Structural integrity of rotator cuff at 16 years following repair: good long-term outcomes despite recurrent tears

Abstract

Background

There are few studies reporting long-term rotator cuff integrity following repair. The present study reports a case series of surgically repaired supraspinatus tendons followed up with clinical outcomes and ultrasound imaging after an average of 16 years.

Methods

The prospectively studied clinical outcomes at short-, medium- and long-term follow-up in 27 shoulders in 25 patients treated with arthroscopic subacromial decompression and mini-open rotator cuff repair have been reported previously. The functional outcomes scores recorded were the University of California Los Angeles (UCLA), American Shoulder and Elbow Surgeons (ASES) and Simple Shoulder Test (SST) measures. These patients then underwent an ultrasound scan with respect to the long-term assessment of the shoulder and the integrity of the repair.

Results

A recurrent tear was noted in 37% of patients at 16.25 years after surgery, of which 50% were small. Two patients required repeat surgery. Patients had a mean UCLA score of 30, an ASES score of 91.3 and a SST score of 9.5 with a 85% level of satisfaction with surgery. Patients with a recurrent tear had outcome scores equivalent to those with an intact cuff with no significant pain. No independent risk factors were identified as predictors for recurrent tear.

Conclusions

Patients showed sustained benefit and satisfaction at long-term follow-up despite a 37% recurrence of full-thickness supraspinatus tear.

Introduction

Rotator cuff tears of the supraspinatus are a common shoulder condition affecting middle to older age adults. The aetiology can be traumatic, degenerative or a combination. Presentation is variable with many full-thickness tears being asymptomatic, whereas others cause pain and weakness.¹ A significant proportion of tears are known to progress in size and larger tears are more likely to cause pain and weakness.² Rotator cuff arthropathy is the final consequence of massive rotator cuff tear and causes stiffness, pain and weakness.

Rotator cuff repair aims to restore cuff integrity, stop pain and preserve power and motion. With increasing evidence suggesting that the condition is degenerative, with a genetic predisposition, it is imperative to investigate whether we are able to alter the natural history of this condition by reattaching the tendons to bone.^3–6

The purpose of the present study was to assess the structural integrity of the rotator cuff at long-term follow-up following superior cuff repair. The secondary objective was to assess clinical outcomes in relation to cuff integrity and identify possible factors associated with poor structural or clinical outcome.

Materials and methods

Medical Ethics Committee approval was obtained and all study participants provided their informed consent. Patients were initially enrolled by the senior surgeon (SNB) over a 4-year period. We have previously reported on the clinical outcomes at medium- and long-term follow-up from this series.^7,8 Patients presented with pain and weakness of the affected shoulder, with clinical findings suggesting rotator cuff pathology and impingement.

Inclusion criteria were the presence of a reparable supraspinatus, or supraspinatus and infraspinatus, tendon tear. Patients were excluded in the presence of glenohumeral or acromioclavicular osteoarthritis, biceps degeneration requiring tenodesis or a massive cuff tear (greater than 5 cm).

The presence of a reparable supraspinatus tear was demonstrated on ultrasound in all cases and subsequently confirmed intra-operatively. Tear size was measured intra-operatively in the sagittal plane at the greater tuberosity and classified as small (<1 cm), medium (1 cm to 3 cm), large (3 cm to 5 cm and involving two or more tendons). Massive (>5 cm and involving two or more tendons) tears were excluded.⁹ All operations were performed by the senior surgeon (SNB).

The patient was placed in a lateral position with the arm in traction. An arthroscopic acromioplasty was performed, including bursectomy. Subsequently, the rotator cuff was approached via a small longitudinal mini-open split in the deltoid muscle attachment to the acromion, taking care not to detach any deltoid muscle from the acromion. The greater tuberosity bone was freshened and a rotator cuff-to-bone repair was performed with use of transosseous Ethibond number-1 suture (Ethicon, Somerville, NJ, USA). The suture was passed through the cuff tendon medially, and then through the tuberosity bone at the articular margin to emerge lateral to the tuberosity. In general, two separate passes through the bone and tendon were carried out for each individual ethibond suture. PDS (polydioxanone) number-1 suture (Ethicon) was used for side-to-side tendon repair and was placed to diminish stress on the tendon-to-bone sutures. No transosseous sutures through the lateral cuff margin or Mason-Allen suture techniques were utilized.

Postoperative rehabilitation was routinely commenced on the first postoperative day. This consisted of pendular exercises with supine passive self-assisted external rotation and forward elevation. A sling was worn for 3 weeks.

Seventy-nine consecutive patients were initially recruited for follow-up. At the 7-year report, this was reduced to 74 as a result of three deaths and two patients loss to follow-up. By 2010, the number of willing and able participants had dropped to 49. Of those that could not participate with respect to questionnaire completion: 11 had died, three had dementia and 16 were completely lost to follow-up. Patients were asked to complete a subjective self-scoring University of California Los Angeles (UCLA) questionnaire. This included visual analogue pain scores at rest and on activity and patient satisfaction with surgery. The outcome results are categorically reported as excellent (34 points to 35 points), good (28 points to 33 points), fair (21 points to 27 points) or poor (< 20 points).¹⁰

American Shoulder and Elbow Surgeons (ASES) and Simple Shoulder Test (SST) measures were also self-assessed. All tests have been validated for self-assessment.^11–13

Patients were invited after their last clinical follow-up to undergo diagnostic ultrasound scan (USS) at a University affiliated radiology practice where scans were dual reported by two out of three experienced musculoskeletal radiologists. A dedicated study report sheet was completed assessing the integrity of the cuff tendons, long head of biceps, bursal tissue and the presence of fluid about the bursa, joint and long head of biceps. Tears of the cuff were measured in the sagittal plane and classified according to thickness and tendons involved. Twenty-seven shoulders underwent ultrasound in 25 patients (two bilateral). Of the 49 patients who were located and completed the self-assessment at this time; 24 did not undergo USS. The reasons for not attending the USS included inconvenience or other obligations such as work or care giving (n = 6), immobility and frailty (n = 3), unacceptable distance to imaging centre (n = 5), did not want to participate (n = 2) and no reason given/loss to follow-up (n = 8) (Figure 1).

Figure 1.

Flowchart demonstrating patient drop out from recruitment to final follow-up.

Patients

The mean age of the patients who underwent USS was 69.2 years. There were 16 male shoulders and 11 female shoulders. Two patients had bilateral repairs. Mean follow-up was 16.25 years (range 14 years to 18 years). The initial tear size was graded as small in one, medium in 22 and large in 4. It was the dominant shoulder in 19 (59%). These demographics were compared to the group that failed to attend USS to assess for responder bias (Table 1). Age, UCLA and SST scores were very similar between the groups (p > 0.05).

Table 1.

Comparison of USS participants and non-USS patients.

Demographics	USS group	Non USS	p value
Male	16 (59%)	12 (50%)	0.58
Female	11	12
Mean age at surgery (years)	52.9	53.6	0.82
Tear size
Small	1	0
Medium	22	17
Large	4	7
Side
Right	16 (59%)	14 (58%)	1.0
Left	11	10
Dominance
Dominant	19 (70%)	14 (58%)	0.40
Nondominant	8	10
UCLA	30	30.6	0.73
Excellent/good	20	17
Fair	2	6
Poor	5	1
ASES	91.3	84.6	0.21
SST	9.5	9.3	0.79
Satisfaction	85%	92%
Repeat surgery	7.4%	8%

USS, ultrasound scan; UCLA, University of California Los Angeles; ASES, American Shoulder and Elbow Surgeons; SST, Simple Shoulder Test.

Statistical analysis

Statistical analysis was performed by an independent statistician using R software (R Foundation for Statistical Computing, Vienna, Austria). Continuous parametric data were assessed with two-sample, two-tailed t-test. Distribution free Kruskal–Wallis (equivalent to the Wilcoxon rank-sum test) was used for nonparametric continuous data. Categorical data was assessed with Fisher’s exact two-sided test. p < 0.05 was considered statistically significant. Subsequent logistic regression analysis was performed.

Results

Full-thickness re-tears (or failure to heal) of the surgically repaired supraspinatus were noted in 10 shoulders. Nine full-thickness tears were noted at USS and the other was noted during revision surgery. The incidence of full-thickness recurrent tear was 37%. Overall, five re-tears were small, three re-tears were medium and two were large in size. Only one tear increased in size, from medium to large; the remainder were the same size or smaller.

Two patients required revision repair of the supraspinatus tendon. One patient required revision repair 16 months after the primary surgery. The revision surgery, which was for acromioclavicular pain, identified failure of healing in the mid portion (< 1 cm, small) of supraspinatus and an identical repair technique was used again, in addition to arthroscopic resection of the distal clavicle. This patient was noted to have an intact cuff on USS at 14 years after revision repair. For the purpose of the present study, this patient was assessed as being a re-tear, despite the cuff being intact on USS, as a result of the need for revision surgery. The other patient underwent revision repair in 2010 by another surgeon (15 years after the initial repair); subsequent USS reported a recurrent 30-mm tear in the supraspinatus, 10 months after revision (see Supporting information, Table S1).

The mean UCLA score was 30. Excellent or Good outcomes were achieved in 74% and fair outcomes were achieved in 7%, whereas 18% had a poor UCLA score. The mean ASES score was 91.3, the mean SST score was 9.5 and 85% of the patients were satisfied with their surgery.

Comparing those with a recurrent tear with those whose supraspinatus was intact, it was noted that the mean age at surgery was higher in those with re-tear, (57.1 versus 50.6, p = 0.12), although this did not reach statistical significance. There was no relationship between re-tear and male and female sex (p = 0.69), arm dominance (p = 1.0) and manual workers (p = 0.23). Initial tear size was not a predictor of subsequent re-tear (p = 0.81). Only one patient in the group reported smoking at the time of surgery, and there were two patients who had diabetes. These numbers are too small to warrant formal analysis. Of note, both diabetics had intact tendons and the smoker had a partial thickness supraspinatus tear.

Those with recurrent tears had a mean UCLA score of 31.7 with nine of 10 patients categorized as good or excellent. The mean UCLA for those without full-thickness recurrent tear was 29.0 (p = 0.52). The mean ASES in the re-tear group was 95.8 compared to those without full-thickness tear with a mean ASES of 88.7 (p = 0.48). The SST scores were also similar (p = 0.73) (Table 2).

Table 2.

Demographics, and outcomes in those with and without retear.

Demographics	No tear	Tear	p value
Mean age at surgery (years)	50.6	57.1	0.12
<45	5	1
45 to 55	6	3
56 to 65	6	4
>65	0	2
Male	11 (65%)	5 (50%)	0.69
Female	6	5
Dominant side	13 (76%)	7 (70%)	1
Nondominant	4	3
Manual worker	9 (53%)	2 (20%)	0.23
Non manual	8	8
Smoker	1 (6%)	0
Diabetes	2 (12%)	0
Workers’ compensation	6 (35%)	0	0.06
Pre-operative cortisone	6 (35%)	6 (60%)	0.26
Mean pre-operative tear	18.2 mm	15 mm	0.81
Small	1	0
Medium	14	8
Large	2	2
UCLA	29	31.7	0.52
Good/excellent	11	9
Fair	2	0
Poor	4	1
ASES	88.7	95.8	0.48
SST	9.3	9.9	0.73
Satisfaction
Satisfied	14 (82%)	9 (90%)	1
Not satisfied	3	1
Subacromial bursa
Thickened	12 (71%)	7 (70%)	0.23
Not thickened	5	3
Long head of biceps
Present	16 (94%)	8 (80%)	0.53
Ruptured/cut	1	2

UCLA, University of California Los Angeles; ASES, American Shoulder and Elbow Surgeons; SST, Simple Shoulder Test.

Satisfaction was 90% for those with a recurrent tear compared to 82% in those without recurrent tear (p = 1.0).

There was no association between bursal thickening and recurrent tear (p = 0.23). There was no association between bursal thickening, subscapularis tendinopathy, long head of biceps presence/absence or supraspinatus tendinopathy and the clinical outcome scores (Table 3).

Table 3.

USS findings and clinical outcomes.

	USS findings		p
	LHB
	Intact	Absent
UCLA	30.2	28.5	0.6
ASES	91.0	94.0	0.7
	Subscapularis
	Intact	Tendinosis/PT
UCLA	31.0	27.6	0.3
ASES	93.4	86.5	0.3
	Supraspinatus
	Intact	Tendinosis/PT/FT
UCLA	28.7	30.5	0.6
ASES	90.0	91.8	0.8
	Bursal thickening
	Absent	Thickened
UCLA	29.5	30.3	0.8
ASES	95.3	89.0	0.3

USS, ultrasound scan; UCLA, University of California Los Angeles; ASES, American Shoulder and Elbow Surgeons; SST, Simple Shoulder Test.

LHB, long head of biceps; PT, partial thickness; FT, full thickness.

No workers’ compensation patient had a re-tear (p = 0.06), resulting in 35% of those without recurrent tear being workers’ compensation patients. The mean UCLA for workers’ compensation patients was 31.3 (p = 0.64) and ASES was 97.5 (p = 0.25).

The mean age, UCLA, ASES and SST scores between the USS participants and total respondents were very similar, as was patient satisfaction and the re-operation rate.

Logistic regression analysis with re-tear as the outcome was performed. No statistically significant association of variables was noted. A logistic regression model was fitted for age and re-tears with a p-value of 0.12 for the co-efficient of age, which resulted in estimates for the probability of re-tear for patients of different ages (Figure 2).

Figure 2.

Scatterplot of probability of re-tear versus age at surgery.

Discussion

We consider that this series in the present study represents the longest published follow-up on structural integrity following a common, contemporary, rotator cuff repair technique.

Only one published series has a longer radiographic follow-up of repair integrity. That study had a magnetic resonance imaging (MRI) follow-up to mean 20 years after surgery. Its relevance is diminished as a result of the fully open technique utilized with open acromioplasty and the high rate of free tendon graft (78%). A 94% re-tear rate was reported.¹⁴ We found a 37% re-tear rate of supraspinatus at 16.25 years following surgery based on imaging studies and a 7% re-operation rate. The re-tears in almost half were only small and probably not important. Therefore, only 19% had a medium or large size retear.

These imaging results are comparable to other previously reported series using modern techniques. Kluger et al.¹⁵ reported a 33% re-tear rate at 11 years after mini-open cuff repair, with an 8.6% re-operation rate. Their technique was similar with transosseous ethibond sutures, although they utilized a lateral row transosseous repair, which the senior author (SNB) now also utilizes. They also used an abduction pillow for 6 weeks postoperatively, whereas the present study used a normal sling for 3 weeks only.

Castagna et al.¹⁶ reported a 38% re-tear rate at 30 months following arthroscopic repair. Their mean UCLA score was also very similar at 30.8.¹⁶ Oh¹⁷ reported a 31.1% re-tear rate at 29 months and Defranco et al.¹⁸ showed a 40% re-tear rate. A recent meta-analysis cited a mean re-tear rate of 26.6% at a mean of 23 months with a range from 6 months to 118.8 months.¹⁹ The present study reports a mean follow-up of 195 months.

Advancing age has consistently been shown to be an independent predictor of failure of cuff repair integrity.^19–22 Those patients with an intact repair were a mean of 6.5 years younger at surgery, although this failed to reach statistical significance.

A recent study with sequential imaging over the first year of recovery has suggested that the majority of repair failures occur between 6 weeks and 26 weeks after surgery. Their overall re-tear rate was 17% at 1 year with only one of 19 tears occurring from 26 weeks to 52 weeks.²³ This is supported by Kluger et al.¹⁵ whose follow-up extended to 11 years. Zumstein²⁴ demonstrated an increase in failure rate from 37% at 3.1 years to 57% at 9.9 years following open repair of massive cuff tears.

We have an understanding of the natural history of the rotator cuff. The tearing is commonly degenerative in nature and its prevalence increases and approaches 46% of the population at 70 years of age.²⁵ Small tears can, but do not always progress, and symptomatic tears are more likely to be to progress and be large.² What has not been established is whether rotator cuff repair can alter the natural history of cuff integrity. It is clear from the present study that cuff repair and decompression improves pain and outcome scores in most patients from the short- to long-term in an enduring fashion. This outcome appears to be independent of the integrity of the cuff providing that re-tears are small to medium. Other studies have demonstrated a persisting positive outcome after cuff repair despite structural failure of the cuff repair.^{17,23,26–30}

It is hoped that repairing the tendon may alter the biological environment and allow healing of the degenerative tendon. In this long-term follow-up, the failure rate (37%) approaches the population incidence at that age; however, 63% were still healed at average of 16.25 years postoperatively. The recurrent tears were predominantly the same size, or smaller. Only one tear, in an elderly gentleman, had enlarged over baseline and no patients had progressed to cuff arthropathy. Our study population selects those prone to cuff degeneration and so we remain optimistic that this surgery does positively alter the natural progression of rotator cuff degeneration for the majority of our patients.

There are several limitations to the present study. The large loss to follow-up (48%) was disappointing. Although this clearly represents a potential source of responder bias, the comparison of demographics confirms that the USS participants had characteristics similar to the total population sample. The mean age, UCLA, ASES and SST scores between the USS participants and total respondents were very similar, as was patient satisfaction and the re-operation rate.

The small sample size makes correlation of variables with outcomes subject to beta error. The rate of reported smoking was very low, as was the incidence of diabetes. Workers’ compensation patients and manual workers did not have inferior structural or clinical outcomes, although numbers were too low to draw any conclusion. We report individual patient demographics so that they may be incorporated into a subsequent meta-analysis.

Initial tear size was graded intra-operatively but sized by ultrasound on follow-up by three experienced musculoskeletal radiologists and the studies were dual read. USS has been shown to be equivalent to MRI when used in the hands of experienced operators. The results were also found to correlate well, although not perfectly, with arthroscopic findings.³¹

Plain radiographs were not taken at follow-up to assess for any arthritic degenerative change that may have contributed to diminished shoulder function. A shoulder effusion was only noted in two of the patients and both of them had excellent outcome scores. Long head of biceps, subscapularis and bursal ultrasound findings were not associated with outcome scores.

Conclusions

Arthroscopic acromioplasty with mini-open supraspinatus cuff repair demonstrates a 37% rate of recurrent supraspinatus tear at 16-year follow-up, with 19% medium to large tears. Clinical outcome scores and satisfaction with surgery remained high and were not influenced by cuff integrity or other ultrasound findings in the shoulder. Further long-term monitoring and reporting is needed to document the effects of rotator cuff surgery.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: None of the authors received payments or services, either directly or indirectly (i.e. via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the 36 months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical review and patient consent

Human Research Ethics Committee, Southern Health, VIC, Australia, issued a certificate of approval (10180B). All patients consented to the anonymous use of their personal information at initial recruitment, and prior to participation in this study.

Level of evidence

Level IV: therapeutic

Supplementary Material

Supplementary material is availabe at journals.sagepub.com/doi/suppl/10.1177/1758573217738198.