Arch Orthop Trauma Surg DOI 10.1007/s00402-015-2239-1
ORTHOPAEDIC SURGERY
The natural history of rotator cuff tears: a systematic review Walid Eljabu1 • Hans Michael Klinger2 • Marius von Knoch3
Received: 22 January 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose To analyse the current scientific evidence regarding the natural history of the clinical and anatomical progression of rotator cuff tears. Methods A broad systematic review of the literature (PubMed database through January 2014) which was guided, conducted and reported according to PRISMA criteria. This article focuses on the rotator cuff tears. Articles had to meet an inclusion criteria. The methodological quality of each study was individually assessed using a recently developed general assessment tool AMQPP (assessing the methodological quality of published papers). Results Seven articles dealing with rotator cuff tears were included, one of them was a high-quality study. Three papers assessed the natural history and the natural course of rotator cuff rupture directly. The other studies indirectly assessed the natural history with reports on non-operative and operative therapy trends. All of these articles had been published in four different top medical journals according to 2013 ranking. We found no articles which clearly referred to the role of regression to the mean of rotator cuff tears.
& Walid Eljabu
[email protected] 1
Department of Traumatology, Hand Surgery and Orthopaedics, Klinikum Bremerhaven Reinkenheide gGmbH, Bremerhaven, Germany
2
Department of Orthopaedic Surgery, Universita¨tsmedizin Go¨ttingen, Go¨ttingen, Germany
3
Department of Shoulder Surgery, Kreiskrankenhaus Osterholz, Osterholz-Scharmbeck, Germany
Conclusion The development of symptoms and anatomical deterioration are often directly correlated. Spontaneous recovery to normal levels of function has been successfully achieved, and standardised non-operative treatment programmes are an effective alternative to surgery for many patients. Follow-up is necessary to avoid irreparable stage. However, surgery is still favoured by young active people and highly professional persons who need to get fit in a short period of time. Further research is still necessary. The AMQPP score system is simple and reliable. It works as a quick quality-checking tool which helps researchers to identify the key points in each paper and reach a decision regarding the eligibility of the paper more easily. Keywords Natural history Rotator cuff tears Supraspinatus tear Non-operative therapy Regression to the mean AMQPP assessment tool
Introduction Rotator cuff tears are the most common tendon injury in the adult population, affecting at least 10 % of people over 60 years of age in the United States alone [1]. They often lead to painful impairment of shoulder function. Although non-surgical intervention can achieve good pain relief and improvement in the range of motion in the majority of patients, there has been little research on the exact anatomical natural history of non-surgically treated tears [2]. The concern that a tear will progress is one of the reasons why surgical repair is considered at an early stage as this procedure can halt progression and degradation of the rotator cuff musculature. Since the first report on surgical reconstruction of a rotator cuff tendon tear by Codman in
123
Arch Orthop Trauma Surg
1911, various open and arthroscopic techniques have been developed [3]. Detailed evaluation of the risks and knowledge of the factors associated with the development of rotator cuff tears would make it easier to identify at-risk individuals and guide successful preventive and therapeutic strategies. This knowledge is particularly important for younger people, as maintaining shoulder strength is essential for an active lifestyle. Therefore, we systematically reviewed the literature to analyse the current status of medical research regarding the clinical and anatomical development of rotator cuff tears and the therapies most frequently applied.
available in full text; (4) the article investigated the natural history or course of shoulder disorders, the outcome of the non-operative management, or the regression of the shoulder symptoms to the mean. Exclusion criteria included: (1) basic science; (2) animal models, cadaver studies or studies of an asymptomatic population; (3) biomechanical studies not reporting clinical outcomes; (4) studies reporting only imaging (X-ray, ultrasound, computed tomography or magnetic resonance imaging with no clinical assessment); and (5) systematic reviews. All criteria were independently applied. In case of disagreement, a consensus method was used to discuss and resolve the disagreement between the authors.
Methods We performed a systematic review of the available medical literature in the English language using the US National Library of Medicine/National Institutes of Health (PubMed) bank data. It was conducted and reported according to the protocol outlined by PRISMA (preferred reporting items for systematic review and meta-analyses) [4]. Identification and selection of the literature The search was independently performed by the first author in January 2014. It started with a generic search strategy to find studies on shoulder disorders and was then combined with a subject-specific strategy addressing the shoulder and its natural history, non-operative therapy and its relation to regression to the mean. Regression to the mean is a statistical phenomenon that occurs whenever there is a nonrandom sample from a population and two measures are imperfectly correlated. To avoid making incorrect inferences, regression toward the mean should be considered when designing scientific experiments and interpreting data. The articles related to rotator cuff tears were part of the final search results along with a number of articles covering other shoulder problems. All relevant article titles and abstracts were independently screened and reviewed applying the inclusion/exclusion criteria. Full-text articles were retrieved if the abstract provided insufficient information to establish its eligibility. All articles that passed the first eligibility screening were fully read and assessed. The bibliographies of identified studies were also assessed in order to seek additional relevant publications that were not identified in the computerised search. Selection criteria Inclusion criteria were: (1) the article was written in English; (2) the level of evidence was 1–4; (3) the article was
123
Quality assessment The methodological quality of each of the included studies was assessed using a newly developed assessment tool (AMQPP). It is based on Greenhalgh’s article (Assessing the Methodological Quality of Published Papers) published in the British Medical Journal (BMJ 1997) [5]. Because of the structural and functional differences between the included articles, it was a difficult task to find a single existing assessment method which could be applied to all papers. Therefore, the authors created the AMQPP assessment tool (Table 1) to evaluate the essential requirements for high-quality papers. The maximum score on the AMQPP tool criteria list is 6. The total score was counted from all the criteria that scored ‘yes’. ‘No’ and ‘unclear’ scored no points. Based on our experience and observations, we used a minimum score of 4 out of 6 to indicate a good-quality study. These four points have to match the first four questions of the AMQPP tool. The AMQPP assessment tool was linked with the level of evidence rating. We used the scientific credibility, impact factor 2013 and the ranking of the publishing journal for 2013 as a relative quality indicator of the included studies (Table 2). Data extraction The first author independently extracted data from the selected studies on the study population, study design, hypothesis, treatments, outcomes and summary of results (Table 3). The senior authors reviewed and confirmed the abstracted results of the first author. During the review, the senior authors were blind to the initial abstracted results. Due to the diversity in the outcomes, the treatments described in the included studies and the differing presentations of data, statistical analysis was not performed.
Arch Orthop Trauma Surg Table 1 A recently developed tool AMQPP (assessing the methodological quality of published papers) as a quality assessment tool
AMQPP
Yes
No
Not clear
1. Is the study original? 2. Does the study make clear what it is about? (hypothesis clearly stated, subjects recruited, inclusion and exclusion criteria, circumstances) 3. Is the design of the study sensible? (What specific intervention or other manoeuvre was considered and compared? How was the outcome measured?) 4. Does the study deal with preliminary statistical questions? (the size of the sample, the duration of follow-up, the completeness of follow-up) 5. Does the study avoid or minimise systematic bias? 6. Was assessment blind? (Did the people assessing the outcome know which group the patient they were assessing was allocated to?)
Table 2 The impact factors and rankings of the publishing journals for 2013 Journal
No. of articles
Rank out of 40 journals (orthopaedics)
Impact factor (IF) 2013
Am J Sport Med.
1
1
4.439
J Bone Joint Surg. Am. Clin Orthop Relat Res.
4
4
3.355
1
8
2.787
1
14
2.319
J Shoulder Elbow Surg.
Results The PubMed search resulted in 1413 citations 265 of which were duplicates. These were removed, leaving 1148 titles with abstracts to review. After the first screening, the fulltext articles of 26 potentially eligible citations were retrieved. A total of seven studies dealing specifically with rotator cuff rupture were finally included in this systematic review after 18 articles reporting on other shoulder disorders had been excluded. In view of the lack of material dealing with the natural history of rotator cuff injury, a high-quality systematic review of this topic seemed a difficult undertaking. However, an alternative approach to this task was to investigate the options for treatment of this shoulder problem. There were no studies that clearly identified the role of regression to the mean in shoulder diseases. The quality assessment tool (AMQPP) revealed one article with the full mark of 6. The remaining six articles scored 4 out of 6, because the papers did not specify whether or not the assessors were blind to the patient’s treatment programme and there was no clear action to minimise the systematic bias. This result correlated well with the level of evidence rating of the included articles. The impact factors and the ranking of the publishing journal for 2013 were also analysed (Table 2). The articles included in this systematic review were published in four different orthopaedic
journals which ranked in the top 15 out of 40 listed journals according to medical journal impact factors in 2013. A summary of the characteristics of each study is presented in Table 3. The samples in the included studies ranged from 19 [6] to 452 [7] patients. The table contains one high-quality study which had level 2 of evidence according to the Oxford Centre for Evidence-Based Medicine [8]. One paper was rated at level 3 of evidence and the remaining five articles were case series or case-controlled studies (level 4). Of these seven articles, three studies were designed to be therapeutic and four articles were aimed to be prognostic. All included publications had directly or indirectly studied the natural history of shoulder diseases and evaluated the effect of the patients’ characteristics on the outcome of the disease and on certain therapeutic measures. In their prognostic level-2 paper Moosmayer et al. [9] studied the natural history of 50 patients with asymptomatic rotator cuff tears for 3 years. Eighteen tears developed symptoms during the follow-up period. There was a significant increase in the mean tear size, a higher progression rate of muscle atrophy, a significant rate of fatty degeneration, and a higher rate of pathology of the long head of the biceps tendon in the symptomatic group in comparison to patients with no clinical problems. This correlation between the increase in tear size, decrease in muscle quality and development of symptoms was also reported by Mall et al. [10] in a cohort of 195 asymptomatic rotator cuff tears in their prognostic level-3 study. During a follow-up period of more than 2 years, 18 % of the full-thickness tears showed an increase of [5 mm, and 40 % of the partialthickness tears had progressed to full thickness. In comparison with the assessment made before the onset of pain, the American Shoulder and Elbow Surgeons scores for shoulder function had significantly fallen and the range of motion had decreased in all planes except for external rotation at 90° of abduction. Compared with the tears that remained asymptomatic, the tears which became painful were found to be significantly larger in size than at the time of first examination. Safran et al. [11] followed 51 patients with 61 full-thickness rotator cuff tears over a mean period
123
123
Sample size
50 patients
452 patients
24 patients
51 patients
195 patients
19 patients
46 patients
References
Moosmayer et al. [9]
Kuhn et al. [7]
Fucentese et al. [13]
Safran et al. [11]
Mall et al. [10].
Zingg et al. [6]
Goldberg et al. [12]
Therapeutic level 4 case series
Prognostic level 4 case series
Prognostic level 3 prospective comparative
Prognostic level 4 case series
Therapeutic level 4 case series
Therapeutic level 4 a multicentre prospective cohort
Prognostic level 2 prospective comparative
Study design
Table 3 Summary of the included papers
To determine the outcome of non-operative management (patient education and home programme of gentle stretching and strengthening) of full-thickness rotator cuff tears
To determine clinical and structural mid-term outcomes in a series of non-operatively managed rotator cuff tears
To identify changes in tear dimensions, shoulder function, and glenohumeral kinematics when an asymptomatic rotator cuff tear becomes painful compared with those which remain asymptomatic
To determine the size change of non-operatively treated fullthickness rotator cuff tears after 3-year follow-up
To assess the clinical and structural outcomes of a consecutive series of patients with symptomatic, isolated full-thickness supraspinatus tears who had declined operative treatment
To study the effectiveness of a specific non-operative physical therapy programme in treating atraumatic full-thickness rotator cuff tears
To assess whether deterioration in rotator cuff tear anatomy correlates with the development of symptoms. Patients with initially asymptomatic full-thickness rotator cuff tears were studied
Hypothesis
About 60 % of patients showed improvement, 30 % experienced worsening, and 11 % remained unchanged
Patients with a non-operatively managed, moderately symptomatic massive rotator cuff tear can maintain satisfactory shoulder function for at least 4 years despite significant progression of degenerative structural joint changes. There is a risk that a reparable tear will progress to an irreparable tear within 4 years
Pain development is associated with an increase in tear size and deterioration of shoulder function and active range of motion
Full-thickness rotator cuff tears tend to increase in size in about half of patients aged 60 years or younger. Surgery should be initially considered for these patients to prevent a probable increase in tear size. Patients treated non-operatively should be routinely monitored for tear size increase, especially if they remain symptomatic
Small isolated full-thickness tears of the supraspinatus in patients under the age of 65 do not necessarily progress over time. The refusal of surgery surprisingly resulted in high clinical patient satisfaction and no increase in the average size of the tear 5.5 years after surgical repair had been recommended
Patients’ outcomes improved significantly. Less than 25 % of patients went on to have surgery
About 40 % of asymptomatic rotator cuff tears became symptomatic and anatomically deteriorated. Increase in tear size and decrease in muscle quality correlated with the development of symptoms
Conclusion
Arch Orthop Trauma Surg
Arch Orthop Trauma Surg
of 29 months. The tears increased in size in 49 %. There was no change in 43 and 8 % decreased in size. No correlation was found between the change in tear size and age of the patient, existence of a prior trauma or bilateral tears. Goldberg et al. [12] documented the functional outcome in a consecutive series of 46 patients with full-thickness rotator cuff tears who underwent non-operative treatment and follow-up for at least 1 year. Treatment consisted only of patient education and a home programme of gentle stretching and strengthening. The patients completed the Simple Shoulder Test at the initial visit and subsequently at 6-month intervals. Almost 60 % of patients experienced improvement, 30 % experienced worsening, and 11 % remained unchanged. The ability to sleep on the affected side and the ability to place the hand behind the head improved significantly. The average score of Simple Shoulder Test functions rose from initially 5.6 ? -3.2 to 7.0 ? -3.8 at the last follow-up. In their multicentre prospective cohort study, Kuhn et al. [7] examined 452 patients with atraumatic full-thickness rotator cuff tears who underwent a physical therapy programme and were evaluated at 6 and 12 weeks. During those visits the patients were divided into 1 of 3 groups: 1, cured (no formal follow-up needed); 2, improved (continued therapy with re-assessment in 6 weeks); 3, no improvement (surgery offered). The dominant arm was affected in 68 % of subjects and the mean age of the study population was 62.6 years. The shoulder function was also assessed using the American Shoulder and Elbow Surgeons score, the Western Ontario Rotator Cuff score, and the Shoulder Activity Scale. Less than 25 % of patients went on to have surgery. Fucentese et al. [13] studied 24 patients with isolated full-thickness supraspinatus tears who were offered rotator cuff repair, but favoured non-operative treatment. Interestingly, refusal of surgery by the patients surprisingly resulted in high clinical patient satisfaction and no increase in the average size of the tear after 5 years. In two shoulders the tear was no longer detectable on magnetic resonance imaging, in nine shoulders the tear was smaller than it had been at the time of initial diagnosis and in nine patients the tear had not changed. Zingg et al. [6] documented 19 consecutive patients with massive rotator cuff tears who were non-operatively managed and followed-up for a mean period of 48 months. The score for pain averaged 11.5 points on a 0to 15-point visual analogue scale in which 15 points represented no pain. The active range of motion did not change over time. The size of the tendon tear increased and fatty infiltration progressed by approximately one stage in all three muscles. Patients with a three-tendon tear showed more rapid progression of osteoarthritis. Four tears graded as reparable at the time of the diagnosis became irreparable at the time of final follow-up.
Discussion The natural course of rotator cuff tears is still not completely understood due to the scarcity of published studies. Better understanding of the various phases of this condition and identification of factors that lead to symptoms is vital in order to establish guidelines for management. The level of evidence of the included papers varied. Case series studies and retrospective uncontrolled studies could not be ignored as they provided a rich foundation for understanding of the mainstream literature. In conjunction with higher-quality papers, case series and uncontrolled studies are crucial for creating a complete picture of the current status of the occurrence of rotator cuff rupture and therapeutic trends. Consequently, however, the purposes and the methodologies of the included articles differed considerably, and it was difficult to find a single existing quality assessment tool which could be applied to all papers. We used Greenhalgh’s article [5] as a guideline and applied its criteria to all the included papers. This article considered six essential questions that should form the basis of every study (Table 1). It assessed the method and design section of each paper, what the study was about, whether the systematic bias was avoided or minimised, and whether the study sample was large enough and continued for long enough to make the results credible. Regression to the mean is a statistical ubiquitous phenomenon in repeated data and should always be considered as a possible cause of an observed change. Its effect can be alleviated through better study design and use of suitable statistical methods. Unfortunately, no clear evidence was found that the studies in this review had investigated the role of regression to the mean of shoulder problems. The impact factor of a journal reflects the frequency with which the journal’s articles are cited in the scientific literature. The use of the impact factor as an index of journal and article quality relies on the theory that citation frequency accurately measures a journal’s importance to its end users. Conceptually developed in the 1960s, the impact factor has gained acceptance as a quantities measure of journal quality. Saha et al. [14] found a strong correlation between impact factor and physicians’ rating of journal quality and therefore the impact factor could be considered as a reasonable indicator of quality for medical journals. Rotator cuff tears are a common occurrence. Degenerative rotator cuff tears are exceedingly rare before 40 years of age, partial-thickness tears usually occur in the sixth decade of life, full-thickness tears in the seventh decade [15]. The onset of shoulder pain in a patient with a known pre-existing asymptomatic tear may indicate an increase in tear size. A few authors reported that patients who developed new shoulder pain showed significant tear size
123
Arch Orthop Trauma Surg
increases and had more severe grades of muscular fatty degeneration compared with baseline assessment, whereas those who remained asymptomatic had no significant change in tear size [9, 16]. Nevertheless, half the patients with symptoms experienced no progression. A study of 120 shoulders with or without symptoms showed that pain was far more closely correlated to subacromial bursitis and long biceps tendinopathy than tear size [17]. It was documented that full-thickness supraspinatus tears in patients under the age of 65 did not necessarily increase in size over a period of 3.5 years. Although the size of the tear increased in 25 % of shoulders, the mean size did not increase significantly [13]. The nature of the patient’s symptoms and expectations are important when making decisions regarding treatment. Most patients usually present with pain as the main complaint. Some surgeons believe that tear size progression and progressive fatty infiltration could cause shoulder disability as well as render the lesion irreparable, therefore they are in favour of early repair of isolated tears. However, it is also known that progression can occur without the development of symptoms [16]; therefore, delaying surgery does not necessarily affect tear reparability. Interestingly, patients who had experienced failed surgical repair of the rotator cuff tears reported satisfaction levels and outcome scores that are almost indistinguishable from those whose repairs were intact [18]. A few studies [6, 19, 20] of non-operative treatment of rotator cuff tears demonstrated satisfactory results regarding functional use of the arm and pain relief at short- to mid-term follow-up but less satisfying results after long-term observation (longer than 6 years). Physical therapy and patient education are vital for effective longterm non-operative treatment of atraumatic full-thickness rupture of the rotator cuff [7, 13].
Limitations Reviews are limited by the quality of the studies included. The methods used by the various authors may have resulted in biased results, particularly if the patient cohorts were small. Most of the studies did not consider regression to the mean. In some of the studies the level of patients’ activity was neither clearly defined nor consistently assessed by means of established scales. There were also some differences regarding the protocols for conservative treatment, the duration of immobilisation either after acute injuries or after surgery, and the length of the follow-up period. It was not possible to address and analyse this clinical heterogeneity within the scope of this systematic review. A further limitation is that in most of the studies it was unclear whether the assessors were blinded or whether an intentionto-treat analysis was performed.
123
Conclusions Rotator cuff tears progress through a number of development stages. The evidence derived from the studies in our review shows that in most cases symptoms and physical findings alone are a reliable indicator of the clinical status. The development of symptoms and anatomical deterioration are often directly correlated. Spontaneous recovery to normal levels of function has been successfully achieved, and standardised non-operative treatment programmes and patients education are an effective alternative to surgery for many patients. However, follow-up is necessary to avoid irreparable stage. Surgery is still favoured by young active people and highly professional persons who need to get fit in a short period of time. Further research is necessary on these two approaches to therapy for rotator cuff tears and other shoulder diseases. The AMQPP score system is simple and reliable. It works as a quick quality-checking tool which helps researchers to identify the key points of each paper. The AMQPP scoring system is open for further development.
References 1. Reilly P, Macleod I, Macfarlane R, Windley J, Emery RJ (2006) Dead men and radiologists don’t lie: a review of cadaveric and radiological studies of rotator cuff tear prevalence. Ann R Coll Surg Engl 88(2):116–121 2. Baydar M, Akalin E et al (2009) The efficacy of conservative treatment in patients with full-thickness rotator cuff tears. Rheumatol Int 29(6):623–628. doi:10.1007/s00296-008-0733-2 Epub 2008 Oct 12 3. Codman (1911) Complete rupture of the supraspinatus tendon. Operative treatment with report of two successful cases. Boston Med Surg J 164:708–710 4. Moher D, Liberati et al (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 339:b2535. doi:10.1136/bmj.b2535 5. Greenhalgh T (1997) Assessing the methodological quality of published papers. BMJ 315(7103):305–308 6. Zingg PO, Jost B et al (2007) Clinical and structural outcomes of nonoperative management of massive rotator cuff tears. J Bone Joint Surg Am 89(9):1928–1934 7. Kuhn JE, Dunn WR et al (2013) Effectiveness of physical therapy in treating atraumatic full-thickness rotator cuff tears: a multicenter prospective cohort study. J Shoulder Elbow Surg 22(10):1371–1379. doi:10.1016/j.jse.2013.01.026 8. DeVries JG, Berlet GC et al (2010) Understanding levels of evidence for scientific communication. Foot Ankle Spec 3(4):205–209. doi:10.1177/1938640010375184 9. Moosmayer S, Tariq R, Stiris M, Smith HJ (2013) The natural history of asymptomatic rotator cuff tears: a three-year follow-up of fifty cases. J Bone Joint Surg Am 95(14):1249–1255. doi:10. 2106/JBJS.L.00185 10. Mall NA, Kim HM et al (2010) Symptomatic progression of asymptomatic rotator cuff tears: a prospective study of clinical
Arch Orthop Trauma Surg
11.
12.
13.
14. 15.
and sonographic variables. J Bone Joint Surg Am 92(16):2623–2633. doi:10.2106/JBJS.I.00506 Safran O, Schroeder J, Bloom R, Weil Y, Milgrom C (2011) Natural history of nonoperatively treated symptomatic rotator cuff tears in patients 60 years old or younger. Am J Sports Med 39(4):710–714. doi:10.1177/0363546510393944 Goldberg BA, Nowinski RJ, Matsen FA 3rd (2001) Outcome of nonoperative management of full-thickness rotator cuff tears. Clin Orthop Relat Res 382:99–107 Fucentese SF, von Roll AL, Pfirrmann CW, Gerber C, Jost B (2012) Evolution of nonoperatively treated symptomatic isolated full-thickness supraspinatus tears. J Bone Joint Surg Am 94(9):801–808. doi:10.2106/JBJS.I.01286 Saha S, Saint S, Christakis DA (2003) Impact factor: a valid measure of journal quality. J Med Libr Assoc 91(1):42–46 Milgrom C, Schaffler M, Gilbert S, van Holsbeeck M (1995) Rotator-cuff changes in asymptomatic adults. The effect of age, hand dominance and gender. J Bone Joint Surg Br 77(2):296–298
16. Yamaguchi K, Tetro AM et al (2001) Natural history of asymptomatic rotator cuff tears: a longitudinal analysis of asymptomatic tears detected sonographically. J Shoulder Elbow Surg 10(3):199–203 17. Brasseur JL, Zeitoun-Eiss D (2005) Ultrasound of acute disorders of the shoulder. JBR-BTR 88(4):193–199 18. Slabaugh MA, Nho SJ et al (2010) Does the literature confirm superior clinical results in radiographically healed rotator cuffs after rotator cuff repair? Arthroscopy 26(3):393–403. doi:10. 1016/j.arthro.2009.07.023 19. Itoi E, Hatakeyama Y et al (2003) A new method of immobilization after traumatic anterior dislocation of the shoulder: a preliminary study. J Shoulder Elbow Surg 12(5):413–415 20. Bokor DJ, Hawkins RJ, Huckell GH, Angelo RL, Schickendantz MS (1993) Results of nonoperative management of full-thickness tears of the rotator cuff. Clin Orthop Relat Res 294:103–110
123