were treated with bedrest and 20% were treated with traction. While the exact values for our cohorts’ surgical treatment plans are not available, we infer that far more patients in our study population were treated with surgical fixation owing to the advances in trauma management since Jones’ publication. Lastly, their study population consisted of 39 cases, so small variations in a few patients could have significantly altered the percentage observed.
Multivariate analysis identified increased age, lower GCS, higher ISS, BP less than 90 mmHg, and alcohol abuse to be independent risk factors for complications. Surprisingly, smokers and patients who tested positive for drugs were less likely to have a complication. This may be a result of insufficient data, or patients’ death before this information was collected. Trauma certification level also failed to predict complication, in contrast to a study by Oliphant et al.,34 which found that matched cohorts of pelvic fracture patients had significantly lower morbidity and mortality rates at Level 1 centers compared to Level 2. The various medical comorbidities that we studied did not significantly impact the risk for complications suggesting the risk of complications may be more influenced by the injury rather than patient specific medical comorbidities. While the presence of an airbag was more common in the non-complication group, all other protective devices had higher use in the complication group. Protective devices likely keep patients alive through injuries because patients who were pronounced dead on arrival were excluded in the analysis for demographics.
During trauma evaluation, increased ISS and lower GCS are accurate indicators for the overall gravity of injuries, and have previously shown to influence prognosis which was consistent with our findings.4,13,22,35 Hypotension on arrival, has been shown to be a risk factor for complication and mortality in trauma, as well as specifically patients with pelvic ring injuries.36,37 While the cutoff for hypotension has been debated, systolic blood pressure less than 90 (SBP<90 mmHg) is most consistently reported in the trauma literature.23 Gabbe et al.36 found that patients with severe pelvic ring injuries hypotensive on arrival were four times more likely to die than patients who were not hypotensive. In comparison to closed pelvic ring fractures, open pelvic fractures are commonly associated with treatment-resistant hemorrhage. This is thought to be from the lack of tamponade effect that occurs with closed injuries.38
As with any study using a large database, there are several limitations that should be noted. The data quality of the National Trauma Data Bank (NTDB) is dependent on the quality of each individual hospital’s data submission.39 Limitations inherently associated with the use of trauma registries include selection bias, inconsistency in the measurement of clinical variables, and differences between hospitals. Participation in the NTDB is voluntary, which can lead to a convenience sample, possibly introducing selection and information bias. Therefore, the data may not be a representation of all hospitals in the United States, and has not been systematically selected to represent a specific population. With this in mind, every effort was made to specify inclusion criteria for analysis to create a homogenous population. Despite the rigorous efforts by NTDB to continuously clean and standardize their data to improve quality, information can be lost or coded incorrectly. Missing information such as specific transfusion requirements and time between procedures was either missing or inconsistent in our dataset and was not included in our analysis. Fracture stability and classification were not coded in our dataset and therefore is a limitation of our study. Specific classifications and severity of pelvic fractures have been shown to correlate with complications and mortality.33,35 Without the ability to analyze diagnostic imaging, we were only able to stratify fractures based on the location of injury in the pelvic ring based on the ICD-9 and ICD-10 codes. Nevertheless, missing information, and associated bias, was minimized with the aforementioned strategies in the methods section. Lastly, although our results evaluate overall risks of inpatient mortality and complications, the use of the NTDB does not track complications and mortality after hospital discharge. Therefore, our results are likely more conservative than a sample with follow up data available for analysis. Despite the limitations, the results of our study add valuable information on outcomes of open pelvic fractures.
5. Conclusion
Morbidity and mortality remain high following open pelvic fractures. The most predictive risk factors that appear to influence outcomes include GCS, ISS, and hemodynamic stability on arrival. Any opportunity to decrease complications and possibly lower mortality, is of vital importance when managing these patients. This study is the largest overview on open pelvic fractures to date, and provides a summary of the complications and risk factors. The sampling strategy, uniformity of incident-level data, and methodology of the NTDB, have allowed us to build on prior literature reliant on data from single trauma systems and smaller sample sizes.
Dataset
A query was performed for patients treated at trauma centers using the 2007 to 2017 American College of Surgeons National Trauma Data Bank (NTDB) Participant Use File/Research Dataset. https://www.facs.org/quality-programs/trauma/tqp/center-programs/ntdb.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Authorship
N.F, C.I, C.B, P.R, G.K, M.L each have made substantial contributions to the conception, design, and drafting of the manuscript. All authors have approved the submitted version of the manuscript.
Declaration of competing interest
M.L has given lectures for Depuy Synthes. The rest of the authors declare that they have no competing interests.
Acknowledgments
The authors have no acknowledgments.
Contributor Information
Nicholas Frane, Email: nfrane@northwell.edu.
Cesar Iturriaga, Email: citurriaga@northwell.edu.
Christine Bub, Email: Cbub@northwell.edu.
Peter Regala, Email: pregala@northwell.edu.
Gus Katsigiorgis, Email: GusKats@aol.com.
Michael Linn, Email: MLinn@northwell.edu.
Appendix 1.
| ICD-9-CM Code | Diagnosis |
|---|---|
| 808.3 | Open fracture of pubis |
| 808.51 | Open fracture of ilium |
| 808.52 | Open fracture of ischium |
| 808.53 | Multiple open pelvic fractures with disruption of pelvic circle |
| 805.7 | Open fracture of sacrum and coccyx |
| 808.54 | Multiple open pelvic fractures without disruption of pelvic circle |
| 808.59 | Open fracture of other specified part of pelvis |
| 808.9 | Unspecified open fracture of pelvis |
ICD-9-CM: International Classification of Disease, Ninth Revision, Clinical Modification.
Appendix 2.
| ICD-10-CM Code | Diagnosis + initial encounter for open fracture |
|---|---|
| S32.10XB | Unspecified fracture of sacrum |
| S32.110B | Zone I fracture of sacrum |
| S32.111B | Minimally displaced Zone I fracture of sacrum |
| S32.112B | Severely displaced Zone I fracture of sacrum |
| S32.119B | Unspecified Zone I fracture of sacrum |
| S32.120B | Zone II fracture of sacrum |
| S32.121B | Minimally displaced Zone II fracture of sacrum |
| S32.122B | Severely displaced Zone II fracture of sacrum |
| S32.129B | Unspecified Zone II fracture of sacrum |
| S32.130B | Zone III fracture of sacrum, |
| S32.131B | Minimally displaced Zone III fracture of sacrum |
| S32.132B | Severely displaced Zone III fracture of sacrum |
| S32.139B | Unspecified Zone III fracture of sacrum |
| S32.14XB | Type 1 fracture of sacrum |
| S32.15XB | Type 2 fracture of sacrum |
| S32.16XB | Type 3 fracture of sacrum |
| S32.17XB | Type 4 fracture of sacrum |
| S32.19XB | Other fracture of sacrum |
| S32.2XXB | Fracture of coccyx, initial encounter for open fracture |
| S32.301B | Unspecified fracture of right ilium |
| S32.302B | Unspecified fracture of left ilium |
| S32.309B | Unspecified fracture of unspecified ilium |
| S32.311B | Avulsion fracture of ilium |
| S32.312B | Displaced avulsion fracture of left ilium |
| S32.313B | Displaced avulsion fracture of unspecified ilium |
| S32.314B | Nondisplaced avulsion fracture of right ilium |
| S32.315B | Nondisplaced avulsion fracture of left ilium |
| S32.316B | Nondisplaced avulsion fracture of unspecified ilium |
| S32.391B | Other fracture of right ilium |
| S32.392B | Other fracture of left ilium |
| S32.399B | Other fracture of unspecified ilium |
| S32.501B | Unspecified fracture of right pubis |
| S32.502B | Unspecified fracture of left pubis |
| S32.509B | Unspecified fracture of unspecified pubis |
| S32.511B | Fracture of superior rim of right pubis |
| S32.512B | Fracture of superior rim of left pubis |
| S32.519B | Fracture of superior rim of unspecified pubis |
| S32.591B | Other specified fracture of right pubis |
| S32.592B | Other specified fracture of left pubis |
| S32.599B | Other specified fracture of unspecified pubis |
| S32.601B | Unspecified fracture of right ischium |
| S32.602B | Unspecified fracture of left ischium |
| S32.609B | Unspecified fracture of unspecified ischium |
| S32.611B | Avulsion fracture of ischium |
| S32.612B | Displaced avulsion fracture of left ischium |
| S32.613B | Displaced avulsion fracture of unspecified ischium |
| S32.614B | Nondisplaced avulsion fracture of right ischium |
| S32.615B | Nondisplaced avulsion fracture of left ischium |
| S32.616B | Nondisplaced avulsion fracture of unspecified ischium |
| S32.691B | Other specified fracture of right ischium |
| S32.692B | Other specified fracture of left ischium |
| S32.699B | Other specified fracture of unspecified ischium |
| S32.810B | Multiple fractures of pelvis with stable disruption of pelvic ring |
| S32.811B | Multiple fractures of pelvis with unstable disruption of pelvic ring |
| S32.82XB | Multiple fractures of pelvis without disruption of pelvic ring |
| S32.89XB | Fracture of other parts of pelvis |
| S32.9XXB | Fracture of unspecified parts of lumbosacral spine and pelvis |
ICD-10-CM: International Classification of Disease, Tenth Revision, Clinical Modification.
References
- 1.Buller L.T., Best M.J., Quinnan S.M. A nationwide analysis of pelvic ring fractures: incidence and trends in treatment, length of stay, and mortality. Geriatr Orthop Surg Rehabil. 2016;7(1):9–17. doi: 10.1177/2151458515616250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Balogh Z., King K.L., Mackay P. The epidemiology of pelvic ring fractures: a population-based study. J Trauma Inj Infect Crit Care. 2007;63(5):1066–1072. doi: 10.1097/TA.0b013e3181589fa4. [DOI] [PubMed] [Google Scholar]
- 3.Gänsslen A., Pohlemann T., Paul C., Lobenhoffer P., Tscherne H. Epidemiology of pelvic ring injuries. Injury. 1996;27(SUPPL. 1):13–20. doi: 10.1016/s0020-1383(96)90106-0. [DOI] [PubMed] [Google Scholar]
- 4.Court-Brown C.M., Caesar B. Epidemiology of adult fractures: a review. Injury. 2006;37(8):691–697. doi: 10.1016/j.injury.2006.04.130. [DOI] [PubMed] [Google Scholar]
- 5.Brenneman F.D., Katyal D., Boulanger B.R., Tile M., Redelmeier D.A. Long-term outcomes in open pelvic fractures. J Trauma Inj Infect Crit Care. 1997;42:773–777. doi: 10.1097/00005373-199705000-00003. [DOI] [PubMed] [Google Scholar]
- 6.Grotz M.R.W., Allami M.K., Harwood P., Pape H.C., Krettek C., Giannoudis P.V. Open pelvic fractures: epidemiology, current concepts of management and outcome. Injury. 2005;36(1):1–13. doi: 10.1016/j.injury.2004.05.029. [DOI] [PubMed] [Google Scholar]
- 7.Maull K.I., Sachatello C.R., Ernst C.B. The deep perineal laceration-an injury frequently associated with open pelvic fractures: a need for aggressive surgical management. A report of 12 cases and review of the literature. J Trauma. 1977;17(9):685–696. http://www.ncbi.nlm.nih.gov/pubmed/894758 [PubMed] [Google Scholar]
- 8.Perry J.F. Pelvic open fractures. Clin Orthop Relat Res. 1980;151:41–45. doi: 10.1097/00003086-198009000-00007. [DOI] [PubMed] [Google Scholar]
- 9.Hermans E., Edwards M.J.R., Goslings J.C., Biert J. Open pelvic fracture: the killing fracture? J Orthop Surg Res. 2018;13(1):83. doi: 10.1186/s13018-018-0793-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cannada L.K., Taylor R.M., Reddix R. The Jones-Powell Classification of open pelvic fractures: a multicenter study evaluating mortality rates. J Trauma Acute Care Surg. 2013;74(3):901–906. doi: 10.1097/TA.0b013e3182827496. [DOI] [PubMed] [Google Scholar]
- 11.Burlew C.C., Moore E.E., Stahel P.F. Preperitoneal pelvic packing reduces mortality in patients with life-threatening hemorrhage due to unstable pelvic fractures. J Trauma Acute Care Surg. 2017;82:233–242. doi: 10.1097/TA.0000000000001324. Lippincott Williams and Wilkins. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dente C.J., Feliciano D.V., Rozycki G.S. The outcome of open pelvic fractures in the modern era. Am J Surg. 2005;190(6):830–835. doi: 10.1016/j.amjsurg.2005.05.050. [DOI] [PubMed] [Google Scholar]
- 13.Siada S.S., Davis J.W., Kaups K.L., Dirks R.C., Grannis K.A. Current outcomes of blunt open pelvic fractures: how modern advances in trauma care may decrease mortality. Trauma Surg acute care open. 2017;2(1) doi: 10.1136/tsaco-2017-000136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Richardson J.D., Harty J., Amin M., Flint L.M. Open pelvic fractures. J Trauma Inj Infect Crit Care. 1982;22(7):533–538. doi: 10.1097/00005373-198207000-00002. [DOI] [PubMed] [Google Scholar]
- 15.Raffa J., Christensen N.M. Compound fractures of the pelvis. Am J Surg. 1976;132(2):282–286. doi: 10.1016/0002-9610(76)90061-1. [DOI] [PubMed] [Google Scholar]
- 16.Rothenberger D., Velasco R., Strate R., Fischer R.P., Perry J.F. Open pelvic fracture: a lethal injury. J Trauma. 1978;18(3):184–187. http://www.ncbi.nlm.nih.gov/pubmed/642044 [PubMed] [Google Scholar]
- 17.Hanson P.B., Milne J.C., Chapman M.W. Open fractures of the pelvis. Review of 43 cases. J Bone Joint Surg Br. 1991;73(2):325–329. doi: 10.1302/0301-620X.73B2.2005166. http://www.ncbi.nlm.nih.gov/pubmed/2005166 [DOI] [PubMed] [Google Scholar]
- 18.Ferrera P.C., Hill D.A. Good outcomes of open pelvic fractures. Injury. 1999;30(3):187–190. doi: 10.1016/s0020-1383(98)00246-0. [DOI] [PubMed] [Google Scholar]
- 19.Jones A.L., Powell J.N., Kellam J.F., McCormack R.G., Dust W., Wimmer P. Open pelvic fractures. A multicenter retrospective analysis. Orthop Clin N Am. 1997;28(3):345–350. doi: 10.1016/s0030-5898(05)70293-5. [DOI] [PubMed] [Google Scholar]
- 20.Giordano V., Koch H.A., Gasparini S., Serrão de Souza F., Labronici P.J., do Amaral N.P. Open pelvic fractures: review of 30 cases. Open Orthop J. 2017;10(1):772–778. doi: 10.2174/1874325001610010772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rommens P.M., Hessmann M.H. Staged reconstruction of pelvic ring disruption: differences in morbidity, mortality, radiologic results, and functional outcomes between B1, B2/B3, and C-type lesions. J Orthop Trauma. 2002;16(2):92–98. doi: 10.1097/00005131-200202000-00004. [DOI] [PubMed] [Google Scholar]
- 22.Arroyo W., Nelson K.J., Belmont P.J., Bader J.O., Schoenfeld A.J. Pelvic trauma: what are the predictors of mortality and cardiac, venous thrombo-embolic and infectious complications following injury? Injury. 2013;44(12):1745–1749. doi: 10.1016/j.injury.2013.08.007. [DOI] [PubMed] [Google Scholar]
- 23.Lalezarzadeh F., Wisniewski P., Huynh K., Loza M., Gnanadev D. Evaluation of prehospital and emergency department systolic blood pressure as a predictor of in-hospital mortality. Am Surg. 2009;75(10):1009–1014. doi: 10.1177/000313480907501032. [DOI] [PubMed] [Google Scholar]
- 24.Moffatt S.E. Hypothermia in trauma. Emerg Med J. 2013;30(12):989–996. doi: 10.1136/emermed-2012-201883. [DOI] [PubMed] [Google Scholar]
- 25.Black E.A., Lawson C.M., Smith S., Daley B.J. Open pelvic fractures: the University of Tennessee Medical Center at Knoxville experience over ten years. Iowa Orthop J. 2011;31:193–198. http://www.ncbi.nlm.nih.gov/pubmed/22096441 [PMC free article] [PubMed] [Google Scholar]
- 26.Jarvis S., Orlando A., Blondeau B. Variability in the timeliness of interventional radiology availability for angioembolization of hemodynamically unstable pelvic fractures: a prospective survey among U.S. level i trauma centers. Patient Saf Surg. 2019;13(1):23. doi: 10.1186/s13037-019-0201-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wang Z., hao Li, nan K. Regional injury classification and treatment of open pelvic fractures. Orthop Surg. 2019;11(6):1064–1071. doi: 10.1111/os.12554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dong J.L., Zhou D.S. Management and outcome of open pelvic fractures: a retrospective study of 41 cases. Injury. 2011;42(10):1003–1007. doi: 10.1016/j.injury.2011.01.032. [DOI] [PubMed] [Google Scholar]
- 29.Jones A.L., Powell J.N., Kellam J.F., McCormack R.G., Dust W., Wimmer P. Open pelvic fractures. A multicenter retrospective analysis. Orthop Clin N Am. 1997;28(3):345–350. doi: 10.1016/s0030-5898(05)70293-5. [DOI] [PubMed] [Google Scholar]
- 30.Holstein J.H., Culemann U., Pohlemann T. Vol. 470. Springer New York LLC; 2012. What are predictors of mortality in patients with pelvic fractures? Trauma; pp. 2090–2097. (Clinical Orthopaedics and Related Research). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hauschild O., Strohm P.C., Culemann U. Mortality in patients with pelvic fractures: results from the German pelvic injury register. J Trauma Inj Infect Crit Care. 2008;64(2):449–455. doi: 10.1097/TA.0b013e31815982b1. [DOI] [PubMed] [Google Scholar]
- 32.Pohlemann T., Tscherne H., Baumgärtel F. [Pelvic fractures: epidemiology, therapy and long-term outcome. Overview of the multicenter study of the Pelvis Study Group] Unfallchirurg. 1996;99(3):160–167. http://www.ncbi.nlm.nih.gov/pubmed/8685720 [PubMed] [Google Scholar]
- 33.Mi M., Kanakaris N.K., Wu X., Giannoudis P.V. Management and outcomes of open pelvic fractures: an update. Injury. February 2020 doi: 10.1016/j.injury.2020.02.096. [DOI] [PubMed] [Google Scholar]
- 34.Oliphant B.W., Tignanelli C.J., Napolitano L.M., Goulet J.A., Hemmila M.R. American College of surgeons committee on trauma verification level affects trauma center management of pelvic ring injuries and patient mortality. J Trauma Acute Care Surg. 2019;86(1):1–10. doi: 10.1097/TA.0000000000002062. [DOI] [PubMed] [Google Scholar]
- 35.Song W., Zhou D., Xu W. Factors of pelvic infection and death in patients with open pelvic fractures and rectal injuries. Surg Infect. 2017;18(6):711–715. doi: 10.1089/sur.2017.083. [DOI] [PubMed] [Google Scholar]
- 36.Gabbe B.J., De Steiger R., Esser M., Bucknill A., Russ M.K., Cameron P.A. Predictors of mortality following severe pelvic ring fracture: results of a population-based study. Injury. 2011;42(10):985–991. doi: 10.1016/j.injury.2011.06.003. [DOI] [PubMed] [Google Scholar]
- 37.Damme C.D., Luo J., Buesing K.L. Isolated prehospital hypotension correlates with injury severity and outcomes in patients with trauma. Trauma Surg Acute Care Open. 2016;1(1) doi: 10.1136/tsaco-2016-000013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Langford J.R., Burgess A.R., Liporace F.A., Haidukewych G.J. Pelvic fractures: Part 1. Evaluation, classification, and resuscitation. J Am Acad Orthop Surg. 2013;21(8):448–457. doi: 10.5435/JAAOS-21-08-448. [DOI] [PubMed] [Google Scholar]
- 39.National trauma Data Bank ® NTDB research data set user manual and variable description list. www.ntdb.org