Knee Surgery, Sports Traumatology, Arthroscopy

, Volume 15, Issue 8, pp 1055–1061

Revisiting open capsuloplasty for the treatment of anterior shoulder instability: 35-year follow-up of the Du Toit procedure

Authors

  • Stefano Zaffagnini
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Leonardo Marchesini Reggiani
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Francesco Iacono
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Giuseppe Filardo
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Marco Delcogliano
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Andrea Visani
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
  • Maurilio Marcacci
    • Department of Orthopaedic and Sport TraumatologyRizzoli Orthopaedic Institute, Biomechanics Laboratory
Shoulder

DOI: 10.1007/s00167-007-0303-2

Cite this article as:
Zaffagnini, S., Russo, A., Reggiani, L.M. et al. Knee Surg Sports Traumatol Arthr (2007) 15: 1055. doi:10.1007/s00167-007-0303-2

Abstract

The Du Toit open capsuloplasty for the treatment of anterior shoulder instability is based on the concept of restoring joint stability by recreating the integrity of the anterior glenoid labrum and inferior gleno-humeral ligament using staples. The long-term validity of this procedure for the treatment of anterior shoulder instability was retrospectively assessed by a clinical or telephone interview and radiographic evaluation in 58 patients with an average 35-year follow-up between 1948 and 1974. The range of motion was evaluated by comparing the treated side with the contralateral one; subjective and objective evaluation was performed according to the ASES, Rowe scales and Constant rating system; A-P and axillary X-rays were performed to evaluate glenohumeral arthrosis by the Samilson criteria. Despite the need for a second operation due to staple loosening (5 of 58 patients in our series), this open procedure for shoulder instability gave a high rate of satisfactory results, thus holding the ground for the current concept of the modern arthroscopic Bankart repair.

Keywords

ShoulderCapsuloplastyInstabilityDu Toit procedure

Introduction

Several surgical options have been introduced over the past 50 years for the treatment of recurrent shoulder instability in patients who do not respond to non-operative treatment. The surgical gold standard is universally considered to be shoulder stabilization to avoid recurrence without over-tightening the reconstructed capsulolabral ligaments to avoid loss of motion and degenerative arthrosis. Numerous methodologies have been implemented with the aim of correcting the pathophysiology of shoulder instability. Some techniques aim to limit external rotation, considered to be responsible for shoulder luxation [3, 13, 19, 28], others aim to avoid luxation by increasing the anterior glenoid surface by bone transplant, such as the Latarjet or Eden–Hibbinette procedure [1, 912, 20, 26] and creating an anterior bone stock; others identify the capsulolabral detachment as the pathologic element of shoulder instability and its correction as the most effective treatment [7, 14, 18, 22, 23].

A precursor of the Bankart repair, uncommon today, was the Du Toit [2, 5] procedure, a technically demanding procedure based on restoring the integrity of the anterior glenoid labrum using staples.

The aim of this study was to evaluate the clinical and radiological outcome of the Du Toit procedure for anterior shoulder instability at a very long-term (35 years) follow-up in order to establish the validity of this procedure in restoring joint stability without impairing the range of motion (ROM) and to evaluate the morbidity of such a procedure at unusual follow-up.

Materials and methods

Sixty-five Du Toit procedures were performed between October 1948 and July 1974 at our institution. Forty-five patients returned for clinical and radiological evaluation, 13 patients were available for a telephone interview, with an average follow-up of 35 years; seven patients were lost to follow-up or died. Office papers and intervention descriptions were reviewed for each patient.

Forty-six patients (79%) were men and 19 (21%) were women; 47 shoulders (81%) were on the dominant side, 11 shoulders (19%) were on the non-dominant side. No patient was treated bilaterally. The mean age at the time of intervention was 24 years (minimum 17 and maximum 40 years), and the mean age at follow-up was 60 years (range 49–79 years). Mean follow-up was 35 years (range 31–57). All data are summarized in Table 1.
Table 1

Clinical data

Patient

Age

Age at surgery

Follow-up (years)

Age of first dislocation

Number of dislocations

Time to surgery (years)

1

54

20

34

18

1

2

2

53

17

36

13

2

4

3

61

28

33

22

2

6

4

62

31

31

15

2

16

5

58

23

35

21

2

2

6

79

38

41

30

1

8

7

67

17

50

16

2

1

8

66

27

39

20

1

7

9

55

19

36

16

1

3

10

62

29

33

26

1

3

11

52

18

34

16

3

2

12

60

22

38

19

1

3

13

78

24

54

23

2

1

14

67

29

38

24

2

5

15

74

17

57

14

3

3

16

52

20

32

15

3

5

17

54

17

37

17

2

0

18

59

19

40

16

3

3

19

67

35

32

31

2

4

20

74

37

37

28

1

9

21

60

23

37

21

3

2

22

52

20

32

16

2

4

23

63

25

38

20

2

5

24

62

27

35

22

3

5

25

57

18

39

11

1

7

26

51

19

32

19

2

0

27

65

22

43

19

2

3

28

55

20

35

15

2

5

29

50

19

31

13

2

6

30

61

28

33

15

3

13

31

79

40

39

24

1

16

32

56

25

31

24

2

1

33

74

35

39

28

2

7

34

56

25

31

20

1

5

35

52

20

32

18

2

2

36

53

21

32

14

2

7

37

60

23

37

21

3

2

38

53

21

32

19

3

2

39

61

30

31

27

2

3

40

60

23

37

21

3

2

41

54

22

32

22

1

0

42

55

21

34

19

2

2

43

67

36

31

31

2

5

44

56

22

34

20

2

2

45

49

18

31

15

2

3

46

54

23

31

20

2

3

47

54

22

32

19

2

3

48

58

23

37

18

2

5

49

61

25

36

16

1

9

50

69

27

42

27

1

0

51

53

21

32

19

1

2

52

54

23

31

21

2

2

53

55

23

32

20

2

3

54

59

26

33

24

1

2

55

55

24

31

15

1

9

56

62

30

32

23

1

7

57

58

27

31

24

3

3

58

52

20

32

17

2

3

Mean

60

24

35

20

2

4

Number of dislocations. 1: from 1 to 3 dislocations; 2: from 4 to 7 dislocations, 3: more than 8 dislocations

Fifty-two patients reported a traumatic dislocation as a consequence of sports activities or a fall, six patients had atraumatic dislocation. The patients were defined as having had dislocations if there had been radiographic documentation of an actual dislocation, or a visible deformity of the shoulder that had led to manipulation for reduction. Operative findings revealed a detachment of the labrum from the glenoid rim and a Bankart lesion in 51 shoulders, and capsule loosening in seven shoulders. No signs of degenerative changes were present on the pre-operative X-rays, and no signs of rotator cuff disease were present at the clinical examination before the intervention.

Surgical technique

All patients underwent the same procedure with the Du Toit technique [13]. Through a deltopectoral approach with the shoulder externally rotated, the subscaplaris tendon was incised transversely in the direction of the fibres at the junction of its middle and lower third. The capsule was identified and also incised transversely. After an accurate insertion of the retractor behind the posterior lip of the glenoid, the Bankart lesion could be easily identified by passing the tip of a small haemostat under the glenoid labrum and around the glenoid rim. If the haemostat passed freely around the glenoid rim, the labrum was detached from the anterior rim of the glenoid.

In our series the Bankart lesion was present in 51 patients. Prior to stapling, the anterior rim of the glenoid was roughened with a curved curette. The staples were then inserted 0.5 cm medially to the anterior rim of the glenoid, parallel to the articular surface of the glenoid. The ideal position for the staples was close to the inferior lip of the glenoid to secure the inferior glenohumeral ligament (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-007-0303-2/MediaObjects/167_2007_303_Fig1_HTML.jpg
Fig. 1

X-ray showing the correct position for the staples

After a satisfactory insertion of the staples, the capsule and the subscapularis tendon were closed anatomically with multiple interrupted sutures without imbrications of capsule and subscapolaris tendon, which can potentially create an excessively tight repair [8, 13]. All the patients had a thoracobrachial cast for 4 weeks after surgery with the shoulder fixed at 70° abduction and 10° external rotation. The rehabilitation protocol had a mean duration of 3 months.

Follow-up evaluation

At follow-up all the patients were subjectively and objectively evaluated by the American Shoulder and Elbow Rating Scale [14] in order to assess pain, function, stability and strength. Active and passive ROM of the treated shoulder was compared with the contralateral side. Stability was assessed by the drawer test and Hawkins grading system [6, 8]. We also evaluated the clinical outcome of the patients with the Rowe [13] rating system and Constant [27] rating system to allow comparison with other reports [4, 21]. On antero–posterior and anxillary X-rays we evaluated degenerative changes of the glenohumeral joint according to the Samilson criteria [24].

Statistical analysis was performed to correlate the effect of age of the first luxation, the number of luxations, and the time from the first luxation to surgery on the clinical and radiological result at long-term follow-up. Moreover, we wanted to test whether there was a correlation between limitation in range of movement and incidence of degenerative arthritis at long-term follow-up.

We also evaluated the correlation between ROM and the scales used for evaluation. StatView 5.0™ (SAS Institute Inc., Cary, NC, USA) was used to perform the statistical analysis. Mean and standard deviation was used as descriptive statistics. Non-parametric methods were used to determine the correlations between parameters (Spearman’s correlation).

Results

Two post-operative dislocations occurred (one anterior and one posterior); they underwent revision of the capsuloplasty: the anterior one with the Bristow–Latarjet technique and the posterior one with capsulorrhaphy plus bone graft, because of insufficient anterior and posterior bone stock. Subjective assessment revealed that 50 of 58 patients (86%) were very satisfied with the treated shoulder and reported normal function and no shoulder pain during daily life activities at the time of follow-up. Two patients (4%) reported pain only after unusual activities, while six patients (10%) reported pain and discomfort during daily life activities.

Low number of dislocations before surgery and short time between first dislocation and surgery were positively correlated with better results at long-term follow-up (P < 0.05). Five of the six patients (8.6%) who reported pain and discomfort during daily life activities were treated surgically a second time, from 1 to 4 years after the primary intervention, due to staple loosening (Fig. 2), and the other patient underwent two operations in order to revise the scar and drain the joint due to deep infection. This patient had grade III arthrosis and severely limited function and pain at follow-up (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-007-0303-2/MediaObjects/167_2007_303_Fig2_HTML.jpg
Fig. 2

X-ray showing the staples loosened

https://static-content.springer.com/image/art%3A10.1007%2Fs00167-007-0303-2/MediaObjects/167_2007_303_Fig3_HTML.jpg
Fig. 3

X-ray showing the grade III arthrosis of the gleno-humeral joint in the patient who underwent staple removal due to the loosening of the staples

This technique was successful at restoring joint stability without impairing the ROM: over 90% of patients presented normal or nearly normal ROM compared with the contra lateral side. As to the active ROM, in the 45 patients physically examined, forward elevation was symmetric with the opposite side in 39 patients (87%); five cases (11%) presented a 5° reduction, and one patient (2%) had a 15° reduction. External rotation with the arm at the side was normal in 28 patients (62%), slightly reduced (5–10°) in 15 patients (33%) and reduced (>15°) in two patients (5%). External rotation with the arm 90° abducted was normal or slightly reduced (<10°) in 40 patients (89%), and significantly reduced (>20°) in five patients (11%).

The internal rotation was evaluated with the thoracic level, which was equal to the contralateral side in 32 patients (T6–T8) (71%), slightly reduced in ten patients (T8–T10) (22%), and severely reduced in three patients (S1) (7%). The complete ROM (mean and standard deviation) is reported in Table 2. Three patients presented posterior subluxation with good functional results.
Table 2

Range of motion (mean ± SD)

Active motion against gravity

Unaffected shoulder

Surgically treated shoulder

Forward elevation

180 ± 6

179 ± 3

External rotation with arm at side

44 ± 7

41 ± 7

External rotation with arm in 90 abducted position

89 ± 7

82 ± 14

Internal rotation

44 ± 8

41 ± 7

According to the Rowe scoring system, 39 patients (86.6%) were satisfied or very satisfied with the Du Toit procedure, while six patients (13.3%) had fair results at 35-year follow-up. A mean score of 95 (range 65–100) on the Rowe scale was achieved.

Strength was evaluated according to the ASES scoring system by comparing the treated shoulder with the contra lateral one: in 33 of 45 patients they were equal, six patients had a mild reduction in forward elevation and abduction and six had a severe reduction in abduction and external rotation (five of them underwent staple removal and one needed two revisions of the scar due to deep infection).

The mean Constant score was 86.2 in the 45 patients examined; six patients had poor results. All data are summarized in Fig. 4.
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-007-0303-2/MediaObjects/167_2007_303_Fig4_HTML.gif
Fig. 4

Graphic reporting mean, min and max values of the self-evaluation, Rowe, Constant scales used for clinical evaluation

According to the Samilson criteria, the X-rays of the 45 patients reviewed revealed that 18 had no arthrosis, 22 had asymptomatic grade I arthrosis with no pain and no function limitation, four had grade II arthrosis correlated with mild limitation of external rotation (<15°) and pain after unusual activities and one patient had grade III arthrosis, pain and discomfort during daily life activities (Fig. 3). Our results showed a correlation between ROM limits and the incidence of degenerative arthritis (tied P < 0.05).

We asked patients retrospectively about their post-surgical ability to achieve their previous level of activity. Fifty-three of 58 patients reported a return to pre-surgical activity in terms of work and sports.

Spearman’s rank correlation showed a statistically significant correlation between ROM (forward elevation, external and internal rotation) of the treated shoulder and the scales used for the evaluation (Rowe, Constant and Self Evaluation scale, all P < 0.01). All the evaluation scales used showed a high correlation in the analysis of the results (Spearman’s correlation all P < 0.0001).

Discussion

Successful treatment for shoulder instability has to fulfil three main conditions: recovering of full ROM, avoiding recurrence and avoiding late degenerative arthrosis.

To authors’ knowledge this is the longest follow-up study following open shoulder stabilization. The limitations of this retrospective study are related to the duration of the follow-up. Authors encountered difficulties in getting complete data of the operations in some cases, and difficulties in assessing whether the intervention was able to restore the pre-pathologic activity levels. On the other hand, the strength of this study at a very long 35-year follow-up is based on the reliable assessment of the durability of the results over time using this technique.

The long-term results of this study showed that the concept of repair the labral tears is successful for the treatment of traumatic anterior shoulder instability in term of restoring joint stability without impairing the ROM. In this series over 90% of patients presented normal ROM compared with the untreated contralateral side. The patients who presented severely reduced ROM were those who had suffered the most frequent complication of staple capsulorrhaphy: staple loosening treated by surgical revision. With 3.5% of recurrence’s rate and 10% of poor result we can assess that this open procedure holds the ground for the current concept of arthroscopic Bankart repair.

Moreover, our results underlined that low number of dislocations and short time between first dislocation and surgery are positive prognostic factors for successful long-term results. In our series we observed a correlation between limited ROM and incidence of degenerative arthritis. It should be acknowledged that the insurgence of symptomatic grades II and III disease occurred in the patients who had undergone revision surgery for staple loosening or deep infection.

Our results are in agreement with those reported by Sisk et al. and Rao et al. in terms of recurrences and problems related to staple loosening [21, 25]. Rao et al. reported, in a group of 79 patients followed-up for 10 years, a 1% rate of recurrence of dislocation, whereas 94% of their results were excellent, 4% good and 2% poor [21]. Sisk et al. reported a series of 239 patients treated with staples capsulorrhaphy with a high rate of excellent results and low rate of complications [25]. They reported seven revisions for improper staple position or staple loosening, two superficial infections and seven recurrences, one in an uncontrolled epileptic patient and six in football players who had suffered post-traumatic shoulder dislocations.

Our results are contrary to those of O’Driscoll and Evans, who reported a series of 204 open staple capsulorrhaphies. They had a recurrence rate of 29% and a 12% rate of complications due to staple loosening [15]. It should be acknowledge that in their series, 24% of the 157 shoulders had physical signs of posterior instability [17, 21, 25] and 15% of the patients were affected by congenital laxity.

Magnusson et al. [16] in a recent revision of open capsulorrhaphy treated with the Bankart technique, reported a recurrence rate of 17% both in terms of dislocations (11%) and subluxations, but they avoided specifying whether recurrences were anterior or posterior. Rahme et al. [20], revisiting the Eden–Hybinette technique, reported a recurrence rate of 20% and again they did not distinguish between anterior or posterior re-dislocations. In the same report Rahme found that moderate and severe arthrosis occurred in 30% of the treated shoulders. In our series, in keeping with O’Driscoll and Evans, staple loosening and infection was associated with pain and reduction of ROM [17]. The need for revision due to hardware loosening and the possible risk of neurovascular damage might have been the main cause for the progressive abandonment of this procedure, as well as its direct correlation with late moderate or severe gleno-humeral arthrosis.

Another factor related to the loss of a few degrees of external rotation and incidence of asymptomatic grade I arthrosis with regards to the unaffected shoulder may be the inadequate post-operative management, which consisted of a thoracobrachial cast for 4 weeks and a totally inadequate rehabilitation protocol. Previous studies on the Du Toit technique reported the recovery of full ROM in most patients within 3 weeks after surgery [21], which highlighted the importance of early mobilization and avoiding the thoracobrachial cast.

The percentage of grade I asymptomatic arthritis cannot be correlated to the surgery alone because at such long-term follow-up grade I arthritis might be a normal joint evolution if we consider the mean age at follow-up.

On the other hand, Hovelius et al. [11, 12] believe that in shoulders treated for traumatic dislocation, it was the primary dislocation that initiated the arthropathy.

The Du Toit procedure was the first technique to eliminate shoulder instability by correcting the pathological defect at the basis of anterior post-traumatic dislocation, i.e. glenoid labrum detachment. In the Du Toit procedure, in fact, the first staple is put in the glenoid rim to repair the labrum detachment and the second is put in the medial aspect of the glenoid in order to reduce pathological elongation of the capsule and tighten the inferior gleno-humeral ligament.

We can conclude that previous experience with this Bankart-like technique holds the ground for the current concept of arthroscopic Bankart repair.

When all the inclusion criteria were respected and the surgical technique was correctly performed the Du Toit technique turned out to be a successful procedure for post-traumatic anterior dislocation with a high rate of excellent results at 35 years of follow-up. In our series it is reasonable to say that the correlation between limitation in external rotation and the incidence of gleno-humeral arthrosis is consistent also with the need for a second intervention due to staple loosening and the post-operative management of the treated shoulder, rather than with the surgical technique itself.

Acknowledgements

Dott. A. Vascellari, Dott.ssa A. Montaperto, M. Bonfiglioli, G. Bernagozzi. The experiments comply with the current law of Italy.

Copyright information

© Springer-Verlag 2007