Patient Selection
We evaluated fifty-three consecutive patients who had undergone
arthroscopic repair of an isolated supraspinatus tear in a standardized
fashion at our institution between 2000 and 2001 by one of the two senior
surgeons (S.L. and P.H.).
Patients were excluded from the study if they had involvement of rotator
cuff tendons other than the supraspinatus, it was not possible to achieve a
watertight tension-free repair, a previous operation had been performed on the
affected shoulder, or the patient had glenohumeral arthritis or inflammatory
arthropathy.
Clinical Evaluation and Operative Technique
The patients were examined clinically on the day before the surgery, and
the clinical results were documented with use of the Constant
score19. Both
surgeons used an identical operative technique. The operations were performed
with the patients in the beach-chair position. After diagnostic arthroscopy
and subacromial decompression (in all but one patient), the supraspinatus tear
was localized and the tear size was documented according to the Bateman
classification20. A
modified Mason-Allen technique with a single row of bio-absorbable suture
anchors preloaded with two number-2 nonabsorbable braided polyester sutures
(Bio-Corkscrew; Arthrex, Naples, Florida) was used to achieve a tension-free
repair of the rotator cuff in all
patients21.
Postoperatively, the shoulder was immobilized in a sling for forty-eight
hours, after which immobilization continued in an abduction pillow for three
weeks. During the first six weeks, physiotherapy consisted of passive
range-of-motion exercises for the shoulder. Range-of-motion limits were
continuously increased from 60° of abduction, 60° of flexion, and
10° of external rotation in the first week to 90° of abduction,
145° of flexion, and 45° of external rotation in the sixth week. At
seven weeks, a free passive range of motion was allowed and active
mobilization was begun. At nine weeks, carefully performed isometric
strengthening exercises, with respect for the patient's pain limit, were
started, and the intensity of these exercises was increased to eccentric
strengthening and weight training in the twelfth week.
After an average duration of follow-up of 26.4 months (minimum, twenty-four
months), all patients were again clinically evaluated with use of the Constant
score and underwent a standardized magnetic resonance imaging examination at
our institution.
Magnetic Resonance Imaging
Examination
Twenty-nine patients (55%) provided preoperative magnetic resonance images
acquired elsewhere, and the other twenty-four (45%) underwent magnetic
resonance imaging at our institution, either because the study had not yet
been performed or it was considered inadequate or too old.
At the time of follow-up, all patients underwent a standardized magnetic
resonance imaging examination at our institution. We used an open low-field
(0.2-T) magnetic resonance imaging system with a shoulder coil. The sequences
used for the examination included an oblique coronal T1-weighted spin-echo
sequence (echo time, 24 msec; repetition time, 770 msec), an oblique coronal
T2-weighted turbo-spin-echo sequence (repetition time, 3000 msec; echo time,
80 msec), an oblique sagittal T1-weighted spin-echo sequence (echo time, 24
msec; repetition time, 770 msec), and an axial T1-weighted spin-echo sequence
(repetition time, 870 msec; echo time, 24 msec).
Evaluation of Images
The preoperative and postoperative magnetic resonance imaging scans were
independently evaluated by each of two observers who were blinded to the
clinical outcome of the patient. Eight (15%) of the preoperative scans either
were not available for reevaluation or were considered inadequate to analyze
all three criteria (described below). This left forty-five preoperative scans
to be evaluated. The findings of each observer were documented, and average
values were calculated.
The three criteria used for the evaluation were the integrity of the
rotator cuff, atrophy of the supraspinatus muscle, and fatty infiltration of
the supraspinatus, infraspinatus, and subscapularis muscles. The integrity of
the repaired supraspinatus tendon was evaluated with use of established
criteria22. The
oblique coronal T2-weighted turbo-spin-echo sequence was carefully inspected
for fluid-equivalent signal in the tendon or complete nonvisualization of the
tendon on at least one sequence. Either finding was considered to represent a
retear of the repaired tendon.
Supraspinatus atrophy was evaluated on the most lateral of the oblique
sagittal images on which the scapular spine was in contact with the scapular
body. The tangent sign introduced by Zanetti et al. was drawn on this
image23. The
position of the supraspinatus muscle belly in relation to the tangent sign was
graded with a qualitative three-stage system corresponding to the quantitative
system introduced by Thomazeau et
al.24. Atrophy was
considered to be Grade 1 if the muscle was superior to the tangent, it was
considered to be Grade 2 if the muscle just touched the tangent, and it was
considered to be Grade 3 if the muscle was clearly below the tangent. Since
there are no established criteria for qualitative assessment of atrophy of the
infraspinatus, teres minor, or subscapularis, only atrophy of the
supraspinatus muscle was evaluated.
Fatty infiltration of the supraspinatus, infraspinatus, and subscapularis
was assessed on the same oblique sagittal image, with use of the
classification system described by Goutallier et
al.13. With this
system, Stage 0 indicates no fatty infiltration of the muscle; Stage 1, some
fatty streaks; Stage 2, less fat than muscle; Stage 3, equal muscle and fat
contents; and Stage 4, more fat than muscle. This system was originally
designed for use in computed tomography examinations but was later correlated
for magnetic resonance
imaging12.
Statistical Analysis
Statistical analysis was performed with use of SPSS statistical software
(version 13.0; SPSS, Chicago, Illinois). The level of significance was set at
0.05. Preoperative and postoperative nonparametric data were analyzed with use
of the Wilcoxon signed-rank test. Comparisons between two groups were
performed with use of the Mann-Whitney U test.
Patient Demographics and Clinical Outcome
Fifty-three consecutive patients met the above-mentioned criteria and were
included in this study. The average age (and standard deviation) of our cohort
was 60.9 ± 7.3 years, and the average duration of follow-up was 26.4
months (minimum, twenty-four months). The study group consisted of thirty-four
men and nineteen women. The dominant side was affected in thirty-seven
patients, and the nondominant side was affected in sixteen. The influence of
the demographics on the clinical outcome is outlined in
Table I. The only factor with
significant influence on the clinical outcome was the patient's age (p =
0.002). The integrity of the repair was also significantly influenced by the
patient's age (p = 0.011) as well as by the duration of symptoms
preoperatively (p = 0.033).
Preoperatively, the average overall Constant score was 53.5 points (range,
23 to 80.8 points). At the time of follow-up, the score was significantly
improved to 83.4 points (range, 62.5 to 99.3 points) (p < 0.001). All
single parameters of the Constant score (pain, activities of daily living,
range of motion, and abduction strength) were also significantly improved
compared with the preoperative scores
(Table II).
Tendon Integrity
Evaluation of standardized postoperative magnetic resonance imaging scans
revealed thirteen retears (25%). Preoperatively, no single parameter of the
Constant score, nor the overall Constant score, differed significantly between
the group with a retear and the group with an intact tendon. Postoperatively,
tendon integrity had a significant influence only on the abduction strength,
which was significantly lower (p = 0.043) in the retear group (8.4 points)
than in the intact-tendon group (13.9 points). The parameters of pain,
activities of daily living, and range of motion were not significantly
influenced by tendon integrity. However, the overall Constant score was still
significantly lower (p = 0.012) in the retear group (78.9 points) than in the
intact-tendon group (86.1 points). The only demographic parameter that had a
significant influence on the clinical result was the patient's age (p =
0.002). Patients with a structural failure of the repair were significantly
older (65.3 years) than those with an intact tendon (59.5 years) (p =
0.011).
Supraspinatus Atrophy
Atrophy of the supraspinatus muscle measured on the preoperative magnetic
resonance imaging scan was a positive predictor of the integrity of the repair
and therefore of the clinical outcome. There was significantly greater
preoperative atrophy of the supraspinatus muscle in the retear group than in
the intact-repair group (p = 0.014). The retear rate associated with Grade-2
atrophy of the supraspinatus muscle was significantly higher than that
associated with Grade-1 atrophy (p = 0.018). There was also a significantly
higher degree of postoperative supraspinatus atrophy in the retear group (p
< 0.001). When we compared preoperative and postoperative magnetic
resonance imaging scans it was evident that atrophy of the supraspinatus was
stabilized in shoulders with a successful repair; however, there was no
evidence of reversal. In the retear group, the amount of postoperative
supraspinatus atrophy had increased significantly compared with the
preoperative state (p = 0.011) (Table
III).
Fatty Infiltration
Overall, regardless of tendon integrity, there was a significant increase
in fatty infiltration in the supraspinatus (p = 0.015) and the infraspinatus
(p = 0.001) from the preoperative to the postoperative examinations. Fatty
infiltration of the subscapularis did not increase significantly (p = 0.059).
Analysis of the available forty-five preoperative magnetic resonance imaging
scans showed differences in the extent of fatty infiltration of the
supraspinatus, infraspinatus, or subscapularis between the retear group and
the intact-tendon group, but these differences did not reach significance with
the small number of patients evaluated in this study. However, the retear rate
was significantly higher for patients with Stage-2 fatty infiltration of the
supraspinatus on the preoperative magnetic resonance imaging scans than it was
for those with Stage-0 or 1 fatty infiltration (p = 0.021)
(Table IV).
Postoperatively, fatty infiltration of the supraspinatus (p < 0.001) and
infraspinatus (p = 0.001) was significantly greater in the retear group than
in the group with an intact repair. Fatty infiltration of the subscapularis
did not differ significantly between the groups (p = 0.067).
Comparison of preoperative and postoperative magnetic resonance imaging
scans revealed no decrease in fatty infiltration in either the intact-repair
group or the retear group. The retear group had a significant increase in
fatty infiltration of the supraspinatus (p = 0.014) and the infraspinatus (p =
0.016) but not of the subscapularis. In the intact-repair group, fatty
infiltration of the infraspinatus increased significantly (p = 0.001) but the
supraspinatus and subscapularis showed no significant difference in fatty
infiltration between the preoperative and postoperative magnetic resonance
imaging examinations (Table
III).
The clinical results of arthroscopic rotator cuff repair have been good to
excellent and comparable with those reported following open or mini-open
rotator cuff
repair25-30.
While the clinical outcomes and patient satisfaction have been shown to be
equal to those following established open techniques, structural changes
associated with arthroscopic repairs of the rotator cuff and the integrity of
these repairs are not well known. Failure has not been exclusive to
arthroscopic repairs, as numerous studies of open rotator cuff repairs have
also shown
failures11,14-16,18,31,32.
The retear rates reported in those studies varied considerably, although some
reports have suggested that there are more retears after arthroscopic rotator
cuff repairs than after open
repairs10.
When different studies are analyzed, it is important to consider the extent
of the tears. When used for isolated tears of the supraspinatus, arthroscopic
repairs are associated with retear rates comparable with those reported after
open repair. In our study, the retear rate following use of a standardized
technique involving a single row of suture anchors with Mason-Allen stitches
was 25% (thirteen of fifty-three). This percentage is in agreement with the
25% rate of retears (six-teen of sixty-five) recently found by Boileau et al.
after treatment of isolated supraspinatus tears with the tension-band suture
technique1. Retear
rates have been considerably higher after repairs of massive tears of the
rotator
cuff10,15.
Muscle atrophy of the supraspinatus was shown to be an important prognostic
factor in our study group. More severe atrophy seen on preoperative magnetic
resonance imaging was associated with a higher percentage of failed repairs.
Furthermore, patients with a retear showed postoperative progression of the
muscular atrophy accompanied by an inferior clinical result. Preoperative
fatty infiltration exceeding Stage 1 was also a prognostic factor for a
recurrent tear. In this regard, our results after use of the arthroscopic
technique are in agreement with the findings following open repair. Goutallier
himself identified Stage-1 fatty infiltration as a cutoff between tendon
integrity and failure following open
repair17. These
findings were confirmed recently by Mellado et al., who reported a higher
prevalence of recurrent defects following open rotator cuff repair if
preoperative fatty infiltration had exceeded Stage
115.
The question of whether fatty infiltration can be reversed by successful
rotator cuff repair is controversial. We found no evidence that this was the
case. Neither muscle atrophy nor fatty infiltration decreased after a
successful repair, but both parameters significantly worsened in patients with
a retear. This finding is in agreement with those of recent studies of open
rotator cuff
repair14, but to
our knowledge this issue has not been previously examined following
arthroscopic rotator cuff repair. It should be noted that there was a
significant increase in fatty infiltration of the infraspinatus in our study
group, regardless of the integrity of the repair. In a recent study addressing
open repair of an isolated tear of the supraspinatus or subscapularis, Fuchs
et al. also found a significant increase (p = 0.018) in fatty infiltration of
the infraspinatus when the supraspinatus had been repaired but not when the
subscapularis had been
repaired14. As
there was no clinical correlation with these magnetic resonance imaging
findings in either the study by Fuchs et al. or our study, no sufficient
explanation can be given. Additional studies examining possible neurologic
damage to the suprascapular nerve during mobilization of the supraspinatus
could improve our knowledge regarding these findings.
The clinical implication of the results of our study is that only a
successful repair can halt the atrophy of the muscle and lead to an optimal
clinical result regarding strength. Because strength accounts for 25% of the
Constant score, which was chosen as the outcome measure in our study, the
retear group had a significantly lower overall Constant score. This finding
was reported by other investigators following open and arthroscopic
repair1. It should
be noted that a recurrent defect of the rotator cuff had no impact on the
other parameters measured by the Constant score (pain, activities of daily
living, and range of motion). It will be interesting to see whether the
difference in the clinical outcome between the retears and intact repairs will
become more evident with longer follow-up. Recently, there has been evidence
that the clinical results associated with structural failures of rotator cuff
repairs do not deteriorate over
time31. An
important finding of our study was the fact that higher age was associated
with a higher prevalence of retears and inferior clinical outcomes. This
concurs with the findings of the recent study on arthroscopic rotator cuff
repair by Boileau et
al.1.
Limitations of this study include the fact that we were not able to analyze
all of the preoperative magnetic resonance imaging scans. This could obviously
have altered our results with regard to associations with preoperative
findings. Also, whether low-field magnetic resonance imaging systems have the
same sensitivity for evaluating the rotator cuff as established high-field
magnetic resonance imaging has been
debated33-35.
We found no study comparing low-field and highfield magnetic resonance imaging
for postoperative evaluation of the rotator cuff in the literature. We believe
that the magnetic resonance images that we acquired were sufficient to
evaluate the above-mentioned three criteria, as these criteria are clearly
defined in the literature and safely reproducible. A known limitation of
measuring supraspinatus atrophy is the fact that retraction of a tear can
alter comparability among different magnetic resonance imaging scans. This
variable could not be eliminated in our study.
In conclusion, the clinical and structural results of arthroscopic repair
of isolated supraspinatus tears were comparable with the previously reported
good and excellent results of open and mini-open repair. Structural changes
such as fatty infiltration and muscle atrophy cannot be reversed by successful
arthroscopic repair. A higher degree of muscular atrophy and fatty
infiltration preoperatively as well as the patient's age are prognostic
factors for a recurrent tear. Finally, a recurrent tear leads to progression
of fatty infiltration and muscular atrophy and to an inferior clinical result.
?