TO THE EDITOR:
In "Acetabular Involvement in Osteonecrosis of the Femoral Head" (81-A: 60—65, Jan. 1999), Steinberg et al. discussed the difficult treatment decision that is faced when a patient has osteonecrosis with early collapse of the femoral head. The authors showed that the acetabular cartilage in such patients is both grossly and histologically abnormal. They pointed out that "there has been renewed interest in … hemi-surface replacements … [but] … studies of greater numbers of patients and with longer follow-up are required." Although this statement is true, no studies regarding hemi-surface replacement were cited.
Several authors have reported on the clinical results of hemi-surface replacement performed for the treatment of osteonecrosis of the femoral head. In 1974, Kerboul et al.6 reported a successful result for 82 percent of sixty-five patients who were followed for one to six years after an adjusted cup arthroplasty, which is essentially a hemi-surface replacement. In 1978, in a study on surface replacement, Wagner14 included fourteen patients who had had a hemi-surface replacement; all of these patients reported substantial relief of pain, and no prosthesis had loosened at a maximum follow-up of four years. In 1982, Townley13 reported a 95 percent rate of survival in a study of 300 patients who were followed for as long as twenty-eight years. In 1987, Scott et al.9 reported an 88 percent rate of good or excellent results and an average Harris hip score4 of 90 points in a study of twenty-five patients who were followed for three years. That same year, Sedel et al.10 reported a good or excellent result in 79 percent of thirty-eight hips that were followed for an average of seven years. In 1993, Krackow et al.7 reported an 84 percent rate of good or excellent results and an average Harris hip score of 84 points in a study of nineteen patients who were followed for an average of three years. In 1997, Nelson et al.8 reported an 82 percent rate of good or excellent results and an average Harris hip score of 87 points at an average of six years in a study of seventeen patients who were an average of thirty-three years old. In 1998, Hungerford et al.5 reported on twenty-five patients (thirty-three hips) who had been followed for as long as fourteen years after a hemi-surface replacement. Thirty prostheses (91 percent) survived for at least five years. At an average follow-up of 10.5 years, thirteen hips (39 percent) had failed, but twenty hips (61 percent) had a good or excellent result, with an average Harris hip score of 91 points. Amstutz described his experience with hemi-surface replacement in a number of studies that were published during a ten-year period. In 1987, he and his colleagues reported a successful result for ten of eleven hips that had been followed for three years12. In a follow-up study performed at an average of eleven years postoperatively, five of these ten hips had a successful result1. In a subsequent retrospective study, hemi-surface replacement was compared with total surface replacement of the hip (performed with and without cement) as well as with total hip arthroplasty as treatment for osteonecrosis3. Hemi-surface replacement was the recommended technique, with a 90 percent rate of survival of the replacement at five years, a 70 percent rate at eight years, and a 58 percent rate at ten years3. In 1997, Amstutz et al.2 reported a 78 percent rate of satisfactory results at an average of six years in a study of twenty-seven patients, including some patients who had been included in previous reports1,3,12.
A formal meta-analysis has not been performed; however, when combined, these series represent 542 patients with an average age of thirty-six years, an average duration of follow-up of six years, an 89 percent rate of good or excellent results, and an average Harris hip score of 89 points. Hemi-surface replacement is not associated with the disadvantages that have caused total surface replacement of the hip to fall out of favor: namely, extensive reconstruction on the acetabular side and the use of a large femoral head with thin polyethylene that leads to large amounts of particulate wear debris and early failure. Revision of a hemi-surface replacement is relatively simple if the problem of osteolysis is avoided. There was essentially no loosening of the hemi-surface replacement prostheses described in these series; the revisions were performed for failures that were due to acetabular wear. Interestingly, the decrease in the rate of survival over time that was reported by Amstutz et al.1-3,12 and by Hungerford et al.5 is explained by the findings of Steinberg et al., who noted abnormal acetabular cartilage in patients with osteonecrosis.
The long-term outcome of hemi-surface replacement performed to treat postcollapse osteonecrosis of the femoral head may be limited by abnormal acetabular cartilage; however, reasonable intermediate-term clinical success has been reported since 1974 in more than 500 patients who had an average age of less than forty years. Although studies of greater numbers of patients with longer durations of follow-up will be useful, hemi-surface replacement is currently a reasonable treatment option in certain patients.
Greg Stocks, M.D.: Fondren Orthopaedic Group, 7401 South Main, Houston, Texas 77030
Dr. Steinberg and Dr. Corces reply:
The main focus of our study was on the changes in the acetabular cartilage of hips with osteonecrosis that had been treated with total hip replacement at a time when the acetabulum still appeared radiographically normal. Although we did discuss some of the clinical implications of the observations, our study was not intended to be a comprehensive discussion of treatment options. The results that have been achieved with femoral endoprostheses were reviewed briefly and were compared with the results of total hip replacement because the study was, in part, prompted by the poor results that have been reported in association with the use of endoprostheses in patients with osteonecrosis.
Dr. Stocks correctly points out that we did not cite specific studies of hemi-surface replacement as a treatment for osteonecrosis. He then cites twelve studies1-3,5-10,12-14 from which he concludes that "reasonable clinical success has been reported since 1974 in more than 500 patients" and that "hemi-surface replacement is currently a reasonable treatment option." However, our evaluation of the literature that Dr. Stocks cites leads us to somewhat different conclusions.
Many of the twelve articles are more than ten years old, and most involved a small number of patients with a short duration of follow-up (average, six years). The only study that had a substantial number of patients was the 1982 report by Townley13, which included 300 patients. Currently, very few surgeons use the prosthesis described in that report. We presume that there is a valid reason for this. The results reported in the other studies are mediocre at best. In 1974, Kerboul et al.6 reported on eighty hips that had been followed for only one to six years. They reported a successful result in 66 percent of the hips, improvement but some pain and a limp in 16 percent, and a failure in 16 percent. In 1987, Sedel et al.10 cited a 79 percent rate of good or excellent results and an 18 percent rate of failure after an average duration of follow-up of seven years. It seems that those authors later abandoned use of the hemi-surface replacement arthroplasty in favor of a ceramic-on-ceramic total hip prosthesis, which was reported to have an eight-year rate of survival of 97.8 percent in 199011. Nelson et al.8 reported failure in seven of twenty-one hips at an average of six years. In 1998, Hungerford et al.5 reported a good or excellent result after only 61 percent (twenty) of thirty-three resurfacing procedures. The thirteen hips (39 percent) that had a fair or poor result subsequently needed a total hip replacement. In 1994, Amstutz et al.1 reported a 50 percent rate of failure in ten hips at an average of 7.8 years, and in 1997, Amstutz et al.2 reported a 22 percent rate of failure in twenty-seven hips at an average of six years. Even the proponents of this technique have described it as "a time buying, first-stage operation" or as "a successful interim procedure."
This raises the question as to whether it is best to perform hemi-surface replacement arthroplasty with the expectation that it may need to be converted to a total hip replacement in a relatively short time or whether it is best to perform total hip replacement initially as the primary procedure.
The recent experiences of several investigators are encouraging and lead us to hope that better results will be possible with newer components and techniques. However, this has yet to be proved. If hemi-surface replacement arthroplasty does become a standard treatment for osteonecrosis, the observations noted in our article should be borne in mind. Better results can be anticipated when the acetabular cartilage is relatively normal; therefore, the surgeon must carefully note the condition of the acetabulum at the time of the operation to decide whether to proceed with resurfacing of the femoral head alone or whether to replace the acetabulum as well.
Marvin E. Steinberg, M.D.; Arturo Corces, M.D.: Department of Orthopaedic Surgery, Hospital of the University of Pennsylvania, 3400 Spruce Street, Silverstein Two, Philadelphia, Pennsylvania 19104
Michael Fallon, M.D.: Deceased