TO THE EDITOR:
We read with interest "Survival Analysis of Hips Treated with Core Decompression or Vascularized Fibular Grafting Because of Avascular Necrosis" (80-A: 1270—1275, Sept. 1998), by Scully et al. We believe that a comparison study of treatment methods for osteonecrosis of the femoral head is an intriguing idea, but we have numerous reservations concerning the construct and conclusions of this work.
The authors did not determine the size or location of the lesion and therefore do not know if the groups were really comparable with respect to the severity of the disease. Size has been shown to be extremely important in many studies of this disease2-5. We suggest that the authors measure the size of the lesions in Ficat stage-II and III hips1 and determine if size is indeed a predictor of which hips will respond well to core decompression or bone-grafting and which will not.
The authors had an excellent opportunity to match the patients from each group with regard to clinical and radiographic parameters. It would have been relatively easy to select patients from the larger group (the 614 hips that were treated with vascularized fibular grafting) to match those from the smaller group (the ninety-eight hips that were treated with core decompression). We would like to see this stratification performed. We performed a statistical analysis of the data delineated in this study and found that, although the two groups were significantly similar with regard to age (p = 0.2) and the prevalence of bilaterality (p = 0.5), they were significantly different with regard to the stage of the lesion (p = 0.00001); gender (p = 0.001); and the associated factors of corticosteroid use (p = 0.000147), trauma (p = 0.001), and the prevalence of idiopathic avascular necrosis (p = 0.005). This possible bias in the data could have been reduced by directly matching the patients from the larger group to the smaller group as described. In addition, the authors evaluated the ages of patients who had different stages of disease and found marked (p = 0.081, stage-II hips) or significant (p = 0.001, stage-III hips) differences between the two treatment groups. The authors merely state that these demographic factors have not been found to be important in previous studies3. Thus, the authors' interpretation of their results is limited because they did not evaluate the impact of these factors despite the opportunity to do so.
A striking feature of this study is that stage-III hips that were treated with vascularized fibular grafting had excellent results. However, it should be noted that fifty months is a short duration of follow-up for a disease that afflicts young patients (mean age, approximately thirty-eight years) who have a life expectancy of at least forty more years. In addition, the use of total hip arthroplasty as the end point for failure is not an ideal outcome measure, particularly since many of these young patients try to avoid hip replacement at any cost. Some type of pain or functionality score would have been useful. Furthermore, any additional procedure, not necessarily a hip arthroplasty, should be considered a failure. A patient who has invested so much time in a grafting procedure might be less likely to have a hip arthroplasty than one who has had a core decompression. We agree with the authors that core decompression should not be used to treat a stage-III lesion unless the lesion is small and central, which is rare. In a recent study4, core decompression had a long-term success rate of less than 30 percent (eighteen of sixty-nine) when used to treat such lesions.
Scully et al. reported that both methods had similar rates of success when used to treat stage-I lesions. If this is the case, then core decompression, which takes fifteen minutes to perform and is done as an outpatient procedure, should be offered as a time-extending alternative to the more extensive procedure of vascularized fibular grafting. The same recommendation could almost be made for stage-II lesions. Because core decompression has low morbidity and does not preclude additional operative intervention, it is not unreasonable to manage a patient with a core decompression first and then to offer vascularized fibular grafting (using the core track that is created at the time of the decompression) if the initial treatment fails. As it stands, the article by Scully et al. justifies the following conclusions. First, grafting should never be performed for stage-I lesions. Second, core decompression should be performed for stage-II lesions (depending on size), and grafting should be performed if, or when, the decompression fails. Third, grafting should be performed for stage-III lesions.
Michael A. Mont, M.D.; Lynne C. Jones, Ph.D.; David S. Hungerford, M.D.: Department of Orthopaedic Surgery, Good Samaritan Professional Building, 5601 Loch Raven Boulevard, Baltimore, Maryland 21239
Dr. Scully, Dr. Aaron, and Dr. Urbaniak reply:
Mont et al. point out that the two treatment groups were not stratified according to the extent and location of the lesion of the femoral head. We are aware that this is a prognostically important variable and wish that the analysis of the paper could have accounted for it; however, the available data did not permit this.
The next point raised by Mont et al. is that the two groups were significantly similar with regard to age and the prevalence of bilateral involvement. In fact, there was a significant difference (p = 0.001 to 0.081) with regard to age, and the data were stratified according to this factor in the analysis. The data were also stratified according to the stage of disease. In our experience with avascular necrosis of the femoral head, we have not found gender, the use of steroids, a history of trauma, or an idiopathic etiology to have prognostic importance6,7. In fact, we have found all etiologies to have equivalent prognoses. Hence, matching for these variables in the statistical comparison is not necessary. We do not believe that the absence of stratification for these variables detracts from the interpretation of the study.
Mont et al. next question the use of total hip arthroplasty as an end point for failure. They point out that a pain or functionality score would have been preferable and suggest that any additional operation, not necessarily a hip arthroplasty, should have been considered a failure. In most of these patients, the next operation after the index procedure was, in fact, a total hip arthroplasty. Thus, the consideration of additional operations as an end point would not have substantially changed our conclusion. We acknowledge that there was a potential inherent bias in the study in that vascularized fibular grafting is associated with a large emotional investment, on the part of both the physician and the patient, which may potentially delay a decision to proceed to total hip arthroplasty.
Mont et al. also propose a treatment algorithm for avascular necrosis of the femoral head. They suggest that core decompression, rather than vascularized fibular grafting, should be offered as a time-extending procedure for the treatment of stage-I lesions. In fact, very few patients who had a stage-I lesion had vascularized fibular grafting at Duke University. Although the number of patients in our study was small, there were no failures among the patients who had a stage-I lesion, suggesting that perhaps neither procedure was necessary. Mont et al. believe that the same argument could be applied to stage-II lesions. They suggest that core decompression should be used as the initial treatment because of its low morbidity and propose that vascularized fibular grafting can be used as a salvage procedure if decompression fails. Certainly, the criteria for failure of the core decompression would be important in such a setting; however, we firmly believe that progression from a stage-II to a stage-III lesion has a substantial adverse effect on the prognosis for long-term survival of the native femoral head. In light of this factor and the data presented in the manuscript, we continue to believe that vascularized fibular grafting is preferable to other existing options for the treatment of stage-II and III avascular necrosis of the femoral head.
Sean P. Scully, M.D., Ph.D.; James R. Urbaniak, M.D.: Duke University Medical Center, Boxes 3312 (S. P. S.) and 2912 (J. R. U.), Durham, North Carolina 27710
Roy K. Aaron, M.D.: Brown University School of Medicine, Southwest Pavilion 524, 593 Eddy Street, Providence, Rhode Island 02903