 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1996 The Journal of Bone and Joint Surgery, Inc. Open Fracture of the Tibia in Children* , CINCINNATI, OHIOInvestigation performed at the Department of Orthopaedic Surgery, Children's Hospital Medical Center, Cincinnati
The records of eighty-three children who had had an open fracture of the tibial metaphysis or diaphysis between January 1983 and July 1993 were studied retrospectively. The average duration of follow-up was fourteen months (range, two to seventy-five months). There were twenty-four grade-I, forty grade-II, thirteen grade-IIIA, six grade-IIIB, and no grade-IIIC fractures, according to the classification scheme of Gustilo et al. Sixty patients (72 per cent) had sustained the fracture when they were struck by an automobile, and forty-eight patients (58 per cent) had other associated major injuries. All fractures were treated with irrigation and débridement, and antibiotics were administered parenterally for a minimum of forty-eight hours. Thirty-two patients were managed with immobilization in a cast only; forty, with transcutaneous fixation with an average of two Steinmann pins followed by immobilization in a cast; nine, with external fixation; one, with open reduction and internal fixation with two screws and two pins; and one, with delayed intramedullary nailing. Fifty-seven wounds were closed primarily (forty-four, over a Penrose drain, and thirteen, without a drain), ten were treated with delayed closure, four were allowed to heal by secondary intention, seven were covered with a soft-tissue flap, and five were treated with skin-grafting (a split-thickness skin graft was used for four, and a split-thickness and a full-thickness skin graft were used for one). The average time to union was fifteen weeks (range, five to sixty-one weeks), with the fracture healing by sixteen weeks in sixty-four patients (77 per cent). Eighteen patients (22 per cent) had delayed union, and only one patient (1 per cent) had non-union. Secondary procedures were necessary to achieve union in only two patients. Two patients had a superficial wound infection, and no patient had osteomyelitis. One patient, who had been managed with external fixation, had a pin-track infection; none of the patients who had had transcutaneous fixation had a pin-track infection. Two patients had a compartment syndrome, and two patients had a transient stretch injury of a nerve (the peroneal nerve in one and the sciatic nerve in the other). Four fractures healed with an angulatory deformity of more than 10 degrees in any plane. Five patients had overgrowth of the limb of one centimeter or more. Physeal arrest did not occur in any patient. We concluded that treatment of unstable open fractures of the tibia in children with débridement and transcutaneous fixation followed by immobilization in a cast leads to good anatomical and functional results. We prefer this technique to external fixation, which is associated with several potential complications. Loose closure of a clean open wound over a Penrose drain is effective and can be safely utilized in selected children.
There have been many studies on the treatment of open fractures of the tibia in adults; however, there have been only a few on the treatment and outcome of these fractures in children3,6,11,21,24,45. Open fractures in children have been reported to heal with few complications in a time frame similar to that for closed fractures20,31. The purpose of the present retrospective study was to analyze the results and complications of treatment associated with open fractures of the tibia in children.
Between January 1983 and July 1993, ninety-two children with an open fracture of the tibial metaphysis or diaphysis were admitted to the Children's Hospital Medical Center in Cincinnati. The records and radiographs of all patients were retrospectively reviewed to evaluate the results and complications of treatment. Nine patients were not included in the study because of inadequate follow-up or incomplete medical records (six patients), primary amputation (two patients), or death as a result of associated injuries (one patient). Eighty-three patients who had been followed until the fracture healed were included in the study. The average age of the eighty-three children was nine years (range, three to seventeen years). All of the patients had open physes at the time of the injury. There were sixty-five boys and eighteen girls. The fractures were due to an automobile-pedestrian accident (forty-four patients), an automobile-bicycle accident (sixteen patients), an all-terrain-vehicle accident (five patients), an automobile accident (four patients), a motorcycle accident (three patients), a sports-related accident (three patients), a fall (two patients), a sledding accident (two patients), a bicycle accident (one patient), a lawn-mower accident (one patient), a crush injury (one patient), or the child being struck by a rock (one patient). There were twenty-seven transverse fractures, forty-five oblique fractures, six spiral fractures, three segmental fractures, one buckle fracture, and one cortical defect. Thirteen fractures were located in the proximal third of the tibia, thirty-three were in the middle third, fifteen were at the junction of the middle and distal thirds, nineteen were in the distal third, and three were segmental. The comminution of the fracture was graded according to the classification system of Winquist and Hansen43. Fifty-seven fractures had no or grade-I comminution; fourteen, grade-II comminution, and twelve, grade-III or IV comminution. Thirty-one fractures were displaced less than 50 per cent of the diameter of the shaft; twenty-three, 50 to 100 per cent; and twenty-nine, more than 100 per cent. Fifty-four patients also had a fracture of the fibula, and seven had a greenstick fracture of the fibula; the fibula was intact in the remaining twenty-two patients. There were twenty-four grade-I, forty grade-II, thirteen grade-IIIA, six grade-IIIB, and no grade-IIIC fractures, according to the classification scheme of Gustilo et al.15,17. Forty-eight patients (58 per cent) had other major injuries. There were forty-seven associated fractures, including seven of the ipsilateral femur, eight of the contralateral femur, and five of the humerus. There were five closed head injuries, seven open joint lacerations, six pneumothoraces, five lacerated tendons or ligaments, and six major soft-tissue lacerations or degloving injuries. In the emergency room, a splint was applied to the limb and antibiotics were administered parenterally. No specific antibiotic protocol was used during this time period, and a variety of antibiotic regimens were prescribed. Prophylaxis against tetanus was administered when indicated2. All fractures were irrigated and debrided in the operating room on the day of the injury, after resuscitation and stabilization of the patient. The type of treatment depended on the severity of the soft-tissue injury and the stability of the fracture. Fractures were considered to be stable when there was less than 10 degrees of angulation in any plane on passive elevation. Thirty-two patients who had a stable fracture with no more than grade-I comminution, and in whom an acceptable reduction could be maintained, were managed with immobilization in an above-the-knee cast Table I). Forty patients who had an unstable fracture without major soft-tissue loss were managed with anatomical reduction and transcutaneous fixation of the fracture, typically with one or two threaded Steinmann pins. An average of two pins was needed to achieve stability. One pin was used in twelve patients; two pins, in twenty-seven; and four pins, in one patient. The size of the Steinmann pin varied according to the age and size of the patient. The pins were deliberately left long so that they could be removed without general anesthesia if there was a problem with the wound. The fractures were then protected with a cast.
Nine patients who had an unstable fracture with loss of soft tissue were managed with external fixation (Table I). The external fixator was chosen according to the surgeon's preference. Two Roger Anderson devices (Kirschner Medical, Timonium, Maryland) and seven Orthofix devices (EBI Medical Systems, Parsippany, New Jersey) were used. One patient was managed with open reduction and internal fixation with two screws and two pins, and one patient had delayed intramedullary nailing. A four-compartment fasciotomy was performed at the time of the operative treatment of the fracture in one patient and twenty-four hours after the injury in a patient who had a severe closed head injury. The type of closure of the wound was chosen on the basis of the extent of the soft-tissue injury and the contamination of the wound. Minimum contamination was defined as only devitalized loose soft tissue and bone debris in the wound, with most of the tissue appearing well vascularized. A wound was considered to be grossly contaminated if there was visible loose debris and devitalized tissue. Primary closure (loose closure, with large nylon sutures, usually over a Penrose drain) after débridement and irrigation was the treatment of choice for wounds with minimum contamination; it was done at the discretion of the surgeon. A window was created in the cast to allow access to and inspection of the wound. This made it possible to remove the sutures if there was any drainage or if infection was suspected. Fifty-seven wounds had primary closure (forty-four, over a Penrose drain, and thirteen, without a drain), ten had delayed closure on the third postoperative day, and four were allowed to heal by secondary intention (Table II). Seven patients had a total of eight soft-tissue-coverage procedures; two rectus flaps, two gastrocnemius flaps, two soleus flaps, one scapulocutaneous flap, and one abductor hallucis flap were used. The average time to coverage of the wound was eight days (range, three to seventeen days). A split-thickness skin graft was used on four wounds, and split-thickness and full-thickness skin grafts were used on one wound.
All patients were admitted to the hospital, and antibiotics were administered parenterally for a minimum of forty-eight hours. When the wound had been closed or closed over a drain, it was evaluated and the drain was advanced or removed on the second or third postoperative day. Stable patients who had an isolated injury were permitted to get out of bed and were sent to the physical therapy department for training in walking with toe-touch weight-bearing. The average duration of hospitalization was ten days (range, two to forty-seven days). Transcutaneous pins were removed in the operating room, and an above-the-knee or below-the-knee cast was applied, when there was evidence of early fracture callus on follow-up radiographs; this occurred at an average of seven weeks (range, two to fifteen weeks). The external fixators were removed at an average of twenty-one weeks (range, nine to thirty-one weeks). Protected weight-bearing was then permitted, except if the patient had an unstable fracture that had not healed completely, in which case a cast was applied. Union was defined radiographically as evidence of bridging callus on plain radiographs and clinically as painless full weight-bearing. Delayed union was defined as incomplete healing at sixteen weeks21, which is four to six weeks longer than the average time to union of closed fractures of the tibia in children20,37. Non-union was defined as no progression of clinical and radiographic healing after six months of immobilization7,15,21,30,34. Malunion was defined as residual angulation of more than 10 degrees in any plane6,14,30. Alignment of the fracture was assessed on radiographs that were made at the time of the most recent follow-up evaluation. A patient was considered to have an infection when there were clinical signs and symptoms of an infection and a culture or gram stain of material obtained from the wound was positive. A superficial infection involved the skin and subcutaneous tissues, and a deep infection extended to the fracture site. All patients were followed at least until the fracture united; the average duration of follow-up was fourteen months (range, two to seventy-five months). Overgrowth or shortening of the lower extremity was assessed clinically, and scanograms were made only if a discrepancy of one centimeter or more was identified (five patients). The data were analyzed, through a series of one and two-factor linear statistical models with use of least-squared means testing, to evaluate whether delayed union was significantly associated with the grade, location, pattern, displacement, or comminution of the tibial fracture; the presence of an associated fracture of the fibula; or the age of the patient. Analyses for each model were assessed for homoscedasticity (having equal statistical variances) and normality. A standard value of p < 0.05 was used for all analyses.
Time to Union
For the thirty-two fractures treated with immobilization in a cast only, the average time to union was twelve weeks. There was routine healing of 88 per cent (twenty-eight) of the fractures, with an average time to union of eleven weeks. Union of 13 per cent (four) of the fractures was delayed, with an average time to union of twenty-two weeks (Table III). For the forty fractures that had transcutaneous fixation followed by immobilization in a cast, the average time to union was sixteen weeks (Figs. 1-A, 1-B, 1-C, 1-D, 1-E through 1-F). Seventy-five per cent (thirty) of these fractures healed routinely, with an average time to union of eleven weeks. Union of 23 per cent (nine fractures) was delayed, with an average time to union of twenty-three weeks, and 3 per cent (one fracture) did not unite until sixty-one weeks after treatment. The nine fractures treated with external fixation united in an average of twenty-one weeks. There was routine healing of four fractures, with an average time to union of fourteen weeks, and delayed union of the other five fractures, with an average time to union of twenty-six weeks. The two fractures that were treated with delayed intramedullary nailing or open reduction and internal fixation with two screws and two pins healed routinely at an average of thirteen weeks.
For the forty-five children who were ten years old or less, the average time to union was thirteen weeks. Routine healing occurred in 84 per cent (thirty-eight) of them, with an average time to union of eleven weeks, and delayed union occurred in 13 per cent (six), with an average time to union of twenty-two weeks. Non-union occurred in one patient. For the thirty-eight children who were more than ten years old, the average time to union was fifteen weeks. Routine healing occurred at an average of eleven weeks in 66 per cent (twenty-five), and delayed union occurred in 34 per cent (thirteen), at an average of twenty-four weeks. Comminution of the fracture (p < 0.01) and segmental injuries (p < 0.004) were significant predictors of delayed union. The grade (p < 0.8), location (p < 0.11), pattern (p < 0.4), and displacement (p < 0.2) of the tibial fracture; the presence of an associated fibular fracture (p < 0.2); and the age of the patient (p < 0.1) were not significant predictors. Statistical analysis revealed no difference (p < 0.3) in the prevalence of delayed union between fractures treated with immobilization in a cast only and those treated with transcutaneous fixation. There was, however, a significant difference between fractures treated with immobilization in a cast only and those treated with external fixation (p < 0.0004) and between fractures treated with transcutaneous fixation and those treated with external fixation (p < 0.02). Although union was delayed in three of the five patients who had a head injury and five of the fifteen patients in whom the fracture was stabilized in a position of malalignment, these findings were not significant because of the small number of patients involved. Secondary procedures were used to achieve union in two patients (per cent). The first patient was a ten-year-old boy who was struck by an automobile and sustained a grade-II open fracture at the metaphyseal-diaphyseal junction of the left tibia. He also had a closed head injury, a pneumothorax, and fractures of the left femur, humerus, radius, and ulna. The tibial component of the floating-knee injury was treated with irrigation and débridement followed by transcutaneous fixation and application of a below-the-knee cast, and the femoral fracture was treated with 90-90 traction for five weeks. After this, the patient wore a hip-spica cast for four weeks and then a walking spica cast for an additional eight weeks, until the femoral fracture had healed. He then wore a patellar-ligament-bearing cast for four weeks. A fibular ostectomy was performed at seventeen weeks because the fracture had not healed. Postoperatively, he wore an above-the-knee weight-bearing cast for two months, followed by a patellar-ligament-bearing brace. An electrical stimulator (EBI) was used for six months to promote healing of the fracture, and union was achieved at sixty-one weeks. The second patient was a fifteen-year-old girl who was struck by an automobile and sustained a grade-II open fracture of the middle third of the left tibia with grade-IV comminution. The fracture was treated with irrigation and débridement and external fixation. Weight-bearing was allowed at six weeks. At nineteen weeks, the fracture had not healed, and posterolateral bone-grafting was performed. The fracture was healed at thirty-one weeks, and the external fixator was removed.
Infection A pin-track infection developed in a patient who had a grade-II open fracture that was treated with external fixation. The infection resolved with administration of cephalexin for two weeks. Slight bone resorption was noted at the pin sites at six weeks, and the fixator was removed at sixteen weeks; no débridement of the pin site was performed. Neither a draining sinus nor osteomyelitis developed after the removal of the external fixator. There was drainage from the pin site in a patient in whom a grade-IIIA open fracture of the proximal part of the tibia with a laceration of the patellar ligament and exposure of the knee joint had been treated with open reduction and internal fixation with two screws and two pins. The infection resolved after removal of the pins and a seven-day course of cephalexin. No patient who had been managed with fixation with transcutaneous pins had a pin-track infection.
Neurovascular Injuries
Angulatory Deformities
Limb-Length Discrepancy
In recent reports of open tibial fractures in children, Yasko and Wilber45, Buckley et al.6, Hope and Cole21, and Cramer et al.11 concluded that the prevalences of compartment syndrome, vascular injury, infection, and delayed union were similar to those found in adults3,6,11,21,45. The prevalence of compartment syndrome associated with open fractures of the tibia in adults was reported by DeLee and Stiehl12 to be 6 per cent (six of 104) and by Blick et al.4 to be 9 per cent (eighteen of 198). Both of these rates are greater than the 2 per cent prevalence (two of eighty-three) in our series of children. When the results of our series are combined with those reported by Buckley et al.6 (two of forty-one) and by Hope and Cole21 (four of ninety-two), the over-all prevalence is 4 per cent (eight of 216). It is important to recognize that this injury occurs in children and that early fascial release is necessary to prevent the long-term consequences of unrecognized compartment syndrome23,26. The prevalence of vascular injury associated with open fractures of the tibia in adults was reported by Caudle and Stern7 to be 4 per cent (nine of 217). During the period of our study, only one child (not included in the series) had a vascular injury. When our results are combined with those reported by Buckley et al.6 (two of forty-one), Cramer et al.11 (four of twenty-four), and Hope and Cole21 (one of ninety-two), the over-all prevalence is 3 per cent (eight of 241). Cole10 postulated that injuries of the popliteal and tibial arteries may be less prevalent in children because the vessels have greater elasticity. The prevention of wound infection involves thorough débridement and irrigation of the wound, stabilization of the fracture, and parenteral administration of antibiotics9,28. The prevalence of infection associated with open fractures of the tibia in adults has been reported to be ten to twenty times greater than that associated with other skeletal injuries29. The prevalence of infection after open fractures of the tibia in adults was reported by Patzakis et al.29 to be 5 per cent (five of 109), by Clancey and Hansen9 to be 15 per cent (fifteen of 102), and by Edwards et al.14 to be 15 per cent (twenty-six of 171). In a series of forty-nine patients who had a grade-III open fracture of the tibia, Caudle and Stern7 found a 33 per cent rate of infection (sixteen patients). The literature on infection in association with open fractures of the tibia in children is limited6,21,45. Patzakis and Wilkins28 reported a 2 per cent rate of infection in children (one of fifty-five) and a 7 per cent rate in adults (seventy-six of 1049). Buckley et al.6 reported that infection occurred in three (7 per cent) of forty-one patients and osteomyelitis, in one (2 per cent). Yasko and Wilber45 reported an over-all rate of infection of 8 per cent in fifty-three patients; infection developed in association with none of the grade-I fractures, 6 per cent of the grade-II fractures, and 33 per cent of the grade-III fractures. Infection developed in 11 per cent (ten) of the ninety-two patients in the series of Hope and Cole21: it was associated with none of the twenty-two grade-I fractures, 12 per cent (six) of the fifty-one grade-II fractures, and 21 per cent (four) of the nineteen grade-III fractures. Osteomyelitis developed in 3 per cent (three) of the ninety-two children. Cramer et al.11 and Blaiser and Barnes3 did not report a rate of infection. In our series, a superficial infection developed in two (2 per cent) of the eighty-three children, and osteomyelitis did not occur in any patient. There were no infections in association with the nineteen grade-III fractures. This rate of infection is lower than that associated with the same injury in adults7,10,14,15,17,18. We believe that our results support the contention that the prevalence of wound infections and osteomyelitis associated with open fractures of the tibia in children is lower than that associated with the same injury in adults. Other large series of open fractures of the tibia in children must be reported on to confirm the validity of this finding. The orthopaedic literature is replete with studies in which the investigators concluded that primary closure of the wound is not indicated in the treatment of open fractures5-7,10,16-18,28,29,35,41. Russell et al.35 reported a 21 per cent prevalence of infection (twelve of fifty-eight) associated with open fractures of the tibia that were closed primarily. There was one infection in the thirty-two wounds that had delayed primary closure. We agree that there is no role for primary closure of the wound in adults who have an open fracture, but we do not believe that this applies rigidly to the treatment of wounds in children who have an open fracture. Hope and Cole21 performed primary wound closure in the treatment of fifty-one open fractures of the tibia, including eleven grade-I, thirty-five grade-II, and five grade-III fractures; infection developed in 8 per cent (four). In the same study, infection developed in 15 per cent (six) of the forty-one wounds that had been left open to close by secondary intention. We believe that the soft-tissue injuries in children are not the same as those in adults and that, in selected patients, loose closure over a Penrose drain is acceptable. With our protocol, fifty-seven wounds (twenty grade-I, thirty grade-II, and seven grade-IIIA open fractures) were closed primarily (forty-four, over a drain, and thirteen, without a drain); only two patients had a superficial infection and no patient had osteomyelitis. Both wounds that became infected were grossly contaminated at the time of the operation and, in retrospect, not suitable for primary wound closure. We believe that our results support the use of primary wound closure in the treatment of selected non-contaminated open fractures in children. If there is any contamination, delay in operative débridement, or concern regarding the adequacy of the débridement, open care of the wound is indicated. Immobilization in an above-the-knee cast was the preferred treatment for stable fractures with minimum soft-tissue injury in children. This technique predictably leads to union with few complications. External fixation is indicated for children who have a complex unstable fracture or loss of soft tissue, or both32. External fixation maintains the length and alignment of the fracture while allowing access to the open wound. It has been shown to be associated with an increased rate of deep infection, pin-track infection, tibial overgrowth, refracture, and delayed union1,6,38-40. In our series, nine patients were managed with external fixation; there was one pin-track infection, no wound infections, no tibial overgrowth, and five delayed unions. Because of selection bias (external fixation was used for patients who had more serious injuries of bone and soft tissue), it is impossible to determine retrospectively the role of external fixation in delayed union of open fractures. There is no standard definition of delayed union and non-union in the pediatric literature. It has been well documented that fractures in children heal faster than those in adults; therefore, it would be inappropriate to apply the definition of delayed union and non-union in adults to children. Closed tibial fractures in children have been reported to unite in a predictable fashion, at an average of four to fourteen weeks after closed treatment19,36,37,42,44. Levy et al.24 reported an average time to union of eleven weeks, and Hope and Cole21 reported that open fractures without infection healed in an average of fourteen weeks. On the basis of these data, we agree with Hope and Cole that incomplete healing at sixteen weeks represents delayed union in children. To our knowledge, Lewallen and Peterson25 reported on the only series of children who had non-union of a long bone, but this series is of limited value for the prediction of which fractures are at risk for delayed union, as eight of the ten non-united open fractures of the tibia were associated with infection. Prospective studies are necessary to better define accurate time frames for the definition of delayed union and non-union in children. Stable fractures that were treated with immobilization in a cast healed at an average of twelve weeks, with delayed union of four (13 per cent) of the thirty-two. Unstable fractures treated with transcutaneous fixation healed at an average of sixteen weeks, with delayed union of 23 per cent (nine) of the forty, whereas unstable fractures with soft-tissue loss treated with external fixation healed at an average of twenty-one weeks, with delayed union in five of the nine. There was no significant difference in the time to healing between fractures treated with immobilization in a cast only and those treated with transcutaneous fixation. Despite an over-all 23 per cent rate of delayed union, the fracture healed in less than six months in 92 per cent (seventy-six) of the eighty-three patients and secondary procedures to achieve healing were necessary in only two patients. These results suggest that the prevalence of delayed union of open fractures of the tibia in children is similar to that reported for adults but the need for secondary procedures is less common. Similar results were found by Levy et al.24. We believe that our results indicate that the prognosis for open fractures of the tibia is more favorable in children than in adults. Limb-length discrepancies due to tibial overgrowth and physeal arrest are unique to fractures in children. Shannak36 and Reynolds33 reported that tibial overgrowth is minor after closed tibial fractures and typically is less than one-half centimeter. Maximum growth stimulation occurs at three months after the injury and returns to normal at forty months33. Buckley et al.6 reported on four patients who had a limb-length discrepancy of one to four centimeters after management with external fixation. Hope and Cole21 reported that both tibial shortening and overgrowth occurred after open tibial fractures in children. In our series, five children had tibial overgrowth, but we believe that this may be an underestimation because of the short duration of follow-up of some patients and the lack of a standard postoperative protocol other than clinical examination to assess limb-length discrepancy. Physeal arrest did not occur in any child in our series, but several instances of physeal arrest in association with non-physeal fractures have been reported22, including two described by Hope and Cole21. Long-term follow-up of at least one year in duration is necessary to obtain an accurate assessment of these delayed complications of open fractures in children. The reports of transcutaneous fixation have been anecdotal in the literature13,21 but, to our knowledge, there have been no studies of large series of patients who were managed with this technique for the treatment of an open fracture of the tibia. We used this technique to treat forty unstable fractures (ten grade I, twenty-five grade II, and five grade III) that, at most centers, would have been treated with external fixation. Although this technique does not provide rigid fixation of the fracture8, it does provide preliminary stabilization that can be supplemented with immobilization in a cast to achieve predictable healing with fewer complications than external fixation. We have had few complications associated with this technique. We recommend transcutaneous fixation as an alternative to external fixation in all but the most unstable open fractures of the tibia in children.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |