TO THE EDITOR:
The article "Neer Hemiarthroplasty and Neer Total Shoulder Arthroplasty in Patients Fifty Years Old or Less. Long-Term Results" (80-A: 464—473, April 1998), by Sperling et al., is an impressive statistical report from the outstanding database of the Mayo Clinic. However, the clinical information inflates the prevalence of complications because it is based on experience from early in the learning curve, which does not represent optimum technique. Regardless of how perfect the mathematics are, we are looking at a technique that is twenty years old.
Many of the patients in this series, who were managed between 1976 and 1985, were reported on previously1,2,6. Consider the technique at that time. The Neer-II prosthesis, which was introduced in 1973, was the first nonconstrained total shoulder system. With the cooperation of the manufacturer (3M, St. Paul, Minnesota), it was field-tested in a multicenter study for nine years before it was released for general use in 19823. The Neer-II components were designed for insertion with cement. In this early series reported on by Sperling et al., however, the humeral component was inserted with cement in only thirteen (11 percent) of the 114 shoulders in the entire study of seventy-eight hemiarthroplasties and thirty-six total shoulder arthroplasties. In contrast, the humeral component was implanted with cement in 336 (95 percent) of the most recent 355 shoulder arthroplasties in my series4. In another recent report from the Mayo Clinic on long-term results with the Neer prosthesis6, a shift in the position of the humeral component occurred in forty (49 percent) of eighty-one humeral components that had been inserted without cement, and none of the eight cemented humeral components in that study had shifted in position or were considered loose. This factor accounts for the higher rate of loosening of the humeral component in this study by Sperling et al. as compared with that in other studies3-5.
Interestingly, the most frequent indication for a subsequent operation in the present series (reported as the reason for eleven of the nineteen revisions) was painful glenoid arthrosis after a hemiarthroplasty. Only two revisions in the entire series were performed because of aseptic loosening of the glenoid component. No shoulder was revised because of breakage, dissociation, or wear of the prosthesis.
The main reason for an unsatisfactory result was active abduction that was less than satisfactory according to the rating scale used; overall, this factor contributed to twenty-seven of the fifty-two unsatisfactory outcomes. The soft tissues and the quality of aftercare are key factors in determining the range of motion that is achieved postoperatively. Fifty-eight (74 percent) of the seventy-eight shoulders that had a hemiarthroplasty had either rheumatoid arthritis or sequelae of trauma, and twenty-six (33 percent) had had a previous procedure. Furthermore, most of these procedures were performed before the extended deltopectoral approach was used and before the present-day quality of aftercare.
In all fairness, it should be pointed out that, despite suboptimum technique and adverse pathological changes, use of this one-piece implant with conforming surfaces was successful enough for the authors to conclude: "The data from the present study indicate that a shoulder arthroplasty provides marked long-term relief of pain and improvement in motion."
Charles S. Neer, II, M.D.: 231 South Miller, Vinita, Oklahoma 74301
Dr. Sperling, Dr. Cofield, and Mr. Rowland reply:
There are perhaps two dozen major issues related to shoulder arthroplasty. We attempted to address three: the effectiveness of the procedure in younger patients, the outcome after replacement of the humeral head compared with that after total shoulder arthroplasty, and the outcome after a longer duration of follow-up. In doing so, the material became somewhat complex, and it is useful that Dr. Neer, who developed this arthroplasty procedure, raised key points on the basis of his experience and perspective.
We found that younger patients differ from the typical group of patients who have an arthroplasty in more respects than just age. Younger patients usually have severe, active, juvenile or early-onset rheumatoid arthritis or have had a previous procedure to treat a complex proximal humeral fracture; as a result, such patients are known to have lower-than-average rates of success with regard to the return of movement and strength.
The current operative technique is indeed different than the one that we used. Various sizes and shapes of implants are now available, as are precision instruments. Fixation is improved by the application of current cementing techniques. Most importantly, releases of the shoulder capsule, particularly those involving the anterior and inferior aspects of the capsule, are extensive. Finally, the deltoid is usually left intact (although no complications related to the deltoid were observed in our study).
We think that the rate of complications in our study was relatively low. The four infections developed in immunocompromised patients and appeared to be unrelated to the initial operative procedure. Two arose from a known active infection at another site. Three humeral and two glenoid components became clinically loose during the long period of follow-up; an additional twenty-two humeral and ten glenoid components became radiographically loose. Otherwise, the revisions were almost entirely limited to shoulders that had been treated with a hemiarthroplasty and in which painful glenoid arthrosis had developed.
The rating system that we used is often employed to assess this procedure. We modified it to include an unsatisfactory result for shoulders that had had a revision. This is a reasonable modification of the system; however, had these shoulders been excluded, the results in this somewhat diverse group of patients would have been better because the unsatisfactory results would have been largely determined by stiffness—that is, failure to regain a certain degree of active motion and strength.
Implant arthroplasty has been dramatically successful in terms of relief of pain and maintenance of variable degrees of improved function. This success naturally produces, as Charnley stated, "[a] level of excitement among many surgeons." Charnley went on to say: "I have always attempted to promote an extremely conservative attitude to this operation and have never been happy [about doing it] in young patients … without the same strict selection we have employed here. I think there is going to be a harvest of very bad results in the young people."7 We adhere to this conservatism but are cautiously optimistic about the outcomes for younger people and believe that the results will continue to improve as understanding, patient selection, and techniques evolve.
John W. Sperling, M.D.; Robert H. Cofield, M.D.; Charles M. Rowland, M.S.: Departments of Orthopedics (J. W. S. and R. H. C.) and Biostatistics (C. M. R.), Mayo Clinic, 200 First Street S.W., Rochester, Minnesota 55905