Research Forum Abstracts

enrolled. The target age range was 18 - 50 years. The following subject characteristicswere collected: age; sex; known peripheral vascular disease (PVD) of the lowerextremities; presence of palpable dorsalis pedis and posterior tibial artery pulses; andpresence of femoral bruits. ABIs were performed and calculated in a consistentmanner using an 8-MHz continuous-wave Doppler ultrasonography probe and anappropriately sized blood-pressure cuff. Subjects were examined by 2 of theinvestigators, each blinded to the measurements obtained by their co-investigator.“Best Case” ABI measurements used the higher of the two ankle readings and thehigher of the two brachial readings to calculate the result. “Worst Case” ABImeasurements used the lower of the two ankle readings and the higher of the twobrachial readings to calculate the result.

Results: 118 subjects were enrolled, with both arms and legs measured in all. Themean age was 35 years (range 19 - 49), 50% were female, none had known PVD, andno femoral bruits were auscultated. When the ABI measurements were performedand calculated correctly (Best Case), inter-rater agreement ranged from 96.6% to98.3%, the single provider false positive rate was 1.7%, and the two provider falsepositive rate was 0%. When the ABI measurements were performed and calculatedincorrectly (Worst Case), inter-rater agreement ranged from 69.5% to 78.8%, thesingle provider false positive rate was 29.2%, and the two provider false positive ratewas 4.2%.

Conclusions: Our investigation suggests that an ABI measurement performed andcalculated incorrectly by a single provider could result in a false positive rate thatapproaches 30%. This error is decreased to a still clinically relevant rate of 4.2% iftwo providers perform the measurement and calculation incorrectly. If a singleprovider performs the measurement and calculation correctly, the false positive rateapproaches 2%. Two providers performing the ABI measurement and calculationcorrectly will essentially eliminate the risk of a false positive result. We recommendstrict adherence to a standardized ABI measurement and calculation protocol. Wealso recommend that an ABI measurement of �0.9 in a patient with lower extremitytrauma have the measurement confirmed by a second provider in order to minimizethe risk of inappropriately implementing invasive diagnostic procedures.

412 Does a Portable, Non-Invasive Hemoglobin MonitorEffectively Correlate With Venous Blood Levels?

Knutson C, Della-Giustina C, Tomich C, Wills B, Leurssen C/Madigan ArmyMedical Center, Tacoma, WA; Virginia Commonwealth University Health System,Richmond, VA

Study Objectives: The Masimo Radical-7® is a medical device recently approvedby the Food and Drug Administration that performs non-invasive oximetry andestimated venous or arterial hemoglobin measurements. A portable, non-invasivedevice that rapidly measures hemoglobin concentration could be useful in bothaustere and modern hospital settings. The objective of this study is to determine thedegree of variation between the device’s estimated hemoglobin measurement and theactual venous hemoglobin concentration in undifferentiated emergency department(ED) patients.

Methods: We conducted a prospective, observational, cross-sectional study ofadult patients presenting to the ED. The subjects consisted of a convenience sampleof adult ED patients who required a complete blood count (CBC) as part of their carein the ED. A simultaneous probe hemoglobin was obtained and recorded.

Results: There were 127 measurements recorded with corresponding laboratoryhemoglobin determinations. Overall, the mean absolute difference between probe andlaboratory measurement was 1.6 g/dL (95%CI 1.4-1.9). For laboratory hemoglobinconcentrations of less than eight, 8-11.5, and greater than 11.5, the mean absolutedifference was 1.6 (95%CI 1.0-2.2), 1.9 (95%CI 1.4-2.4), and 1.6 (95%CI 1.2-2.0)respectively. Statistical equivalence testing between the two methods for hemoglobinconcentrations less than eight were considered not equivalent using a differencethreshold of 1 g/dL (p � 0.1) and were considered equivalent for hemoglobinconcentrations of 8-11.5 (p � 0.004) and �11.5 (p � 0.0004).

Conclusion: For hemoglobin values in the normal range, variances of 1-2 g/dLare not likely to result in altered decision-making. For extremes in hemoglobin,modest variances could be more important. Occasionally large variances occurred. Forthe majority of hemoglobin determinations, the non-invasive device was accurate to

within 1.6 g/dL.

Volume , . : September

413 Tactile Feedback Comparison of Three Types ofIntraosseous Access Devices for Needle InsertionAccuracy

Miller L, Philbeck T, Bolleter S, Garcia G/Vidacare Corporation, San Antonio, TX;San Antonio AirLife, San Antonio, TX; University of Texas Health Science Centerat San Antonio, San Antonio, TX

Study Objectives: Intraosseous (IO) catheter placement is an accepted standardfor vascular access when peripheral intravenous catheter placement is difficult orimpossible. There are also IO applications for bone marrow sampling and therapiessuch as vertebroplasty. For over 85 years clinicians have placed IO catheters into boneusing the manual technique of twisting and pushing, or the hammer method, withvariable results. Within the last decade rotary powered IO devices have beenintroduced, enabling clinicians to access the IO space with minimum effort and ahigh degree of accuracy. Increased interest in this technology has raised questionsconcerning procedural control and the ability of clinicians to discern catheter tiplocation within the bone. Tactile feedback is an important primary means fordetermining correct IO placement by clinicians for vascular access, especially for out-of-hospital and emergent applications. A study was designed by Vidacare Corporationto determine the relative precision of needle placement using only tactile feedback.The study also assessed the ability to access simulated osteoporotic bone withoutdamage using the 3 methods.

Methods: Three 11-gauge needles (manual, hammer-assisted, and rotary power)were inserted into simulated human bones by 11 operators (“Participants”). The totalthickness of each block was 9cm, with each consisting of 3 varying layers of simulatedcompact bone and 3 layers of cancellous bone. The simulated bone was stacked withan artificial skin and subcutaneous overlay further simulating varied resistanceencountered in actual insertions. Lastly, a foundation layer of cancellous material wasplaced beneath the stacked test layers to compensate for the varied thicknesses of theblocks and provide space for the catheter tip to extend during over-penetration.Blocks were draped to prevent visual bias from the participant. Time was measured inseconds, started at the beginning of bone insertion, and ended when the participantindicated the needle had been placed in the target area. Participants rated perceivedaccuracy of insertion on a 0-10 scale. Accuracy was assessed by fluoroscopy and resultswere classified as either success or failure. Success was defined as the needle tipresiding precisely in the correct test block layer. Participants also inserted needles intoraw eggs, simulating osteoporotic bones. Ability to place the needle without shelldamage was assessed.

Results: In simulated bones, mean insertion times in seconds were Manual20.7�10.1, Hammer 12.7�5.9, Power 8.7�2.8. Differences were significant(p�.001). Insertion success was Manual 48.5%, Hammer 69.7%, Power 97.0%;statistically significant (p�.001). Mean insertion certainty levels were: Manual 48%,Hammer 61%, Power 91%; statistically significant (p�.05). Simulated osteoporoticbone insertion rates (without damage) were Manual 19.2%, Hammer 36.4%, Power100%; statistically significant (p�.001).

Conclusions: Using tactile feedback only, rotary power may allow precise IOcatheter placement with greater success and confidence than manual or hammer-assisted devices. Powered insertion may facilitate penetration into fragile bonewithout damage. These findings may have far reaching clinical applications whendetermining the best method for accessing the IO space for vascular access and otherclinical procedures.

414 A Comparison of Point of Care Cerebral SpinalFluid Glucose Measurements With LaboratoryCerebral Spinal Fluid Glucose Measurements

Kennedy P, Mullins M/Washington University School of Medicine, St. Louis, MO

Study Objectives: Compare glucose levels measured by bedside point of care(POC) glucometer with laboratory glucose measurements of cerebral spinal fluid(CSF) in patients undergoing lumbar puncture (LP) for diagnosis of meningitis.

Methods: We enrolled eighteen adult emergency department (ED) patients who wereundergoing LP. The patients were approached prior to the procedure to obtain consent.The LP was performed by the patient’s emergency physician provider according to thestandard method of CSF collection. After CSF collection was completed for routinetesting, a single drop of CSF was placed directly from the spinal needle onto the POCglucometer test strip and the glucose level was recorded. This value was then compared tothe glucose result obtained from the CSF sent to the laboratory.

Results: CSF glucose levels obtained via POC glucometer generally overestimate

the actual glucose values as determined by laboratory testing. With the exception of

Annals of Emergency Medicine S133