Association of Mechanism of Injury with Risk for Venous Thromboembolism After Trauma
Total Page:16
File Type:pdf, Size:1020Kb
Research JAMA Surgery | Original Investigation | ASSOCIATION OF VA SURGEONS Association of Mechanism of Injury With Risk for Venous Thromboembolism After Trauma Charles A. Karcutskie, MD, MA; Jonathan P. Meizoso, MD; Juliet J. Ray, MD; Davis Horkan, MD; Xiomara D. Ruiz, MD; Carl I. Schulman, MD, PhD, MSPH; Nicholas Namias, MD, MBA; Kenneth G. Proctor, PhD Invited Commentary page 41 IMPORTANCE To date, no study has assessed whether the risk of venous thromboembolism CME Quiz at (VTE) varies with blunt or penetrating trauma. jamanetworkcme.com OBJECTIVE To test whether the mechanism of injury alters risk of VTE after trauma. DESIGN, SETTING, AND PARTICIPANTS A retrospective database review was conducted of adults admitted to the intensive care unit of an American College of Surgeons–verified level I trauma center between August 1, 2011, and January 1, 2015, with blunt or penetrating injuries. Univariate and multivariable analyses identified independent predictors of VTE. MAIN OUTCOMES AND MEASURES Differences in risk factors for VTE with blunt vs penetrating trauma. RESULTS In 813 patients with blunt trauma (mean [SD] age, 47 [19] years) and 324 patients with penetrating trauma (mean [SD] age, 35 [15] years), the rate of VTE was 9.1% overall (104 of 1137) and similar between groups (blunt trauma, 9% [n = 73] vs penetrating trauma, 9.6% [n = 31]; P = .76). In the blunt trauma group, more patients with VTE than without VTE had abnormal coagulation results (49.3% vs 35.7%; P = .02), femoral catheters (9.6% vs 3.9%; P = .03), repair and/or ligation of vascular injury (15.1% vs 5.4%; P = .001), complex leg fractures (34.2% vs 18.5%; P = .001), Glasgow Coma Scale score less than 8 (31.5% vs 10.7%; P < .001), 4 or more transfusions (51.4% vs 17.6%; P < .001), operation time longer than 2 hours (35.6% vs 16.4%; P < .001), and pelvic fractures (43.8% vs 21.4%; P < .001); patients with VTE also had higher mean (SD) Greenfield Risk Assessment Profile scores (13 [6] vs 8 [4]; P Յ .001). However, with multivariable analysis, only receiving 4 or more transfusions (odds ratio [OR], 3.47; 95% CI, 2.04-5.91), Glasgow Coma Scale score less than 8 (OR, 2.75; 95% CI, 1.53-4.94), and pelvic fracture (OR, 2.09; 95% CI, 1.23-3.55) predicted VTE, with an area under the receiver operator curve of 0.730. In the penetrating trauma group, more patients with VTE than without VTE had abnormal coagulation results (64.5% vs 44.4%; P = .03), femoral catheters (16.1% vs 5.5%; P = .02), repair and/or ligation of vascular injury (54.8% vs 25.3%; P < .001), 4 or more transfusions (74.2% vs 39.6%; P < .001), operation Author Affiliations: Ryder Trauma time longer than 2 hours (74.2% vs 50.5%; P = .01), Abbreviated Injury Score for the Center, Division of Trauma and Surgical Critical Care Services, DeWitt abdomen greater than 2 (64.5% vs 42.3%; P = .02), and were aged 40 to 59 years (41.9% Daughtry Family Department of vs 23.2%; P = .02); patients with VTE also had higher mean (SD) Greenfield Risk Assessment Surgery, University of Miami Leonard Profile scores (12 [4] vs 7 [4]; P < .001). However, with multivariable analysis, only repair M. Miller School of Medicine, Miami, and/or ligation of vascular injury (OR, 3.32; 95% CI, 1.37-8.03), Abbreviated Injury Score for Florida (Karcutskie, Meizoso, Ray, Horkan, Ruiz, Schulman, Namias, the abdomen greater than 2 (OR, 2.77; 95% CI, 1.19-6.45), and age 40 to 59 years (OR, 2.69; Proctor); Ryder Trauma Center, 95% CI, 1.19-6.08) predicted VTE, with an area under the receiver operator curve of 0.760. Division of Burns, DeWitt Daughtry Family Department of Surgery, University of Miami Leonard M. Miller CONCLUSIONS AND RELEVANCE Although rates of VTE are the same in patients who School of Medicine, Miami, Florida experienced blunt and penetrating trauma, the independent risk factors for VTE are different (Karcutskie, Meizoso, Ray, Horkan, based on mechanism of injury. This finding should be a consideration when contemplating Ruiz, Schulman, Namias, Proctor). prophylactic treatment protocols. Corresponding Author: Kenneth G. Proctor, PhD, Ryder Trauma Center, Divisions of Trauma, Surgical Critical Care, and Burns, DeWitt Daughtry Family Department of Surgery, University of Miami Leonard M. Miller School of Medicine, 1800 NW 10th JAMA Surg. 2017;152(1):35-40. doi:10.1001/jamasurg.2016.3116 Ave, Ste T-215 (Box D40), Miami, FL Published online September 28, 2016. 33136 ([email protected]). (Reprinted) 35 Copyright 2017 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/01/2021 Research Original Investigation Mechanism of Injury and Risk for Venous Thromboembolism After Trauma rauma patients are typically classified according to the mechanism of injury—blunt or penetrating. However, Key Points advanced trauma life support guidelines do not distin- T Question Does mechanism of injury influence the independent 1 guish management based on the mechanism of injury. Sev- risk factors for venous thromboembolism after trauma? eral studies have found significantly different outcomes be- Finding This cohort study finds that, although the Greenfield Risk tween patients who experienced blunt and penetrating trauma Assessment Profile accurately predicts venous thromboembolism who were resuscitated to similar end points,2-5 even when se- after blunt and penetrating trauma, independent risk factors are 6-12 verity of the injuries are closely matched. different based on the mechanism of injury. Venous thromboembolism (VTE) is a significant cause Meaning Risk assessment and prophylactic strategies for venous of late morbidity and mortality after trauma13 despite thromboembolism may improve if mechanism of injury is thromboprophylaxis.14-27 At least 2 models stratify risk of VTE considered. specifically in trauma patients to guide surveillance and pro- phylaxis: the Trauma Embolic Scoring System28 and the Green- field Risk Assessment Profile (RAP).18 To our knowledge, no groups using χ2 or Fisher exact test, as appropriate. Paramet- study has assessed the differences in risk of VTE from the in- ric data are expressed as mean (SD) and compared using t tests dependent risk factors of blunt and penetrating trauma. for 2 independent samples. Nonparametric data are ex- We stratified patients in the intensive care unit by either pressed as median (interquartile range) and compared using blunt or penetrating trauma mechanism, and then deter- the Mann-Whitney test. Variables demonstrating statistical sig- mined which RAP risk factors are associated with VTE. We hy- nificance on univariate analyses were entered into a multi- pothesized that the mechanism of injury alters the indepen- variable logistic regression model. Odds ratios (ORs) and dent VTE risk factors. 95% CIs are reported. Area under the receiver operator curve (AUROC) was used to assess predictive values of models. P ≤.05 was considered statistically significant for all results. Methods This was a retrospective cohort study conducted at the Ryder Results Trauma Center in the University of Miami and Jackson Memo- rial Medical Center and was approved by the University of Mi- The study population comprised 1137 patients: 813 with blunt ami Institutional Review Board with a waiver of informed con- trauma and 324 with penetrating trauma. Patient demograph- sent. Adult patients sustaining blunt or penetrating trauma and ics and injury characteristics are shown in Table 2. admitted to the Ryder Trauma Center intensive care unit from There was no difference in rates of VTE, deep vein throm- August 1, 2011, to January 1, 2015, were included. Patients who bosis, or pulmonary embolism based on mechanism of in- were younger than 18 years, incarcerated, pregnant, or died less jury. In patients with blunt trauma, the overall VTE rate was than 72 hours after admission were excluded. 9% (n = 73), with a deep vein thrombosis rate of 6.6% (n = 54) Thirty percent of the population was concurrently en- and a pulmonary embolism rate of 2.7% (n = 22). In patients rolled in a prospective observational trial evaluating hyperco- with penetrating trauma, the overall VTE rate was 9.6% (n = 31), agulability and VTE in patients who have undergone trau- with a deep vein thrombosis rate of 7.7% (n = 25) and a pul- matic injury. This group was deemed to be at high risk for VTE, monary embolism rate of 2.5% (n = 8). using a RAP score of 10 or more, based on previous work.25,29 In patients with blunt trauma, those with VTE had a worse All high-risk patients enrolled in this concurrent study re- mean (SD) Injury Severity Score (29 [14] vs 21 [11]; P ≤ .001), ceived weekly surveillance venous duplex ultrasonography admission heart rate (102 [26] vs 95 [23] beats/min; P = .03), (VDU) of the lower extremities, in addition to chemical and me- admission systolic blood pressure (122 [30] vs 135 [32]; P = .01), chanical prophylaxis. Deep vein thrombosis was diagnosed admission Glasgow Coma Scale score (11 [5] vs 13 [4]; P ≤ .001), with VDU. Pulmonary embolism was identified by computed and base deficit (–4 [5] vs –2 [5]; P = .007). As expected, mean tomography with angiography of the chest. The VDU proto- (SD) RAP score (13 [6] vs 8 [4]; P ≤ .001) and median (inter- cols at our institution have been previously reported.30 quartile range) hospital length of stay (40 [22-78] vs 12 [7-26] Briefly, all VDUs were performed by certified ultrasonog- days; P ≤ .001) were also higher in patients with VTE. There raphy technologists and interpreted by an attending radiolo- was no difference in time to first prophylaxis between pa- gist. The deep venous systems of both lower extremities are tients with VTE and those without VTE.