190 Fear of Brain Herniation From Lumbar Puncture: DoHistory and Physical Exam Indicate Abnormalitieson Head Computed Tomography?

O’Laughlin KN, Go S, Gabayan GZ, Iqbal E, Merchant G, Lopez-Freeman R,Zucker MI, Hoffman J, Mower W/Harvard Medical School, Boston, MA; Universityof Kentucky Medical Center, Lexington, KY; Greater Los Angeles Veterans’ Affairs,Los Angeles, CA; UCLA School of Medicine, Los Angeles, CA; UCLA Medical Center,Los Angeles, CA; University Hospital in Cincinnati, Cincinnati, OH

Study Objectives: There is a fear that performing a lumbar puncture (LP) on a patientwith increased intracranial pressure (ICP) may lead to tonsillar herniation. Because of this,many physicians first check a screening head computed tomography (CT) looking forabnormalities suggestive of elevated ICP. Checking a head CT on every patient prior toLP is time consuming, costly, and exposes patients to significant radiation. Our goal wasto define clinically significant head CT abnormalities and then to analyze the ability ofcertain history and physical exam findings to predict those radiographic findings.

Methods: This was a secondary analysis of a prospectively maintained head CTdatabase of patients presenting to the UCLA Medical Center Emergency Departmentbetween April 2006 and February 2007. We used the Delphi method to defineclinically significant radiographic head CT abnormalities. We then analyzed the testcharacteristics for history and physical exam findings in predicting clinicallysignificant head CT abnormalities.

Results: When analyzed individually, the history and physical exam findings did notpredict significant head CT abnormalities well. The most sensitive individual findingswere: the presence of a neurological deficit, sensitivity 68.9% (CI 53.4, 81.8) and NPV93.4% (CI 89.1, 96.3); altered level of consciousness, sensitivity 67.39% (CI 52.0, 80.5)and NPV 91.8% (CI 86.8, 95.3); and lack of proper orientation, sensitivity 65.1% (CI49.1, 79.0) and NPV 92.1 (CI 87.3, 95.5). Using the combined criteria of all of thehistory and physical exam findings together, the sensitivity for predicting clinicallysignificant head CT abnormalities was good but not perfect. The sensitivity for thecombined criteria was 95.7% (CI 85.5, 99.5) and the NPV was 96.1% (86.5, 99.5).

Conclusions: Our data suggests that history and physical exam alone may beinadequate to detect the subtle head CT changes that could indicate potential forbrain herniation as defined by the Delphi criteria. Despite that finding, weacknowledge the major limitation that our outcome measure was radiographicabnormalities rather than actual brain herniation, which makes it difficult toextrapolate concrete conclusions regarding the clinical relevance of this information.We think that the likelihood of herniation is much lower than the Delphi criteriacaution and that because of this the head CT criteria we used are too sensitive.

191 Disease and Non-Battle Traumatic InjuriesEvaluated by Emergency Physicians in a U.S.Tertiary Combat Hospital

Bebarta VS, Mason PE, Ferre RM, Eadie JS, Muck AE, Joseph J, Pitotti RL/Wilford Hall Medical Center, San Antonio, TX; University of Pittsburgh,Pittsburgh, PA

Study Objectives: Emergency physicians have played a central role in medical caredelivery in Operation Iraqi Freedom. Medical war planning has focused on combat-related injuries; however, since 2004 nearly half of the patients treated in a tertiaryU.S. combat hospital in Iraq are not combat related. In order to plan for future warsand properly train emergency physicians, the common emergent and urgent non-combat diagnoses and complex procedures performed should be identified. Thesedata have not been previously collected or studied to determine the types of non-combat injuries and illnesses seen at a tertiary combat hospital and the types ofprocedures performed to manage these patients.

Methods: In our institutional review board-approved study, we enrolled all non-combat injured patients over one year who were evaluated in the emergencydepartment (ED) of a US military tertiary hospital in Iraq. The treating emergencyphysician used a standard data collection form to enroll all patients who arrived to theED whose injury or illness was unrelated to combat. Data collected included age,time of visit, ED diagnoses, emergency procedures, disposition, and consultations.The diagnosis and procedure lists were defined before study start.

Results: Data were gathered on 1745 patients with a mean age of 30.2 years (range 6months – 72 years). 1460 (83%) patients were male and 1316 (75.4%) were U.S. militarypersonnel, with Iraqis, foreign military members and third country nationals making upthe remaining. The most common diagnoses evaluated in the ED were abdominaldisorders (302 cases, 17.3%), orthopedic injuries (209, 12%), headache (108, 6.2%),ophthalmologic injuries (106, 6.1%), lacerations or abrasions (99, 5.7%), soft tissue

infections (94, 5.4%), and renal colic (88, 4.7%). Emergent medical diagnoses consistedof 10.2% (179) of cases (aortic dissection, meningitis, altered mental status, overdose,pulmonary embolism, intestinal bleeding, acute myocardial infarction, chest pain, atrialfibrillation). Urgent medical diagnoses consisted of 4.4% (78) of cases (pneumonia,diverticulitis, deep venous thrombosis, seizure, syncope). Emergent surgical diagnosesconsisted of 10.7% (186) of cases (appendicitis, cholecystitis, bowel obstruction,peritonsillar abscess, fracture, dislocation, gunshot wound, eye trauma, brain injury). 52%(909 cases) required IV access, 45% (793) received laboratory testing, 31% (537) receivedradiographs, and 28% (488) received CT scans. Complex procedures include ECGinterpretation (9%, 160 cases), lumbar puncture (1.1%, 20), procedural sedation (0.8%,14), endotracheal intubation (10), and central line (4). 4.5% of cases were admitted to theoperating room, 6.5% to the ICU, and 21.6% to the ward. 12.6% of patients wereevacuated out of Iraq.

Conclusion: Life-threatening diseases and non-battle traumatic injuries arecommon in a tertiary-combat hospital emergency department. Providers working insimilar settings should have diagnostic and procedural skills to evaluate and treat arange of emergently ill patients.

192 The Utility of HbA1C Screening in Low Risk ChestPain Patients in the Emergency DepartmentObservation Unit

Wiederhold H, Swor R, Robinson D, Clark CL/William Beaumont Hospital, RoyalOak, MI

Background: The American Heart Association and The American DiabetesAssociation recommend early identification of pre-diabetic and undiagnosed diabeticpopulations. HbA1C has been shown to be an independent predictor of Type 2diabetes. Previous studies have shown low yield for the use of HbA1C as a screeningtool in the general emergency department population.

Study Objective: To identify the proportion of non-diabetic patients withabnormal HbA1C within an Emergency Department Observation Unit (EDOU)population of low risk chest pain patients.

Methods: We performed a retrospective review of patients admitted to the EDOU ata large academic community hospital for low risk chest pain evaluation from August 1,2008 through September 18, 2008. Patient data was extracted for patient demographics,known history of diabetes, current treatment for hyperglycemia, random glucose andHbA1C levels. We dichotomized values of random glucose levels � 110 mg/dl andHbA1C levels greater than 5.9%, per recommended screening levels, and calculated theproportion of possibly undiagnosed diabetic patients. We assessed whether abnormalrandom glucose levels might identify a similar proportion of non-diabetic patients asfound by abnormal HbA1C levels.

Results: During the study period, 457 low risk chest pain patients were admittedto the EDOU. Of these, 342 received HbA1C and random glucose testing. Thesepatients were 60.8% male, their mean age was 57.9 �/� 15.38 years and 279patients had no known history of diabetes. Of patients with no known history ofdiabetes, 66 (23.7%) were found to have elevated HbA1C and 85 (30.5%) werefound to have elevated random glucose. 36 (54.5 %) patients with elevated HbA1Chad random glucose levels within the normal range.

Conclusion: Screening evaluation by HbA1C of EDOU low risk chest pain patientsappears to identify patients with impaired glucose metabolism not found by randomglucose screening alone. Prospective studies of low risk chest pain patients should be doneto validate and assess the value of using elevation of HbA1C as a screening tool.

193 A New Study of Intraosseous Blood for CBC andChemistry Profile

Miller L, Philbeck T, Montez D, Spadaccini C/Vidacare Corporation, San Antonio,TX; Ameripath, Inc., San Antonio, TX

Study Objectives: Recent studies have indicated a resurgence in using the intraosseous(IO) route for vascular access because of improved devices that enable providers to delivercritically needed drugs systemically as quickly as central lines, and faster than peripherallines. Increased awareness of IO vascular access has led to questions about other uses of theIO space, including the viability of IO blood for routine laboratory analysis. Despiteearlier studies, some laboratory personnel have voiced concern over the adequacy of thedata supporting the use of IO-derived blood for routine laboratory tests. This study wasdesigned to validate earlier studies and address these concerns.

Methods: The study was approved by IntegReview Institutional Review Board and 10healthy adult volunteers were recruited. After written consent, standard phlebotomy bloodsamples were obtained. Following phlebotomy, an IO catheter (EZ-IO®, Vidacare

Research Forum Abstracts

Volume , . : September Annals of Emergency Medicine S59

Corporation, San Antonio, TX, USA) was placed in the proximal humerus and two setsof IO blood samples were obtained; one following 2ml of marrow/blood waste, and onefollowing 6ml of waste. All three samples sets were sent to a reference laboratory forchemistry profile and complete blood count (CBC) analysis. Means for each value for thethree blood draws (designated IV, IO-1, and IO-2) were calculated and compared withthe intravascular (IV) blood serving as a control for the IO draws.

Results: For IO-1, mean red blood cells (RBC), hemoglobin (Hgb), hematocrit,glucose, blood urea nitrogen (BUN), sodium, chloride, total protein and albumin levelswere within 5% of mean values from IV blood. For IO-2, mean Hgb, glucose, BUN,sodium, chloride, total protein and albumin levels were within 5% of mean values fromIV blood. For both IO samples, most other values were within 10% of IV blood.

Conclusion: In the first study of its kind in 15 years, we have found that theintraosseous space is a reliable source for blood used for laboratory analysis commonlyperformed in emergency medicine, including CBC and chemistry profile. Results maybe moderately reliable for carbon dioxide, but unreliable for WBC counts that appearto be elevated and platelets counts that appear lower.

194 The Use of a Subcutaneous Insulin Aspart Protocolfor the Treatment of Hyperglycemia in the EmergencyDepartment: A Randomized Clinical Trial

Harper JB, Barnard J, Munoz C, Baldwin D/Rush University, Chicago, IL

Study Objective: Emergency physicians have a unique and challenginginteraction with patients who have diabetes. Such patients commonly present withsignificant hyperglycemia. Patients presenting with a specific diabetes-associatedillness routinely have the complaint directly addressed and treated. However, whenpatients present with a non-diabetic chief complaint, there are no establishedprotocols for how best to address the common issue of hyperglycemia. Weprospectively evaluated a SQ insulin protocol for use in the emergency department(ED) in patients with known Type 2 diabetes mellitus (DM) and hyperglycemia.

Methods: Patients with Type 2 diabetes had a point-of-care blood glucose (BG)measured soon after ED presentation; those with BG �200mg/dL were randomizedto an intervention group (INT) vs. usual care (UC). All INT subjects (n�66)received subcutaneous insulin aspart (0.05 U/kg for BG 200–299mg/dL, 0.1 U/kgfor BG 300–399mg/dL, 0.15 U/kg for BG � 400mg/dL) every 2 hours until BG�200 mg/dL. Insulin aspart was chosen for this protocol because its onset of action is30 minutes and it can be redosed every 2 hours, thus making it easy forimplementation in a busy ED. Emergency physicians treated UC subjects (n�72) attheir discretion, and 49% did not receive insulin. Subsequent blood glucose wasmeasured every 2 hours in the ED until discharge home or hospital admission.

Results: Mean initial ED BG for all subjects enrolled in the study was 299 �78 mg/dL. At ED discharge either to home or the hospital, the mean BG decreased by 76 �67mg/dL with INT, and by 82 �77 mg/dL with UC (ns). 47 UC subjects were admitted tothe hospital, and 43% of those had received SQ insulin aspart in the ED, while 40% ofINT subjects were admitted. When the first BG after admission to the hospital wasassessed, however, the mean decreases from the first ED BG were greater: UC subjectswho received insulin decreased 131 �104 mg/dL, while UC subjects who did not receiveinsulin only decreased 33 �67 mg/dL (p� 0.04). INT subjects decreased 104 �80 mg/dL. Only 1 subject had a BG reading less than 100 mg/dL (76 mg/dL). Mean ED lengthof stay was similar, INT 5.4 �1.8 hours, UC 4.8 �1.8 hours (ns).

Conclusion: A weight-based protocol for dosing SQ insulin aspart every 2 hoursin the ED for the treatment of hyperglycemia in patients with Type 2 diabetes wassafe and effective. This protocol was easy for physicians to determine insulin dosingand nursing to administer without significantly increasing ED length of stay. Theinsulin treatment of ED patients with hyperglycemia achieves rapid and significantlowering of BG. A higher unit per kg insulin dosing algorithm can achieve greaterdecreases in BG; however, our previous pilot study found an excess of hypoglycemia.Thus these doses seem optimal. Further study is required to delineate possible benefitsto patients who are subsequently admitted to the hospital.

195 Refusals of Medical Aid in the Out-of-HospitalSetting

Waldron R, Finalle C, Mogelof D/New York Hospital Queens, Flushing, NY

Study Objective: Our research examined the characteristics of both patients andEMTs who are involved in the Refusal of Medical Aid (RMAs.) As well, we studiedthe timing of RMAs both by shift and within an individual shift.

Methods: The study was carried out using data from the New York Hospital

Queens (NYHQ) ambulance service, which is a large urban ambulance serviceproviding 9-1-1 basic life support and advanced life support care. This was aretrospective chart review of all patient charts in which the patient RMA’d for thetime period 8/1/05 through 7/31/06, a one-year period. These patients were thencompared to a control set of patients that was created by reviewing every chart in a24-hour period for ten randomly selected days within the same one-year period. Thedata was obtained from the patient care reports that are scanned by NYHQ EMTsinto HealthEMS database. Data analysis was performed using SAS 9.1 for Windows.For continuous variables, the Student’s t test was used to test for differences betweenthe control and research groups. For categorical variables, the Chi-Square test wasused and the Fisher’s Exact test was used if cell counts were less than 5.

Results: The RMA data set had a total of 238 patients, 58% female and 42% male,with a mean age of 56. The control data set had a total of 303 patients, 53% female and47% male, with a mean age of 53. There was no difference in the sex distribution betweenthe RMA and control groups (P �� 0.2965.) There was also no difference in average ageof the RMA and control patients on the day and evening shifts (P � 0.1764 & 0.0711).However, on the night shift the patients in the RMA group were significantly younger -mean age of 47 in the research group versus 55 in the control group (P � 0.0160).

The EMS team consists of two EMTs. The presence of a male on the team increasedthe likelihood of an RMA. The EMT teams in the RMA set were 0.42% female/female,7.1% male/female, and 92.4% male/male. The EMT teams in the control set were 4.6%female/female, 34.7% male/female, and 60.7% male/male. The higher percentage of maleEMTs in the RMA set achieved statistical significance (P � 0.0001). The sex of the EMTteam versus the sex of the patient had no effect on increasing RMAs (P � 0.9936).

The patient’s chief complaint was significantly different in the RMA versuscontrol groups. The RMA group had more neurological, psychiatric, and social chiefcomplaints. (P � 0.0001) This difference holds true for both day and evening shifts(P � 0.0003, 0.0001); however, on night shifts there is no significant difference inchief complaints (P � 0.0812).

In the RMA group, the shifts were 35.7% day, 41.6% evening, and 22.7% night.In the control group, the shifts were 51% day, 30% evening, and 18.5% night.RMAs were more common on the evening and night shifts (P � 0.0011).

The calls were also broken down as to whether occurring in the first two hours ofthe shift, the middle four, or the last two hours. There was no difference in thefrequency of RMAs based on the timing within the shift. (P � 0.5488).

Conclusions: The call most likely to generate an RMA contains the following setof characteristics: a younger patient with a chief complaint falling out of the usualmedical categories like trauma or cardiac, two male EMTs on the team, and anevening or night shift. The commonly held perception that RMAs are more commonat the end of a shift (to avoid working late) was proven to be untrue.

196 Intubation Success Rates in Helicopter EmergencyMedical Services: A Prospective MulticenterAnalysis

Patel P, Melissa S, Brunko M, Domeier R, Funk D, Greenberg R, Judge T,Lowell M, MacDonald R, Madden J, Thomas S, Howard Z/Harvard AffiliatedEmergency Medicine Residency, Boston, MA; Massachusetts General Hospital,Boston, MA; Flight for Life Colorado, Denver, CO; Midwest MedFlight, St. Joseph’sMercy Hospital, Ypsilanti, MI; LifeNet of New York, Albany Medical College, Albany,NY; PHI Air Medical StatAir, Texas A&M University, Temple, TX; LifeFlight of Maine,Bangor, ME; Survival Flight, University of Michigan Health Sciences Center, AnnArbor, MI; ORNGE Transport Medicine, University of Toronto, Toronto, Ontario,Canada; Christiana Care, Wilmington, DE; University of Oklahoma School ofCommunity Medicine, Tulsa, OK; Brigham and Women’s Hospital, Boston, MA

Study Objectives: The Critical Care Transport Collaborative Outcome ResearchEffort (CCT-CORE) Airway Study is a multicenter analysis of air medical programs’performance on airway management variables defined by the National Association ofEmergency Medical Services Physicians (NAEMSP). This study examines the success rateof endotracheal intubation (ETI) in air transport programs across a variety of settings. Italso examines whether there are lower success rates for air medical crews attempting ETIin patients in whom ETI by non-air medical providers has already failed.

Methods: This was a prospective consecutive-case series of patients undergoingair medical transport in whom advanced airway management was attempted. Therewere 11 participating sites, and all crews had access to RSI drugs. Eligible subjectsincluded all patients in whom air transport crews attempted advanced airwaymanagement. Prospectively defined data points were collected and entered into asecure Web-based data entry system. The primary analysis for this report wasdescriptive, focusing on ETI success rates (reported with exact binomial 95%

Research Forum Abstracts

S60 Annals of Emergency Medicine Volume , . : September