In 1974 and 1982, Neer et al. described the results of humeral head
replacement in patients with rheumatoid arthritis and osteoarthritis of the
glenohumeral
joint1,2.
Despite the increasing number of shoulder arthroplasties performed over the
last decade, the indications and contraindications for humeral head
replacement without prosthetic resurfacing of the glenoid are not clearly
defined. In instances when either total shoulder arthroplasty or humeral head
replacement was used routinely, good results have been reported in the
literature, suggesting that hemiarthroplasty of the glenohumeral joint can be
an effective method to treat glenohumeral
arthritis3-11.
The situations in which hemiarthroplasty may be preferred over total shoulder
arthroplasty, however, are poorly defined in the literature. Additionally, few
studies have documented the results of hemiarthroplasty at a minimum follow-up
of five
years10,12,13.
The present study describes the results of glenohumeral hemiarthroplasty in
fifty shoulders with a minimum follow-up of five years. The purpose of this
study was to describe the patient selection strategy, to discuss the technical
methods that affected patient outcomes, and to test the hypothesis that
hemiarthroplasty can provide sustained functional improvement and pain relief
for at least five years.
Between July 1991 and November 1997, fifty-seven consecutive patients
(sixty-four shoulders) affected by primary or secondary osteoarthritis of the
glenohumeral joint were treated with a hemiarthroplasty. All of the
arthroplasties were performed with a DePuy Global Shoulder modular prosthesis
(DePuy, a Johnson and Johnson company, Warsaw, Indiana). The indication for
each operation was a limitation in the activities of daily living caused by
glenohumeral arthritis, combined with shoulder pain refractory to improvement
by conservative measures.
The inclusion criteria for the study were a diagnosis of glenohumeral
arthritis, a concentric glenoid consisting of eburnated bone, and a humeral
head that was centered in the glenoid preoperatively or at the time of
surgery. The humerus was considered to be centered when the distance between
the center of the humeral head and the center of the glenoid was estimated to
be <25% of the humeral head diameter preoperatively or intraoperatively
after placement of a trial humeral component and soft-tissue balancing.
Moderate subluxation was defined as a difference of 25% to 50% between these
two points, and severe subluxation was defined as a difference of >50%.
Patients presenting with a nonconcentric glenoid were included if,
intraoperatively, the glenoid fossa could be converted to a smooth concentric
surface by reaming, and patients with moderate or severe glenohumeral
subluxation were included when the subluxation could be corrected through
soft-tissue balancing. The exclusion criteria were a cystic or pitted
osteopenic glenoid, a nonconcentric glenoid that was not correctable by
reaming, and moderate or severe glenohumeral subluxation that persisted
following glenoid preparation and attempts at soft-tissue balancing (these
patients were managed with a total shoulder arthroplasty).
Rotator cuff tears, and the specific tendons that were torn, were
documented immediately postoperatively. Clinical assessments were performed
preoperatively and at one-year intervals postoperatively for at least five
years. Active range-of-motion variables, including elevation, external
rotation, and internal rotation, were measured. Active elevation and external
rotation (with the arm adducted) were measured by the examiner with use of a
2-ft (61-cm)-long handheld goniometer and were recorded in 5° increments.
Internal rotation was recorded as the highest vertebral level that the patient
could reach with the thumb extended.
Outcome data were derived from patient self-assessment instruments,
including the American Shoulder and Elbow Surgeons questionnaire (an overall
shoulder scoring system based on ten questions concerning the functional
activities of daily living, pain, and patient satisfaction), and the validated
Simple Shoulder
Test14,15.
Visual analog scales were used preoperatively and postoperatively to evaluate
the following six parameters: (1) pain, (2) shoulder comfort (with the arm at
rest), (3) shoulder comfort (during sleep), (4) ability to use the arm full
time (work or play), (5) overall shoulder function, and (6) overall quality of
life. For these scales, the patients were asked to place a mark on a 10-cm
line, which was later measured to the nearest millimeter and was recorded as a
score from 0 to 100 (with zero reflecting a pain-free or problem-free
response). Analyses of the data were performed without any knowledge of
patient identity.
Forty-three patients (fifty shoulders) were followed for a minimum of five
years. Forty patients returned for examination and radiographs at regular
intervals. Three patients had a local orthopaedic surgeon send in the results
of a clinical examination and recent radiographs. The mean duration of
follow-up was 7.5 years (range, five to ten years). Fourteen additional
patients were excluded. Of the fourteen patients (fourteen shoulders), ten
were lost to follow-up, three had died because of unrelated causes, and one
had severe dementia and was unable to respond to the questionnaire. The
fourteen excluded patients were similar to the study group in terms of age,
sex, dominant hand, and diagnosis.
The study group included eighteen women and twenty-five men with an average
age of sixty-three years (range, forty-four to seventy-eight years).
Thirty-one (62%) of the fifty shoulders were on the dominant side. Thirty-two
shoulders in twenty-eight patients had primary osteoarthritis, thirteen
shoulders in eleven patients had secondary osteoarthritis, and five shoulders
in four patients had osteonecrosis of the humeral head. Of the shoulders with
secondary osteoarthritis, nine had posttraumatic arthritis and four had
failure of previous rotator cuff surgery. The average age of the patients with
primary osteoarthritis was sixty-nine years (range, fifty-five to
seventy-eight years) compared with an average age of fifty-nine years (range,
forty-four to seventy-six years) for patients with secondary osteoarthritis.
The risks and benefits of operative treatment were discussed in detail. All
patients in the study gave full verbal and written consent for operative
treatment and participation in this study. Our institutional review board
approved this study.
Operative Technique
Subscapularis and Capsular Release
The operation was performed with the patient under general anesthesia and
in the beach-chair position. A deltopectoral approach that emphasized
preservation of the entire deltoid origin and insertion was used.
The subscapularis was detached from its insertion on the lesser tuberosity
and was separated from the underlying anterior capsule. Adhesions from the
base of the coracoid process were released, and the entire anterior capsule
was excised. These releases were performed in order to free the subscapularis
muscle-tendon unit for 360° about its long axis. Subscapularis
reattachment was accomplished with use of 1-mm Cottony Dacron sutures
(Deknatel, Fall River, Massachusetts) passed through 2-cm-long bone tunnels
from the neck of the humerus to the osteotomy site. At the conclusion of the
arthroplasty, the subscapularis tendon was reattached in a more medial
location at the proximal humeral osteotomy site, thereby functionally
lengthening the muscle-tendon unit by 1 to 2 cm.
Resection of the Humeral Head
The articular surface of the humeral head was resected in a plane just
inside the rotator cuff insertion to the level of the greater tuberosity. The
humeral osteotomy varied between 20° and 30° of retroversion. The
location of the varus-valgus angle of the humeral head resection was
determined by placing a humeral osteotomy template along the anterior aspect
of the arm parallel to the shaft of the humerus. The superior lateral portion
of the mark was located at the junction of the articular surface with the
attachment of the rotator cuff on the greater tuberosity.
A glenoid that was nonconcentric and/or demonstrated moderate-to-severe
bone loss was reamed with a power reamer to normalize glenoid version with
respect to the scapula, and to provide a concentric glenoid articular surface.
Reamers were available in five different sizes, and the selection was based on
a set of trial glenoid components used for total shoulder arthroplasty.
Orientation of the reamer was facilitated by palpating the lateral extent of
the subscapularis fossa at a point midway between the superior and lateral
borders of the scapula. This point is described as the "centering
point"7
(Fig. 1). A line connecting
this point to the center of the glenoid face is the normalized glenoid
centerline. Orienting the reamer to the normalized glenoid centerline enables
the surgeon to correct pathologic glenoid version.
Assessment of Soft-Tissue Balancing and Closure
Periarticular osteophytes were removed from both sides of the shoulder
joint, and an anatomically sized trial humeral component was seated in place.
Soft-tissue balancing was judged to be optimal when the following criteria
were met: (1) posterior drawer testing demonstrated 40% to 60% translation of
the center of the prosthetic head relative to the center of the glenoid, (2)
=75° of internal rotation was present with the arm positioned in
90° of abduction, (3) the hand on the involved side could be placed on the
superior aspect of the contralateral shoulder without protraction of the
scapula, and (4) there was 45° of external rotation with the subscapularis
approximated to the proximal humeral osteotomy site.
Postoperative Protocol
A physician-directed postoperative rehabilitation program was used in all
patients. On the afternoon of the day of surgery, the physician performed
passive forward flexion of the arm up to 90°, or as far as was
comfortable, for each patient. On the first postoperative day, the patient
performed passive forward flexion with a pulley attached to an overhead frame,
passive external rotation with a meter stick, and pendulum exercises. Each
exercise involved five repetitions and was performed three to four times a
day, seven days a week. The patients were encouraged to use the arm for gentle
activities of daily living and were usually discharged on the second
postoperative day.
If the shoulder did not have sufficient passive motion by four to six weeks
after surgery, then additional stretching exercises were instituted. If
weakness of the deltoid was encountered, then a specific exercise program with
the patient in the supine position was instituted. The strengthening phase of
this program was added after the shoulder had achieved sufficient passive
range of motion (120° to 140° of flexion and 20° to 40° of
external rotation). This occurred typically by six to eight weeks after
surgery. Additionally, the patient performed resisted shoulder shrug exercises
for the trapezius. The serratus anterior and rhomboids were strengthened in
all patients with the use of wall push-ups; in male patients, the wall
push-ups were progressed to knee push-ups as symptoms allowed. The stretching
exercises were continued throughout the rehabilitation and were decreased to
one or two times a week in a maintenance program phase at four to six months
after surgery.
Radiographic Analysis
The preoperative, initial postoperative, and most recent radiographs were
evaluated. The radiographic evaluation included an axillary radiograph and an
anteroposterior radiograph made perpendicular to the plane of the scapula with
the humerus externally rotated 30°. A limited axial computed tomographic
scan was performed preoperatively in eight shoulders because the axillary
lateral radiograph was inadequate for evaluation. The radiographs were
reviewed to determine the presence of glenohumeral subluxation, periprosthetic
radiolucency, and glenoid bone loss (Fig.
2).
Loss of glenoid bone was measured on the axillary radiograph or computed
tomographic scan following the method of Iannotti and
Norris16. Zonal
analysis of humeral radiolucent lines was determined with use of an adaptation
of the system described by Gruen et
al.17. For the
evaluation, the area of humeral fixation was divided into eight zones
(Fig. 3). Zone 8 was the area
of the proximal part of the humerus beneath the collar of the humeral
component, zones 1 and 7 were the lateral and medial aspects of the proximal
third of the area containing the stem, zones 2 and 6 were the lateral and
medial aspects of the middle third of the area containing the stem, zones 3
and 5 were the lateral and medial aspects of the distal third of the area
containing the stem, and zone 4 was the area at the tip of the stem. The
maximum thickness of any radiolucent line in each of these zones was recorded
as the radiolucency for that zone.
Glenohumeral subluxation was evaluated on an axillary lateral radiograph
with respect to the direction and amount of translation of the center of the
prosthetic head relative to the center of the glenoid after a modification of
the method of Iannotti and
Norris16. The
center of the head was determined with use of concentric best-fit circles. The
center point between the anterior and posterior aspects of the glenoid rim was
marked on a line perpendicular to the scapular body. The distance from the
center of the prosthetic head to the center of the glenoid was measured and
divided by the diameter of the head. Varus-valgus orientation of the humeral
component was determined by radiographic measurement of the angle between the
humeral axis and the component. The difference in humeral head offset between
the immediate postoperative and most recent radiographs was used to determine
the presence and extent of glenoid erosion in
millimeters18.
Statistical Analysis
Statistical analysis was performed with SAS software (version 8.2; SAS
Institute, Cary, North Carolina). Within-patient comparisons of continuous
variables between different time-points were tested with two-tailed
matched-pairs t tests unless otherwise stated. A p value of =0.05 was
considered to be significant; if p = 0.1, it was considered to be a
statistical trend. Comparisons between groups were tested with use of
two-tailed t tests, and F tests were conducted in each instance to determine
whether group variances differed. When the variances differed
(a-critical = 0.05), the Satterthwaite method was used.
Clinical Findings
All patients reported substantial pain and functional limitations
preoperatively. At an average follow-up of 7.5 years, physical examination
indicated improvement in the pain score and the shoulder range of motion. The
mean active forward elevation was 94° (range, 40° to 150°)
preoperatively and improved to 120° (range, 10° to 170°) at the
final follow-up examination (p < 0.0001). External rotation improved from a
mean of 8° (range, -45° to 60°) to 33° (range, 0° to
70°); the difference was significant (p < 0.0001). The internal
rotation, measured as the highest vertebral level that the patient could reach
with the thumb extended, improved an average of three spinal segments (the
preoperative range was from the lateral aspect of the hip to T7, and the
postoperative range was from the posterior aspect of the hip to T8) (p <
0.0001, Wilcoxon matched-pairs signed-rank test). The scores on the American
Shoulder and Elbow Surgeons system were available for twenty-nine shoulders at
the time of the final follow-up. The mean score was 85 (range, 50 to 100).
An interim analysis of forty-five shoulders for which data were available
was conducted at a nominal follow-up period of approximately two years
(between twenty-two and twenty-six months). The average Simple Shoulder Test
was 9.7 positive responses, which was a significant improvement from the
preoperative mean of 4.2 positive responses (p < 0.0001). The improvement
in the visual analog scale scores at the time of the two-year follow-up
(Table I) also represented a
significant improvement over the preoperative values (p = 0.0001 for
each).
An evaluation of the results of the Simple Shoulder Test and visual analog
scale scores at the final follow-up evaluation (average, 7.5 years) indicated
that the significant improvements over the preoperative results observed at
the two-year postoperative follow-up had been sustained
(Table I). For the Simple
Shoulder Test, and for every visual analog scale measure, the results at the
time of the final follow-up were better than the preoperative results (p <
0.0001 for each).
Changes between the results at the two-year evaluation and the final
follow-up evaluation were examined. Despite the fact that the patients had
aged more than five years between these sets of measurements, all changes
observed were small in terms of clinical importance. Only one, overall
shoulder function, reached significance. The mean score on the Simple Shoulder
Test at the time of the final follow-up was 9.4 positive responses compared
with 9.7 positive responses at two years (p = 0.32). The mean visual analog
scale scores were 18.6 points for pain compared with 14.9 points (p= 0.45),
6.9 points for pain at rest compared with 3.1 points (p = 0.09), 12.4 points
for pain during sleep compared with 11 points (p = 0.79), 20.3 points for
overall shoulder function compared with 14 points (p = 0.04), 21 points for
ability to use the arm at work or play compared with 15.7 points (p = 0.12),
and 14.7 points for quality of life compared with 13.3 points (p = 0.84).
Patients with primary and secondary osteoarthritis were stratified into
groups and analyzed. The difference in the improvement in the scores on the
Simple Shoulder Test and the visual analog scale scores between patients with
primary and secondary osteoarthritis was not found to be significant, with the
numbers available.
Shoulders that presented with a normal rotator cuff were compared with
shoulders with thinning or tears of the rotator cuff. The mean values at the
time of the final follow-up for the Simple Shoulder Test, forward flexion,
external rotation, and the six visual analog scale parameters were tested with
two-tailed t tests. With the numbers tested, no trend suggesting superiority
of one group was observed with respect to the mean results for any of the
measures tested.
The same series of measurement tests were subsequently evaluated for only
the shoulders with rotator cuff abnormalities, and the results for those that
were repaired were compared with the results for those that were not repaired.
Again, with the numbers evaluated, no significant difference was observed
between these groups for any of the variables.
When all sixty-four shoulders originally considered for inclusion in the
study were included, the point estimate for survival at the eight-year
follow-up evaluation was 98.4%, according to Kaplan-Meier survival analysis,
with twenty at-risk (uncensored) shoulders remaining. The 95% confidence
interval for survival is 95.4% to 100% up to eight years postoperatively
(Fig. 4). At six years
postoperatively, forty-two shoulders were at risk; at seven years, thirty-five
shoulders were at risk; at nine years, eight shoulders were at risk; and at
ten years, three shoulders were at risk. There was one failure (conversion to
a total shoulder arthroplasty) at 9.3 years postoperatively, with seven
at-risk patients remaining (which was too few for the Kaplan-Meier 9.3-year
survival estimate of 86% to be considered
reliable)19.
Radiographic Findings
The most recent radiographs for fifty shoulders that had been followed for
an average of 7.5 years were assessed. Twenty-two shoulders had no lucency in
any zone. Three had a lucency in one zone; nine had lucencies in two zones;
twelve, in three zones; three, in four zones; and one, in five zones. The
majority of the lucencies occurred in zones 3, 4, and 5
(Table II).
Seventeen shoulders (34%) had erosion of the glenoid at the time of the
most recent follow-up. The mean amount of glenoid erosion was 2 mm (range, 1
to 4 mm).
Preoperatively, the glenoid bone loss (which was manifested as a
nonconcentric glenoid or a glenoid with increased retroversion due to
eccentric posterior erosion) was graded as mild in twenty shoulders (40%),
moderate in fourteen (28%), and severe in sixteen (32%). After glenoid
reaming, the glenoid bone loss was graded as mild in twenty-seven shoulders
(54%), moderate in fourteen (28%), and severe in nine (18%). Preoperatively,
thirty-six shoulders (72%) had no subluxation, thirteen (26%) had moderate
subluxation, and one had severe subluxation. Postoperatively, the results had
improved, with forty-six shoulders (92%) that had no subluxation and four that
had moderate subluxation (Figs. 5-A and
5-B).
Shoulders that had either moderate or severe glenoid bone loss, or moderate
or severe subluxation, or both, at the preoperative examination were placed
into one of two groups. Group I included twelve shoulders in which the
preoperative deficit had been corrected to mild bone loss and there was no
subluxation, and Group II included eighteen shoulders that had retained a
deficit with respect to bone loss, subluxation, or both. Preoperatively, a
difference between these groups was evident, as the mean pain score (and
standard deviation) on the visual analog scale was 52 ± 32 for the
twelve shoulders in Group I and 75 ± 16 for the eighteen shoulders in
Group II. Thus, Group I had more variation in the preoperative pain scores,
and the mean pain score was significantly less than that for Group II (p =
0.038, two-tailed Satterthwaite t test for unequal variances).
Postoperatively, a statistical trend was observed with respect to the mean
external rotation (and standard deviation), which was 39° ± 19°
for the Group-I shoulders compared with 27° ± 14° for the
Group-II shoulders (p = 0.06, two-tailed t test). A significant difference was
detected with respect to the mean improvement in external rotation, which was
38° ± 19° for Group-I shoulders compared with 19° ±
21° for Group-II shoulders (p = 0.015, two-tailed t test). There was no
difference in the mean postoperative pain scores between these groups.
Complications
Postoperative complications developed in four (8%) of the fifty shoulders.
One patient (who was one of the original fifty-seven patients) required
revision surgery to a total shoulder arthroplasty less than two years after
primary hemiarthroplasty. The patient had a gradual onset of pain with medial
migration of the prosthetic head. No evidence of loosening of the humeral stem
was noted on radiographic evaluation, and no evidence of infection was found
on culture of specimens obtained during the revision operation. Two patients
sustained early postoperative subscapularis ruptures and were treated with
pectoralis major tendon transfer. The first of these two patients sustained a
subscapularis rupture secondary to overly aggressive rehabilitation at a
physical therapy center at approximately four weeks after surgery. The patient
subsequently had development of anterior instability and underwent pectoralis
major transfer at ten months. The other patient noted a slipping sensation of
the humeral head at two months and underwent pectoralis major transfer at
three months. A nonfunctioning, attenuated subscapularis was noted at the time
of reoperation. The fourth patient had idiopathic brachial plexus neuritis
develop seven months following the operation.
Sperling et
al.10 reported on
the long-term results of shoulder arthroplasty in patients followed for a
minimum of five years. Of seventy-eight shoulders that had a hemiarthroplasty,
seventy-four that had been followed for a mean of 11.7 years were available
for clinical analysis. Three shoulders had been followed less than five years;
forty-nine shoulders, between eleven and fifteen years; and fourteen
shoulders, between sixteen and twenty years. Thirty-three (49%) of the
sixty-eight shoulders that were available for radiographic evaluation had an
unspecified degree of glenoid erosion preoperatively. At the time of the most
recent follow-up, forty-six shoulders (68%) had erosion of the glenoid, which
was graded as mild in nineteen, moderate in fourteen, and severe in thirteen
shoulders.
Patient selection in the study by Sperling et
al.10 and in the
present study was different in several respects. First, the mean age of the
patients in the study by Sperling et al. was thirty-nine years compared with
sixty-three years in this study. Second, there were differences with respect
to the diagnoses, with sequelae of trauma diagnosed in thirty shoulders (38%)
in the study by Sperling et al. and in nine shoulders in the present series;
rheumatoid arthritis and other inflammatory arthropathies, in thirty-three
shoulders (42%) and none, respectively; and primary osteoarthritis in eight
shoulders (10%) and thirty-two (64%), respectively. Last, there were more
shoulders that had a previous operation in the study by Sperling et al.
(twenty-six shoulders) than in our study (four shoulders).
Levine et al.20
retrospectively reviewed the cases of thirty-one patients who had been treated
with a hemiarthroplasty for glenohumeral arthrosis. All patients had an intact
rotator cuff and had been followed for an average of 2.4 years. Good or
excellent results were obtained in thirteen of fifteen shoulders with a smooth
concentric glenoid and in ten of sixteen shoulders with a nonconcentric
glenoid. A later study of the same patients at an average follow-up of 7.9
years revealed deterioration of the results with
time13. The number
of shoulders with a good or excellent result had decreased to eight of twelve
shoulders with a concentric glenoid and to nine of sixteen shoulders with a
nonconcentric glenoid. However, the report did not provide a detailed
description of the glenoid in terms of the quality of the articular surface,
degree of glenoid erosion, or presence of glenohumeral subluxation.
Conversely, the good and excellent results in the current study did not
deteriorate appreciably at a minimum follow-up of five years, which may be due
to several factors. First, the patients with degraded glenoid cartilage and
non-eburnated bone were managed with a total shoulder arthroplasty and were
excluded from the study. Second, moderate or severe posterior glenohumeral
subluxation that persisted following glenoid preparation and soft-tissue
balancing was improved by the placement of a prosthetic glenoid; these
patients were excluded and treated with total shoulder arthroplasty. Third,
spherical reaming of the glenoid along the glenoid centerline was performed in
all patients with moderate or severe posterior bone loss in order to optimize
glenohumeral
stability21.
Finally, every effort was made to provide optimal soft-tissue balancing. In
our experience, this is of paramount importance in hemiarthroplasty. As in
many other conditions that require shoulder arthroplasty, typically the
capsule and ligaments are found to be contracted. Shoulder arthroplasty tends
to tighten the capsule further; unless anatomically sized components are used
and sufficient capsular releases are performed, the joint is
"overstuffed," which further compounds shoulder stiffness and
glenohumeral
subluxation7,22.
An analysis of zonal lucencies showed that the majority of patients had
either no lucency or lucencies only near the tip of the stem. These findings
were similar to those in previous reports pertaining to cementless
hemiarthroplasty10,23.
For the shoulders that had preoperative glenoid bone loss or subluxation
that had been corrected, the mean improvement in external rotation (39°)
was similar to the result for patients managed with total shoulder
arthroplasty (32°) in the report by Iannotti and
Norris16. In
contrast, the mean improvement in external rotation for the shoulders in this
study that had bone loss or subluxation, or both, that remained
postoperatively (19°) was similar to the result for the shoulders treated
with hemiarthroplasty (16°) in the study by Iannotti and
Norris16.
Sperling et
al.12 found the
postoperative range of motion and the improvement in strength did not vary
markedly with respect to the presence of a tear of the rotator cuff that was
repaired at the time of hemiarthroplasty. Additionally, a tear of the rotator
cuff was not found to be a risk factor for an unsatisfactory clinical result.
Reflecting the findings of Sperling et al., the present study demonstrated no
significant differences between shoulders with a normal rotator cuff and those
that underwent repair of the rotator cuff with respect to the Simple Shoulder
Test, visual analog scale scores, or motion. Moreover, there was also no
difference between the shoulders that had a rotator cuff tear repaired and
those that did not.
The inferences that may be drawn from the current study are subject to
limitations: (1) the patient population was limited to those who had primary
or secondary osteoarthritis, (2) the parameters established for optimal
soft-tissue balancing involved intraoperative assessment with the patient
under anesthesia, (3) the evaluation of glenoid bone characteristics was
limited to radiographic studies and intraoperative qualitative assessment, and
(4) the patient sample was from one institution. Accordingly, the results may
not be applicable to other patient populations. Despite these limitations, our
results support the hypothesis that hemiarthroplasty is a good option for
carefully selected patients with glenohumeral arthrosis. When the selection
and management principles described in this study are used to guide surgical
decisions, shoulder hemiarthroplasty appears to be effective in providing
good-to-excellent results that are maintained five to ten years
postoperatively. ?