Epidemiology and natural history of rotator cuff disease

Introduction

The management of degenerative rotator cuff disease is controversial, and surgical indications vary widely. With increasing utilization of surgical management for rotator cuff disease, ensuring a complete understanding of the epidemiology and natural history of rotator cuff disease is essential. Defining the natural history of degenerative cuff tears helps to refine surgical indications and provides a baseline to define the potential impact of surgery in altering the natural course of the disease. For degenerative rotator cuff tears, the natural history is specifically related to tear enlargement, progression of muscle degeneration, and the development of cuff tear arthropathy. Establishing an accurate prognosis and patient-specific treatment plan, ^, absent an understanding of the potential for disease progression without intervention, will fail to identify optimal candidates for surgery. This chapter provides a review of the current knowledge regarding the natural history and epidemiology of rotator cuff disease, yet much remains unknown.

To start, it is important to establish distinguishing characteristics of rotator cuff pathology. First, patients with rotator cuff pathology may be asymptomatic or present with pain, weakness, or symptoms of instability. Second, the onset of symptoms associated with rotator cuff disease may have a distinct originating event or be insidious. A distinct origin ranges from simple provocation, such as lifting overhead, to glenohumeral dislocation. More commonly, rotator cuff disease presents with an insidious atraumatic onset. This history is suggestive of degenerative rotator cuff pathology. These processes (traumatic vs. degenerative) are not mutually exclusive and, in many cases, injuries can present in the presence of previously asymptomatic tears. Third, the natural history of rotator cuff disease is also dependent on the disease state. Rotator cuff disease may present as tendinosis, partial-thickness tears, or full-thickness tears. In the setting of a rotator cuff tear, the dimensions of the tear and quality of the remaining tendon/muscle play an important role in the prognosis and natural history. Lastly, rotator cuff pathology cannot be managed without regard for patient-specific factors. Patient age, sex, hand dominance, comorbidities, family history, and social history play a role in the development and/or progression of rotator cuff disease. In the process of history taking, physical exam, and review of radiographic imaging, it is essential that these characteristics are established to develop an appropriate prognosis and treatment plan for each individual.

Epidemiology

Degenerative tears

When discussing the epidemiology of rotator cuff disease, it is important to distinguish traumatic rotator cuff pathology from degenerative. Degenerative rotator cuff pathology likely represents a disease spectrum ranging from cuff tendinopathy, associated with aging, to partial- and full-thickness cuff tears. Multiple factors have been identified as risk factors for atraumatic rotator cuff disease. It cannot be overstated that the strongest predictor is the patient’s age. In a systematic review, Teunis et al. identified 30 separate studies assessing 6112 shoulders for rotator cuff disease in association with patient age. In their analysis, they found that patients younger than 20 years had a 9.7% prevalence of rotator cuff disease compared with 62% of those older than 80 years ( Fig. 51.1 ). This association was maintained independent of shoulder symptoms or history of shoulder dislocation. Minagawa et al., in a large clinical exam and ultrasound-based analysis of a single village, found that the incidence of full-thickness tears increased from 10.7% of those in their 50s to 36.6% of individuals in their 80s. If identified, rotator cuff tears in this sample were more likely to be symptomatic in those younger than 60 years (50%) compared with those older (<33%). Tear size, not just the presence of pathology, may also be dependent upon age. ^, Minagawa et al. found that the proportion of tears classified as large were more common in individuals older than 60 years old. Yamaguchi et al. demonstrated both advancing age and tear severity in subjects with unilateral shoulder pain to be predictive of the presence and severity to cuff disease in the contralateral asymptomatic shoulder. In this series the average age of subjects with no cuff tear on ultrasound was 49 years compared with 59 years in subjects with a unilateral tear and 68 years in subjects with bilateral cuff tears. The chance of having a full-thickness cuff tear in the opposite shoulder was 36% if a full-thickness tear was diagnosed in the painful shoulder. Analysis demonstrated age to be predictive of tear severity in the contralateral shoulder; the mean age of a subject with a 50% chance of a contralateral full-thickness tear when presenting with a painful full-thickness tear in the opposite shoulder was 66 years. In a case-control study of surgically repaired symptomatic rotator cuff tear patients matched to asymptomatic individuals by sex, side, and age, Liem et al. found that 67% of patients with symptomatic rotator cuff tears had a supraspinatus tear in the contralateral shoulder compared with 11% in the control group. However, the majority of the contralateral tears were partial-thickness tears. These studies demonstrate that regardless of age, existence of a symptomatic rotator cuff tear predisposes the patient to a contralateral rotator cuff tear.

The strong association of age-related rotator cuff degeneration combined with the frequent presence of bilateral disease points to a physiologic change in tendon integrity secondary to intrinsic degeneration, likely related to decreased nutritional supply of the tendon. Histologic studies have suggested the presence of a zone of relative hypovascularity of the lateral supraspinatus, ^, particularly at the articular margin, which is the frequent site of degenerative rotator cuff tear initiation. Rudzki demonstrated this using contrast-enhanced ultrasound in 31 rotator cuff intact shoulders. They found that patients older than 40 years had a 47% to 53% decrease in blood flow to the insertion of the supraspinatus. Benson et al. analyzed samples from rotator cuffs with varied levels of disease and found that full-thickness rotator cuff tears were associated with increased evidence of hypoxia. The level of hypoxia in the tendon was also correlated with patient age, regardless of tendon degeneration. Although cadaveric studies have suggested a poor blood supply to the lateral aspect of the cuff, some in vivo studies have contradicted these findings, noting an increase in blood supply at the edge of a full-thickness rotator cuff tear. ^, A recent in vivo study evaluating rotator cuff blood supply in normal and pathologic shoulders using laser Doppler ultrasonography demonstrated the region of greatest tendon blood supply to be at the musculotendinous junction with decreasing flow in the most lateral aspect of the tendon. Although the tendinous blood flow in a normal shoulder was greater than those with cuff disease, there were no differences in regional flow between shoulders with “subacromial impingement syndrome” compared with full-thickness cuff tears.

Patient sex has previously been investigated as a risk factor for rotator cuff disease, with conflicting findings. ^, ^, ^, ^, In their analysis of a single village, Minagawa et al. found that men in their sixth and seventh decade of life had significantly greater rates of full-thickness rotator cuff tears compared with women. This difference faded beyond the seventh decade of life. In contrast, using ultrasound analysis of 90 asymptomatic adults, Milgrom et al. identified no difference in the rate of rotator cuff disease between the sexes. In a separate analysis of a Japanese village, Yamamoto found a higher rate of rotator cuff tears among men (25% vs. 18%); however, gender was not found to be an independent predictor of rotator cuff tears in multivariate analysis.

With evidence emerging of the age-related changes in the rotator cuff, it has been postulated that development of rotator cuff disease is an inheritable condition. Harvie et al. established the familial association of rotator cuff disease by performing ultrasounds on 129 siblings and 150 spouses of patients with known full-thickness rotator cuff tears. They found that siblings were twice as likely to have a full-thickness rotator cuff tear and five times as likely to develop a symptomatic rotator cuff tear. This finding held across all age groups. Five years later, Gwilym et al. revisited this cohort. They found that siblings had greater progression of tear size or tear initiation over the 5-year period. Separately, Tashjian et al. confirmed the familial link of rotator cuff pathology—specifically in those younger than 40 years—using a genealogic database in conjunction with an administrative claims database in Utah. Although it has been established that there is a familial component of rotator cuff pathology, identifying specific predisposing genes has proven more difficult. Two separate studies have linked estrogen-related receptor beta to the development of rotator cuff pathology (both tears and tendinosis). ^, This gene has been associated with cellular adaptation in hypoxic environments. This is an environment, as previously described, common to the rotator cuff tendons in an aging population. In a genome-wide association study, Tashjian et al. identified two single-nucleotide polymorphisms in genes related to cell apoptosis that were associated with rotator cuff tears. However, in an analysis of rotator cuff pathology identified through administrative data, a genome-wide association study by Roos et al. was unable to validate the previous findings.

Hand dominance is often considered in the clinical evaluation of patients with potential rotator cuff pathology. However, in their analysis of 90 asymptomatic adults, Milgrom et al. found no difference in the incidence of rotator cuff pathology based on hand dominance. This is in contrast to Yamamoto et al., who found that the dominant arms of individuals were more likely to have rotator cuff tears than the nondominant arms. In an analysis of 420 asymptomatic volunteers, Moosmayer et al. challenged the high prevalence of asymptomatic rotator cuff tears: reporting a prevalence of only 15% for those in their eighth decade of life while finding that more than 80% of unilateral tears (17 of 21) occurred in the dominant shoulder. Park et al. confirmed this finding with an analysis of more than 600 rural residents by magnetic resonance imaging (MRI). They reported arm dominance to be an independent predictor of rotator cuff tear (partial or full thickness) with an odds ratio of 2. In an analysis of patients with bilateral rotator cuff tears and unilateral pain, Keener et al. found a correlation between hand dominance and the presence of pain. Given this limited evidence, there does appear to be an association between hand dominance and ipsilateral rotator cuff pathology.

Another factor often considered in the pathogenesis of degenerative cuff tears is shoulder activity or occupational level. Patients with repetitive heavy labor jobs are often seen with shoulder pain associated with their work demands. In their single-village analysis, Yamamoto et al. established that patients with more labor-intensive jobs had a higher prevalence of rotator cuff tears prior to adjusting for other covariates. In a similar analysis, Minagawa et al. found that those working in forestry had a significantly higher rate of rotator cuff tears compared with the unemployed (38 vs. 20%). From these two analyses, there may be an association between higher demand work and the development of asymptomatic, degenerative rotator cuff tears.

Scapula morphology has long been associated with the development of rotator cuff pathology. The theory of subacromial impingement causing rotator cuff disease was first proposed by Neer. Early studies seemed to validate this explanation for rotator cuff disease. Bigliani et al. classified acromial morphology and found that 70% of patients with a hooked acromion had a full-thickness rotator cuff tear. Multiple studies have examined acromial morphology in relation to cuff disease; however, it should be emphasized that this classification system suffers from very limited interobserver reliability. Toivonen et al. found that a flat acromion (type I) had no associated rotator cuff tears, half of patients with a curved acromion (type II) had a rotator cuff tear, and 90% of hooked acromion (type III) had a rotator cuff tear. Unfortunately, these early studies did not truly establish causation, only association. In 2002 Gill et al. identified this shortcoming and performed a separate analysis controlling for patient age and sex when assessing the relationship between acromial morphology and rotator cuff disease. They found a strong association between increasing age and the incidence of a hooked acromion. While controlling for age and sex, acromial morphology was still identified as an independent predictor of rotator cuff pathology. The presence of a hooked anterior acromion can be interpreted in some shoulders to be equivalent to the development of an anterior acromial “spur,” essentially ossification of the coracoacromial ligament ( Fig. 51.2 ). Hamid et al. examined various acromial features in a cohort of 216 shoulders with degenerative rotator cuff tears, noting a correlation between the presence of an anterior acromial osteophyte and the presence of a full-thickness tear after controlling for age, sex, and hand dominance. This relationship has also been demonstrated by Ogawa et al. Various acromial characteristics appear to have an association with degenerative cuff disease in some cases; however, given the lack of prospective studies it remains difficult to imply causation.

In recent years, alternate descriptions of acromial characteristics have been proposed to possess importance in the development of rotator cuff disease. Nyffeler et al. proposed the acromial index that is a proportional measure of the lateral extent of the acromion compared with the lateral dimension of the humerus ( Fig. 51.3 ). In this series, a higher acromial index showed a strong correlation with full-thickness rotator cuff tear compared with controls (0.73 ± 0.06 vs. 0.64 ± 0.06). In contrast, Hamid et al. failed to find a relationship between acromial index and the presence of an asymptomatic rotator cuff tear in subjects with a painful cuff tear in the contralateral shoulder or to a control group of subjects with adhesive capsulitis. Moor et al. provided a novel method for assessing the scapular morphology: the critical shoulder angle (CSA). A composite measure of the lateral acromion overhang and glenoid inclination, an increased CSA was found to be strongly associated with the existence of full-thickness rotator cuff disease ( Fig. 51.4 ). In a subsequent study, Moor et al. examined several acromial anatomic characteristics and found the CSA to be most predictive of rotator cuff tears compared with age- and sex-matched controls. This association has been confirmed by multiple other authors, yet causation remains elusive. Chalmers found a significantly larger CSA (24 ± 20) in a cohort of degenerative rotator cuff tears compared with a control group of shoulders with adhesive capsulitis (32 ± 20). However, the CSA did not correlate with the tear size nor the risk of tear enlargement in this cohort at a median of 4 years’ longitudinal follow-up. Furthermore, this study highlights the limitations of this measurement when imperfect radiographs are analyzed. An extrinsic explanation for rotator cuff pathology, namely subacromial impingement of the posterosuperior rotator cuff, neglects the intrinsic pathology associated with aging.

Variables that may hasten the intrinsic degenerative changes associated with rotator cuff tears include modifiable risk factors such as diabetes mellitus, dyslipidemia, smoking, and obesity. Consideration of these factors is also important when identifying appropriate surgery candidates given their potential influence on tendon healing capacity. Diabetes, a microvascular disease that may negatively impact the nutritional supply to the tendon insertion, has been associated with rotator cuff disease in multiple studies. ^, The strongest of these studies, a population-based study of a rural region of Korea, established diabetes as an independent predictor of degenerative rotator cuff tear. Another potential culprit in vascular disease is dyslipidemia. Dyslipidemia has repeatedly been correlated with rotator cuff disease. ^, ^, Using a cohort of patients undergoing shoulder surgery for varied diagnoses, Abboud et al. found that patients with rotator cuff tears had higher levels of triglycerides and low-density lipoprotein compared with those with intact rotator cuffs. Nearly two-thirds of patients with rotator cuff tears had elevated total cholesterol compared with only one-third of patients with intact rotator cuff tendons. In their population survey study, Park et al. found low high-density lipoprotein (good cholesterol) to be an independent predictor of rotator cuff tear. In a separate case-control study comparing patients with arthroscopically repaired full-thickness rotator cuff tears to a control group with asymptomatic shoulders, Djerbi et al. found that dyslipidemia was an independent predictor of rotator cuff tear. They also reported that the presence of dyslipidemia increased the tear severity among those with a rotator cuff tear.

Djerbi et al. also associated another predictor of microvascular disease—smoking—with rotator cuff tear presence and severity, which has been reproduced by other studies. ^, ^, ^, Baumgarten et al. found that there was a time- and dose-dependent relationship between smoking and the presence of partial or full-thickness rotator tear in patients with unilateral shoulder pain. Carbone et al. used a cohort of 408 patients undergoing arthroscopic rotator cuff repair for full-thickness defects to demonstrate that smoking resulted in larger tears at the time of surgical intervention.

Obesity, a marker of malnourishment, has been investigated as a potential risk factor for rotator cuff disease. ^, ^, Gumina et al. compared a cohort of surgical patients with rotator cuff tears with patients with asymptomatic shoulders comparing their body mass index (BMI) and percentage body fat. They found that patients with increased BMI had increased risk of rotator cuff tear that was dose dependent (the higher the BMI, the greater the risk). They also identified that, among those with known rotator cuff tears, an increased BMI increased the risk of a larger tear. Park et al. found that BMI was an independent risk factor for rotator cuff tear when controlling for age, sex, diabetes, dyslipidemia, hand dominance, and manual labor, among other factors. These radiographic studies were confirmed to have clinical significance by Wendelboe et al., who demonstrated that an increasing BMI was a risk factor independent of age for surgical intervention for rotator cuff pathology. These studies establish the modifiable risk factors that contribute to the development of rotator cuff pathology. However, it should be emphasized that the improvement of a modifiable risk factor has not been demonstrated, as of yet, to slow cuff tear progression or improve tendon healing after surgery.

The evidence reviewed here provides a snapshot of the current understanding of the epidemiology of degenerative rotator cuff disease. Rotator cuff disease is highly prevalent and often asymptomatic. The etiology is multifactorial, with some conflicting evidence. For many of the variables presented earlier, a positive finding in one case-control study was not identified in another. This represents the multifactorial nature of rotator cuff disease. Yet again, the strongest predictor of degenerative rotator cuff tears is chronologic age, with obvious genetic and external contributions. The association of trauma with rotator cuff disease is complex and yet to be discussed. The following section reviews the epidemiology associated with traumatic rotator cuff tears. It is important to remember the principals established here, as a history of trauma does not likely lead to rotator cuff injury without the background of rotator cuff degeneration.

Traumatic tears

A history of injury is often provided by patients when presenting with shoulder pain consistent with rotator cuff tear. However, it is unclear the extent to which minor trauma leads to the development of an acute rotator cuff tear versus development of symptoms related to an existing, degenerative rotator cuff tear. In many cases, acute injury with functional decline can occur in the presence of a minimally symptomatic preexisting tear. In their ultrasound survey, Yamamoto et al. found that a history of trauma significantly increased the risk of having a full-thickness rotator cuff tear. Among patients with a rotator cuff tear, more than 7% reported a history of trauma. This is compared with 3% in those with intact rotator cuffs. In addition, the development of traumatic rotator cuff tears is more common in older patients, suggesting tendon degeneration does play a role. Despite this, reports including both traumatic and atraumatic rotator cuff tears have generally found that patients with traumatic rotator cuff tears are younger. ^, This suggests that the development of traumatic rotator cuff tears, like degenerative tears, is multifactorial.

The prevalence of traumatic tears is difficult to quantify given the often-delayed presentation of patients with a history of trauma and subsequently identified rotator cuff tears. Both MRI and ultrasound are capable of distinguishing acute traumatic tears from chronic degenerative tears, ^, and the implications for timely treatment of acute tears are significant. ^, Due to the importance of timely identification of acute rotator cuff tears, Sørensen et al. performed a prospective study of patients presenting with acute shoulder trauma without prior shoulder pain and limited active elevation. They found that nearly 60% of patients included had a partial- or full-thickness (one-third of cases) rotator cuff tear and that the likelihood of rotator cuff pathology was associated with increasing age. In a similarly designed study, Aagaard et al. found that 23% of patients presenting had full-thickness rotator cuff tears, and they estimated an annual incidence of 16 per 100,000 adults. In the setting of shoulder trauma, glenohumeral dislocation is not uncommon. In older patients, a concomitant rotator cuff tear is often identified. When reviewing 87 patients older than 40 years with glenohumeral dislocation, Simank et al. identified a rotator cuff tear in 54% of cases. When isolating the cohort to those older than 70 years, the incidence of rotator cuff tears was 100%. In addition to age, the only other identified demographic risk factor for traumatic rotator cuff tear is male sex. It is unknown if this is a result of higher-risk activities or variation in seeking care.

Existing research has attempted to segregate rotator cuff pathology into acute or chronic tears. Often this is not clinically accurate and neglects patients presenting with acute changes of a chronic rotator cuff tear. Further research is required to accurately quantify this scenario and identify specific risk factors for the development of traumatic rotator cuff tears. There also exists a need for improvement of soft tissue diagnostic imaging to help discern the extent of the acute component of a traumatic cuff tear. When evaluating patients with a traumatic history and acute-onset shoulder pain, it is critical to consider acute rotator cuff pathology as the intrinsic healing capacity of these tears is likely superior to purely degenerative tears. A traumatic rotator cuff tear is less common than degenerative pathology but not infrequent. More importantly, identification and timely treatment of an acute traumatic tear alters the clinical outcomes.

Natural history: Evolution of rotator cuff tears

Presentation

Patients with rotator cuff disease present in varied stages of its natural history. Because most degenerative tears are asymptomatic, there are highly varied differences in symptom onset associated with rotator cuff pathology. Appreciating the natural history of rotator cuff tears requires an understanding of the initial onset of the disease and its typical clinical presentation. Perhaps the most elusive component of defining the natural history of degenerative rotator cuff disease is identification of the factors most important for pain development.

Traumatic rotator cuff tears, as described earlier, are often considered to be rotator cuff tears in shoulders that were pain free prior to a recent trauma. If performed in a timely manner, imaging will demonstrate characteristics unique to an acute rotator cuff tear. Sørensen et al. found that among 47 patients presenting with traumatic rotator cuff tears, 5 (10%) had a full-thickness tear of multiple tendons, 28 (60%) had a full-thickness tear of a single tendon, and 14 (30%) had a partial-thickness tear. They also noted that clinical exam had a low sensitivity (45%) in diagnosing full-thickness rotator cuff tears when evaluating patients with recent trauma. When patients incur a rotator cuff tear due to a glenohumeral dislocation (approximately 10% of dislocations), Robinson et al. found that 14.2% of patients had tear smaller than 1 cm. Meanwhile, 48.2% of patients had a tear between 1 and 3 cm, 25% had a tear between 3 and 5 cm, and 12.3% had a near complete avulsion of the rotator cuff. All of these tears involved the supraspinatus, more than half involved the infraspinatus, and 15.3% were determined to be acute-on-chronic injuries with fatty infiltration.

Without a history of trauma, the development of degenerative tears is often insidious. To understand the clinical significance of degenerative rotator cuff tears, it is necessary to appreciate the pattern with which they develop. As discussed previously, the rotator cuff insertion site within the crescent is a vascular watershed. In a cohort of 360 shoulders, half of which were asymptomatic, Kim et al. demonstrated that the site of initiation does not occur at the anterior supraspinatus as previously described with subacromial impingement. Rather, they found that both partial- and full-thickness tears initiated at the center of the rotator crescent, approximately 1.5 cm posterior to the biceps tendon ( Fig. 51.5 ). This finding lends credence to the rotator cable, and its importance in maintaining function of the posterosuperior rotator cuff despite degenerative tearing within the crescent. In this series, only 30% of full-thickness tears disrupted the anterior cable attachment of the supraspinatus tendon. The development of fatty muscle degeneration of the rotator cuff muscles occurs primarily in the presence of full-thickness rather than partial-thickness cuff tears. In their analysis of symptomatic rotator cuff tears less than 3 cm in size, Namdari et al. found that patients with disrupted anterior cables had greater incidence of supraspinatus fatty atrophy. Patients with disrupted anterior cables also had larger tears but no difference in their clinical presentation. Kim et al. examined 251 shoulders with full-thickness cuff tears, noting the importance of both tear size and location to the development of muscle degeneration. Regression analysis identified disruption of the anterior supraspinatus tendon to be most predictive of supraspinatus muscle degeneration and larger tear size to be most predictive of infraspinatus muscle degeneration ( Fig. 51.6 ).