Blunt popliteal artery injury following tibiofemoral trauma: vessel-first and bone-first strategy

All surgically treated BPAI patients ≥ 18 years old with concomitant tibiofemoral dislocations and fractures, admitted from the emergency department at two level 1 trauma centers between January 2000 until January 2019, were included in this study. The institutional research patient data registry (RPDR) was queried for corresponding International Classification of Disease (ICD) 9 and 10 codes for popliteal artery injury and tibiofemoral dislocations and fractures. Patients younger than 18 years old, those with incomplete medical records, those with penetrating popliteal artery injury, and those that did not undergo revascularization (e.g., because of mangled extremities that required primary amputation) were excluded.

Demographic and clinical patient characteristics were extracted from medical records (Table 1). Knee dislocations were graded based on the Schenck classification, and tibial plateau fractures were graded based on the Schatzker classification, both using available radiological images and reports and clinical data [16, 17]. BPAI grade was determined based on computed tomography angiography (CTA) imaging and radiology reports. Lower leg ischemia was determined according to the Rutherford grading scale for acute limb ischemia, using clinical notes [18]. Due to the limitations of retrospectively reviewing clinical notes for determining the Rutherford sub classification [marginally threatened (IIa) and immediately threatened (IIb)], we limited classification of acute extremity ischemia to three grades: not immediately threatened (I), threatened (II, combining the sub classification), and irreversible (III). Total delay duration was defined as time between the reported time of injury and operating room arrival time. In-hospital delay duration was defined as time between the initial encounter noted in the emergency department records and operating room arrival time, including patients that received revascularization at a different hospital. Infectious complications were defined as surgical site infections or osteosynthesis-related infections. Thrombo-embolic complications were defined as new pulmonary embolisms or deep venous thrombo-embolisms detected on imaging while admitted to hospital. The decision to either perform an initial vascular repair or initial external fixation was made according to preference of the attending surgeon, based on the total injury burden, extent of associated orthopedic injuries, and the total delay duration. In case of isolated extremity injury in the absence of severe associated injuries and in case of polytrauma patients requiring a damage-control approach, a VF strategy was preferred. If severe associated orthopedic injury or knee instability was present, a BF strategy was preferred if allowed by the total ischemia duration. To illustrate potential differences between patients that were transferred from scene versus those that were transported from other hospitals, a comparison between these patients is displayed in the supplementary materials Tables 1, 2 and 3.

All out-patient encounters of all identified patients were reviewed, in order to determine postdischarge outcomes. Time to follow-up was defined as time between hospital discharge and most recent out-patient visit, rather than a standardized follow-up moment. Vascular re-interventions were defined as all surgical or endovascular procedures performed to restore vascular patency. Claudication was defined as (intermittent) claudication complaints of the affected leg, excluding those with vascular re-interventions.

Data were analyzed using STATA^® statistical software (StataCorp. 2017. Release 15. College Station, TX: StataCorp LLC.) Baseline and clinical patient characteristics were, in the case of non-parametrical distributed categorical explanatory variables and dichotomous outcome variables, analyzed using the two-sided Chi-square test or two-sided Fisher’s exact test. In the case of dichotomous explanatory variables and continuous outcome variables, the Mann–Whitney U test was used. In the case of multiple non-parametric categorical explanatory variables and continuous outcome variables, the Spearman-Rank correlation was used. A p value of < 0.05 was considered significant.

Twenty-seven patients with BPAI that met the inclusion criteria were identified and included. One patient with a penetrating popliteal artery injury was excluded. Three patients that had BPAI and did not undergo arterial repair were excluded. Two of these patients had mangled extremities that were deemed unsalvageable, requiring primary amputation. One patient was transferred from a different hospital approximately 17 days after the initial knee injury, latent ischemia caused by a previously undiagnosed popliteal artery dissection resulted in primary amputation. Baseline characteristics of all 27 included patients, the 15 patients that received primary arterial repair first (VF, 56%), and the 12 patients that received primary external fixation first (BF, 44%) are depicted in Table 1. Median (IQR) body mass index (BMI) was 36 (28–48) and the most common mechanism of injury was a low-energy fall from height less than 2 m in eight cases (30%) followed by crush injuries in seven cases (26%). As 5/27 patients were polytrauma patients (19%) and median (IQR) ISS was 9 (4–10), most patients suffered from isolated lower extremity injuries. Sixteen patients initially presented at a different hospital (59%), of which two received vascular repair at that hospital prior to transfer to our institution (7%). All patients showed signs of lower leg ischemia, of which most legs were threatened (67%). Twenty-three patients had knee dislocations (85%), of which seven (26%) had Schenck grade V injuries (multiligamentous injury with peri-articular fracture). Nine patients had associated tibial plateau fractures of which most had Schatzker type 4 fractures (n = 4), followed by Schatzker type 6 fractures (n = 3). The most common popliteal artery injuries were occlusions in 67% of cases. Injury mechanisms and associated injuries are depicted in Table 2.

Median (IQR) total delay duration in the VF group was 5.5 (4.0–8.5) [95% confidence interval (CI) 4.0–9.9] h and median (IQR) total delay duration in the BF group was 6.8 (4.0–7.5) (95% CI 3.8–7.8) h. Median (IQR) in-hospital delay duration in the VF group was 2.6 (2.0–2.9) (95% CI 2.0–3.0) h and median (IQR) in-hospital delay duration in the BF group was 2.7 (1.0–3.0) (95% CI 0.8–3.2) h. Both total delay duration and in-hospital delay duration were not significantly different between groups (p = 1.00 and p = 0.85), as is depicted in Table 1. Median (IQR) total delay duration was higher in those that initially presented at different hospitals compared to those that initially presented at the studied institutions [7.5 (5.7–10.0) vs. 4.0 (3.8–4.6) hours; p < 0.01] which is depicted in supplementary materials, Table 1.

Eight patients in the VF group had isolated ligamentous injuries (53%), compared to six in the BF group (50%), which was not different between groups (p = 0.86). Five patients in the VF group had tibial plateau fractures (33%), compared to four in the BF group (33%), which did not differ between groups (p = 1.00), as is depicted in Table 2. The most commonly observed Schatzker type of tibial plateau fracture in the VF group was type 3 (n = 3), and in the BF group, this was type 6 (n = 2) (p = 0.35). No primary amputations were observed. The only observed secondary amputation was observed in the BF group. This patient initially had successful revascularization approximately 7.5 h after the initial injury and was discharged after 5 days. During follow-up, the patient suffered from an iatrogenic injury to the popliteal artery bypass during meniscal surgery and ligament reconstruction at a different hospital. This resulted in multiple vascular re-interventions to restore patency, after which eventually an above the knee amputation was performed due to chronic pain and knee instability (approximately 5.5 years after initial presentation). Because this secondary amputation was the result of an iatrogenic injury rather than the initial management strategy, this patients was excluded for this analysis. No in-hospital mortalities were found. Two patients in the VF group had local infections, compared to one in the BF group (p = 0.54). Twelve patients in the VF group received fasciotomies (80%), compared to 10 in the BF group (83%, p = 0.83). Nineteen out of 22 fasciotomies were prophylactic (86%). No thrombo-embolic complications were observed. No differences in terms of studied outcomes were observed when comparing those that presented at other hospitals to those that initially presented at the studied institutions (supplementary materials, Table 3).

Vascular repair details and outcomes of both groups are depicted in Table 3. In 16 patients, revascularization was achieved by popliteal artery bypass (59%), followed by interposition grafts in nine patients (33%) and endovascular repair in two patients (7%). All popliteal artery bypasses were performed using reversed greater saphenous vein (GSV) grafting. Two interposition grafts were performed using reversed GSV and the rest utilized synthetic graft materials. Intraluminal shunting was used during surgery in two patients (7%), of which one shunt occluded. This patient was treated by initial external stabilization (BF) and intraluminal shunting. Shortly after the initial surgery, shunt occlusion occurred, requiring immediate popliteal artery bypass.

The popliteal artery was the most common proximal anastomosis site (59%) as well as the most common distal anastomosis site (70%). The two observed popliteal vein injuries, a traumatic occlusion and a transection, were treated conservatively and by ligation, respectively. Neither injury resulted in complication.

Twenty-five patients were available for follow-up during out-patient visits (93%). During the median (IQR) follow-up period of 2.7 years (1.1–5.9), three patients required ipsilateral popliteal artery re-interventions to restore vascular patency. One of these patients experienced acute occlusion of the polytetrafluoroethylene (PTFE) interposition graft, while on acetylsalicylic acid (ASA) 325 mg, vascular patency was restored by percutaneous transluminal angioplasty and stent placement. The second patient experienced a proximal stenosis of the GSV bypass, requiring vein patch angioplasty after which vascular patency was restored. The third patient is the previously described patient that experienced an iatrogenic injury to the popliteal artery bypass during a subsequent surgery and then required multiple vascular re-interventions ultimately resulting in amputation due to pain. In total, 20 patients were started on ASA medication (in most cases dosed 325 mg once daily) to prevent thrombus formation and occlusion of the vascular graft (74%). After exclusion of the patients that required vascular re-interventions, no patients experienced claudication complaints at long-term follow-up.

International guidelines regarding the management of lower extremity trauma do not provide unambiguous recommendations whether revascularization or bone stabilization should be performed first in tibiofemoral trauma patients with popliteal artery injury [12, 19]. The Western Trauma Association guideline recommendation regarding management of the mangled extremity states that in hemodynamically stable patients, initial reduction of fractured bones using either splinting or traction fixation will alleviate kinking of the vasculature and improve subsequent perfusion. However, in case of vascular injury in addition to orthopedic injury in patients with isolated extremity injuries or limited associated injuries, no recommendations regarding a VF or BF strategy are made. In case of hemodynamic instability and an extremity which is deemed salvageable, initial use of an intraluminal shunt is recommended [12].

As multiple studies emphasize the influence of total ischemia duration on clinical outcomes, initial revascularization is often preferred [10, 20]. This VF strategy has been associated with good clinical outcomes [21]. The absence of external fixation materials during a VF strategy may provide better popliteal artery exposure during surgery. During a damage-control approach in polytrauma patients, a VF strategy (utilizing intraluminal shunting) may be preferable as it reduces operating time. However, we speculate that initial bone stabilization (BF) may be preferred in patients with severe isolated orthopedic injuries, to prevent traction injuries to the vascular graft and to prevent sharp fracture elements from damaging the vascular graft. Furthermore, tibiofemoral alignment may aid the correct positioning and lengthening of the vascular graft.

In addition to current literature, we were able to compare both the VF and BF strategy in BPAI patients (mostly consisting of isolated lower extremity trauma patients) and determined that a BF strategy could be safely performed in those where primary osseous stabilization was deemed necessary.

Although a few studies have described external fixation before or after popliteal artery repair, the studied patients suffered mostly from penetrating (popliteal) arterial injuries in addition to blunt injuries [22, 23]. Neither specifically studied popliteal artery injuries, since all lower extremity vascular injuries were combined. One of these studies, consisting of 17 lower extremity vascular injuries, made a recommendation regarding the optimal strategy, advocating initial revascularization over orthopedic stabilization [22]. A retrospective study focusing strictly on blunt injuries to the popliteal artery identified multiple factors associated with limb salvage, of which prehospital delay > 6 h was frequent among amputees, but did not study the effects of a VF or BF strategy [24].

Penetrating and blunt injuries to the popliteal artery result in different types of arterial damage (e.g., transection with free extravasation and tourniquet application vs. dissections and near-occlusions), caused by different injury mechanisms (e.g., knife/gunshot wounds vs. mechanical falls/crush injuries) with different associated injuries. Of these, BPAIs are associated with a higher frequency of secondary amputations, due to higher-energy trauma [20]. Due to these differences in both the arterial injury and the associated injuries, the need for a VF or BF may be different. Therefore, the outcomes of these strategies for BPAI patients remained unknown thus far.

Although median (IQR) total delay duration in the BF group of 6.8 h (4.0–7.5) was greater than the 5.5 h (4.0–8.5) in the VF group, total delay duration was not significantly different between groups (p = 1.00). Median (IQR) in-hospital delay in the BF group of 2.7 h (1.0–3.0) and the VF group of 2.6 h (2.0–2.9) was not different, as well (p = 0.82). When taking the total delay duration and population characteristics into account, we conclude that both the VF as well as the BF strategy have comparable good outcomes. We speculate that during treatment of BPAI patients with severe tibiofemoral fractures and dislocations, a BF strategy may be preferred in order to provide stability for the vascular repair. However, caution should be used when extrapolating these results to individual patients with other ischemia durations and total trauma burdens. In case of a longer total ischemic duration, irrespective of associated tibiofemoral injuries, a VF strategy (using intraluminal shunting if necessary) or a strategy where external fixation is preceded by vascular shunting before definite popliteal artery repair may be preferred, as total ischemia duration is a predictor for (secondary) amputation [10, 20].

Most patients received lower leg revascularization by popliteal artery bypass (59%) or interposition graft (33%), as transections and (severe) vessel-wall contusion limited endovascular treatment options. Therefore, these long-term outcomes are applicable to revascularization utilizing these surgical procedures. A total of 25 patients were available for long-term follow-up (93%), with a median (IQR) follow-up of 2.7 years (1.1–5.9). During this period, three vascular re-interventions were performed in these 25 patients, of which one was related to the reverse GSV bypass (4%). Excluding those that had re-interventions, no claudication complaints were observed. A total of 20 patients were started on ASA antiplatelet therapy (74%), mostly dosed 325 mg once daily, which was continued during the follow-up period. Using antiplatelet therapy, including ASA, as postoperative antithrombotic regimen to assure vascular patency, is in accordance with regimens of infrainguinal bypass surgery for (non-traumatic) peripheral arterial disease [25, 26]. Based on these results, we conclude that popliteal artery bypasses and interposition grafts for BPAI, using reversed GSV grafting in combination with antiplatelet therapy, appear to result in good vascular patency after a follow-up period of almost 3 years. Larger studies focusing on complications, re-interventions, and long-term vascular patency should be performed to determine if long-term outcomes are similar to the ones described.

Our study has several limitations. First, this was a retrospective study with a small number of patients. Although our results indicate comparably good outcomes between the VF and BF strategy for BPAI, larger studies are required to determine if (long-term) outcomes are similar to the ones we found. Of course, this is limited by the rare occurrence of these injuries and suggests that a multi-institutional effort may be needed for a larger study. In addition, the retrospective identification of patients during a 19-year period may potentially have resulted in missing cases.

Although no primary amputations were observed in our cohort, literature states that BPAIs are associated with high amputation rates. We speculate that this difference may be explained by undiagnosed BPAIs among the polytrauma patient population that did not survive emergency surgery (e.g., for intra-cranial or abdominal bleeding) and did not receive CTA scanning/peri-operative angiography (due to hemorrhagic shock). Polytrauma patients that required primary amputation of their mangled extremities due to the additional high total trauma burden may have had undiagnosed BPAIs, as well. This form of selection bias/survivorship bias may have resulted in better reported outcomes. Second, as patients received treatment according to the VF or BF strategy by surgeon preference (rather than randomization), selection bias is likely to be present. Although we speculate that patients with more severe orthopedic injuries were treated according to the BF strategy, we were unable to illustrate these differences with statistical significance. These differences may have been present and possibly have affected outcomes. Third, we were unable to accurately determine total ischemia duration, since the operative records did not literally mention the moment in time where an anastomosis was completed, and clamps were released. We, therefore, had to use total delay duration (time of injury—operating room arrival time) as the most accurate substitute for total ischemia duration. Fourth, as clinical outcomes of a VF or BF strategy for BPAI management are highly dependent on total ischemia duration (therefore ischemic damage to the lower leg), our results cannot be seen separate from the observed total delay durations. Caution should be used when extrapolating the results of both strategies to a setting with different (pre-)hospital trauma care logistics.