The young patient with glenohumeral arthritis who has disabling pain and
loss of motion but wants to return to active sporting activities or a job
requiring heavy lifting or overhead use poses a difficult treatment problem.
Traditionally, these patients, especially those who perform heavy labor, have
been denied total shoulder arthroplasty and have been considered candidates
for glenohumeral arthrodesis. Although pain relief following glenohumeral
arthrodesis is satisfactory in many cases, persistent pain and stress on the
scapulothoracic musculature and acromioclavicular joint combined with marked
loss of motion make this an unattractive option for many patients, especially
those who want to do overhead
activity1.
Accelerated loosening of polyethylene glenoid components and increased
polyethylene wear are theoretical concerns when offering total shoulder
arthroplasty to younger patients. Although total shoulder arthroplasty with
porous-coated implants seemed an attractive alternative for the younger
patient, polyethylene dissociation, especially in patients with preexisting
instability, and increased polyethylene wear associated with the metal-backed
components make this option less
appealing2-8.
An alternative is to perform humeral head replacement alone in these
patients. However, in several large series, hemiarthroplasty, when compared
with total shoulder arthroplasty, was found to be inferior with respect to
pain relief and return of motion and
function9-13.
This result has been especially true in our experience with younger male
patients who have post-traumatic or postreconstructive arthritis and glenoid
erosion.
In an effort to offer patients the pain relief and range of motion achieved
in association with total shoulder arthroplasty, without the complications of
polyethylene wear, cement fragmentation, and glenoid loosening or
dissociation, a technique has been developed to resurface the glenoid
biologically in conjunction with humeral head replacement. In the present
report, we describe the intermediate-term results of this procedure.
Between November 1988 and November 2003, the senior authors (W.Z.B. and
S.G.K.) performed humeral hemiarthroplasty with biologic resurfacing of the
glenoid on thirty-six shoulders in thirty-four patients with glenohumeral
arthritis. Fourteen of these thirty-six shoulders were initially studied
between 1989 and 1992, and the early results for some of those patients were
reported
previously14. All
thirty-four patients were selected for this technique on the basis of age
and/or a high functional demand of the shoulder that would raise substantial
concern about potential loosening of a conventional polyethylene glenoid
component. All shoulders that were treated during the time-period of the study
were treated with this technique, and the patients were followed
prospectively. The study group included thirty men and four women. A
bilateral, staged procedure was performed in two of the men. Thirty-three
shoulders were on the dominant side, and three were on the nondominant side.
The mean age of the patients was fifty-one years (range, thirty to
seventy-five years). Only one patient (a seventy-five-year-old man) exceeded
the age of sixty years, and he did not desire conventional polyethylene
glenoid replacement because he performed heavy manual labor as an active
rancher and cattle roper. The diagnoses included primary osteoarthritis
(eighteen shoulders; 50%), arthritis after reconstruction for instability
(twelve), posttraumatic arthritis (five), and osteonecrosis (one). No patient
demonstrated inflammatory arthropathy. Sixteen of thirty-six shoulders had
undergone previous surgery. All patients had been followed for one to four
years prior to the glenoid resurfacing procedure, with conservative care
including anti-inflammatory medications, injections, and a home
range-of-motion exercise program. All patients had ceased overhead activities
for two to five years before the arthroplasty and had both severe night pain
and pain with activities of daily living.
Three types of biologic resurfacing materials were studied: anterior
glenohumeral capsule (seven shoulders; 19%), autogenous fascia lata (eleven
shoulders; 31%), and Achilles tendon allograft (eighteen shoulders; 50%), the
latter of which was used in the most recent cases. Depending on which surgeon
performed the procedure, the glenoid resurfacings were all performed in
combination with the use of either a modular uncemented porous-coated
(twenty-six shoulders) or a cemented (ten shoulders) humeral head stemmed
replacement.
All thirty-six shoulders in the present study were followed for a minimum
of two years (mean, eighty-four months; range twenty-four to 180 months). All
patients were prospectively followed, both clinically and radiographically, at
three, six, and twelve weeks; at six and twelve months; and at yearly
intervals thereafter until the time of the most recent follow-up. Clinical
evaluation was performed with use of both the Neer
score15 and the
scoring system of the American Shoulder and Elbow Surgeons
(ASES)16. The
subjective pain level was documented with use of a visual analog scale (with 0
indicating no pain and 10 indicating the worst pain ever). Patient
satisfaction was recorded with use of a similar 10-point visual analog scale
(with 0 indicating that the patient was unsatisfied and 10 indicating that the
patient was completely satisfied). Preoperative and postoperative radiographs
(including anteroposterior views in neutral and external rotation, an axillary
lateral view, and a supraspinatus outlet view) were available for all
patients. Statistical analysis was performed with analysis of variance for
quantitative values and with a chi-square test for qualitative values. The
level of significance was set at p < 0.05.
Surgical Technique
The patient was placed in the beach-chair position after the induction of
general anesthesia. A modified straight deltopectoral approach was performed.
In all thirty-six cases, the glenoid was prepared and resurfaced prior to
preparation of the humerus for stem insertion.
Glenoid Resurfacing with Anterior Capsule or Fascia Lata (First
Eighteen Shoulders)
A subscapularis tendon and anterior shoulder capsule z-plasty as described
by Rockwood17 was
performed in the first eighteen shoulders. Osteophytes were removed from the
humeral head, and the articular margin was identified. A humeral head
osteotomy at the anatomic neck was performed, and then attention was turned to
the glenoid surface, which was prepared to bleeding subchondral bone with use
of either a spherical reamer with a guiding pilot tip or a power burr along
the glenoid centerline to normalize glenoid version perpendicular to the
scapular spine. Multiple drill-holes with a 2.0-mm drill-bit (usually five to
seven holes to a depth of 5 mm each) were then made in the glenoid surface. In
seven shoulders, the anterior glenohumeral capsule was placed over the
recontoured glenoid and was sutured to the posterior labrum with use of
number-2 Ethibond (Ethicon, Somerville, New Jersey). In the other eleven
shoulders, a free autogenous fascia lata graft (2 × 4 cm) was obtained
from the ipsilateral thigh. The fascia lata graft was sutured to the existing
labrum with its anatomic outer surface as the new articular surface
(Fig. 1).
Glenoid Resurfacing with Achilles Tendon Allograft (Second Eighteen
Shoulders)
In the second eighteen shoulders, an Achilles tendon allograft was used as
the resurfacing graft in order to reduce graft harvest-site morbidity and to
increase the thickness of tissue available for resurfacing. In these cases,
the subscapularis was released either directly from its lateral insertion on
the lesser tuberosity or with a small fleck of bone from the lesser
tuberosity. The humeral head osteotomy was performed after osteophyte
removal.
The Achilles tendon resurfacing technique began with thawing of the graft
and sharp removal of the osseous calcaneal attachment immediately at the
bone-tendon junction (Fig. 2).
The tendon was then contoured by folding it back onto itself several times (a
minimum of two times and preferably three times) to create a shape of
appropriate diameter for the involved glenoid, with a thickness varying from 5
to 8 mm, depending on the caliber of the allograft tendon
(Fig. 3). A running everting
mattress stitch of number-5 Ethibond (Ethicon) was used around the periphery
to secure the shape of the graft.
The glenoid surface was then decorticated with a burr or standard glenoid
reamer to create a bleeding surface. At this point, four bioabsorbable anchors
(Panalok 3.5 anchors; Mitek, a Johnson and Johnson Company, Norwood,
Massachusetts) with number-2 Ethibond suture (Ethicon) were placed at the
so-called cruciate corners of the glenoid—the positions on the glenoid
that correspond to the positions on a clock-face of 12, 3, 6, and 9 o'clock
(Fig. 4). Metal anchors were
avoided because of potential loosening in the soft bone and erosion through
the graft into the joint.
The Achilles tendon graft was then secured to the bleeding glenoid surface
with suture anchors that were tied in a horizontal mattress fashion. The
periphery of the graft was then sutured directly to the glenoid rim with use
of a minimum of four transosseous Bankart-type sutures of number-2 Ethibond
(Fig. 5).
Humeral Stem Preparation (All Shoulders)
After glenoid resurfacing, final humeral reaming and broaching was
performed. The humeral implant was inserted (either with or without cement),
and the modular humeral head was then applied. In the first eighteen cases,
the subscapularis z-plasty was repaired end-to-end with the arm in as much
external rotation as possible with number-1 cottony Dacron suture (Deknatel,
Fall River, Massachusetts). In the remaining eighteen cases, the subscapularis
(and, if present, the bone fleck) was reattached at its anatomic insertion on
the lesser tuberosity through transosseous drill-holes with similar heavy
nonabsorbable sutures.
Postoperative Care
All patients were managed with a Velpeau arm sling or shoulder immobilizer
for four weeks and immediate passive range-of-motion exercises. After four
weeks, the sling was removed and full active motion was allowed, with
resistance exercises begun by the eighth week after surgery.
The mean ASES score was 39 points (range, 28 to 50 points) preoperatively
and 91 points (range, 63 to 100 points) at the time of the latest follow-up at
a minimum of two years (mean, seven years) postoperatively (p < 0.01).
According to Neer's criteria, the result was excellent for eighteen shoulders
(50%), satisfactory for thirteen (36%), and unsatisfactory for five (14%).
Pain relief was excellent (twenty-two shoulders) or good (nine shoulders) for
86% of the shoulders in this series, with an overall decrease in the level of
pain from 7.7 to 2.1 (p < 0.03). The mean level of patient satisfaction was
9.0 on the 10-point scale. Thirty-one patients (91%) stated that they would
undergo the procedure again. Twenty-nine patients (85%) returned to premorbid
activities, including polo, firefighting, police duty, farming, ranching,
roping, heavy work (pushing and pulling as much as 200 lb [91 kg] or overhead
lifting of >50 lb [23 kg]), push-ups and weight-lifting, golf, and
tennis.
The mean active anterior elevation was 70° (range, 40° to 120°)
preoperatively and 140° (range, 50° to 170°) postoperatively, for
an overall mean improvement of 70° (p < 0.01). The mean external
rotation was 5° (range, —30° to 15°) preoperatively and
50° (range, 0° to 75°) postoperatively, for an overall mean
increase of 45° (p < 0.02). The mean internal rotation was to the
sacroiliac joint (range, greater trochanter to the L5 spinous process)
preoperatively and to the T12 spinous process (range, L5 to T8)
postoperatively, for a mean increase of six spinal segments (p < 0.05).
At the time of the most recent radiographic evaluation, all twenty-six
shoulders with a cementless humeral stem had a stable interference fit with
sclerosis proximally and no progressive radiolucent lines in the cancellous
bone adjacent to the porous-coated pads. All cementless stems were classified
as Engh grade IA or
IB18, with no
radiolucent lines in the region of the porous-coated proximal body. All ten
cemented humeral stems demonstrated no evidence of progressive radiolucent
lines at the cement-bone interface or radiographic evidence of loosening at
the time of the most recent review. There had been no revisions for humeral
loosening.
At the time of the most recent review, glenoid erosion (measured as the
decrease in distance from the coracoid base to the glenoid rim on neutral
rotation anteroposterior radiographs) averaged 7.2 mm (range, 3 to 9 mm). For
the twenty shoulders with radiographic follow-up of longer than sixty months,
glenoid erosion appeared to stabilize at approximately five years with no
further progression of medial erosion toward the coracoid base
(Fig. 6). Preoperatively, all
thirty-six shoulders demonstrated complete obliteration of the glenohumeral
joint space on the axillary lateral radiograph. Immediately postoperatively,
the glenohumeral joint space on the axillary lateral radiograph averaged 2.9
mm (range, 2.0 to 7.1 mm), and, at the time of the most recent follow-up, it
averaged 1.3 mm (range, 0.5 to 3.0 mm). The glenohumeral joint space decreased
from the immediate postoperative period to the final radiographic evaluation
in all thirty-six shoulders, but it appeared to stabilize with no further
decrease at approximately five years in the twenty shoulders with more than
sixty months of follow-up. With the numbers available, there was no apparent
difference between the eighteen shoulders in the autogenous graft group and
the eighteen shoulders in the allograft group with regard to the mean
glenohumeral joint space (0.7 compared with 1.9 mm, respectively). However, no
shoulder in the allograft group demonstrated a joint space of =1 mm whereas
seventeen of the eighteen shoulders in the autogenous graft group demonstrated
a joint space of =1 mm.
Complications
The complications in the present series included three cases of
postoperative instability (all three after anterior capsular resurfacing), two
infections (one in a shoulder that had been resurfaced with anterior capsular
tissue and one in a shoulder that had been resurfaced with fascia lata), one
case of brachial plexitis, and one upper extremity deep-vein thrombosis. No
donor-site complications occurred at the autograft fascia lata harvest area.
No adverse reactions to the Achilles tendon allograft tissue or infections in
the allograft group were experienced.
Unsatisfactory Results and Reoperations
Factors associated with unsatisfactory results included the use of anterior
capsule as the resurfacing graft and infection. Of the five shoulders with a
poor result, three had had pain and instability after resurfacing with
anterior capsule, one had had development of an infection after resurfacing
with anterior capsule, and one had had development of an infection following
resurfacing with fascia lata (Table
I). Of the two infections, one was treated successfully with
intravenous and oral antibiotics alone; the other required operative wound
débridement and intravenous antibiotic therapy with successful
retention of components and graft.
Three of the five shoulders with a poor result (all three of which had had
anterior capsular resurfacing without development of an infection) were
revised to total shoulder arthroplasty with a cemented polyethylene glenoid
component within the first five years after the index arthroplasty, two with
an excellent result and one with a poor result.
Another reoperation was performed in a patient who initially was classified
as having an excellent result but had development of persistent anterior
shoulder pain within the first twelve months after the initial arthroplasty.
Injections into the region of the biceps tendon consistently relieved the pain
and, subsequently, a biceps tenodesis was performed. At the time the of
tenodesis, arthroscopy of the resurfaced glenohumeral joint (which had been
resurfaced with autogenous fascia lata) was performed. Although visualization
was difficult, a small biopsy sample of tissue on the glenoid surface was
obtained. This sample showed only fibrous tissue and no fibrocartilage.
However, arthroscopy confirmed the maintenance of a fibrous layer between the
humeral head implant and the bone of the glenoid. This patient subsequently
was returned to the "excellent" category at the time of the final
review.
Results According to Type of Biologic Graft: Autogenous Tissue
Compared with Allograft Tissue
When the eighteen shoulders that had been treated with autogenous graft
were compared with the eighteen shoulders that had been treated with
allograft, there was no difference between the groups, with the numbers
available, with regard to postoperative range of motion. However, the average
pain score was 4.6 in the autogenous graft group, compared with 1.4 in the
allograft group (with 0 indicating no pain and 10 indicating the worst pain)
(p < 0.04). Glenoid erosion, as measured radiographically, averaged 8.0 mm
in the autogenous group compared with 5.8 mm in the allograft group, but this
difference did not reach significance (p = 0.07).
As stated previously, five (28%) of the eighteen shoulders in the
autogenous group had an unsatisfactory result. Four of the seven shoulders
that had been treated with anterior capsular tissue had an unsatisfactory
result, and three of these four shoulders underwent revision to a conventional
polyethylene glenoid component. Only one of the eleven shoulders that had been
treated with fascia lata had an unsatisfactory result, and none of these
eleven shoulders underwent revision to a conventional glenoid component.
Finally, none of the eighteen shoulders that had been treated with Achilles
tendon allograft had an unsatisfactory result or revision to a conventional
glenoid component.
The initial results of biologic resurfacing of the glenoid with
hemiarthroplasty of the shoulder for the treatment of arthritis were reported
in 1995 by Burkhead and
Hutton14. Fourteen
of the thirty-six shoulders in the current series initially had been studied
between 1989 and 1992, and six of these fourteen shoulders had been followed
for at least two years at that time. The biologic resurfacing was done with
either autogenous fascia lata or anterior shoulder capsule; Achilles tendon
had not yet been utilized as a graft source during this early time-period.
According to Neer's criteria, there were five excellent results and one
satisfactory result among the six shoulders that had been followed for at
least two years. All six shoulders had relief of pain. The intermediate to
long-term results reported in the current series speak to the durability of
biologic resurfacing of the glenoid when combined with humeral
hemiarthroplasty for young, active patients with glenohumeral arthritis. As
demonstrated by the subjective results and the objective measures of motion
and function, biologic glenoid resurfacing can provide pain relief similar to
total shoulder arthroplasty for these patients. It allows young, active
patients the ability to maintain an active lifestyle, including weight-lifting
and manual work, without the worrisome risk of polyethylene wear of a
conventional total shoulder glenoid component.
Objective radiographic review of the patients in this series, while limited
to plain radiographic evaluation and not computed tomographic evaluation,
demonstrated stabilization of glenoid erosion by five years postoperatively
with a persistently maintained joint space. Limited examination at the time of
arthroscopy (one shoulder) or conversion to polyethylene glenoid components
(three shoulders) demonstrated that a fibrous tissue interface had formed on
the glenoid surface. While these studies were only performed for patients with
an autogenous graft (anterior capsule or fascia lata), it does appear that
glenoid bone stock may be better maintained in shoulders that are treated with
a biologic glenoid graft than in those that are treated with an isolated
hemiarthroplasty. Thus, eventual conversion to a conventional glenoid
component may be relatively uncomplicated. This is encouraging because the
literature documents less satisfactory outcomes after conversion of an
isolated humeral hemiarthroplasty to a total shoulder arthroplasty than after
initial primary total shoulder
arthroplasty13.
On the basis of the results in the current series, we believe that
autogenous tissue (either anterior capsule or fascia lata) is not the optimal
resurfacing tissue. While good results (up to fifteen years of durability) can
be obtained with the use of autogenous fascia lata, we do not believe that it
provides a durable bearing surface for reproducible outcomes with this
operation. Furthermore, several patients in the present series who were
managed with an anterior capsule interposition reported higher levels of
postoperative pain as compared with the levels reported by those who were
managed with the other two methods. Because the capsular resurfacing technique
involves leaving a medial soft-tissue attachment, we presume that persistent
nerve fibers in the capsule can act as a potential source of pain. This
problem can therefore be avoided with the use of allograft tissue. In
addition, three of the shoulders that had been treated with anterior capsular
resurfacing had development of postoperative instability in addition to the
recalcitrant pain, requiring conversion to a conventional glenoid component.
On the basis of these findings, it does not appear that anterior capsule is a
reliable and/or durable bearing surface for successful biologic glenoid
resurfacing.
Because of its relatively lower cost, easy availability, and large bearing
surface, our current graft of choice for biologic glenoid resurfacing is a
cadaveric Achilles tendon, which provides a much thicker graft material as
compared with anterior capsule, fascia lata, or other potential allograft
bearing surfaces. The intermediate-term results reported in the present series
confirm the durability of this procedure in terms of both pain relief and
functional restoration (which appear to be comparable with those of
conventional total shoulder arthroplasty) in younger and highly active
patients with end-stage glenohumeral arthritis. Currently, we believe that
this procedure is appropriate for active patients with an age of less than
sixty years and/or patients with high functional (either recreational or
occupational) demands that would increase the potential for mechanical
loosening of a conventional glenoid component. ?