One hundred and two patients who had a unilateral, displaced, isolated closed fracture of the tibial shaft were managed by the two senior ones of us (L. B. B. and P. M. S.) between 1987 and 1991 at the Erie County Medical Center in Buffalo, New York. Forty-eight patients were managed with closed intramedullary nailing, and fifty-four had closed reduction and application of an above-the-knee cast. There was no protocol to determine which fractures would be treated with a cast or an intramedullary nail; however, the two senior ones of us believed that isolated closed tibial fractures should be treated, whenever possible, with a cast and a fracture-brace, and we chose the treatment for each fracture on the basis of the stability and the initial amount of shortening of that fracture. Closed reduction and a cast tended to be chosen for the more stable fractures, and intramedullary nailing tended to be chosen for the unstable injuries. Three patients were lost to follow-up: one patient who had been managed with a cast died of unrelated causes, and two patients (one who had been managed with a cast and one who had had nailing) were lost to follow-up before the fracture had united. This left a total of ninety-nine patients: forty-seven who had been managed with nailing and fifty-two who had been managed with a cast.
The criteria for inclusion in the study were an age of at least sixteen years, a closed displaced fracture of the tibial diaphysis (from five centimeters distal to the tibial tuberosity to seven centimeters proximal to the ankle joint), and an absence of associated compartment syndrome or neurovascular injury.
Hospital inpatient and outpatient records and radiographs were used to identify the patients and to assess the time to union of the fracture and the rate of complications.
The intramedullary nailing procedures were performed between three and seven days after the injury. After the fracture had been reduced and the limb had been placed in a splint and elevated, the timing of the operation was determined according to the degree of soft-tissue swelling. All of the nails were inserted after reaming, and eleven millimeters (range, ten to thirteen millimeters) was the most common size of nail that was used.
The use of locking screws was determined by the location and the stability of the fracture. Rotationally or axially unstable fractures were locked statically with proximal and distal screws to prevent either shortening or malrotation. Eighteen of the forty-seven patients managed with nailing had static locking.
The time to weight-bearing was determined on the basis of the stability of the fracture, the degree of cortical continuity, and the comfort of the patient, as well as evidence of fracture-healing. No patient who had had nailing was able to bear full weight on the extremity before six weeks.
The patients who were managed with a cast initially had reduction of the fracture, application of a splint, and elevation of the limb. An above-the-knee cast was applied within five days after the injury. The reduction was considered acceptable if there was at least 50 per cent cortical apposition as seen on anteroposterior and lateral radiographs, no more than 10 degrees of angulation in any direction, and no more than 1.5 centimeters of shortening. The above-the-knee cast was worn for a mean of seven weeks (range, four to twelve weeks). Forty patients had conversion to a patellar ligament-bearing cast, and twelve had application of a fracture-brace. Weight-bearing while wearing the cast or the brace was begun within two weeks after it was applied and at a mean of eight weeks (range, six to fourteen weeks) after the injury.
All casts were applied and all operations were performed by attending staff with the assistance of residents.
Radiographs were used to determine the location, amounts of comminution and displacement, classification, and time to healing of the fractures. The fractures were graded according to the AO/Orthopaedic Trauma Association classification11 and the classification system of Winquist and Hansen for comminution of fractures of long bones. The location of the fracture was defined as the proximal, middle, or distal third of the tibia. Radiographic healing was defined as the presence of bridging callus as seen on both anteroposterior and lateral radiographs. There was no standardization of the radiographs. Rotational alignment was assessed both clinically and with use of radiographs of the knee and the ankle. A fracture that had not united at nine months was considered a non-union. The time to healing was determined by one of the senior two of us.
There was no protocol for the routine follow-up of the patients. In general, patients who had been managed with a cast were seen every two weeks for a maximum of six to eight weeks after the injury and at monthly intervals thereafter, while those who had had nailing were seen two weeks after the operation and monthly thereafter.
Of the ninety-nine patients, twelve were lost to follow-up, nine were contacted and refused to be seen, and eleven were not contacted. Thus, sixty-seven patients (thirty-two who had had nailing and thirty-five who had been managed with a cast) returned for long-term follow-up, and from this group twenty-five matched pairs of patients were identified on the basis of age (within ten years) and the location and type of the fracture. Matches could not be found for the remaining seventeen patients (seven who had had nailing and ten who had been managed with a cast). However, with the numbers available, we found no discernible difference in the outcomes for these patients compared with those for the twenty-five matched pairs.
The matched pairs, consisting of twenty-five patients who had had nailing and twenty-five who had been managed with a cast, were assessed at a mean of 4.4 years (range, two to 6.5 years) (Table I). Each patient completed the Iowa Knee Evaluation9 and the Ankle-Evaluation Rating System9 questionnaires as well as the Medical Outcomes Study Short Form-36 Health Survey (SF-36)18. The SF-36 is a general health-status measure designed to assess health-related outcomes18 with regard to the quality of life. With use of this thirty-six-item, multidimensional, well validated questionnaire, seven different health concepts are evaluated according to eight separate subscales: physical functioning, physical role functioning, bodily pain, general health, vitality, social functioning, mental health, and emotional role functioning (not completed for the present study). The scores for the SF-36 were tabulated with use of standard scoring methodology and were transformed into values ranging from 0 to 100 (with larger numbers indicating better health).
All patients were examined by one of us (L. B. B. or P. M. S.) to determine the degrees of alignment and shortening and to complete the Iowa Knee Evaluation9 and the Ankle-Evaluation Rating System9 questionnaires. The time to union was compared for the matched pairs. In addition, all fifty patients were asked when they had returned to work after the injury.
Statistical Analysis
Statistical analyses were performed with use of the paired Student t test to compare differences between the two groups with respect to the time to union, the scores for the ankle and the knee, the functional outcome scores, and the time until the patients returned to work. The chi-square test was used to assess differences in the rates of complications between the two groups. Significance was defined as p < 0.05.
The mean time to radiographic union was twenty-six weeks (range, sixteen to fifty-two weeks) for the fifty-two patients who had been managed with a cast compared with eighteen weeks (range, twelve to fifty-two weeks) for the forty-seven patients who had been managed with nailing (p = 0.02). A non-union occurred in five patients (10 per cent) who had been managed with a cast and in one patient (2 per cent) who had had nailing. Four of the five patients who had a non-union after management with a cast had an associated fracture of the fibula. Fourteen patients (27 per cent) who had been managed with a cast had shortening of more than 1.5 centimeters compared with one patient (2 per cent) who had been managed with nailing. Of the ten patients who had shortening of more than two centimeters after management with a cast, only two had had shortening at the time of presentation. Five patients (10 per cent) had varus or valgus malalignment of more than 10 degrees, and five (10 per cent) had 10 degrees of recurvatum after management with a cast. No patient had varus, valgus, or sagittal malalignment after intramedullary nailing. These differences were significant (p < 0.05, chi-square test). No patient in either group had rotational malalignment of more than 10 degrees, an infection, or a compartment syndrome. No patient who had had nailing had failure of the hardware. Twenty-six of the forty-seven patients who had had nailing elected to have the nail removed more than eighteen months after the injury, usually because of pain in the knee.
For the twenty-five matched pairs of patients, the mean time to union was twenty-six weeks after management with a cast and twenty weeks after management with nailing (p = 0.04) (Table II). According to the Iowa Knee Evaluation9 and the Ankle-Evaluation Rating System9, the patients who had been managed with a cast had mean scores of 89 points (range, 62 to 100 points) and 84 points (range, 62 to 100 points), respectively, of a possible score of 100 points, and those who had had nailing had mean scores of 96 points (range, 68 to 100 points) and 97 points (range, 74 to 100 points) (Table II). The differences between the two groups were significant (p < 0.05).
In the matched-pairs group, the over-all score on the SF-36 questionnaire18 was significantly higher after nailing than it was after management with a cast (mean, 85 points [range, 27 to 99 points] compared with 74 points [range, 20 to 97 points]; p < 0.05). The scores after nailing were higher in each of the categories (Table III).
There were nine smokers in each matched-pair group. The mean time to union was twenty-two weeks (range, twelve to fifty-two weeks) for the smokers who had had nailing compared with twenty-three weeks (range, sixteen to thirty-six weeks) for the smokers who had been managed with a cast. With the numbers available for study, we could not detect a significant difference between the two groups (p > 0.05).
The twenty-five patients who had had nailing returned to work significantly sooner than did the twenty-five who had been managed with a cast (mean, four compared with 6.5 months; p < 0.05).
Closed fractures of the tibial shaft traditionally have been treated with closed reduction and an above-the-knee cast, with conversion to a patellar ligament-bearing cast or a cast-brace at about one month13,15. Sarmiento et al. reported that closed treatment with use of a prefabricated functional below-the-knee brace was effective in a study of 1000 closed diaphyseal fractures of the tibia14. However, those authors had strict criteria for use of the fracture-brace. An intact fibula is a relative contraindication to functional bracing because angular deformity is more likely to develop14. Fractures with more than one to two centimeters of initial shortening also are a contraindication to functional bracing because the initial amount of shortening indicates the ultimate amount of shortening after healing14. Additionally, an inability to bear weight while wearing the patellar-ligament brace or the fracture-brace is predictive of a longer time to union. Sarmiento et al. had thirty years of experience with the treatment of tibial fractures with a functional brace, and the results of their study14 were certainly excellent—the best reported in the literature, to our knowledge. However, their results have not been reproduced by other authors5,12,17. In fact, in an earlier study by Sarmiento et al., eighty-five (28 per cent) of 306 patients managed with functional bracing had angulation of more than 5 degrees15. Böstman reported that closed reduction is difficult to maintain when there is initial displacement of more than half of the width of the tibia, especially with spiral fractures of the distal third3.
There have been few studies comparing the results of closed non-operative treatment with those of closed intramedullary nailing of fractures of the tibial shaft. Hooper et al. performed a prospective, randomized study of patients who had a closed fracture of the tibial shaft with at least 50 per cent displacement or at least 10 degrees of angulation in any direction. Thirty-three patients (Group A) were managed with closed reduction and an above-the-knee cast for four weeks, followed by use of a patellar ligament-bearing cast until union. Full weight-bearing was encouraged. Twenty-nine patients (Group B) had closed intramedullary nailing of the fracture. The mean time to union was 18.3 weeks for Group A compared with 15.7 weeks for Group B (p < 0.05), and the mean time until the patients returned to work was twenty-three weeks for Group A and 13.5 weeks for Group B (p < 0.01); both of these differences were significant. There was significantly more angular deformity in Group A, with nine patients having varus or valgus angulation of more than 5 degrees and no patient in Group B having an angular deformity (p < 0.01). Shortening of one to two centimeters occurred in six patients in Group A and in one patient in Group B (p < 0.01). No infections were reported in either group.
The results reported by Hooper et al. were similar to those in the current study. As was pointed out by Sarmiento et al.14, performance of a randomized study may result in treatment with a cast-brace for a fracture for which such treatment is not recommended—for example, a fracture with an intact fibula or one associated with more than two centimeters of shortening as seen on the initial radiograph. Thus, the study may be biased toward the group managed with intramedullary nailing.
Puno et al. reported markedly improved results for ankles that had been treated with intramedullary nailing. Their study established a direct relationship between the extent of residual malalignment of the limb and the clinical result for the ankle but not that for the knee12. The responses to the SF-36 questionnaire in our study showed a considerable difference, in the patients' perceptions of bodily pain, general health, and physical role functioning, between the group that had been managed with a cast and the group that had had nailing.
Our study had several drawbacks in that it was a retrospective review of closed fractures of the tibial shaft that had been treated by two surgeons on the basis of their personal preference and the characteristics of the individual fracture rather than according to protocol. The time to union was significantly longer after treatment with a cast, but this could have been affected by the physicians not encouraging the patients to bear weight fully as early as they could have. Five patients who had been managed with a cast had a non-union; however, one of them had an intact fibula, which could have contributed to the non-union. By identifying a matched group of patients who had had long-term follow-up, we tried to minimize some of the selection bias in our study.
We recognize that, when the proper indications for intramedullary nailing or use of a cast are present, the choice of treatment of a closed fracture of the tibial shaft should be made by the informed patient. The results of the current study show that closed intramedullary nailing may yield better functional results than use of a cast for many patients who have a displaced closed fracture of the tibial shaft.