17.18 Limitations Of PPV-Driven Models Of ACS Trauma Triage Criteria

Posted on January 18, 2016

M. Mukit1, B. A. Asif1, M. Subrize1, H. C. Thode1, M. C. Henry1 1Stony Brook University Medical Center,Stony Brook, NY, USA

Introduction:

From 2003-2007, trauma prehospital care reports (TPCR) were used to delineate trauma data from the field; the TPCR includes the American College of Surgeon’s (ACS) triage criteria for traumatic events. We used these reports together with the Suffolk Country Trauma Registry Data to examine which criteria had the greatest positive predictive value for trauma resource need.

Methods:

Retrospective structured study. Cases were matched with the County’s regional Trauma Registry. The study was approved by our IRB. All adult patients with a TPCR from 2003 to 2007 were eligible for inclusion.The occurrences, sensitivity, specificity, PPV and NPV of each trauma triage criteria for the major outcomes of ISS > 15 and major OR/death were noted. Models were created to determine which trauma triage criteria would allow the greatest sensitivity, specificity, and NPVfor predicting ISS >15 or major OR/death while maintaining a minimum PPV of 20%. The decision rules for both models were to select variables with the largest PPV and to stop selection when the PPV of the overall model was < 20.

Results:
17120 TPCR’s were analyzed. The criteria that were significant in the ISS>15 model, in order,were flail chest, RR < 10 or > 28, amputation, multitrauma, long bone fracture, GCS <14, SBP <90, death of same vehicle occupant, ejection from vehicle and pelvic fracture. This model achieved a sensitivity of 46%, specificity of 97%, PPV of 20% and NPV of 99%. The criteria that were significant in the major OR/death, in order, were flail chest, long bone fracture, amputation, RR < or > 28, multitrauma, SBP < 90, GCS < 14 and death of samevehicle occupant. This model achieved a sensitivity of 38%,specificity of 97%, PPV of 22% and NPV of 99%.

Conclusion:

The goal of trauma triage guidelines is to minimize both under triage (so patients can receivethe specialized trauma care they require) and over triage (so patients do not consume scarce resources unnecessarily). In the 2011 Guidelines for Field Triage of Injured Patients, the panel decided to only include criteria that had a minimum PPV of 20% for predicting severe injury.Our study shows that models that place PPV at the lower limit of 20 are able to achieve high specificity and NPV but poor sensitivity. Such models would result in little over triage but significant under triage. Additionally, the first eight criteria for both models (ISS>15 and major OR/death) were the same, indicating that the same criteria are predictive for different outcomes.

17.19 Comparison of Clinical Outcomes after Extremity Vascular Repair Performed by Different Services

Posted on January 18, 2016

A. Hassan1, P. Rhee1, A. A. Haider1, N. Kulvatunyou1, T. O’Keeffe1, A. Tang1, R. Latifi1, L. Gries1, G. Vercruysse1, F. Randall1, B. Joseph1 1University Of Arizona,Trauma,Tucson, AZ, USA

Introduction:
Studies have associated surgeon’s volume with improved outcomes after elective vascular procedures. However, the difference in outcomes of vascular repairs performed by surgeons from different services has never been studied in trauma. The aim of this study was to compare outcomes between extremity vascular repair performed by vascular and trauma surgeons at a level I trauma center.

Methods:
We performed a 5-year retrospective analysis of all patients with Brachial, Femoral or Popliteal artery injury who presented to our level I trauma center. The operating surgeon was identified from patient medical records along with other details of vascular injuries, radiological findings, operative repairs and complications. Regression analysis was performed to compare outcomes of amputations, complications, mortality and hospital length of stay.

Results:
A total of 76 patients were included. Mean age was 31 ± 13 years; median extremity AIS [IQR] score was 3 [3 – 3], and overall complication, amputation, and mortality rates were 35.5%, 9.2%, and 5.3% respectively. Vascular repairs in 53.9%(41/76) patients were performed by trauma surgeons and in 46.1%(35/76) by vascular surgeons. The details of the injuries and repairs are as in Tabel 1.
After adjusting for possible confounders in regression analysis, there was no difference in outcomes between the two surgeon groups for amputations (OR [95% CI] = 0.50 [0.06 – 4.23], p=0.52), complications (OR [95% CI] = 0.71 [0.24 – 2.10], p=0.54), mortality (OR [95% CI] = 0.49 [0.02 – 13.92], p=0.68) or hospital length of stay (β [95% CI] = -1.25 [-7.18 – 4.67], p=0.67).

Conclusion:

Extremity vascular injuries remain a relatively scarce entity in trauma. Despite differences in the type of injuries and operative repairs performed for extremity vascular injuries, trauma and vascular surgeons provide equivalent level of care for extremity vascular injuries.

17.14 Predicting The Need For Surgical Intervention In Acute Pancreatitis

Posted on January 18, 2016

M. Magarakis1, N. M. Vranis1, B. Bruns1, R. B. Tesoriero1, S. Sivaraman1, C. A. Sadler1, H. Desai1, J. Diaz1 1University Of Maryland,Acute Care Emergency Surgery,Baltimore, MD, USA

Introduction:

Hospital admissions for Acute Pancreatitis in the United States exceed 300.000 per year, with annual costs of 2 billion dollars. Approximately 15-25% of patients develop a severe course with mortality as high as 50%. Predicting which patients will require surgical intervention remains challenging, and delays to definitive care may worsen outcomes. We aimed to identify risk factors that predict the need for surgical intervention (SI) in this patient population

Methods:

A retrospective review of patients admitted with acute pancreatitis from January 2011 to December 2013 was performed from a prospectively collected Emergency General Surgery Registry. Charts were reviewed for patient demographics, admission type, laboratory values, etiology of pancreatitis, CT severity index, presence of pancreatic necrosis or infection, ICU length of stay, development of organ failure, and need for SI

Results:

386 patients were admitted with acute pancreatitis. There were 68% males and 32% females. Fifty-eight patients received SI. Using multivariable regression analysis, factors associated with SI included: direct admission (OR=3.6, 95% CI 1.3, 9.7, p<0.05), WBC>16 (OR=2.8, 95% CI 1.2, 6.6, p<0.05), respiratory failure (OR=5.5, 95% CI 2.3, 13.4, p<0.001), and presence of necrosis/infection (OR=9.3, 95% CI 3.4, 25.2, p<0.0001)

Conclusion:

Acute pancreatitis is a complex disease with an unpredictable course. Identification of patients who will require SI remains difficult despite multiple severity classification systems. In our review; direct admission, WBC > 16, presence of respiratory failure, and pancreatic necrosis/infection significantly increased the odds of receiving SI. Identification of risk factors for SI may allow more time efficient surgical consultation/admission and improve outcomes

17.15 Cirrhosis and Traumatic Brain Injury: Survival and Discharge to Independent Living are Unlikely

Posted on January 18, 2016

B. Chung1, P. Miller1, J. J. Hoth1 1Wake Forest University School Of Medicine,General Surgery,Winston-Salem, NC, USA

Introduction: Trauma patients with cirrhosis are known to have a poor prognosis. The subset of injured patients (pts) with traumatic brain injury (TBI) and cirrhosis would be expected to do poorly, but the impact of this combination on outcome and resource utilization has not been addressed. The purpose of this study is to evaluate the outcomes of cirrhosis in conjunction with TBI.

Methods: A retrospective review was conducted through our trauma registry from 2010-2014. Cirrhotic pts with TBI were identified and compared to a non-cirrhotic group that was matched 1:1 for gender, age, type of TBI, Abbreviated Injury Score for the respective TBI, Injury Severity Score, and other traumatic injury. Student’s t-test and χ 2 were used for analysis and a p<0.05 considered significant.

Results: 24 pts with cirrhosis and TBI were identified and matched to 24 non-cirrhosis control pts. Pts with cirrhosis had a higher mortality (37% vs. 4%, p=0.01), longer length of stay (11.25 days vs. 5.79 days, p=0.008), ICU days (5.67 vs. 1.91, p=0.007), and ventilator days (4.04 vs. 0.75, p=0.003). In addition, the cirrhosis group had increased incidence of infection/sepsis (14 vs. 2, p=0.005) and need for transfusions (13 vs. 2, p=0.001). Furthermore, 66% of surviving cirrhosis pts were discharged to a skilled nursing or rehabilitation facility vs. 13% in the control group (p=0.0001).

Conclusion: Cirrhosis in TBI is an exceptionally poor prognostic indicator, with more than 1 in 3 dying of their injuries. The average length of stay, ICU stay, and ventilator days in cirrhotics is also significantly longer compared to a matched non-cirrhosis group. Furthermore, they are larger consumers of resources since they are at higher risk of sepsis and need for transfusions. This indicates an increased utilization of hospital resources. In addition, 2/3 of pts that survive require a skilled nursing or acute rehabilitation facility on discharge. Overall, only 21% (5/24 pts) from the cirrhosis group are discharged to home, compared to 83% (20/24 pts) in the non-cirrhosis group. For cirrhotic patients with TBI that survive their hospitalization, there is further increased utilization of outpatient health resources. These data can be used to inform discussions of realistic goals of care in this difficult population.

17.17 Variation In MRI Use For Cervical Spine Clearance In Obtunded Blunt Trauma Patients

Posted on January 18, 2016

A. Albaghdadi1, J. Canner1, E. B. Schneider1, C. B. Feather1, J. Odden1, E. R. Haut1 1Johns Hopkins University School Of Medicine,Baltimore, MD, USA

Introduction:

Controversy remains about the ideal approach for cervical spine clearance in obtunded, blunt, adult trauma patients. An EAST practice management guideline was recently published suggesting that MRI is not necessary in hopes of standardizing care. We aimed to identify national variation of MRI use to clear the c-spine in obtunded trauma patients and describe patient- and hospital-level factors associated with its use.

Methods:

We conducted a retrospective review of the NTDB from 2007 to 2012. We included blunt trauma patients >=18 years, treated at level 1 or 2 trauma centers (TCs), with a GCS<=8, Head AIS>3 and mechanically ventilated >72 hours. The proportion of patients undergoing MRI at each hospital was calculated. Multi-level modeling was used to identify patient- and hospital- level factors associated with MRI use.

Results:

We included 32,125 patients treated at 395 unique TCs. The mean proportion of MRI over the entire sample was 9.9%. Amongst the 181 hospitals (57.8% of all admissions) that performed and reported MRIs, the proportions of patients who received MRI per hospital ranged from 0.5-68.4%. (Figure) Younger patients, MVC and pedestrian injuries were more likely to receive an MRI. Injury severity (ISS) was not associated with MRI use. Hospitals in the northeast, level 1 TCs, and non-teaching hospitals were more likely to perform MRI.

Conclusion:

After controlling for patient-level characteristics, variation remained in MRI use based on hospital specific, geographic, trauma center level, and teaching status characteristics. Cervical spine clearance protocol implementation based on the new EAST guideline may standardize care, reduce variation in practices, and decrease healthcare costs.

17.09 The Influence of Poverty Level on Urban Versus Rual Trauma Patients

Posted on January 18, 2016

P. J. Schenarts1, R. Cooper1, L. Schlitzkus1, K. Buesing1, C. Evans1, M. Goede1, J. Stothert1 1University Of Nebraska,Acute Care Surgery,Omaha, NE, USA

Introduction: There is a strong relationship of poverty in a rural setting with increased trauma deaths. To date, there are no studies characterizing the influence of poverty on trauma patients from urban vs. rural environments. Therefore our purpose was to determine if there were differences in urban vs. rural trauma patients based upon poverty level.

Methods: Data from 46 hospitals collected in the Nebraska Trauma Registry between 2007-2012 was geocoded by patient’s home address. Census information regarding income level was used to stratify patients into three income groups (low, middle & high), based on percentage of the census tract below the poverty line. Rural was defined on the county level, using county health district designations. Differences in demographics, insurance status, mode of transport, injury severity score (ISS), hospital length of stay (LOS), need for health services after hospital discharge and mortality were compared between urban and rural patients within each income group. Statistical analysis: ANOVA and chi-square

Results: Of 54,592 patients, 36,309 were urban; 18,283 were rural. In table at bottom, data is presented as mean ± SD, or percent.. In all income groups urban patients are significantly younger (p<0.001) (Low income 44 ± 25 vs 52 ± 29 yrs, Middle Income 52 ± 28 vs 56 ± 28 yrs, High income 47 ± 28 vs 52 ± 28 yrs); utilize EMS more frequently (p<0.001) (Low income 70 vs 61%, Middle Income 68 vs 63%, High income 72 vs 64%) and have longer hospital LOS (p<0.001) (Low income 3.8 ± 9 vs 2.6 ± 11 days, Middle Income 3.7 ± 5.1 vs 2.9 ± 4.4 days, High income 3.6 ± 5.1 vs 2.2 ± .4 days).

Conclusion: ~~Differences between urban and ruralprimarily occur in the low and middle income groups. Across all income levels, fewer rural patients are self-pay. In the low and middle income groups, urban patients have more penetrating injuries and are hospitalized more frequently for minor injuries (ISS<14). Mortality is higher in the rural low income group, but higher in the urban middle income group.

17.11 Resuscitative Thoracotomy for Traumatic Arrest, an Analysis of a National Trauma Dataset

Posted on January 18, 2016

K. Williams1, S. N. Zafar1, Z. Hashmi2, N. Changoor1, S. A. Zafar3, J. Hwabejire1, A. Haider4, E. E. Cornwell1 1Howard University College Of Medicine,Surgery,Washington, DC, USA 2Sinai Hospital,Baltimore, MD, USA 3University Hospital Limerick,Surgery,Limerick, , Ireland 4Brigham And Women’s Hospital,Boston, MA, USA

Introduction: Emergency resuscitative thoracotomy (ERT) is a potentially life-saving procedure that is infrequently performed. Evaluation of its efficacy is controversial and has been limited to bivariate analyses. Studying outcomes from a large national database allows for more robust analytic techniques and a better understanding of its current use in both penetrating and blunt trauma.

Methods: We analyzed the National Trauma Data Bank for the period 2007 to 2012. We selected all patients who underwent an emergency thoracotomy for traumatic arrest (no vital signs after sustaining an injury). Logistic regression analyses were utilized to assess factors associated with survival after ERT. Propensity score matching was used to match patients receiving ERT with similarly injured patients in traumatic arrest that did not undergo ERT. Separate analyses were performed for blunt and penetrating injuries.

Results: Of the 4 million records in the dataset, 31,528 (0.9%) patients presented in traumatic arrest. ERT was performed on 2,223 (7%) patients. The mean age was 30.6 (±20) years, 85% were male, 45% were Black, 25% White and 21% were Hispanic, 55% were uninsured, 75% suffered severe thoracic injuries, 75% arrived via ground ambulance. Penetrating injury occurred in 80% of patients and the overall survival was 2.2%. On multivariate analysis patients more likely to survive were Black compared with White, insured, with stab wounds, severe chest injuries, and those transported by private vehicle versus ground ambulance. After propensity matching, patients with penetrating injuries were 44% more likely to survive if they underwent an ERT (OR = 0.56, 95% CI = 0.38-0.83). However, for blunt injuries there was no survival benefit of ERT (OR= 0.73 95% CI = 0.39-1.36).

Conclusion: Outcomes for ERT among patients with traumatic arrest have been reviewed in a large trauma dataset. Survival among patients presenting with no vital signs is very low and the utility of ERT in this population may need reconsideration especially for blunt injuries.

17.12 Sports and Recreation-Related Ocular Injuries

Posted on January 18, 2016

R. S. Haring1,2,3,4, I. D. Sheffield5, J. K. Canner3, A. H. Haider1,2, E. B. Schneider1,2,3,4 1Harvard School Of Medicine,Brookline, MA, USA 2Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 3Johns Hopkins University School Of Medicine,Johns Hopkins Surgery Center For Outcomes Research,Baltimore, MD, USA 4Johns Hopkins Bloomberg School Of Public Health,Dept. Of Health Systems And Policy,Baltimore, MD, USA 5Brigham Young University,Provo, UT, USA

Introduction: Ocular injuries can have long-term sequelae and substantially impact quality of life. Currently available data on the incidence and overall burden of sports-related ocular trauma are limited or outdated. In 2015, the Centers for Disease Control and Prevention published a comprehensive model for the classification of sports and recreation-related injury. We applied this model to estimate and characterize the burden of sports and recreation-related (S/R) ocular injury in the United States.

Methods: Using the Nationwide Emergency Department Sample, we identified patients presenting with a diagnosis of ocular trauma from 2006-2012. We then examined ICD-9CM codes to identify individuals with S/R ocular injuries using the CDC’s comprehensive classification system. Age, sex, external mechanism of injury, type of S/R activity, and other factors were used to characterize and stratify injuries. Comparisons were made within and between injury strata across time. Data specific to individual team sports were not readily available until 2010; a subset was created to characterize those injuries from 2010-2012.

Results: A total of 287,718 ED visits associated with a diagnosis of S/R-related ocular injury occurred from 2006-2012. Males represented 78.8% of cases, and that proportion did not vary significantly across the 7-year period. The proportion of all ocular trauma cases that were S/R-related rose a relative 36.3%, from 3.8% of all injuries in 2006 to 5.2% in 2012. Overall, the leading single cause of S/R ocular injury was pedal cycling—an activity resulting in 34,965 (12.2%) S/R ocular injuries. The number of patients presenting cycling-related ocular injuries increased from 4,076 in 2006 to 5,623 in 2012. Among team sports, basketball resulted in the highest number of ocular injuries, with 17,018 patients presenting to the ED between 2010 and 2012. The next most common sport was baseball (12,734), followed by soccer (5,787), football (4,844), and watersports (2,063). During the study period, 291 patients were hospitalized with baseball-related ocular injury, 82 for basketball-related eye injury, 59 for soccer, and 45 for football-related injuries.

Conclusions: Despite an overall reduction in the number of all-cause ocular trauma cases reporting to the ED across the study period, the absolute number of S/R ocular trauma cases presenting for care increased significantly. The observed increase in S/R ocular trauma presentations appears to be driven in part by a 38.0% increase in the number of bicycle-related ocular injuries, which tend to be more severe (23.9% of cases resulting in hospitalization). Basketball remains the leading cause of S/R ocular injuries among the team sports, but hospitalization rates for baseball are 4.6 times higher than those for basketball (2.3% vs 0.5%). Efforts aimed at preventing serious vision-threatening injury may be most effectively focused on high energy S/R activities such as cycling and baseball.

17.13 Implanted Cardioverter Defibrillators and Pacemakers: Prevalance and Outcomes After Geriatric Trauma

Posted on January 18, 2016

M. Altieri1, I. Almasry2, M. J. Shapiro1, J. A. Vosswinkel1, J. E. McCormack1, E. C. Huang1, P. Eckardt3, R. S. Jawa1 1Stony Brook University Medical Center,Trauma,Stony Brook, NY, USA 2Stony Brook University Medical Center,Cardiology,Stony Brook, NY, USA 3Stony Brook University Medical Center,Nursing,Stony Brook, NY, USA

Introduction: The prevalence of Implanted cardioverter defibrillators (ICDs) and pacemakers (PMs) in the US population is increasing. However, the prevalence and outcomes of patients with these devices after trauma is unknown.

Methods: The trauma registry at a regional trauma center was retrospectively reviewed for adult blunt trauma patients with age≥ 60 years admittd between January 2007 and June 2014. Deaths in the emergency department were excluded.

Results: 145 patients were admitted with ICDs, 234 patients with PMs, and 3,814 patients with no device (ND). ICD patients had ≥ 2 National Trauma Data Standard (NTDS) comorbidities more often than the other groups (93.1%# vs 78.6%PM vs 58.3%ND*,^). The most frequent major injury (AIS≥3) location was head/neck (43.4% ICD vs 35.0% PM vs 31.9%^ND) followed by extremities (24.1% ICD vs 32.5% PM vs 30.7% ND) in all groups. ICD patients had significantly longer hospital (8 vs 7 days) and ICU LOS (6 vs 4 days) than PM patients. Major surgery was required in similar frequency in all groups on admissions, with orthopedic surgery being the most common operation (31.0% ICD vs 35.3% PM vs 29.5%ND, p=ns). ICD patients had significantly more frequent NTDS defined complications (38.6%) than no device patients (29.6%^), but not compared to PM (33.8%, p=ns). Significantly fewer ICD were discharged to home as compared to no device patients (24.8% vs 38.8%^), but not as compared to PM patients (31.2%, p=ns). Significantly more ICD patients were discharged to acute rehabilitation as compared to no device patients (52.4% vs 44.2%^), but not as compared to PM patients (42.3%). Patients with ICDs had significantly higher mortality (12.4%# versus 9% in PM and 5.6%*,^ in ND group). However, in multivariate logistic regression, ICDs and pacemakers were not independent predictors of in-hospital mortality.

Conclusion: To our knowledge, this is the first report on the prevalence and outcomes of ICDs and PMs in the trauma population. We noted a substantial presence (9%) of PMs or ICDs in geriatric trauma patients. Over 1/3 of cardiac device patients developed complications during hospitalization and more than 10% expired expire during the index hospitalization, despite sustaining low energy trauma, predominantly low-level falls, and having low injury severity scores. Few survived to discharge home. The overall in-hospital mortality rate of cardiac device patients was double the ND rate. Greater vigilance is suggested in the triage and management of patients with ICDs or PMs, as these devices are markers of adverse outcomes following trauma.

17.06 Not as Futile as We Think: Age Adjusted Outcomes After Craniotomy for T.B.I. in the Elderly

Posted on January 18, 2016

S. R. Moradian1,2, Q. Dang1,2, I. Puente2, J. Catino2, F. Habib2, L. Zucker2, M. Bukur2 1Larkin Community Hospital,General Surgery,Miami, FL, USA 2Delray Medical Center,Trauma Surgery,Delray Beach, FL, USA

Introduction: Traumatic Brain Injury (TBI) continues to be a leading cause of death and disability particularly in the elderly population. Age is considered a risk factor for adverse outcomes after TBI however the impact of age on outcomes after surgical therapy for TBI is less well understood. We hypothesized that age alone would not be a risk factor for unfavorable outcomes after surgical treatment for TBI.

Methods: This was a retrospective review of all survivable head injuries undergoing craniotomy at a Level I trauma center from 2008-2013. Patients were stratified by age into young and elderly groups using the age of 40 as a reference group based upon current Brain Trauma Foundation guidelines. Logistic regression was used to adjust for baseline differences in demographic and injury variables to determine the effect of age on outcomes. The primary outcomes were in-hospital mortality, worsening discharge GCS, and discharge disposition.

Results: 265 patients met inclusion criteria. Overall rate of craniotomy for the 6 year period was 12.4% and was not significantly different between groups. Elderly patients were more likely to be victims of falls and present with a higher GCS despite having a higher Head AIS. Immediate need for operative intervention was similar between groups. After adjusting for differences in characteristics, there was no significant difference in overall mortality (8%vs 15%, p=0.383), worsening discharge GCS (21%, p=0.187).or rehabilitation placement (32.9% vs. 23.15, p=0.740). Stratification of mortality by decade revealed similar results.

Conclusion: Favorable outcomes after craniotomy for TBI appear to be obtainable in the elderly population. Using age alone as criterion for surgical intervention after TBI should be avoided.

17.07 Should Prior Anticoagulant use After Blunt Minor Head Trauma Result in Immediate Trauma Evaluation?

Posted on January 18, 2016

J. A. Simon1,2,3, J. Catino1,2, L. Zucker1, S. Evans1, I. Puente1,2, F. Habib1,2, M. Bukur1,2 1Delray Medical Center,Trauma Surgery,Delray Beach, FL, USA 2Broward General Medical Center,Trauma Surgery,Fort Lauderdale, FL, USA 3Larkin Community Hospital,General Surgery,South Miami, FL, USA

Introduction: With the elderly population increasing patients presenting on pre-existing antiplatelet or anticoagulant medications after traumatic injury is more common. The current CDC field triage criteria suggest that anticoagulant history should be considered in patients not meeting physiologic or mechanistic criteria for trauma activation. We hypothesized that using anticoagulant history alone as a criteria after minor head trauma would result in a high over-triage rate and that patients transferred with documented Traumatic Brain Injury (TBI) after minor head trauma would have similar outcomes to patients evaluated immediately by the trauma team.

Methods: This was a retrospective review of all trauma patients seen at a Level I trauma center (TC) during the past 3 years. History of antiplatelet or anticoagulant medication was abstracted from chart review as well as admission demographic, physiologic, and outcome characteristics. Minor head injury was defined as an admission GCS≥13, SBP>90, as well absence of tachypnea or mechanistic criteria set forth in the CDC field triage guidelines. Isolated TBI was defined as radiographic TBI+ other AIS≤2. Over-triage was defined as patients that were discharged or admitted to medical services without traumatic injuries. Our primary outcomes were the over-triage rate as well as the comparison of in-hospital mortality, need for craniotomy, worsening GCS, and discharge disposition for those patients with TBI not directly seen in our trauma center (NTC). Chi-Square and Fishers exact test were used to compare categorical outcomes, continuous variables were compared using the students t-test or Mann-Whitney test for non-parametric variables.

Results: During the 3 year period a total of 273 patients with minor head injury were directly evaluated in TC with 108 (39.5%) patients having radiographic TBI resulting in an over-triage rate of 60.5%. During the same period 54 NTC anticoagulated patients were transferred with documented TBI. Patients sustaining TBI were similar with respect to age, race, mechanism of injury, GCS, and type of intracranial bleed. Median Head AIS (4) was similar as well as type of anticoagulant (ASA TC 63% vs. NTC 57%,p=0.500, Plavix 24%vs.15%,p=0.221, Coumadin 31%vs.46%,p=0.057). Therapeutic INR (14%vs.20%,p=0.352) and platelet function abnormality (50%vs.48%,p=0.869) were similar between TC/NTC patients. Need for craniotomy (16%vs22%,p=0.385), in-hospital mortality(1.9%,p=1.000), worsening GCS (16%vs.17%,p=1.000), and discharge disposition (Home 36%vs.28%,p=0.337,Rehab 24%vs.35%, p=0.142,SNF 27%vs.26%,p=1.000) were also equivalent.

Conclusion: Use of anticoagulant status alone after minor head trauma results in a high rate of over triage and unnecessary utilization of trauma resources. Initial NTC evaluation of these patients even when TBI is discovered is safe with equivalent outcomes. Triage criteria incorporating anticoagulant status may offer an opportunity for process improvement.

17.08 Effect of Geospatial Organization of Trauma System Resources on Injury Mortality in Pennsylvania

Posted on January 18, 2016

J. B. Brown1, M. R. Rosengart1, T. R. Billiar1, A. B. Peitzman1, J. L. Sperry1 1University Of Pittsburgh,Surgery,Pittsburgh, PA, USA

Introduction: Trauma systems improve outcome; however it is unclear how geographic organization of these resources affects outcome. The objective was to evaluate correlation of geospatial trauma system resources and injury mortality in Pennsylvania (PA).

Methods: Trauma centers and helicopter emergency medical service (HEMS) bases in PA were mapped. Scene mortality rates were obtained at the county-level from the PA trauma registry. Service areas were generated for HEMS and ground EMS (GEMS) transport to a trauma center within 60min, accounting for response and scene times. Trauma system resources in each county including number of trauma centers or HEMS bases, proportion covered by HEMS and GEMS service areas were categorized as high or low by median values. Mortality was compared between high and low trauma system resources. Local Moran's I was used to identify geographic outlier clusters. Geographically weighted regression (GWR) evaluated association of mortality with trauma system resources, adjusting for county-level demographics, injury severity, and socioeconomic factors. Correlation of weighted mortality and resources from GWR was examined.

Results: 63,706 blunt and 6,260 penetrating patients were included. Blunt mortality was spatially autocorrelated (Moran I 0.22, p<0.01). Fig 1 demonstrates blunt mortality with trauma system resources. High HEMS coverage was associated with lower mortality (4.2% v 5.7%, p=0.04). Mortality was not associated with high number of trauma centers/HEMS bases (p=0.20) or high GEMS coverage (p=0.07). Outlier clusters with high mortality and low resources were seen in southcentral and northwest PA (p<0.01), while clusters of low mortality and high resources were seen in southwest and southeast PA (p<0.05). In GWR, mortality was inversely correlated with HEMS coverage (ρ=-0.71, p<0.01), GEMS coverage (ρ=-0.69, p<0.01), and number of trauma centers/HEMS bases (ρ=-0.38, p<0.01). Penetrating mortality was also autocorrelated (Moran I 0.36, p<0.01). High number of county trauma centers/HEMS bases was associated with lower mortality (13% v 26%, p<0.01), as was high HEMS coverage (13% v 25%, p<0.01) and high GEMS coverage (13% v 21%, p<0.01). Outlier clusters were similar to blunt injury. In GWR, mortality was inversely correlated with number of trauma centers/HEMS bases (ρ=-0.64, p<0.01) and HEMS coverage (ρ=-0.47, p<0.01), but not GEMS coverage (ρ=-0.23, p=0.07).

Conclusion: Geospatial organization of trauma system resources is correlated with injury mortality in PA. Different geospatial resources may be more important by mechanism. These results may aid trauma system planning and identify local areas to target trauma system resources.

17.02 Method of Hypertonic Saline Administration: Effects on Osmolality in Traumatic Brain Injury Patients

Posted on January 18, 2016

K. Maguigan1, B. M. Dennis2, S. Hamblin1, O. Guillamondegui2 1Vanderbilt University Medical Center,Department Of Pharmaceutical Services,Nashville, TN, USA 2Vanderbilt University Medical Center,Division Of Trauma And Surgical Critical Care,Nashville, TN, USA

Introduction:

Hypertonic saline (HTS) is an effective therapy for reducing intracranial pressure (ICP) in patients with traumatic brain injury (TBI). The ideal method of administration is unknown. The purpose of this study was to evaluate the method of HTS infusion and time to goal osmolality. We hypothesized that patients receiving bolus only HTS would reach goal osmolality more rapidly than patients receiving continuous infusion HTS.

Methods:

This is an IRB-approved, retrospective cohort analysis of severe TBI patients at a Level 1 Trauma Center from January 2008 to May 2014 who received at least 2 doses of 3% sodium chloride. Inclusion criteria were ICP monitor or external ventricular drain while on HTS and greater than 48 hour ICU admission. Craniotomy patients were excluded. Eligible patients were identified using the Trauma Registry of the American College of Surgeons. Patients were divided into bolus versus continuous infusion HTS cohorts. The primary outcome was median time in hours to the first osmolality value within goal range of ≥ 310mOsm/kg. Secondary outcomes included percentage of patients reaching goal osmolality, percent time at goal osmolality, mean cerebral perfusion pressure (CPP) and ICP, ICU length of stay, and ICU mortality. Statistical analysis included, the Mann-Whitney U test for continuous data and the Fisher’s Exact test for categorical data.

Results:

Of the 162 patients included (30 bolus and 132 continuous), baseline characteristics were similar. The median injury severity score for the bolus and continuous group was 33.5 (IQR 26-42.8) and 37 (IQR 29-45) respectively. The median Glasgow Coma Scale score was 3 for each group. The median volume of HTS given was significantly higher in the continuous group (1250mL vs 2735mL, p<0.001), with insignificant median total doses of mannitol. Time to goal osmolality was similar between the two groups (bolus 9.78 hr vs continuous 11.4 hr, p=0.817). A significant difference in the percentage of patients reaching goal osmolality favoring the continuous group was found (93.9% vs 73.3%, p=0.003). The continuous group was also maintained at goal osmolality for a higher percentage of osmolality values after reaching goal (80% vs 50%, p=0.032). Median CPP and ICP were similar on HTS therapy, and no differences were observed in ICU length of stay or mortality.

Conclusion:

Although no difference in time to goal osmolality was observed when comparing bolus or continuous infusion HTS, continuous HTS was associated with a higher percentage of patients achieving goal osmolality. Due to the limited sample size, we were unable to find a difference in ICP or CPP. Further studies are needed to definitively assess these methods of administration.

17.03 Morphometric Roadmaps to Improve Device Delivery For Fluoroscopy-Free Balloon Occlusion of the Aorta

Posted on January 18, 2016

J. N. MacTaggart1, W. Poulson1, M. Akhter1, A. Seas1, K. Thorson1, N. Phillips1, A. Desyatova1, A. Kamenskiy1 1University Of Nebraska College Of Medicine,Surgery,Omaha, NE, USA

Introduction:
Uncontrolled hemorrhage from large vessel injuries within the torso remains a significant source of prehospital trauma mortality. Resuscitative endovascular balloon occlusion of the aorta can effectively control hemorrhage, but this minimally invasive technique relies heavily upon imaging not available in the field. Our goal was to develop morphometric roadmaps to enhance the safety and accuracy of fluoroscopy-free endovascular navigation of hemorrhage control devices.

Methods:
Three-dimensional reconstructions of thin-section, contrast-enhanced computed tomography angiography scans from n=122 subjects (mean age 47±24 years, range 5-93 years old, 64 Male/58 Female) were used to measure centerline distances from femoral artery access sites to the major aortic branch artery origins. Morphometric roadmaps were created using multiple linear regression analysis to predict distances to the origins of the major arteries in the chest, abdomen and pelvis using torso length, demographics, and risk factors as independent variables. A hypothetical 40-mm long occlusion balloon was then virtually deployed into Zones 1 and 3 of the aorta using these roadmaps, with placement accuracy assessed by comparing predicted versus actual measured distances to the target locations within the aorta.

Results:
Torso length and age were the strongest predictors of centerline distances from femoral artery access sites to the major aortic branch artery origins. Male gender contributed to longer distances to aortic arch branches, while diabetes and smoking history were associated with shorter distances. Hypertension, dyslipidemia and coronary artery disease had no effect. Using morphometric roadmaps, virtual occlusion balloon placement accuracy was 100% for Zone 3 of the aorta, compared to 87% accuracy when using torso length alone (Figure).

Conclusion:
Morphometric roadmaps demonstrate potential for improving the safety and accuracy of fluoroscopy-free aortic occlusion balloon delivery. Continued development and refinement of minimally invasive hemorrhage control techniques could lead to improvements in prehospital mortality for all age groups and both genders of patients with noncompressible torso hemorrhage.

17.04 Optimization of Surgeon Resources: Reducing Overtriage Rates in Pediatric Trauma

Posted on January 18, 2016

H. Naseem2, N. Pantano1, K. D. Bass1,2 1State University Of New York At Buffalo,College Of Medicine,Buffalo, NY, USA 2Women And Children’s Hospital Of Buffalo,Department Of Pediatric Surgery,Buffalo, NY, USA

Introduction: The American College of Surgeons (ACS) has described criteria for highest-level (Level I) trauma team activation (TTA) and presented guidelines for limited trauma team activation (Level II) and trauma consults. Evidence based literature is lacking on data supporting Level II activation and consult guidelines for pediatric trauma patients. At our center, the current criteria for Level II TTA leads to a high rate of overtriage, especially with a large percentage of ‘motor vehicle crashes (MVC) greater than 20 mph’ and ‘pedestrian or bike versus vehicle’ being discharged from the emergency department (ED). This study was designed to evaluate possible options to optimize surgeons as resources by revising our current Level II activation criteria. We hypothesized that allowing ED to evaluate ‘MVC>20mph’ and ‘pedestrian or bike versus vehicle’ prior to the Level II activation would significantly reduce the overtriage rate while maintaining an acceptable undertriage rate.

Methods: Trauma registry data was reviewed for a five-month period. Overtriage was defined as Level I and II TTAs that were discharged from the ED. Undertriage was defined as trauma consults that were admitted to the intensive care unit or had an injury severity score (ISS) greater than 16. The over and undertriage rates were calculated using the matrix method defined by ACS. The two most common criteria resulting in overtriage were then identified, and the over and undertriage rates were recalculated with these criteria moved to the consult category. Charts were reviewed to verify that there were no missed injuries or diagnosis. Over and undertriage rates before and after criteria revision were statistically analyzed using the paired t-test.

Results: The over and undertriage rates using current criteria were 54.0% and 2.1%, respectively. The two criteria leading to the most overtriage were ‘MVC>20mph’ and ‘pedestrian or bike versus vehicle’. When the rates were recalculated moving these two criteria under the consult category, the overtriage rate was 37.3% (p <0.001) and the undertriage rate was 2.7% (p 0.16). After chart review no missed injuries or diagnosis were identified.

Conclusion: The overtriage rate and use of surgeons as resources in trauma activations may be reduced by allowing the ED to evaluate ‘MVC>20mph’ and ‘pedestrian or bike versus vehicle’ prior to Level II activation, eliminating surgeon response for non-injured children. The protocol will be changed and the data will be prospectively collected and analyzed for follow-up to confirm that the overtriage and undertriage rates continue to meet the expected rates.

17.05 Timing of Tracheostomy and its Association with Length of Stay and Mortality

Posted on January 18, 2016

Z. Wells2, M. Bascom2, Z. Gauthier2, M. Suh2, B. Meloro2, A. Lopez2, M. Goodwin4, A. H. Parsikia1, J. A. Ortiz4, P. S. Leung1, A. R. Joshi1,3 1Einstein Healthcare Network,Surgery,Philadelphia, PA, USA 2Philadelphia College Of Osteopathic Medicine,Philadelphia, PA, USA 3Jefferson Medical College,Surgery,Philadelphia, PA, USA 4University Of Toledo,Surgery,Toledo, OH, USA

Introduction:
Patients requiring long-term mechanical ventilation may benefit from early tracheostomy. However, the most appropriate timing of tracheostomy is still not well-determined.

Methods:
We reviewed 500 consecutive tracheostomies over four years. After exclusion criteria were applied, 417 patients remained. We divided the cohort according to two benchmarks: tracheostomy before and after 7 days, and before and after 10 days on mechanical ventilation. We then analyzed ICU length of stay (LOS), overall LOS, mortality, laboratory data, demographics, type of tracheostomy, post-operative complications, type of ICU, type of surgeon, and disposition after discharge. We compared the variables with uni- and multi-variate logistic regression tests to determine if any were prognostic.

Results:
The early and late cohorts were comparable with regard to pre-operative characteristics and post-operative complications. Pre-operative albumin was significantly higher in the early group. Earlier tracheostomy was associated with shorter ICU and overall LOS. Survival rates at 1, 6, and 12 months were not statistically different. In multivariate logistic regression, higher albumin predicted decreased in-hospital mortality (p=0.003) and higher INR predicted increased in-hospital mortality (p=0.002)—better than timing of tracheostomy.

Conclusion:
Early tracheostomy was associated with decreased ICU LOS (by 9-10 days) and overall LOS (by 8-10 days). Timing of tracheostomy was not associated with a difference in mortality or disposition after discharge. Preoperative albumin and INR are predictive for in-hospital mortality among patients who undergo tracheostomy. Early tracheostomy does appear to be highly significantly associated with decreased LOS in the ICU and hospital, and therefore should be encouraged.

16.20 Injury as a Risk Factor for Overall Mortality in Elderly Americans

Posted on January 18, 2016

C. M. Psoinos1, J. M. Flahive1, F. Anderson1, H. P. Santry1 1University Of Massachusetts Medical School,Surgery,Worcester, MA, USA

Introduction: Survival to discharge is increasingly considered a poor measure of post-injury outcome. We studied post-discharge mortality in a national sample of elderly trauma patients.

Methods: A 5% random sample of Medicare Beneficiaries (n=864,604) was queried for patients admitted (1/1/2009-12/31/2010) with a primary diagnosis code (ICD9) for traumatic injury. We examined inpatient and post-discharge mortality until 12/31/2011 using univariate tests of association and multivariable logistic regression models to determine predictors (demographic factors, Elixhauser co-morbidity index, critical injury) of EARLY post-discharge (0-30 days) and LATE (31-365 days) mortality. Critical injury was defined as intensive care unit (ICU) stay ≥1 day and an abbreviated injury score (AIS) in any one region ≥3.

Results: Of 60,016 elderly trauma patients, 1,832 died during index hospitalization (3.1%). Age, sex, Elixhauser index, and critical injury (all p≤0.01) were independent predictors of inpatient mortality.

Of the 58,184 patients discharged alive, 2,318 (4.9%) died EARLY and 9,123(15%) died LATE. Figure 1 shows unadjusted rates of EARLY and LATE mortality for factors associated with mortality in univariate analyses. In multivariable models, age ≥85, male sex, Elixhauser index ≥3 critical injury and discharge disposition (all p≤0.01) were independent predictors of EARLY mortality. Age, sex, Elixhauser index, and discharge disposition (all p≤0.01) predicted LATE mortality. For the EARLY model, significant discharge locations for increased odds of mortality (referent group discharge home) were home with services (OR 1.9, 95%CI 1.6, 2.2), LTAC (OR 11, 95%CI 8.8, 14), and rehabilitation facility (OR 2.4, 95%CI 1.9, 2.9). Discharge dispositions predictive of LATE mortality were home with services (OR 1.6, 95%CI 1.5, 1.7), LTAC (OR 5.0, 95% CI 4.2, 6.0), other in-patient (OR 2.1, 95%CI 1.9, 2.4), and rehabilitation facility (OR 1.5, 95%CI 1.4, 1.7).

Conclusion: Nearly a quarter of elderly trauma patients were dead by 1 year after injury, with deaths during index hospitalization accounting for ~3% of this mortality. The association between age, sex, comorbidities, severity of injury & condition at discharge and post-discharge mortality suggests that, baseline patient condition, factors leading up to injury and physiologic & functional consequences of injury all play a role in long-term mortality. While further research is needed, our findings have implications for framing expectations for post-discharge outcomes after injury in the elderly.

16.21 Elucidation of the Temporal Response to Tranexamic Acid Administration following Shock

Posted on January 18, 2016

M. Diebel1, D. Liberati1, J. Liberati1, L. Diebel1 1Wayne State University,Department Of Surgery/School Of Medicine,Detroit, MI, USA

Introduction: Early adminstration of tranexamic acid (TXA) appears to reduce mortality in bleeding trauma patients. Potential mechanisms include decreased blood loss and antiinflammatory effects with TXA administration. Hemorrhagic shock (HS) related gut ischemia leads to disruption of the mucosal barrier. This has been attributed to entry of activated digestive enzymes into the intestinal wall and subsequent activation of mucosal associated matrix metallopeptidase 9 (MMP-9). Both "systemic" and enteral administration of TXA may protect the gut barrier following HS. The temporal response and mechanism(s) are unknown and served as the basis of the current study.

Methods: Caco-2 (intestinal cells) + HT29-MTX (mucus cells) intestinal epithelial cell (IEC) co cultures were established in a two-chamber cell culture system. IEC co cultures were then exposed to hypoxia-reoxygenation (HR) challenge ± trypsin added to the apical media. TXA was added in some experimental groups immediately after hypoxic challenge (Time 0) and then 60 and 120 minutes following reoxygenation. Trypsin activity was measured using a substrate specific assay (N α-Benzoyl-L-arginine 4-nitroanilide hydrochloride, BAPNA) at times corresponding to 60 minute intervals +/- TXA administration. MMP-9 was determined from intestinal lysates at corresponding times using Western blot analysis and quantified by relative density calculation.

Results: Please see table.

Conclusion: The protective effect of TXA on gut barrier function may be due to inhibition of trypsin and subsequent MMP-9 activation. These effects were noted only with TXA administration within 60 minutes following the 90 minute IEC hypoxic challenge. These results support the early administration of TXA in bleeding trauma patients and may represent an additional benefit of TXA in these patients.

17.01 Scooter and Moped Crash Injuries: Time Matters

Posted on January 18, 2016

K. E. Moldowan1, D. R. Fraser1, N. D. Fulkerson1, D. A. Kuhls1 1University Of Nevada School Of Medicine,Division Of Acute Care Surgery,Las Vegas, NV, USA

Introduction:

In 2014, 33,528 scooters were sold across the United States. A recent study conducted in the southeastern region of the United States found that scooter and moped crashes (SMCs) that occurred at night resulted in more severe injuries. However, there were no recently published studies that looked at SMCs in the western region of the United States and furthermore there were no studies in the United States that investigated temporal aspects as they related to SMCs. To understand the safety risks and to guide injury prevention, the objective of this study was to analyze temporal risk factors associated with SMCs in the Las Vegas metropolitan area.

Methods:

We reviewed all patients injured by SMCs admitted to the only Level 1 Trauma Center in Nevada, located in Las Vegas, over an 8-year period from 2006-2013 (n=404). Demographics, hospital utilization data, injury characteristics including Abbreviated Injury Scale (AIS), and temporal aspects, based on when the patient was admitted (day versus night, weekday versus weekend, and spring/summer versus autumn/winter), of the crash were analyzed. There were patients with missing data points that were excluded from our analysis. Data was analyzed in SPSS 22 with significance set at p<0.05.

Results:
More than half (54%) of SMC patients were admitted during the day time. Those admitted during the day had: 1) more severe abdominal injuries 2) higher revised trauma scores (RTS), 3) increased Emergency Department Glasgow Coma Scores (ED GCS) (Table 1, p=0.004, p=0.001, p=0.010). Meanwhile, those admitted at night had more severe brain and/or spinal injuries (Table 1, p=0.047). Interestingly, of all patients that were suspected of alcohol consumption, 51% were admitted at night (p<0.001). Those admitted on the weekend had a longer hospital length of stay (HLOS; average: 7.8 vs. 4.8 days) (p=0.014). Additionally, 52% of those admitted to the Intensive Care Unit were admitted during autumn and winter (p=0.012). From our binary logistic regression model, time of day is the strongest predictor for having a brain and/or spinal injury after a SMC, followed by time of week (controlling for age, race, gender, and safety). Those admitted at night and during the weekend are 2.6 and 2.4 times more likely to have a brain and/or spinal injury as a result of the SMC (p=0.031).

Conclusion:
In Las Vegas, there are temporal risk factors associated with SMCs. Weekend admissions resulted in longer HLOSs (p=0.014). Those admitted at night and on a weekend are 2.6 and 2.4 times more likely to have a brain and/or spinal injury (p=0.031). These findings should guide injury prevention efforts and resources.

16.17 TRAUMATIC BRAIN INJURIES: OUTCOMES COMPARING INTRACRANIAL PRESSURE (ICP) MONITORING VERSUS PbtO2

Posted on January 18, 2016

J. C. Haynes1, C. Lawson1, B. Daley1 1University Of Tennessee Medical Canter At Knoxville,Surgery,Knoxville, TN, USA

Introduction: Traumatic Brain Injuries (TBI) cause significant burden on society with an estimated 1.5-2 million injuries with 50,000 associated deaths. Unfortunately there is no true consensus on the best way to treat patients who suffer from severe TBIs. Following brain tissue oxygenation (PtbO2) using special monitors (Licox) has been espoused as beneficial and our center has followed an existing guideline for the last seven years involving close ICP monitoring and in specific patients (chosen by the neurosurgeon) directed therapies to keep PbtO2 greater than 20. We sought to determine if the addition of PbtO2 improved outcome over standard ICP monitoring.

Methods: With IRB approval, a retrospective chart review of the trauma database at a Level 1 center was performed of TBI since 2009. We compared the outcomes of those with had ICP Monitors (‘bolts’ or ventriculostomies) and Licox monitors. We specifically looked at ICU stay, time on the ventilator, length of stay, Injury Severity Scores, and deaths employing T testing and Chi squared analysis.

Results:We found 150 patients meeting inclusion criteria: the Licox group had 41 patients, with an average age of 32, LOS 32.3, ICU days 18.75, vent days 15.9, ISS of 32.6 and 15 deaths. The ICP group had 109 patients, average age of 35.2, LOS 22.3, ICU days 14.8, ISS of 34.5, and 48 deaths. There was no significant difference between the two groups with respect to Age, ICU and Vent days, ISS. There was a difference in the total length of stay; with the ICP group have a shorter LOS. Using Chi-square testing, there was no difference in deaths between groups.

Conclusion: We found that in our current mature treatment schema for TBI, Licox monitors did not improve outcome, but did extend length of stay. Because the Licox paradigm involves aggressive and prolonged mechanical ventilation and frequent transfusion, both known to adversely affect injured patients outcomes, we feel unless its value is validated prospectively, it should be used sparingly.

« Previous Page
Next Page »