Secondary Logo

Journal Logo


Hard signs gone soft: A critical evaluation of presenting signs of extremity vascular injury

Romagnoli, Anna Noel MD; DuBose, Joseph MD; Dua, Anahita MBChB; Betzold, Richard MD; Bee, Tiffany MD; Fabian, Timothy MD; Morrison, Jonathan MD; Skarupa, David MD; Podbielski, Jeanette MD; Inaba, Kenji MD; Feliciano, David MD; Kauvar, David MD; AAST PROOVIT Study Group

Author Information
Journal of Trauma and Acute Care Surgery: January 2021 - Volume 90 - Issue 1 - p 1-10
doi: 10.1097/TA.0000000000002958


The management of extremity arterial trauma has been evolving for the past 150 years. The days of near-universal arterial ligation and subsequent amputation have been replaced by successful complex vascular limb salvage. Modern diagnostic modalities, advances in open surgical revascularization techniques, and the emergence of endovascular intervention have broadened the treatment options for extremity arterial injury. While the management of these injuries has progressed, the decades-old dogma of “hard” and “soft” signs of vascular injury remains gospel and constitutes the unchanged foundation of bedside vascular evaluation of the injured limb.

Indications for exploration of a presumed arterial injury were reported by Sinkler and Spencer1 in 1960 and included distal arterial insufficiency, expanding hematoma, active bleeding, pulsatile bleeding, and location of wound. Hard and soft signs of vascular injury appeared in the lexicon in the 1984 second edition of Rutherford’s Vascular Surgery in a chapter by Snyder et al.2 as a bifurcation of these indications, without reference to prior origin. In 1992, Frykberg et al.3 described hard signs as absent distal pulse, distal ischemia, active hemorrhage, large hematoma, bruit, or thrill and stated that it was widely agreed that these should prompt immediate surgical exploration. They further described soft signs as history of hemorrhage, small hematoma, hypotension, or deficit of associated peripheral nerve and concluded that they had no clinically useful predictive value for the presence of arterial injury requiring surgical intervention.3 There seem to be little to no objective data driving the distinction between, and clinical use of hard and soft signs. Despite decades of diagnostic and surgical advances, we continue to rely on descriptive findings from the distant past to evaluate patients for extremity arterial injuries. Shortly after his description of indications for surgical intervention, in 1962, Spencer4 reported that hemorrhagic signs of peripheral arterial injury (pulsatile bleeding and expanding hematoma) had a 100% and 75% correlation, respectively, with major vascular injury, while an absent pulse was associated with vascular injury 76% of the time; however, this observation did not gain traction as a significant diagnostic differentiation.

The distinction between hard and soft signs of vascular injury is founded upon little more than observation, and these signs do not suggest or imply anything other than additional workup; they have little to no value in planning injury management. In the current era of ubiquitous multidetector computed tomography imaging in trauma, the use of decades-old “signs” to select limbs for additional imaging is likely obsolete. The objective of this analysis was to explore the distinction between hemorrhagic and ischemic signs of extremity vascular injury as the first step toward defining a modern, more clinically relevant screening paradigm for the workup of a patient with suspected extremity vascular injury.


The PROspective Observational Vascular Injury Treatment (PROOVIT) registry is a multicenter, nationwide vascular injury registry sponsored by the American Association for the Surgery of Trauma (AAST). After obtaining institutional review board approval, enrolled trauma centers submit data directly to the PROOVIT study via an AAST portal. Approval for this analysis was granted by the PROOVIT study review panel, and deidentified data for admissions between September 1, 2012, and November 11, 2019, were used as this was when accrued data were last reported.

Patients sustaining an arterial injury to the upper extremity (subclavian, axillary, brachial, radial, or ulnar arteries) or lower extremity (common femoral, superficial femoral, popliteal, or tibial arteries) were included and divided into two groups, those presenting with hard signs (hemorrhage, expanding hematoma, or ischemia, as defined in PROOVIT data collection) and those presenting without hard signs of vascular injury. Soft signs of vascular injury included in the PROOVIT database were limited to wound proximity, reduced pulses, and fracture/dislocation pattern. Management and outcomes were analyzed, specifically with regard to preoperative imaging with computed tomography angiography (CTA) and definitive management with open repair (OR), or endovascular or hybrid repairs (EHRs). The data were subsequently regrouped into hemorrhagic signs (hemorrhage, expanding hematoma) and ischemic signs (absent or diminished pulses) of vascular injury, and reanalyzed. Patients were included in the hemorrhagic signs group if they had any hemorrhagic sign, regardless of concurrent presence of ischemic sign. Primary outcomes were hospital length of stay, amputation rate, reintervention rate, packed red blood cell (pRBC) transfusion, and in-hospital mortality.

Data were collected using a spreadsheet program (Excel; Microsoft, Redmond, WA), and analyses were performed using GraphPad Prism 8.4.3 (GraphPad Software, La Jolla, CA). Categorical data are reported as frequencies and percentages and compared using χ2 statistics. Continuous variables were assessed for normality with the Shapiro-Wilk test. Nonnormally distributed data are reported as medians with interquartile ranges, and comparisons were performed with the Mann-Whitney U test. Normally distributed continuous data are reported as means with SDs, and comparisons were performed with Student’s t tests. If data points were missing, they were excluded from that calculation and the denominator was reduced. p Values of <0.05 were considered statistically significant.


Data were collected from 25 participating centers, all of which had American College of Surgeons and/or state trauma center designations. Median number of limbs submitted to the database per institution was 33 (12.5, 146.5). There were a total of 1,910 limbs with an extremity arterial injury included in the study. Of these, 1,108 (58%) had hard signs of vascular injury and 802 (42%) did not. When evaluated on a center-by-center level, some predominantly low-volume institutions demonstrated different proportions of hard and absent hard signs in comparison with the entire study population (Supplemental Digital Content, Supplementary Table 1,

Further review of the entire study group demonstrated no difference in median age between the hard and absent hard signs groups, but there were differences in sex and injury mechanism. Penetrating injury predominated in the hard signs group, while blunt injury was more common in patients without hard signs. Between these two groups, there was no difference in Injury Severity Score (ISS), admission systolic blood pressure (SBP), admission pH, or admission international normalized ratio (INR) (Table 1). Extremity Abbreviated Injury Scale (AIS) was slightly higher in the hard signs group, as was admission lactate. Admission hemoglobin and platelet counts were lower in the hard signs group (Table 1).

TABLE 1 - Demographics
Hard Signs Absent Hard Signs p
Demographics 1,108 802
 Male, n [%] 928 83.75% 587 73.19% <0.0001
 Age, median y 31 [23, 44] 31 [23, 47] 0.1731
Injury type
 Blunt 305 27.53% 289 36.03% <0.0001
 Penetrating 745 67.24% 350 43.64% <0.0001
 Mixed 48 4.33% 17 2.12% 0.01
 Not specified 1 0.09% 146 18.20% <0.0001
Injury parameters
 ISS 10 [9, 17] 10 [6, 18] 0.2525
 Extremity AIS 3 [2, 3] 3 [2, 3] 0.0049
 Admission SBP 123 [101, 142] 125 [108, 141] 0.0638
 Admission Hgb 12 [11,14] 13 [11, 14] <0.0001
 Admission pH 7.31 [7.23, 7.38] 7.32 [7.24, 7.38] 0.1453
 Admission lactate 3.5 [2.2, 6.1] 3.2 [2, 5] 0.0036
 Admission INR 1.1 [1, 1.3] 1.1 [1, 1.2] 0.6159
 Admission platelets 223 [178, 277] 231.5 [187, 275.5] 0.0268
Hgb, hemoglobin.

When compared with patients without hard signs, those with hard signs had a higher rate of arterial transection, while the rates of occlusion and pseudoaneurysm were lower. There was no difference between the two groups with regard to partial transection/flow limiting defect (Table 2). In patients presenting with hard signs of vascular injury, CTA alone was performed for diagnosis in 180 (16.2%), while operative exploration alone was performed in 721 (65.1%). In the group of 802 patients who presented without hard signs of vascular injury, 231 (28.8%) underwent CTA alone and 242 (30.2%) underwent operative exploration alone, both with p values of <0.0001.

TABLE 2 - Injury Pathology
Hard Signs Absent Hard Signs p
n 1,108 830
Transection 721 65.07% 221 26.63% <0.0001
Occlusion 114 10.29% 119 14.34% 0.0073
Partial occlusion/flow limiting lesion 126 11.37% 118 14.22% 0.0719
Pseudoaneurysm 21 1.90% 32 3.86% 0.0109

There was no difference in hospital length of stay between the hard signs and non–hard signs groups. Patients with hard signs received more units of pRBC in the first 24 hours and throughout the duration of their hospitalizations. Hard signs patients had a higher in-hospital mortality rate, amputation, and reintervention rate (Table 3). Among the entire cohort, 10.7% of patients who underwent CTA also underwent EHR, while only 1.5% of patients who underwent operative exploration for diagnosis underwent EHR. Patients with hard signs of vascular injury who underwent exploration without imaging had an OR rate of 68% compared with EHR rate of 1.4%.

TABLE 3 - Outcomes
Hard Signs Absent Hard Signs p
n 1,108 830
Need for reoperation/reintervention 125 11.28% 44 5.30% <0.0001
Need for amputation during hospitalization 132 11.91% 27 3.25% <0.0001
Mortality 54 4.87% 21 2.53% 0.0086
pRBC first 24 h 2 [0, 6] 0 [0, 3] <0.0001
pRBC total 0 [1, 4] 0 [0, 2] <0.0001
Hospital LOS 7 [3, 16] 7 [3, 15] 0.1732
LOS, length of stay.

The same data set was then reanalyzed to compare presence of hemorrhagic signs (n = 915) and presence of ischemic signs (n = 490). Patients presenting with both hemorrhagic and ischemic signs were grouped with the hemorrhagic signs group (570 of 915 patients) for the purpose of the analysis, as the bleeding extremity with concurrent shock is an overriding clinical concern. The majority (403, 70.7%) of these patients had reduced pulses, while 167 (29.3%) had frank ischemia. All patients with hemorrhagic signs presented with traditional hard signs of vascular injury. Of 490 patients with ischemic signs, 174 (35%) presented with traditional hard signs of vascular injury, while the remainder 316 (64.5%) did not. When evaluated on participating center level, some predominantly low-volume institutions demonstrated different proportions of hemorrhagic and ischemic signs in comparison with the entire study population (Supplemental Digital Content, Supplementary Table 1,

There was no difference in sex or age between the hemorrhagic and ischemic signs groups, while patients with hemorrhagic signs were more likely to have incurred penetrating injury compared with patients with ischemic signs. There was no difference in extremity AIS, admission INR, and admission platelets between these two groups. Patients with ischemic signs presented with a slightly higher ISS, SBP, hemoglobin, and pH; however, patients with hemorrhagic signs presented with a higher admission lactate (Table 4).

TABLE 4 - Demographics
Hemorrhagic Signs Ischemic Signs p
Demographics 915 490
 Male, n [%] 759 82.95% 411 83.88% 0.7078
 Age, y 31 [23, 44] 31 [22, 45] 0.27
Injury type
 Blunt 200 21.86% 246 50.20% <0.0001
 Penetrating 674 73.66% 229 46.73% <0.0001
 Mixed 40 4.37% 15 3.06% 0.2509
 Not specified 1 0.11% 0 0.00%
Injury parameters
 ISS 10 [8, 17] 10 [9, 19] 0.0009
Extremity AIS 3 [2, 3] 3 [2, 3] 0.259
 Admission SBP 122 [100, 142] 127 [110, 142.5] 0.003
 Admission Hgb 12 [11, 14] 13 [11, 14] <0.0001
 Admission pH 7.3 [7.22, 7.38] 7.32 [7.24, 7.38] 0.0426
 Admission lactate 3.65 [2.25, 6.3] 3.2 [2.2, 4.5] 0.0005
 Admission INR 1.1 [1, 1.3] 1.1 [1.01, 1.21] 0.1927
 Admission platelets 222 [175, 276] 227 [185, 276] 0.1015
Hgb, hemoglobin.

Of the 915 patients presenting with hemorrhagic signs, 133 (14.54%) underwent CTA only for diagnosis of a vascular injury, and 631 (68.96%) underwent operative exploration alone. Patients who underwent CTA were less likely to have initial operative management of their injury and to undergo OR. They were more likely to undergo EHR (14.29% vs. 1.58%, p = 0.0001). Patients who underwent CTA alone for diagnosis were also more likely to undergo initial observation of their extremity vascular injuries (Table 5). There were also no differences in reoperation/reintervention rates, pRBC in the first 24 hours, pRBC during admission, or amputation rate. Patients with hemorrhagic signs undergoing CTA alone for diagnosis compared with those who underwent operative exploration experienced a lower mortality rate; however, there was a significantly longer hospital length of stay for these patients (Table 6).

TABLE 5 - Type of Repair
Hemorrhagic Signs CTA Only Exploration Only p
n 133 631
Initial operative management 83 62.41% 502 79.56% <0.0001
OR 75 56.39% 495 78.45% <0.0001
EHR 19 14.29% 10 1.58% <0.0001
Initial observation 21 15.79% 6 0.95% <0.0001
Ischemic Signs CTA Only Exploration Only p
n 155 216
Initial operative management 70 45.16% 168 77.78% <0.0001
OR 68 43.87% 165 76.39% <0.0001
EHR 10 6.45% 3 1.39% 0.0182
Initial observation 58 37.42% 7 3.24% <0.0001
Hgb, hemoglobin.

TABLE 6 - Outcomes Hemorrhagic/Ischemic Signs
Hemorrhagic Signs CTA Only Exploration Only p
n 133 631
Need for reoperation/reintervention 12 9.02% 63 9.98% 0.8728
Need for amputation during hospitalization 17 12.78% 65 10.30% 0.7691
Mortality 1 0.75% 37 5.86% 0.0081
pRBC first 24 h 2 [0, 6.5] 2 [0, 5] 0.6584
pRBC total 2 [0, 4] 2 [0, 5] 0.4874
Hospital LOS 10 [5, 18] 4 [2, 10.5] <.0001
Ischemic Signs CTA Only Exploration Only p
n 155 216
Need for reoperation/reintervention 10 6.45% 27 12.50% 0.0778
Need for amputation during hospitalization 9 5.81% 25 11.57% 0.0682
Mortality 2 1.29% 4 1.85% 1
pRBC first 24 h 0 [0, 2] 1 [0, 5] 0.0126
pRBC total 0 [0, 2] 1 [0, 4] 0.0947
Hospital LOS 10.5 [5, 18] 9 [4, 19] 0.5538
LOS, length of stay.

Of the 490 patients presenting with ischemic signs, 155 (31.6%) underwent CTA only for diagnosis of vascular injury, while 216 (44%) underwent operative exploration alone. Patients who underwent CTA were less likely to undergo initial operative management. Patients in the CTA group were less likely to undergo OR and more likely to undergo EHR. They were also more likely to undergo initial observation of their injuries (Table 5). There was no difference in need for reoperation/reintervention, amputation, pRBC during hospitalization, mortality, or hospital length of stay. Patients with ischemic signs undergoing exploration alone for diagnosis received more pRBC during the first 24 hours (Table 6).

Patients with hemorrhagic signs had a higher rate of arterial transection compared with patients with ischemic signs (67.76 vs. 44.29, p < 0.0001), while patients with ischemic signs had a higher rate of occlusive pathology (26.36% vs. 7.32%, p < 0.0001). There was no difference between these two groups regarding partial transection/flow limiting defect or pseudoaneurysm (Table 7).

TABLE 7 - Injury Pathology
Hemorrhagic Signs Ischemic Signs p
n 915 490
Transection 620 67.76% 217 44.29% <0.0001
Occlusion 67 10.81% 129 59.45% <0.0001
Partial occlusion/flow limiting lesion 110 164.18% 69 53.49% 0.2759
Pseudoaneurysm 17 15.45% 12 17.39% 0.4397


Sinkler and Spencer’s1 initially reported indications for surgical exploration were reworked by Perry et al.5 in 1971 to also include history of arterial bleeding, major hemorrhage with hypotension or shock, bruit at or distal to the suspected site of injury, and injury of anatomically related nerves. These findings were considered absolute indications for immediate surgical exploration without preoperative angiography. This practice of routine operative exploration in the setting of penetrating proximity extremity trauma (PPET) lasted into the 1980s.5,6 Subsequent studies examining this practice of mandatory operation reported only a 25% to 50% positive exploration rate.7,8

Recognition that the combination of findings on physical examination and routine operative exploration yielded suboptimal results prompted the reconsideration of the benefit of preoperative arteriography in particular settings. These included blunt trauma, associated fractures, multiple missiles, proximity of missile tract to a major vessel, preexisting vascular insufficiency, and suspected pseudoaneurysm or arteriovenous fistula.7 Subsequently, this so-called exclusion arteriography evolved into the gold standard for evaluation of peripheral arterial injury in hemodynamically normal patients.7 In 1992, the pendulum reversed its course as Frykberg et al.3 reported that, even though PPET patients who underwent exclusion contrast arteriography had a 10% incidence of injury to major arteries, there was only an overall 1.8% incidence of injury that required surgical repair. This low prevalence of surgically relevant injuries in combination with the 1% to 2% reported false-negative rate of angiography at the time (said to be comparable with that of physical examination) led to a widespread decrease in the use of exclusion arteriography.3

In the early 1990s, Doppler indices (ankle brachial indices [ABI], injured extremity indices [IEI]) of <0.9 were demonstrated to have a sensitivity and specificity of 95% and 97%, respectively, for detection of major arterial injuries. In follow-up studies, the authors discontinued screening arteriography for patients with a Doppler index of >0.9, in favor of selective arteriography for patients at risk of occult injury based on a Doppler index of <0.9. They reported missing no major injuries.9 These findings were redemonstrated with subsequent prospective study,10 and it was later shown that asymptomatic PPET patients with a Doppler pressure index of >1.011 and, later, ≥0.912 could safely be discharged home without additional imaging or a period of inpatient observation.

By the early 2000s, CTA emerged as a reliable stand-alone imaging modality for evaluation of peripheral arterial trauma.13 Seamon et al.14 performed a prospective comparison of CTA and subsequent contrast angiography or operative exploration in patients with potential extremity vascular injuries and an ankle brachial index of <0.9. They reported that CTA had a 100% sensitivity and specificity for detection of clinically relevant vascular injuries, as well as substantial cost savings in comparison with contrast angiography.14 This cemented the role of CTA as the new gold standard for the evaluation of the potentially injured extremity.

Currently, it is recommended that patients with hard signs of an arterial injury should undergo operation without need for preoperative angiography, and it is frequently taught that imaging beyond plain films (including CTA) is contraindicated in this setting.15,16 When imaging is required, CTA should be used as the primary diagnostic study for evaluation of a possible extremity arterial injury.15 Patients without hard signs of vascular injury who have an abnormal physical examination (ankle brachial indices, <0.9) should have further evaluation, typically a CTA.15,16

In comparing the hard signs with non–hard signs group, there were differences in injury type, extremity AIS, admission hemoglobin, lactate, and platelet count. Nevertheless, the majority of parameters generally thought of as correlating with depth of shock or physiologic derangement (ISS, admission SBP, admission pH and admission INR) showed no difference. Not surprisingly, the group with hard signs had a higher rate of transection or occlusion. While the presence of hard signs correlates with the likelihood of arterial intervention, it does not permit differentiation between which type of injury has occurred. Patients with hard signs received more units of pRBCs both in the first 24 hours of admission and throughout their hospitalization and, as noted previously, had higher rates of amputation and reintervention, and had a greater in-hospital mortality rate. While the outcome differences certainly support the fact that patients presenting with and without hard signs of vascular injury take different clinical courses, this division does not have much predictive utility as far as the necessity of operative intervention.

Analyzing the clinically relevant distinction between hemorrhagic and ischemic signs of vascular injury, the patients with hemorrhagic signs had predictably lower admission SBP, hemoglobin, and pH, with higher admission lactate. While their ISS was also slightly lower than the ischemic signs group, these data suggest that their degree of physiologic derangement was more severe, likely due to hemorrhage from the injured vessel. In the ischemic group, higher admission SBP, hemoglobin, and pH and lower lactate suggest an overall lesser degree of physiologic derangement despite a higher injury burden resulting from blunt mechanism polytrauma.

Early treatment decisions are based not only on the presence of a vascular injury but also on the vascular pathology and physiology that are involved. Unlike hard signs, which do not suggest vascular pathology, we observed that hemorrhagic signs were more likely to have a vascular transection, while limbs with ischemic signs were more likely to have arterial occlusion. Early initiation of therapeutic anticoagulation is a standard of care for nontraumatic acute limb ischemia,17 has been associated with improved outcomes in select cases of ischemic arterial injury,18 and is recommended at diagnosis of ischemic arterial injury.19 However, a recent survey of surgeons managing peripheral arterial trauma showed minimal consensus regarding heparinization in extremity arterial injury.20 The ability to reliably distinguish between these two presentations at the bedside could facilitate not only diagnostic workup steps but also expedite management and treatment decisions, including therapeutic anticoagulation, which may have significant effect on limb salvage.

The limbs with hemorrhagic signs of vascular injury were more likely to undergo operative exploration alone (68.96%) versus CTA alone (14.54%). While this could certainly be attributable to continued hemodynamic instability due to severity of injury, or perceived severity of injury, it might also suggest that preoperative CTA allowed for consideration of alternative management strategies, reflected by a significant increase in EHR in the CTA group (14.29% vs. 1.58%, p = 0.0001) and higher rate of initial observation of injury. Despite pursuing this diagnostic study before proceeding to the operating room, there was no difference in pRBC requirement overall, pRBC in first 24 hours, reintervention rate, or amputation rate. The lower mortality rate observed in patients with hemorrhagic signs undergoing CTA (0.75% vs. 5.86%, p = 0.0081) is likely reflective of the overall injury burden and hemodynamic stability of patients taken for operative exploration but unrelated to the decision to obtain CTA.

A larger percentage of limbs presenting with ischemic signs underwent CTA in comparison with those presenting with hemorrhagic signs (31.6% vs. 14.54%). Computed tomography angiography can provide important anatomic information relevant to the planned vascular reconstruction, demonstrating the value of the designation “ischemic signs” in the planning of the imaging workup. Hard signs are not useful for this purpose because their presence only dictates operative exploration. The operative exploration arm of the ischemic signs group, compared with the CTA only group, did receive significantly more pRBCs during the first 24 hours (1 [0, 5] vs. 0 [0, 2], p = 0.0126). This suggests that CTA and the preoperative planning it affords may reduce intraoperative and immediate postoperative transfusion of blood products.

The ischemic signs group demonstrated the same trend as mentioned previously, with a lower rate of initial operative management and OR as well as a higher rate of EHR than the hemorrhagic signs group (6.45% vs. 1.39%, p = 0.0182). As was the case with exclusion angiography in the past, CTAs will identify lesions that are not surgically significant.3,21 This suggests that patients undergoing CTA may be overtreated for injuries that might resolve without intervention, such as vasospasm, intimal flap, and small pseudoaneurysms.22,23 In these patients, there is a small risk of injury progression. Late hemorrhage or ischemia may occur, or asymptomatic progression may be detected on follow-up imaging. Because there is no universally accepted grading system for extremity arterial injury, prescribed protocol for serial imaging, or evidence-based management strategies for these injuries, it is challenging to predict who is at risk for progression and failure and who may benefit from early intervention.

Most limbs presenting with hemorrhagic signs of vascular injury sustained penetrating trauma. Management of limb hemorrhage in the field or trauma bay may include direct manual pressure, tourniquet, or resuscitative endovascular balloon occlusion of the aorta. These potentially lifesaving maneuvers alter or occlude blood flow through the zone of injury, limiting the utility of CTA. In the setting of continued hemodynamic instability, urgent operative exploration with proximal and distal vascular control should be expeditiously pursued (Fig. 1). If, however, the patient can be stabilized and extremity hemorrhage can be controlled without disrupting blood flow, preprocedural CTA should be strongly considered. These preoperative images are invaluable in identifying the extent of the vascular lesion, anatomic variants, inflow and outflow vessels, and potential targets for endovascular intervention. In contrast to hemorrhagic signs, the urgency of hemostasis measures and early interventions are not dictated by the presence of hard signs of vascular injury.

Figure 1
Figure 1:
Suspected extremity vascular injury algorithm. *Consider tourniquet loosening or removal prior to CTA. **Consider preoperative systemic heparinization in isolated extremity injuries with low soft tissue injury burden (i.e., posterior knee dislocation, midshaft humeral fracture).

Limbs presenting with ischemic signs were more likely to have sustained blunt trauma. While associated skeletal injury (fractures, dislocations) can help localize the area of arterial injury, the vascular pathology can be occult and far more extensive than the soft tissue defect might suggest. Reconstructions of these injuries are often extensive, lengthy, and involve multidisciplinary teams. Preoperative CTA is critical in this scenario for both operative planning and sequencing (fasciotomy, shunt, harvest of conduit, fracture fixation, definitive repair) and is suggested by the presence of ischemic signs. Once more, hard signs (even pulselessness) do not suggest anything other than immediate operative exploration. In the multiply injured patient with blunt trauma and an ischemic extremity, continued hemodynamic instability after initial resuscitation should again be considered an indication for immediate exploration and preservation of life over limb (Fig. 1). When life-threatening injuries have been managed and preemptive fasciotomies performed, if the vascular pathology remains unclear, intraoperative angiography or postoperative CTA should be used as diagnostic tools. For patients with hemorrhagic and/or ischemic signs of vascular injury, the importance of emergent operative exploration in the presence of continued hemodynamic instability cannot be understated.

To pursue CTA or immediate operative exploration of a limb with a suspected extremity vascular injury represents the primary decision in most trauma evaluations performed today. Transfer of the patient to a hybrid operating room with advanced fluoroscopic imaging capabilities is emerging in some centers as an additional option, and in the setting of a dedicated surgical endovascular trauma service, has been demonstrated to decrease time to hemostasis.24 In contrast to the Japanese hybrid emergency room system,25 however, there is a paucity of data regarding outcomes and full utilization of the hybrid capability in the North American model, with a recent article citing only 18% utilization of hybrid capabilities in their dedicated operating room.26 Modern fixed imaging systems found in hybrid operating rooms can perform standard angiography, digital subtraction angiography, and cone beam computed tomography. Cone beam computed tomography can be used for imaging intracranial as well as maxillofacial injuries, blunt cerebrovascular injury, blunt thoracic aortic injuries, solid organ, pelvic injuries, and extremity vascular injuries.27 It is important, however, to note that the hybrid room itself is only one part of the equation; without around-the-clock access to physicians and technologists capable of operating the equipment, it is of dubious benefit. Reliance on out-of-hospital interventionalist teams after hours and on weekends has been demonstrated to be associated with both increased time to intervention and increased mortality in trauma patients.28 When such imaging systems, physicians, and technologists exist in an operating room easily accessible from the initial trauma bay, the hybrid imaging and open and endovascular treatment capabilities available in one place can expand the treatment options for injured extremities.

While this article specifically explored the validity of hard and soft signs of vascular injury in extremities, this terminology and paradigm, although not identical, are used to describe penetrating injuries to the neck. The management of penetrating neck trauma has evolved significantly over the last decades, with CTA, bronchoscopy, and esophagogastroduodenoscopy safely preceding operative exploration in most cases.29 Given the invasive nature of zone 1 operative repair and technical challenge of zone 3 repair, preoperative CTA in patients without signs of hemodynamic instability or airway compromise facilitates endovascular intervention. Signs of ischemia resulting from a cervical vascular injury represent a stroke in evolution and require emergent neurorescue maneuvers. In nontraumatic stroke, advanced neuroimaging (CTA or magnetic resonance angiography [MRA]) is considered a critical component in determining which patients are candidates for endovascular therapy and is routinely obtained while observing a door-to-needle target time of 30 minutes.30 The presenting characteristics of traumatic vascular injury of the neck should be further studied because it is unlikely that the results reported here can be directly extrapolated to the workup of cervical injuries.

The PROOVIT data included in this study are the result of voluntary contributions from 25 centers across the United States, which represents a small fraction of institutions participating in trauma care and submitting data to the National Trauma Data Bank. All institutions participating have American College of Surgeons and/or state trauma center designation. Consequently, the results reported may not be indicative of practice patterns in the country at large. Ability to obtain CTA in an expeditious fashion and to perform endovascular interventions at any time of day may not be representative of all institutions managing trauma patients. While these data are more granular and specific to vascular injury than that collected by larger databases like the National Trauma Data Bank, there continue to be significant inherent limitations. PROspective Observational Vascular Injury Treatment data are received via an AAST portal in a deidentified fashion. Definitive determination of which data entries represent multiple extremity injuries in the same patient was not possible, making the database susceptible to patient-level clustering. The multi-institutional nature of this database is also a potential source of cluster bias. When evaluated on an institutional level, there were outlier centers regarding proportions in both the hard/absent hard signs analysis and the hemorrhagic/ischemic signs analysis. While the variability of the lower volume institutions could be secondary to less experience in the diagnosis and management of extremity trauma, there are other clinically relevant variables not accounted for in the PROOVIT database that are potential sources of institutional disparity, that is, transport time from the scene, CTA timing, time to the operating room, time to intervention, specific surgical/interventional team variation (vascular, interventional radiology, trauma), and concurrent procedures performed, among others. Because of the observational nature of this study and the inability to further parse out factors, which directly contribute to the clinical course and outcome of the patients, adjusted statistical analysis to account for institutional clustering was therefore not performed.

Regardless of admission markers of physiologic derangement included in the database, it is likely that patients who proceeded directly to the operating room showed signs of extremis not captured by the database, while those who underwent CTA did not. Because presence of vascular injury is a requirement for inclusion in the database, there are no adequate data to describe sensitivity, specificity, or predictive value of hemorrhagic and ischemic signs.


We have demonstrated that hard signs of vascular injury have significant limitations in the identification and characterization of extremity arterial injuries. In the era of readily available CTA imaging, using hard and soft signs as a decision-making strategy is outdated. These signs do not provide useful clinical distinctions, and a strategy of using hemorrhagic and ischemic signs of vascular injury is of far greater clinical utility. Further prospective study is needed to validate this proposed redefinition of presentations of extremity arterial injury.


A.N.R., J.D., A.D., and D.K. were responsible for initial planning. R.B., T.B., T.F., J.M., D.S., J.P., K.I., D.F., J.D., and D.K. were involved in database design and data acquisition. J.D. was responsible for database maintenance. A.N.R. drafted the article. A.D., R.B., T.B., T.F., J.M., J.P., K.I., D.F., J.D., and D.K. provided expertise and advice, and editing. A.N.R. and D.K. take responsibility for the content of the article.


We thank the AAST PROOVIT Study Group: John Sharpe, MD, Tiffany Bee, MD, Timothy Fabian, MD, University of Tennessee Health Sciences Center—Memphis, Memphis, TN; Jonny Morrison, MD, PhD, David Feliciano, MD; Thomas M. Scalea, MD, University of Maryland, R Adams Cowley Shock Trauma Center, Baltimore, MD; David Skarupa, MD; Jennifer A. Mull, RN, CCRC; Yohan Diaz Zuniga, MD, University of Florida—Jacksonville, Jacksonville, FL; Jeanette M. Podbielski, RN, CCRP, Garrett Jost, University of Texas Health Sciences Center—Houston, Houston, TX; Richard D. Catalano, MD, Ahmed M. Abou-Zamzam Jr., MD, Xian Luo-Owen, PhD, Loma Linda University Medical Center, Loma Linda, CA; Jennie Kim, MD; Kenji Inaba, MD, Los Angeles County + University of Southern California Hospital, Los Angeles, CA; Nathaniel Poulin, MD, East Carolina Medical Center, Greenville, NC; John Myers, MD; Michael Johnson, MD; Kristin Rocchi, RN, The University of Texas Health Sciences Center at San Antonio, San Antonio, TX; John K. Bini, MD; Joshua Pringle, MD; Karen Herzing, BSN, RN; Kailey Nolan, BS, Wright State Research Institute, Miami Valley Hospital, Dayton, OH; Ramyar Gilani, MD; Tikesha Smith; Reginva Knight, Ben Taub General Hospital/Baylor College of Medicine, Houston, TX; Peter Hammer, MD, Indiana University School of Medicine, Indianapolis, IN; Nicholas Namias, MD, MBA, Ryder Trauma Center, University of Miami/Jackson Memorial, Miami, FL; Juan Asensio, MD; Creighton University School of Medicine, Omaha, NE; Joseph M. Galante, MD; Misty Humphries, MD, University of California—Davis, Sacramento, CA; Ravi R. Rajani, MD, Grady Memorial Hospital, Emory University School of Medicine, Atlanta, GA; George Dulabon, MD; Riyad Karmy-Jones, Peace Health Southwest Washington Medical Center, Vancouver, WA; Andreas Larentzakis, MD; George Velmahos, MD, Massachusetts General Hospital, Boston, MA; Suresh Agarwal, MD, University of Wisconsin, Madison, WI; Jayraan Badiee, MPH; Michael Sise, MD, Scripps Mercy Hospital, San Diego, CA; Alan Cook, MD; Annette Taylor, RN, BS, CCRC; John Zumbuhi, RN, MSHA, CCRC, CRCP; Antonia Greigo, CCRC, Chandler Regional Medical Center, Chandler, AZ; Fausto Y. Vinces, DO; Salvatore Docimo, DO, Lutheran Medical Center, Brooklyn, NY; Matthew M. Carrick, MD; Kathy Rodkey, RCIS, CCRC, Medical City Plano, Plano, TX; Sameer Hirji, MD; Reza Askari, MD, Brigham and Women’s Hospital, Boston, MA; Forrest O. Moore, MD; Richard Butler, MD, John Peter Smith Hospital, Fort Worth, TX. Also, special thanks to David Feliciano for continued mentorship and access to his files.


The authors declare no conflicts of interest.


1. Sinkler WH, Spencer AD. The value of peripheral arteriography in assessing acute vascular injuries. Arch Surg. 1960;80:300–304.
2. Snyder W, Thal E, Perry M. Peripheral and abdominal vascular injuries. In: Rutherford RB, ed. Vascular Surgery. Philadelphia, PA: W.B. Saunders Company; 1984:462.
3. Frykberg ER. Arteriography of the injured extremity: are we in proximity to an answer? J Trauma. 1992;32(5):551–552.
4. Spencer AD. The reliability of signs of peripheral vascular injury. Surg Gynecol Obstet. 1962;114:490–494.
5. Perry MO, Thal ER, Shires GT. Management of arterial injuries. Ann Surg. 1971;173(3):403–408.
6. Fomon JJ, Warren WD. Late complications of peripheral arterial injuries. Arch Surg. 1965;91(4):610–616.
7. Sirinek KR, Levine BA, Gaskill HV 3rd, Root HD. Reassessment of the role of routine operative exploration in vascular trauma. J Trauma. 1981;21(5):339–344.
8. Turcotte JK, Towne JB, Bernhard VM. Is arteriography necessary in the management of vascular trauma of the extremities? Surgery. 1978;84(4):557–562.
9. Johansen K, Lynch K, Paun M, Copass M. Non-invasive vascular tests reliably exclude occult arterial trauma in injured extremities. J Trauma. 1991;31(4):515–519; discussion 519–522.
10. Nassoura ZE, Ivatury RR, Simon RJ, Jabbour N, Vinzons A, Stahl W. A reassessment of Doppler pressure indices in the detection of arterial lesions in proximity penetrating injuries of extremities: a prospective study. Am J Emerg Med. 1996;14(2):151–156.
11. Conrad MF, Patton JH Jr., Parikshak M, Kralovich KA. Evaluation of vascular injury in penetrating extremity trauma: angiographers stay home. Am Surg. 2002;68(3):269–274.
12. Sadjadi J, Cureton EL, Dozier KC, Kwan RO, Victorino GP. Expedited treatment of lower extremity gunshot wounds. J Am Coll Surg. 2009;209(6):740–745.
13. Wallin D, Yaghoubian A, Rosing D, Walot I, Chauvapun J, de Virgilio C. Computed tomographic angiography as the primary diagnostic modality in penetrating lower extremity vascular injuries: a level I trauma experience. Ann Vasc Surg. 2011;25(5):620–623.
14. Seamon MJ, Smoger D, Torres DM, Pathak AS, Gaughan JP, Santora TA, Cohen G, Goldberg AJ. A prospective validation of a current practice: the detection of extremity vascular injury with CT angiography. J Trauma. 2009;67(2):238–243; discussion 243-4.
15. Fox N, Rajani RR, Bokhari F, Chiu WC, Kerwin A, Seamon MJ, Skarupa D, Frykberg E; Eastern Association for the Surgery of Trauma. Evaluation and management of penetrating lower extremity arterial trauma: an Eastern Association for the Surgery of Trauma practice management guideline. J Trauma Acute Care Surg. 2012;73(5 Suppl 4):S315–S320.
16. Feliciano DV, Moore FA, Moore EE, West MA, Davis JW, Cocanour CS, Kozar RA, McIntyre RC Jr. Evaluation and management of peripheral vascular injury. Part 1. Western Trauma Association/critical decisions in trauma. J Trauma. 2011;70(6):1551–1556.
17. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG; TASC II Working Group. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). J Vasc Surg. 2007;45(Suppl S):S5–S67.
18. Daugherty ME, Sachatello CR, Ernst CB. Improved treatment of popliteal arterial injuries using anticoagulation and extra-anatomic reconstruction. Arch Surg. 1978;113(11):1317–1321.
19. Feliciano DV, Moore EE, West MA, Moore FA, Davis JW, Cocanour CS, Scalea TM, McIntyre RC. Western Trauma Association critical decisions in trauma: evaluation and management of peripheral vascular injury, part II. J Trauma Acute Care Surg. 2013;75(3):391–397.
20. Romagnoli A, DuBose J, Feliciano D. Through thick or thin: disparities in perioperative anticoagulant use in trauma patients. Am Surg. 2019;85(9):1040–1043.
21. Gonzalez RP, Falimirski ME. The utility of physical examination in proximity penetrating extremity trauma. Am Surg. 1999;65(8):784–789.
22. Stain SC, Yellin AE, Weaver FA, Pentecost MJ. Selective management of nonocclusive arterial injuries. Arch Surg. 1989;124(10):1136–1140; discussion 1140-1.
23. Dennis JW, Frykberg ER, Veldenz HC, Huffman S, Menawat SS. Validation of nonoperative management of occult vascular injuries and accuracy of physical examination alone in penetrating extremity trauma: 5- to 10-year follow-up. J Trauma. 1998;44(2):243–252; discussion 242-3.
24. Morrison JJ, Madurska MJ, Romagnoli A, et al. A surgical endovascular trauma service increases case volume and decreases time to hemostasis. Ann Surg. 2019;270(4):612–619.
25. Founding members of the Japanese Association for Hybrid Emergency Room System (JA‐HERS). The hybrid emergency room system: a novel trauma evaluation and care system created in Japan. Acute Med Surg. 2019;6(3):247–251.
26. Carver D, Kirkpatrick AW, D’Amours S, Hameed SM, Beveridge J, Ball CG. A prospective evaluation of the utility of a hybrid operating suite for severely injured patients: overstated or underutilized? Ann Surg. 2020;271(5):958–961.
27. Gupta S, Martinson JR, Ricaurte D, Scalea TM, Morrison JJ. Cone-beam computed tomography for trauma. J Trauma Acute Care Surg. 2020;89:e34–e40.
28. Schwartz DA, Medina M, Cotton BA, Rahbar E, Wade CE, Cohen AM, Beeler AM, Burgess AR, Holcomb JB. Are we delivering two standards of care for pelvic trauma? Availability of angioembolization after hours and on weekends increases time to therapeutic intervention. J Trauma Acute Care Surg. 2014;76(1):134–139.
29. Sperry JL, Moore EE, Coimbra R, et al. Western Trauma Association critical decisions in trauma: penetrating neck trauma. J Trauma Acute Care Surg. 2013;75(6):936–940.
30. Zerna C, Thomalla G, Campbell BCV, Rha JH, Hill MD. Current practice and future directions in the diagnosis and acute treatment of ischaemic stroke. Lancet. 2018;392(10154):1247–1256.


ENRIQUE GINZBURG, M.D. (Miami, Florida): I would like to thank the AAST Program Committee for this opportunity to discuss “Hard Signs Gone Soft”, a critical evaluation of presenting signs of extremity vascular injury, a multicenter AAST PROOVIT registry-based study.

As the authors expressed clearly in the paper, the objective of this study was for the distinction between hemorrhagic and ischemia signs of extremity vascular injury as the first step towards defining a modern, more clinically-relevant, screening paradigm for the workup of a patient with a suspected extremity vascular injury.

Hard and soft signs and their definition serve well in an age where cumbersome angiography and open exploration were the only options in diagnosis and management of vascular injury, until the adjunct of complex ultrasonography as demonstrated in our institution for neck trauma, followed in the mid-90s with the use of ABIs and then, obviously, the use of CT angio.

Fast forward now to 2020 and this study redefines hard and soft signs in an attempt to compare the utility in the age of CTA and endovascular techniques to the new classification of hemorrhagic and ischemic signs.

In this paper hard signs were redefined as hemorrhage, expanding hematoma, and ischemia, basically a la Frykberg, Eric Frykberg, may he rest in peace. We also included (indistinguishable) as well as just large hematomas.

Soft signs were redefined in this paper as proximity injury, reduced pulses, fracture/dislocation (indistinguishable). They compared this to the new redefined classification of hemorrhagic signs as bleeding and expanded hematoma and ischemic signs comprised of absent or reduced pulses.

My first question in light of the above, can you please define “reduced pulses”? Is it, one, unequal decreased pulses by physical exam, which is operator dependent? Or, two, ABI less than .9? And/or three, only Doppler-able signals?

In reality the advent of CTA and endovascular procedures, substituting the use of open surgery, is the real (indistinguishable) to advocating new nomenclature that directs management algorithms.

As the study points out, the majority of hemorrhagic patients and approximately half of ischemic patients underwent surgery; However, 14.5 percent of hemorrhagic patients went to CT angio.

My second question, why would a patient who is bleeding or have an expanding hematoma and continued risk of bleeding while going in and obtaining CT angio have the CT angio be perfomed?

And, am I correct to assume from this, that these patients were all hematomas? And if not, which I doubt, highly, what method was employed to control bleeding or prevent bleeding without obstructing distal flow during the process of obtaining the CT angio?

According to the study, 31 percent of patients with ischemia underwent CT angio and 44 percent went directly to the OR. The only outcomes difference between these groups were in packed red blood cells units transfused, supporting the use of CT angio as not having increased morbidity and mortality in the delay of intervention.

The difference between those that went to CT angio versus surgery in this group was the amount of endovascular repairs being performed – 6.45 percent in CT angio versus 1.5 percent in open cases – so my third question is shouldn’t this dictate that all trauma services should have a hybrid room with concomitant vascular or interventional radiologists who are available to potentially increase the rate of endovascular repair for both the hemorrhagic and the ischemia groups?

In light of the above, would it be more useful to be more granular in the ischemia group and reanalyze (glitch) to see if there were differences in outcomes between patients with absent pulses versus reduced pulses and potentially change the definitions back to: 1. Hard, bleeding, expanding large hematoma (glitch) pulses; or soft, those with reduced pulses or hemorrhagic with bleeding, expanding large hematomas, and absent pulses to the ischemic just being reduced pulses?

The obvious – excuse me, the last question that I have, where do you categorize patients with posterior displaced tibial plateau fractures and dislocations or patients with a (indistinguishable) where none of the current signs include them?

Should we exclude them until they develop either ischemia changes or hemodynamic alterations associated with increased morbidity and potential limb loss?

The obvious and most compelling finding in this study demonstrates that hemodynamic stability and instability define the management algorithm, whether you use hard, soft, hemorrhagic, or ischemia signs.

Overall, I commend the author and the coauthors for a very interesting and comprehensive and potentially paradigm changing reclassification of suspected vascular injury and decision making. It’s easy to remember, especially for residents and aging surgeons.

ANNA ROMAGNOLI, M.D. (Boston, Massachusetts): Thank you for your comments, sir. And I’d like to thank the AAST for the opportunity to present this work.

In response to your first question, the reduced pulses group included in the ischemic signs of vascular injury analysis included both reduced pulses by physical exam, ABI and by Doppler. We included all of these in order to maximize catchment of this poorly defined group. While the PROOVIT database did differentiate between patients who had ABIs and Doppler exams performed, the database is not granular enough to be able to determine complete absence of signal, phasicity of signal and number of run-off vessels assessed. Consequently, any report of decreased pulses based on physical exam, Doppler or ABI was included.

The PROOVIT database, while more granular than larger databases like the National Trauma Databank (NTDB), does not differentiate between history of bleeding or current active bleeding. Patients with hemorrhagic signs were included in that group for analysis, however the temporal element of their hemorrhagic sign is unknown. As a result we are unable to really comment on the individual decision making of the receiving trauma surgeons regarding imaging patients with hemorrhagic signs. Regardless of whether it is reflected in the captured datapoints, it can be presumed that patients who underwent CTA prior to exploration were thought to be stable enough to do so. There is likely also an element of institutional variability at play, as some institutions are able to obtain a scan on the way to the operating room with minimal delay, while others are not. Both of these elements are certainly potential sources of bias in our dataset.

Regarding injuries with a high association of sometimes occult vascular trauma including posterior knee dislocations, tibial plateau fractures, and midshaft humeral fractures; these warrant additional study and likely should be included in a separate arm of the flow sheet. Even if these patients present without ischemic signs, it is likely prudent to obtain CTA in these scenarios. A missed injury of this sort in a young patient without collateral vessel development or ischemic preconditioning of the extremity can be catastrophic. Additionally, due to the often small soft-tissue injury burden associated with these injuries, it may be appropriate to initiate systemic heparinization prior to proceeding to the operating room, which could significantly affect limb salvage rates.

Holcomb’s group demonstrated several years ago that there were two different standards of care being delivered, as well as a difference in mortality rate, when patients with pelvic fracture requiring angioembolization presented during business hour or after hours. As the Endovascular Trauma Service continues to mature at Shock Trauma, it has completely changed the way the institution conceptualizes and manages complex polytrauma. The ability to concomitantly diagnose and treat vascular injuries at any hour of the day without relying on a consultant service has certainly upped our game. So yes, I think the future of trauma management involves a hybrid room and an Endovascular Trauma Service. As Shock Trauma continues to demonstrate success with this model, I hope it will become more widespread.

Thank you again for your time and attention.


Extremity arterial injury; hard signs; hemorrhagic signs; ischemic signs; vascular injury

Supplemental Digital Content

Copyright © 2020 American Association for the Surgery of Trauma.