Nonoperative management of rotator cuff disease

Introduction

While descriptions of rotator cuff tears have been noted at least as early as 1788, the first surgical repair of a rotator cuff tear is credited to Muller in 1898, who used cat gut sutures to repair a tear of the supraspinatus and infraspinatus. A variety of techniques were later developed by surgical pioneers in the 19th and 20th centuries, and repair of the torn rotator cuff became the standard of care for patients with symptomatic rotator cuff tears.

Despite the fact that by the 20th century rotator cuff repair had become established dogma for the treatment of rotator cuff tears, there was some early evidence that not all rotator cuff tears required surgery. Ten case series were reviewed in a systematic review by Ainsworth, representing 272 patients. This systematic review concluded that despite surprisingly little information in the literature on physical therapy as a treatment for rotator cuff tears, exercise could potentially be a successful treatment method. There is now ample evidence to support nonoperative treatment for patients with full-thickness rotator cuff tears. This evidence comes from many sources, including prevalence data, the finding that rotator cuff anatomy does not correlate well with patient symptoms, and multiple clinical studies.

Rotator cuff tear prevalence data

Prevalence data would suggest that patients with rotator cuff tears do not require surgery. Cadaveric and imaging studies of the shoulders of patients who are asymptomatic or only minimally symptomatic and those with shoulder pain would suggest that rotator cuff disease, including full-thickness rotator cuff tears, is extremely common. ^, A very conservative estimate is that 10% of adults older than 65 years have full-thickness rotator cuff tears (estimates range from 10% to 40%). ^, According to the 2010 US Census, there are at least 57 million citizens older than 65 years. This would equate to a conservative estimate for a prevalence rate of 5.7 million Americans with rotator cuff tears.

Approximately 275,000 rotator cuff repairs were performed in the United States in 2006. By using this information with the prevalence data above, fewer than 4.8% of people with full-thickness rotator cuff tears have surgery in a given year. The overwhelming majority of people with rotator cuff tears (>95%) are not undergoing surgery.

The number 5.7 million is enormous, and yet it represents a conservative estimate. To gain some perspective on the massive number of patients living with rotator cuff tears, it is interesting to note that there are approximately 25,000 practicing orthopedic surgeons in the United States who performed 5.3 million orthopedic surgeries in 2010. If every patient with a full-thickness rotator cuff tear required surgery, then for more than 1 year, every US orthopedic surgeon’s practice would consist of nothing but repairing rotator cuff tears!

Symptoms do not correlate with rotator cuff tear severity

As a reflection of the demographic data above, it is not surprising that the relationship between symptoms and rotator cuff tear size is not robust. A number of studies have found that pain levels do not correlate with the structural anatomy or the magnitude of the rotator cuff tear. In the Multicenter Orthopaedic Outcomes Network (MOON) Shoulder Cohort, a cross-sectional study of nearly 400 patients with symptomatic, atraumatic full-thickness rotator cuff tears, neither activity level, pain, nor duration of symptoms correlated with the severity of the patient’s rotator cuff tear. Interestingly, emotional health correlated with pain levels and the Single Assessment Numerical Evaluation (SANE) score. Similarly, Wylie et al. demonstrated that there is no correlation between the severity of the rotator cuff tear and the visual analog scale for pain, SF-36 function, Simple Shoulder Test score, or the American Shoulder and Elbow Surgeons (ASES) score. Interestingly, a patient’s mental health status as measured on the SF-36 Mental Health Component has a very strong correlation with these measures. It is not clear if mental health status influences a patient’s decision to have surgery.

Furthermore, if the rotator cuff tear was responsible for the patient’s symptoms, then failed rotator cuff repairs should be accompanied by a recurrence of symptoms. Interestingly, this does not seem to be the case. It is well known that rotator cuff repairs fail at high rates (>20%), yet multiple studies comparing the outcomes of patients whose rotator cuff repairs remain intact to those who have had a failure of surgery demonstrate little difference in the outcomes. In systematic reviews and meta-analyses, patients whose rotator cuff repairs healed had outcome measures that were indistinguishable for those whose repairs failed. The patient essentially does not know his or her repair has failed. Interestingly, in studies where strength is measured, the reviews demonstrate that strength is better if the repair heals. It is important to note that in every trial, patients who had rotator cuff repair underwent some form of postoperative physical therapy, which strongly suggests that the postoperative rehabilitation may have been the effective treatment in providing symptom resolution in patients whose surgery failed.

These data strongly indicate that the patient’s symptoms do not correlate well with the findings of a full-thickness rotator cuff tear on magnetic resonance imaging (MRI) or ultrasound. This disconnect between symptoms and the anatomy may be why there is little consensus regarding treatment and the indications for surgery are not uniform.

Indications for surgery are not clear

Systematic reviews of the literature designed to identify the features to use as indications for surgery for treating rotator cuff tears demonstrate great variability. ^, In general, these reviews suggest that patients with traumatic rotator cuff tears and those with weakness or functional loss may be best suited for surgery. The lack of consensus in the indications for surgery is further reflected in the wide geographic variation noted for rotator cuff repair rates and in the variation in approaches and the belief in the effectiveness of treatment when surgeons are surveyed with different patient scenarios.

Features that predict failure of nonoperative treatment could potentially be used as indications for surgery. The MOON Shoulder Group conducted a multicenter prospective cohort study ^, of 433 patients with atraumatic, symptomatic, full-thickness rotator cuff tears who were all treated with an evidence-based synthesized physical therapy protocol. Patients were followed for 2 years after enrollment and could elect to have surgery at any time. In this cohort, 87 (20%) elected to have surgery. Most patients who elected to have surgery did so within the first 12 weeks after enrollment. Multivariate modeling demonstrated statistically significant associations with the decision to have surgery correlating with nonsmoking and a higher activity level. However, patient expectations regarding the effectiveness of physical therapy were by far the strongest predictor of having surgery. In essence, patients who believe therapy would not work had surgery, and patients who believed therapy would work did not. There was no correlation with having surgery and tear size or retraction, pain, or weakness.

Keener et al. prospectively followed 395 patients who had symptomatic rotator cuff tearing on one shoulder and were asymptomatic with cuff disease on the contralateral side. Of 169 who developed pain in the previously asymptomatic side, 48 (28%) of these underwent surgery. Features that predicted having surgery included earlier onset of the development of pain, greater intensity of pain, a decline in the ASES score, and a history of having surgery in the contralateral symptomatic side. Interestingly, the anatomy of the rotator cuff tear (type, size, or the presence of enlargement of the previously asymptomatic tear) did not predict surgery.

Progression of rotator cuff tears

Patients with shoulder pain and rotator cuff tears who are treated nonoperatively may experience enlargement of the rotator cuff tear over time. Moosmayer et al. followed a cohort of 49 patients treated nonoperatively for full-thickness rotator cuff tears for 8.8 years. On average, the tear progressed by 8.3 mm in the sagittal plane and 4.5 mm in the coronal plane. A subset of this cohort (16%) had more extensive progression (>20 mm), and Constant scores were worse in patients with more extensive progression.

Others have demonstrated that patients with shoulder pain and rotator cuff tears are capable of progressing. Interestingly, not everyone will experience progression, and some smaller rotator cuff tears appear to heal. ^, One recent systematic review designed to compile the reports of rotator cuff tear progression reviewed eight studies of 411 patients and determined that there was no statistical difference in proportion of tears that progress when comparing symptomatic patients to people without symptoms. Overall, tear progression was seen in 40.6% of patients with asymptomatic tears at an average of 46.8 months follow-up and in 34% of symptomatic patients at an average of 37.8 months.

It has been shown in natural history studies of patients with asymptomatic tears that some will enlarge with time, and those that do enlarge are statistically more likely to cause pain. However, as other studies have shown, this relationship is not robust. Patients can have pain without rotator cuff enlargement, and rotator cuff tear enlargement can occur without increases in pain. As a result, the Positive Likelihood Ratio to use pain as a predictor for rotator cuff tear progression is only 1.75; therefore it is difficult to use new onset of pain as a sign of tear progression ( Table 53.1 ). Again, these data seem to emphasize the disconnect between the patient’s symptoms and the imaging finding of a full-thickness rotator cuff tear. Despite the disconnect between symptoms and rotator cuff tear progression, we do know that in some people progression leads to significant problems and may ultimately require reverse shoulder arthroplasty. It would be helpful to know who are at risk for extensive progression of their rotator cuff tears. Although the literature is limited, smoking, age greater than 60 years, larger tears, ^, ^, and more fatty infiltration have been associated with higher rates of rotator cuff tear enlargement. Unfortunately, other studies failed to show a relationship between age and tear size and rotator cuff tear progression. We do not know if surgery can change the natural history of rotator cuff tear progression. When followed for up to 4 years, progression could be seen in up to 40% of patients treated without surgery. It would be useful to know if surgery is capable of changing the natural history. Failure of rotator cuff repairs is high: up to 35% for smaller tears and 94% for large tears. Like the nonsurgically treated rotator cuff tears, the failed rotator cuff repair enlarges over time. Moosmayer reported a 10-year follow-up of 47 patients with small or medium-sized rotator cuff tear repairs. Over time, the number of rotator cuff repair failures increased, with 10 (21%) at 1 year, 13 (28%) at 5 years, and 16 (34%) at 10 years. Interestingly, for the 10 patients with failed rotator cuff repair within the first year, Constant scores remained stable over the 10-year follow-up period. Jost et al. followed 20 patients with failed rotator cuff repairs for a mean of 7.6 years. While 19 of the 20 patients were satisfied and their Constant scores seemed stable over time, 25% of the tears enlarged. Paxon et al. published the 10-year ultrasound evaluations of 11 patients whose large or massive rotator cuff repairs had failed. They reported that the ASES scores remained stable; however, progression seemed to occur in most patients, with only 4 of the 11 having unchanged rotator cuff tear size at 10 years. Interestingly, these data would suggest that failed rotator cuff repairs are at fairly high risk for progression, but patient-reported outcomes do not seem to be affected.

TABLE 53.1

Pain as a Sign of Rotator Cuff Tear Progression

Data from Keener JD, Galatz LM, Teefey SA, et al. A prospective evaluation of survivorship of asymptomatic degenerative rotator cuff tears. J Bone Joint Surg Am . 2015;97(2):89–98.

	Enlargement	No Enlargement	Total
New pain	63	37	100
Asymptomatic	45	74	119
Total	108	111	219

Sensitivity = 0.58; specificity = 0.67; positive predictive value = 0.63; negative predictive value = 0.38; accuracy = 0.62; positive likelihood ratio = 1.75; negative likelihood ratio = 0.62.

To know if surgical repair of smaller tears can change the natural history of rotator cuff tear progression, large datasets of patients who have surgery that can be compared to those who did not have surgery would be required. Because this study has not yet been done, we cannot yet make claims that surgery may change the natural history of rotator cuff disease.

Nonoperative treatment is highly effective in treating symptoms

Nonoperative treatment for full-thickness rotator cuff tear is highly effective. The MOON Shoulder Group, in a prospective cohort study of 452 patients with symptomatic full-thickness rotator cuff tears treated with an evidence-based physical therapy exercise program, found that 80% of patients were treated successfully.

One could argue that physical therapy may not be required and time alone may produce a resolution of the patient’s symptoms; however, Dickinson et al., in a prospective cohort study of 55 patients with rotator cuff tears treated without surgery, demonstrated that patients who did work with a physical therapist had better Shoulder Pain and Disability Index (SPADI) scores at 3 months after diagnosis. A dose effect was noted demonstrating that more than 16 visits with the therapist did not improve SPADI scores.

General components of nonoperative treatment of rotator cuff disease

A number of randomized trials have clearly demonstrated that exercise is effective in treating rotator cuff disease. In order to develop a protocol to use in the MOON Cohort Study, randomized trials were systematically reviewed and the elements of the effective rehabilitation protocols were synthesized. The elements of this protocol are described below.

Modalities

Both heat and cold have shown benefit in treating patients with rotator cuff disease. However, ultrasound has not been shown to be beneficial.

Manual therapy

Joint and soft tissue mobilization techniques may augment the exercise program. While it has been shown that patients may benefit from some supervised physical therapy, at some point home exercises can be effective. The protocol should include supervised physical therapy for a period of time during which manual therapy may be used. The patient will transition to a home exercise program when ready as deemed by the therapist.

Range of motion

Range-of-motion exercises should be performed daily and include working on posture with shrugs and shoulder retraction. Glenohumeral motion begins with pendulum exercises ( Fig. 53.1 ) and progresses as comfort allows to active assisted motion using a cane, pulley system ( Fig. 53.2 ), and/or the uninvolved arm ( Fig. 53.3 ). Active motion begins as comfort allows and can be done in front of a mirror or with the opposite hand on the trapezius to prevent hiking of the shoulder.