Lack of elevation and external rotation in the shoulder: Reverse total shoulder arthroplasty combined with latissimus dorsi transfer to the humerus versus the greater tuberosity

Abstract

Background

The purpose of this study was to prospectively compare clinical outcomes of reverse total shoulder arthroplasty (RTSA) combined with latissimus dorsi transfer by modified L’Episcopo (Group I) versus the greater tuberosity (Group II) in patients with lack of elevation and external rotation in the shoulder.

Materials and methods

Eighteen patients participated in the study. They were placed randomly into two groups of nine patients in each group. The results were evaluated prior to surgery and at the end follow-up.

Results

The average follow-up was 33.2 months. The mean ASES scores improved from 13 points to 79 in Group I and from 15 to 73 in Group II. The mean Constant shoulder scores improved from 19 to 66 in Group I and from 18 to 67 in Group II. Mean active elevation increased from 59° to 147° in Group I and from 58° to 148° in Group II. The mean external rotation scores increased from −43° to 7° in Group I and from −41° to 23° in Group II.

Conclusion

Comparing clinical outcomes and active elevation, there were no significant differences. The latissimus dorsi transfer to the greater tuberosity provided greater external rotation than did the modified L’Episcopo transfer.

Introduction

Reverse total shoulder arthroplasty (RTSA) was first described by Paul Grammont in the mid-1980. ¹ RTSA can improve pain and restore abduction and anterior flexion of shoulders with pseudoparesis that is secondary to rotator cuff insufficiency.^2–5 In patients with large irreparable posterosuperior rotator cuff tears and loss of active external rotation, RTSA in combination with latissimus dorsi transfer (LDT) is reported to restore active external rotation.^6,7

In 1934, LDT was first described by L’Episcopo for the treatment of brachial plexopathy in children. ⁸ In 1988, Gerber et al. described LDT as a means of restoring external rotation in young patients with irreparable posterosuperior rotator cuff tears. ⁹

In 2007, Boileau et al. ⁶ and Gerber et al. ⁹ first reported an RTSA outcome combined with LDT in a patient with a large, irreparable posterosuperior cuff tears, loss of active external rotation and cuff arthropathy – Boileau with a modified L’Episcopo latissimus dorsi and teres major transfer, performed using a single deltopectoral approach and Gerber with a LDT to the greater tuberosity using a two-incision approach.

Favre et al. ¹⁰ investigated several insertion sites for the LDT combined with RTSA. They found it best to insert the tendon into the greater tuberosity at the teres minor insertion site rather than at other insertion sites.

The purpose of this study was to prospectively compare RTSA clinical outcomes combined with LDT using either modified L’Episcopo or transfer to the greater tuberosity in patients with lack of elevation and external rotation in the shoulder. We hypothesized that LDT to the greater tuberosity would provide better clinical and functional results.

Materials and methods

Between May 2015 and October 2017, 18 patients underwent RTSA combined with LDT. The inclusion criteria were painful pseudoparesis secondary to an irreparable rotator cuff lesion in the presence of a functional deltoid muscle as well as severe active external rotation deficiency, documented by a positive external rotation lag sign ¹¹ and lack of elevation with dropping sign. ¹² Eight of the 18 shoulders had undergone previous surgery, including failed rotator cuff repair (three by arthroscopic and five by open surgery). Likewise, antecedents existed in a further three patients of traumatic anterior glenohumeral dislocation.

The power analysis was performed by an independent statistician using the GRANMO program. The external rotation was used like primary variable. Accepting an alpha risk of 0.05 and beta of 0.2 in a bilateral contrast, nine patients in each group were required to detect a difference equal to or greater than 20°, assuming that the standard deviation is of 15°. This was based on the mean and standard deviation of the rotations external agreed to other authors. A loss rate of 0% was estimated.^7,13

The 18 patients meeting the inclusion criteria were assigned according to a table of random numbers to one of two groups depending on the repair technique. Of these patients, nine underwent LDT with modified L’Episcopo, fixing the transferred tendon to the posterolateral humerus (Group I) and nine underwent LDT with fixation to the posteroinferior of the greater tuberosity, at the insertion site of the teres minor (Group II). All the surgeries were performed by a single surgeon. The nature of study was explained to the patients in detail and both verbally and by written informed consent were obtained from the patients for their anonymised information to be published in this article. The study protocol and informed consent document were approved by the Local Ethical Committee (S232/2015).

Standard anteroposterior radiographs were used preoperatively to determine the acromio-humeral distance and to classify the radiographic changes, according to Hamada et al. ¹⁴ Preoperative magnetic resonance imaging (MRI) was available for all patients. The rotator cuff tears and the fatty infiltration of the rotator cuff muscles were assessed according to the Goutallier et al. ¹⁵ classification system.

There were 14 women (7 in each group) and 4 men (2 in each group). The dominant extremity was involved in 14 patients (8 in Group I and 6 in Group II). The mean age at the time of surgery was 70 years of age (range: 55–82) in Group I and 69 years of age (range: 60–78) in Group II. The mean time between symptom development and surgery was 13.6 months (range: 5–24) in Group I and 16.4 months (range: 8–36) in Group II.

Clinical evaluations were performed preoperatively and the end follow-up by a physician assistant specialized in shoulder pathology who was blinded to the surgery type and the outcome were assessed and compared between the two groups. Clinical assessment consisted of the measurement of active external rotation, internal rotation, abduction and elevation using a goniometer, the American Shoulder and Elbow Surgeons score (ASES) and Constant shoulder score. Pain was recorded using the visual analogue scale (VAS); a score of 0 indicated no pain whereas 10 points indicated the worst possible pain. Besides this, all patients were asked to subjectively describe the result of the intervention as excellent, good, fair, or poor.

Surgical technique

All the patients received preoperative antibiotic prophylaxis and underwent a preoperative interscalene block. After induction of the general anaesthesia, patients were placed in the beach-chair position.

The patients in Group I underwent a single incision using an extended deltopectoral approach. After releasing parts of the pectoralis major tendon to gain access to the LD, the tendon end was tagged with 3 thick Ethibond No. 5 (Ethicon Inc, Sommerville, NJ). Blunt dissection was used to create a channel around the posterior humeral diaphysis. The tendon-tagging was transferred at the anterolateral aspect of the humerus passing below the Teres Mayor (TM) with the help of a curved clamp. With the arm in internal rotation, three drill holes were made along the perpendicularly orientated humeral diaphysis. The same RTSA (Delta III, Depuy, Warsaw, IN) was used in all cases. The humeral component was cemented into position at 0°–10° of retroversion. After reduction, one end of each tagging suture was passed from the posterior to the anterior then tied firmly with the arm positioned at 30° of abduction and around 0° of external rotation (Figure 1). The subscapularis tendon, in cases where we find a remnant, we reinsert transosseously. The pectoralis major parts were reattached. The deltopectoral incision was closed and drainage was inserted for 48 h.

Figure 1.

(a) The LD tendon, in its original anatomical position, is located at the medial edge of the humerus next to the TM. (b) The Group I surgical technique: RTSA combined with latissimus dorsi transfer to the humerus (modified L’Episcopo). The LD is reinserted into the anterolateral zone of the humeral diaphysis. LD: latissimus dorsi; TM: teres major.

For the patients in Group II, a two-incision approach was used. A deltopectoral and another posterolateral incision were made parallel to the extreme edge of the scapula and the posterior axillary crease, with the arm held in elevation and internal rotation (Figure 2). The latissimus dorsi tendon was released from its humeral insertion and using a posteriorlateral approach, the muscle was dissected and freed; any fascial connections between the muscle bellies of the teres major and the latissimus dorsi ¹⁶ could then be freed. Two Ethibond No. 5 sutures were used to grasp the latissimus dorsi in its musculotendon union with a distal braid for better tendon fastening. A large blunt clamp was then passed through the deltopectoral interval into the plane between the teres minor and the deltoid until it exited underneath the posterior part of the deltoid. The sutures were passed through the superior incision. Two drill holes were made into the posteroinferior aspect of the greater tuberosity at the teres minor insertion site; the suture was passed through and after placing the humeral component, tied firmly (Figure 3). Two drains were inserted, one anterior and one posterior, for 48 h.

Figure 2.

Image of the second incision performed in the Group II surgical technique: With the patient in the beach-chair position and with the arm held in elevation and internal rotation, a posterolateral incision is made parallel to the extreme edge of the scapula and the posterior axillary crease. This incision allows the dissection and disinsertion of the latissimus dorsi from the humeral diaphysis, and its reorientation towards the greater tuberosity.

Figure 3.

The Group II surgical technique: RTSA combined with latissimus dorsi transfer to the greater tuberosity. The LD is redirected and passed through the deltopectoral interval into the plane between the teres minor and the deltoid. Then, it is transferred to the greater tuberosity at the teres minor insertion. LD: latissimus dorsi; TM: teres major.

All 18 patients followed the same postoperative rehabilitation protocol. A shoulder brace in neutral external rotation was applied postoperatively to all patients for five weeks. Passive gentle range-of-shoulder motion exercise commenced at two weeks for physiotherapy. Active assisted range-of-shoulder motion exercises began at five weeks up until three months.

Statistical methodology

For the descriptive study, absolute and relative frequencies were calculated for the qualitative and mean variables and the typical deviation for the quantitative variables. Confidence intervals were calculated with 95% certainty.

To compare the mean values for the quantitative variable evaluated before and after the intervention, a mixed analysis of variance (ANOVA) was carried out with the Sidak correction.

All the contrasts were considered significant when p < 0.05.

The data were collected, processed and analysed using the IBM SPSS v.25 program.

Results

Standard radiographs using the Hamada classification showed grade 1 in eight cases, grade 2 in six cases and grade 3 in four cases.

Preoperative MRI confirmed an irreparable supraspinatus and infraspinatus rotator cuff tear in all patients. A partial tear of the teres minor was noted in 2 patients and a complete tear in 16 patients. The subscapularis had a partial tear in 7 patients and a complete tear in 11 patients. Fatty infiltration of the supraspinatus was grade 3 in 5 patients and grade 4 in 13 patients, according to Goutallier et al. ¹⁵ Fatty infiltration of the infraspinatus muscle was grade 3 in 9 patients and grade 4 in 9 patients. Teres minor fatty infiltration was grade 2 in one patient, grade 3 in 11 and grade 4 in 6 patients. Subscapularis fatty infiltration was grade 1 in 2 patients, grade 2 in 7, grade 3 in 3, and grade 4 in 6 patients (Table 1).

Table 1.

Fatty infiltration of muscle in MRI scan and functional results.

Patient no.	Transfer type	Fatty infiltration Goutallier preo Sub Supr Infra Ter.				FF		ABD		ER		IR*		ASES		Constant
						Preo	Post	Preo	Post	Preo	Post	Preo	Post	Preo	Post	Preo	Post
1	Gerber	3	4	4	4	60°	160°	50°	110°	−40°	40°	8	2	13	53	24	59
2	Episcopo	2	3	3	3	80°	160°	50°	130°	−40°	0°	6	10	13	93	24	72
3	Gerber	3	4	3	3	60°	110°	40°	90°	−40°	0°	10	8	22	42	20	42
4	Episcopo	1	4	4	4	80°	160°	60°	120°	−50°	20°	2	6	2	88	12	75
5	Gerber	4	4	4	3	40°	160°	10°	150°	−40°	30°	6	2	15	95	13	85
6	Gerber	2	3	3	3	80°	160°	80°	140°	−60°	40°	8	8	18	95	18	86
7	Episcopo	4	4	4	4	30°	120°	50°	110°	−60°	0°	2	2	17	85	13	61
8	Gerber	4	4	3	3	50°	160°	60°	100°	−40°	20°	6	2	8	73	14	65
9	Episcopo	2	4	4	4	80°	110°	80°	90°	−20°	0°	8	2	25	53	26	51
10	Episcopo	4	4	4	4	30°	160°	30°	130°	−30°	10°	6	2	3	73	14	56
11	Episcopo	2	3	3	3	30°	160°	40°	160°	−60°	10°	8	8	22	80	21	74
12	Gerber	4	4	3	3	40°	120°	50°	90°	−60°	0°	8	2	13	53	17	34
13	Gerber	2	3	3	3	80°	160°	80°	110°	−40°	30°	8	8	7	70	16	72
14	Episcopo	2	3	3	3	80°	160°	60°	140°	−30°	−20°	2	2	18	75	20	72
15	Gerber	3	4	4	2	70°	140°	60°	120°	−10°	40°	8	2	18	86	17	82
16	Episcopo	2	4	3	3	40°	140°	40°	90°	−60°	0°	2	2	15	76	18	65
17	Gerber	4	4	4	4	45°	160°	40°	150°	−40°	10°	8	2	18	90	19	77
18	Episcopo	1	4	4	3	80°	150°	40°	90°	−40°	40°	8	2	2	90	19	72

Preo: preoperative; Post: postoperative; FF: forward flexion; ABD: abduction; ER: external rotation; IR*: internal rotation (10: D7, 8: T12, 6: L3, 4: sacrum, 2: gluteus); Sub: subscapular; Supr: supraspinatus; infra: infraspinatus; Ter: teres minor.

The average follow-up duration was 33.2 months (range 18–47 months); 32 months in Group I and 34 in Group II.

A statistically significant improvement between preoperative and by the end of the follow-up VAS pain scores was found in both groups (Tables 2 and 3). There was no statistically significant difference in the final results between the two techniques (Table 4).

Table 2.

Comparison of clinical outcomes based on the repair technique: Group I.

	Preoperative	Final follow-up	p-Value test
Pain score (VAS)	8.7 ± 1.4	1.4 ± 1.9	<0.0001
Elevation (°)	59 ± 25	147 ± 19	<0.0001
External rotation (°)	−43 ± 15	7 ± 17	<0.0001
Abduction (°)	50 ± 15	118 ± 25	<0.0001
Internal rotation (°)	7.3 ± 1.3	4 ± 3.2	<0.0003
ASES score	13 ± 9	79 ± 12	<0.0001
Constant score	19 ± 5	66 ± 9	<0.0001

VAS: Visual Analogic Scale; °: degrees; ASES score: American Shoulder and Elbow Surgeons score.

Table 3.

Comparison of clinical outcomes based on the repair technique: Group II.

	Preoperative	Final follow-up	p-Value test
Pain score (VAS)	8.8 ± 1	2.3 ± 2	<0.0001
Elevation (°)	58 ± 16	148 ± 20	<0.0001
External rotation (°)	−41 ± 15	23 ± 17	<0.0001
Abduction (°)	52 ± 22	118 ± 24	<0.0001
Internal rotation (°)	7.8 ± 1.2	4 ± 3	<0.003
ASES score	15 ± 5	73 ± 20	<0.0001
Constant score	18 ± 3	67 ± 19	<0.0001

VAS: Visual Analogic Scale; °: degrees; ASES score: American Shoulder and Elbow Surgeons score.

Table 4.

Comparison of clinical outcomes at the final follow-up: Group I and Group II.

	Final follow-up: Group I	Final follow-up: Group II	p-Value test
Pain score (VAS)	1.4 ± 1.8	2.3 ± 2	0.899 (NS)
Elevation (°)	147 ± 19	148 ± 20	0.956 (NS)
External rotation (°)	7 ± 17	23 ± 17	0.049
Abduction (°)	118 ± 25	118 ± 24	1 (NS)
Internal rotation (°)	4 ± 3	4 ± 3	1 (NS)
ASES score	79 ± 12	73 ± 20	0.436 (NS)
Constant score	66 ± 9	67 ± 19	0.949 (NS)

VAS: Visual Analogic Scale; °: degrees; ASES score: American Shoulder and Elbow Surgeons score; NS: no statistically significant.

Functional outcomes were significantly improved in both groups after final follow-up. The mean ASES scores increased from 13 points (range: 8–18) preoperatively to 79 (range: 68–91) at the end of the follow-up in Group I (p < 0.0001) and from 15 (range: 10–20) to 73 (range: 61–85) in Group II (p < 0.0001).

The mean constant shoulder scores increased from 19 (range: 16–21) to 66 (range: 56–77) in Group I (p < 0.0001) and from 18 (range: 15–20) to 67 (range: 57–77) in Group II (p < 0.0001). There was no significant difference in the final results between both techniques. In both groups, 15 patients (83.3 %) were much better and 3 better (16.7 %) following the surgery.

The mean active elevation increased from 59° (range: 44°–74°) to 147° (range: 133°–161°) in Group I (p < 0.0001) and from 58° (range: 43°–73°) to 148° (range: 134°–162°) in Group II (p < 0.0001). There was no significant difference between the two groups at the end of the follow-up. The mean external rotation scores increased from−43° (range: −54° to −33°) to 7° (range: −5° to 18°) in Group I (p < 0.0001) and from −41° (range: −52° to −31°) to 23° (range: 12° to 35°) in Group II (p < 0.0001). There was more external rotation following LDT to the greater tuberosity (p < 0.05).

There was only one complication: a re-operation for hematoma evacuation and a drain reinsertion.

Discussion

In patients with massive irreparable posterosuperior cuff tears and combined loss of active abduction and external rotation, RTSA combined with LDT has been described for restoring these functions.^6,7,17–20

Clinical indications are based on shoulder pain, abduction and an anterior flexion of no more than 90° as well as a positive dropping arm sign and a horn blower sign in the absence of residual teres minor function to provide external rotation.

The LD extends, internally rotates, and adducts the humerus; LD is well situated for transfer because it has the largest potential excursion. ¹⁶ Several different fixation techniques and fixation sites have been reported. Boileau et al. ⁶ reinserted the tendon to the lateral humerus, such that the function of the LD changes from internal to external rotation. All the works published using this technique associate LD and teres major transfer with different fixation types using a single incision, extended deltopectoral approach. Other authors reinserted the LD tendon into the greater tuberosity at the teres minor insertion site^7,17,19 performing both single and two-incision approaches. The biomechanical rationale for the procedure is to restore the posterior force when the infraspinatus and the teres minor are about or fatty-infiltrated. The line or pull of the LD transfer is more vertical than that of the native posterior cuff, which may explain the varied results in restoring external rotation. In a biomechanical study by Favre et al., ¹⁰ the authors experimented using several insertion sites for the LDT combined with RTSA. They found it best to insert the tendon at the teres minor insertion site on the greater tuberosity compared to other insertion sites.

The extended deltopectoral approach requires dissection between the LD and the teres major tendons and, in 10% of cases, there are muscular and fascial connections between both muscles that can make separation difficult. ¹⁶ The transfer of both tendons^6,18 strengthens the external rotation. However, the consequent myotendinous pedicle limitation of the teres major ²¹ has led some authors to recommend lateral fixation on the humerus. ²

The double approach allows better LD dissection but requires a posterior incision that risks scarring.

Regarding complications, some clinical studies have reported a complication rate of 26%.^13,17–20 The most common complication being transient nerve palsy (radial, axillary and plexus). In this study there was one complication, a hematoma that required re-operation with evacuation and drain reinsertion. The risks of anterior glenohumeral dislocation following LDT diminish when associated with RTSA; however, in cases where we find a remnant of the subscapular tendon, we reinsert it on the humeral metaphysis to minimise this complication.

The functional improvement on the Constant scale, according to Boileau et al., ¹³ with extended anterior LD and TM transfer is 35 points while for flexion and external rotation, it is 75° and 34°, respectively. For Bougheribri et al., ¹⁸ the Constant scale improvement was 47° and for flexion and external rotation was 64° and 35.7°, respectively. Both authors considered the inclusion of the TM necessary to achieve significant improvement in active external rotation.

With LD transfer to the greater tuberosity using the double approach, Puskas et al. ¹⁹ refer to a 35-point improvement in the functional Constant scale and an improvement in flexion and external rotation of 56° and 23°, respectively. Regarding LD transfer to the greater tuberosity but via a single extended anterior approach, Ortmaier et al. ¹⁷ refer to a 44-point Constant scale improvement and improved flexion and external rotation of 83° and 37°, respectively.

In Group I, with an LD transfer using a single extended anterior approach, we find an ASES scale improvement of 66 points and a Constant scale improvement of 48 points. In Group II, with LD transfer to the greater tuberosity using a double approach, there was an improvement in the ASES clinical scale of 58 points and in the Constant scale of 47 points. No significant clinical differences were found between the two groups.

In Group I, the active mobility in flexion increased to 88° and in external rotation to 50° while in Group II, the active mobility in flexion increased to 88° and in external rotation to 64°. No clinical differences were found between the two groups in active flexion mobility. However, we did find a significant difference in the external rotation in favour of LD transfer to the greater tuberosity.

From our prospective work on patients with lack of elevation and external rotation in the shoulder performing RTSA combined with LDT, we can partially confirm our working hypothesis given that, although both techniques provided similar clinical results in the functional ASES and Constant scales, and active elevation mobility, the LD transferred to the greater tuberosity provided more external rotation.

The limitations of this study were the small number of patients and the short-term follow-up. On the other hand, one of its strengths is the clinical outcomes of two surgical transfer techniques used to treat patients with large irreparable posterosuperior rotator cuff tears and loss of active external rotation. RTSA in combination with LDT is reported in randomized patients who were compared prospectively.

Conclusion

In patients with lack of elevation and external rotation in the shoulder, RTSA combined with LDT improved the clinical outcome in terms of the ASES and Constant scales for both techniques at 12 months. No statistically significant differences existed in the final results between the two techniques. The active mobility in elevation and external rotation improved with both groups. There were no significant differences in active elevation. LD transferred to the greater tuberosity provided better external rotation compared to the modified L’Episcopo transfer.