Historically, microsurgical reconstruction of the lower extremity has been associated with failure rates that are higher than for other areas of the body, such as breast, head and neck, and upper extremity.1–3 However, recent advances in the fields of both orthopedic and plastic surgery have led to increased rates of limb salvage in patients with severe lower extremity injuries.4,5 In particular, free tissue transfer has revolutionized the management of even routine open fractures of the lower leg with soft-tissue deficits. Crucial to the success of any microsurgical procedure is the identification of suitable recipient vessels.4–6 In many respects, this is perhaps the most challenging portion of a lower extremity free flap reconstruction, even by the most experienced of microsurgeons.6,7 Regardless of the method, preoperative identification of arterial injury with traditional or computed tomographic angiography is essential to identify a vascular injury and facilitate the design and execution of a successful reconstructive plan.8
When initially the classification of open tibia-fibula fractures was described by Gustilo in his landmark articles in 1976 and 1984, he correctly recognized that patients with ischemic limbs requiring emergent revascularization (type IIIC) had the worst prognoses of all patients requiring limb salvage.9,10 Unfortunately, though common, the Gustilo classification is most suited for orthopedic surgeons, never being intended to guide reconstructive treatment. His classification makes no attempt to stratify outcomes for patients with open fractures and concomitant non–limb-threatening arterial injuries. Recent data on lower extremity salvage have demonstrated that the presence of arterial trauma does indeed indicate a more severe injury.5 Compared with three-vessel limbs, the risk of unplanned return to the operating room and total flap failure was increased among both two-vessel and single-vessel legs (the risk of flap failure increased 1.6 and 2.2 times for two-vessel and one-vessel limbs, respectively).5
This has prompted a call to substratify the Gustilo type IIIB category with the proposed 3-2-1 modifier, whereby the number of patent vessels is reported in a patient with a type IIIB open tibia-fibula fracture, thus improving its reconstructive value when evaluating patients.5 When injured arteries are identified preoperatively in patients requiring lower extremity free tissue transfer, a wide array of treatment options are indeed possible: end-to-end or end-to-side anastomosis can be performed on either the injured vessel in question or to an adjacent uninjured vessel. This study aims to compare the outcomes of these different treatment methods based on the number of injured vessels identified and determine whether there is an optimal microsurgical treatment algorithm for patients with concomitant arterial injuries in the setting of Gustilo type IIIB open fractures of the tibia-fibula.
PATIENTS AND METHODS
A retrospective review was performed of a prospectively maintained database of all free tissue transfers performed at our institution from 1976 to 2016. Flaps were performed at three affiliated sites: a private university hospital, a Veterans Health Administration hospital, and a public trauma center. This study was approved by the institutional review board at our medical center. Initial query of the registry identified a total of 806 flaps performed for lower extremity reconstruction. Patients who underwent vascularized osteocutaneous flaps, those with injuries extending above the knee, or those with incomplete information were excluded, yielding a total of 481 patients. Within this group, to reduce the heterogeneity among patients for comparison of microvascular treatments, only Gustilo type IIIB fractures with complete lower extremity imaging that underwent soft-tissue free flap coverage were compared (n = 331). Patients with Gustilo type IIIA and type IIIC injuries were excluded from final analysis, as were any patients with incomplete vascular imaging.
Charts were reviewed for demographics, mechanism of injury, location of injury (i.e., proximal, middle, or distal third of the leg), timing of reconstruction, Gustilo classification, limb vascular status, flap type, recipient vessel site, anastomosis type, and perioperative outcomes. The primary outcome of interest was total flap failure, and secondary outcomes were partial flap failure and unplanned return to the operating room. Partial flap failure was defined as that requiring an additional surgical procedure related to wound breakdown or need for flap débridement during the first 3 months after free flap coverage. Total flap failure was defined as flap compromise requiring complete débridement during the index hospitalization. Major perioperative complications were defined as any flap loss, unplanned return to the operating room, amputation, death, or major perioperative cardiovascular event.
Statistical analyses were performed using Microsoft Excel (Microsoft Corp., Redmond, Wash.) and GraphPad (GraphPad Software, Inc., La Jolla, Calif.). Categorical values were assessed using Fisher’s exact test and continuous variables by the t test. Patients were grouped based on the number of patent arteries (in keeping with the 3-2-1 modification of lower extremity injuries) seen on preoperative imaging studies. Within each group, patients were further stratified based on the type of microsurgical anastomosis performed. Variables, including our outcomes of interest, were compared according to treatment within a specific arterial injury pattern group. A value of p < 0.05 was deemed statistically significant.
Overall, 481 patients were identified who had below-knee lower extremity trauma requiring reconstruction with a nonosteocutaneous free flap. Applying the original Gustilo classification, there were 40 patients with type IIIA injuries (8.3 percent), 409 patients with type IIIB injuries (85.1 percent), and 32 patients (6.6 percent) with an ischemic distal limb (type IIIC injury) (Table 1). The majority of patients were male [n = 363 (75.3 percent)], and the average age was 36.3 years at injury. The most common mechanism of injury was a motor vehicle accident [n = 105 (21.7 percent)]. Most patients [n = 368 (76.3 percent)] were treated at the public hospital (Table 1). In accordance with traditional teachings regarding lower extremity reconstruction, the distal third of the leg was the most common site of injury requiring a free flap [n = 311 (64.5 percent)] (Table 1).
Myocutaneous flaps [n = 362 (75.1 percent)] were more common than fasciocutaneous flaps in this cohort, and this relationship held true over the 40-year duration of these results. The breakdown of timing from injury to coverage was as follows: less than 7 days, 28.1 percent; 8 to 30 days, 22.8 percent; 31 to 90 days, 17.8 percent; and more than 90 days, 31.3 percent. Similarly, the time to coverage remained relatively constant over the course of this study, with slightly more flaps being performed in the acute and subacute periods over the time course of these results. Preoperative angiography was performed in 363 patients (75.5 percent). Of those, three-vessel runoff was demonstrated in 224 patients (61.8 percent), two-vessel runoff was demonstrated in 68 patients (18.7 percent), and single-vessel runoff was demonstrated in 39 patients (10.7 percent). The remaining 8.8 percent of patients (n = 32) had no below-knee runoff (type IIIC injuries) (Table 1). Among all flaps, there were 71 unplanned returns to the operating room (14.7 percent), 37 flaps that suffered total failure (7.7 percent), and 45 flaps that suffered a partial failure (9.3 percent) (Table 1). The total major perioperative complications numbered 111 (23.1 percent). The calculated flap salvage rate was 47.8 percent.
Gustilo Type IIIB-3 Patients (Three-Vessel Runoff)
There were 224 patients with Gustilo type IIIB-3 injuries and no identified vascular injury. Treatment options included end-to-end or end-to-side anastomosis into an uninjured vessel. There were 96 patients (42.8 percent) treated with an end-to-end anastomosis and the remaining 128 (57.2 percent) treated with an end-to-side anastomosis (Table 2). Consistent with prior published reports, complication rates in this group were fairly low, with generally good outcomes.5 There were no major differences between the two treatment options with respect to average patient age, flap type, mechanism of injury, need for vein grafts, or intraoperative anastomosis revision rate (Table 2). There were 14 major complications in the end-to-end group (14.6 percent), including 10 unplanned returns to the operating room (10.4 percent), four instances of partial flap loss (4.2 percent), and two total flap losses (2.1 percent). In the end-to-side group, there were 13 unplanned returns to the operating room (10.2 percent), nine partial flap failures (7.0 percent), and eight total flap failures (6.2 percent) observed. There was no statistically significant difference between these two groups with respect to overall major complications, unplanned return to the operating room, partial flap loss, or total flap loss (Table 2).
Gustilo IIIB-2 Patients (Two-Vessel Runoff)
Based on preoperative imaging, there were 68 patients identified with Gustilo type IIIB-2 injuries (two-vessel runoff below the knee). Three potential treatment options were analyzed for this patient population: (1) end-to-end anastomosis on an injured artery proximal to the zone of injury, (2) end-to-end anastomosis to an adjacent uninjured artery, or (3) end-to-side anastomosis onto an adjacent uninjured artery. There were 23 flaps (33.8 percent) with anastomoses performed end-to-end on an injured artery proximal to the zone of injury; three flaps (4.4 percent) with anastomoses performed end-to-end to an adjacent uninjured artery; and the majority, 42 flaps (61.8 percent), treated with anastomosis performed end-to-side onto an adjacent uninjured artery (Table 3). There were no major differences between the two treatment options with respect to patient age, flap type, mechanism of injury, or need for intraoperative anastomosis revision (Table 2). There was an increased use of vein grafts among the third group of patients who underwent end-to-side anastomosis into an uninjured vessel when compared to the other two groups (7.1 percent; p = 0.001) (Table 3).
In the group of flaps with end-to-end anastomosis performed on an injured vessel proximal to the zone of injury, there were six major perioperative complications (26.1 percent), three unplanned returns to the operating room (13.0 percent), four instances of partial flap failure (17.4 percent), and no complete flap losses (Table 3). In the end-to-end anastomosis into an uninjured adjacent vessel group, there were two major perioperative complications (66 percent): one unplanned return to the operating room (33 percent) and one total flap loss (33 percent) (Table 3). Finally, in the end-to-side anastomosis into an adjacent vessel group, there were 11 major perioperative complications (26.2 percent), four unplanned returns to the operating room (9.6 percent), seven instances of partial flap loss (16.6 percent), and two total flap losses (4.7 percent). There was no statistically significant difference between any of the three treatment options with respect to overall complications, unplanned return to the operating room, or flap loss rate (both partial or complete) (Table 3).
Gustilo Type IIIB-1 Patients (Single-Vessel Runoff)
There were a total of 39 patients with Gustilo type IIIB-1 injuries (single-vessel runoff below the knee) that underwent reconstruction. Treatment options for this group included an end-to-end anastomosis into an injured vessel proximal to the zone of injury or an end-to-side anastomosis into the remaining patent artery. Of these patients, 22 (56.4 percent) had end-to-end anastomoses performed on an injured vessel proximal to the zone of injury, whereas 17 (43.5 percent) had end-to-side anastomoses performed on the uninjured vessel (Table 4). There were no major differences between the two treatment options with respect to average patient age, type of flap, mechanism of injury, need for vein grafts, or intraoperative anastomosis revision (Table 4).
There were seven major complications in the end-to-end anastomosis group (31.8 percent), including five unplanned returns to the operating room (22.7 percent), two instances of partial flap loss (9.1 percent), and three total flap losses (13.6 percent). In the end-to-side anastomosis group, there were eight major perioperative complications (29.4 percent) observed, with five unplanned returns to the operating room (29.4 percent), three partial flap failures (17.6 percent), and five total flap failures (29.4 percent) (Table 4). There were no statistically significant differences between these groups with respect to overall complications, unplanned return to the operating room, partial flap loss, or total flap loss. However, there was a trend toward increased rates of unplanned take-backs and partial and total flap loss in the end-to-side group (Table 4).
The lower extremity flap loss rate has remained high compared with other types of microsurgical reconstruction, where failure rates of 5 percent or less are commonplace.3,5,11–13 One prominent issue in assessing and performing reconstruction in patients with lower extremity injuries is the use of an archaic classification system, the Gustilo scale, as a proxy for injury severity.14,15 Unfortunately, though common, the Gustilo classification is most suited for orthopedic surgeons, never being intended to guide reconstructive treatment. When Gustilo first introduced his grading system, microsurgical techniques were in their infancy and vascular reconstruction was reserved only for a devascularized limb (type IIIC).9,10 This classification scheme does not account for the presence of any arterial injury in an otherwise grossly perfused limb—directly resulting in heterogeneity among type IIIB injuries in terms of severity. Recent data have demonstrated that complication rates increase with decreasing number of patent vessels.5 Stranix et al. reported that when compared to three-vessel limbs, the relative risk of flap failure was increased 1.6 and 2.2 times for two-vessel and one-vessel limbs, respectively.5 Other reconstructive groups have noted this clinical discrepancy as well, although none have previously fleshed it out completely.16
Our similar institutional experience has led to the adoption of the evidence-based 3-2-1 modification of the Gustilo system, denoting the number of intact arteries perfusing the lower leg. However, simply using this modified classification system is not enough; its true power lies in the ability to better define which treatment options are superior for each group of patients. To date, there is no consensus on the optimal microsurgical anastomosis or vessel to use for lower extremity injuries.17–19 Many surgeons use either an end-to-end or end-to-side anastomosis based on nothing more than how they were trained or personal preferences. Current data suggest that neither is superior to the other.17–19 The 3-2-1 modification allows for a more precise categorization of these injuries and enables us to determine whether there are indeed superior (or inferior) treatment options for each subgroup based on the arterial inflow. To that end, we aimed to compare all potential treatment options based on preoperative vascular status (Fig. 1).
In type IIIB-3 injuries, all options for arterial inflow are available without increased risk of flap complications or significant vascular compromise to the limb (Fig. 1). Consistent with previous reports, neither end-to-side nor end-to-end anastomoses were superior in terms of major perioperative complications, take-backs, or flap loss (partial or total) (Table 2). However, there was a nonsignificant trend toward increased partial (7.0 percent versus 4.2 percent) and complete flap loss (6.2 percent versus 2.1 percent) in the end-to-side group, likely related to overall familiarity with end-to-end anastomoses among most microsurgeons and the increased ease of execution (Table 2).
For patients with type IIIB-2 injuries, there are several considerations to be made. First, surgeons may choose to use an injured recipient vessel proximal to the zone of injury as an end-to-end recipient to avoid compromise to the remaining flow to the leg; however, the theoretical concern is a potential increase in flap complications. To minimize this issue, surgeons may perform end-to-end anastomoses to an uninjured vessel, at the risk of compromising the perfusion to an already damaged leg. Given these concerns, surgeons often shy away from this option; which proved true in our data set, where only three of the 68 type IIIB-2 patients received this treatment. A third treatment option in type IIIB-2 injuries, and the most common in this group (61.8 percent), involves an end-to-side anastomosis on one of the remaining vessels, allowing for the preservation of distal perfusion and avoidance of the need to use an injured recipient (Fig. 1).
Importantly, the concern for increased complications among patients undergoing end-to-end anastomoses into injured vessels was not demonstrated in our patient cohort; in fact, similar rates of major perioperative complications (26.1 percent versus 26.2 percent) and partial flap losses (17.4 percent versus 16.6 percent) were observed when compared to patients undergoing end-to-side anastomoses. Surprisingly, a higher (though not statistically significant) increase in total flap loss (4.7 percent versus 0.0 percent) was seen with the end-to-side group, perhaps related to surgeons’ lack of experience with this type of anastomosis. The lack of increased complications in the end-to-end group into a previously injured vessel means that these procedures are safe to perform, provided the surgical team makes an accurate assessment of the zone of injury and cuts these injured vessels back to the point where they appear healthy and have good inflow, even if vein grafts are needed.
Patients suffering type IIIB-1 injuries remain the most challenging group, given the single-vessel inflow into the leg, with an end-to-end anastomosis proximal to the zone of injury on an injured vessel or an end-to-side anastomosis to the single remaining patent vessel. As with type IIIB-2 injuries, the theoretically increased risk of complications associated with using an injured recipient vessel for an end-to-end anastomosis must be weighed against the potential for creating an ischemic limb by performing an end-to-side anastomosis on the only remaining patent vessel providing distal limb perfusion (Fig. 1). Rates of complications and flap failures were higher in this group than in the type IIIB-2 group, but interestingly, a similar pattern of outcomes was seen. When compared with end-to-end anastomosis into an injured vessel, patients treated with end-to-side anastomoses demonstrated similar rates of major perioperative complications (29.4 percent versus 31.8 percent), with a nonsignificant trend toward higher rates of partial (17.6 percent versus 9.1 percent) and total flap loss (29.4 percent versus 13.6 percent). In this type IIIB-1 patient group particularly, the trend toward increased complications in the end-to-side group could be related to surgeons’ lack of experience with this type of anastomosis, but more likely is attributable to the technical difficulty with performing such anastomoses in single-vessel legs.
Despite its large size, the retrospective nature of our study comes with some limitations. Primarily, we focused on major complications and perioperative outcomes within our cohort, as our patient retention for long-term follow-up was unfortunately poor. In addition, microsurgical techniques and wound management strategies have evolved over the 40-year period encompassed by our free flap registry, likely leading to an impact on our results. Interestingly, as a result of the long time span of this project, we had a significant proportion of our patients who underwent delayed reconstruction (>90 days). Many of these reconstructions came at the beginning of our series, in the infancy of microsurgery (before Godina’s work), and some of the patients in this series underwent attempts at failed local tissue or local flap options, leading to a delayed reconstruction. Finally, many of the patients treated at our hospital are recent immigrants or are indigent, often leading to a delay in the time to flap coverage. Lastly, this research is further limited by the small number of patients within each of the treatment groups, requiring larger prospective projects to be conducted in the future. Taking these limitations into consideration, however, this study still constitutes perhaps the largest series of soft-tissue free flap coverage of the lower extremity to provide significant insight into the effect of vascular injury and subsequent treatment options on lower extremity reconstruction.
Historically, the Gustilo classification served as an indication for lower extremity injury severity; however, it should be amended with the 3-2-1 modification to identify the number of patent vessels in the leg after injury. Analysis of treatment patterns of microsurgical reconstruction based on the 3-2-1 modification did not demonstrate any techniques to be superior to others. Importantly, performing an end-to-end anastomosis to an injured vessel outside of the zone of injury in both type IIIB-2 and IIIB-1 injury patterns was demonstrated to be a feasible option for some patients, with complication rates comparable to and in some cases lower than those with an end-to-side anastomosis. In addition, consistent with historical teaching, end-to-side anastomoses can be safely and reliably performed to uninjured vessels regardless of the number of patent vessels in the leg.
1. Medina ND, Fischer JP, Fosnot J, Serletti JM, Wu LC, Kovach SJ III. Lower extremity free flap outcomes using an anastomotic venous coupler device. Ann Plast Surg. 2014;72:176–179.
2. Harashina T. Analysis of 200 free flaps. Br J Plast Surg. 1988;41:33–36.
3. Ricci JA, Crawford K, Ho OA, Lee BT, Patel KM, Iorio ML. Practical guidelines for venous thromboembolism prophylaxis in free tissue transfer. Plast Reconstr Surg. 2016;138:1120–1131.
4. Stranix JT, Lee ZH, Anzai L, et al. Optimizing venous outflow in reconstruction of Gustilo IIIB lower extremity traumas with soft tissue free flap coverage: Are two veins better than one? Microsurgery 2018;38:745–751.
5. Stranix JT, Lee ZH, Jacoby A, et al. Not all Gustilo type IIIB fractures are created equal: Arterial injury impacts limb salvage outcomes. Plast Reconstr Surg. 2017;140:1033–1041.
6. Duymaz A, Karabekmez FE, Vrtiska TJ, Mardini S, Moran SL. Free tissue transfer for lower extremity reconstruction: A study of the role of computed angiography in the planning of free tissue transfer in the posttraumatic setting. Plast Reconstr Surg. 2009;124:523–529.
7. Chen HC, Chuang CC, Chen S, Hsu WM, Wei FC. Selection of recipient vessels for free flaps to the distal leg and foot following trauma. Microsurgery 1994;15:358–363.
8. Haddock NT, Weichman KE, Reformat DD, Kligman BE, Levine JP, Saadeh PB. Lower extremity arterial injury patterns and reconstructive outcomes in patients with severe lower extremity trauma: A 26-year review. J Am Coll Surg. 2010;210:66–72.
9. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: Retrospective and prospective analyses. J Bone Joint Surg Am. 1976;58:453–458.
10. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: A new classification of type III open fractures. J Trauma 1984;24:742–746.
11. Ricci J, Lin SJ. Reply: Comparing the outcomes of different agents to treat vasospasm at microsurgical anastomosis during the papaverine shortage. Plast Reconstr Surg. 2017;140:229e–231e.
12. Ricci JA, Koolen PG, Shah J, et al. Comparing the outcomes of different agents to treat vasospasm at microsurgical anastomosis during the papaverine shortage. Plast Reconstr Surg. 2016;138:401e–408e.
13. Khouri RK, Shaw WW. Reconstruction of the lower extremity with microvascular free flaps: A 10-year experience with 304 consecutive cases. J Trauma 1989;29:1086–1094.
14. Lachica RD. Evidence-based medicine: Management of acute lower extremity trauma. Plast Reconstr Surg. 2017;139:287e–301e.
15. Medina ND, Kovach SJ III, Levin LS. An evidence-based approach to lower extremity acute trauma. Plast Reconstr Surg. 2011;127:926–931.
16. Chummun S, Wigglesworth TA, Young K, et al. Does vascular injury affect the outcome of open tibial fractures? Plast Reconstr Surg. 2013;131:303–309.
17. Ahmadi I, Herle P, Miller G, Hunter-Smith DJ, Leong J, Rozen WM. End-to-end versus end-to-side microvascular anastomosis: A meta-analysis of free flap outcomes. J Reconstr Microsurg. 2017;33:402–411.
18. Cho EH, Garcia RM, Blau J, et al. Microvascular anastomoses using end-to-end versus end-to-side technique in lower extremity free tissue transfer. J Reconstr Microsurg. 2016;32:114–120.
19. Dotson RJ, Bishop AT, Wood MB, Schroeder A. End-to-end versus end-to-side arterial anastomosis patency in microvascular surgery. Microsurgery 1998;18:125–128.