Is Arterial Proximity a Valid Indication for Arteriography in Penetrating Extremity Trauma? A
Prospective Analysis
Fred A. Weaver, MD; Albert E. Yellin, MD; Madeline Bauer, PhD; Jon Oberg, MD; Navid Ghalambor; Richard P. Emmanuel; Robert M. Applebaum, MD; Michael J. Pentecost, MD; Robert M. Shorr, MD
\s=b\Three hundred seventy-three patients with a penetrating extremity injury were studied to assess the yield of arteriography. Patients underwent arteriography if any of the following was present: bruit, history of hemorrhage or hypotension, fracture, hematoma, decreased capillary refill, major soft-tissue injury, or nerve or pulse deficit. In the absence of these findings, arteriography was performed if the injury was in "proximity" to a major neurovascular bundle. In 216 patients, arteriography was performed when an abnormal finding was noted. Sixty-five injuries were identified, 19 requiring intervention. Proximity was the indication for arteriography in 157 patients. Seventeen injuries were identified, of which one required repair. In penetrating extremity trauma, the need for arteriography is based on clinical findings. The use of arteriography to screen for an arterial injury when proximity alone is the indication rarely identifies a significant injury and should be abandoned. (Arch Surg. 1990;125:1256-1260)
Rather practice
than risk the medical and possible legal conse¬ quences of a missed arterial injury, it has become com¬ for surgeons to employ arteriography as a mon device in patients with penetrating extremity trau¬ screening ma. The value of arteriography in patients who do not require immediate operative intervention but have obvious signs of an arterial injury, is readily apparent. An arteriogram per¬ formed in either the operating room or radiology suite will identify the location and nature of an injury and assist in planning the subsequent treatment. The value of arteriogra¬ phy in patients in whom there are no definite signs of an arterial injury is less clear. The recognition that occult arteri¬ al injuries may be present despite the absence of overt physi¬ cal signs of arterial trauma has led many surgeons to advocate the routine use of diagnostic arteriography for injuries in "proximity" to a major neurovascular bundle.1"7 The clinical usefulness and diagnostic yield of arteriogra¬ phy for penetrating extremity injuries in proximity to a neu¬ rovascular bundle remains controversial. A recent report by Frykberg et al8 suggests that the ultimate benefit of routine arteriograms is minimal and represents overuse of costly medical resources. Others, however, have maintained that if arteriography for proximity were abandoned, major arterial injuries would be missed. This would result in serious medi¬ cal, legal, and financial consequences.4'8,9 This study was de¬ signed to address these concerns by prospectively determin¬ ing the diagnostic and therapeutic yield of arteriography Accepted for publication June 9,1990. From the Departments of Surgery (Drs Weaver, Yellin, Bauer, Oberg, Applebaum, and Shorr and Messrs Ghalambor and Emmanuel) and Radiology (Dr Pentecost), University of Southern California School of Medicine, Los Angeles. Read before the 61st Annual Meeting of the Pacific Coast Surgical Association, Laguna Niguel, Calif, February 19,1990. Reprint requests to Department of Surgery, LAC/USC Medical Center, 1200 N State St, Rm 9442, Los Angeles, CA 90033 (Dr Weaver).
when routinely performed for proximity alone as well as for other signs more directly suggestive of an arterial injury.
PATIENTS AND METHODS
During the 18-month period from July 1, 1987, to December 31, 1988, 373 patients with a unilateral penetrating injury to an upper or lower extremity were entered into a prospective study designed to assess the value of routine diagnostic arteriography in identifying clinically occult arterial injuries. Patients with a penetrating injury distal to the deltopectoral groove or inguinal ligament underwent arteriography if a distal pulse deficit, nerve deficit, hematoma, histo¬ ry of hemorrhage or hypotension, bruit, fracture, major soft-tissue injury, or delayed capillary refill was present; or if in the absence of these findings, the path of the penetrating object was judged by the admitting surgeon to be in proximity to a major neurovascular bun¬ dle. All arteriograms were performed in the radiology department via a femoral puncture and confined to the injured limb. Patients with bilateral penetrating injuries, injuries outside the defined anatomic boundaries, or arterial injuries that required operation because of severe limb ischemia or active hemorrhage were excluded. Before the arteriogram was performed, information on the age and
patient, injury type (stab wound, shotgun wound, gunshot other), extremity injured (arm or leg), and side injured (left or right) was obtained. A history of hemorrhage or hypotension, the quadrants (medial anterior, medial posterior, lateral anterior, and lateral posterior) traversed by the penetrating object, and the presence of a bruit, pulse deficit, hematoma, fracture, major softtissue injury, neurologic deficit (sensory, motor, combined, or glove distribution), or decreased capillary refill was noted. In 210 patients, posterior tibial/dorsalis pedis or radial/ulna Doppler pressures of the injured and contralateral uninjured extremity were obtained and indexed (ankle-arm/brachial index [ABI]) using the brachial Doppler pressure of an uninjured extremity. Each arteriogram was interpreted as "normal" or "injured." The arterial segment injured was described. Arterial injuries were de¬ fined as "major" or "minor." Major injuries were injuries to arterial segments that, if interrupted, would likely result in clinically signifi¬ cant limb ischemia. All other arterial injuries were termed minor. The arterial injury was further classified as an intimai defect, intimai flap, pseudoaneurysm, arteriovenous fistula, stenosis, or occlusion as defined in a previous report.10 Therapy was based on the artério¬ graphie findings and included observation, artériographie embolization, arterial ligation, or repair. The artériographie findings, arterial segment injured, and treat¬ ment were compiled for patients in whom diagnostic arteriography was performed for abnormal clinical findings and for patients without abnormal findings in whom proximity alone was the indication. The sensitivity and specificity of the Doppler ABI in predicting an artériographie arterial injury were determined for the 210 patients in whom Doppler data were available. The minimum of the two ABIs sex of each
wound,
or
( ) at the ankle or wrist were used for these determinations. Three different ranges, less than or equal to 0.99 and less than or equal to 0.90, were arbitrarily chosen as abnormal for these calcula¬ tions. Sensitivity represents the percent of artériographie arterial and specificity injuries correctly predicted by an abnormal represents the percent of normal arteriograms correctly predicted by a normal MAB I.
Downloaded From: http://archsurg.jamanetwork.com/ by a Monash University Library User on 06/17/2015
For all patients, a multivariate stepwise logistic regression analy¬ sis was performed to determine whether any abnormal findings correlated with the artériographie demonstration of a major or minor arterial injury.11 The following variables were considered for this analysis: age, sex, injury type, extremity, side, projectile path by quadrant, pulse deficit, fracture, major soft-tissue injury, hemato¬ ma, neurologic deficit, bruit, history of hemorrhage, hypotension, decreased capillary refill, and of the injured and contralateral
uninjured extremity. Three hundred seventy-three patients were entered into the proto¬ col; 345 were male and 28 were female. Median age was 24 years (range, 10 to 73 years). There were 229 (61%) lower and 144 (39%) upper extremity injuries. The right extremity was injured in 172 (46%) cases and the left in 201 (54%) cases. The quadrants traversed by each injury are shown in Table 1. The incidence of injury type, history of hemorrhage or hypotension, fracture, hematoma, neuro¬ logic deficit, pulse deficit, soft-tissue injury, decreased capillary refill, and bruit are found in Table 2. Two hundred sixteen patients (58%) underwent arteriography because one or more abnormal find¬ ings were present, and 157 (42%) had proximity alone as the indication.
RESULTS
=
=
Artériographie Findings
Injury vs
Artériographie Findings, No.
Quadrants*
No. of Patients
Normal
Med/ant-med/post
26
21
5
Med/ant-lat/ant
32
19
13
Med/ant-lat/post
34
28
6
Med/ant-no exit
113
94
19
Med/post-lat/ant
24
13
11
Med/post-lat/post Med/post-no exit Lat/ant-lat/post
15
10
5
29
26
3
56
44
12
40
31
9
373
291
82
Lat/ant-no exit Lat/post-no exit Total *Ant indicates anterior;
Eighty-two patients (22%) had an arterial injury demon¬ strated arteriographically (Table 3). The quadrants traversed by the penetrating object and the incidence of artériographie injuries for each combination are found in Table 1. The inci¬ dence of a positive arteriogram for given injury or abnormal in finding is found in Table 2. The mean and median extremities with an arterial injury were 0.81 and 0.92, respec¬ tively. In extremities in which an arterial injury was not was 0.94 and the median, demonstrated, the mean 0.99. The sensitivity/specificity of the in predicting an arterial injury was 82%/40% at less than or equal to 0.99, 71%/63% at less than or equal to 0.95, and 49%/76% at less than or equal to 0.90. In the 216 patients with abnormal findings, 65 arterial injuries were found, of which 22 were major and 43 were minor. Nineteen injuries in this cohort required arterial re¬ pair or embolization. Proximity alone was the indication for arteriography in 157 patients. In this group, 17 arterial inju¬ ries (four major and 13 minor) were demonstrated. One in¬ jured artery underwent ligation (Table 3). Overall, 20 arterial injuries were treated by repair or arté¬ riographie embolization. The arterial repair was end-to-end in 11 and by vein interposition in four. Embolization was per¬ formed for four injuries and arterial ligation for one. All other injuries were treated nonoperatively. The major artery inju¬ ries treated nonoperatively included intimai defects (n 7), occlusions (two profunda femoris, one profunda brachii) (n 3), pseudoaneurysm (n 1), and stenosis (n 1). Thirtytwo minor artery occlusions, five intimai defects, five intimai flaps, five small pseudoaneurysms, and three small arteriovenous fistulae were treated nonoperatively (Table 3). Following arterial repair, all patients were discharged with a patent repair and intact vascular examination. Embolizations were successful in all instances. No untoward events occurred in those patients treated nonoperatively and all were discharged with a stable vascular examination and in¬ tact limb. Twenty-one patients treated nonoperatively have been fol¬ lowed up from 1 to 8 months (median, 2 months) after injury and remain clinically stable. Ten patients have undergone repeated arteriography at 4 days to 2 months after injury to reassess five intimai injuries, three arteriovenous fistulae, and two pseudoaneurysms. All injuries had resolved with the exception of an intimai defect in the radial artery, which was unchanged, and a persistent medial geniculate arteriovenous fistula, which was embolized. For the stepwise logistic regression analysis, the following variables were initially considered for the model: age, sex, =
Table 1 .—Quadrants Traversed by Penetrating
Injury (%)
(19) (40) (18) (16) (46) (33) (10) (21) (22)
Lat, lateral; Med, medial; and Post, posterior.
Table 2.—Incidence of Arterial Injury for Each Abnormal Finding No. of Patients
Injury Type or Arterial
Injury, No. (%)
Injury
Stab wound
76
(16) (46) 57 (22) 1 (8) 25 (26) 21 (28)
27
14 3
Motor/sensory
13 22
Glove All
Shotgun
79 24
Gunshot Other
256
Fracture Hematoma
95
Neurologic deficit Sensory Motor
13
13 11
8
63
25
Pulse deficit Soft-tissue Injury
38 19
22 4
Hemorrhage history
10
2
Decreased refill
3
Hypotension
2 1
Bruit
(52) (2) (36) (40) (56) (21) (20) (38) (50) (100)
=
injury type, extremity, side, distal pulse deficit, fracture, major soft-tissue injury, hematoma, neurologic deficit, bruit, history of hemorrhage, hypotension, and decreased capillary refill. Those variables that occurred in less than 10 patients, ie, bruit, history of hemorrhage, hypotension, and decreased capillary refill, were subsequently not entered. A pulse defi¬ cit, neurologic deficit, and shotgun injury were shown to correlate significantly (P
Prospective Analysis
Fred A. Weaver, MD; Albert E. Yellin, MD; Madeline Bauer, PhD; Jon Oberg, MD; Navid Ghalambor; Richard P. Emmanuel; Robert M. Applebaum, MD; Michael J. Pentecost, MD; Robert M. Shorr, MD
\s=b\Three hundred seventy-three patients with a penetrating extremity injury were studied to assess the yield of arteriography. Patients underwent arteriography if any of the following was present: bruit, history of hemorrhage or hypotension, fracture, hematoma, decreased capillary refill, major soft-tissue injury, or nerve or pulse deficit. In the absence of these findings, arteriography was performed if the injury was in "proximity" to a major neurovascular bundle. In 216 patients, arteriography was performed when an abnormal finding was noted. Sixty-five injuries were identified, 19 requiring intervention. Proximity was the indication for arteriography in 157 patients. Seventeen injuries were identified, of which one required repair. In penetrating extremity trauma, the need for arteriography is based on clinical findings. The use of arteriography to screen for an arterial injury when proximity alone is the indication rarely identifies a significant injury and should be abandoned. (Arch Surg. 1990;125:1256-1260)
Rather practice
than risk the medical and possible legal conse¬ quences of a missed arterial injury, it has become com¬ for surgeons to employ arteriography as a mon device in patients with penetrating extremity trau¬ screening ma. The value of arteriography in patients who do not require immediate operative intervention but have obvious signs of an arterial injury, is readily apparent. An arteriogram per¬ formed in either the operating room or radiology suite will identify the location and nature of an injury and assist in planning the subsequent treatment. The value of arteriogra¬ phy in patients in whom there are no definite signs of an arterial injury is less clear. The recognition that occult arteri¬ al injuries may be present despite the absence of overt physi¬ cal signs of arterial trauma has led many surgeons to advocate the routine use of diagnostic arteriography for injuries in "proximity" to a major neurovascular bundle.1"7 The clinical usefulness and diagnostic yield of arteriogra¬ phy for penetrating extremity injuries in proximity to a neu¬ rovascular bundle remains controversial. A recent report by Frykberg et al8 suggests that the ultimate benefit of routine arteriograms is minimal and represents overuse of costly medical resources. Others, however, have maintained that if arteriography for proximity were abandoned, major arterial injuries would be missed. This would result in serious medi¬ cal, legal, and financial consequences.4'8,9 This study was de¬ signed to address these concerns by prospectively determin¬ ing the diagnostic and therapeutic yield of arteriography Accepted for publication June 9,1990. From the Departments of Surgery (Drs Weaver, Yellin, Bauer, Oberg, Applebaum, and Shorr and Messrs Ghalambor and Emmanuel) and Radiology (Dr Pentecost), University of Southern California School of Medicine, Los Angeles. Read before the 61st Annual Meeting of the Pacific Coast Surgical Association, Laguna Niguel, Calif, February 19,1990. Reprint requests to Department of Surgery, LAC/USC Medical Center, 1200 N State St, Rm 9442, Los Angeles, CA 90033 (Dr Weaver).
when routinely performed for proximity alone as well as for other signs more directly suggestive of an arterial injury.
PATIENTS AND METHODS
During the 18-month period from July 1, 1987, to December 31, 1988, 373 patients with a unilateral penetrating injury to an upper or lower extremity were entered into a prospective study designed to assess the value of routine diagnostic arteriography in identifying clinically occult arterial injuries. Patients with a penetrating injury distal to the deltopectoral groove or inguinal ligament underwent arteriography if a distal pulse deficit, nerve deficit, hematoma, histo¬ ry of hemorrhage or hypotension, bruit, fracture, major soft-tissue injury, or delayed capillary refill was present; or if in the absence of these findings, the path of the penetrating object was judged by the admitting surgeon to be in proximity to a major neurovascular bun¬ dle. All arteriograms were performed in the radiology department via a femoral puncture and confined to the injured limb. Patients with bilateral penetrating injuries, injuries outside the defined anatomic boundaries, or arterial injuries that required operation because of severe limb ischemia or active hemorrhage were excluded. Before the arteriogram was performed, information on the age and
patient, injury type (stab wound, shotgun wound, gunshot other), extremity injured (arm or leg), and side injured (left or right) was obtained. A history of hemorrhage or hypotension, the quadrants (medial anterior, medial posterior, lateral anterior, and lateral posterior) traversed by the penetrating object, and the presence of a bruit, pulse deficit, hematoma, fracture, major softtissue injury, neurologic deficit (sensory, motor, combined, or glove distribution), or decreased capillary refill was noted. In 210 patients, posterior tibial/dorsalis pedis or radial/ulna Doppler pressures of the injured and contralateral uninjured extremity were obtained and indexed (ankle-arm/brachial index [ABI]) using the brachial Doppler pressure of an uninjured extremity. Each arteriogram was interpreted as "normal" or "injured." The arterial segment injured was described. Arterial injuries were de¬ fined as "major" or "minor." Major injuries were injuries to arterial segments that, if interrupted, would likely result in clinically signifi¬ cant limb ischemia. All other arterial injuries were termed minor. The arterial injury was further classified as an intimai defect, intimai flap, pseudoaneurysm, arteriovenous fistula, stenosis, or occlusion as defined in a previous report.10 Therapy was based on the artério¬ graphie findings and included observation, artériographie embolization, arterial ligation, or repair. The artériographie findings, arterial segment injured, and treat¬ ment were compiled for patients in whom diagnostic arteriography was performed for abnormal clinical findings and for patients without abnormal findings in whom proximity alone was the indication. The sensitivity and specificity of the Doppler ABI in predicting an artériographie arterial injury were determined for the 210 patients in whom Doppler data were available. The minimum of the two ABIs sex of each
wound,
or
( ) at the ankle or wrist were used for these determinations. Three different ranges, less than or equal to 0.99 and less than or equal to 0.90, were arbitrarily chosen as abnormal for these calcula¬ tions. Sensitivity represents the percent of artériographie arterial and specificity injuries correctly predicted by an abnormal represents the percent of normal arteriograms correctly predicted by a normal MAB I.
Downloaded From: http://archsurg.jamanetwork.com/ by a Monash University Library User on 06/17/2015
For all patients, a multivariate stepwise logistic regression analy¬ sis was performed to determine whether any abnormal findings correlated with the artériographie demonstration of a major or minor arterial injury.11 The following variables were considered for this analysis: age, sex, injury type, extremity, side, projectile path by quadrant, pulse deficit, fracture, major soft-tissue injury, hemato¬ ma, neurologic deficit, bruit, history of hemorrhage, hypotension, decreased capillary refill, and of the injured and contralateral
uninjured extremity. Three hundred seventy-three patients were entered into the proto¬ col; 345 were male and 28 were female. Median age was 24 years (range, 10 to 73 years). There were 229 (61%) lower and 144 (39%) upper extremity injuries. The right extremity was injured in 172 (46%) cases and the left in 201 (54%) cases. The quadrants traversed by each injury are shown in Table 1. The incidence of injury type, history of hemorrhage or hypotension, fracture, hematoma, neuro¬ logic deficit, pulse deficit, soft-tissue injury, decreased capillary refill, and bruit are found in Table 2. Two hundred sixteen patients (58%) underwent arteriography because one or more abnormal find¬ ings were present, and 157 (42%) had proximity alone as the indication.
RESULTS
=
=
Artériographie Findings
Injury vs
Artériographie Findings, No.
Quadrants*
No. of Patients
Normal
Med/ant-med/post
26
21
5
Med/ant-lat/ant
32
19
13
Med/ant-lat/post
34
28
6
Med/ant-no exit
113
94
19
Med/post-lat/ant
24
13
11
Med/post-lat/post Med/post-no exit Lat/ant-lat/post
15
10
5
29
26
3
56
44
12
40
31
9
373
291
82
Lat/ant-no exit Lat/post-no exit Total *Ant indicates anterior;
Eighty-two patients (22%) had an arterial injury demon¬ strated arteriographically (Table 3). The quadrants traversed by the penetrating object and the incidence of artériographie injuries for each combination are found in Table 1. The inci¬ dence of a positive arteriogram for given injury or abnormal in finding is found in Table 2. The mean and median extremities with an arterial injury were 0.81 and 0.92, respec¬ tively. In extremities in which an arterial injury was not was 0.94 and the median, demonstrated, the mean 0.99. The sensitivity/specificity of the in predicting an arterial injury was 82%/40% at less than or equal to 0.99, 71%/63% at less than or equal to 0.95, and 49%/76% at less than or equal to 0.90. In the 216 patients with abnormal findings, 65 arterial injuries were found, of which 22 were major and 43 were minor. Nineteen injuries in this cohort required arterial re¬ pair or embolization. Proximity alone was the indication for arteriography in 157 patients. In this group, 17 arterial inju¬ ries (four major and 13 minor) were demonstrated. One in¬ jured artery underwent ligation (Table 3). Overall, 20 arterial injuries were treated by repair or arté¬ riographie embolization. The arterial repair was end-to-end in 11 and by vein interposition in four. Embolization was per¬ formed for four injuries and arterial ligation for one. All other injuries were treated nonoperatively. The major artery inju¬ ries treated nonoperatively included intimai defects (n 7), occlusions (two profunda femoris, one profunda brachii) (n 3), pseudoaneurysm (n 1), and stenosis (n 1). Thirtytwo minor artery occlusions, five intimai defects, five intimai flaps, five small pseudoaneurysms, and three small arteriovenous fistulae were treated nonoperatively (Table 3). Following arterial repair, all patients were discharged with a patent repair and intact vascular examination. Embolizations were successful in all instances. No untoward events occurred in those patients treated nonoperatively and all were discharged with a stable vascular examination and in¬ tact limb. Twenty-one patients treated nonoperatively have been fol¬ lowed up from 1 to 8 months (median, 2 months) after injury and remain clinically stable. Ten patients have undergone repeated arteriography at 4 days to 2 months after injury to reassess five intimai injuries, three arteriovenous fistulae, and two pseudoaneurysms. All injuries had resolved with the exception of an intimai defect in the radial artery, which was unchanged, and a persistent medial geniculate arteriovenous fistula, which was embolized. For the stepwise logistic regression analysis, the following variables were initially considered for the model: age, sex, =
Table 1 .—Quadrants Traversed by Penetrating
Injury (%)
(19) (40) (18) (16) (46) (33) (10) (21) (22)
Lat, lateral; Med, medial; and Post, posterior.
Table 2.—Incidence of Arterial Injury for Each Abnormal Finding No. of Patients
Injury Type or Arterial
Injury, No. (%)
Injury
Stab wound
76
(16) (46) 57 (22) 1 (8) 25 (26) 21 (28)
27
14 3
Motor/sensory
13 22
Glove All
Shotgun
79 24
Gunshot Other
256
Fracture Hematoma
95
Neurologic deficit Sensory Motor
13
13 11
8
63
25
Pulse deficit Soft-tissue Injury
38 19
22 4
Hemorrhage history
10
2
Decreased refill
3
Hypotension
2 1
Bruit
(52) (2) (36) (40) (56) (21) (20) (38) (50) (100)
=
injury type, extremity, side, distal pulse deficit, fracture, major soft-tissue injury, hematoma, neurologic deficit, bruit, history of hemorrhage, hypotension, and decreased capillary refill. Those variables that occurred in less than 10 patients, ie, bruit, history of hemorrhage, hypotension, and decreased capillary refill, were subsequently not entered. A pulse defi¬ cit, neurologic deficit, and shotgun injury were shown to correlate significantly (P
