Eciency of the Damage Control Orthopedic Strategy in the Management of Military Ballistic Limb Trauma.

Background: In conicting areas, orthopaedic surgeons adopted this concept of damage control orthopaedic (DCO) to face limb fracture due to ballistic trauma because of the gravity of the hurts, the limitation of equipment, and precarious conditions of asepsis. They use external xation as an initial treatment at the nearest health centre. They delay the denitive treatment to be realized in better conditions. Our study aims to assess the outcome of the damage control orthopaedic (DCO) strategy in military ballistic limb trauma according to the experience of the Military Hospital of Tunisia. Materials and methods: This study is a retrospective study on patients who were hospitalized for a limb ballistic fracture. It includes military patients treated urgently with primary external xation at the nearest health centre. We adapted the Gustilo classication to describe wound opening, the Winquist and Hansen classication to dene fracture comminution and the grading system for bone loss to evaluate bone loss. Then, all victims were evacuated secondary to the military hospital to receive the denitive treatments. The conversion to internal osteosynthesis had taken place according to some criteria. They were the absence of local sepsis, a haemoglobin level (> 10 g/dl), a serum protein level (> 50 g/l) and negative or falling CRP kinetics. We studied the delay of conversion from external xation to internal osteosynthesis, bone healing time and complications. Result: Our study included 32 patients, all men, with a mean age of 31 years. The average follow-up was 33.2 months. Eighty-eight percent of trauma concerned the lower limb. The average delay for conversion from external xation to internal osteosynthesis was 7.8 days. Bone union was achieved in 26 cases, with an average delay of 4.23 months. The observed general complications were anaemia, pulmonary embolism and rhabdomyolysis. Local complications were essentially sepsis on osteosynthesis material and pseudarthrosis. These complications were signicantly associated with a type III Gustilo skin opening, a type III and IV Winquist fracture comminution, a type II and III Grading system for bone loss, and a delay in conversion from external xation to internal osteosynthesis. Conclusion: DCO is a global strategy that involves all measures participating in the acceleration of wound healing and ghting against infection. These measures shorten the delay of conversion from external xation into internal osteosynthesis, which constitutes a key parameter in the management of limb fracture due to ballistic trauma.

Then, all victims were evacuated secondary to the military hospital to receive the de nitive treatments. The conversion to internal osteosynthesis had taken place according to some criteria. They were the absence of local sepsis, a haemoglobin level (> 10 g/dl), a serum protein level (> 50 g/l) and negative or falling CRP kinetics. We studied the delay of conversion from external xation to internal osteosynthesis, bone healing time and complications.
Result: Our study included 32 patients, all men, with a mean age of 31 years. The average follow-up was 33.2 months. Eighty-eight percent of trauma concerned the lower limb. The average delay for conversion from external xation to internal osteosynthesis was 7.8 days. Bone union was achieved in 26 cases, with an average delay of 4.23 months. The observed general complications were anaemia, pulmonary embolism and rhabdomyolysis. Local complications were essentially sepsis on osteosynthesis material and pseudarthrosis. These complications were signi cantly associated with a type III Gustilo skin opening, a type III and IV Winquist fracture comminution, a type II and III Grading system for bone loss, and a delay in conversion from external xation to internal osteosynthesis.
Conclusion: DCO is a global strategy that involves all measures participating in the acceleration of wound healing and ghting against infection. These measures shorten the delay of conversion from external xation into internal osteosynthesis, which constitutes a key parameter in the management of limb fracture due to ballistic trauma.

Introduction:
Damage control (DC) is not a medical concept; Navy (1) used it in the Second World War to describe all the temporary measures used in the ght to prevent a ship from sinking while pursuing its mission. The damage control surgery was developed to face abdominal traumas with massive bleeding, shorten the operating time and avoid the lethal set of three: hypothermia, coagulopathy and acidosis (2). Thus, it opposes early total care de ned as an ideal and complete coverage of all the hurts from the rst passage in the operating room. The DC was then applied to manage polytraumatic fractures of the long bones and pelvis and de ned as damage control orthopedics (DCO). It is a relatively new concept (3) using external xation because of the gravity of hurts, the limitation of equipment, and precarious conditions of asepsis in con ict areas. The de nitive synthesis is secondarily made as soon as possible, after regression of the in ammatory reaction and oedema.
Our study aims to assess the outcome of the DCO strategy in military ballistic limb trauma according to the experience of the Military Hospital of Tunisia.

Materials And Methods:
This study is a retrospective, single-centre, descriptive study on patients who were hospitalized for gunshot limb trauma at the Orthopaedic and Traumatology Department of the Main Military Hospital of Instruction in Tunis for seven years from 2011 to 2017. It included military patients with ballistic fractures requiring emergency primary external xation followed secondarily by conversion to internal synthesis, with a minimum follow-up of 12 months. Patients requiring exo xation as de nitive treatment and patients lost to follow-up were excluded from this study.
All patients had urgent surgical management at the nearest health centre. It consists of soft tissue debridement, lavage and bone stabilization by external xation (Fig. 1A, B, C). The elaboration of lesion assessment. We adapted the Gustilo classi cation to describe wound opening, the Winquist and Hansen classi cation to de ne fracture comminution and the grading system for bone loss to evaluate bone loss.
A rst-line antibiotic therapy based on amoxicillin-clavulanic acid and aminoglycoside with tetanus prophylaxis was administered to all our patients.
Victim evacuation to the military hospital was carried out within 24 hours by land transport, except three victims, presenting vascular lesions, were transferred urgently by air transport.
At J2 post-traumatic, all patients had second-look surgery with debridement. Three patients with vascular lesions required immediate revision surgery after radiological exploration.
All the wounds were left open. Wound debridement was repeated once every two days until the wound was closed. Vacuum-assisted closure or VAC® therapy was used to accelerate healing. This procedure was changed every two days following debridement. Hyperbaric oxygen therapy was performed in patients with Gustilo III wound opening.
Bacteriological samplings from deep soft tissues were carried out at each debridement. Antibiotic therapy was adapted to the results of the bacteriological examination. Victims underwent biological monitoring with an assessment, including complete blood count (CBC), C reactive protein (CRP), serum protein, and prothrombin time (PT) at a rate of 1 day / 3.
The conversion to internal osteosynthesis had taken place at different times in our patients. The criteria for determining the time to conversion were the absence of local sepsis, a haemoglobin level (> 10 g/dl), a serum protein level (> 50 g/l) and negative or falling CRP kinetics.
Our study protocol is resumed in the diagram (Fig. 2).
We noted all the reports of the examinations carried out at the time of the accident throughout the hospitalization and postoperative follow-up. We studied the delay of conversion from external xation to internal osteosynthesis, bone healing time and complications.
The data were collected using a standardized analytical sheet and then entered using SPSS software. We calculated frequencies and relative frequencies (percentage) for the qualitative variables. We evaluated means and determined the range (extreme values = minimum and maximum) for the quantitative variables. The parametric Pearson correlation coe cient was used to study the relation between two quantitative variables. The link between two qualitative variables was evaluated by the chi-2 test and the Fisher test in case of chi-2 test invalidity. The association between a binary qualitative variable and a quantitative variable was assessed by Student's t-test. The signi cance level was set at p ≤ 0.05 for all tests performed.

Results:
Our study included 32 patients. All of our patients were male. The mean age was 31 years (21-51 years). Nineteen patients had a smoking history.
Ballistic trauma mainly concerned the lower limb (88% of cases) and especially the femur, which represents 50% of lesions of the lower limbs.
The distribution of lesions according to the classi cation of Gustillo, Winquist and Hansen and the grading system for bone loss are represented by Table 1. Grading system for boneloss 12 6 12 2 Three patients had vascular damage treated by an emergency revascularization procedure. One patient had nerve damage (neurapraxia).
VAC therapy was used in all our patients. Hyperbaric oxygen therapy was performed in twelve patients with Gustilo III wound opening. Bacteriological analysis isolated "Acinetobacter baumani" in twelve cases and "Staphylococcus" in seven cases. In the rest of the cases, the result was multibacterial.
The conversion from external xation to internal osteosynthesis took place in an average delay of 7.8 days (1 day -15 days). Intramedullary nailing was the most commonly used osteosynthesis material in 72% of cases. In the remaining 28%, we used an osteosynthesis plate (Fig. 3A, B, C). A pin-site infection was observed in 3 cases at the conversion surgery, requiring a systemic curettage of the pin sites before any material placement. Finally, the average hospital stay was 13.4 days (10 days -18 days).
Most observed general complications were anaemia, pulmonary embolism and rhabdomyolysis.
Anaemia (Hb < 10 g/dl) was observed in sixteen patients treated by transfusion. Pulmonary embolism was detected in three patients. It has progressed well under anticoagulant treatment, without recourse to reanimation. Rhabdomyolysis was noticed in three patients without an impact on renal function and responded well to medical treatment. No fat embolism, acute respiratory distress syndrome, or multiple organ dysfunction syndrome were observed in our series.
Local complications are essentially summarized in sepsis on osteosynthesis material and pseudarthrosis. Five patients suffered from sepsis on osteosynthesis material, and six patients evolved into non-union treated by a spongy graft with internal synthesis.
The analysis of these complications objective a signi cant association between sepsis on osteosynthesis material and four settings: smoking (p = 0.005), pin-site infection (p = 0.009), a Gustillo type III skin opening (p = 0.03) and the delay of conversion to internal osteosynthesis, as shown in Table  2. Table 2 The association between the delay of conversion and sepsis on osteosynthesis material Presence of sepsis Absence of sepsis Signi cance Average conversion delay 12.6 days 6.9 days P = 0.003 The non-union rate was also associated with smoking (p = 0.001), type III and IV of the Winquist fracture comminutions (p = 0.013), and type II and III of the Grading system for bone loss (p = 0.016). The delay of conversion to internal osteosynthesis is shown in Table 3. Table 3 The association between the delay of conversion and a non-union rate.

Fracture non-union
Fracture union Signi cance Average conversion delay 12.1 days 6.8 days P = 0.008

Discussion:
Our study objective is that the DCO strategy in the management of fractures due to ballistic trauma is e cient, with a consolidation rate > 80% and a relatively low percentage of general complications. These complications had minor severity, and they were easy to manage.
Local complications, mainly sepsis on osteosynthesis material and pseudarthrosis, had relatively low rates. Our analyses showed that they were associated with several parameters. These parameters can be divided into nonmodi able risk factors, such as a history of smoking, wound opening, fracture comminution, and bone loss. Modi cable risk factors are mainly the delay of conversion from external xation to internal osteosynthesis. The local complication rate was lower when this delay of conversion was shorter.
Our study is distinguished by strict inclusion and exclusion criteria. It was carried out in an institution that has good experience in the eld of ballistics. Its protocol was interested not only in the results of the DCO strategy but also in the parameters that could affect this strategy. However, our study also has some limitations, mainly a selection bias due to its retrospective nature and the relatively small number of patients.
DCO is a relatively new concept (3) adopted in the management of ballistic limb trauma and appears to solve the defects of an older approach, which consists of early stabilization of skeletal lesions called early total care.
Early total care was the principal strategy for the management of polytrauma and war wounded in the 1980s and 1990. However, recent studies have shown that adoption of this strategy in groups of patients with hemodynamic instability is more associated with signi cant complications such as pulmonary embolism and acute respiratory distress syndrome and multiple organ dysfunction syndrome. This was exceptionally observed in intramedullary femur nailing (4). Studies have associated the occurrence of these complications with changes in pro-in ammatory markers (5). Indeed, the initial accident causes in ammatory and immunological reactions proportional to the severity of the trauma called the " rst hit" (6). This reaction is characterized by local and systemic proliferation of various pro-in ammatory mediators, such as cytokines, complement, coagulation proteins and others. In addition, the prolonged duration of the surgery and bleeding lead to another signi cant in ammatory and immunological reaction called "Second Hit". This second hit potentiates the effect of the rst hit, and this may lead to a severe consequence in patients.
From these observations, a new strategy, based on minimizing the impact of the second hit by shortening the initial operating time and delaying the de nitive treatment, was adapted to manage limb ballistic trauma. This strategy was called the DCO. Therefore, the treatment of long bone fractures of soldiers wounded on the battle eld is based on temporary external xation, whose objective is controlling haemorrhage, restoring perfusion of the limb, debridement of necrotic soft tissue and ensuring bone stability. Additionally, without disturbing resuscitation care measures (7).
However, after hemodynamic stabilization, control of in ammatory phenomena and improvement of the local wound condition, this external xation would be better converted into internal osteosynthesis.
Although external xation does not always allow anatomical reduction, it is often associated with a high rate of pin-site infection and low-quality bone callus. Respet et al attempted to determine the time between the realization of external xation and the onset of pin site infection. They found that pin bacteriological cultures were positive in 50% of the cases in 2 weeks and 67% in 4 weeks (8). These results were also con rmed later by Clasper et al. (9). Additionally, Sigurdsen et al (10,11) experimented on rats.
To study the quality of the bone callus after an osteotomy, the callus was initially treated with an external xator and then converted to internal synthesis with different delays. This conversion delay was seven days in the A group, 14 days in the B group, 30 days in the C group and control group D without conversion. All the groups had better consolidation than the control group, but only group A had a signi cant difference. Biomechanically, the rigidity and quality of the callus in group A were better than those in the other groups.
These studies not only show the interest of the conversion from external xation to internal osteosynthesis but also prove the interest of the delay of this conversion. The shorter it is, the better the quality of bone callus, and the lower the risk of pin site infection and sepsis. This result is consistent with our study, which objecti ed that the delay of conversion is a risk factor for both septic and non-union complications.
In our study, this conversion time was relatively short compared to other papers in the literature (12)(13)(14)(15). Care aimed at accelerating different phases of the wound healing process and ghting infection was possible. Indeed, VAC therapy allows a permanent elimination of exudates from the wound bed. Associated with repetitive debridement, it accelerated the in ammatory phase of the wound healing process. Hyperbaric oxygen therapy improves tissue oxygenation. Additionally, it enhances broblast and collagen synthesis, neovascularization, and the closure of arterial-venous shunts (16), which shortens the time to granulation formation, especially in open wound fracture Gustillo III (17). Additionally, several studies have proven the effect of VAC therapy to accelerate granulation tissue formation (18)(19)(20). With normal haemoglobin and serum protein levels, they hasten the proliferative phase of wound healing. Furthermore, in addition to adapted antibiotic therapy, hyperbaric oxygen therapy and VAC therapy had a con rmed role against infection.

Conclusion:
DCO is a management strategy not only limited to shortening the initial management and delaying the de nitive treatment but also aimed to reduce severe or even fatal complications. However, it is a global strategy that involves all measures participating in the acceleration of wound healing and ghting against infection. These measures shorten the delay of conversion from external xation into internal osteosynthesis, which constitutes a key parameter in the management of limb fracture due to ballistic trauma. Availability of data and materials

Abbreviations
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
There were no sources of support from any funding agency.