Keywords rotator cuff injuries - tendon transfer - shoulder - orthopedic procedures
Introduction
Chronic irreparable rotator cuff tendon tears lead to atrophy, fatty infiltration,
and functional impairment of their respective muscle bellies.[1 ] When they affect the posterosuperior rotator cuff (supra and infraspinatus tendons),
they mostly result in pain, reduction of shoulder strength for external rotation and
forward flexion, and difficulty in positioning the hand in space.[1 ]
[2 ]
Due to the irreparability of these tears, several surgical techniques were described
to improve shoulder function. Reverse total shoulder arthroplasty has been shown to
provide good and predictable functional outcomes at short and medium-term follow-up.[3 ] Nevertheless, two important concerns suggest that it should be reserved for older
patient populations (65 years or older): (1) there is evidence showing that after
10 years postoperatively, these good functional results deteriorate and the complication
rates increase significantly[4 ]; (2) reverse total shoulder arthroplasty has been widely used for a relatively short
period (since 2004 in the USA[5 ]) and, therefore, there is little evidence of its long-term results.
Because of these reasons, other techniques are recommended for younger and more active
patients.[6 ] The three most commonly used techniques are: (1) latissimus dorsi tendon (LDT) transfer
to the greater tuberosity,[1 ]
[2 ]
[7 ]
[8 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ] (2) superior capsular reconstruction,[14 ]
[15 ] and (3) lower trapezius tendon transfer to the greater tuberosity.[16 ]
[17 ]
Despite some publications regarding the outcomes[14 ]
[15 ] in the last few years for superior capsular reconstruction, its medium and long-term
results are still unknown. Similarly, the works concerning trapezius tendon transfers
outcomes are still very few and recent.[16 ]
[17 ]
Conversely, LDT transfer was initially described by Gerber et al.[1 ] in 1992, and there are several publications about its results, including long-term
ones.[2 ]
[11 ]
[18 ] Among them, several authors have demonstrated the possibility of obtaining good
outcomes.[2 ]
[9 ]
[10 ]
[11 ]
[19 ]
[20 ] However, up to 36% of clinical failures (as well as 38% of transfer tears) were
found.[21 ] In addition to that, findings about the predictive factors of bad results are still
scarce and controversial.[2 ]
[4 ]
[22 ]
[23 ]
There is, however, evidence that most failures are due to one of the following reasons:
(1) transfer rupture from the greater tuberosity,[12 ]
[13 ] and (2) deltoid origin disruption.[24 ] As such, some authors have recently proposed changes to the original Gerber technique.
The use of arthroscopy[8 ]
[12 ]
[21 ] or a single approach,[9 ] for instance, has been recommended to prevent deltoid violation. Latissimus dorsi
detachment along with some bone fragments of the humeral cortex has been recommended
to reinforce the healing of the transfer along the greater tuberosity.[13 ]
Nonetheless, the current literature lacks a technique that simultaneously prevents
these two complications. With this in mind, we have proposed and recently published[25 ] the following modifications to the surgical technique: the LDT is elongated and
reinforced with a tendinous allograft, enabling its transfer to be performed through
a single deltopectoral approach.
The primary objective of the present paper is to assess, through a retrospective analysis
of cases, the short-term functional results of shoulders that have been submitted
to this technique. The secondary objective is to analyze possible pre and intraoperative
variables that may influence the outcomes.
Materials and Methods
The surgical technique used, and the rehabilitation protocol followed were the same
as those of the technique previously published.[25 ] It is summarized in [Figs. 1 ]
[2 ] through [3 ].
Fig. 1 Illustration of a right shoulder (anterior view) depicting the anatomical structures
(the pectoralis major has been tenotomized and retracted; the latissimus dorsi tendon
humerus insertion has already been detached).
Fig. 2 (A ) Illustration of a right shoulder (anterior view), showing the surgical step in which
the tendinous allograft has already been sutured to the native latissimus dorsi tendon
and a surgical tweezer has been passed posterior to the humeral shaft. (B ) Intraoperative photograph of a right shoulder during this surgical step.
Fig. 3 (A ) Illustration (anterior view) and (B ) photograph (lateral view) showing the final aspect of surgery, where the elongated
latissimus dorsi tendon transfer is sutured to the posterolateral aspect of the greater
tuberosity.
Between November 2013 and August 2018, we performed the aforementioned technique[25 ] in 20 patients with irreparable posterosuperior rotator cuff tears (RCTs). Their
medical records were retrospectively evaluated for data related to epidemiological
as well as pre and postoperative clinical findings.
The surgery was proposed to patients after failure of conservative treatment with
physiotherapy, and only after explaining that there would be no expectation of full
functional recovery.
Diagnosis was performed by MRI and confirmed intraoperatively. Fatty infiltration
of subscapularis, supraspinatus, and infraspinatus were classified according to the
Goutallier Classification, modified by Fuchs et al.[26 ] Likewise, preoperative radiographs were used for the Hamada classification.[27 ]
Irreparability of tendons was determined by intraoperative impossibility of its reduction
to the greater tuberosity with the shoulder at 45° of abduction, after proper tendon
release.[6 ] In every patient in this study, no partial repair was possible due to tendon retraction.
Both pre and postoperatively active and passive range of motions (ROMs) were registered
according to the method proposed by the American Academy of Orthopedic Surgeons.[28 ] Pre and postoperative functions were evaluated according to the University of California,
Los Angeles (UCLA)'s score, as modified by Ellman et al.[29 ]
The inclusion criteria were adult patients with irreparable RCTs treated with this
technique, and with a minimum postoperative follow-up of 1 year. The exclusion criteria
were no consent to participate in the study, noncompliance to rehabilitation protocol,
and previous shoulder surgery.
Out of the 20 operated patients, 16 met the inclusion criteria (three exclusions occurred
due to short follow-up, and one due to noncompliance to rehabilitation protocol [suffered
a stroke 3 months after surgery]). Six (37.5%) were male and 10 (62.5%) were female.
The average age at surgery was 65 years (56 to 82 years). Thirteen were dominant shoulders
(81%). The etiology was traumatic in 8 patients (50%) and degenerative in 8 (50%)
([Table 1 ]).
Table 1
Goutallier
Case #
Gender
Age
Dominance
Trauma?
Symptoms (in months)
SUPRA
INFRA
SUBSCAP
HAMADA
Graft
Follow-up (in months)
1
M
77
+
−
36
3
3
1
1
Patellar
13
2
M
56
−
−
24
2
3
1
2
Fascia Lata
13
3
M
61
−
+
25
3
4
0
2
Patellar
21
4
F
52
+
+
10
3
4
1
1
Patellar
31
5
F
63
+
+
24
4
4
0
1
Patellar
43
6
F
68
+
−
36
4
4
4
1
Fascia Lata
12
7
F
64
+
−
96
4
4
4
3
Patellar
47
8
F
68
+
+
52
2
2
2
4B
Quadricipital
19
9
M
68
+
+
7
3
3
0
1
Fascia Lata
15
10
F
59
+
−
17
2
3
0
1
Fascia Lata
15
11
F
67
+
−
36
3
3
1
2
Fascia Lata
16
12
F
56
+
−
24
4
4
0
1
Fascia Lata
14
13
F
61
+
+
23
3
3
2
1
Patellar
28
14
F
82
+
+
24
4
4
2
1
Patellar
19
15
M
67
−
+
7
3
3
1
1
Fascia Lata
12
16
M
67
+
+
72
3
3
1
1
Fascia Lata
15
AVERAGE
65
32
21
The average period from symptoms until surgery was 32 months (7 to 96 months), with
a follow-up period of at least 12 months and no longer than 47 (average of 21 months)
([Table 1 ]).
Supraspinatus Goutallier classifications were: three cases as grade II, eight as grade
III, and eight as grade IV ([Table 1 ]).
There were two subscapularis tears, and they were all limited to the proximal half
of their respective tendons. They were all repairable tears. The proximal half of
the muscle bellies of these two cases were classified as Goutallier grade IV (the
distal halves were grade 0) ([Table 1 ]).
Eleven shoulders were classified as Hamada grade 1, 3 as grade 2, 1 as grade 3, and
1 as grade 4B ([Table 1 ]).
Eight of the allografts used were patellar tendons, six were fasciae latae, and one
was a quadricipital tendon. All of them were frozen allografts ([Table 1 ]).
The average preoperative results were compared with the average postoperative results
([Table 2 ]). Furthermore, patients were split into subgroups for average results comparison
concerning the following variables:
Age at the time of surgery
Preoperative active forward flexion (six pseudoparetic [from zero to 90°] versus 10
nonpseudoparetic [higher than 90°]) ([Table 3 ]).
Time span between 1st symptoms and surgery (10 shorter than 26 months versus 6 longer
than 35 months)
Type of allograft used (eight fascia latae versus seven patellar tendon).
Etiology of RCT (atraumatic versus traumatic).
Postoperative follow-up period (9 shorter than 17 months versus 7 longer 18 months).
Table 2
Case #
ROM
UCLA
FLEX PRE
FLEX POST
EXT ROT PRE
EXT ROT POST
INT ROT PRE
INT ROT POST
PAIN PRE
PAIN POST
FUNCTION PRE
FUNCTION POST
FLEX PRE
FLEX POST
STRENGTH PRE
STRENGTH POST
SATISF PRE
SATISF POST
TOTAL PRE
TOTAL POST
1
70
160
45
70
Gluteus
T12
4
8
4
8
3
5
3
3
0
5
14
29
2
70
150
0
45
Gluteus
T3
2
8
2
8
2
5
2
3
0
5
8
29
3
120
160
0
50
T9
T5
6
8
2
4
3
4
2
3
0
5
13
24
4
140
150
50
60
T10
T12
2
6
2
6
4
5
3
4
0
5
11
26
5
120
160
40
45
T10
T8
4
8
4
8
4
5
3
4
0
5
15
30
6
120
140
45
45
T12
L1
4
8
4
6
4
4
3
3
0
5
15
27
7
70
160
45
60
T12
T10
1
5
4
8
2
5
2
3
0
5
9
26
8
130
140
30
60
Gluteus
T7
2
8
2
8
4
5
3
4
0
5
11
30
9
120
140
20
50
T12
T12
1
10
1
8
3
3
2
4
0
5
7
30
10
80
160
60
60
T12
T12
4
8
2
8
1
5
2
4
0
5
9
30
11
130
130
45
45
T10
T10
2
2
8
8
3
3
3
4
0
0
16
17
12
140
140
50
60
T7
T7
2
10
8
8
3
3
3
4
0
5
16
30
13
125
140
30
70
T9
T10
2
10
4
8
3
5
2
4
0
5
11
29
14
90
130
0
60
L5
T12
4
6
2
6
1
3
1
3
0
5
8
23
15
100
130
20
40
T12
T9
4
8
4
8
3
4
2
4
0
5
13
29
16
70
130
0
40
Gluteus
T12
4
8
2
8
1
4
3
4
0
5
10
29
AVERAGE
105.94
145
30
53.75
L1
T10
3.00
7.56
3.44
7.38
2.75
4.25
2.44
3.63
0.00
4.69
11.63
27.38
p
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
Table 3
Average gain of
PSEUDOPARETIC (6 cases)
NONPSEUDOPARETIC (10 cases)
p
ROM - FLEX
73.33°
18.5°
< 0.001 ***
ROM - EXT ROT
30.83°
19.5°
0.27
ROM - INT ROT
6.17°
1.7°
0.09
UCLA - PAIN
4
4.9
0.48
UCLA - FUNCTION
5
3.3
0.12
UCLA - FLEX
2.83
0.7
< 0.001 ***
UCLA - STRENGTH
1.17
1.2
0.925
UCLA - SATISF
5
4.5
0.458
UCLA - TOTAL
18
14.4
0.183
ROM - FLEX PREOP
75°
124.5°
< 0.001 ***
ROM - FLEX POSTOP
148.33°
143°
0.41
Comparisons of continuous variables within the same group were done through paired
t -tests. Comparisons of continuous variables between different groups were done through
unpaired t -tests. Correlations between continuous variables were determined by calculating the
Pearson correlation coefficient. Values of p ≤ 0.05 were considered significant.
Results
The average postoperative follow-up was 21 months (12–47 months) ([Table 1 ]). The passive range of motion (ROM) in all shoulders was normal, equal to the contralateral
shoulder, including case 8 ([Table 1 ]), classified as Hamada 4B.
The average active forward flexion (FF) was 106° preoperatively (70–104°) and 145°
postoperatively (50–160°) (p < 0.001). The average preoperative active external rotation was 30° (0–60° and postoperative
average was 54° (40–70°) (p < 0.001). The active internal rotation averaged at L1 preoperatively (range gluteus
to T5), and at T10 postoperatively (p = 0.017) ([Table 2 ]).
According to the UCLA's score, there were 71% (12 patients) of good results (≥ 28
points) and 29% (5 patients) of fair and poor results (≤ 27 points). The average preoperative
score was 11.63 (range 8 to 16) while the postoperative average was 27.38 (range 17–30)
(p < 0.001). The comparisons of pre and postoperative average results for each of the
UCLA subscores were statistically significant in all cases (for all five subscores
p < 0.001) ([Table 2 ]).
Comparing the results from subgroups, we found the following:
- For age at the time of surgery, time span between first symptoms and surgery, type
of allograft used, etiology of RCT, and postoperative follow-up period, the comparisons
of average gains (postoperative results subtracted from preoperative results) from
all variables related to the active ROM and the UCLA score did not result in any statistically
significant difference.
Those with preoperative pseudoparesis (average forward flexion [FF] at 75°, ranging
from 70–90°) and 10 without (average FF at 124.5°, ranging from 100–140°) (p < 0.001). The comparisons of both groups' average gains (postoperative results subtracted
from preoperative results) did not result in any statistically significant difference
concerning external rotation, internal rotation, or the following UCLA score criteria:
pain, function, strength, satisfaction, and total score. However, there was a statistically
significant difference concerning the FF range improvement, which was 73.33 degrees
for pseudoparetic patients and 18.5 degrees for nonpseudoparetic (p < 0.001), as well as in the UCLA's forward flexion score, which was 2.83 points for
pseudoparetic patients and 0.7 for nonpseudoparetic patients (p < 0.001). As shown in the last line of [Table 3 ], this difference in FF gain was able to promote statistically equivalent results
in both groups, which was 148.33 degrees for pseudoparetic patients and 143 degrees
for nonpseudoparetic patients (p = 0.41), therefore showing that pseudoparesis was not a predictor of worse outcome
([Table 3 ]).
Discussion
The main objectives of LDT transfer for treating irreparable RCTs are improvements
in shoulder's active ROM, pain, function, and strength, as well as improvement in
patient satisfaction. As shown in the results, all of these objectives have been met.
The improvements obtained in active external rotation, internal rotation, and FF ROM
were similar to those of other published works,[1 ]
[2 ]
[9 ]
[10 ]
[11 ]
[19 ]
[20 ] with average gains of 39 degrees (p < 0.001), 3 vertebral levels (p = 0.017) and 24 degrees (p < 0.001), respectively ([Table 2 ]).
This technique seems effective in improving pain relief, considering the average gain
of 4.53 points (p < 0.001) to the UCLA score criterion for pain (which ranges from 0–10), besides also
being efficient to improve shoulder function, as we can see with the average gain
of 3.71 points (p < 0.001) to the UCLA score criterion for function (which ranges from 0–10). Both
findings were similar to those of most other works,[1 ]
[2 ]
[9 ]
[10 ]
[11 ]
[19 ]
[20 ] such as Grimberg et al.,[8 ] which reported an average gain of 28.4 points (from 37.0–65.4) to the overall Constant
score, as well as an average gain of 10.9 points (from 1.7 preop to 12.6 postop) to
pain as a Constant subscore.
The improvements to FF strength occurred in all patients, except in the first two
cases, in whom strength remained the same. However, in no case did it return to normal
(grade 5), remaining at grade 3 in 6 patients, and at degree 4 in 5 patients. One
concern is that the elongation of the transferred tendon could harm the shoulder's
FF strength. However, analysis of the literature shows that the strength recovery
was very similar to that of other series (no graft used),[2 ]
[8 ]
[9 ]
[10 ]
[11 ]
[19 ]
[30 ] such as Gerber et al.,[30 ] which obtained mean strength improvement from 0.9 kg preoperatively to 1.8 kg postoperatively
(p < 0.0001) in the 65 shoulders that were evaluated.
Surprisingly, and contrary to the findings of Iannotti et al.[10 ] (conventional technique), pseudoparesis was resolved in all cases. We believe this
occurred because the tendon grafts used were thicker than the native LDT and, therefore,
could also have worked as a subacromial spacer. Thus, it would be easier to maintain
stability to the shoulder center of rotation, favoring the deltoid's FF function.
According to the UCLA score, we have obtained 71% of good results and 29% of fair
and poor results. Such findings are similar to those in the literature, where around
30% of unsatisfactory results are found, such as the findings of Kany et al.[21 ] Despite these unsatisfactory results, we believe these patients will have fewer
complications at medium and long terms, since this technique is possibly less prone
to new ruptures, as Kany et al.[12 ] have shown from 15 to 46% of ruptures in arthroscopic LDTTs at 3 months of postoperative
follow-up.
Reverse shoulder arthroplasty (RSA) is arguably the last surgical resource after failure
of a tendon transfer. However, for an RSA to work, a functioning deltoid is required,
because it is the last motor unit available capable of generating upward force vectors
that would result in FF.[2 ] Thus, we can argue that it is extremely wise to avoid any deltoid surgical violation,
as it is performed in superior approaches. Therefore, only three deltoid-sparing surgical
approaches are available: a single posterior approach; an axillary approach followed
by an arthroscopic one; and a single anterior approach.
The major shortcoming of a single posterior approach is that it precludes treatment
of lesions to the long head of the biceps and to the subscapularis tendon.
Similarly, an axillary approach does not allow an eventual primary repair of the RCT,
which should always be attempted before performing any tendon transfer.
The advantages of a single deltopectoral approach are that it is a customary approach;
it allows primary repair of the RCT when possible; and that long head of the biceps
and subscapularis lesions can be easily treated. Its only disadvantage is the impossibility
of dissecting the latissimus dorsi muscle belly, therefore preventing it from being
pulled to the upper portion of the humeral head. However, given that this tendon is
extremely thin and friable, it is recommended that its insertion to the greater tuberosity
be somehow reinforced, since its postoperative rupture, which usually occurs in the
first 2 years of postoperative follow-up, is one of the few statistically proven predictors
of poor functional results.[2 ] Therefore, we believe that the use of a tendon graft may be able to simultaneously
solve both problems: it allows the use of a single (deltoid-sparing) approach and
it reinforces the insertion of the tendon to the greater tuberosity.
There are concerns about three possible complications: (1) fragility between the graft
and the native tendon; (2) the allograft's ability to heal to the bone; and (3) the
possibility of the LDT being too elongated, overly reducing the resulting final strength
on the humerus. Although the present work hasn't directly evaluated these three possible
complications, they did not seem to occur, since all patients have shown improvements
in the UCLA strength score, as previously discussed.
As a secondary objective, we have assessed whether some preoperative (age, pseudoparesis,
symptom duration, and etiology of RCT), intraoperative (the type of allograft used),
and postoperative (“follow-up period”) variables have influenced the results. Despite
the limited number of patients, the results have shown that none of them have significantly
influenced the outcomes.
It is worth pointing out cases 6 and 7, which had complete tears of the proximal subscapularis
half. Both had their tears repaired (at the same time of the transfer), and both had
noticeable improvement of shoulder motion and UCLA score ([Tables 1 ] and [2 ]). Likewise, cases 7 and 8 had advanced joint degeneration (Hamada 3 and 4B, respectively),
and they also were able to recover shoulder ROM and achieve noticeable improvement
of UCLA score ([Tables 1 ] and [2 ]).
Lastly, it is worth highlighting the flaws of this work: as a retrospective analysis
of cases, there was no control group and, therefore, all of the comparisons had to
be done with preoperative findings or findings from other published works that have
used different methodologies. Also, even though the sample loss was of only 20% (or
4 cases), the total number of evaluated cases was small, compromising the power of
statistical analysis.
Conclusion
At short and medium follow-ups, the studied technique was effective as a treatment
for irreparable posterosuperior RCTs, with 71% of satisfactory results. None of the
evaluated variables, including preoperative pseudoparesis, has influenced the final
outcomes.
