Advancements in medicine and surgery have greatly improved the ability to reconstruct severely injured lower extremities and improve functional outcome. However, limb salvage and major limb reconstructions often result in a protracted course of treatment and rehabilitation with long-term functional, psychosocial, and financial consequences. Some investigators have suggested that the functional outcome after a successful limb reconstruction is often poorer than that after treatment with early amputation and a functional below-the-knee prosthesis1-5. Others have found the results of reconstruction to be equivalent or better than those after amputation6,7. Reconstruction is associated with a longer rehabilitation time and a higher risk of complications, additional surgical procedures, and rehospitalization6,7, whereas amputation is associated with high dissatisfaction rates because of discomfort caused by the prosthetic device, phantom pain, and residual-limb skin problems8,9.
Several authors have reported on outcomes and decision-making associated with acute lower-extremity-threatening injuries1-5,10, but long-term outcomes after treatment of late posttraumatic lower-extremity complications such as complex malunion or nonunion of the tibia or femur have not been thoroughly documented.
The aim of the present study was to evaluate the long-term functional and psychosocial outcomes in a series of patients who had been treated with reconstructive surgery for complex malunion or nonunion of the tibia or femur by one surgeon.
Inclusion and Exclusion Criteria
Between 1982 and 1996, more than 200 patients were referred to our institution (Massachusetts General Hospital, Boston, Massachusetts) because of a complex malunion or nonunion of the tibia or femur. The criteria for inclusion in the present retrospective case series were (1) skeletal maturity and (2) treatment, at our institution by the senior author (J.B.J.), of complex malunion or nonunion of the tibia or femur, with complex defined as concurrence of one of the following factors: osteomyelitis, a segmental defect larger than 1 cm, nonunion or malunion for twelve months or more, two or more prior surgical procedures, or moderate or severe arthritis of adjacent joints. The exclusion criteria included a duration of follow-up of less than sixty months after the index treatment, mental disability, and an inability to return for clinical evaluation.
A manual search of medical records identified 152 patients who had been treated for complex malunion or nonunion of the tibia or femur by the senior author. The institutional review board approved the study protocol. Of the 152 patients, forty-six (30%) were located and 106 had been lost to follow-up. All patients were invited to return for assessment of the lower extremities. Of the forty-six located patients, seven had died and ten declined to participate. In total, 123 patients were excluded because of an inability or unwillingness to return for clinical evaluation. The remaining twenty-nine patients met the inclusion criteria and were included in the study. All participants gave consent to participate.
To investigate if the study cohort was a representative sample of the full cohort (n = 152), several demographic and injury characteristics were compared between the in-sample (n = 29) and out-of-sample (n = 123) cohorts, and no significant differences were found (Table I).
Patients
The study group included ten women and nineteen men with an average age of fifty-six years (median, fifty-three years; range, thirty-one to eighty-five years) at the time of the latest follow-up examination. The initial injury resulted from a motor-vehicle accident in twelve patients, a motorcycle accident in four, a collision involving a pedestrian and a motor vehicle in three, and a fall in ten. Complications involved the tibia in twenty-three patients and the femur in six. Eight patients had a complex malunion, and three of them had a united segmental defect (see Appendix). Twenty-one patients had a complex nonunion, and seven of them had osteomyelitis.
The majority of patients (twenty-three, 79%) had internal fixation as their initial fracture treatment and were referred to our department because of implant failure, deformity, and/or infection. Two patients were initially treated with an external fixator; one, with percutaneous pinning; and three, with cast immobilization. The choices of surgical procedures for the treatment of the postinjury complications were based on the location and nature of the complications and are summarized in Table II. Seven patients had an infected nonunion of the tibia (five) or femur (two), and six of the seven had had at least one prior surgical procedure (range, one to twenty-eight procedures). The treatment protocol used in these patients, as previously reported by the senior author, consisted of sequential debridement11,12, external fixation13, and free-flap soft-tissue coverage followed by autogenous bone-grafting14,15 (Figs. 1-A through 1-E, and Appendix).
Fourteen of the twenty-nine patients underwent additional surgery beyond the index (complication) surgery, including implant removal (four), revision internal fixation for persistent nonunion or implant failure (four), total hip arthroplasty (two), total knee arthroplasty (two), Achilles tendon lengthening and capsular release for equinus contracture (one), and internal fixation because of a refracture (one). There were no additional surgical procedures for persistent infections or deformities.
Clinical Evaluation
At the time of the latest follow-up, all patients were assessed by an independent physician (G.A.B.)—i.e., one who was not involved in the patient's care. Evaluation consisted of a personal interview and a physical examination of the lower extremities. The patients were interviewed about pain, the overall function of the lower extremities, and the psychosocial consequences of their injury and treatment. Physical examination of the lower extremities consisted of assessment of the hip, knee, and hindfoot; range of motion of the adjacent joints; muscle strength; neurovascular status; and presence or absence of tenderness and swelling. Muscle strength was rated subjectively with use of a 6-point Likert scale ranging from 0 (no contraction) to 5 (normal strength). The contralateral extremity was used as a control. Gait was recorded on videotape and assessed by both the treating physician and an independent physician (see Appendix).
Patient-Based Assessment Questionnaires
Patients completed the Lower Extremity Functional Scale (LEFS), the Short Form-36 (SF-36), and 10-point visual analog scales for patient satisfaction and pain.
The LEFS was used to measure patient-rated overall lower-extremity function16. The overall score ranges from 0 to 80, with higher scores indicating better function. The SF-36 was used to measure health-related quality of life17. The physical and mental component summary scores, as well as the overall summary score, range from 0 to 100, with higher scores indicating better health.
Radiographic Evaluation
An independent observer assessed anteroposterior and lateral radiographs of the adjacent joints to determine alignment and radiographic findings of degenerative changes in these joints. With use of radiographic criteria, osteoarthritis was classified according to the grading system of Knirk and Jupiter18 as Grade 0 (none), Grade 1 (slight joint-space narrowing), Grade 2 (marked joint-space narrowing and osteophyte formation), and Grade 3 (bone on bone, osteophyte formation, and cyst formation).
Statistical Analysis
Basic descriptive statistics were used to summarize study sample demographics. Statistical comparisons of the in-sample and out-of-sample patient cohorts were performed with use of an independent t test for continuous data and a chi-square test for categorical data, with the level of significance set at p < 0.05. Pearson and Spearman correlations were performed (for continuous and categorical data, respectively) in order to correlate the outcome scores of the LEFS, the SF-36, pain, and patient satisfaction with demographic, treatment, and outcome parameters.
Source of Funding
Financial support was received from the Netherlands Organisation for Scientific Research (NWO).
A chart review revealed that 121 of the 123 patients in the out-of-sample cohort had healing following treatment of the complex malunion or nonunion of the tibia or femur and that none of the patients had died either prior to healing or of related causes. The remaining two patients underwent successful below-the-knee amputation after multiple failed reconstructive procedures. The in-sample cohort of twenty-nine patients was evaluated at an average of twenty years (median, nineteen years; range, twelve to thirty-five years) after the injury. The study group was divided in two subgroups, depending on whether the complex malunion or nonunion involved the femur or tibia. The complications, treatment, and outcomes of all patients included in the study group are summarized in Table II. All malunions and nonunions healed after treatment.
Clinical Evaluation
The joint ranges of motion, with comparison with the contralateral side, are summarized in a table in the Appendix. Lower-extremity muscle strength was normal (5/5) in all patients except two: one with a traumatic peroneal nerve injury had a score of 3/5 for ankle motion of the injured lower extremity, and one with a sedentary lifestyle because of osteogenesis imperfecta had a score of 4/5 for both knee extension and ankle motion of the injured lower extremity. With the exception of the patient with the peroneal nerve injury, all patients had intact neurovascular function. Two patients had tenderness and swelling in the knee and five patients, in the ankle.
Five of the six patients in the femur cohort had a slight limp. Of the twenty-three patients in the tibia cohort, twelve had no limp, nine had a slight limp, one had a moderate limp, and one had a severe limp (requiring bilateral crutches). All patients were able to bear full weight and to walk at least one block. Four patients used a walking aid: two used a cane, one used bilateral crutches, and one used a walker. Two patients wore a brace on a daily basis. The remaining twenty-three patients did not use any support for walking. Six patients (three in each cohort) had a slight-to-moderate limb-length discrepancy (ranging from 1 to 4 cm), which could be corrected with a lift; only one patient with a 4-cm tibial discrepancy was symptomatic.
Patient-Based Assessment Questionnaires
The LEFS outcome tool revealed that twenty patients (69%) experienced moderate-to-severe difficulties (a score lower than 60) with carrying out activities because of their lower-limb disability. The mean LEFS score was 59 (median, 62; range, 35 to 70) in the femur subgroup and 53 (median, 53; range, 26 to 79) in the tibia subgroup.
The SF-36 results revealed that most patients were limited secondary to their general health perceptions with regard to their participation in vigorous activities (twenty-five), participation in moderate activities (fifteen), and walking more than 1 mile (1.6 km) (fifteen). Twenty-three patients experienced bodily pain to some extent, and in nearly all of them (twenty-two) it interfered with their normal work. The SF-36 scores for each health component subscale are compared with the scores for the general U.S. population in a table in the Appendix19.
General health was comparable with the U.S. norm, indicating that there was minimal influence of other substantial health problems on the outcomes measured with the other scales. Differences in comparison with the general U.S. population were largest for SF-36 scales sensitive to differences in physical health status, particularly physical functioning, role limitations due to physical health, and bodily pain. The role physical domain had the lowest average scores, with a substantially lower average score in the tibia subgroup than in the femur subgroup. The mean SF-36 score for patients with femoral malunion or nonunion was 75 (median, 79; range, 34 to 96), with a mean physical component score of 71 (median, 75; range, 32 to 94) and a mean mental component score of 75 (median, 76; range, 40 to 95). The mean SF-36 score in the tibia subgroup was 60 (median, 58; range, 29 to 95), with a mean physical component score of 54 (median, 54; range, 26 to 92) and a mean mental component score of 68 (median, 71; range, 35 to 97).
The mean visual analog scale score for satisfaction was 9.0 (median, 10; range, 1 to 10), with 10 representing complete satisfaction. With the exception of one patient who had a protracted treatment course and higher expectations, all patients were moderately to very satisfied with the final result. Notably, several patients with a moderate-to-poor functional outcome were very satisfied as they had faced amputation and found salvage of the lower extremity to be more important than good function. The mean pain rating on the visual analog scale was 3.0 (median, 3; range, 0 to 10), with 0 representing no pain and 10 indicating the worst pain imaginable. Two patients who had undergone major femoral reconstruction had continuous moderate-to-severe pain (7 and 8 of 10), but although the pain dominated their daily functioning, they were very satisfied with the outcome (8 and 10 of 10).
Radiographic Evaluation
Of the six patients in the femur subgroup, four had Grade-1 osteoarthritis in the hip and the remaining two had undergone total hip arthroplasty. One of the patients in the femur subgroup had undergone total knee arthroplasty for osteoarthritis at forty-four years of age, and four of the remaining five patients in that group had osteoarthritis in the knee, which was Grade 1 in two, Grade 2 in one, and Grade 3 in one. The femur had excellent alignment in all but one patient, who had 10° of valgus at the distal part of the femur.
In the tibia subgroup, six patients had osteoarthritis in the knee, which was Grade 1 in two, Grade 2 in one, and Grade 3 in three, and six had osteoarthritis in the ankle, which was Grade 1 in two, Grade 2 in three, and Grade 3 in one. This excludes patients who could not be graded for knee (one) or ankle (six) osteoarthritis because of total knee arthroplasty and ankle arthrodesis, respectively. Tibial malalignment was present in 39% of the patients and consisted of a mean deviation of 17.2° (median, 20°; range, 5° to 30°) from the tibial axis.
Return to Work and Prior Activities
Twenty-two patients were able to return to work within a mean of twenty-two months (median, six months; range, 0 to 120 months) after the time of injury, three were retired, and four patients remained unemployed. However, most patients had changed their career to one more suitable to their lower-extremity impairment. Nine patients could return to their prior form of athletics, within a mean of twenty-eight months (median, 18 months; range, 6 to 102 months) after the fracture; one returned to (and exceeded) the prior level of athletics despite both protracted medical treatment and a rehabilitation course totaling 102 months as well as a lower-extremity length discrepancy of 4 cm.
Predictors of Outcome and Satisfaction
A correlation analysis showed which parameters are predictors of functional and psychosocial outcome and satisfaction (see Appendix). A longer duration of a complication significantly correlated with a lower LEFS functional outcome score (rSpearman = —0.4; p < 0.05). Decreased lower-extremity muscle strength significantly correlated with lower outcome scores for the LEFS (rSpearman = 0.46; p < 0.05), SF-36 (rSpearman = 0.41; p < 0.05), and satisfaction (rSpearman = 0.52; p < 0.01). No other demographic, treatment, or outcome parameter that was not based on a score showed a significant correlation with a functional or psychosocial outcome or satisfaction.
There remains debate regarding the functional outcomes following reconstruction compared with those after amputation treatment of severe trauma-related lower-extremity injuries and complications. Return to work has been used as an indicator of outcome in many previous investigations20-22. In the present study, we assessed the long-term outcomes solely in patients who had undergone reconstructive treatment of a complex malunion or nonunion of the tibia or femur. The results show that the majority of patients experienced moderate-to-severe functional lower-limb limitations. SF-36 physical health profile scores of the subjects were found to be substantially lower than those of the general U.S. population, particularly in the physical functioning, role limitations due to physical health, and bodily pain domains. It is more meaningful to compare the long-term outcome scores of our subjects with those of patients who had undergone a trauma-related lower-extremity amputation, as reported in the study by Pezzin et al.20. The majority of the SF-36 health profile scores were very similar in the two studies. The largest difference between the two groups was in physical functioning. Compared with the patients who had had an amputation, the patients in our study group had substantially better scores for physical functioning (61 versus 52) and worse average scores for role limitations from physical pain (47 versus 49) and for bodily pain (54 versus 59). Approximately one-fourth of the patients in the study by Pezzin et al. experienced severe problems with the residual limb, with pain (including phantom pain) accounting for the majority of the problems.
Regardless of the levels of pain and disability, all but one patient were moderately to very satisfied with the final outcome. In contrast to a recent study on outcomes after severe traumatic lower-extremity injuries, which showed a highly significant (p < 0.001) correlation between pain and satisfaction23, our study did not demonstrate a correlation between pain and satisfaction (p = 0.06). Moreover, most patients expressed a high appreciation of the counseling regarding realistic outcome expectations that they had received during the process of decision-making. In particular, patients who had faced amputation of the lower extremity reported very high satisfaction rates, even with less optimal functional outcomes. Among all patients who had faced and/or considered posttraumatic amputation, only one reported regret about his previous decision to salvage the lower extremity, as he continued to experience severe pain in the lower extremity. Chronic pain is highly prevalent among persons with limb loss, regardless of the time since the amputation24.
The fact that only one patient had regrets regarding his decision to salvage the limb is most likely related to the high success rate (no amputations after a minimum of twelve years of follow-up) and overall high satisfaction rate. The average time to presentation to us was forty months, and the subsequent time to healing was another year or more. Despite this length of treatment time, the patients were generally satisfied. A meta-analysis of observational outcome studies on patient preference in retrospect showed highly heterogeneous results, with one study demonstrating that 75% of patients with failure of a limb salvage wished that they had opted for primary amputation25.
The patients’ general and mental health ratings were comparable with the U.S. norms, indicating that there was substantial impact on their physical health status only. Seventeen of the twenty-two patients who returned to work did so prior to their referral visit at our institution. However, most of these patients either had desk-based positions or had adapted their employment positions prior to referral to our institution. The time to return to work may therefore not be representative of treatment effectiveness in this series.
The Lower Extremity Assessment Project (LEAP), a study of the largest prospective cohort to date with severe lower-limb trauma, provided evidence that functional outcomes of reconstruction and amputation were comparable at the time of midterm and long-term follow-up6,26. Like previous studies, it consistently showed that physical and psychological outcomes were negatively influenced by specific demographic and socioeconomic factors, including poverty, lower education levels, and limited social support25,27-30. These risk factors as well as psychological distress and inappropriate coping mechanisms should be adequately screened for, and rehabilitation programs should focus more extensively on psychosocial aspects. Psychological rehabilitation and practical social support after severe injury are likely to be at least as important as physical rehabilitation.
The results of the current study should be interpreted cautiously. The major limitation is the high attrition rate, which raises the question of whether this study cohort was representative of the full cohort. The fact that no significant differences were found between the in-sample and out-of-sample groups provides some evidence suggesting that our sample was likely representative. Other limitations of our study are its retrospective design and the small study group, which was heterogeneous regarding the types and complexity of the malunions and nonunions. The strengths are that the observational cohort reflects the senior surgeon's strong preference to attempt a reconstruction in any patient and a follow-up that was long enough for us to assume that the reported functional results were close to the end results. Thus, the results of the present study add relevant clinical data to the current debate about the optimal management of these complications.
The results of the present study support our opinion that reconstruction can be a worthwhile endeavor and should be considered for all patients with a complex malunion or nonunion of the tibia or femur. Patients should be counseled well regarding realistic outcome expectations before they undergo surgery.