Background: Limb-salvage surgery and segmental reconstruction for the treatment of lower extremity osseous tumors in the pediatric population have been described in the literature, but there is little consensus regarding the optimal surgical treatment for this patient population.
Methods: A systematic review of the literature was performed to identify studies focusing on limb-salvage procedures in pediatric patients who were managed with one of three reconstructions with use of a metallic endoprosthesis, allograft, or allograft-prosthesis composite. Data were segregated according to the excised and reconstructed anatomical location (proximal part of the femur, total femur, distal part of the femur, proximal part of the tibia) and were collated to assess modes of failure and functional outcomes of each reconstruction type for each anatomic location.
Results: Sixty articles met the inclusion criteria; all were Level-IV evidence, primarily consisting of small, retrospective case series. Infection was a primary mode of failure across all reconstruction types and locations, whereas allograft reconstructions were susceptible to structural failure as well. The rate of failure in the pediatric population correlated well with previously published results for adults. The incidence of subsequent amputation was lower in the pediatric population (5.2%) than has been reported in adults (9.5%) (p = 0.013). Meaningful growth of expandable metallic endoprostheses was reported in the literature, with an overall rate of leg-length discrepancy of 13.4% being noted at the time of the latest follow-up. The Musculoskeletal Tumor Society (MSTS) questionnaire was the most consistently used outcome measure in the literature, with average scores ranging from 71.0% to 86.8%, depending on reconstruction type and anatomic location.
Conclusions: The current state of the literature detailing the surgical and functional outcomes of segmental reconstruction for the treatment of pediatric bone tumors is limited to Level-IV evidence and is complicated by under-segregation of the data by age and anatomical location of the reconstruction. Despite these limitations, pediatric limb-salvage surgery demonstrates satisfactory initial surgical and functional outcomes.
Level of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Prior to the advent of therapeutic chemotherapy and radiation, bone sarcoma of the lower extremity was routinely treated with radical excision. As the majority of bone sarcomas occur about the knee and hip, patients were often managed with above-the-knee amputation or hip disarticulation1-6. Even with this drastic intervention, the survival rate following amputation was consistently poor, especially for children. For example, in 1975, Price et al. reported a 12% five-year disease-free survival rate after early amputation in a study of 125 children7. That same year, Campanacci and Cervellati presented their experience with the treatment of extremity sarcoma, demonstrating a ten-year survival rate of 11% following radical amputation; in that study, no children under the age of fifteen years survived8. Within a year after those reports, chemotherapy and radiation began to show promise9-12. Five-year survival rates quickly increased from 20% to 50%-65%13-19. With this development providing improved ability to control part of the disease process, orthopaedic oncologists were able to more consistently attempt limb salvage, rather than amputation, as a component of multimodal treatment intervention.
It is now well established that patient survival following surgical and adjuvant treatment is not different when limb-salvage surgery is compared with amputation for the majority of bone sarcomas19-22. The transition from amputation to limb salvage, allowed by the development of chemotherapy and radiation, occurred under the assumption that patient outcomes would be better with salvage, both functionally as well as psychologically23-28. For adults, even with rates of surgical failure approaching 25%, limb salvage is now the prevailing treatment for most bone tumors about the hip and knee29-31. With this acceptance of limb salvage, a growing body of literature is being developed, focusing on functional outcomes, gait characteristics, and quality of life32-37. As a result, new prostheses are being developed to address the surgical complications and functional limitations of previous reconstructions.
Historically, however, limb salvage has been less eagerly adopted for children as compared with adults. This difference is partly due to the opinion that children, particularly those who are far from skeletal maturity, can more readily adapt to amputation than adults can. In this light, a single amputation procedure with a reliable and functionally acceptable outcome may be preferable to a limb-salvage procedure that has a high likelihood of complication, often requiring future surgical interventions38-47. Additionally, surgical reconstructions in children are complicated by the fact that a physis is typically sacrificed and clinically important leg-length discrepancy can ensue. In this view, the younger the patient, the more appropriate an amputation seems to be. Conversely, advocates of limb preservation have countered with data in an attempt to demonstrate acceptable complication rates and outcomes48-53. Additionally, in order to address limb-length inequality following resection of a physis, expandable prostheses have been developed51,54-60. Overall, however, the evidence supporting the optimum treatment for limb salvage in a pediatric patient with a tumor has been limited and inconclusive. This lack of evidence is in part due to the relative rarity of bone sarcoma in children. The vast majority of the published data are derived from small case series. As an additional limitation, many manuscripts pool data from a wide range of ages and across anatomical sites. As Henderson et al. demonstrated, reconstructions at the proximal part of the femur, distal part of the femur, proximal part of the tibia, and upper extremity are unique and behave differently in terms of postoperative complications and failure31. Whether there are unique differences in surgical failures between pediatric subjects and adults has yet to be determined. The purpose of the present article is to systematically review the literature on children and adolescents undergoing limb salvage because of a lower extremity bone tumor. Outcomes are assessed in an anatomically site-specific manner, including modes of failure and functional outcomes, and an attempt was made to pool the data to identify any potential trends or specific deficiencies associated with the use of limb salvage for the treatment of tumors in pediatric patients.
A computer-based literature search was performed to identify all studies examining the outcomes for adolescent patients who had undergone limb salvage following tumor excision. Our search utilized the following databases: PubMed (1947 to present), CINAHL (1981 to present), RECAL Legacy (1900 to 2007), Cochrane Database of Systematic Reviews (1992 to present), Cochrane Central Register of Controlled Trials (CENTRAL) (1992 to present), EMBASE (1974 to present), PMC-NIH (2000 to present), Web of Science (1900 to present), and Google Scholar (early 1900s to present).
The search terms for the PubMed search are displayed in a table in the Appendix. All databases were searched from their inception point to June 9, 2014.
Three individuals reviewed the articles to determine whether they satisfied the inclusion criteria. When disagreement occurred, the third reviewer served as a tiebreaker. The following criteria were applied for article inclusion to ensure that the question of outcomes for adolescents undergoing limb salvage following tumor excision was addressed. First, the article must have been an original research study in a peer-reviewed journal that was published or translated into English. Second, the subjects must have undergone lower extremity salvage surgery because of a tumor by the age of seventeen years. Third, the investigators must have attempted to determine a functional outcome after a minimum duration of follow-up of twelve months; if a failure occurred prior to twelve months of follow-up, it was included in an attempt to capture all reported failures.
Additional subject definition and exclusion criteria were applied to allow for the most homogeneous comparisons across studies. Subjects must have had a reconstruction that was clearly identified as a proximal femoral reconstruction, total femoral reconstruction, distal femoral reconstruction, or proximal tibial reconstruction. Reconstructions were included if they had been performed with use of a metallic endoprosthesis, allograft, or allograft-prosthesis composite. Surgical outcome and functional data must have been clearly attributable to each subject. Articles or data were excluded if the anatomical site was not clearly identifiable, subjects were over seventeen years of age at time of surgery, or the indication for surgery was non-oncologic. Intercalary resections, arthrodesis, rotationplasties, and amputations were excluded. When multiple studies from a single institution were found with overlapping inclusion dates and similar anatomic locations, the study with the greatest number of subjects or the most clearly reported data was included. When studies included reconstructions involving multiple anatomical sites or included adult and pediatric patients, an effort was made to differentiate the data; if the data were not clearly attributable to the studied reconstruction and subject, the data were excluded.
Using the above criteria, one reviewer extracted information from the included studies. Information was collected on subject demographics, failure mode, lengthening, and function after limb salvage. Information regarding the study design and methods was also collected to assist in interpretation and quality assessment.
Assessment of Study Quality
As the quality of a study can be assessed according to many different criteria and the validity of these criteria have not been determined, quality scores were not used to exclude studies but rather to assist in the explanation of the review findings61. Given the relatively rare nature of the condition being examined and the current lack of homogeneity across studies, we included all evidence types above Level V.
Assessment of Heterogeneity
We assessed the studies for homogeneity to allow for meta-analysis and decided that, given the heterogeneity of all criteria and the level of evidence across studies, the review would be descriptive, except when comparisons between similar outcomes could be made.
After extraction, the data were compiled on a spreadsheet, with each type of reconstruction and anatomical location separated by study. The data were then compiled for averages, totals, and ranges (Table I).
Descriptive data were calculated for patient demographic characteristics. We collected frequency data on failure mode and location, patient function, and lengthening characteristics according to anatomic location. Failure proportions were compared between prosthetic types with use of the Fisher exact test. A lack of sufficient data in some anatomic and prosthetic categories prevented comparison across all types. We then compared the rates of failure and subsequent amputation between pediatric patients with metallic endoprostheses and adults as reported in the previously published literature. The functional outcomes as determined with the Musculoskeletal Tumor Society (MSTS) score were not compared because of inconsistency in reporting but are presented descriptively.
The literature search returned 1322 articles (Fig. 1)62. The titles and abstracts were reviewed against the inclusion and exclusion criteria. After this initial review, 971 articles were excluded. The remaining 351 articles were read, and an additional 289 were excluded because they did not meet the inclusion criteria. The most common reason for exclusion at this stage was failure of the manuscript to clearly differentiate the data on outcome or failure by age or anatomic location of the reconstruction. This criterion alone led to the exclusion of 175 manuscripts. Sixty distinct manuscripts met final inclusion criteria: forty-one focusing on metallic endoprostheses, eleven focusing on allograft, and ten focusing on allograft prosthesis composites36,48-51,53,63-116. Two studies assessed both allograft and allograft-prosthesis composites with sufficient clarity to remain included in this review101,108.
Study Designs and Content
All of the included studies were classified as Level-IV evidence under The Journal of Bone & Joint Surgery guidelines117, with the overwhelming majority being retrospective case series. No randomized study was identified. Sixteen of the sixty selected studies evaluated multiple anatomical sites, including nine studies on metallic endoprostheses36,49-51,53,85,86,88,97, four on allografts48,102,106,108, and three on allograft-prosthesis composites102,109,116. These studies were included because the outcome data were clearly listed for each reconstruction. After the data were collected, 393 patients with a metallic endoprosthesis were identified, along with 105 patients with allografts and twenty-three patients with allograft-prosthesis composites. Table I summarizes the demographic data according to prosthetic type, anatomic location, age, sex, and duration of follow-up. When analyzed by site-specific reconstruction, the majority of the studies had fewer than ten subjects. The average number of subjects was 6.6 for studies on metallic endoprostheses, 7.0 for studies on allografts, and 1.8 for studies on allograft-prosthetic composites. A figure in the Appendix demonstrates the distribution of the number of subjects per study.
Failure was defined by the need to revise the prosthesis, implant, or surrounding soft tissue following the index procedure. Failure modes were categorized as described by Henderson et al. and are listed in Table II31. Fifty-seven reports described failure characteristics for at least one subject according to prosthetic type and anatomic location. The rate of revision of metal endoprostheses ranged from 15.4% to 55%, and subsequent amputation was reported in 3.4% to 10.5% of cases, depending on the anatomic location. Infection was a main mode of failure of metallic endoprostheses across all anatomic sites, whereas proximal femoral reconstructions and distal femoral reconstructions demonstrated additional susceptibility to aseptic loosening. Allograft revisions were reported in association with 38.7% of distal femoral replacements and 57.1% of proximal tibial replacements. Only one subject was identified as having had a proximal femoral allograft replacement, and no total femoral allografts were found in the literature. The most common mode of failure for allografts was structural failure, followed by infection. The rate of revision of allograft-prosthesis composites ranged from 25% to 40% depending on anatomic location, with no subsequent amputations across locations. The total number of subjects identified as having an allograft-prosthesis composite, however, was limited (Table I and Table II). The most common modes of failure across all anatomic sites for allograft-prosthesis composites were aseptic loosening and infection. Sufficient numbers were available so that allograft proximal femoral and proximal tibial reconstructions could be compared with metallic endoprosthesis proximal femoral and proximal tibial reconstructions; likewise, allograft-prosthetic composite proximal femoral reconstructions were compared with metallic endoprosthesis proximal femoral reconstructions (Table III, Fig. 2). The only difference in failure rates that reached significance was between metallic endoprosthesis proximal tibial reconstructions and allograft proximal tibial reconstructions (33.4% compared with 57.1%; p = 0.004).
Henderson et al. performed an anatomically specific systematic review of metallic endoprosthesis reconstructions in the adult population31. The current systematic review of pediatric metallic endoprosthesis reconstructions was compared with that previously published review (Table IV). Compared with their adult counterparts, the pediatric population demonstrated a significant increase in the failure rate at the proximal part of the femur (55% compared with 20.1%; p < 0.001) but a significant decrease in failure rates of total femoral reconstructions (15.4% compared with 47.6%; p = 0.004). Metallic endoprosthesis distal femoral and proximal tibial reconstructions demonstrated no difference in failure rates between the adult and pediatric populations according to the available literature. The need for subsequent amputation in the metallic endoprosthesis group was analyzed in the current series and compared with the findings reported by Jeys et al. (Table V)118. A significant decrease in the rate of amputation following index metallic endoprosthesis limb salvage was seen for distal femoral replacements in the pediatric population (p = 0.014).
Twelve studies had extractable data on lengthening procedures for metallic endoprosthesis expandable prostheses (Table VI)49-51,53,67,72,85,88-90,96,97. The number of expansions per subject was greatest for the total femoral replacement group (average, 6.9 expansions), for an average growth of 84.8 mm. The number of expansions per subject was lowest for the proximal femoral replacement group (average, 2.95 expansions), for an average growth of 29.9 mm. The likelihood of using the expansion mechanism of an implanted proximal or distal femoral expandable reconstruction was 52% and 51%, respectively. Expandable total femoral replacements were lengthened 82% of the time. Failure of the lengthening mechanism requiring explantation and revision occurred in 0% to 3.4% of the reported cases, depending on the anatomical location (Table II).
Forty-four studies evaluated functional outcomes. The MSTS questionnaire was the only functional measure that was used with enough consistency across studies; these scores are summarized across prosthetic types and anatomical sites in Table VII119. The average MSTS scores ranged from 76% to 82.5% across anatomical sites for metallic endoprostheses, with the total femoral group having the lowest functional score (76%) and the proximal tibial group having the highest (82.5%)36,50,51,53,63,67,69-75,77,80-88,90-94,96-98. The average scores ranged from 75.9% to 85.0% across anatomical sites for allograft reconstructions, with the distal femoral group having the lowest functional score (75.9%) and the proximal tibial group having the highest (85.0%)48,102-104,106,107. The average scores ranged from 71.0% to 86.8% across anatomical sites for allograft-prosthesis composite reconstructions, with the distal femoral group having the lowest functional score (71.0%) and the proximal femoral group having the highest (86.8%)101,102,110,111,115,116. The number of subjects was low for all groups, and the results of statistical analysis are not presented because of the lack of uniform reporting and concern over selection bias.
Range of motion was reported in eighteen studies of metallic endoprosthesis reconstructions, one study of allograft reconstructions, and three studies of allograft-prosthesis composite reconstructions50,63,67,69,70,73,75,77,81,84,90-92,94-96,98,101,109,113,120,121. Gait characteristics were described in eight studies of metallic endoprosthesis reconstructions, predominantly consisting of subjective reports, such as the presence or absence of a limp or the need for an assistive device51,63,82,87,89,94-96. We identified no studies describing gait characteristics of pediatric patients managed with allograft or allograft-prosthesis composite reconstructions. The data on range of motion and gait are listed in Table VII.
The recommended treatment following wide resection of a lower extremity osseous tumor in the skeletally immature patient remains a matter of debate. The present systematic review presents both the state of the literature as well as the available outcome data for three types of pediatric limb salvage: metallic endoprostheses, allografts, and allograft-prosthesis composites.
Overall, the literature on pediatric limb salvage is lacking in both quality and quantity. No study presented more than Level-IV evidence, and 82.6% of the reviewed studies included ten or fewer subjects per anatomic location. No study of allograft-prosthesis composite reconstructions included more than ten subjects, and no study of allograft reconstructions included more than twenty subjects per anatomic location. The largest series was a study on metallic endoprosthesis reconstructions that included thirty-three distal femoral replacements76. There were a total of twenty-six articles with one subject per anatomic location after the data were acceptably parsed. One hundred and seventy-five studies were excluded because of over-inclusion, the inclusion of pediatric and adult subjects simultaneously, or the evaluation of multiple anatomical locations without the ability to attribute outcome data to a given subject or group. In part, these results reflect the rare nature of bone sarcoma in children, but they also highlight the need for standardized and segregated reporting. Many pediatric patients have undergone tumor resections and subsequent reconstructions without clear representation in the literature.
Despite the above-mentioned limitations, we believe that an initial assessment of the outcomes of pediatric limb salvage can be made on the basis of the present systematic review. Three tentative observations can be made: (1) the rates of failure are high but are mostly in line with those in adults, (2) meaningful metallic endoprosthesis lengthening can be achieved for skeletally immature patients, and (3) reports of functional outcomes are limited but have shown satisfactory results on MSTS questionnaires.
Infection led to failure of each of the three types of reconstructions across the studied anatomical locations. Insufficient numbers of patients precluded statistical comparison of failure across all pediatric groups, but, when comparisons were possible, a significant risk difference was found between metallic endoprosthesis reconstructions and allograft reconstructions of the proximal part of the tibia. This difference was due to the increased incidence of structural failure in the allograft group (32.9%).
When compared with their adult counterparts as reported in the literature, pediatric patients undergoing metallic endoprosthesis reconstructions fared well. Comparison of the current review with that performed by Henderson et al.31 indicated that only metallic endoprosthesis proximal femoral reconstructions were at higher risk of failure in the pediatric population. The total femoral reconstruction failure rate was 15.4% in pediatric patients, compared with 47.6% in adults; this difference reached significance (p = 0.004). With regard to the need for subsequent amputation, the results of the current review compare favorably with the adult data provided by Jeys et al.118. Of note, Jeys et al. did not report on subject age other than to state that it had no relationship with amputation after endoprosthetic replacement118; the results of the current study suggest otherwise. The overall rate of amputation following metallic endoprosthesis reconstruction was 5.2% in the current review, compared with a rate of 9.5% following similar anatomic reconstructions in the study by Jeys et al.118. This significant difference was due to the decreased frequency of amputation following distal femoral reconstruction. Such comparison was not feasible with the allograft and allograft-prosthesis composite reconstructions because of the limited numbers of patients in the current review and the lack of comparable systematic reviews involving the adult population. It must be noted that the failure rates reported here should be viewed as minimums because of a potential selection bias as well as inconsistent follow-up criteria between studies. The minimum accepted follow-up in this review (twelve months) may miss many local recurrences or late failures. We included this relatively short follow-up to ensure capture of early failures that were then no longer followed or reported in the literature. Overall, the average duration of follow-up ranged from twenty-five to eighty-nine months, depending on reconstruction type and anatomic location (Table I).
Meaningful growth for pediatric patients with metallic endoprosthesis expandable prostheses was demonstrated in the literature49-51,53,67,72,85,88,89,96,97. The majority of the patients with an expandable prosthesis achieved limb-length equality, with failure of the expansion mechanism being reported ≤3.4% of the time, depending on anatomical location (Table II). The definition of limb-length inequality was not uniform, but, according to the various authors’ definitions, was typically noted as a 2 to 3-cm side-to-side difference. With these broad inclusion criteria, the overall reported likelihood of a limb-length difference was 13.4% in patients with an expandable prosthesis, with proximal femoral reconstruction accounting for the greatest risk (23.5%). Average expansion varied by site of reconstruction, but episodes of total lengthening of >10 cm were reported in all but the proximal tibial cohort. This finding is particularly meaningful as the youngest subjects in the metallic endoprosthesis proximal femoral reconstruction, total femoral reconstruction, distal femoral reconstruction, and proximal tibial reconstruction groups were three, two, five, and four years old, respectively. The included studies involved various methods of lengthening, but an evolution from open lengthening procedures to minimally invasive to closed lengthening devices was observed in the present review.
Data on functional outcome as measured with the MSTS questionnaire suggested satisfactory results, with average scores ranging from 71% to 87%, depending on anatomical location and reconstruction type. More objective measures of function were scarce in the literature, and the majority of studies with such measures had to be excluded because of over-inclusion. For example, Winter et al. measured the overall amount and intensity of activity throughout the day for young patients undergoing lower extremity limb salvage122. That study effectively demonstrated the extent of surgery-induced limitations when the patients were compared with a control group on the basis of objective data, but the study had to be excluded from the current review because of the inclusion of different anatomical locations (the proximal part of the femur, distal part of the femur, and proximal part of the tibia) in one group. A universal or readily used objective measure for function was not found in the literature. The presence or absence of a limp was reported in twenty-two patients with metallic endoprosthesis reconstruction, with ten of these patients demonstrating no residual limp. When reported, quadriceps lag was noted in association with 16.7% of metallic endoprosthesis distal femoral reconstructions and 33.3% of metallic endoprosthesis proximal tibial reconstructions. Again, this finding suggests a potential for adequate function in this cohort of patients, but more rigorous reporting needs to be performed.
In addition to the deficiencies with the available literature noted above, several other limitations of the present study should be noted. First, homogeneity within the data is assumed, whereas experience informs us otherwise. That is, the patients were grouped according to general age, anatomic location, and prosthetic type, but additional segregation may be appropriate to determine trends of surgical and functional outcomes. For example, the outcomes of reconstructions of the distal part of the femur may be influenced by the extent of soft-tissue excision and maintenance of the extensor mechanism. Furthermore, soft-tissue failure in a patient with a metallic endoprosthesis may be different from that in a patient with an allograft-prosthesis composite. The available literature does not permit analysis of such nuance, both because of limitation in number of patients and because of limitation in detailed descriptions of failure type. The inability of the literature to permit nuanced analysis holds true for subanalysis of different metallic endoprosthesis types as well as technical differences (e.g., the use of cemented versus ingrowth stems). While these subanalyses were beyond the capacity of the current literature and beyond the intent of the present review, they warrant consideration in future investigations.
Next, patients may be further subdivided by age into the very young, the skeletally immature, and the nearly skeletally mature, which may provide additional insight into the future needs for these patients. This subanalysis was not completed in the present study as it would have further restricted inclusion and prevented achieving the study objectives; however, future studies should treat these groups as distinct and should clearly segregate the subjects accordingly. Second, evolution of treatment technique also may have had an effect on limb-salvage outcomes. The current review included patients who were enrolled from 1970 through 2014. Trends in reconstruction type and improvements in outcomes may have been made in those forty-four years, but, again, the reviewed data did not lend themselves to such analysis.
It is important to recognize that not all conditions that necessitate study lend themselves to the traditional hierarchies of evidence. Individuals cannot be randomized to amputation, limb salvage, or different types of limb salvage. These complicated decisions are made after consultation with the patient and family and are moderated by the presentation of the disease and the resources available to the health-care team at the time of diagnosis. Therefore, it is perhaps more important for conditions that have treatment options that cannot be randomized to have standardized reporting methods that allow for more robust analysis of data across individual studies. This statement is not intended to be a criticism of previous literature. In following the history of advances in both cancer treatment and limb salvage, it is appropriate that earlier studies focused more on salvage success than on function of the limb. Now that the survival of the limb is less in question, the literature may focus on the characteristics associated with minimizing complications and maximizing function.
Moving forward, we suggest age-specific and anatomy-specific reporting of data. In addition, collaborative work between institutions would improve the power of the studies presented. Finally, ongoing use of the MSTS score is encouraged because of its historic value as the established outcome measure, but the addition of objective measures such as range of motion, self-selected gait velocity, total activity volume throughout the day, and the ability to negotiate stairs would add insight into the daily function of these patients.
The optimum surgical reconstruction for the treatment of osseous tumors of the lower extremity in pediatric patients remains unclear. The present systematic review demonstrates that while failure rates of reconstruction remain high, they are largely comparable with those seen in adults. Satisfactory functional outcome scores and low rates of amputation indicate that limb salvage is a suitable and stable intervention in the pediatric population. Additional study with standardized outcome measures and collaborative pooling of data should greatly advance our understanding of orthopaedic reconstructions following en bloc tumor excision and should also assist in improving prosthesis design.
Note: The authors acknowledge the efforts of Robert H. Riffenburgh, PhD, in bringing this investigation to fruition.
Source of Funding: No funding was provided in the development or completion of this manuscript.
Investigation performed at the University of South Florida, Tampa, and the H. Lee Moffitt Cancer Center, Tampa, Florida
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated