The investigational review board of our institution approved this
re-review, with use of the same methodology as we used in our previous
follow-up study11,
of the results in the twenty patients in that previous study. The patients
provided informed consent to participate in the present study. The average age
at the rotator cuff repair was fifty-nine years (range, forty-seven to
seventy-one years). There were thirteen men and seven women, and seventeen
patients had involvement of the dominant arm. Initially, eight patients had
had a tear of one tendon (the supraspinatus), ten patients had had a tear of
two tendons (the supraspinatus and subscapularis in five and the supraspinatus
and infraspinatus in five), and two patients had had a tear of three tendons.
Open rotator cuff repair with transosseous tendon reinsertion had been
performed in all twenty patients, as described in detail in the previous
report11.
Clinical Assessment
The follow-up clinical assessments for this study were performed by two
examiners (M.Z. and B.J.), neither of whom had performed the operations, which
were done by another author (C.G.). Clinical assessment was carried out in the
same standardized fashion as
previously11 and
included a structured interview, a detailed physical examination, and
assignment of a Constant
score12,13.
Abduction strength was measured with an Isobex dynamometer (Cursor, Bern,
Switzerland). If 90° of abduction could not be reached, abduction strength
was automatically considered to be 0°. The total score, in points, was
also related to the age and sex-matched normal values identified by
Constant12 and the
resultant percentage was called the relative Constant
score14.
In addition, patients estimated the value of the operatively treated shoulder
as a percentage of the value of an entirely normal shoulder (which would be
considered 100%). This value was called the subjective shoulder
value.
Radiographic Assessment
Postoperative standardized radiographs (a true anteroposterior view with
the arm in neutral rotation and an axillary lateral view) were made for all
twenty patients. The acromiohumeral distance was measured in millimeters on
the anteroposterior radiograph, with use of fluoroscopic control, with the arm
in neutral rotation as standardized in the
literature15-26.
Glenohumeral osteoarthritic changes were assessed on the anteroposterior and
axillary lateral radiographs according to the classification system of
Samilson and
Prieto27.
Magnetic resonance images were made for all twenty patients with use of the
same 1.0-T scanner (Siemens, Erlangen, Germany) and with the same protocol as
had been used at the time of the 3.2-year
follow-up11.
Continuity or rerupture of the tendon was determined on coronal oblique
T2-weighted and proton-density-weighted as well as short inversion recovery
(STIR) sequences according to established magnetic resonance imaging
criteria28-30.
When a fluid-equivalent signal or nonvisualization of the supraspinatus,
infraspinatus, or subscapularis tendon was found on one or more T2-weighted or
fat-suppressed sections, the diagnosis of a full-thickness rerupture was
made30. Additional
parasagittal T1-weighted turbo spin-echo images were acquired parallel to the
glenohumeral joint for qualitative and quantitative assessment of the rotator
cuff muscles31. The
slices covered the rotator cuff from the humeral tuberosities to the medial
third of the scapula. Intramuscular fatty degeneration was assessed as
described by Goutallier et
al.3 for computed
tomography and as validated by Fuchs et
al.32 for magnetic
resonance imaging. The sizes of the tears were assessed exactly as they had
been at 3.2
years11, on the
basis of the maximal mediolateral and anteroposterior diameters, so that we
could compare the sizes of the rotator cuff defects at the earlier and later
follow-up examinations.
Magnetic resonance images and radiographs were assessed with standardized
methods that were the same as those used to evaluate the images at 3.2 years
by the same experienced musculoskeletal radiologist (C.W.A.P.), who was
blinded to the names of the patients and the clinical findings. The
radiologist evaluated the magnetic resonance images and radiographs made at
the time of the 7.6-year follow-up, then reassessed the images made at the
time of the 3.2-year follow-up, and then compared the two sets of findings.
Questionable or unexpected findings were then presented to a second
experienced musculoskeletal radiologist, who interpreted the images while
blinded to any clinical information. The two radiologists subsequently
determined the radiographic diagnosis by reaching a consensus concerning the
interpretation of the findings.
Statistical Analysis
The Mann-Whitney test was used to compare unpaired groups, and the Wilcoxon
test was used to compare paired groups. The level of significance was set at p
< 0.05. The Spearman correlation coefficient was used to test relationships
between variables.
All twenty patients were reexamined after a mean of 7.6 years (range, 6.4
to 9.8 years) postoperatively. The average age at the time of this follow-up
was sixty-six years (range, fifty-five to seventy-seven years). There were no
additional complications, and no patient had had a reoperation between the two
follow-up examinations.
Clinical Results
Subjective Values and Constant Scores
At the time of the 7.6-year follow-up, thirteen patients were very
satisfied with the result, six were satisfied, and one was not satisfied.
Compared with the findings at the time of the 3.2-year follow-up, four
patients had a higher level of satisfaction (two patients changed from
satisfied to very satisfied, and two changed from disappointed to satisfied).
Three of these four patients also had improvement in the relative Constant
score, with gains of 8% (Case 3; see Appendix), 26% (Case 8), and 28% (Case
20). The fourth patient (Case 6) had a relative Constant score of 100% at both
follow-up examinations. The Constant scores and the subjective shoulder values
for each patient are depicted in the Appendix.
The mean subjective shoulder value did not change over time: it was 74%
(range, 30% to 100%) of the value for a normal shoulder at 7.6 years and 75%
(range, 50% to 100%) at 3.2 years (p = 0.5417). The mean relative Constant
score was 88% (range, 40% to 100%) at the most recent follow-up examination
compared with 83% (range, 49% to 100%) at 3.2 years (p = 0.0785). Eleven
patients had a Constant score of =90% of a normal shoulder at 7.6 years
compared with nine patients at 3.2 years. At 7.6 years, the patients still had
a significant improvement in all clinical parameters, except for external
rotation, compared with the preoperative state and there was no significant
difference in any of the parameters compared with the findings at 3.2 years.
The Constant scores before the surgery and at 3.2 and 7.6 years are depicted
in Table I. At the most recent
follow-up examination, eight patients had a relative Constant score of 100%
and five patients were completely pain-free (a Constant pain score of 15
points) compared with seven and eight patients, respectively, at 3.2
years.
Postoperatively, the eight patients who had originally had a one-tendon
tear did not have a significantly higher mean relative Constant score than did
the ten patients who had originally had a two-tendon tear either at 7.6 years
(95% compared with 84%, p = 0.1457) or at 3.2 years (92% compared with 80%, p
= 0.1457).
Activity Status
At 3.2 years, fifteen patients had returned to their original occupation.
Two patients had to change to a less strenuous job, two patients were
receiving a disability pension, and one patient had been retired since before
the operation. At 7.6 years, six patients were retired, two patients received
a disability pension, and the remaining twelve patients were still working in
their original occupation, although two of them had had to modify their work
by excluding repetitive overhead shoulder activity. Preoperatively, eighteen
patients were severely restricted in their daily and recreational activities
because of the involved shoulder, and two patients were mildly restricted. At
3.2 years, twelve patients reported no restrictions and eight patients
reported mild restrictions. At 7.6 years, there were no substantial changes:
fourteen patients had no restrictions, five had mild restrictions, and one had
severe restrictions (Case 17, who had had mild restrictions at 3.2 years).
Imaging
Radiographs
The mean acromiohumeral distance decreased significantly (p = 0.006) from
8.4 mm (range, 4 to 12 mm) at 3.2 years to 7.4 mm (range, 4 to 12 mm) at the
time of final follow-up. A distance of <7 mm, reflecting an irreparable
tear22, was
observed in five patients at 7.6 years compared with three patients at 3.2
years. This change involved three of the ten patients who had originally had a
two-tendon tear and both patients who had originally had a three-tendon tear.
On the axillary lateral radiograph, the humeral head appeared centered in the
horizontal plane in nineteen of the twenty patients. Only the patient (Case
17) who had already had a slight posterior subluxation associated with a
traumatic supraspinatus and infraspinatus tear at 3.2 years still had mild
static posterior subluxation at the time of final follow-up.
There was no progression in the stage of the glenohumeral osteoarthritis
between 3.2 and 7.6 years (p = 0.062). At the time of the most recent
follow-up, seven of the twenty patients had stage-2 (four patients) or stage-3
(three patients) osteoarthritic changes in the glenohumeral joint. At 3.2
years, only four patients had had stage-2 osteoarthritis; three of them had
progression to stage-3 osteoarthritis at the time of the latest follow-up.
Magnetic Resonance Imaging
At the time of the most recent follow-up, a rerupture was again identified
in twelve of the twenty patients, with use of the same established magnetic
resonance imaging
criteria28-30
as had been used at 3.2 years. In eight patients (Cases 1, 4 through 7, 9, 14,
and 20), a rerupture could no longer be identified; the rotator cuff was in
continuity without any detectable full-thickness structural defect (Figs.
1-A and 1-B). The tissue
bridging the former rerupture had a signal intensity ranging from that of scar
tissue to that of normal tendon tissue, but absolute differentiation between
one and the other was not possible. The fibrous layer was very thin in four of
the eight patients (Cases 4, 5, 7, and 20). Five of the eight patients had
originally had a one-tendon supraspinatus tear; two, a two-tendon tear; and
one, a three-tendon tear.
The eight reruptures that were healed at 7.6 years had been a mean of 317
mm2 (range, 25 to 1256 mm2) at 3.2 years, and seven of
the eight had been <400 mm2. Seven of the eight patients with a
healed rerupture had originally had a tear of the supraspinatus only, and one
(Case 14) had had a rerupture of the supraspinatus and infraspinatus tendons.
At the time of surgery in Case 14, reinsertion of the entire supraspinatus to
the greater tuberosity was not possible because of tendon inelasticity, and an
anterosuperior residual defect was accepted. The structural supraspinatus
defect was confirmed with magnetic resonance imaging at 3.2 years, and the
patient was included in the study (and regarded as having had an immediate
failure). At 7.6 years, this patient had a completely intact rotator cuff as
seen on magnetic resonance imaging, with normal tendon signal at the site of
the previous defect.
At 7.6 years, eight of the twenty patients no longer had a rerupture, seven
had a smaller rerupture compared with that seen at 3.2 years, and two had a
rerupture of the same size as that seen at 3.2 years. Only three patients had
a larger rerupture than they had had at 3.2 years; two of them had originally
had a two-tendon tear, and one had originally had a three-tendon tear. The
overall average size of the reruptures did not change significantly from the
3.2 to the 7.6-year follow-up examination (p = 0.066). In the eight patients
who had originally had a one-tendon tear, the mean size of the rerupture
decreased significantly from 3.2 to 7.6 years (p = 0.007). In the ten patients
who had originally had a two-tendon tear, the rerupture size did not change
significantly over time (p = 0.312). Of the twelve patients who still had a
rerupture at the 7.6-year follow-up examination, five had a rerupture that was
larger than the original tear. Thus, one more patient had a rerupture that was
larger than the original tear compared with the number at 3.2 years.
At the time of final follow-up, six patients had an extension of the
rerupture into the infraspinatus, with a mean relative Constant score of 75%,
which was significantly lower than the score for the fourteen patients with an
intact infraspinatus (94%, p = 0.035).
Whereas fatty infiltration of the infraspinatus muscle progressed
significantly between 3.2 and 7.6 years (p = 0.015), fatty infiltration of the
supraspinatus and subscapularis muscles did not (p = 0.156 and p = 0.078,
respectively). However, compared with the findings before the surgery, fatty
degeneration of the supraspinatus (p = 0.001), infraspinatus (p = 0.0003), and
subscapularis (p = 0.003) muscles increased significantly. At 3.2 years, fatty
infiltration of the supraspinatus (p = 0.002) and the infraspinatus (p =
0.0005), but not of the subscapularis (p = 0.147), had progressed with respect
to the preoperative state (Figs. 2-A, 2-B,
and 2-C). Details regarding the locations and sizes of the tears
and fatty infiltration of the supraspinatus and infraspinatus muscles are
depicted in the Appendix.
Comparison Between Healed and Persistent Reruptures
The clinical outcomes for the eight patients with a healed rerupture at the
time of final follow-up did not differ significantly from the outcomes for
those patients at 3.2 years (mean relative Constant scores, 94% compared with
91%; p = 0.875). At 7.6 years, the mean relative Constant score of 94% (range,
82% to 100%) for the eight patients with a healed rerupture was not
significantly different from the score for the twelve patients with a
persistent rerupture (85% [range, 40% to 100%], p = 0.3516). There were also
no significant differences in any of the clinical parameters of the Constant
score between the patients with a healed rerupture and those with a persistent
rerupture.
Radiographically, there was no significant difference, with the numbers
studied, between the patients with a healed rerupture and those with a
persistent rerupture in terms of glenohumeral osteoarthritis, and the
osteoarthritis did not progress significantly in either group. The
acromiohumeral distance did not differ significantly between the healed and
persistent-rerupture groups preoperatively (10 mm compared with 10 mm, p = 1),
at 3.2 years (8.9 mm compared with 8 mm, p = 0.5359), or at 7.6 years (8.5 mm
compared with 6.6 mm, p = 0.0689). In the eight patients with a healed
rerupture, the acromiohumeral distance did not change significantly from the
preoperative status to 7.6 years (p = 0.0625), from the preoperative status to
3.2 years (p = 0.125), or from 3.2 to 7.6 years (p = 0.375). However, the
twelve patients with a persistent rerupture showed a significant decrease in
the acromiohumeral distance from the preoperative status to 7.6 years (p =
0.0156) and from 3.2 to 7.6 years (p = 0.0156), although not from the
preoperative status to 3.2 years (p = 0.0938).
When we compared the healed and persistent reruptures with magnetic
resonance imaging, we found no difference in the amount (grade) of fatty
infiltration of the different muscles preoperatively, at 3.2 years, or at 7.6
years. In the eight patients with a healed rerupture, only the infraspinatus
showed significant progression of fatty infiltration (p = 0.0031) between the
preoperative and final follow-up studies. The twelve patients with a
persistent rerupture showed progression in the supraspinatus (p = 0.0117) as
well as in the infraspinatus (p = 0.002) from the preoperative status to 7.6
years and from the preoperative status to 3.2 years (p = 0.0078 for the
supraspinatus and p = 0.008 for the infraspinatus). Details of the progression
of fatty infiltration of the different muscles are depicted in the
Appendix.
Correlation Between Preoperative and 7.6-Year Functional and Imaging
Parameters
Few and only weak correlations could be found between the preoperative and
7.6-year functional and imaging parameters (see Appendix).