TRAUMA PATIENTS ARE AT PARTICULARLY great risk forthe development of deep venous thrombosis(DVT) because of hypercoagulability, immobiliza-tion, and venous injury.1-11 Geerts et al12 reporteda DVT incidence of 58% in trauma patients whodid not receive prophylaxis. The incidences ofDVT vary, depending on whether patients routine-ly undergo ultrasound scanning. Spain et al13
recently reported a 5% incidence of DVT in high-risk patients but only performed duplex Doppler
scanning when there was clinical suspicion. Otherinvestigators who used routine surveillance duplexscanning in high-risk patients report an incidencefrom 5% to 23%, with asymptomatic DVTs com-prising the majority.1-8,12 Screening of asymp-tomatic patients may significantly reduce the riskof unanticipated pulmonary embolism (PE).Multiple studies have demonstrated that prophy-laxis can significantly reduce the incidence of DVTin the severely injured patient; however, depend-ing on the method of prophylaxis used, 5% to 45%of these patients still experience the developmentof a venous thromboembolus (VTE).1-8,12
Many risk factors have been historically associ-ated with the development of DVT; however, it isunclear which individual or combination of riskfactors defines a high-risk group. Greenfield et al1
developed a scoring system that weighted risk fac-tors that are commonly associated with an
The risk assessment profile score identifies trauma patients at risk for deep vein thrombosisMichelle M. Gearhart, PharmD, Fred A. Luchette, MD, Mary C. Proctor, MS, Dave M. Lutomski, MS,Christine Witsken, MS, Laura James, MS, Kenneth Davis, Jr, MD, Jay A. Johannigman, MD, James M.Hurst, MD, and Scott B. Frame, MD, Cincinnati, Ohio, and Ann Arbor, Mich
Background. The identification of trauma patients at risk for the development of deep venous thrombosis(DVT) at the time of admission remains difficult. The purpose of this study is to validate the risk assess-ment profile (RAP) score to stratify patients for DVT prophylaxis.Methods. All patients admitted from November 1998 thru May 1999 were evaluated for enrollment. Weprospectively assigned patients as low risk or high risk for DVT using the RAP score. High-risk patientsreceived both pharmacologic and mechanical prophylaxis. Low-risk patients received none. Surveillanceduplex Doppler scans were performed each week of hospitalization or if symptoms developed. Hospitalcharges for prophylaxis were used to determine the savings in the low-risk group. Statistical differencesbetween the risk groups for each factor of the RAP and development of DVT were determined by the chi-squared test, with significance at a probability value of less than .05.Results. There were 102 high-risk (64%) and 58 low-risk (36%) individuals studied. Eleven of thehigh-risk group (10.8%) experienced the development of DVT (asymptomatic, 64%). None of the low-risk group was diagnosed with DVT. Five of the 16 RAP factors were statistically significant for DVT.Eliminating prophylaxis and Doppler scans in low-risk patients resulted in a total savings of $18,908in hospital charges.Conclusions. The RAP score correctly identified trauma patients at increased risk for the development ofDVT. Despite prophylaxis, the high-risk group warrants surveillance scans. Withholding prophylaxis inlow-risk patients can reduce hospital charges without risk. (Surgery 2000;128:631-40.)
From the Division of Trauma and Critical Care, the Departments of Surgery and Pharmacy Services, Universityof Cincinnati College of Medicine, Cincinnati, Ohio, and University of Michigan Medical Center, Ann Arbor,Mich
Presented at the 57th Annual Meeting of the Central SurgicalAssociation, Chicago, Ill, March 2-4, 2000.
Reprint requests: Fred A. Luchette, MD, Department ofSurgery, ML 0558, 231 Bethesda Ave, Cincinnati, OH 45267-0558.
increased incidence of DVT. This tool is referred toas the Risk Assessment Profile (RAP) score and is asimple means of stratifying patients within 24 hoursof admission according to their potential for thedevelopment of DVT. In their study, patients with aRAP score of 5 or more were at high-risk for thedevelopment of DVT. Cost analysis of the risk/ben-efit for VTE revealed prophylaxis was beneficial inthis high-risk group. In contrast, a RAP score of lessthan 5 had a low incidence of DVT, and prophylax-is was not cost effective.1 Cafferata et al14 reportedtheir efforts to calculate a risk factor score for trau-ma patients using a slightly modified version of theInternational Society for Cardiovascular Surgery/Society for Vascular Surgery reporting standards.15
They used only 5 parameters (age, long bone frac-ture, immobilization, anesthesia time, previousVTE) with each weighted from 0 to 4. Risk wasdetermined to be low, moderate, or high. Theseinvestigators could not make a recommendationfor prophylaxis or a need for surveillance venousduplex scans.14 The purpose of the current studywas to assess and validate the RAP score for theidentification of injured patients at risk for thedevelopment of DVT and to evaluate guidelinesregarding DVT prophylaxis and surveillance scan-ning in trauma patients.
PATIENTS AND METHODSStudy protocol. This was a prospective study to
evaluate a protocol that used the RAP score toidentify patients who are at risk for the develop-ment of DVT and thus who might benefit from pro-phylaxis (Figure). Inclusion criteria were 18 yearsof age or older, expected length of stay greaterthan 36 hours, and a RAP score of more than 2.Patients were excluded if they had a known hyper-coagulable state (anti-thrombin III deficiency, pro-tein C or S deficiency, or anti-phospholipid-proteinantibodies), if they were pregnant, lactating, or
incarcerated, or if they had previously undergoneheparin, warfarin, or low–molecular-weight he-parin (LMWH) therapy.
All patients were evaluated for study enrollmentwithin 24 hours of admission. Once inclusion andexclusion criteria were met, a RAP score (Table I)was calculated, and the patients were classified aslow risk (RAP score, >2 and <5) or high risk (RAPscore, ≥5) for the development of DVT. Patientswere followed for 30 days.
The high-risk group received both pharmaco-logic and mechanical prophylaxis. Either unfrac-tionated-heparin (5000 units, subcutaneously every12 hours) or enoxaparin (30 mg, subcutaneouslyevery 12 hours) was prescribed by the individualtrauma surgeon. Compression devices includedbelow-the-knee single-sleeve pneumatic compres-sion devices (Jobst Athrombic Pump-System 2000;Jobst, Charlotte, NC) or arteriovenous foot pumps(Kendall A-V Impulse System; Kendall Co,Mansfield, Mass). A Doppler scan was performedduring the first 7 days of admission and then eachweek during the hospitalization on all studyenrollees. Patients underwent a final duplex scanon day 30 either in the hospital or as an outpatient.Low-risk patients did not receive any form of pro-phylaxis.
Data collection included age, gender, mecha-nism of injury, Injury Severity Score (ISS),16 fre-quency of each RAP score variable, methods of pro-phylaxis, and the date and results of Doppler scans.Once diagnosed with an acute DVT or PE, thepatient began the appropriate treatment and wasremoved from the study.
Venous duplex ultrasonography. A registered orregistry-eligible vascular technologist in an Inter-Societal Commission for Accreditation of VascularLaboratories accredited laboratory performed allvenous duplex examinations. Each technician hadat least 2 years of experience. Static images of per-
Figure. Algorithm used to determine use of DVT prophylaxis based on the RAP score and neurosurgical con-sultation.
Surgery Gearhart et al 633Volume 128, Number 4
tinent findings were made on color print film forinterpretation by board-certified vascular surgeons.Complete studies were recorded for detailedreview.
Examinations were performed with the use of anAdvance Technology Laboratories HDI 3000CVUltrasound System (ATL Ultrasound, Bothell,Wash) with either a linear 7.4-mHz or curved array4.2-mHz transducer probe. With the patient in thesupine position, the deep venous system of bothlower extremities was examined from the proximalcommon femoral vein to at least the distal poplitealvein. The peroneal, posterior, and anterior tibialveins were also visualized when there were no dress-ings or splints in place. With the use of a transverseimaging plane, the entire length of the deep veinswere assessed for compressibility and the presenceof intraluminal echoes. Spontaneous and phasicaugmentation of the Doppler signal and intralumi-nal defects was assessed with longitudinal images.
A normal study was characterized by completevenous compressibility with a Doppler velocity sig-nal that was spontaneous and phasic. Lack of com-pressibility with the presence of low-level intralumi-nal echoes and an absent or continuous Dopplervelocity signal defined an abnormal study.
Hospital charges. The hospital charge for a com-plete bilateral venous duplex Doppler examinationwas $371.00. Charges for unfractionated heparinand LMWH were $15.86 per day and $130.50 perday, respectively. There were 2 charges for the calfpneumatic compression device and arteriovenousfoot pump: a 1-time charge of $37.50 for the sleeveor foot pump and a daily charge of $75.42 for theair pump. These charges were used to determinesavings for the low-risk group, which did notreceive chemical or mechanical prophylaxis.
Statistical analysis. Data are reported as mean ±SEM and analyzed with the Student t test. The chi-squared test was used for determining differencesamong frequencies for each variable of the RAPscore and the development of DVT. The rate ofDVT for the high-risk and low-risk groups was test-ed for significance with Fisher’s exact test.Statistical significance was set at .05%.
RESULTSThree hundred fourteen patients were admitted
and seen by 1 of 5 trauma surgeons betweenNovember 15, 1998, and May 20, 1999. Of the totaladmissions 184 individuals (59%) met the study cri-teria and were enrolled. Twenty-four patients didnot complete the study because a final Duplex studywas not obtained or because of a protocol violation.The study sample consisted of 160 participants
(87%) who completed all phases of the study. Fifty-eight patients (36.3%) with RAP scores between 2and 5 were stratified as low risk. The remaining 102patients (63.8%) comprised the high-risk group.Age, ISS, and mechanism of injury for the high-riskgroup were statistically greater (P < .05) comparedwith the low-risk group (Table II). However, subsetanalysis of the high-risk individuals did not revealage, ISS, or mechanism of injury to be associatedwith the development of DVT (Table II).
As expected, significant differences were foundbetween risk factors for the high-risk and low-riskgroups (Table III). Obesity, transfusion of morethan 4 units of packed red cells in the first 24hours, operations that lasted longer than 2 hours,severe head injury (Abbreviated Injury Score17
[AIS], > 2), and complex fractures of the lowerextremities were the only factors that were statisti-cally significant in a comparison of the high-riskand low-risk groups (Table III). In the high-riskgroup, 11 of 102 patients (10.8%; 95% confidenceinterval [CI], 5.5%-18.5%) had a DVT diagnosedby Doppler scan versus zero of 58 patients (0%;95% CI, 0.0%-6.2%) in the low-risk group (P = .0077). Only 4 DVTs (36%) were symptomatic
Table I. Individual risk factors and points allottedto calculate the RAP score
Points
Underlying conditionObesity 2Malignancy 2Abnormal coagulation 2History of thromboembolism 3
Iatrogenic factorsFemoral venous line 2Transfusion, >4 units 2Operation, >2 h 2Major venous repair 3
although most (n = 7 patients; 64%) were asympto-matic and identified by surveillance Doppler scans.Within the high-risk group, there was no statisticaldifference in the RAP score associated withoutDVT or the presence of DVT (7.5 ± 0.25 vs 7.8 ±0.64, respectively; P = .7163). The patients in whoma DVT developed were slightly younger than theother high-risk patients who did not experience thedevelopment of a DVT (40.5 ± 4.78 vs 43.7 ± 2.05years; NS). Seventy-three percent of the patients (n= 8 patients) who experience the development ofDVT were male. The frequency of the individualRAP score risk factors associated with the 11 diag-nosed DVTs were more than 4 units of blood in thefirst 24 hours (n = 6 patients; 55%), a surgical pro-cedure longer than 2 hours (n = 5 patients; 45%),obesity (n = 4 patients; 36%), a repair of venousinjury (n = 3 patients; 27%), a head injury with anAIS of more than 2 (n = 3 patients; 27%), a pelvic
fracture (n = 3 patients; 27%), age between 40 and60 years (n = 3 patients; 27%), an abdominal AIS ofmore than 2 (n = 2 patients; 18%), a Glascow ComaScore of less than 8 (n = 2 patients; 18%), agebetween 60 and 74 years (n = 2 patients; 18%), afemoral venous line (n = 1 patient; 9%), and coag-ulopathy (n = 1 patient; 9%). The number and fre-quency of each risk factor for the low-risk and high-risk groups are shown in Table IV. Segregation ofthe high-risk group into individuals in whom riskfactors were present or absent and its relationshipto the occurrence of DVT are given in Table V. Thetransfusion of 4 or more units of blood during theinitial 24 hours was the only independent variablefor the development of DVT (P ≤ .001). Althoughmany of the groups are small, it is noteworthy thatno DVTs developed in any patients with complexlower extremity fractures or with chest trauma.
Doppler scan reports were reviewed for com-pleteness and limitations of the examinations inview of the high frequency of complex lowerextremity fractures in the high-risk group (30%)and in 21% of the entire study population. Allpatients in the lowrisk group underwent a com-plete duplex study of the venous system in bothlegs. In contrast, only 75% of the 102 high-riskpatients underwent complete Doppler scans. Thevascular technician was able to perform only a lim-ited study in the remaining 25 patients for 1 ormore of the following reasons: plaster casts, dress-ings, splint devices, restricted mobility of theextremity, or pain. The importance of the techni-cian’s ability to perform a complete evaluation ofboth lower extremities is evident in that most of theDVTs (n = 9 patients; 82%) were diagnosed aftercomplete Doppler scans of the entire venous sys-tem of each extremity were obtained. Most ofvenous thrombi (n = 6 patients; 55%) were in theproximal femoral system. The locations of distalthrombi were popliteal (n = 2 patients; 18%), tibial(n = 2 patients; 18%), and gastrocnemius (n = 1patient; 9%). The only patient with a PE had a nor-mal duplex scan of both lower extremities but anacute left subclavian DVT. Nine prophylactic venacaval filters (Stainless Steel Greenfield Vena CavalFilter; Meditech, Watertown, Mass) were inserted
634 Gearhart et al SurgeryOctober 2000
Table II. Demographics for high-risk and low-risk groups
Mechanism of injuryAge (y) Gender (F/M) ISS (Blunt/penetrating)
in study patients through the femoral route. Only 1instance of proximal DVT at the insertion siteoccurred in this group. All of the high-risk patientswho received a diagnosis of an acute DVT under-went standard anticoagulation with either weight-based unfractionated heparin or LMWH. Warfarin(Coumadin) therapy was begun as dictated by theclinical situation.
The mortality rate for this study was 1.3% (n = 2patients). Both deaths were patients in the high-risk group; 1 patient had been treated for DVT.The cause of death was attributed to their injuriesand not related to VTE complications.
All but 2 high-risk patients received a pharma-cologic agent for DVT prophylaxis, with most ofthe patients receiving unfractionated heparin (n =83 patients; 81%). The remainder of the patients(n = 17 patients; 17%) underwent prophylaxis withenoxaparin. This included patients with an AISgreater than 2 for head injury (n = 26 patients;93%), spinal cord injury (n = 6 patients; 100%),pelvic fracture (n = 15 patients; 100%), and solidorgan injuries (abdominal AIS, > 2; n = 21 patients;100%) managed nonoperatively. None of thepatients had bleeding complications, hypersensitiv-ity reaction, or thrombocytopenia attributed to theheparin agents. One of the 2 patients with severehead injuries who did not receive any pharmaco-
logic prophylaxis experienced the development ofDVT. Mechanical compression was used in 74% ofthe patients in the high-risk group. Most DVTs(8/11 patients; 73%) occurred despite both meth-ods of prophylaxis. Seventy-three percent of thisgroup received unfractionated heparin therapy,and 17% received LMWH, which was similar to theoverall use of heparin in the high-risk group. Tenof 11 DVTs occurred despite pharmacologic pro-phylaxis. Of these 10 DVTs, 8 patients receivedunfractionated heparin therapy and a compressiondevice, and 2 patients underwent prophylaxis withonly enoxaparin therapy.
The elimination of prophylaxis and surveillanceduplex Doppler scans for the low-risk group result-ed in a total savings of $18,908 in hospital charges.
DISCUSSIONThis study supports the results of the pilot study
by Greenfield et al,1 which showed patients with aRAP score of 5 or more were 3 times more likely toexperience the development of VTE than patientswith a RAP score of less than 5. There was a 10.8%incidence of DVT in the high-risk patient popula-tion, which is comparable to the rate that othersreported when prophylaxis and surveillance scanswere used. Although only 1 patient experiencedthe development of a PE (1/102 patients), this is
Surgery Gearhart et al 635Volume 128, Number 4
Table IV. Occurrence of risk factors in high-risk and low-risk groups
Risk factor High risk* (%) Low risk† (%) Total‡ (%)
lower than other studies reported, and we believethat this is due to early detection and treatment ofasymptomatic DVTs. This study also reaffirms theconclusion by Greenfield et al1 that it is not costeffective to perform prophylaxis on patients in thelow-risk population. After injury, patients should begiven a RAP score calculation within the first 24hours of admission to determine their risk fordevelopment of DVT. If the score is low-risk (RAPscore, <5), there is minimal probability for thedevelopment of DVT such that the cost of prophy-laxis and monitoring with routine Doppler scan isnot warranted. The elimination of both in the low-risk individual allows for a considerable reductionin hospital charges (at least $35,000 annually for116 low-risk admissions). There were no DVTsfound on the 1-month follow-up Duplex scans;therefore, eliminating this follow-up scan after dis-charge may be another cost-saving effort.Additionally the complications that were reportedwith heparin prophylaxis include bleeding, hem-atomas, and thrombocytopenia could be averted.Knudson et al18 and Geerts et al12 reported a rateof 5% and 1.7%, respectively, for hemorrhagiccomplications as the result of LMWH.
We were able to document and follow the modeof chemical and mechanical prophylaxis for allhigh-risk patients. However, because the method ofprophylaxis was not controlled (it was left to the
physician’s discretion) and because of the lownumber of patients who received LMWH, we wereunable to draw any conclusions about whichmethod of prophylaxis was more effective. We werealso unable to determine whether compressiondevices in conjunction with chemical prophylaxiswere of any benefit compared with either alone.
Many groups have attempted to classify patientsaccording to risk for VTE events on the basis ofscores derived from individual risk factors. Cafferataet al14 stratified patients as low, moderate, and highrisk and developed guidelines based on the patient’srisk. The score and risk category have the potentialto change during the hospital course. In contrast, theRAP score used in the current study takes intoaccount 16 risk factors, several that were not includ-ed in the scoring system used by Cafferata et al. TheRAP score is determined within 24 hours of admis-sion as high or low risk and does not change.Theoretically, this deletion of additional informationcould affect a patient’s outcome; however, in the cur-rent study, it did not. All patients in whom a DVTdeveloped were in the high-risk patient sample andwere correctly identified by the RAP score. Nopatient in the low-risk group experienced the devel-opment of a DVT during the hospital stay.
Similar to our findings, Dennis et al5 reportedthat their high-risk patients who experienced thedevelopment of a DVT had a higher ISS and were
636 Gearhart et al SurgeryOctober 2000
Table V. Frequency of risk factors in high-risk patients and relationship to the development of DVT
Risk factor No. with DVT/with risk factor (%) No. with DVT/without risk factor (%)
PRBC, Packed red blood cells; AIS, abbreviated injury score.*Chi-squared test, 5.320; P = .021.
older than the high-risk patients who did not expe-rience the development of a DVT. Geerts et al12
identified the following factors as significant associ-ations with thrombosis: age, lower extremity frac-ture, spinal cord injury, blood transfusion withinthe first 24 hours after admission, and prolongedoperation. Our findings for risk factors in the high-risk group are similar to the results Geerts et al;however, the only statistically significant risk factorwe identified that correlated with the developmentof DVT was transfusion of more than 4 units ofblood within 24 hours of admission. We identified5 DVTs in 19 patients with prolonged operativetime, but this failed to reach statistical significance(p = .056) as an independent variable. This may bedue to the small sample size.
Another topic of controversy is whether headinjury, spinal cord injury, and pelvic fracture shouldreceive chemical prophylaxis because of the poten-tial risk of bleeding complications. Greenfield et al,1
Knudson et al,18 and Cafferata et al14 each excludedthis subset of patients for study. Yet, these are thepatients who are at increased risk for the develop-ment of a DVT or a PE. This was true in the currentstudy. There was a total of 28 patients with a severeclosed head injury (AIS, >2) in the present study.After consultation with the neurosurgical service, allbut 2 of these patients received either unfractionat-ed-heparin or LMWH therapy. Despite the aggressiveuse of pharmacologic prophylaxis in this subset ofhigh-risk patients, 8% of the patients (2/26 patients)experienced the development of a DVT without asecondary brain injury related to bleeding from theuse of either heparin agent. In contrast, a DVT wasidentified in 1 of the 2 patients who did not receivechemical prophylaxis therapy.
In conclusion, this study independently validatesthe RAP score as an effective tool for the stratifica-tion of trauma patients as high-risk or low-risk forthe development of a DVT. The protocol used inthis study avoids prophylaxis therapy in the low-riskgroup without increased patient risk. The elimina-tion of surveillance Duplex scans and pharmaco-logic and mechanical prophylaxis in the low-riskgroup reduces hospital charges. In contrast, the10.8% DVT rate in the high-risk group despite bothmodes of prophylaxis justifies routine weeklyDoppler scans because most of the DVTs wereasymptomatic. Additional studies of new methodsof DVT prophylaxis are needed to identify moreefficacious agents to further reduce this commoncomplication after injury.
We thank Michael Ottaway, MS, and Clyde Miyagawa,PharmD, for technical assistance.
Cipolle MD, Cho J. Posttrauma thromboembolism prophy-laxis. J Trauma 1997;42:100-3.
2. Napolitano LM, Garlapati VS, Heard SO, et al.Asymptomatic deep venous thrombosis in the traumapatient: is an aggressive screening protocol justified? JTrauma 1995;39:651-7.
3. Shackford SR, Davis JW, Hollingsworth-Fridlund P, Brewer,Hoyt D, Mackersie RC. Venous thromboembolism inpatients with major trauma. Am J Surg 1990;159:365-9.
4. Knudson MM, Lewis FR, Clinton A, Atkinson K, MegermanJ. Prevention of venous thromboembolism in traumapatients. J Trauma 1994;37:480-7.
5. Dennis JW, Menawat S, Thron JV, et al. Efficacy of deepvenous thrombosis prophylaxis in trauma patients and iden-tification of high-risk groups. J Trauma 1993;35:132-8.
6. Kudsk KA, Fabian TC, Baum S, Gold RE, Mangiante E,Boeller G. Silent deep vein thrombosis in immobilized mul-tiple trauma patients. Am J Surg 1989;158:515-9.
7. Geerts WH, Jay RM, Code KI, et al. A comparison of low-dose heparin with low–molecular-weight heparin as pro-phylaxis against venous thromboembolism after major trau-ma. N Engl J Med 1996;335:701-7.
8. Piotrowski JJ, Alexander JJ, Brandt CP, McHenry CR, YuhasJP, Jacobs D. Is deep vein thrombosis surveillance warrantedin high-risk trauma patients? Am J Surg 1996;172:210-3.
9. Chen JP, Rowe DW, Enderson BL. Contrasting post-trau-matic serial changes for D-dimer and PAI-1 in criticallyinjured patients. Thromb Res 1999;94:175-85.
10. Engelman DT, Gabram SG, Allen L, Ens GE, Jacobs LM.Hypercoagulability following trauma. World J Surg1996;20:5-10.
11. Owings JT, Bagley M, Gosselin R, Romac D, Disbrow E.Effect of critical injury on plasma antithrombin activity: lowantithrombin levels are associated with thromboemboliccomplications. J Trauma 1996;41:396-405.
12. Geerts WH, Code KI, Jay RM, Chen E, Szalai JP. A prospec-tive study of venous thromboembolism after major trauma.N Engl J Med 1994;331:1601-6.
13. Spain DA, Richardson JD, Polk HC Jr, Bergamini TM,Wilson MA, Miller FB. Venous thromboembolism in thehigh-risk trauma patient: do risks justify aggressive screen-ing and prophylaxis? J Trauma 1997;42:463-7.
15. Porter JM, Moneta GL, and an international consensuscommittee on chronic venous disease. Reporting standardsin venous disease: an update. J Vasc Surg 1995;21;635-45.
16. Baker SP, O’Neil B, Haddon W, et al. The injury severityscore: a method for describing patients with multipleinjuries and evaluating emergency care. J Trauma1974;14:187-96.
17. American Association for Automotive Medicine. TheAbbreviated Injury Scale (AIS): 1990 revision. Des Plaines,Illinois.
18. Knudson MM, Morabito D, Paiement GD, Shackford S. Useof low molecular weight heparin in preventing throm-boembolism in trauma patients. J Trauma 1996;41:446-59.
DISCUSSION
Dr David A. Spain (Louisville, Ky). Discussing DVTprophylaxis with trauma surgeons is worse than dis-cussing religion or politics; discussions usually deterio-
Surgery Gearhart et al 637Volume 128, Number 4
rate into emotional ranting. So I congratulate DrLuchette on bringing some objective data to the table.
One, transfusion of more than 4 units and operatingroom time of more than 2 hours were associated withDVT. Did you perform logistic regression analysis to seewhether these were independent factors?
Two, how soon after admission did you apply yourcompression devices or start heparin? How much of ahead, spine, or solid organ injury is a contraindication tosubcutaneous heparin? I ask this because complicationsmay be rare, but they are potentially catastrophic.
Three, is a combination of compression devices andheparin better than either one alone?
Four, your incidence of 6% proximal DVT and 1%PE is identical to that in studies that use injury patternsto identify your high-risk patients. So has the RAP scorereally improved your patient selection for those at highrisk?
Five, I am not convinced that serial duplex scans havechanged the management of these patients. What is thenatural history of an asymptomatic DVT found on scanin a patient who is already receiving subcutaneousheparin? Could we not just use some form of prophylax-is and assume it is going to work well for 94% to 99% ofour patients and avoid the serial duplex scanning at anenormous cost?
Finally, you referred to the study by Geerts et al,7
which I think makes the point that DVT is an enormousproblem in trauma patients. However, in that study ofpatients who did not receive prophylaxis, the incidenceof pulmonary embolism was only 1%. Can we reallyimprove on that, and is it cost effective?
Dr Sidney F. Miller (Dayton, Ohio). Could you com-ment about any correlation with femoral lines in yourstudy? Some of us feel fairly strongly about that.
From your study, is there a combination of findingson your Doppler ultrasound scans and your RAP scoresthat might suggest the earlier use of Greenfield filters inthese patients?
Dr Thomas J. Howard (Indianapolis, Ind). Do youhave any data on the sensitivity and specificity of duplexscanning in your institution with regards to some othergold standard, either venography or helical computedtomography scanning? How have these results impactedyour data?
Dr William D. Turnipseed (Madison, Wis). Who doesyour venous testing? Is it the Department of Radiology oris it your vascular laboratory? The philosophy of mostradiology groups is that doing infrapopliteal surveillanceis inappropriate, and they do not even look at the tibiallevel of the venous system. Most peripheral vascular lab-oratories consider a complete examination to includethe tibial veins. So much would depend, in interpretingthese data, on what the definition of a complete exami-nation is and by whom it is performed.
Dr Bruce A. Harms (Madison, Wis). Have you lookedat the pelvic fractures in this group of patients? I guess Iwould be concerned about your classification of low-riskpatients if your patient had no other risk factors than anisolated pelvic fracture. There is a difference in pelvicfractures, and all are not created equal. If there is poste-rior element involvement that might impact on operativeintervention or delay in ambulation, I think that thattype of patient should undergo prophylaxis.
Dr Mark A. Malangoni (Cleveland, Ohio). First, areyou really comfortable recommending that low-riskpatients not receive prophylaxis? You had fewer than 60patients in that group. I am not so comfortable. Pleasecomment.
Second, Are you going to revise the RAP score basedon what you have now found to be significant predic-tive parameters and make it a little easier to use andperhaps even have better predictability than you foundin your study?
Dr Luchette. Dr Spain, you asked whether we per-formed multiple logistic regression to identify the trans-fusion of more than 4 units and the operating room timeof more than 2 hours as independent risk factors. We didnot because of the small numbers of patients in the high-risk and low-risk groups. We recognize this work is essen-tially a pilot study, and we are not going to make any rec-ommendations on changing the RAP score or any otheraspects of the design.
You also asked how soon heparin was started, particu-larly in the patients with head and spine injuries. Wedelayed 24 hours from the time of admission and/orafter coagulation abnormalities were corrected. The rea-son our neurosurgeons have been a bit more cooperativeis a recent study in the neurosurgical literature that usedprophylactic heparin in patients who underwent electivecraniotomy. In that study, there were no significant com-plications reported. Our preliminary data would supportthat heparin in prophylactic doses is safe in the patientwith head injury. We did not do a subset analysis of thepatients with head, spine, and solid organ injury for com-plication because of the small number of patients in eachgroup. I do agree with you that, although complicationsmay be rare, they do carry potentially catastrophic effectsfor the patient.
I do believe that the combination of a compressiondevice with a pharmacologic agent is better than eitherone alone. In the study by Greenfield et al,1 we random-ized for compression devices and pharmacologic agentsalone or in combination. The lowest incidence of DVTwas in the group that received both a mechanical com-pression device and a heparin agent.
You asked whether the RAP score really improvespatient selection for individuals at high risk for the devel-opment of DVT in contrast to injury patterns. Sixteen riskfactors allow individualization of the patient’s specific risk
638 Gearhart et al SurgeryOctober 2000
factors. Although the incidence of DVT and pulmonaryembolism that we are reporting may be identical to resultsfrom other larger studies, I cannot differentiate the subtledifferences between the 2 methods of identifying patientsat risk for the development of DVT. An additional benefitfrom the use of the RAP score is that each variable (riskfactor) is weighted and can be further analyzed to helpdefine the significance of each risk factor.
Your next question regards the natural history of anasymptomatic DVT. When a venous thromboemboliccomplication occurs in a trauma patient, it is typicallyclassified as a preventable event. Thus, it behooves us asthe treating physician to minimize the rate of this com-plication in our patients. An example of the worst-casescenario is a young patient who sustains a moderate headinjury and a femur fracture in a motor vehicle crash. Thepatient is bedridden and unresponsive for 1 to 2 weeks,and an asymptomatic DVT that is not diagnosed devel-ops. Several months later, the patient has recovered andis ambulatory only to be plagued by the postphlebitic syn-drome for the rest of his/her life because surveillancescans were not performed. I think you would agree thattimely diagnosis (surveillance scans) and appropriatetreatment should reduce the disability from the post-phlebitic syndrome, if not eliminate it all together.
Is routine screening for an asymptomatic DVT costeffective? The real reason for using the RAP score, whichallows further refinement of specific risk factors for thedevelopment of VTE, is because elimination of prophy-laxis and routine duplex scanning in low-risk patients isa cost-effective practice. Indeed, we are not performingprophylaxis in all trauma patients; our protocol limitsprophylaxis to only high-risk patients (RAP score, ≥5).These are the same patients who receive serial screeningduplex scans. As we note in the article, avoiding prophy-lactic measures and routine scanning in low-risk patientsled to a significant reduction in hospital charges.Furthermore, we participated in a second study with DrGreenfield and Mary Proctor, looking at DVT prophylax-is in the high-risk population. In that paper, a cost analy-sis was performed for DVT in low-risk and high-riskpatients. Prophylaxis and routine screening duplex scan-ning was beneficial in reducing overall hospital costs forvenous thromboembolic complication in only the high-risk patients.
Your last question is in reference to the first study byGeerts et al7 and the 1% incidence of pulmonaryembolism that they observed in patients who were notreceiving DVT prophylaxis. You are correct that 3 of the30 deaths were attributed to saddle embolism, butremember that none of these patients had any symptomsof venous thromboembolic complication until their firstsymptom, sudden cardiopulmonary arrest as the result ofa saddle embolism. These 3 deaths were clearly pre-ventable deaths. Ideally, we would like to eventually elim-
inate or at least minimize the risk of this complication forour patients.
Dr Miller, we had 3 patients who received a femoralline during resuscitation. Only 1 patient was eventuallydiagnosed with a DVT. We did not specifically look at thefindings of the Doppler scan and correlate them with theRAP score to identify patients who might benefit from aprophylactic Greenfield filter. However, 9 patients didreceive a prophylactic filter. Only 1 patient experiencedthe development of a DVT at the insertion site.
Drs Howard and Turnipseed asked about the sensitiv-ity and specificity of duplex scanning in our vascular lab-oratory compared to some other gold standard such as avenography or helical computed tomography scan andhow this difference in sensitivity and specificity mayimpact our data. Several years ago, Dr Greenfield pro-posed a multiinstitutional study for DVT prophylaxis tovalidate the work of Geerts et al. The only differencebetween the 2 studies was the use of duplex scanning byDr Greenfield, rather than ascending venography, as thediagnostic tool for DVT. This was because most of thepotential study centers stated that duplex scanning hadessentially replaced venography for diagnosing DVT intheir hospitals. Our vascular laboratory, which isapproved by the Inter-Societal Commission forAccreditation of Vascular Laboratories, only employs reg-istered vascular technicians and registry-eligible techni-cians. In this particular study, all scans were performedby individuals with at least 2 years of experience. Thedirector of our vascular laboratory (a vascular surgeonwith added qualifications) states the laboratory’s accura-cy for diagnosing lower extremity DVT is greater than95%. A detailed description of the technique used forperforming the lower extremity duplex scans is includedin the article. For these reasons, we did not perform anyconfirmatory studies with dynamic helical computedtomography scans or ascending venography. I am confi-dent that our vascular laboratory’s use of duplex scans ishighly sensitive and specific for diagnosing a DVT.
Dr Harms asked about the patients with pelvic frac-tures potentially being stratified to low risk. We all knowthat there is definite differences in pelvic fractures andthat they are not created equal. However, the individualwith an isolated pelvic fracture who requires immobiliza-tion is rather unusual. Most isolated pelvic fractures areweight-bearing on 1 extremity and the patient can ambu-late with a walker or crutches. In contrast, more complexpelvic fracture patterns that require operative stabiliza-tion are rarely isolated injuries and would likely have asecond risk factor, which would then place the patient athigh risk and prophylaxis would be administered.
Dr Malangoni, your first question dealt with the low-risk group, which included only 58 patients, and whetherwe were comfortable not giving these patients any formof prophylaxis. This is the second study that validates the
Surgery Gearhart et al 639Volume 128, Number 4
incidence of DVT being so low in the low-risk patientthat prophylaxis is not cost effective. Thus, we do not useprophylaxis in any of the low-risk patients.
Your second question asked if we are going to revisethe RAP score based on what we have now found to besignificant predictive parameters to make it easier to use
or perhaps even more predictive for the identification ofhigh-risk patients. I think any revisions to the originalRAP score would be premature. This is a pilot study andnot intended to be definitive regarding each specific riskfactor. In fact, several of the risk factors were not evenidentified in any of the 160 patients in this study
![[Intestinal ischaemia caused by acute mesenteric vein thrombosis]](https://static.documents.pub/doc/80x56/635a2bbf9d85dc43cb06d7e9/intestinal-ischaemia-caused-by-acute-mesenteric-vein-thrombosis.jpg)
