Institutional Review Board approval and a waiver of informed consent were obtained prior to our review of any data in the patient chart. Between April 2004 and January 2007, sixty-five patients with sixty-seven type-III tibial fractures sustained during Operation Enduring Freedom and Operation Iraqi Freedom were identified. Patients treated to completion with monoplanar external fixation, those treated with conversion of monoplanar external fixation to hybrid external fixation or internal fixation, or those treated definitively with an amputation were not included in our results. The twenty-two patients (twenty-two tibiae) who were excluded are described below.
Fourteen patients had an amputation performed as definitive treatment. One patient with a type-IIIB fracture was treated with open reduction and internal fixation immediately prior to flap coverage. The wound subsequently became infected and fixation was converted to a ring external fixator followed by a below-the-knee amputation for a chronic infected nonunion. One patient with a type-IIIA fracture had a hybrid external fixator placed, and the fracture subsequently healed. Three patients (three tibiae) were treated with an intramedullary nail: one tibia with a type-IIIA fracture became infected and required nail removal for treatment of the infection, one tibia with a type-IIIA fracture healed as a malunion in 15° of valgus, and one tibia with a type-IIIB fracture healed uneventfully. In three patients (three tibiae), the fractures were treated with monoplanar external fixation. Two of the tibiae (one type-IIIA and one type-IIIB fracture) healed uneventfully. The third tibia (a type-IIIC fracture) had development of an infected nonunion that was treated with irrigation and débridement and placement of antibiotic beads, followed by bone-grafting and ring external fixation, and it subsequently healed.
Forty-five tibiae in forty-three patients underwent placement of ring external fixation as definitive management of the injury and are the subject of this review. Of the forty-three patients who were initially treated at our hospital with ring external fixation, seven subsequently completed treatment at other military facilities where different treatment protocols were followed and they were excluded, leaving a total of thirty-eight tibial shaft fractures in thirty-six patients (an 84% rate of follow-up).
Of the seven patients who continued their treatment at other institutions, three went on to achieve union in the circular frame following planned delayed bone-grafting procedures. Two patients had frame removal at six months, and the circular frames were converted to intramedullary nails after a short period of casting; both fractures healed uneventfully. One patient had osteomyelitis develop after the circular frame was removed, and a planned open bone-grafting and plating was performed. He was completing one year of suppressive antibiotic treatment at the time of the latest follow-up. The remaining patient had a presumed culture-negative, infected nonunion and was treated with block resection. At the time of the latest follow-up, he was completing monoplanar external fixator bone transport alongside a plate as salvage for the subsequent bone deficit.
The patients who sustained combat-related injuries in the theater of war were rapidly assessed and evacuated to local medical treatment units close to the battlefield. Initial wound management included gross decontamination of foreign materials and débridement of nonviable tissue followed by copious irrigation and placement of clean wound dressings with splinting. An extensile fasciotomy of the extremity, as well as placement of monoplanar external fixation to stabilize the bone injury and facilitate management of the soft tissues, was performed frequently (Fig. 1).
Wound débridement and irrigation was continued throughout the evacuation chain, ideally at least every forty-eight hours, until the patients arrived in the United States for definitive care. Our institution is the tertiary-care referral center for the majority of war-injured U.S. Marines.
Upon arrival, patients received a thorough evaluation by the trauma surgery service, orthopaedic surgery, and plastic surgery. Vascular surgeons were consulted for previously reported vascular repair or deficient peripheral pulses. On admission, patients were started on broad-spectrum antibiotics. A combination of cefazolin and meropenem was routinely prescribed in our institution to cover gram-positive pathogens and target Acinetobacter baumannii, which was frequently found on surveillance cultures. The extent of osseous and soft-tissue injury was evaluated in the operating room, usually with direct inspection of vascular and nerve continuity made possible by the extensive open wounds and fasciotomies. Additional ipsilateral extremity injury was also evaluated and considered. Aggressive débridement of nonviable muscle and soft tissue was performed sharply. Wound cultures were not typically performed. External fixators were loosened, and the bone ends were delivered out of the wound for curettage. Nonarticular, devitalized bone was discarded regardless of size (Figs. 2-A and 2-B).
Copious irrigation of at least 9 L of normal saline solution per wound was then delivered throughout the wound by pulsatile lavage. For fractures with substantial segmental bone loss, antibiotic-impregnated spacer beads were placed. We typically used one bag of bone cement (Palacos; Zimmer, Warsaw, Indiana) with 1 g of vancomycin and 500 mg of imipenem powder added prior to mixing. Wounds were then managed with the use of negative-pressure wound therapy (V.A.C. Kinetic Concepts, San Antonio, Texas). Balanced overhead elevation of the extremity from the external fixator frame was standard. Over-the-counter ankle-foot orthoses were used to help to prevent ankle equinus contracture. Subsequent to this initial assessment, we held a team meeting to solicit input from the treating services and to encourage discussion with patients and their families regarding limb salvage feasibility. Peers with an amputation and patients who had undergone limb salvage surgery were made available for counsel. Except in the most extreme circumstances, ultimate limb salvage decisions were made by the patient and his family on the basis of data extrapolated from the civilian trauma literature16-20. Patients with a type-IIIC fracture who desired limb salvage and those who required extensive muscle flap coverage were particularly made aware of the challenges of limb salvage21. The majority of patients treated with amputation at our institution were selected because of extensive soft-tissue damage, associated trauma of the ipsilateral foot, unreconstructible nerve injury, and/or a failed vascular repair.
Typically, wound irrigation and débridement continued at forty-eight-hour intervals until the wound was stable and considered clean. No attempt was made to quantitate bacterial load within the wound. If the wounds supported delayed primary closure without undue tension, they were closed over a drain with number-2.0 nylon sutures. Larger wounds requiring skin-grafting or local or free muscle transfer were covered at this time by the plastic surgery service (Fig. 3).
Wounds without associated bone loss that were closed by one author (R.C.A.), during 2006, received rhBMP-2 (INFUSE; Medtronic Sofamor Danek, Memphis, Tennessee) around the fracture site immediately prior to wound closure.
The patients and their wounds continued to be assessed over the course of the following week. If there were no clinical signs of deep wound infection as defined below, the previously placed monoplanar frame was removed and a ring external fixator was applied. We utilized either traditional Ilizarov circular fixators (Fig. 4) or the Taylor Spatial Frame (TSF; Smith and Nephew, Memphis, Tennessee). A combination of 6-mm hydroxyapatite-coated half-pins and fine wires was utilized with at least three points of fixation in each segment. With the exception of limb-lengthening procedures, any residual deformity was acutely corrected with use of the TSF web-based correction program22 in the operating room, on the basis of the immediate postoperative radiographs.
Following frame placement, the patients were encouraged to walk as soon as tolerable with full weight-bearing. Delayed bone-grafting at six weeks following frame placement was performed routinely for bone defects or extensive comminution. Patients continued increasing their activities and were enrolled in aggressive physical therapy programs for ankle and knee range of motion and strengthening. Limbs suspected of deep infection at any point in treatment were evaluated with routine laboratory studies, including a complete blood-cell count with differential, an erythrocyte sedimentation rate, and a C-reactive protein level. At-risk wounds with a workup indicating possible infection were evaluated at the time of surgery, and deep bone and soft-tissue cultures were obtained. Infections not affecting frame stability were treated with débridement and intravenous antibiotics without frame removal. A deep infection was defined with use of the Centers for Disease Control and Prevention publication "Guideline for Prevention of Surgical Site Infection, 1999" for both deep incisional and organ-space (osteomyelitis) infection23.
When osseous consolidation was evident on orthogonal radiographs and the patient exhibited minimal pain to palpation at the fracture site, the frames were dynamized by unlocking the struts or unfastening the threaded rods with the half-pins and the rings left in place. Following a successful two to three-week trial of full activity with no evidence of loss of fracture reduction and no pain at the fracture site with walking, the frame was removed with the patient under sedation. Final radiographic alignment was assessed by measuring the alignment of the long axis of the tibial shaft on standard anteroposterior and lateral orthogonal digital radiographs made following frame removal.
A blast mechanism accounted for thirty-five injuries, while the remaining three were from high-velocity gunshot wounds. There were twenty-one type-IIIA, thirteen type-IIIB, and four type-IIIC fractures (Table I). An average of 6.4 (range, four to ten) irrigation and débridement procedures was performed prior to frame conversion. Rotational flaps were performed on thirteen patients, and two patients had free flaps for soft-tissue reconstruction while two underwent acute osseous shortening with subsequent lengthening to facilitate management of the soft-tissue component of the injury. The average delay to flap coverage from the time of injury was 15.9 days (range, seven to thirty-eight days).
All patients were treated with ring external fixation by either the Ilizarov technique (eight tibiae) or placement of a Taylor Spatial Frame (thirty tibiae). Frames were applied an average of 23.8 days from the time of injury. Forty-eight additional surgical procedures were performed on twenty-five patients (see Appendix). Eighteen of these patients received planned delayed bone-grafting. Nine fractures with minimal or no bone loss received placement of rhBMP-2 at the time of final wound closure, and no bone graft was placed.
Two patients with a type-IIIC fracture required a total of twenty-one additional procedures. The remaining patients had no more than three additional procedures. Of the eight patients receiving unplanned additional procedures (including one who had had planned delayed bone-grafting), two required débridement and subsequent bone-grafting for an infected nonunion under a muscle flap. Another patient developed a wound dehiscence around a muscle flap following planned delayed bone-grafting and subsequently required débridement. One patient underwent a pin revision, and one patient required knee manipulation for a joint contracture. Without removal of the circular frame, one patient with an atrophic nonunion was treated with bone-grafting to achieve union. One hypertrophic nonunion healed following insertion of an intramedullary nail after the frame was removed prematurely, and the fracture angulated with weight-bearing. The remaining patient had a protracted course because of a complete loss of the anterior and lateral muscle compartments following complications associated with reverse saphenous vascular grafting of the popliteal artery soon after injury. Fracture-healing required late bone-grafting and frame revision to gain a successful union. Subsequently, a refracture occurred through a previous distal pin site surrounded by necrotic bone, and the patient chose to have a below-the-knee amputation.
The overall deep infection rate was 8% (three of thirty-eight tibiae) and 3% (one of thirty-four) with type-IIIC fractures excluded. All infections were treated successfully with serial débridement and intravenous antibiotics without removal of the frame. All fractures healed with <5° of coronal or sagittal malalignment. The average time to frame removal, which typically occurred shortly after documentation of clinical and radiographic healing, was 220.8 days (range, 102 to 339 days) (Table I). The average duration of clinical and radiographic follow-up was 493 days (range, 157 to 1013 days).
Numerous authors have concluded that intramedullary nail fixation of type-III tibial fractures is the preferred method of stabilization for patients at a civilian trauma center2-8,24-28. Kakar and Tornetta4 recently reported the results of 143 open tibial fractures treated with protocol-driven wound management and immediate unreamed intramedullary nail fixation, which resulted in minimal complications and a low rate (3%) of deep infection.
Similarly, monoplanar external fixation has been considered for use in tibial fractures characterized by severe soft-tissue injury because of its ease of placement and the preservation of existing blood supplies to the tibia. However, Henley et al.24 and others25-28 have found that use of monoplanar external fixation often leads to higher rates of complications, including malunion, infection, and an increased number of operative interventions, and this type of fixation is generally reserved for temporary stabilization.
Compared with injuries sustained in the civilian trauma setting, wounds sustained on the battlefield resulting from an "outside-in" injury mechanism are considered to be more contaminated29-31. Moreover, the typical early débridement and coverage protocols espoused by Fischer et al.32, which lead to a lower rate of infectious complications, are not always possible because of delays associated with evacuation to definitive treatment facilities during wartime. These realities combined with the common occurrence of extensive retained blast fragments throughout the soft tissues led us to apply circular external fixators as our definitive method of fracture fixation. Similar to many other institutions, we have observed the benefits of a protocol-driven approach to the treatment of these limb-threatening injuries, which includes frequent aggressive soft-tissue and osseous débridement, expeditious soft-tissue coverage, and delayed supplemental bone-grafting when needed3-8.
Utilizing a comparable wound management protocol, Lacap and Frisch33 recently reported a deep infection rate of 14.3% (five tibiae) in a study of thirty-five type-III wartime tibial fractures treated with intramedullary nail fixation at a U.S. military hospital. Four of the five patients required removal of the nail to eradicate the infection. In contrast, Lerner et al.10 reported the results of a staged treatment of sixty-four high-energy open wartime tibial fractures treated with ring fixators in Israel. They reported a 90.6% rate of union and a 1.6% rate of deep infection, and they stated that only one patient required an amputation. Similarly, Atesalp et al.34 described their experience with the Ilizarov technique in the treatment of 163 tibiae with a Gustilo type-IIIA extremity fracture. All fractures healed with adequate alignment, and they had a low rate (3.1%) of deep infection. The most frequently reported complication was pin track infection, which occurred in 50.7% of the patients. Inan et al.35, in a study of sixty-one type-IIIA tibial fractures, compared the radiographic and clinical outcomes of thirty-two patients treated with Ilizarov external fixation and twenty-nine patients treated with unreamed intramedullary nailing. The Ilizarov group had a significantly shorter time to union (nineteen weeks) compared with the group that had intramedullary nail fixation (twenty-one weeks) (p = 0.039). However, the Ilizarov technique carried a notable prevalence of pin track infection and joint contracture. The group that had unreamed intramedullary nailing had a higher rate of deep infection compared with the Ilizarov group (10% and 6.3%, respectively). However, this difference was not significant.
Our protocol-driven use of ring external fixation also yielded acceptable results. We utilized hydroxyapatite-coated tapered half-pins, which have been associated with both a lower prevalence of pin track infection and improved pullout strength36,37. In another study, we reported loosening and revision rates of 6% and 7%, respectively, for 222 consecutive half-pins used at our institution on multiple extremity fractures with risk factors for pin loosening associated with pins placed within the zone of injury or traversing free or local soft-tissue flaps38. We attribute the low rate of malalignment in our patients, compared with previous reports in the literature in which monoplanar external fixation was used, to the control of fracture alignment afforded by the versatility of the Taylor Spatial Frame and the stability afforded by multiplanar fixation. The frame allowed the ability to dynamize without additional operative intervention. Moreover, deep infections and nonunion complications were effectively managed without frame removal.
We consider the infection rate in the present study to be quite low in light of the extent of contamination in these wounds. Furthermore, our results are at least equivalent to previously reported infection rates that have ranged from 0% to 66% for tibial fractures treated initially with external fixation and converted to intramedullary nail fixation39-42. Johnson et al.15 reported on infectious complications in a similar cohort of war-related open tibial fractures. According to their report, thirteen of thirty-five patients had union times of greater than nine months, which appeared to be associated with infection. In their patients, although initial wound cultures from admission showed a high frequency of gram-negative pathogens, most notably Acinetobacter, cultures from the nonunions all showed at least one staphylococcal organism. Our patients also had a high prevalence of Acinetobacter and other gram-negative pathogen contamination on admission surveillance cultures. It is unclear why the infection rates in the two studies differ. Our institution typically provided broad-spectrum coverage to most patients who were febrile and had positive Acinetobacter surveillance cultures. This practice was instituted because we decided to provide broad coverage with antibiotics known to be active against this species of bacteria. Some of our patients received rhBMP-2 at the time of wound closure, and in one report this has been associated with a decreased risk of infection43.
There are several limitations to our study. Foremost is the retrospective nature of our review and the lack of a comparison group. Currently, we know of no published data supporting our protocol in the civilian literature, and only level-IV data have been reported in the literature on war-related fractures33-35. We also do not report functional outcome data. Ideally, clinical decision-making should be based on prospective functional outcome data. We are attempting to collect retrospective functional data on all war-related open tibial fractures from the recent conflicts in the Middle East to better define best practices for blast-related fracture care and to provide outcomes comparing different modes of fracture fixation for these injuries.
In conclusion, we present a treatment protocol that has resulted in a low rate of complications and a relatively high rate of fracture union for a subset of severe open tibial shaft fractures resulting from high-energy blast injuries in a war zone. Ring external fixation in these patients appears to be a valuable form of treatment for tibial shaft fractures and allows the added benefit of immediate weight-bearing without hardware retention after fracture-healing.