an intensive insulin therapy protocol that required glucose determi-nations at least every 2 - 4 hours. Hypoglycemia was defined as �60mg/dl. Results: Data in the table are shown as means � SEM.

Patient Characteristics and Outcomes

Day 3Mean

Glucose(mg/dL) N

GenderM/F

Age(years) BMI Mortality % Sepsis

HypoglycemicEvents (�60

mg/dL)

SICU �150

8 2/6 67 � 5.3 33.9 � 5.2 0 0 0

SICU�150

12 6/6 71 � 2.8 37.2 � 3.0 7 (58%) 7 (58%) 4

BICU �150

4 1/3 50 � 12.3 26.4 � 1.4 0 2 (50%) 2

BICU�150

9 8/1 55 � 4.4 34.0 � 3.1 5 (56%) 7 (78%) 1

Admission glucose levels and mean glucose levels for day 1, 2 and3 for the groups below or equal to 150 mg/dL were 173, 146, 139 and127 mg/dL in the burn ICU; the corresponding levels in the surgicalICU were 260, 151, 118 and 111 mg/dL. Only the admission levelswere significantly different (p�0.01). The levels for the groups above150 mg/dL were 212, 211, 151 and 227 mg/dL in the burn ICU and205, 165, 158 and 192 mg/dL in the surgical ICU. Percent total burnsurface area in the burn ICU patients was 12 and 46 in the �150 and�150 mg/dL groups, respectively. Daily amounts of insulin admin-istered were similar in the two units for all groups. Overall only 36%of surgical ICU/burn ICU patients achieved glycemic goal of �150mg/dL by day 3 (40% in the surgical ICU and 31% in the burn ICU,p�NS), and hypoglycemic events (�60 mg/dL) were seen in 20% ofpatients (none with medical sequellae). Poor glycemic control wasexclusively associated with mortality in both burn ICU and surgicalICU patients (57%), and correlated highly with septic events. Con-clusions: Burn mortality is associated with poor glycemic control,percent total body surface area burn, and elevated BMI. We concludethat both surgical ICU and burn ICU patients benefit equally fromtight glycemic control obtained by intensive insulin therapy; sepsis-related deaths are reduced by intensive insulin therapy in the burnICU.

QS191. QUANTITATIVE BACTERIOLOGY AND WOUNDFAILURE IN HIGH ENERGY PENETRATING WARINJURIES. Jason Hawksworth1, Matthew R. Kasper2,Fred Gage3, Trevor Brown3, Douglas K. Tadaki3, Philip Per-due4, John Forsberg7, Alexander Stojadinovic8, Eric A. El-ster9; 1Department of Surgery, Walter Reed Army MedicalCenter, Washington D.C., DC; 2Department of Microbiology,National Naval Medical Center, Bethesda, MD; 3Regenera-tive Medicine Department, Combat Casualty Care, NavalMedical Research Center, Silver Spring, MD; 4Departmentof Surgery, National Naval Medical Center, Bethesda, MD;7Department of Orthopedics, National Naval Medical Cen-ter, Bethesda, MD; 8Department of Surgery, Walter ReedArmy Medical Center; Uniformed Services University ofHealth Sciences; Combat Wound Initiative, WashingtonD.C., DC; 9Department of Surgery, National Naval MedicalCenter; Naval Medical Research Center; Uniformed Ser-vices University of Health Sciences; Combat Wound Initia-tive, Bethesda, MD

Introduction: High-energy penetrating war injuries generatecomplex, contaminated wounds. Patients with these devastatinginjuries arrive at tertiary care military medical facilities heavilycolonized with fastidious environmental bacteria, such as Acineto-bacter baumannii. The cornerstone of war wound managementremains surgical debridement to remove devitalized tissue and

decrease bacterial colony counts. However, despite aggressive sur-gical management, some wounds fail, often secondary to ongoinginfection. Quantitative bacteriology has not previously been stud-ied in acute war wounds and may influence surgical decisionmaking. Methods: Patients with high-energy penetrating extrem-ity wounds sustained during combat in Operation Iraqi Freedom(OIF) were prospectively studied, and followed for 30 days afterdefinitive wound closure. The primary outcome was wound fail-ure, defined as delayed wound closure (�21 days from injury) orwound dehiscence. Surgical debridement with VAC applicationwas repeated every 48-72 hours until wound closure. Timing ofclosure was at the discretion of the attending surgeon. Wound bedtissue biopsy (1 cm3) and VAC effluent (5 mL) were collected priorto each wound debridement and evaluated for quantitative micro-bial counts, reported in colony forming units (CFU) per cm3 oftissue or effluent. Associations between categorical variables werestudied with Fisher’s exact test or Chi-squared test, asappropriate. Results: Nineteen penetrating extremity warwounds (mean size 687 �873 cm3) in thirteen male patients (meanage 22 years) were investigated. The mean Injury Severity Score(ISS) was 25 �13. There were seven (36.8%) wound failures,including four (21.1%) delayed wound closures and three (15.8%)wound dehiscences. Bacterial colonization (in order of prevalence:Acinetobacter, Staphylococcus Aureus, Enterococcus faecalis,Pseudomonas, and Escherichia coli) was evident in 68.4% and78.9% of patient wound tissue and effluent, respectively. Thepresence of associated vascular injury, high ISS (�25), and woundcolonization correlated with wound failure (p�0.05). At time ofdefinitive wound closure, tissue biopsy CFU �105 was associatedwith 67% rate of wound dehiscence (p�0.05). Sensitivity, specific-ity, positive and negative predictive value of CFU �105 for deter-mining wound failure was 92%, 40%, 81%, 67% respectively. Con-clusions: Tissue quantitative bacteriology is predictive of woundfailure in high-energy penetrating war injuries. Additionally,while common in war wounds, microbial colonization in associa-tion with concomitant vascular injury and higher overall injuryseverity is an important clinical determinant of wound outcome.Given these findings, war wound failure may represent infectiousmorbidity related to the compensatory counter-inflammatory re-sponse (CARS) to injury.

QS192. THE IMPACT OF METHYLATION DEMAND AND HO-MOCYSTEINE METABOLISM ON WOUND GRANU-LATION TISSUE FORMATION. Conrad Hamilton, KellyKrier, Naveed Saqib, Thomas R. Howdieshell; University ofNew Mexico HSC, Albuquerque, NM

Introduction: Methionine is adenylated to form S-adenosylmethionine,the methyl donor for virtually all known biological methylation reac-tions. The end products of these methyltransferase reactions are amethylated substrate and S-adenosylhomocysteine, which is hydro-lyzed to homocysteine and adenosine. Homocysteine, a known indepen-dent risk factor for atherothrombosis, has not been identified as a riskfactor for impaired wound repair. Recent reports suggest that supple-mental arginine may accelerate atherothrombosis via increased homo-cysteine generation. This study was designed to examine the effects ofarginine-induced methylation stress and homocysteine generation onwound granulation tissue formation. Methods: A ventral hernia,surgically created in the abdominal wall of 12 swine, was repairedwith silicone sheeting and skin closure. An osmotic infusion pump,inserted in a remote subcutaneous pocket, continuously deliveredsaline solution (n�6) or L-Arg (n�6) into the wound environment.Granulation tissue thickness was determined by ultrasonography.Fluid was aspirated serially from the developing wound compart-ment for measurement of nitrite/nitrate (NOx), homocysteine, andadditional amino acid concentrations. On day 14, the animalswere sacrificed, and the abdominal wall was harvested for histo-logic analysis. Data are expressed as mean � SEM. Repeated-

343ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS

measures ANOVA was used to determine differences amonggroups, and between baseline and subsequent time points, withsignificance at p�0.05. Results: In animals that received saline, a4-fold increase in granulation tissue thickness was measured dur-ing the 14 day interval. In contrast, in L-Arg treated animals, day14 granulation tissue thickness was unchanged from day 4 valuesof saline treated animals (10.1 � 1.1 mm vs. 20.2 � 1.7 mm at day14, p�0.05). Wound vessel count and vascular surface area esti-mates derived from image analysis of histologic sections were 2 to3-fold lower in L-Arg animals compared to controls (p�0.05).Progressive and sustained increases in wound fluid homocysteinelevels were noted in L-Arg treated animals compared to controls(peak, 23.2 �m/L vs. 17.3 �m/L at day 7, p�0.05). Methionine andcysteine concentrations were significantly higher in L-Arg treatedanimals compared to controls (peak, 31 �m/L vs. 16 �m/L at day7, p�0.05; 50 �m/L vs. 25 �m/L at day 9, p�0.05, respectively).Progressive and sustained increases in wound fluid NOx levelswere noted in L-Arg treated animals compared to controls (230�m/L vs. 75 �m/L at day 14, p�0.05). Arginine concentrationsremained 2 to 3-fold greater in L-Arg treated animals comparedwith controls over the entire time course (p�0.05). There were nosignificant differences in concentrations of ornithine, proline, orhydroxyproline noted between groups over the 14 day period.Conclusion: Supplemental arginine appears to be preferentiallymetabolized to homocysteine and nitric oxide rather than orni-thine, with consequent reductions in wound angiogenesis andcollagen production. Homocysteine is remethylated to methionineand eliminated by transsulfuration to cysteine. As endothelialcells are not capable of transsulfuration, the resulting increase inlocal homocysteine concentration might contribute to wound en-dothelial dysfunction and injury, and ultimately to impaired gran-ulation tissue formation.

QS193. NKT CELLS INFILTRATE CUTANEOUS WOUNDSAND REGULATE LOCAL INFLAMMATORY SIG-NALS. David F. Schneider, Jessica L. Palmer, Julia M.Tulley, Richard L. Gamelli, Douglas E. Faunce; Loyola Uni-versity Medical Center, Maywood, IL

Here we report the novel finding that Natural Killer T (NKT) cellscontribute to cutaneous wound healing as constituents of the earlyinflammatory infiltrate and regulators of inflammatory signals.Briefly, wild-type (WT) BALB/c and NKT cell-deficient J281komice were given dorsal 3 mm excisional punch wounds. At day 1,punch wounds in WT mice were 32.6% closed while wounds fromJa81ko mice were 52.3% closed. Compared to WT mice, woundclosure continued at a significantly faster rate in J281ko micethrough day 3 (36.4 vs. 66.5%, respectively). To determine if NKTcells, an infrequent lymphocyte subset, infiltrated excisionalwounds, we created wound cell suspensions to identify NKT cellsby flow cytometry. CD3� Ly49c� CD1d: IgG fusion protein�(NKT cells) were identified in WT wounds as early as 12 hourspost-injury. NKT cell infiltration peaked at 24 hours post-wounding, comprising 4.96% of live cells or approximately 92,000cells. Since NKT cells participate in the inflammatory infiltrate,we hypothesized that they also influenced local inflammatorysignals. To examine the inflammatory response in the absence vs.presence of NKT cells, wound homgenates were assayed for che-mokine content by ELISA. Compared to WT mice, day 1 woundsfrom Ja81ko mice had 5 to 7-fold higher levels of the neutrophilchemoattractants MIP-2 and KC and 3 and 8-fold higher levels ofthe monocyte/macrophage chemoattractants MIP-1a and MIP-1B,respectively. Due to their chemoattractant properties, these mol-ecules are traditionally associated with local immunity. However,when we examined wound neutrophil and monocyte/macrophagecontent by flow cytometry, we found that WT and Ja81ko woundsdid not differ in the magnitude or kinetics of either cell infiltratethrough day 5. Our results support the concept that absence of

NKT cells accelerates wound closure despite enhanced chemokinecontent. We have taken a novel approach to demonstrating thatNKT cells infiltrate cutaneous wounds with the early inflamma-tory infiltrate. The NKT infiltration correlates with transientenhancement of a select subset of locally produced neutrophil andmacrophage chemokines without a corresponding increase in theinfiltration of either cell type. Thus NKT cells appear to negativelyregulate certain local inflammatory signals. Modulating NKT cellactivity could become a useful therapeutic strategy in patientssuffering from delayed wound healing.

QS194. INFLAMMATORY CYTOKINE AND CHEMOKINE EX-PRESSION IS ASSOCIATED WITH WOUND FAIL-URE IN HIGH ENERGY PENETRATING WAR INJU-RIES. Jason Hawksworth1, Alexander Stojadinovic2, FredGage3, Douglas K. Tadaki3, Philip W. Perdue4, John Fors-berg6, Tom Davis3, Trevor Brown3, Eric A. Elster7; 1Depart-ment of Surgery, Walter Reed Army Medical Center, Wash-ington D.C., DC; 2Department of Surgery, Walter ReedArmy Medical Center; Uniformed Services University ofHealth Sciences; Combat Wound Initiative, WashingtonD.C., DC; 3Regenerative Medicine Department, Combat Ca-sualty Care, Naval Medical Research Center, Silver Spring,MD; 4Department of Surgery, National Naval Medical Cen-ter, Bethesda, MD; 6Department of Orthopedics, NationalNaval Medical Center, Bethesda, MD; 7Department of Sur-gery, National Naval Medical Center; Naval Medical Re-search Center; Uniformed Services University of HealthSciences; Combat Wound Initiative, Bethesda, MD

Introduction: The majority of modern war wounds are secondaryto high-energy penetrating ballistics. Current treatment of thesecomplex injuries includes multiple surgical debridements, sys-temic antibiotics, and liberal use of the vacuum assisted closure(VAC) device. However, despite technological advances, appropri-ate timing of war wound closure remains subjective. Objectivecriteria for decision algorithms to define timing of wound closureare lacking, and reliable predictors of wound healing outcomeremain undefined. Methods: Patients with high-energy penetrat-ing extremity wounds sustained during combat in Operation IraqiFreedom (OIF) were prospectively studied, and followed for 30days after definitive wound closure. The primary outcome waswound failure, defined as delayed wound closure (�21 days frominjury) or wound dehiscence. Surgical debridement with VACapplication was repeated every 48-72 hours until wound closure.Timing of closure was at the discretion of the attending surgeon.Wound bed tissue biopsy (1 cm3), VAC effluent (5 mL), and serum(2 mL) were collected prior to each wound debridement. Effluentand serum were analyzed for relevant cytokines and chemokines,and tissue for 190 wound-healing associated genes. Correlationsbetween wound outcome and potentially important clinical covari-ates were analyzed. Analysis of variance with repeated measureswas used to determine if molecular expression differed at eachtime point. Results: Nineteen penetrating extremity war wounds(mean size 687 �873 cm3) in thirteen male patients (mean age 22years) were investigated. The mean Injury Severity Score (ISS)was 25 �13. The majority (89.5%) of the wounds were secondaryto blast injury and nine (47.4%) of the wounds were traumaticamputations. There were seven (36.8%) wound failures, includingfour (21.1%) delayed wound closures and three (15.8%) wounddehiscences. The presence of associated vascular injury and highISS (�25) correlated with wound failure (p�0.05). Increased lev-els of serum (IL-6, IL-8, IL-10, MCP-1 and MIP-1�) and woundeffluent (IL-1�, IL-1�, IL-2, IL-10, MCP-1 and MIP-1�) inflamma-tory cytokines and chemokines were associated with wound failure(p�0.05). Wound tissue relative increase in marker-specific mRNAtranscription correlated with wound effluent cytokine expression. Con-

344 ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS