Arteriography for Proximity of Injury in Penetrating Extremity Trauma1 John A. Kaufman, M D James E. Parker, M D David L. Gillespie, M D Alan J. Greenfield, MD2 Jonathan Woodson, MD James 0.Menzoian, MD
Index terms: Arteries, injuries, 91.41, 92.41 Extremities, angiOgaphy, 91.122, 92.122 Extremities, injuries, 91.41, 92.41 Trauma, 91.41,92.41
JVIR 1992; 3:719-723 Abbreviations: AVF = arteriovenous fistula, DSA = digital subtraction angiography
Arteriography for proximity of injury was studied prospectively at a trauma center. Findings in 85 patients with penetrating extremity wounds were analyzed to determine the prevalence and types of vascular abnormalities seen with these injuries. Ninety-two limb segments were studied for 77 gunshot and 15 stab wounds. Arteriographic findings were positive in 24% overall but in only 5% for injuries confined to major vessels. A 60% positive rate was seen in a small subgroup of 10 patients with fractures due to gunshot wounds. The most frequently injured vessels were muscular branches of the deep femoral artery (59%); the most common injury was focal, nonocclusive spasm (42%). All patients were treated conservatively, without sequelae at follow-up.In this study, the vascular injuries found at arteriography for proximity of injury in penetrating trauma due to bullets or knives, particularly in the thigh, did not require surgical or radiologic intervention.
A R T E R l o G R M H Y performed
1 From the Departments of Vascularilnterventional Radiology (J.A.K., J.E.P., A.J.G.) and Vascular Surgery (D.L.G.,J.w., J.O.M.), Boston University MedicalCenter2 Boston. Received March 12, 1992; revision requested ~ ~ ~ i lrevision 1 7 ; received June 5; accepted June 12. Address reprint requests to J.A.K., Section of Vascular Radi-
~''gy, Massachusetts Hospital, 32 Fruit St, Boston, MA 02114. Current address: ~ ~ ventional Radiology, T U ~ ~New S England Medical Center, Boston. SCVIR, 1992
to exclude vascular injury when extremity trauma is in proximity to a major artery but physical examination is negative has had a changing role in the management of the trauma patient (1-4). In the middle 1980s, Menzoian et a1 recommended the routine use of proximity arteriography after finding a 16%yield of occult vascular injuries (5,6).Subsequent studies in both the radiology and trauma literature have supported restricting the use of proximity arteriography, due in part to the adoption of a more conservative surgical approach toward minor vascular injuries (7-9). Meticulous evaluation of the patient by means of physical examination for "hard" signs of vascular injury has been a crucial element in these series (7-9). The vascular surgery and radiology services at our hospital recently conducted a reevaluation of routine proximity arteriography for all mechanisms of injury. ~ We reviewed d the angiographic findings in the subset of patients with penetrating trauma due to gunshot and stab wounds, all of
whom were treated with conservative management.
PATIENTS AND METHODS Review of the protocol by the institutional Human Studies Committee was not required because randomization was not performed and patients were treated according to current practice at the time of the study. Patients with trauma to the extremities were entered prospectively into the study over a 14-month period starting July 1989. Upon presentation to the Boston City Hospital, patients were evaluated by senior surgical residents for the following inclusion criteria: ( a )trauma within 1cm of the expected course of a major artery and ( b )absence of "hard" signs of vascular injury-bruit, loss of pulses, limb ischemia, active bleeding, expanding hematoma, or history of pulsatile bleeding. ~ ~ ~ This analysis was limited to the subset of patients with penetrating injuries due to bullets or sharp ob-
719
~
720
Journal of Vascular and Interventional Radiology
--
November 1992
jects. Fractures and neurologic deficits due to penetrating trauma were noted but were not considered a basis for exclusion. Patients with blunt trauma or dislocations, or patients who had undergone surgery on the affected extremity prior to angiography were not included. Patients with multiple pellet injuries were also excluded because only one of the three patients met the conditions outlined above. Eighty-five patients met the criteria for this review: 82 men (age range, 17-59 years; mean, 26 years), and three women (age range, 20-32 years; mean, 27 years). All 85 patients underwent angiography within 24-48 hours of presentation in the radiology department after informed consent was obtained. Femoral arterial catheterization was accomplished by using standard retrograde angiographic technique and a 5-F straight catheter with multiple side holes, a 5.5-F C-2 cobra end-hole catheter, or a 5.5-F H-1 catheter (Cook. Bloomington, Ind). ~ e ~ e n don' i nthe ~ location of the injury, contrast material was injected into the ipsilateral external iliac arterv. " contralateral external iliac artery, or subclavian artery. Iohex01 (350 mg of iodine per milliliter) (Omnipaque 350; Sanofi Winthrop, New York) was used for cut-film angiography; iothalamate meglumine (202 mg of iodine per milliliter) (Conray 43; Mallinckrodt, St Louis) was used for occasional supplemental digital subtraction angiography (DSA). At least two views were obtained in all studies. Although one study was performed entirely with digital imaging, we considered that the quality of the images obtained on our equipment (pixel matrix, 256 x 256) precluded the routine use of this modality. A comparison of cut-film and DSA images was not performed. Limb segments were defined as thigh (inguinal ligament to the femoral condyles), knee (the popliteal fossa), and lower leg (proximal head of the fibula to the ankle) for the lower extremity, and upper arm (anterior axillary line to the humeral condyles) and forearm (humeral conr
dyles to the ulnar styloid) for the upper extremity. Angiograms were considered positive for injury if any of the following were seen in any opacified arterial structure: arteriovenous fistula (AVF),extravasation, intimal flap, laceration, mural hematoma, occlusion, pseudoaneurysm, spasm, or transection. Mass effect (deviation of the course of an otherwise normal artery), parenchymal stains, and standing waves were not considered positive findings. Major vessels were considered to be arteries that if occluded acutely would result in an ischemic, nonviable extremity. All other vessels were termed minor. After arteriography, patients were monitored with serial physical examinations and measurements of anklebrachial indexes on the surgical service for a minimum of 24 hours. Surgical clinic appointments were scheduled on discharge. If patients remained hospitalized for other reasons, their vascular examinations were monitored by the surgical service throughout the duration of the admission. Follow-up ranged from 24 hours to 178 days. Only clinical follow-up was considered adequate. The constraints of practice in a busy trauma center did not permit either blinding of the investigators or randomization to a nonangiographic treatment group.
RESULTS A total of 92 limb segments were evaluated with arteriography for 77 (84%)gunshot wounds and 15 (16%) stab wounds. The distribution of injuries to the extremities was 59 thigh (64%), 15 upper arm (16%), 13 knee (14%),four lower leg (4%),and one forearm (1%).There were 70 (76%) negative limb segments in 64 patients (six patients underwent study of two limb segments) and 22 (24%) positive limb segments in 21 patients (one patient had two positive segments). The mechanisms of injury in the 22 positive studies were gunshot wounds in 19 (25% of all gunshot
Figure 1. Transection and occlusion of the descending deep femoral artery trunk due to a gunshot injury (arrow). Spasm and a small filling defect are present in the adjacent muscular branch. The patient refused follow-up angiography.
wounds, 86% of all positive limb segments) and stab wounds in three (20% of all stab wounds, 14% of all positive limb segments). The majority of the injuries involved unnamed branches of the deep femoral artery (59%, 1 3 of 22 studies), followed in frequency by the superficial femoral artery (18%, four of 221, the deep femoral artery (9%, two of 22), the popliteal artery (5%, one of 221, a n unnamed muscular branch of the circumflex humeral artery (5%,one of 221, and a peroneal artery (5%, one of 22). Of the 22 positive extremity segments, 18 (82%)of the injured vessels were in the thigh. A total of 24 vascular abnormalities were identified in 22 vessels. These included 10 areas of focal nonocclusive spasm (42%),six vascular occlusions (25%),two areas of minimal extravasation (8%),two intimal flaps (8%),two AVFs (8%),one small mural hematoma (4%),and one transection (4%).Two occlusions occurred in conjunctiorl with other abnormalities: an intimal flap in one case and a transection in another. The majority of the injuries, 15 of 24 (63%),occurred in unnamed muscular branches, including both AVFs.
Kaufman et a1
72 1
Volume 3 Number 4
c. a. b. Figure 2. AVF of a muscular branch of the deep femoral artery due to multiple stab wounds, with spontaneous closure at 36 days. (a)Arrow identifies early filling of a profunda vein on the original proximity angiogram. Standing waves are present in the superficial femoral artery. (b)Dense venous opacification is seen later in the same run. (c) Arrow identifies former site of AVF, which
closed spontaneously, on follow-up angiogram at 36 days.
Among named arteries, focal nonobstructing spasm was the most common abnormality, identified in three superficial femoral arteries and in the only abnormal popliteal artery. The small mural hematoma occurred in a superficial femoral artery, a small intimal flap was found in a deep femoral artery origin, a short occlusion was noted in a peroneal artery, and a descending deep femoral artery trunk was transected and occluded distal to the origin of the first unnamed muscular branch below the lateral circumflex femoral artery (Fig 1). Neurologic deficits occurred in five patients (6%), most commonly radial nerve palsies (three of five). These were found in association with one stab wound, one simple gunshot wound, and one gunshot wound with a humeral fracture. The remaining two deficits were trivial. consisting of focal hypoesthesia over t h e laterar thigh and the dorsum of the foot. All five of these patients had negative arteriograms. Within the gunshot wound group, 10 patients had fractures due to bullet impact on bone: femur in six, humerus in two, and fibulae in two. All fractures were comminuted, closed,
and minimally displaced. Six patients (60%)had positive studies involving seven vessels: muscular branches of the deep femoral artery in four, superficial femoral artery in one, peroneal artery in one, and a muscular branch of the circumflex humeral artery in one. Focal spasm was present in a muscular branch of the deep femoral artery, the circumflex humeral artery, and the superficial femoral artery. In three cases occlusions were noted, involving two muscular branches of the deep femoral artery and a tibia1 runoff vessel (peroneal artery). One intimal flap was found in conjunction with an occlusion of a muscular branch of the deep femoral artery. One angiographic complication occurred-a small groin hematoma a t the arterial puncture site. This was noted on return of the patient to the ward after angiography. The hematoma remained stable in size, requiring no intervention. The overall rate of procedure-related complications was therefore 1%. No patients developed vascular symptoms within the immediate postangiographic period. Follow-up longer than 24 hours was available in
28 of 64 patients (44%)with negative examinations a t a mean of 2 1days after arteriography (range, 5-107 days). Among the patients with positive angiograms, 16 of 21 (76%)had a mean duration of follow-up of 20 days (range, 3-178 days). The group of ~ a t i e n t with s mnshot wounds and fractures had a longer follow-up, with a mean of 33 days (range, 5-178 days). Follow-up consisted of visits in surgical or orthopedic clinics, although some patients were seen during prolonged hospitalizations for treatment of nonvascular injuries or a t subsequent emergency room admissions for unrelated trauma. One patient, with an AVF of a muscular branch of the deep femoral artery, underwent repeat angiography at 36 days demonstrating spontaneous closure (Fig 2). The other patient with an AVF was lost to follow-UD.None of the patients who were seen in follow-up developed symptoms or evidence of vascular injury. u
DISCUSSION The history of the treatment of proximity injuries has been well iter-
722
Journal of Vascular and Interventional Radiology -
-- -
November 1992
ated in the literature (10-12). In patients without evidence of vascular trauma on physical examination, manv of the vascular iniuries identified at angiography a r e k i n o r (9,101. Recently, a general trend toward the expectant surgical management of minor vascular injuries has led to a reconsideration of the role of proximity angiography. Although this conservative approach is not universally accepted (13-171, it is amply supported in the literature (18-24). Penetrating extremity trauma due to gunshot and stab wounds is common at our institution. We were interested in identifying the prevalence, types, and locations of vascular abnormalities found in patients whose angiograms were obtained solely because of the proximity of injury. This information is useful in determining the necessity and priority of peripheral vascular imaging studies in these individuals, many of whom have additional, more pressing injuries. Furthermore, a knowledge of the expected angiographic findings in this setting may be helpful when proximity studies are requested at centers where these injuries are less common. Our greatest angiographic experience was with gunshot wounds77 in all. However. we were not able to segregate low-velocity from highvelocity bullet wounds. The second large grouping was 15 stab wounds, all of which were due to knives. In this analysis, we excluded shotgun injuries, as we did not believe that our very limited numbers would be meaningful. The reported rates of positive findings for this mechanism at proximity angiography are higher than those for gunshot and stab wounds (10,22). The results described in this article should not be extrapolated to proximity arteriography performed for dislocations, fractures due to blunt trauma, or injuries caused by multiple pellets. The thigh was the most commonly injured limb segment in our study and also had the highest overall rate of vascular injury (32%).This has not been true for other authors. Dennis
et a1 found the highest rate in upper arm injuries, 15.7% (10);Francis et al, in the forearm, 6.3% (22); and Anderson et al, in the calf, 22.9% (21). The explanation for this variation is not clear but may be a function of study size rather than of differences in patient populations. Our positive rate of 24% is a t the upper end of the range reported in the literature (11). This a reflection of our definition of a "normal" proximity angiogram as demonstrating no evidence of a vascular injury in any arterial structure. It is important to note that the majority of our angiographic findings, 19 of 24 (79%),were in named muscular arteries or unnamed terminal muscular branches. These injuries occurred in vessels that were not essential for limb viability. In addition, they are difficult to approach surgically. The abnormalities were considered too small to warrant intervention by either the vascular radiologists or the vascular surgeons. Specifically, these findings included six of nine areas of spasm, six of six occlusions, both cases of minimal extravasation, both AVFs, both intimal flaps, and the single transection. Had we excluded injuries to minor vessels from our analysis, as do some authors (22,24,25),our positive rate would have been 5% (five of 92). The most common abnormal finding in the five injured major vessels was nonobstructing spasm seen in four areas in three superficial femoral arteries and in one popliteal artery. A small mural hematoma in a superficial femoral artery constituted the remaining injury. None of these patients were considered surgical candidates on the basis of the angiographic findings. None of these patients went on to develop vascular complications. When confronted with a ~ a t i e n t with a proximity wound but no clinical signs of vascular injury, it would be beneficial if other features of the injury could be used to help identify those individuals a t increased risk for occult arterial trauma. In our study, patients with associated nerve inju-
ries had a 0% prevalence of vascular injury. Conversely, vascular abnormalities were found in 60% of the patients with fractures due to gunshot wounds. This suggests that a greater amount of energy was deposited into the soft tissues when a bullet i m ~ a c t e don bone than when it passed cleanly through. However, a larger study is necessary before conclusions can be drawn regarding the likelihood of positive angiographic findings in patients with nerve or bone injuries associated with penetrating trauma. Although it is important to characterize the types of injuries revealed a t arteriography for proximity of injury, this information is of diminished value without knowledge of the outcome of patients undergoing this procedure. Unfortunately, these data can be notoriously difficult to obtain (8,261. Typical of inner city trauma services, most of our patients were young men with gunshot wounds who demonstrated poor compliance with scheduled clinic visits. Telephone numbers, next of kin, and addresses given on admission were frequently inaccurate. Despite these difficulties, we were able to obtain follow-up in 42% of negative studies and in 70% of positive studies, with a mean of 21 days for the former and 20 days for the latter. All follow-ur,involved ~ a t i e n t visits, rather than direct or indirect verbal or written data collection. No patients developed vascular signs or symptoms nor required surgical intervention while in follow-up. A longer duration of follow-up would have been preferable, but this was not possible for reasons described previously. We do not know the true prevalence of missed injuries or delayed complications. In the literature, the reported time to the clinical manifestation of initiallv missed injuries varies considerably: as early as 12 hours for patients undergoing only physical examination (8) to a median of 10 days for those undergoing angiography (17). For geographic and economic reasons, our patients are most likely to return to the Bos-
Kaufman e t a1
-
-
--
723 --
Volume 3 Number 4
ton City Hospital for additional care. To our knowledge, none of these patients have been readmitted with vascular complaints related to their injuries. In conclusion, a t routine angiography for proximity of injury in penetrating extremity trauma due to gunshot and stab wounds, we found a n overall positive rate of 24%but only 5% for injury to major vessels. The vascular abnormalities seen in all vessels were minor in degree. The majority of the injured vessels were in the thigh (86%)as the result of gunshot wounds (85%).There were no early complications associated with a conservative approach to the types of injuries described herein. Routine angiography solely for proximity of injury (in the absence of "hard" physical signs of vascular injury) in penetrating gunshot and stab wounds to the extremities is no longer performed a t our institution. References 1. Snyder WH, Thal ER, Bridges RA, Gerlock AJ, Perry MO, Fry WJ. The validity of normal arteriography in penetrating trauma. Arch Surg 1978; 113:424-428. 2. O'Gorman RB, Feliciano DV, Bitondo CG, Mattox KL, Burch JM, Jordan GL Jr. Emergency center arteriography in the evaluation of suspected peripheral vascular injuries. Arch Surg 1984; 119:568-573. 3. Richardson JD, Vitale GC, Flint LM J r . Penetrating trauma: analysis of missed vascular injuries. Arch Surg 1987; 122:67&683. 4. Sclafani SJA, Cooper R, Shaftan GW, Goldstein AS, Glanz S, Gordon DH. Arterial trauma: diagnostic and therapeutic angiography. Radiology 1986; 161:165-172. 5. Menzoian JO, Doyle JE, Cantelmo NL, LoGerfo FW, Hirsch E. A comprehensive approach to extremity vascular trauma. Arch Surg 1985; 120:801-805.
Menzoian JO, Doyle J E , LoGerfo FW, Cantelmo N, Weitzman F, Sequiera JC. Evaluation and management of vascular injuries of the extremities. Arch Surg 1983; 118: 93-95. Frykberg ER, Dennis JW, Bishop K, Laneve L, Alexander RH. The reliability of physical examination in the evaluation of penetrating extremity trauma for vascular injury: results at one year. J Trauma 1991; 31:502-511. Reid JDS, Redman HC, Weigelt JA, Thal ER, Francis H. Wounds of the extremities in proximity to major arteries: value of angiography in the detection of arterial injury. AJR 1988; 151:1035-1039. Rose SC, Moore EE. Trauma angiography: the use of clinical findings to improve patient selection and case preparation. J Trauma 1988; 28:240-245. Dennis JW, Frykberg ER, Crump JM, Vines FS, Alexander RH. New perspectives on the management of penetrating trauma in proximity to major limb arteries. J Vasc Surg 1990; 11:85-93. Rose SC. Arteriographic evaluation of peripheral trauma. In: Kim D, Orren DE, eds. Peripheral vascular imaging and intervention. St Louis: Mosby-Year Book, 1992; 155164. Snyder WH 111, Thal ER, Perry MO. Vascular injuries of the extremities. In: Rutherford RB, ed. Vascular surgery. 3rd ed. Philadelphia: Saunders, 1989; 613-637. Smyth SH, Pond GD, Johnson PL, Rauch RF, McIntyre KE. Proximity injuries: correlation with results of extremity arteriography. JVIR 1991; 2:451-454. Spigos DG, Al-Hammami G, Dunnes PM, Langer BG. Arterial injury to the lower extremities: experience at a level I trauma center (abstr). Radiology 1991; 181(P):151. King TA, Perse JA, Marmen C, Darvin HI. Utility of arteriography in penetrating extremity trauma. Am J Surg 1991; 162:163-165.
Sclafani SJA. Comment. JVIR 1991; 2:454-456. Feliciano DV, Cruse PA, Burch JM, Bitondo PA. Delayed diagnosis of arterial injuries. Am J Surg 1987; 154:579-584. Stain SC, Yellin AE, Weaver FA, Pentecost MJ. Selective management of nonocclusive arterial injuries. Arch Surg 1989; 124:11361141. Johnasen K, Lynch K, Paun M, Copass M. Non-invasive vascular tests reliably exclude occult arterial trauma in injured extremities. J Trauma 1991; 31:515-522. Anderson RJ, Hobson RW, Padberg FT, et al. Reduced dependency on arteriography for penetrating extremity trauma (PET): influence of wound location and non-invasive vascular studies. J Trauma 1990; 30:1059-1065. Anderson RJ, Hobson RW 11, Padberg FT Jr, et al. Penetrating extremity trauma: identification of patients at risk requiring arteriography. J Vasc Surg 1990; 11:544-548. Francis H, Thal ER, Weigelt JA, Redman HC. Vascular proximity: is it a valid indication for arteriography in asymptomatic patients? J Trauma 1991; 31:512-514. Hartling RP, McGahan JP, Blaisdell FW, Lindfors KK. Stab wounds to the extremities: indications for angiography. Radiology 1987; 162: 465-467. Frykberg ER, Vines FS, Alexander AH. The natural history of clinically occult arterial injuries: a prospective evaluation. J Trauma 1989; 29:577-583. Rose SC, Moore EE. Angiography in patients with arterial trauma: correlation between angiographic abnormalities, operative findings, and clinical outcome. AJR 1987; 149:613-619. Rose SC, Moore EE. Emergency trauma angiography: accuracy, safety, and pitfalls. AJR 1987; 148: 1243-1246.
Index terms: Arteries, injuries, 91.41, 92.41 Extremities, angiOgaphy, 91.122, 92.122 Extremities, injuries, 91.41, 92.41 Trauma, 91.41,92.41
JVIR 1992; 3:719-723 Abbreviations: AVF = arteriovenous fistula, DSA = digital subtraction angiography
Arteriography for proximity of injury was studied prospectively at a trauma center. Findings in 85 patients with penetrating extremity wounds were analyzed to determine the prevalence and types of vascular abnormalities seen with these injuries. Ninety-two limb segments were studied for 77 gunshot and 15 stab wounds. Arteriographic findings were positive in 24% overall but in only 5% for injuries confined to major vessels. A 60% positive rate was seen in a small subgroup of 10 patients with fractures due to gunshot wounds. The most frequently injured vessels were muscular branches of the deep femoral artery (59%); the most common injury was focal, nonocclusive spasm (42%). All patients were treated conservatively, without sequelae at follow-up.In this study, the vascular injuries found at arteriography for proximity of injury in penetrating trauma due to bullets or knives, particularly in the thigh, did not require surgical or radiologic intervention.
A R T E R l o G R M H Y performed
1 From the Departments of Vascularilnterventional Radiology (J.A.K., J.E.P., A.J.G.) and Vascular Surgery (D.L.G.,J.w., J.O.M.), Boston University MedicalCenter2 Boston. Received March 12, 1992; revision requested ~ ~ ~ i lrevision 1 7 ; received June 5; accepted June 12. Address reprint requests to J.A.K., Section of Vascular Radi-
~''gy, Massachusetts Hospital, 32 Fruit St, Boston, MA 02114. Current address: ~ ~ ventional Radiology, T U ~ ~New S England Medical Center, Boston. SCVIR, 1992
to exclude vascular injury when extremity trauma is in proximity to a major artery but physical examination is negative has had a changing role in the management of the trauma patient (1-4). In the middle 1980s, Menzoian et a1 recommended the routine use of proximity arteriography after finding a 16%yield of occult vascular injuries (5,6).Subsequent studies in both the radiology and trauma literature have supported restricting the use of proximity arteriography, due in part to the adoption of a more conservative surgical approach toward minor vascular injuries (7-9). Meticulous evaluation of the patient by means of physical examination for "hard" signs of vascular injury has been a crucial element in these series (7-9). The vascular surgery and radiology services at our hospital recently conducted a reevaluation of routine proximity arteriography for all mechanisms of injury. ~ We reviewed d the angiographic findings in the subset of patients with penetrating trauma due to gunshot and stab wounds, all of
whom were treated with conservative management.
PATIENTS AND METHODS Review of the protocol by the institutional Human Studies Committee was not required because randomization was not performed and patients were treated according to current practice at the time of the study. Patients with trauma to the extremities were entered prospectively into the study over a 14-month period starting July 1989. Upon presentation to the Boston City Hospital, patients were evaluated by senior surgical residents for the following inclusion criteria: ( a )trauma within 1cm of the expected course of a major artery and ( b )absence of "hard" signs of vascular injury-bruit, loss of pulses, limb ischemia, active bleeding, expanding hematoma, or history of pulsatile bleeding. ~ ~ ~ This analysis was limited to the subset of patients with penetrating injuries due to bullets or sharp ob-
719
~
720
Journal of Vascular and Interventional Radiology
--
November 1992
jects. Fractures and neurologic deficits due to penetrating trauma were noted but were not considered a basis for exclusion. Patients with blunt trauma or dislocations, or patients who had undergone surgery on the affected extremity prior to angiography were not included. Patients with multiple pellet injuries were also excluded because only one of the three patients met the conditions outlined above. Eighty-five patients met the criteria for this review: 82 men (age range, 17-59 years; mean, 26 years), and three women (age range, 20-32 years; mean, 27 years). All 85 patients underwent angiography within 24-48 hours of presentation in the radiology department after informed consent was obtained. Femoral arterial catheterization was accomplished by using standard retrograde angiographic technique and a 5-F straight catheter with multiple side holes, a 5.5-F C-2 cobra end-hole catheter, or a 5.5-F H-1 catheter (Cook. Bloomington, Ind). ~ e ~ e n don' i nthe ~ location of the injury, contrast material was injected into the ipsilateral external iliac arterv. " contralateral external iliac artery, or subclavian artery. Iohex01 (350 mg of iodine per milliliter) (Omnipaque 350; Sanofi Winthrop, New York) was used for cut-film angiography; iothalamate meglumine (202 mg of iodine per milliliter) (Conray 43; Mallinckrodt, St Louis) was used for occasional supplemental digital subtraction angiography (DSA). At least two views were obtained in all studies. Although one study was performed entirely with digital imaging, we considered that the quality of the images obtained on our equipment (pixel matrix, 256 x 256) precluded the routine use of this modality. A comparison of cut-film and DSA images was not performed. Limb segments were defined as thigh (inguinal ligament to the femoral condyles), knee (the popliteal fossa), and lower leg (proximal head of the fibula to the ankle) for the lower extremity, and upper arm (anterior axillary line to the humeral condyles) and forearm (humeral conr
dyles to the ulnar styloid) for the upper extremity. Angiograms were considered positive for injury if any of the following were seen in any opacified arterial structure: arteriovenous fistula (AVF),extravasation, intimal flap, laceration, mural hematoma, occlusion, pseudoaneurysm, spasm, or transection. Mass effect (deviation of the course of an otherwise normal artery), parenchymal stains, and standing waves were not considered positive findings. Major vessels were considered to be arteries that if occluded acutely would result in an ischemic, nonviable extremity. All other vessels were termed minor. After arteriography, patients were monitored with serial physical examinations and measurements of anklebrachial indexes on the surgical service for a minimum of 24 hours. Surgical clinic appointments were scheduled on discharge. If patients remained hospitalized for other reasons, their vascular examinations were monitored by the surgical service throughout the duration of the admission. Follow-up ranged from 24 hours to 178 days. Only clinical follow-up was considered adequate. The constraints of practice in a busy trauma center did not permit either blinding of the investigators or randomization to a nonangiographic treatment group.
RESULTS A total of 92 limb segments were evaluated with arteriography for 77 (84%)gunshot wounds and 15 (16%) stab wounds. The distribution of injuries to the extremities was 59 thigh (64%), 15 upper arm (16%), 13 knee (14%),four lower leg (4%),and one forearm (1%).There were 70 (76%) negative limb segments in 64 patients (six patients underwent study of two limb segments) and 22 (24%) positive limb segments in 21 patients (one patient had two positive segments). The mechanisms of injury in the 22 positive studies were gunshot wounds in 19 (25% of all gunshot
Figure 1. Transection and occlusion of the descending deep femoral artery trunk due to a gunshot injury (arrow). Spasm and a small filling defect are present in the adjacent muscular branch. The patient refused follow-up angiography.
wounds, 86% of all positive limb segments) and stab wounds in three (20% of all stab wounds, 14% of all positive limb segments). The majority of the injuries involved unnamed branches of the deep femoral artery (59%, 1 3 of 22 studies), followed in frequency by the superficial femoral artery (18%, four of 221, the deep femoral artery (9%, two of 22), the popliteal artery (5%, one of 221, a n unnamed muscular branch of the circumflex humeral artery (5%,one of 221, and a peroneal artery (5%, one of 22). Of the 22 positive extremity segments, 18 (82%)of the injured vessels were in the thigh. A total of 24 vascular abnormalities were identified in 22 vessels. These included 10 areas of focal nonocclusive spasm (42%),six vascular occlusions (25%),two areas of minimal extravasation (8%),two intimal flaps (8%),two AVFs (8%),one small mural hematoma (4%),and one transection (4%).Two occlusions occurred in conjunctiorl with other abnormalities: an intimal flap in one case and a transection in another. The majority of the injuries, 15 of 24 (63%),occurred in unnamed muscular branches, including both AVFs.
Kaufman et a1
72 1
Volume 3 Number 4
c. a. b. Figure 2. AVF of a muscular branch of the deep femoral artery due to multiple stab wounds, with spontaneous closure at 36 days. (a)Arrow identifies early filling of a profunda vein on the original proximity angiogram. Standing waves are present in the superficial femoral artery. (b)Dense venous opacification is seen later in the same run. (c) Arrow identifies former site of AVF, which
closed spontaneously, on follow-up angiogram at 36 days.
Among named arteries, focal nonobstructing spasm was the most common abnormality, identified in three superficial femoral arteries and in the only abnormal popliteal artery. The small mural hematoma occurred in a superficial femoral artery, a small intimal flap was found in a deep femoral artery origin, a short occlusion was noted in a peroneal artery, and a descending deep femoral artery trunk was transected and occluded distal to the origin of the first unnamed muscular branch below the lateral circumflex femoral artery (Fig 1). Neurologic deficits occurred in five patients (6%), most commonly radial nerve palsies (three of five). These were found in association with one stab wound, one simple gunshot wound, and one gunshot wound with a humeral fracture. The remaining two deficits were trivial. consisting of focal hypoesthesia over t h e laterar thigh and the dorsum of the foot. All five of these patients had negative arteriograms. Within the gunshot wound group, 10 patients had fractures due to bullet impact on bone: femur in six, humerus in two, and fibulae in two. All fractures were comminuted, closed,
and minimally displaced. Six patients (60%)had positive studies involving seven vessels: muscular branches of the deep femoral artery in four, superficial femoral artery in one, peroneal artery in one, and a muscular branch of the circumflex humeral artery in one. Focal spasm was present in a muscular branch of the deep femoral artery, the circumflex humeral artery, and the superficial femoral artery. In three cases occlusions were noted, involving two muscular branches of the deep femoral artery and a tibia1 runoff vessel (peroneal artery). One intimal flap was found in conjunction with an occlusion of a muscular branch of the deep femoral artery. One angiographic complication occurred-a small groin hematoma a t the arterial puncture site. This was noted on return of the patient to the ward after angiography. The hematoma remained stable in size, requiring no intervention. The overall rate of procedure-related complications was therefore 1%. No patients developed vascular symptoms within the immediate postangiographic period. Follow-up longer than 24 hours was available in
28 of 64 patients (44%)with negative examinations a t a mean of 2 1days after arteriography (range, 5-107 days). Among the patients with positive angiograms, 16 of 21 (76%)had a mean duration of follow-up of 20 days (range, 3-178 days). The group of ~ a t i e n t with s mnshot wounds and fractures had a longer follow-up, with a mean of 33 days (range, 5-178 days). Follow-up consisted of visits in surgical or orthopedic clinics, although some patients were seen during prolonged hospitalizations for treatment of nonvascular injuries or a t subsequent emergency room admissions for unrelated trauma. One patient, with an AVF of a muscular branch of the deep femoral artery, underwent repeat angiography at 36 days demonstrating spontaneous closure (Fig 2). The other patient with an AVF was lost to follow-UD.None of the patients who were seen in follow-up developed symptoms or evidence of vascular injury. u
DISCUSSION The history of the treatment of proximity injuries has been well iter-
722
Journal of Vascular and Interventional Radiology -
-- -
November 1992
ated in the literature (10-12). In patients without evidence of vascular trauma on physical examination, manv of the vascular iniuries identified at angiography a r e k i n o r (9,101. Recently, a general trend toward the expectant surgical management of minor vascular injuries has led to a reconsideration of the role of proximity angiography. Although this conservative approach is not universally accepted (13-171, it is amply supported in the literature (18-24). Penetrating extremity trauma due to gunshot and stab wounds is common at our institution. We were interested in identifying the prevalence, types, and locations of vascular abnormalities found in patients whose angiograms were obtained solely because of the proximity of injury. This information is useful in determining the necessity and priority of peripheral vascular imaging studies in these individuals, many of whom have additional, more pressing injuries. Furthermore, a knowledge of the expected angiographic findings in this setting may be helpful when proximity studies are requested at centers where these injuries are less common. Our greatest angiographic experience was with gunshot wounds77 in all. However. we were not able to segregate low-velocity from highvelocity bullet wounds. The second large grouping was 15 stab wounds, all of which were due to knives. In this analysis, we excluded shotgun injuries, as we did not believe that our very limited numbers would be meaningful. The reported rates of positive findings for this mechanism at proximity angiography are higher than those for gunshot and stab wounds (10,22). The results described in this article should not be extrapolated to proximity arteriography performed for dislocations, fractures due to blunt trauma, or injuries caused by multiple pellets. The thigh was the most commonly injured limb segment in our study and also had the highest overall rate of vascular injury (32%).This has not been true for other authors. Dennis
et a1 found the highest rate in upper arm injuries, 15.7% (10);Francis et al, in the forearm, 6.3% (22); and Anderson et al, in the calf, 22.9% (21). The explanation for this variation is not clear but may be a function of study size rather than of differences in patient populations. Our positive rate of 24% is a t the upper end of the range reported in the literature (11). This a reflection of our definition of a "normal" proximity angiogram as demonstrating no evidence of a vascular injury in any arterial structure. It is important to note that the majority of our angiographic findings, 19 of 24 (79%),were in named muscular arteries or unnamed terminal muscular branches. These injuries occurred in vessels that were not essential for limb viability. In addition, they are difficult to approach surgically. The abnormalities were considered too small to warrant intervention by either the vascular radiologists or the vascular surgeons. Specifically, these findings included six of nine areas of spasm, six of six occlusions, both cases of minimal extravasation, both AVFs, both intimal flaps, and the single transection. Had we excluded injuries to minor vessels from our analysis, as do some authors (22,24,25),our positive rate would have been 5% (five of 92). The most common abnormal finding in the five injured major vessels was nonobstructing spasm seen in four areas in three superficial femoral arteries and in one popliteal artery. A small mural hematoma in a superficial femoral artery constituted the remaining injury. None of these patients were considered surgical candidates on the basis of the angiographic findings. None of these patients went on to develop vascular complications. When confronted with a ~ a t i e n t with a proximity wound but no clinical signs of vascular injury, it would be beneficial if other features of the injury could be used to help identify those individuals a t increased risk for occult arterial trauma. In our study, patients with associated nerve inju-
ries had a 0% prevalence of vascular injury. Conversely, vascular abnormalities were found in 60% of the patients with fractures due to gunshot wounds. This suggests that a greater amount of energy was deposited into the soft tissues when a bullet i m ~ a c t e don bone than when it passed cleanly through. However, a larger study is necessary before conclusions can be drawn regarding the likelihood of positive angiographic findings in patients with nerve or bone injuries associated with penetrating trauma. Although it is important to characterize the types of injuries revealed a t arteriography for proximity of injury, this information is of diminished value without knowledge of the outcome of patients undergoing this procedure. Unfortunately, these data can be notoriously difficult to obtain (8,261. Typical of inner city trauma services, most of our patients were young men with gunshot wounds who demonstrated poor compliance with scheduled clinic visits. Telephone numbers, next of kin, and addresses given on admission were frequently inaccurate. Despite these difficulties, we were able to obtain follow-up in 42% of negative studies and in 70% of positive studies, with a mean of 21 days for the former and 20 days for the latter. All follow-ur,involved ~ a t i e n t visits, rather than direct or indirect verbal or written data collection. No patients developed vascular signs or symptoms nor required surgical intervention while in follow-up. A longer duration of follow-up would have been preferable, but this was not possible for reasons described previously. We do not know the true prevalence of missed injuries or delayed complications. In the literature, the reported time to the clinical manifestation of initiallv missed injuries varies considerably: as early as 12 hours for patients undergoing only physical examination (8) to a median of 10 days for those undergoing angiography (17). For geographic and economic reasons, our patients are most likely to return to the Bos-
Kaufman e t a1
-
-
--
723 --
Volume 3 Number 4
ton City Hospital for additional care. To our knowledge, none of these patients have been readmitted with vascular complaints related to their injuries. In conclusion, a t routine angiography for proximity of injury in penetrating extremity trauma due to gunshot and stab wounds, we found a n overall positive rate of 24%but only 5% for injury to major vessels. The vascular abnormalities seen in all vessels were minor in degree. The majority of the injured vessels were in the thigh (86%)as the result of gunshot wounds (85%).There were no early complications associated with a conservative approach to the types of injuries described herein. Routine angiography solely for proximity of injury (in the absence of "hard" physical signs of vascular injury) in penetrating gunshot and stab wounds to the extremities is no longer performed a t our institution. References 1. Snyder WH, Thal ER, Bridges RA, Gerlock AJ, Perry MO, Fry WJ. The validity of normal arteriography in penetrating trauma. Arch Surg 1978; 113:424-428. 2. O'Gorman RB, Feliciano DV, Bitondo CG, Mattox KL, Burch JM, Jordan GL Jr. Emergency center arteriography in the evaluation of suspected peripheral vascular injuries. Arch Surg 1984; 119:568-573. 3. Richardson JD, Vitale GC, Flint LM J r . Penetrating trauma: analysis of missed vascular injuries. Arch Surg 1987; 122:67&683. 4. Sclafani SJA, Cooper R, Shaftan GW, Goldstein AS, Glanz S, Gordon DH. Arterial trauma: diagnostic and therapeutic angiography. Radiology 1986; 161:165-172. 5. Menzoian JO, Doyle JE, Cantelmo NL, LoGerfo FW, Hirsch E. A comprehensive approach to extremity vascular trauma. Arch Surg 1985; 120:801-805.
Menzoian JO, Doyle J E , LoGerfo FW, Cantelmo N, Weitzman F, Sequiera JC. Evaluation and management of vascular injuries of the extremities. Arch Surg 1983; 118: 93-95. Frykberg ER, Dennis JW, Bishop K, Laneve L, Alexander RH. The reliability of physical examination in the evaluation of penetrating extremity trauma for vascular injury: results at one year. J Trauma 1991; 31:502-511. Reid JDS, Redman HC, Weigelt JA, Thal ER, Francis H. Wounds of the extremities in proximity to major arteries: value of angiography in the detection of arterial injury. AJR 1988; 151:1035-1039. Rose SC, Moore EE. Trauma angiography: the use of clinical findings to improve patient selection and case preparation. J Trauma 1988; 28:240-245. Dennis JW, Frykberg ER, Crump JM, Vines FS, Alexander RH. New perspectives on the management of penetrating trauma in proximity to major limb arteries. J Vasc Surg 1990; 11:85-93. Rose SC. Arteriographic evaluation of peripheral trauma. In: Kim D, Orren DE, eds. Peripheral vascular imaging and intervention. St Louis: Mosby-Year Book, 1992; 155164. Snyder WH 111, Thal ER, Perry MO. Vascular injuries of the extremities. In: Rutherford RB, ed. Vascular surgery. 3rd ed. Philadelphia: Saunders, 1989; 613-637. Smyth SH, Pond GD, Johnson PL, Rauch RF, McIntyre KE. Proximity injuries: correlation with results of extremity arteriography. JVIR 1991; 2:451-454. Spigos DG, Al-Hammami G, Dunnes PM, Langer BG. Arterial injury to the lower extremities: experience at a level I trauma center (abstr). Radiology 1991; 181(P):151. King TA, Perse JA, Marmen C, Darvin HI. Utility of arteriography in penetrating extremity trauma. Am J Surg 1991; 162:163-165.
Sclafani SJA. Comment. JVIR 1991; 2:454-456. Feliciano DV, Cruse PA, Burch JM, Bitondo PA. Delayed diagnosis of arterial injuries. Am J Surg 1987; 154:579-584. Stain SC, Yellin AE, Weaver FA, Pentecost MJ. Selective management of nonocclusive arterial injuries. Arch Surg 1989; 124:11361141. Johnasen K, Lynch K, Paun M, Copass M. Non-invasive vascular tests reliably exclude occult arterial trauma in injured extremities. J Trauma 1991; 31:515-522. Anderson RJ, Hobson RW, Padberg FT, et al. Reduced dependency on arteriography for penetrating extremity trauma (PET): influence of wound location and non-invasive vascular studies. J Trauma 1990; 30:1059-1065. Anderson RJ, Hobson RW 11, Padberg FT Jr, et al. Penetrating extremity trauma: identification of patients at risk requiring arteriography. J Vasc Surg 1990; 11:544-548. Francis H, Thal ER, Weigelt JA, Redman HC. Vascular proximity: is it a valid indication for arteriography in asymptomatic patients? J Trauma 1991; 31:512-514. Hartling RP, McGahan JP, Blaisdell FW, Lindfors KK. Stab wounds to the extremities: indications for angiography. Radiology 1987; 162: 465-467. Frykberg ER, Vines FS, Alexander AH. The natural history of clinically occult arterial injuries: a prospective evaluation. J Trauma 1989; 29:577-583. Rose SC, Moore EE. Angiography in patients with arterial trauma: correlation between angiographic abnormalities, operative findings, and clinical outcome. AJR 1987; 149:613-619. Rose SC, Moore EE. Emergency trauma angiography: accuracy, safety, and pitfalls. AJR 1987; 148: 1243-1246.
