Massive irreparable rotator cuff tear and Superior Capsular Reconstruction
This post is intended to give both physios and members of the public more of an insight into the management of rotator cuff tears, and introduce readers to the role and scope of superior capsular reconstruction. If you are reading this as a patient then it may help to read the previous blogs on tendinopathy and throwing pain, as well as our anatomy page. This should help to give you some background information as the following post has more of an academic flavour.
Rotator cuff tear is a common musculoskeletal pathology affecting the shoulder. It is considered a spectrum of disease process and may be related to wear and tear, genetic or traumatic factors (Littlewood et al 2013). Many studies have looked at the incidence and prevalence of rotator cuff tear in the general population with wide ranging results. Yamamoto et al (2009) conducted a large cohort study looking at 1366 people in a Japanese mountain village and found overall prevalence of rotator cuff tear of 20.7% in the whole population. This ranged 2.5% prevalence in people in their thirties, to 50% in people in their eighties. Rotator cuff tears in the general population were most commonly associated with elderly patients, males, affected the dominant arm, those engaged in heavy labour and having a history of trauma. In a much smaller study, Sher et al (1995) found similar results with 4% of nineteen to thirty five year olds having radiographic evidence of a tear, increasing to 54% in those over sixty five in an American population.
It is widely accepted that the symptoms of rotator cuff tear include pain, weakness and functional limitation (Carr et al 2017). These symptoms can have profound social and economic implications, particularly in younger people (Lazarides et al 2015). Symptoms include pain of variable intensity, and, with posterosuperior cuff (Supraspinatus or Infraspinatus) involvement, can cause an inability to elevate and externally rotating the arm; a condition sometimes referred to as pseudoparalysis. It must also be noted that many people with rotator cuff tears have little or no symptoms. Why?
In 1994, Stephen Burkhart, an eminent American surgeon, proposed that the nature of symptomatic rotator cuff tears was due, in part, to failure of the rotator cuff cable; a thickened band of tissue running perpendicular to the fibres of supraspinatus, anchored anteriorly and posteriorly to the humerus. See image below.
He postulated that the function of the cable was to transfer load and stress across the cuff, similar to the cable of a suspension bridge (see image below), maintaining force couples in the coronal and transverse planes (Burkhart 1993), thus maintaining a stable fulcrum in the shoulder.
He proposed that even in the presence of a rotator cuff tear, if the cable remained intact, the shoulder would have capacity to produce movement, and may be entirely symptom free. This Coined the phrase ‘anatomically deficient but biomechanically intact’. This type of rotator cuff deficiency, if symptomatic, can potentially be managed conservatively with physiotherapy and/or steroid injection or pain relieving surgical procedures like arthroscopic debridement. He proposed that where the shoulder is biomechanically deficient, i.e. it does not move well, then rotator cuff repair should be undertaken.
Rotator cuff repair aims to restore the rotator cuff to near normal anatomy, i.e. a healed tendon-bone interface in its normal anatomical position. This is achieved by anchoring the rotator cuff tendon back on to the bone using special anchors. This restores the action of the rotator cuff to ’centre’ the humeral head onto the glenoid, reducing proximal migration of the humeral head which can lead to a progressive degenerative process called rotator cuff arthropathy (Jost et al 2000). It also restores the anterior/posterior force couples as described by Burkhart (1993) and returns the cuff to being a biomechanically intact unit. Between 1996 and 2006, the unadjusted volume of rotator cuff repairs in the USA increased by 141%, with a 600% rise in arthroscopic procedures (Colvin et al 2012).
Massive irreparable rotator cuff tear poses a significant surgical treatment dilemma, particularly in the younger patient. The recent UKUFF trial has shown that both arthroscopic and open rotator cuff repair yields significant improvement in the Oxford shoulder score (OSS – an outcome measure used to score improvement) compared to baseline (Carr et al., 2017). This was a well-constructed randomised control trial but only included small, medium and large tears. They found no significant difference in subgroup analysis based on size of tear. However, OSS was worse in those patients in whom repair was not possible. Another interesting finding was the high re-tear rate (46.4% in the arthroscopic group versus 38.6% in the open group). Within this group there was still a significant improvement in OSS compared to baseline. Some postulate that this improvement may be due to restoration of anterior and posterior force couples through achieving a partial repair. It could also be attributed to placebo, but is more likely a result of the lengthy period of rehabilitation post-operatively.
The mainstay of surgical procedures for massive rotator cuff tear is arthroscopic debridement with biceps tenotomy. In his 1994 paper ‘Reconciling the Paradox of Rotator Cuff Repair Versus Debridement: A Unified Biomechanical Rationale for the Treatment of Rotator Cuff Tears’, Burkhart proposed the following criteria for debridement and decompression to manage massive rotator cuff tear; 1. No improvement after 4 months of nonoperative treatment 2. Age >60 years 3. Large tear that is not easily reparable 4. Active forward elevation of at least 100 °, even though painful 5. Strong external rotators 6. Subscapularis intact (negative lift-off) 7. Balanced force couples. In cases where these conditions are met, debridement and decompression to minimise impingement and reduce edge instability can be highly effective. It helps to reduce pain to allow the pre-existing functional capability to be utilised. In 1991 Burkhart published a case series of ten people with massive rotator cuff tears as per Cofield (1985) classification. He claimed that pain levels were significantly improved in all patients, and all but one had ‘normal’ range of elevation. This is obviously a small case series and may well be susceptible to reporting bias. However, several studies, including a prospective case series by Berth (2010) have supported the findings of significant pain relief following debridement.
Debridement can be considered as a salvage procedure, and, as Burkhart suggested, it is best deployed in those over sixty years of age. In younger patients, Latissimus Dorsi (LD) tendon transfer has been proposed as a treatment alternative for patients with massive rotator cuff tear (Gerber 2006). Transfer of the LD tendon onto the humeral head, or remaining rotator cuff tendon, provides an interposition graft to reduce pain and can also restore external rotation strength. Literature suggests that outcomes are favourable in the short to mid term in well selected patients (Gerber 2013). Espinosa et al (2006), in a case series, looked at clinical outcomes of sixty nine shoulders (sixty seven patients) who underwent LD transfer for massive irreparable tear of at least two tendons. The cohort comprised of fifty two men and fifteen women with a mean age of 61 (range 49-72). Thirteen patients were also deficient in subscapularis. Inclusion criteria required a period of six months failed conservative management and an external rotation lag of at least fifteen degrees. Patients with relatively painless pseudoparalysis (elevation of no more than thirty degrees) and those who had anterior-superior escape were excluded. This study supported previous findings (Gerber 1992 and Aoki 1996) that LD transfer does not produce favourable outcomes in those with Subscapularis deficiency.
Reflecting on this study, Gerber himself comments that reported improvements in subjective shoulder value seem disproportionately good when compared to objective metrics, where flexion and external rotation range improved by a mean 18 degrees and 10 degrees respectively. Mean isometric strength improved from 0.9kg to 1.8kg – 25% normal male strength in age matched controls in a neutral position.
When comparing outcomes in primary versus revision surgery, outcome scores for revision procedures were lower than those in primary procedures, however, the degree of increase in score was comparable in both groups.
Of some concern was the progression of arthritic change noted over the fifty-three month follow up period. Progression of osteoarthritic change of more than 2 stages was noted in seven shoulders, with progression of one stage noted in a further fourteen shoulders, representing progression in 30% of the cohort. This contrasted with Gerbers previous, smaller, 1996 study where no progression of OA was noted. Aoki et al (1996) noted comparable findings to Gerbers later study with progression of OA in 41% of shoulders. In the 2006 case series, acromiohumeral distance decreased by an average of 1.5mm post operatively, and correlated with the progression of arthritis. Reported levels of pain and function did not, however, correlate with arthritis. This should obviously be considered in the decision-making process when listing patients for LD transfer. This point is particularly salient in younger patients where revision to reverse geometry arthroplasty in the arthritic and cuff deficient shoulder has a reported complication rate of up to 38% (Ek, Neukom, Catanzaro & Gerber 2013 & Hartzler 2015).
At ten year follow up the mean acromiohumeral distance had decreased from 7.4mm to 4.9mm and a slight, but significant increase in osteoarthritic change was noted (Gerber et al., 2013). Inferior results occurred in shoulders with insufficiency of the subscapularis muscle and fatty infiltration of the teres minor muscle. Superior functional results were observed in shoulders with a small postoperative critical shoulder angle.
Recently, superior capsule reconstruction has been proposed as another alternative surgical treatment option for patients with massive rotator cuff tear by Mihata in 2007. His technique used a tensor fascia lata (TFL) autograft as an interposition graft anchored between the superior glenoid and greater tuberosity, restoring the structural integrity of the superior capsule. This procedure evolved from a circumstance where reverse shoulder arthroplasty was not available in Japan until 2014, and Japanese surgeons still do not have access to allografts.
Mihata theorised that restoration of the superior capsular construct would improve glenohumeral biomechanics by limiting proximal migration of the humeral head. In turn contact pressure between the humeral head and acromion would be reduced, decreasing pain symptoms arising due to impingement. This is supported by the work of Ishihara et al (2014). Their cadaveric study demonstrated a significant increase in humeral head translation in all directions, and significant increase in acromio-humeral contact pressure, in the presence of a superior capsule defect compared to anatomically intact models. This superior capsule defect is deemed to be the same as those seen in massive rotator cuff tear.
In 2012 Mihata et al published their short to mid term results of SCR in a retrospective case series of 24 shoulders. All had undergone arthroscopy to confirm irreparable tear, eleven were classified as large (3-5cm) and 13 as massive (>5cm). Nine had involvement of subscapularis. All underwent SCR +/- subscapularis repair with margin convergence of the autograft to any remaining Infraspinatus, Supraspinatus or Subscapularis tissue. Those who had ‘severe’ proximal migration which did not correct with traction were excluded though it was not quantified precisely how this was measured or how many were excluded on this basis. All were assessed for level of pre-operative osteoarthritis using the Hamada scale (Hamada, Yamanaka, Uchiyama, Mikasa & Mikasa 2011). Twenty were primary procedures, with four being revisions of previously failed surgeries.
Results were very favourable with significant improvement in all three patient reported outcome measures (ASES, JOA and UCLA scores). Average pre-operative acromiohumeral distance measured 4.6±2.2mm, this was found to increase by an average 4.1±1.7mm. Elevation improved by an average of 64 degrees, with pre op average being 84 degrees to a post operative average of 157 degrees. External rotation active range improved by an average of 14 degrees. Graft or tendon repair failure rate at follow up (range 24 to 51 months) was 16.7%, there were no other reported complications.
In their study, Mihata et al (2012) had no control or comparison group and as such we are unable to extrapolate what the outcome for this cohort would have been with different intervention, particularly as a high proportion were primary procedures. Also, when drawing comparisons between results with Gerbers LD transfer study, Mihata lacks a degree of objectivity. He reported an increase in external rotation strength from 3+ on the oxford scale to 5-, with an identical increase in abduction strength. Cuthbert and Goodheart (2007) published a literature review on the reliability and validity of manual muscle testing. This concluded that there is evidence for reliability and validity, though use of a more scientifically rigorous measure like an isokinetic machine, as used by Gerber, would be desirable.
As previously discussed, Mihatas technique utilises a TFL autograft as allografts are not available for use in Japan. Although Mihata et al (2013) have reported no complications associated with this, there are obvious potential concerns of morbidity associated with a second surgical site. As such Hirahara and Adams (2015) proposed a technique for SCR using a decellulated extracellular matrix (ECM) allograft (Human transplant tissue). This reduces the risk associated with opening a second surgical site to harvest an autograft, but it does introduce the risk of reaction associated with residual DNA in the dermal graft.
Hirahara reports a significant increase in load to failure with an ECM patch (550N) compared to the TFL autograft (180N), though this evidence has not been published. It must also be remembered that his work is industry funded by the manufacturers of the ECM graft, raising a potential conflict of interest. To date there are no published outcomes of using ECM in SCR.
Massive rotator cuff tear remains a difficult problem to manage despite significant quantities of published literature on the topic. It is an extremely heterogenous group where individual variables may dictate surgical options, and make research in the form of true randomised controlled trials difficult.
Further surgical treatment options for massive rotator cuff tears include surgery such as transposition of subscapularis, transplantation of Teres Major, Deltoid flap procedure, Pectoralis Major tendon transfer, interposition balloon and reverse geometry shoulder arthroplasty.
As previously discussed, arthroscopic debridement has good outcomes for pain when function is maintained pre-operatively. However, it is not a solution in the case of biomechanical impairment as per Burkharts (1994) suggestions, and inclusion criteria should be rigorously implemented. If good functional outcome is desirable then it should not be considered in patients with pseudoparalysis, external rotation lag or subscapularis deficiency.
Fundamentally, the efficacy of SCR in managing massive rotator cuff tear depends on the clinical relevance of the superior capsule ligament and its role in maintaining a stable fulcrum around which the arm can move. If, as Ishihara et al (2014) suggest, the superior capsule does have a profound role in stability of the glenohumeral joint, which is compromised in massive rotator cuff tear, then SCR may have a significant role in the surgical management of massive rotator cuff tear.
On reviewing the literature, surgery which does not involve reconstruction of the superior capsule, such as LD transfer, is associated with a progression of osteoarthritis. This may well be due to a failure to control proximal migration of the humeral head, causing a progressive degenerative joint disease as seen in cuff tear arthropathy, as evidenced by progressive narrowing of the acromiohumeral distance as seen in Gerber et al (2013) case series. The same may be said of repair with augmentation patch, where the patch may simply serve as an interposition. As the superior capsule remains deficient, superior stability is compromised allowing attrition between the humeral head and acromion, as demonstrated by Audenaerts (2006) findings of failure to restore acromiohumeral distance with repair using synthetic interposition graft. This attrition may well be the cause of high rates of failure seen in patch augmentation. Unfortunately, the Mihata et al (2012) case series makes no comment on progression of osteoarthritis in his cohort; it will be interesting to learn the outcomes in this regard in future research. His work showed significant and maintained increase in acromiohumeral distance with SCR. As such, if progressive degenerative disease is related to proximal migration of the humeral head, this should be reduced with SCR, reducing the long-term risk of arthropathy.
When considering the role and scope SCR, we can draw some conclusions from literature. Namdari et al (2012) published a systematic review on LD transfer for massive rotator cuff tear and concluded that Subscapularis muscle insufficiency, advanced teres minor muscle atrophy, and the need for revision surgery were correlated with poor functional outcomes in some studies. In contrast, in the Mihata et al (2012) series, patients with subscapularis deficiency had equally good outcomes. There was no progression of fatty degeneration in any cases and a small number showed improvement in fatty degeneration. Subjects undergoing revision surgery had poorer outcomes but still had significant improvement. As such Mihata et al (2012 & 2013) concluded that SCR, as well as restoring the superior capsular construct, helped to restore force couples across the shoulder, improving the biomechanical function of the rotator cuff.
As such, early evidence points to SCR having good outcomes both in terms of pain and function. SCR does not seem to be affected by the limitations which hamper outcome from LD transfer. When considering the small but significant risk of osteoarthritis with LD transfer we cannot make any direct comparisons with SCR, but biomechanical studies suggest the likelihood of superior outcomes with SCR. As SCR can be performed as an arthroscopic procedure there are theoretically less risks when compared with LD transfer which can be a lengthy, open procedure.
In conclusion, early evidence supports SCR as a good surgical treatment for massive rotator cuff in the young population when compared with alternative surgical options. Anecdotally, post operative patients that we have seen following SCR have all shown significant improvement compared to preoperatively.
SCR is still a relatively new procedure and within the r