The authors are grateful for Dr. Malawer’s comprehensive consideration of this manuscript. Clearly, the depth of his clinical experience is evident in the issues he has raised. Dr. Malawer has raised concerns about factors including endoprosthetic design, surgical approach, retention and repair of the hip capsule, and their effect on clinical outcomes including dislocation rates. In a prospective study, criteria for quantitating these parameters could be used to evaluate the effect of partial hip capsule retention and integrity of other soft tissue stabilizing structures of the hip. Also, in a prospective design, it may be possible to standardize both the implant used and the surgical approach. Unfortunately, in a retrospective study that spans almost three decades, the evaluation of these factors is unlikely to have the necessary accuracy to provide a meaningful result. The authors, therefore, avoided evaluating incomplete data.

We felt confident that we were able to determine the integrity of the abductor mechanism with the greater trochanter and the resection length from the medical records and, hence, these were the parameters that were statistically evaluated. We were able to compare the rate of dislocation for tumor related hemiarthroplasty with that performed for non-tumor conditions and to report a statistically significant difference. In a study design such as this, it may not be possible to determine each of the contributing factors for this difference. We also feel that it is not appropriate to directly compare dislocation rates in different manuscripts because there are so many variables that are uncontrolled.

Lastly, Dr. Malawer is disappointed with the Discussion section and absence of acknowledgement of appropriate references including his own(1). Clearly, each of the articles cited by Dr. Malawer address either the technical aspects of reconstruction of the hip after hemiarthroplasty or a heterogenous group of patients including those undergoing total femoral replacement or APCs. The purpose of the current manuscript was to compare the overall dislocation rate in patients undergoing hemiarthroplasty for tumor-related conditions with a comparable group of non tumor patients. We did not attempt to validate each of the reconstructive aspects, e.g., the use of Dacron tape, as important mainly because reliable data was not uniformly available. In an attempt to compare the rates of dislocation following hemiarthroplasty in patients with tumor related conditions with those with non-tumor conditions, we sought to keep the groups as comparable as possible and, hence, procedures such as total femoral replacement are not germane to the question we attempted to address.

We fully acknowledge the limitations of our retrospective study that Dr. Malawer points out. We think many of these issues were not able to be addressed. However, these limitations are balanced by the ability to study a large group of patients treated at one institution with fairly accurate medical records. As such, the authors consider this manuscript a valuable contribution to the literature.

Reference:

1. Bickels J, Meller I, Henshaw RM, Malawer MM. Reconstruction of hip stability after proximal and total femur resections. Clin orthop. 2000; 375:218-30.

To The Editor:

With great interest we read the paper by Schneiderbauer, et al. The aim of the paper was to determine and analyze the dislocation rate of hemiarthroplasties after resection of the proximal femur due to primary or metastatic malignant tumors. The study was based on the Mayo Clinic data- base which is a good resource, well known for its quality and comprehensiveness, and indeed, the number of patients in this study is unprecedented. However, we would like to comment on two disturbing issues.

There seems to have been very little surgical input into the design of the study. Hemiarthroplasty of the hip is a generic definition for any prosthetic replacement of the proximal femur not involving the acetabulum. The dislocation analysis of such a diverse group of procedures, especially when oncologic resection precedes the prosthetic reconstruction, should be much more comprehensive. The only issues addressed by the authors were the extent of osseous resection and the preservation of the greater trochanter. Many other important factors which might have predisposed to dislocation, such as the type of the articulating part of the prosthesis (bipolar vs monopolar),the surgical approach, whether the hip capsule was preserved or sacrificed, and whether there was any attempt to reconstruct the hip capsule were totally ignored .

Bipolar prostheses are considered to be more stable than monopolar prostheses. Different approaches might entail different dislocation rates. The hip capsule is an important stabilizer of the hip joint and its resection due to onconlogical considerations might jeopardize stability. Reconstruction of the capsule, reinforcing it with the external rotators and the iliopsoas, was found to increase hip stability after oncologic resection of the proximal femur. (1) Without this important missing information and analysis, little can be learned from the result of 10.9% dislocation rate.

We were very disappointed to read in the introduction about the authors’ unawareness of any similar published data. Orthopedic oncologists have been reconstructing the proximal femur for the past two decades. We published a paper(1)that reported the long and mid-term follow-up of 57 patients who underwent prosthetic replacement after proximal femoral resection (n=39) or total femur resection (n=18) due to primary or secondary malignant tumors. All procedures were hemiarthroplasties, the prostheses were all bipolars, and the acetabulum was spared in all 57 cases. By suturing the remaining hip capsule with a Dacron tape around the neck of the prosthesis, forming a noose, and reinforcing it by attaching the external rotators to its posterolateral aspect and the iliopsoas muscle to its anterior aspect, we were able to substantially improve the stability of the prosthesis and reduce the prevalence of dislocation to one hip(1.7%). A simple electronic search of the literature was all that was needed to find our manuscript as well as five additional papers which report the dislocation rate of prosthetic replacements of the proximal femur after oncologic resections (either hemiarthroplasties or total joint replacement).(2,3,4,5,6). Any serious discussion would have to address the findings of at least some of these papers.

References:

1. Bickels J, Meller I, Henshaw RM, Malawer MM. Reconstruction of hip stability after proximal and total femur resections. Clin orthop. 2000; 375:218-30.

2. Masterson EL, Ferracini R, Griffin AM, Wunder JS, Bell RS. Capsular replacement with synthetic mesh: Effectiveness in preventing postoperative dislocation after wide resection of proximal femoral tumors and prosthetic reconstruction. J Arthroplasty. 1998; 13(8):860-6.

3. Ilyas I, Pant R, Kurar A, Moreau PG, Younge DA. Modular megaprosthesis for proximal femoral tumors. Int Orthop. 2002; 26(3):170- 3.

4. Zehr RJ, Enniking WF, Scarborough MT. Allograft-prosthetic composite versus megaprosthesis in proximal femoral reconstruction. Clin Orthop. 1996; 322: 207-223.

5. Kabukcuoglu Y, Grimer RJ, Tillman RM, Carter SR. Endoprosthetic replacement for primary malignant tumors of the proximal femur. Clin Orthop. 1999; 358: 8-14.

6. Johnsson R, Carlsson A, Kisch K Moritz U, Zetterstrom R, Persson BM. Function following mega total hip arthroplasty compared with conventional total hip arthroplasty and healthy matched controls. Clin Orthop. 1985; 192:159-67.

Editor's Note: The Editors solicited a response to Dr. Springfield's letter from James G. Wright, MD, Associate Editor for Evidence-Based Orthopaedics. His response follows:

Dr. Springfield questioned the Levels of Evidence rating assigned to an article written by Schneiderbauer, et al.(1) Levels of Evidence ratings of clinical articles published in JBJS are assigned according to the primary research question. To quote from that article: “The current study was performed to compare the dislocation rate following hemiarthroplasty performed in patients without tumor involvement with the rate following hemiarthroplasty in patients with tumor involvement of the surgical site.” Thus, as defined in the Levels of Evidence table, this study investigated the effect of patient characteristics, (i.e. the presence or absence of the tumor involvement of the proximal femur) on the outcome of disease, (i.e. dislocation of the hip). The design was retrospective, and therefore a Level 2 prognostic study. The assignment of Levels of Evidence has been shown to have acceptable reliability.(2)

Dr. Springfield expressed the concern that the study was not controlled. However, the study compared patients receiving hemiarthroplasty for fracture and osteonecrosis compared to those where the resection was performed for tumor.

Dr. Springfield also asserts that no evidence was provided for the statement “the greater trochanter should be retained”. The authors compared the dislocation rates when trochanter was not preserved in 125, where it was re-attached in four, partially preserved in 13, and completely preserved in 175 patients. Their results did not show a statistically significant difference with regards to preservation of the greater trochanter. Thus, Dr. Springfield is correct that results of this study did not demonstrate statistically significant support for preserving the greater trochanter. However, this was not an issue with Levels of Evidence but due to the authors extrapolating their results beyond the usual statistical significance of 0.05.

Yours truly,

James G. Wright, MD, MPH, FRCSC

References:

1. Schneiderbauer, M. M.; Sierra, R. J.; Schleck, C.; Harmsen, W. S.; and Scully, S. P.: Dislocation rate after hip hemiarthroplasty in patients with tumor-related conditions. J Bone Joint Surg, 87(8): 1810-1815, 2005.

2. Bhandari, M.; Swiontkowski, M. F.; Einhorn, T. A.; Tornetta, P.R.; Schemitsch, E. H.; Leece, P.; Sprague, S.; and Wright, J. G.: Interobserver agreement in the application of levels of evidence to scientific papers in the American volume of the Journal of Bone and Joint Surgery. J Bone Joint Surg, 86-A(8): 1717-1720, 2004.

To the Editor:

The level of evidence assigned to this article is "Prognostic Level II." I would have thought this was a "Therapeutic Study" and at best a Level III but more likely a Level IV. Who assigns the category and level? It seems that if the Journal is to use Levels of Evidence the process of assigning them should be as accurate as possible and I am sure that is the intention of the Editor. I question the accuracy of this assignment.

I specifically looked at the level of evidence because I do not agree with the author's recommendations and wondered what weight the article was assigned. The patient population was retrospectively collected over 27 years, includes a variety of diagnoses, is not controlled. They were treated by a variety of surgeons, most likely with a variety of equipment and techniques. The authors conclude that the dislocation rate is higher than after a more conventional hemiarthoplasty (maybe true in their series but not in my personal experience) and that the greater trochanter should be retained (no real evidence for this statement). The last statement is "Expert opinion" or Level V.