Patients
From January 1, 1976, to December 31, 1985, seventy-eight Neer hemiarthroplasties and thirty-six Neer total shoulder arthroplasties were performed by the senior one of us (R. H. C.) in ninety-eight patients who were fifty years old or less. All of these 114 procedures were included in a Kaplan-Meier survivorship analysis, with a revision procedure defined as the end point. Four hemiarthroplasties and two total shoulder arthroplasties that were included in the survivorship analysis were excluded from the clinical and radiographic analysis because two patients had died and four had been lost to follow-up. All six patients had had a unilateral procedure. Two of the hemiarthroplasties were performed for the treatment of rheumatoid arthritis and two, for the treatment of the sequelae of trauma. The two total shoulder arthroplasties were performed because of rheumatoid arthritis. These six patients had no complications at the most recent evaluation.
Seventy-four hemiarthroplasties (95 per cent) in sixty-four patients and thirty-four total shoulder arthroplasties (94 per cent) in thirty-one patients for which complete preoperative and operative records were available and after which the patient had been followed for at least five years (mean, 12.3 years) or until revision were included in the clinical analysis. A complete set of radiographs was available for sixty-eight (92 per cent) of the seventy-four shoulders that had a hemiarthroplasty and for thirty-two (94 per cent) of the thirty-four shoulders that had a total shoulder arthroplasty. Sixteen patients in the entire study group had a bilateral shoulder arthroplasty: ten patients had a bilateral hemiarthroplasty, three patients had a bilateral total shoulder arthroplasty, and three patients had a total shoulder arthroplasty in one shoulder and a hemiarthroplasty in the other.
The primary indication for the shoulder arthroplasty in these patients was painful glenohumeral arthritis, arthrosis, or incongruity that did not respond to non-operative treatment. The benefits and risks of both procedures, including the risk of loosening of the glenoid component after a total shoulder arthroplasty and the risk of less reliable relief of pain after a hemiarthroplasty, were discussed with the patient. If the patient did not have a strong preference, a hemiarthroplasty was performed to replace the humeral head when the joint was well balanced and the prosthesis rotated smoothly. Total shoulder arthroplasty was considered primarily for patients who had more advanced arthritis or arthrosis of the glenoid.
The mean age of the patients who had a hemiarthroplasty was thirty-nine years (range, nineteen to fifty years). There were thirty-six men and thirty-two women; six men and four women had a bilateral hemiarthroplasty. The indications for hemiarthroplasty were the sequelae of trauma (thirty shoulders), rheumatoid arthritis (twenty-eight), osteoarthrosis (eight), avascular necrosis (five), previous septic arthritis (two), rheumatoid arthritis and avascular necrosis (two), traumatic arthrosis and rheumatoid arthritis (one), rheumatoid arthritis and psoriatic arthritis (one), and lupus erythematosus (one). Of the thirty shoulders treated for the sequelae of trauma, ten had a fracture; ten, a chronic recurrent dislocation; eight, a fracture-dislocation; and two, a remote history of trauma that was not specified.
Four hemiarthroplasties were performed less than one year after the initial injury. Two of those hemiarthroplasties were done two and four months after the initial injury because an open reduction and attempted internal fixation of a fracture-dislocation of the proximal part of the humerus had failed. One hemiarthroplasty was performed two weeks after a four-part fracture of the proximal part of the humerus, and another was done four months after a gunshot wound with a fracture of the proximal part of the humerus. Twenty-six hemiarthroplasties were performed because of traumatic arthrosis more than one year after the initial injury.
Twenty-six (33 per cent) of the seventy-eight shoulders that had a hemiarthroplasty had had a previous procedure. Fifteen shoulders had had one procedure; five shoulders, two procedures; four shoulders, three procedures; one shoulder, six procedures; and one shoulder, nine procedures. Previous operations included an open reduction and internal fixation (nine shoulders), a stabilization procedure (nine), exploration (five), débridement (two), and an open reduction and internal fixation with subsequent removal of instrumentation (one).
The mean duration of clinical follow-up after the seventy-four hemiarthroplasties that were included in the clinical analysis was 11.7 years. Three shoulders that had a revision had been followed less than five years; eight shoulders were followed between five and ten years; forty-nine shoulders, between eleven and fifteen years; and fourteen shoulders, between sixteen and twenty years.
The mean age of the patients who had a total shoulder arthroplasty was forty-one years (range, twenty-two to fifty years). There were twenty-four women and nine men: two women and one man had a bilateral total shoulder arthroplasty. The indications for total shoulder arthroplasty were rheumatoid arthritis (twenty-five shoulders), traumatic arthrosis (seven), osteoarthrosis (two), avascular necrosis (one), and previous arthrodesis (one). The seven shoulders treated for the sequelae of trauma had a recurrent dislocation (three shoulders), a fracture-dislocation (three), or a fracture (one). The total shoulder arthroplasty was performed in these seven shoulders more than one year after the time of the initial injury.
Five shoulders had had at least one previous procedure before the total shoulder arthroplasty: four shoulders had had one procedure and one shoulder had had two procedures. Previous procedures included stabilization (two shoulders), open reduction and internal fixation with subsequent removal of instrumentation (one shoulder), arthrodesis (one shoulder), and open reduction without fixation (one shoulder).
The mean duration of clinical follow-up after the thirty-four total shoulder arthroplasties that were included in the clinical analysis was 13.6 years. Six shoulders had been followed five to ten years; eleven shoulders, eleven to fifteen years; sixteen, sixteen to twenty years; and one shoulder, more than twenty years.
The patients were categorized with respect to occupation at the time of the replacement procedure. Of the patients who had a hemiarthroplasty, ten (twelve shoulders) were disabled; twelve (thirteen shoulders) were homemakers; twenty-eight (thirty-two shoulders) were administrators, students, or clerical workers; twelve (fifteen shoulders) were employed at light manual labor; four (four shoulders) were employed at strenuous manual labor; and the occupation of two (two shoulders) was unknown. Of the patients who had a total shoulder arthroplasty, four (four shoulders) were disabled; eighteen (twenty shoulders) were homemakers; ten (ten shoulders) were administrators, students, or clerical workers; one (two shoulders) was employed at light manual labor; and none were employed at strenuous manual labor.
The patients were identified with the use of a computerized database that contained the files of all patients who had had a joint arthroplasty at our institution since 1969. The patients who were eligible for participation in our study were asked to return for an examination, an interview, and a radiographic evaluation at regular intervals. The patients who were unable to return for an examination were sent a questionnaire designed to evaluate function of the shoulder and satisfaction with the result. In addition, these patients were requested to have a local orthopaedic surgeon send in the results of a clinical examination and recent radiographs. The study was approved by the Institutional Review Board of the Mayo Foundation.
Radiographic Analysis
The preoperative, initial postoperative, and most recent radiographs were evaluated. An axillary radiograph, a 40-degree posterior oblique radiograph with external rotation of the humerus, and a 40-degree posterior oblique radiograph with internal rotation of the humerus were used for radiographic analysis. The radiographs of the patients who had a hemiarthroplasty were reviewed to determine the presence of glenohumeral subluxation, a periprosthetic radiolucent line, a shift in the position of the component, erosion of the glenoid, and loss of glenoid cartilage. The radiographs of the patients who had a total shoulder arthroplasty were reviewed to determine the presence of glenohumeral subluxation, a periprosthetic radiolucent line, and a shift in the position of the component.
Periprosthetic radiolucency was rated as grade 0 if there was no radiolucent line, grade 1 if the line was one millimeter wide or less and incomplete, grade 2 if the line was one millimeter wide and complete, grade 3 if the line was 1.5 millimeters wide and incomplete, grade 4 if the line was 1.5 millimeters wide and complete, and grade 5 if the line was two millimeters wide and complete. Glenohumeral subluxation was evaluated with respect to the direction and amount of translation of the center of the prosthetic head relative to the center of the glenoid or the glenoid component. The subluxation was recorded as none if it was not present, as mild if there was less than 25 per cent translation, as moderate if there was 25 to 50 per cent translation, and as severe if there was more than 50 per cent translation. Erosion of the glenoid was graded as none, mild, or moderate. A shift of the component was either present or absent.
Grading of Pain and Strength
At our institution, the clinical assessment of all patients who have an operation on the shoulder is recorded with use of a standard shoulder-analysis form. Pain and strength were rated according to the system of Neer et al. and of Cofield. No pain in the shoulder was assigned 1 point; slight pain, 2 points; pain after unusual activity, 3 points; moderate pain, 4 points; and severe pain, 5 points. Paralysis of the shoulder was given 0 points; a trace of strength, 1 point; poor strength with gravity eliminated, 2 points; fair strength against gravity, 3 points; good strength against gravity, 4 points; and normal strength, 5 points.
Grading of the Results
We graded the results with use of a modification of the rating system of Neer et al. and of Cofield—that is, we measured the motion of the involved shoulder in terms of degrees rather than comparing the motion with that on the contralateral side. Also, we did not incorporate a scale to assess the ability to perform the activities of daily living, as the necessary information was not available for all patients. The result was considered to be excellent if the patient had no or slight pain, had external rotation to at least 45 degrees, had active abduction to at least 140 degrees, and was satisfied with the result. The result was satisfactory if the patient had no or slight pain or moderate pain only with vigorous activity, had external rotation to at least 20 degrees, had active abduction to at least 90 degrees, and was satisfied with the procedure. If any of the criteria were not met, the result was unsatisfactory. Patients who had an additional operative procedure were considered to have an unsatisfactory result.
A subgroup of seven patients (three hemiarthroplasties and four total shoulder arthroplasties) who had limited rehabilitative goals because of dysfunction of the rotator cuff or the deltoid were evaluated separately2,6. These patients participated in an exercise program that was directed at maintaining stability with a smaller range of motion. The result was considered to be successful if the patient had no or slight pain or moderate pain only with vigorous activity, had external rotation to more than 20 degrees, and had active abduction to more than 70 degrees.
Operative Data
Eight shoulders that had a hemiarthroplasty had repair of a tear of the rotator cuff and five had reconstruction of the greater tubercle. The humeral component was press-fit in seventy shoulders, and it was inserted with cement in eight. Bone graft was used in the humerus in six shoulders and in the glenoid in one. The mean operative time was 2.3 hours, and the mean estimated blood loss was 600 milliliters. All patients received a Neer humeral component (Kirschner Medical, Fairlawn, New Jersey).
Seven shoulders that had a total shoulder arthroplasty had repair of a tear of the rotator cuff and two had reconstruction of the greater tubercle. The humeral component was press-fit and the glenoid component was inserted with cement in thirty-one shoulders. Both components were inserted with cement in five shoulders. Bone graft was used in the glenoid in one shoulder. The mean operative time was 3.1 hours, and the mean estimated blood loss was 860 milliliters. All patients received Neer humeral and glenoid components (Kirschner Medical).
Statistical Methods
For ordinal and continuous measurements, the Wilcoxon signed-rank test was performed to detect differences between preoperative and postoperative values. The Wilcoxon rank-sum test was performed to detect differences between the groups of patients. A Fisher exact test was used to test for significant relationships between categorical variables. The estimated rates of survival of the implants were calculated with use of the method of Kaplan and Meier. Comparisons of survival distributions between groups were performed with use of the log-rank test.
Complications and Revisions
Hemiarthroplasty
A postoperative complication that did not necessitate a revision procedure developed in four shoulders. These complications (in one shoulder each) included an immediate postoperative hematoma that resolved after aspiration, subacromial bursitis that developed nine months after the hemiarthroplasty and resolved after injection of steroids, a brachial plexus neuropathy that resolved postoperatively, and an infection that developed two months after the hemiarthroplasty and resolved after arthroscopic irrigation and intravenous antibiotic therapy. The infection did not necessitate an additional procedure.
Fifteen (19 per cent) of the seventy-eight hemiarthroplasties were revised. The estimated survival of the prostheses (with 95 per cent confidence intervals) was 92 per cent (86 to 98 per cent) at five years, 83 per cent (75 to 93 per cent) at ten years, and 73 per cent (59 to 88 per cent) at fifteen years. With the numbers available for study, we could not detect an association between the need for a revision and preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation of the joint, gender, or a tear of the rotator cuff.
Analysis of the risk of revision associated with the two major diagnoses revealed a significantly higher risk for the shoulders that had a hemiarthroplasty because of the sequelae of trauma than for those that had the procedure for the treatment of rheumatoid arthritis (p = 0.017) (Fig. 1). In addition, the risk of revision was significantly higher for the shoulders that had had a previous operation than for those that had not (p = 0.038). The estimated survival of the prostheses (with 95 per cent confidence intervals) in the shoulders that had had a previous procedure was 84 per cent (70 to 100 per cent) at five years, 71 per cent (55 to 92 per cent) at ten years, and 57 per cent (33 to 95 per cent) at fifteen years. The estimated survival of the prostheses (with 95 per cent confidence intervals) in the shoulders that had not had a previous procedure was 96 per cent (91 to 100 per cent) at five years, 90 per cent (81 to 99 per cent) at ten years, and 80 per cent (63 to 97 per cent) at fifteen years.
Eleven of the fifteen revision procedures were performed because of painful glenoid arthrosis. In ten of these eleven shoulders, the revision consisted of a total shoulder arthroplasty. In the eleventh shoulder, the humeral component was revised without placement of a glenoid component because the glenoid bone stock was inadequate. The mean time from the hemiarthroplasty to the revision operation was 7.8 years for these eleven shoulders.
The four revision operations that were done for a reason other than painful glenoid arthrosis included a revision hemiarthroplasty, done ten years after the index hemiarthroplasty, because of aseptic loosening of the humeral component; a total shoulder arthroplasty, done four years after the hemiarthroplasty, also because of aseptic loosening of the humeral component; removal of the prosthesis, done two years after the hemiarthroplasty, because of an infection; and a total shoulder arthroplasty, done at another institution three years after the hemiarthroplasty, because of pain of an unknown etiology.
In order to evaluate the rate of mechanical failure, the data on survival also were calculated with the exclusion of the revision performed because of infection. This analysis revealed an estimated survival of the prostheses (with 95 per cent confidence intervals) of 93 per cent (88 to 99 per cent) at five years, 85 per cent (77 to 93 per cent) at ten years, and 75 per cent (60 to 89 per cent) at fifteen years.
Total Shoulder Arthroplasty
Postoperatively, a hematoma developed in one of the shoulders and resolved without operative treatment. Four shoulders (11 per cent) had a revision operation. The estimated survival of the prostheses (with 95 per cent confidence intervals) was 97 per cent (92 to 100 per cent) at five years, 97 per cent (91 to 100 per cent) at ten years, and 84 per cent (70 to 100 per cent) at fifteen years. With the numbers available for study, we could not detect a significant relationship between gender, diagnosis, preoperative subluxation, or a previous procedure and the risk of revision. However, the risk of revision was higher for the seven shoulders that had had a tear of the rotator cuff at the time of the total shoulder arthroplasty than for the twenty-nine that had not had one (p = 0.029).
Both components were removed from two shoulders because of infection ten and thirteen years after the total shoulder arthroplasty. The glenoid component was removed from two shoulders because of aseptic loosening eleven months and ten years after the total shoulder arthroplasty; the patient who had a revision of the glenoid component at ten years also had loosening of the humeral component.
Excluding the revisions that were done because of infection, the estimated survival of the components of a total shoulder arthroplasty (with 95 per cent confidence intervals) was 97 per cent (92 to 100 per cent) at five years, 97 per cent (91 to 100 per cent) at ten years, and 93 per cent (82 to 100 per cent) at fifteen years.
Comparison of Hemiarthroplasty and Total Shoulder Arthroplasty
The log-rank test did not reveal a significant difference, with the numbers available, between hemiarthroplasty and total shoulder arthroplasty with regard to the overall survival of the components (p = 0.214) (Fig. 2). We also could not detect a significant difference when revisions for infection were excluded from the analysis (p = 0.07). The shoulders treated with hemiarthroplasty had a significantly higher rate of revision at ten years (p = 0.007), but, with the numbers available, we could not detect a significant difference between the rates associated with the two procedures at five years (p = 0.231) or at fifteen years (p = 0.280).
Pain
The hemiarthroplasties resulted in a significant relief of pain, with the mean score decreasing from 4.6 points preoperatively to 2.4 points postoperatively (p < 0.0001). The decrease in pain did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, or a previous operation. Postoperative pain did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral component, postoperative erosion of the glenoid, loss of glenoid cartilage, or subluxation.
The total shoulder arthroplasties also resulted in a significant relief of pain. The mean score for pain decreased from 4.7 points preoperatively to 2.1 points postoperatively (p < 0.0001). Relief of pain did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative subluxation, or a previous procedure. Postoperative pain did not vary markedly between the patients who had and those who did not have a radiolucent line about the humeral or glenoid component or postoperative subluxation.
We could not detect a significant difference, with the numbers available, between the shoulders that had a hemiarthroplasty and those that had a total shoulder arthroplasty with respect to preoperative pain (p = 0.219), postoperative pain (p = 0.235), or relief of pain (p = 0.168).
Range of Motion
The mean active abduction of the shoulders that had a hemiarthroplasty improved from 80 degrees preoperatively to 124 degrees postoperatively (p < 0.0001). The improvement did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, or a previous operation. Postoperatively, active abduction did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral component, postoperative erosion of the glenoid, loss of glenoid cartilage, or subluxation. The mean external rotation increased from 20 degrees preoperatively to 46 degrees postoperatively (p < 0.0001). The improvement in external rotation did not vary markedly between the patients who had and those who did not have a tear of the rotator cuff, preoperative erosion of the glenoid, loss of glenoid cartilage, or subluxation or with respect to gender. However, the shoulders that were treated with a hemiarthroplasty because of the sequelae of trauma had significantly less external rotation preoperatively than did those that had a hemiarthroplasty for the treatment of rheumatoid arthritis (6 compared with 24 degrees) (p = 0.01). With the numbers available, we could not detect a significant difference between the postoperative values in these groups (p = 0.81).
In addition, the shoulders that had had a previous procedure had significantly less external rotation preoperatively than did those that had not had a previous procedure (p = 0.0002). Postoperatively, the shoulders that had had a previous procedure had significantly greater improvement in external rotation (p = 0.026), which resulted in a similar amount of external rotation after the operation for both groups (p = 0.221). Postoperative external rotation did not vary markedly between the patients who had and those who did not have a radiolucent line adjacent to the humeral component, postoperative erosion of the glenoid, loss of glenoid cartilage, or subluxation.
For the shoulders that had a hemiarthroplasty, the mean internal rotation (measured as the most cephalad level that the thumb could reach on the spine) increased significantly from the fourth lumbar vertebra preoperatively to the second lumbar vertebra postoperatively (p < 0.0001). The improvement in internal rotation did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, or a previous procedure. Postoperatively, internal rotation did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral component, postoperative erosion of the glenoid, loss of glenoid cartilage, or subluxation.
The mean active abduction of the shoulders treated with a total shoulder arthroplasty improved from 62 degrees preoperatively to 104 degrees postoperatively (p < 0.0001). The improvement did not vary markedly with respect to gender, the presence of a tear of the rotator cuff, preoperative subluxation, or a previous operation. The shoulders that had a total shoulder arthroplasty because of rheumatoid arthritis had had significantly less active abduction preoperatively than the shoulders that had a total shoulder arthroplasty because of traumatic arthrosis (55 compared with 81 degrees) (p < 0.042). However, with the numbers available, we could not detect a significant difference between the postoperative values in these groups (p = 0.730). Postoperative active abduction did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral or glenoid component, postoperative subluxation, or a previous operation. The mean external rotation increased from 17 degrees preoperatively to 43 degrees postoperatively (p < 0.0001). The mean internal rotation increased from the fourth lumbar vertebra preoperatively to the third lumbar vertebra postoperatively, but, with the numbers available, we could not detect a significant difference between these values (p = 0.20). Changes in external and internal rotation did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative subluxation, or a previous operation. Postoperatively, external and internal rotation did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral or glenoid component, postoperative subluxation, or a previous operation.
The mean values for active abduction before and after the total shoulder arthroplasties were significantly lower than the values before and after the hemiarthroplasties (p = 0.028 for both). However, we could not detect a significant difference between the groups with respect to the mean improvement in active abduction (p = 0.777), the mean external rotation before (p = 0.680) or after (p = 0.552) the operation, the mean improvement in external rotation (p = 0.717), the mean internal rotation before (p = 0.721) or after (p = 0.069) the operation, or the mean improvement in internal rotation (p = 0.414).
Strength
We could not detect a significant improvement, with the numbers available, in the strength of flexion (p = 0.141), abduction (p = 0.446), or external rotation (p = 0.118) after the hemiarthroplasties. The improvement in strength did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, or a previous operation. Postoperative strength did not vary markedly between the patients who had and those who did not have a radiolucent line around the humeral component, postoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, or a previous operation.
The total shoulder arthroplasties resulted in a significant improvement in the strength of flexion (from 3.5 points preoperatively to 4.3 points postoperatively; p = 0.016), abduction (from 3.4 to 4.0 points; p = 0.02), and external rotation (from 3.5 to 4.1 points; p = 0.007). The improvement in strength did not vary markedly with respect to gender, diagnosis, the presence of a tear of the rotator cuff, preoperative subluxation, or a previous operation. Postoperatively, strength did not vary markedly with respect to the presence of a radiolucent line around the humeral or glenoid component, postoperative subluxation, or a previous operation. The mean values for the strength of flexion (p = 0.0002), abduction (p = 0.0004), and external rotation (p = 0.001) before the total shoulder arthroplasties were significantly lower than the values before the hemiarthroplasties. This finding may be due to the fact that patients who had a total shoulder arthroplasty had more advanced disease. Postoperatively, however, there was a greater improvement in the strength of flexion (p = 0.008), abduction (p = 0.018), and external rotation (p = 0.006) after the total shoulder arthroplasties than after the hemiarthroplasties. We could not detect a significant difference between the two procedures with regard to the strength of flexion (p = 0.566), abduction (p = 0.409), or external rotation (p = 0.640) at the most recent follow-up examination.
Radiographic Findings
Hemiarthroplasty
At the time of the most recent follow-up, thirty (44 per cent) of the sixty-eight shoulders that could be evaluated radiographically had some degree of glenohumeral subluxation. Twenty-four shoulders had superior subluxation: it was mild in seventeen, moderate in five, and severe in two. Five shoulders had posterior subluxation: it was mild in one and moderate in four. One shoulder had mild anterior subluxation. We could not detect a relationship between the presence of subluxation and gender, diagnosis, or the presence of a tear of the rotator cuff.
Erosion of the glenoid was seen in thirty-three (49 per cent) of the sixty-eight shoulders preoperatively. Forty-six shoulders (68 per cent) had erosion of the glenoid at the time of the most recent follow-up: it was mild in nineteen, moderate in fourteen, and severe in thirteen. We could not detect a relationship between erosion of the glenoid and gender, diagnosis, or the presence of a tear of the rotator cuff. Loss of glenoid cartilage was seen in fifty-two shoulders (76 per cent) preoperatively and in fifty-seven (84 per cent) at the most recent follow-up examination. We could not detect a significant relationship between the loss of glenoid cartilage and gender, diagnosis, or the presence of a tear of the rotator cuff.
A radiolucent line around the humeral component was seen in sixteen shoulders (24 per cent). Fifteen of those shoulders had a press-fit component, and one had a component inserted with cement. The line was grade 1 in six shoulders, grade 2 in five, grade 3 in one, grade 4 in one, and grade 5 in three. Eight shoulders had a shift in the position of the humeral component. A radiolucent line around the humeral component was more likely to develop in men than in women (p = 0.022). We could not detect a relationship between the development of a radiolucent line around the humeral component and the presence of a tear of the rotator cuff or the diagnosis.
Total Shoulder Arthroplasty
At the time of the most recent follow-up, nineteen (59 per cent) of the thirty-two shoulders that could be evaluated radiographically had some degree of glenohumeral subluxation. Seventeen shoulders had superior subluxation: it was mild in eight, moderate in six, and severe in three. Two shoulders had posterior subluxation: it was mild in one and moderate in the other. We could not detect a relationship between the presence of subluxation and gender, diagnosis, or the presence of a tear of the rotator cuff.
A radiolucent line about the glenoid component was seen in nineteen (59 per cent) of the thirty-two shoulders. The glenoid component had been inserted with cement in each of these shoulders. The line was grade 1 in two shoulders, grade 2 in five, grade 3 in one, grade 4 in seven, and grade 5 in four. A shift in the position of the glenoid component was seen in seven shoulders (22 per cent). We could not detect a relationship between the development of a radiolucent line around the glenoid component and gender, diagnosis, or the presence of a tear of the rotator cuff.
A radiolucent line around the humeral component was evident in seventeen shoulders (53 per cent). Fifteen of these shoulders had a press-fit component, and two had a component inserted with cement. The line was grade 1 in five shoulders, grade 2 in four, grade 3 in two, grade 4 in two, and grade 5 in four. Ten shoulders (31 per cent) had a shift in the position of the humeral component. We could not detect a relationship between the development of a radiolucent line adjacent to the humeral component and gender, diagnosis, or the presence of a tear of the rotator cuff.
Comparison of Hemiarthroplasty and Total Shoulder Arthroplasty
The prevalence of a radiolucent line about the humeral component was significantly higher in the shoulders that had a total shoulder arthroplasty (53 per cent) than in those that had a hemiarthroplasty (24 per cent) (p = 0.006).
Rating of the Clinical Results
Hemiarthroplasty
Overall, fifteen shoulders had an excellent result; twenty-four, a satisfactory result; and thirty-two, an unsatisfactory result. A separate analysis of the three patients (three shoulders) who had limited rehabilitative goals revealed that each had an unsuccessful result. With the numbers available, we could not detect a significant difference among the diagnostic groups with respect to the ratings of the results of the hemiarthroplasties (Table I). Preoperative erosion of the glenoid, loss of glenoid cartilage, subluxation, a tear of the rotator cuff, gender, and diagnosis were not risk factors for an unsatisfactory result. In addition, we could not detect an association between postoperative erosion of the glenoid, loss of cartilage, subluxation, or a radiolucent line around the humeral component and an unsatisfactory result. However, the group that had not had a previous operation had significantly better results than did the group that had had a previous operation (p = 0.017). Of the forty-nine shoulders that had not had a previous operation, twelve (24 per cent) had an excellent result; nineteen (39 per cent), a satisfactory result; and eighteen (37 per cent), an unsatisfactory or unsuccessful result. Of the twenty-five shoulders that had had a previous operation, three (12 per cent) had an excellent result; five (20 per cent), a satisfactory result; and seventeen (68 per cent), an unsatisfactory or unsuccessful result.
More than one factor contributed to nineteen (54 per cent) of the thirty-five unsatisfactory or unsuccessful results. The most common factors were limited active abduction (seventeen shoulders), a revision procedure (fifteen), limited external rotation (eleven), pain (eleven), and a lack of satisfaction on the part of the patient (seven).
Total Shoulder Arthroplasty
Four of the thirty-four shoulders had an excellent result; thirteen, a satisfactory result; and thirteen, an unsatisfactory result. A separate analysis of the four patients (four shoulders) who had limited rehabilitative goals revealed that each had an unsuccessful result. With the numbers available, we could not detect a significant difference among the diagnostic groups with respect to the ratings of the results of the total shoulder arthroplasties (Table I). A previous operation, preoperative subluxation of the joint, a tear of the rotator cuff, gender, and diagnosis were not risk factors for an unsatisfactory or unsuccessful result. In addition, we could not detect an association between a radiolucent line around the glenoid or humeral component postoperatively or subluxation of the joint and an unsatisfactory or unsuccessful result.
More than one factor contributed to eleven of the seventeen unsatisfactory or unsuccessful results. The most common factors were limited active abduction (ten shoulders), limited external rotation (nine), pain (five), a revision operation (four), and a lack of satisfaction on the part of the patient (three).
Comparison of Hemiarthroplasty and Total Shoulder Arthroplasty
Limited active abduction was the most common factor that contributed to an unsatisfactory or unsuccessful result in both groups. With the numbers available, we could not detect a significant difference between the results of hemiarthroplasty and those of total shoulder arthroplasty (p = 1.00).