Injury, Int. J. Care Injured 33 (2002) 765–769

Limb loss following lower extremity arterial trauma:what can be done proactively?

Alejandro Guerreroa, Kathleen Gibsona, Kurt A. Kralovichb, Iraklis Pipinosb,Petros Agnostopolousb, Yvonne Cartera, Eileen Bulgera,

Mark Meissnera, Riyad Karmy-Jonesa,∗a Department of Surgery, Harborview Medical Centre, 352 Ninth Avenue, Box 359796, Seattle, WA 98104, USA

b Henry Ford Hospital, Detroit, MI, USA

Accepted 3 September 2001

Abstract

We performed a retrospective review of patients admitted to two Level I trauma centres over a 15-year period with arterial injuries(excluding primary amputations). Preoperative factors analysed included mechanism of injury, site and type of arterial and venous injuryand repair, time to operating room, initial blood pressure, evidence of ipsilateral limb fracture and/or extensive tissue damage, status ofpreoperative pulses and angiographic data. One hundred and fifty-one arterial injuries were treated (80 penetrating). Overall mortality was10 (6.6%) and limb loss 16 (10.6%). Only two factors that might possibly be modified by specific interventions were noted. The incidenceof limb loss was higher in patients who developed compartment syndrome (41% versus 7% without,P = 0.003) and in those who didnot receive intra- or immediately postoperative anticoagulation (15% without versus 3% with,P = 0.02). Unfortunately, no factor wasfound that reliably predicted the risk of compartment syndrome. In addition, patients who did not receive peri-operative anticoagulationwere more severely injured than those that did were. Despite this, there were no bleeding complications associated with anticoagulation.These findings suggest that the primary interventions that may improve limb salvage include liberal use of fasciotomy (recognising thatany patient may require this) as well as early use of anticoagulation.© 2002 Published by Elsevier Science Ltd.

Keywords: Limb loss; Anticoagulation; Arterial injury

1. Introduction

Limb loss following lower extremity arterial injury hasbeen variously ascribed to extent of tissue damage, durationof ischaemia prior to revascularization, associated venousinjuries, popliteal artery involvement and/or developmentof compartment syndrome[1–7]. We reviewed our experi-ence at two trauma centres to determine the risk factors forlimb loss, with the goal of identifying what factors mightbe preventable.

2. Methods

We performed a retrospective review of patients admittedto Harborview Medical Centre and Henry Ford Hospital witharterial injuries from the common iliac to popliteal arteries

∗ Corresponding author. Tel.:+1-206-731-2857; fax:+1-206-731-3656.E-mail address: karmy@u.washington.edu (R. Karmy-Jones).

between 1 January 1985 and 1 January 1999. Only, vesselsthat underwent operative intervention were included.

The diagnosis of compartment syndrome was made bydocumentation of tense calf compartments demonstrated byswelling at the time of fasciotomy. Extensive tissue damageincluded: massive loss of soft tissue, gross contamination,major muscle injury and/or avulsion. This correlated witha soft tissue/skeletal score of≥3 according to the MangledExtremity Score (MESS) described by Johanson et al.[7].Factors analysed for each outcome are listed inAppendix A.

Univariate analysis was performed using logistic regres-sion. Factors identified to as being significant at theP ≤0.1 level were then evaluated with a stepwise multivariatelogistic regression analysis with backward elimination.

3. Results

Over the study period, 151 patients underwent revascu-larization or ligation for arterial injury. Presenting charac-

0020-1383/02/$ – see front matter © 2002 Published by Elsevier Science Ltd.PII: S0020-1383(01)00175-9

766 A. Guerrero et al. / Injury, Int. J. Care Injured 33 (2002) 765–769

Table 1Characteristics of study populationa

DemographicsAge (years) 29.0± 12.86Male (gender) 125 (82.8)

MechanismGunshot 69 (45.7)Stab wound 10 (6.6)Impalement 1 (0.7)Motor vehicle crash 18 (11.9)Fall 8 (5.3)Pedestrian struck 22 (14.6)Motor cycle crash 12 (7.9)Explosion 1 (0.7)Blunt assault 10 (6.6)

Clinical findingsPulse deficit on presentation 104 (68.9)Hematoma 20 (13.2)Active bleeding 23 (15.2)Neurologic deficit 24 (15.9)Extensive tissue damage 14 (9.3)Ankle brachial index 0.5± 0.29

Associated injuriesIpsilateral fracture 78 (51.5)Abdominal injury 34 (22.6)Pelvic fracture 19 (12.6)Closed head injury 8 (5.3)Injury severity score 15.2± 9.5Presenting systolic blood pressure 111.8± 35.3Time (from injury to operation) (h) 9.5± 21.8Time (from admission to operation) (h) 4.0± 10.1

a Values within parentheses in percent.

teristics and sites of arterial injuries are listed inTables 1and 2. Angiograms were obtained in 56 (37.1%) prior tooperation and 27 (17.9%) in the operating suite. Findingsincluded flap in six (10.8%), occlusion in 61 (73.5%),

Table 3Comparison of level of injury with repair

Artery Primary repair(N = 43)

Syntheticgraft (N = 5)

Reversed saphenousvein (N = 93)

Vein patch(N = 5)

Extra-anatomicbypass (N = 2)

Ligation(N = 3)

Common iliac 5 1 1External iliac 3 2 1Internal iliac 1 1Common femoral 4 4Common femoral bifurcation 1Superficial femoral 26 2 43 3 1 1Popliteal 4 1 43 2 1

Table 4Correlation of method of intervention and outcomea

Procedure (N) Limb loss (%) PSA (%) Graft infection (%) Occlusion (%) Ipsilateral DVT (%)

Primary repair (43) 2 (5) 2 (5) 1 (3) 1 (3) 1 (3)Artificial graft (5) 2 (40) 0 1 (20) 1 (20) 0Saphenous vein (93) 11 (12) 2 (2) 3 (3) 4 (4) 15 (16)Vein patch (5) 0 0 0 0 0Extra-anatomic bypass with graft (2) 1 (50) 0 0 0 0Ligation (3) 0 0 0 0 0

a PSA: pseudoaneurysm; DVT: deep venous thrombosis; values within parentheses are percentage of particular operative intervention.

Table 2Associated venous injury and limb loss according to arterial levela

Artery injured N Vein(N) (%)

Limb loss(N) (%)

Common iliac 7 5 (71) 1 (14)External iliac 6 4 (67) 2 (33)Internal iliac 2 2 (100) 0Common femoral 8 2 (25) 0Common femoral bifurcation 1 0 1 (100)Superficial femoral 76 21 (28) 3 (4)Popliteal 51 8 (16) 9 (18)

a Vein: associated venous injury; values within parentheses are per-centage of associated arterial injury.

pseudoaneurysm in five (6%), fistula in six (7.2%) andactive extravasation in two (2.4%). Correlation betweenartery injured and method of repair is listed inTable 3. Ofnote, shunts were used in 10 (6.6%), thrombectomy in 5(3.3%) and fasciotomy performed at the initial surgery in 51(33.8%) of cases. Two cases of wound infection occurred inthe fasciotomy group. Association between method of revas-cularization and outcome are listed inTable 4. In 40 casesassociated venous injuries were discovered. These weremanaged by ligation in 23, repair in 12 and vein graft in fiveinstances.

The overall mortality was 10 (6.6%). Death occurredintraoperatively in two cases due to unremitting shock andpostoperatively in eight secondary to multiple organ fail-ure. Postoperative compartment syndrome occurred in 17(11.3%) cases. There were four cases (2.6%) of wound in-fection, five (3.3%) cases of graft infection and four (2.6%)cases of postoperative pseudoaneurysm formation. Reop-eration for occlusion was required in six (3.9%) of cases(Table 4). A further two patients had evidence of decreasedflow across the graft by Doppler ultrasound, but have been

A. Guerrero et al. / Injury, Int. J. Care Injured 33 (2002) 765–769 767

Table 5Correlation between anticoagulation and limb loss

Perioperative anticoagulation Limb loss (%) Postoperative compartmentsyndrome (%)

Given intraoperatively or in recovery room (N = 60) 2 (3) 9 (15)Given later than recovery room (N = 31) 5 (16) 5 (16)Never given (N = 60) 9 (15) 3 (5)

followed without need for further intervention. Ipsilateraldeep venous thrombosis was documented in 16 (10.6%)(Table 5).

The overall incidence of limb loss was 16 (10.6%).Patients with severe soft tissue injury (mangled extremity)had a markedly higher incidence of limb loss (10/14 or71%) than those without (6/137 or 4%,P < 0.001). Graftinfection occurred in five patients, four of whom lost limbs.There were no significant differences in the time fromaccident to operation (13.4 ± 7.2 h with limb loss versus9.0± 21.8 h without limb loss,P = 0.5). Factors identifiedas being significant toP < 0.1 were entered into backwardlogistic regression analysis (Table 6). Because mangledextremity and injury severity score (ISS) played such asignificant role in determining limb loss, these factors werenot included in the regression analysis[8]. Of note, theretrospectively calculated MESS score of patients who lostlimbs (6.26±1.88) was significantly greater than those whodid not (3.81 ± 2.21, P = 0.001). Injuries involving thepopliteal artery were associated with an eight-fold increasedrisk of limb loss compared to other vessels (Table 6).

Seven of 17 (41%) patients who developed compartmentsyndrome postoperatively lost the injured limb as opposed to9 of 134 (6.7%) who did not develop postoperative compart-ment syndrome (P < 0.001). Four of 51 patients (8%) whounderwent fasciotomy at the initial operation subsequentlydeveloped compartment syndrome as compared to 13 outof 100 (13%) who did not undergo fasciotomy at the initialoperation (P = 0.35). Of the four patients who developedcompartment syndrome despite fasciotomy, three occurred

Table 6Analysis of factors associated with limb lossa

Factor Univariate analysis Multivariate analysis

OR (95% CI) P-value OR (95% CI) P-value

ISS (per increase of five units) 1.31 (0.98–1.76) 0.02Mechanism (blunt vs. penetrating) 2.7 (1.5–4.7) 0.08Systolic blood pressure (per decrease of 30 mmHg) 1.9 (1.4–2.7) 0.02Pseudoaneurysm 9.5 (3.4–26.8) 0.03Graft infection 44.3 (13.9–141.2) 0.001Postoperative compartment syndrome 14.0 (7.8–25.0) <0.001 81.5 (24.5–270.4) 0.003Peri-operative anticoagulation 1.9 (1.4–2.7) 0.03 5.0 (2.5–9.8) 0.02Mangled extremity 54.6 (25.3–112.1) <0.001Ipsilateral fracture 9.0 (4.2–19.5) 0.004 20.1 (5.5–73.8) 0.02Politeal artery injury 2.9 (1.7–4.9) 0.05 8.2 (3.3–20.1) 0.02Pelvic fracture 3.9 (2.1–7.2) 0.03

a OR: odds ratio; ISS: injury severity score; CI: confidence interval; multivariate model created using stepwise logistic regression with backwardselimination: variable withP-values<0.1 were retained in the model. Mangled extremity and ISS were not included in the model.

after popliteal artery injury. These were limited fascio-tomies.

Perioperative anticoagulation (started in operating roomor postoperative recovery room) was used in 60 cases (as-pirin in five, subcutaneous heparin in eight, intravenousheparin in 31, low molecular weight heparin in three andintravenous dextran in 13). The incidence of limb loss in pa-tients who were started on perioperative anticoagulation (2or 3.3%) was significantly lower than if anticoagulation wasstarted later (14/91 or 15%,P = 0.03). There were no bleed-ing complications associated with anticoagulation. Of note,however, patients who were placed on perioperative antico-agulation had significantly lower ISS (12.1±6.8) comparedto those started later or never (17.3±10.5, P = 0.001). Theprimary stated reason in all cases for not starting anticoagu-lation in the operating suite or immediately postoperativelywas concern of bleeding from associated injuries.

A separate analysis of postoperative compartment syn-drome was performed using the same data points listed inAppendix A. The only factors that appeared significantlyassociated with postoperative compartment syndrome at aP ≤ 0.1 level were preoperative pulse deficit (15/104, 14%versus 2/47, 4% without deficit,P = 0.08), and develop-ment of post operative graft infection (2/5, 40%,P = 0.09)or pseudoaneurysm (2/4, 50%,P = 0.07). These wereentered into backward regression analysis and only docu-mentation of a preoperative pulse deficit approached signif-icance (P = 0.06). While the sensitivity of this finding forpredicting postoperative compartment syndrome was 88%,the specificity was only 34% and the overall accuracy 40%.

768 A. Guerrero et al. / Injury, Int. J. Care Injured 33 (2002) 765–769

4. Discussion

A number of factors have been associated with the riskof limb loss following lower extremity trauma. The majorfactor is the degree of associated tissue destruction, greaterfollowing blunt mechanisms[1,7]. Popliteal artery injuries,particularly when combined with venous injuries, are as-sociated with the highest rate of limb loss[4,5,9,10]. Themost common options for reconstruction include primaryrepair, use of artificial material, or venous bypass[11–15].A concern regarding artificial graft material has been thepotential risk of infection, however, there are large reviewsthat suggest that infection is not a major risk with the useof artificial grafts, as long as adequate coverage is possible[12]. In fact, the centres that advocate the use of artificialmaterial over vein feel that when infection does occur, it in-volves only the anastomoses rather than the entire graft andis thus less life threatening[12,16]. Furthermore, there isdata to suggest that artificial material may be useful as a tem-porising graft when complex venous injuries are discovered[12].

The early (intraoperative or immediately postoperative)use of anticoagulation (in particular heparin) has been linkedto a reduction in limb loss, presumably due to greater graftpatency[2]. The risk of bleeding has been argued to be min-imal. However, the connection between graft patency andlimb loss has not been clearly defined. In addition, still thereis a tendency to withhold anticoagulation in severely injuredpatients who present a risk of bleeding at sites other thanthe extremity injury. The experience derived from electivearterial reconstruction/bypass suggests that some anticoag-ulation is beneficial, but the data specifically documentingwhich form is the “best” is lacking[17].

A major risk factor for limb loss is the developmentof compartment syndrome[3,5,18,19]. Compartment syn-drome has itself been linked to delay in restoration of bloodflow; presence of associated venous injuries and lower ex-tremity fractures. There have been studies which suggestthat prophylactic fasciotomy should be performed in cer-tain circumstances, including delay of revascularization ofgreater than 6 h and/or in the presence of combined arterialand venous injury[3]. Fasciotomy is associated with a smallrisk of wound infection, which may, in association withextensive soft tissue injury, increase the risk of amputationbeing required.

Our data again demonstrated the increased risk of limbloss in association with popliteal artery injury. We did notidentify any significant differences in outcome betweenchoice of repair or graft conduit. There was a suggestionthat the early use of anticoagulation was associated withreduced limb loss. However, perioperative anticoagulationwas started in patients who were less severely injured. Fur-thermore, our data did not identify any one agent as beingspecifically useful. We currently favour intravenous hep-arin, an agent that is relatively easy to titrate in the acutesetting.

We were not able to identify any features that predictedthe occurrence of compartment syndrome. The significanceof this is that any patient can develop this complication, andall patients require careful monitoring following operativerepair of arterial injuries. In particular, four patients who diddevelop compartment syndrome had undergone fasciotomyat the time of the original surgery. This underlines the im-portance of performing a complete fasciotomy, as well asrecognising that compartment syndrome can still occur afterwhat appears to be an adequate decompression.

5. Conclusion

Patients who suffer lower extremity arterial trauma shouldbe started on anticoagulation as soon as possible, preferablyintraoperatively. Compartment syndrome may affect any pa-tient, is associated with a significant risk of limb loss, andearly fasciotomy is warranted if there is any question of itsoccurrence.

Appendix A

Factors studied by statistical analysis to determine anyrelation to limb loss or compartment syndrome

PreoperativeAbdominal injury requiring laparotomyAngiographic findingsAgeClinical findings

HematomaActive bleedingPulse deficit

Sensory or motor deficitClosed head injuryGenderInitial ankle-brachial indexInjury severity scoreIpsilateral limb fractureLevel of arterial injuryMassive tissue destructionMechanismPelvic fractureSystolic blood pressure on presentationSystolic blood pressure on arrival in ORTime from injury to operationTime from admission to operationVenous injury

Operative and postoperative dataGraft infectionGraft failureIntraoperative fasciotomyOperative intervention (arterial and venous)

A. Guerrero et al. / Injury, Int. J. Care Injured 33 (2002) 765–769 769

Appendix A (Continued)

Use of shuntUse of thrombolyticsTiming of anticoagulationChoice of anticoagulationPostoperative DVTWound infection

References

[1] Shah PM, Ivatury RR, Babu SC, Nallathambi MN, Clauss RH, StahlWM. Is limb loss avoidable in civilian vascular injuries? Am J Surg1987;154(2):202–5.

[2] Melton SM, Croce MA, Patton JH, Jr, Pritchard FE, Minard G,Kudsk KA, et al. Popliteal artery trauma. Systemic anticoagulationand intraoperative thrombolysis improves limb salvage. Ann Surg1997;225(5):518–27(discussion 527–9).

[3] Fields C, Senkowsky J, Hollier L, Kvamme P, Saroyan R, Rice J,et al. Fasciotomy in vascular trauma: is it too much, too often? AmSurg 1994;60(6):409–11.

[4] Andrikopoulos V, Antoniou I, Panoussis P. Arterial injuries associatedwith lower-extremity fractures. Cardiovasc Surg 1995;3(1):15–8.

[5] Fainzilber G, Roy-Shapira A, Wall MJ, Jr, Mattox KL.Predictors of amputation for popliteal artery injuries. Am J Surg1995;170(6):568–70 (discussion 570–1).

[6] Faris IB, Raptis S, Fitridge R. Arterial injury in the lower limb fromblunt trauma. Aust NZ J Surg 1997;67(1):25–30.

[7] Johansen K, Daines M, Howey T, Helfet D, Hansen ST, Jr. Objectivecriteria accurately predict amputation following lower extremitytrauma. J Trauma 1990;30(5):568–72 (discussion 572–3).

[8] Baker S, O’Neill B, Haddon W, Long Jr W. The injuryseverity score: a method for describing patients with multipleinjuries and evaluating emergency care. J Trauma 1974;14(3):187–96.

[9] Fabian TC, Turkleson ML, Connelly TL, Stone HH. Injury to thepopliteal artery. Am J Surg 1982;143(2):225–8.

[10] Peck JJ, Eastman AB, Bergan JJ, Sedwitz MM, Hoyt DB,McReynolds DG. Popliteal vascular trauma. A communityexperience. Arch Surg 1990;125(10):1339–43 (discussion 1343–4).

[11] Feliciano D, Herskowitz K, O’Gorman R, Cruse R, Cruse P, BrandtM, et al. Management of vascular injuries in the lower extremities.J Trauma 1988;28:319–28.

[12] Feliciano DV, Mattox KL, Graham JM, Bitondo CG. Five-yearexperience with PTFE grafts in vascular wounds. J Trauma1985;25(1):71–82.

[13] Sriussadaporn S. Arterial injuries of the lower extremity from blunttrauma. J Med Assoc Thai 1997;80(2):121–9.

[14] Waikakul S, Sakkarnkosol S, Vanadurongwan V. Vascular injuries incompound fractures of the leg with initially adequate circulation. JBone Joint Surg Br 1998;80(2):254–8.

[15] Burch JM, Richardson RJ, Martin RR, Mattox KL. Penetrating iliacvascular injuries: recent experience with 233 consecutive patients. JTrauma 1990;30(12):1450–9.

[16] Bricker D, Beall Jr A, DeBakey M. The differential response toinfection of autogenous vein versus Dacron arterial prosthesis. Chest1970;58:566–70.

[17] Watson H, Belcher G, Horrocks M. Adjuvant medical therapy inperipheral bypass surgery. Br J Surg 1999;86:981–91.

[18] Patman RD, Thompson JE. Fasciotomy in peripheral vascular surgery.Report of 164 patients. Arch Surg 1970;101(6):663–72.

[19] Patman RD. Compartmental syndromes in peripheral vascularsurgery, Clin Orthop 1975;(113):103–10.