1775

C OPYRIGHT
2019 BY T HE J OURNAL OF B ONE AND J OINT S URGERY, I NCORPORATED

A commentary by Grant L. Jones, MD, is

linked to the online version of this article at
jbjs.org.

Surgical Versus Nonsurgical Management

of Rotator Cuff Tears
A Matched-Pair Analysis
Austin J. Ramme, MD, PhD, Christopher B. Robbins, PhD, Karan A. Patel, MD, James E. Carpenter, MD,
Asheesh Bedi, MD, Joel J. Gagnier, ND, MSc, PhD, and Bruce S. Miller, MD, MS

Investigation performed at The University of Michigan, Ann Arbor, Michigan

Background: Rotator cuff disease is a major medical and economic burden due to a growing aging population, but
management of rotator cuff tears remains controversial. We hypothesized that there is no difference in outcomes between
patients who undergo rotator cuff repair and matched patients treated nonoperatively.
Methods: After institutional review board approval, a prospective cohort of patients over 18 years of age who had a full-thickness
rotator cuff tear seen on magnetic resonance imaging (MRI) were retrospectively evaluated. After clinical evaluation, each patient
elected to undergo either rotator cuff repair or nonsurgical treatment. Demographic information was collected at enrollment,
and self-reported outcome measures (the Normalized Western Ontario Rotator Cuff Index [WORCnorm], American Shoulder and
Elbow Surgeons score [ASES], Single Assessment Numerical Evaluation [SANE], and pain score on a visual analog scale [VAS])
were collected at baseline and at 6, 12, and >24 months. The Functional Comorbidity Index (FCI) was used to assess health
status at enrollment. The size and degree of atrophy of the rotator cuff tear were classified on MRI. Propensity score analysis was
used to create rotator cuff repair and nonsurgical groups matched by age, sex, symptom duration, FCI, tear size, injury mecha-
nism, and atrophy. The Student t test, chi-square test, and regression analysis were used to compare the treatment groups.
Results: One hundred and seven patients in each group were available for analysis after propensity score matching.
There were no differences between the groups with regard to demographics or rotator cuff tear characteristics. For all
outcome measures at the time of final follow-up, the rotator cuff repair group had significantly better outcomes than the
nonsurgical treatment group (p < 0.001). At the time of final follow-up, the mean outcome scores (and 95% confidence
interval) for the surgical repair and nonsurgical treatment groups were, respectively, 81.4 (76.9, 85.9) and 68.8 (63.7,
74.0) for the WORCnorm, 86.1 (82.4, 90.3) and 76.2 (72.4, 80.9) for the ASES, 77.5 (70.6, 82.5) and 66.9 (61.0, 72.2)
for the SANE, and 14.4 (10.2, 20.2) and 27.8 (22.5, 33.5) for the pain VAS. In the longitudinal regression analysis, better
outcomes were independently associated with younger age, shorter symptom duration, and rotator cuff repair.
Conclusions: Patients with a full-thickness rotator cuff tear reported improvement in pain and functional outcome scores
with nonoperative treatment or surgical repair. However, patients who were offered and chose rotator cuff repair reported
greater improvement in outcome scores and reduced pain compared with those who chose nonoperative treatment.
Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

R
otator cuff tears are a common shoulder disorder that patient visits per year and a $3 to $5 billion annual economic
lack a consensus treatment algorithm1-9. It has been burden in the U.S.10,11. With increases in the number of
estimated that rotator cuff tears account for 4.5 million physically active aging persons, rotator cuff disease is quickly

Disclosure: The authors indicated that no external funding was received for any aspect of this work. On the Disclosure of Potential Conflicts of Interest
forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial
relationship in the biomedical arena outside the submitted work and “yes” to indicate that the author had other relationships or activities that could be
perceived to influence, or have the potential to influence, what was written in this work (http://links.lww.com/JBJS/F448).

J Bone Joint Surg Am. 2019;101:1775-82 d http://dx.doi.org/10.2106/JBJS.18.01473

 1776
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

becoming a major medical and economic burden2. Manage- who received operative treatment for a rotator cuff tear with the
ment of full-thickness rotator cuff tears remains controversial outcomes of patients who had nonoperative management. We
because of its variable presentation, ranging from an incidental hypothesized that there would be no difference in outcomes
finding to substantial pain and disability12-14. Advanced patient between these 2 groups.
age and a high body mass index (BMI) are predictors of whether
a patient will be given nonsurgical treatment15. Furthermore, Materials and Methods
patient expectation of the effectiveness of physical therapy is Michigan Shoulder Registry
predictive of the outcome of nonsurgical treatment16. Registry Enrollment
Both surgical repair and nonoperative management are
accepted treatment strategies for patients with full-thickness
rotator cuff tears. Rotator cuff repair is successful in many
T he Michigan Shoulder Registry is an institutional review
board-approved prospective shoulder cohort including
patients with a partial-thickness rotator cuff tear, a full-thickness
patients11,17-19. Nonsurgical treatment of rotator cuff tears is rec- rotator cuff tear, impingement, or tendinosis treated by 3 expe-
ognized as an alternative to surgery with high patient satisfaction rienced shoulder surgeons. Patients with a subscapularis tear,
and good clinical outcomes in selected patients2,3,5,10,20-29. However, glenohumeral osteoarthritis, a history of infection, fracture,
the unpredictability of cuff tear progression, the possibility that a inflammatory arthritis, adhesive capsulitis, or a history of
reparable tear will become irreparable, the burden of postoper- ipsilateral shoulder surgery were not eligible for enrollment.
ative recovery, and the time and cost of unsuccessful nonopera- The cohort includes 597 patients treated from July 2009 until
tive management add complexity to the treatment strategy8,26,30-33. February 2016 and includes demographic and comorbidity
Only a few clinical trials with limited sample sizes have data, magnetic resonance imaging (MRI) measurements, and
directly compared patient outcomes of operative and nonop- patient-reported outcome scores.
erative treatment of rotator cuff tears17,18,34. Propensity score
matching is a statistical method of causal inference that mini- Registry Variables
mizes the effect of confounding bias by balancing covariates in Age, sex, and BMI were recorded at enrollment. The Functional
nonrandomized studies to approximate a randomized con- Comorbidity Index (FCI) was calculated to evaluate the
trolled trial when one is not practical35. In this study, we used patient’s overall health with regard to physical function36,37. The
statistical propensity score matching, after a minimum of 2 index requires self-reporting of 18 chronic conditions ranging
years of follow-up, to compare the clinical outcomes of patients from anxiety to congestive heart failure and has been shown to

Fig. 1
Flow diagram describing patient enrollment, treatment (Tx) allocation, and propensity score analysis.
 1777
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

be predictive of the Short Form-36 (SF-36) physical function

TABLE II Baseline Outcome Scores in Treatment Groups
subscale. The injury mechanism was classified as either known
or unknown. The symptom duration was recorded as £1 year Nonsurgical
or >1 year. Specific patient-reported outcome measure ques- Surgical Repair* Treatment*
tionnaires were collected via mail at the initial visit and at the (N = 107) (N = 107) P Value
6-month, 12-month, and final follow-up (>24 months). These WORCnorm 41.8 (38.3, 45.4) 48.8 (44.5, 53.2) 0.015
measures included:
ASES 49.8 (46.2, 53.5) 58.8 (54.7, 62.9) 0.001
 Normalized Western Ontario Rotator Cuff Index SANE 30.3 (25.6, 35.0) 35.8 (31.1, 40.6) 0.066
(WORCnorm)38,39, a rotator cuff disease-specific quality- Pain VAS 52.9 (47.8, 58.0) 45.3 (40.3, 50.3) 0.048
of-life self-assessment measurement that includes pain
and physical symptoms, sports and recreation, work *Values are shown as the mean (95% CI).
function, social function, and emotional function.
The best raw score is 2,100 points, with larger scores
indicating worse function; however, a normalized  Visual analog scale (VAS) for pain42,43, a measure of
100-point score, with larger scores indicating better pain intensity on a continuum ranging from no pain
function, is often reported. (0) to the worse pain imaginable (100).
 American Shoulder and Elbow Surgeons score
The WORCnorm was the primary outcome in this study,
(ASES)40, which measures shoulder function by self-
while the ASES, SANE, and pain VAS were secondary outcomes.
assessment of pain, disability, and activities of daily
living. The best score is 100 points.
Registry MRI Measurements
 Single Assessment Numerical Evaluation (SANE)41,
The rotator cuff tears were measured on MRI in the anterior-
which is a single patient response regarding the
posterior direction and recorded in 2 groups: £3 cm (small and
percentage of normal function that can be accom-
medium tears) and >3 cm (large and massive tears). Rotator
plished with the affected shoulder. The best score is
cuff atrophy was assessed with tangent sign evaluation, with a
100, with higher scores indicating better patient
positive tangent sign corresponding to substantial rotator cuff
function.
atrophy44. This was measured on the lateral-most image on
which the scapular spine came into contact with the scapula on
TABLE I Descriptive Baseline Characteristics of Treatment T1-weighted sagittal MRI. MRIs were evaluated by 2 blinded
Groups reviewers, and any discrepancy in their findings was resolved by
a third independent reviewer.
Surgical Nonsurgical
Repair* Treatment*
(N = 107) (N = 107) P Value Treatment Groups
Treatment allocation was determined by agreement between the
Age (yr) 61.3 (59.9, 62.8) 61.2 (59.5, 63.0) 0.949
patient and surgeon after clinical evaluation and thorough dis-
Sex 1.000
cussion of the risks and benefits of surgical and nonsurgical
Male 63.6% 63.6%
treatment options for full-thickness rotator cuff tears. Rotator
Female 36.4% 36.4%
cuff repair was performed arthroscopically and was followed by
BMI (kg/m2) 30.2 (29.1, 31.4) 30.2 (29.1, 31.3) 0.977 4 to 6 weeks of immobilization and a standardized postoperative
Injury mechanism 0.088 physical therapy protocol (see Appendix). The repair technique
Known 29.9% 30.8% was not standardized, but tears of £1 cm were repaired using a
Unknown 70.1% 69.2% single-row technique and tears of >1 cm were repaired using a
Symptom duration 1.000 double-row technique. Nonsurgical treatment included a stan-
£1 yr 67.3% 67.3% dardized physical therapy protocol with anti-inflammatory
>1 yr 32.7% 32.7% medication and corticosteroid injections for symptom control
Tear size 0.776 (see Appendix). Physical therapy compliance was established
£3 cm 64.5% 62.6% during follow-up physician visits but was not a registry variable.
>3 cm 35.5% 37.4%
Atrophy (tangent 0.753 Retrospective Treatment Evaluation
sign) From the Michigan Shoulder Registry, we selected patients who
Positive 24.3% 26.2%
were at least 18 years of age and had an MRI diagnosis of a full-
Negative 75.7% 73.8%
thickness rotator cuff tear (Fig. 1). To establish the best pos-
FCI 1.1 (0.8, 1.3) 1.1 (0.9, 1.4) 0.938 sible matching of patients in the surgical repair and nonop-
erative treatment groups, nearest-neighbor without replacement
*Categorical variables are shown as the percentage. Continuous variables
are shown as the mean (95% CI). (caliper = 0.3) propensity score analysis (SPSS, version 25.0;
IBM) was employed with patients matched by factors thought
 1778
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

Fig. 2
Graph depicting the relationship of the WORCnorm score over time to treatment (Tx) group. The scores for the surgical repair and nonsurgical treatment
groups were significantly different at all time points (p < 0.05). Horizontal lines in bars = median, top and bottom of bars = interquartile range, and whiskers =
maximum and minimum data points.

to influence the outcome of rotator cuff tears, including age, effect (0.8 to 1.1), very large effect (1.2 to 1.9), or huge effect
sex, symptom duration, FCI, tear size, injury mechanism, and (‡2.0)45.
rotator cuff muscle atrophy35. Patients who were unmatched by A priori power and sample size analyses were conducted
the propensity score analysis were excluded. using G*Power46. Power and sample size calculations were
based on an expected medium to large effect size (r = 0.5) for
Statistical Evaluation the difference in WORCnorm scores between the surgical repair
Statistical analysis was performed using SPSS version 25.0. and nonsurgical treatment groups that would allow both a
Comparisons between the surgical repair and nonsurgical statistically significant and clinically relevant difference to be
treatment groups for each independent variable of interest were detected. It was determined that, to detect a difference at an a
carried out with the Student t test for continuous or interval level of 0.05 and maintain a power of 0.80, it would be nec-
variables or the Mann-Whitney U test when normality was not essary to obtain 74 patients per group.
shown. The chi-square test was used to assess differences
between categorical variables. Four models were conducted Results
and analyzed using a random effects repeated-measures mixed
linear regression to examine the predictive effect of demo-
graphic and clinical factors on each outcome of interest
O f the 597 patients in the registry, 422 were evaluated,
including 200 with nonsurgical treatment and 222 with
surgical repair (Fig. 1). After propensity score analysis, 107
(WORCnorm, ASES, SANE, and pain VAS) at baseline and patients per group were matched. No patients crossed between
longitudinally. Additionally, the Cohen d effect size was cal- groups during the study period, and there were no patient-
culated and was classified as a very small effect (0.01 to 0.19), reported complications. The average times to the last follow-up
small effect (0.2 to 0.49), medium effect (0.5 to 0.79), large (and 95% confidence interval [CI]) were 2.35 (2.11, 2.59) and

TABLE III Outcome Scores at Final Follow-up (>24 Months) in Treatment Groups

Surgical Repair* (N = 107) Nonsurgical Treatment* (N = 107) P Value Cohen d (Effect Size)

WORCnorm 81.4 (76.9, 85.9) 68.8 (63.7, 74.0) <0.001 0.56

ASES 86.1 (82.4, 90.3) 76.2 (72.4, 80.9) <0.001 0.52
SANE 77.5 (70.6, 82.5) 66.9 (61.0, 72.2) <0.001 0.43
Pain VAS 14.4 (10.2, 20.2) 27.8 (22.5, 33.5) <0.001 0.56

*Values are shown as the mean (95% CI).

 1779
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

The minimal clinically important difference (MCID) for

TABLE IV Difference Between Treatment Groups with Regard
to Change in Outcome Scores Over Time*
the WORCnorm score has been reported to be between 245 and
282.6, which corresponds to a WORCnorm score between 11.6
Surgical and 13.547,48. The differences seen in the present study were
Repair 2 Nonsurgical Cohen d clinically relevant as the changes in the WORCnorm score
Treatment† P Value (Effect Size)
between baseline and final follow-up were 39.6 and 20.0 in the
D WORCnorm 19.6 (14.0, 28.9) <0.001 0.856 surgical repair and nonoperative treatment groups, respec-
D ASES 18.9 (11.9, 26.4) <0.001 0.794 tively. In the surgical repair group, 91% to 92% of the patients
reached the MCID of 11.6 to 13.5, 2% to 3% had improvement
D SANE 16.1 (6.5, 25.9) 0.001 0.497
but did not reach the MCID, and 6% did not have improve-
D Pain VAS 21.0 (12.6, 31.4) <0.001 0.695
ment. In the nonoperative treatment group, the respective
values were 58% to 59%, 19% to 20%, and 22%.
*Final follow-up – baseline. †Values are shown as the mean
(95% CI). The MCID for the ASES score has been reported to be
between 9 and 21.948,49. In the surgical repair group, the change
in the ASES score from baseline to final follow-up was 36.2,
2.30 (2.04, 2.56) years for the surgical repair and nonsurgical which surpassed the MCID. In the nonsurgical treatment
treatment groups, respectively. There were no significant dif- group, the change in the ASES score from baseline to final
ferences in the patient or injury characteristics between the 2 follow-up was 17.4, which has been reported to be clinically
treatment groups (Table I). The 2 groups differed signifi- important in some studies but not others48,49. In the surgical
cantly with regard to the baseline WORCnorm (p = 0.015), ASES repair group, 77% to 89% of the patients reached the MCID of
(p = 0.001), and pain VAS (p = 0.048) scores, with the non- 9 to 21.9 for the ASES, 5% to 17% had improvement but did
surgical treatment group having higher baseline functional not reach the MCID, and 6% did not have improvement. In the
outcome scores and lower pain VAS scores (Table II). nonoperative treatment group, the respective values were 40%
Significant differences in the primary outcome to 64%, 16% to 40%, and 20%.
(WORCnorm) were found between the surgical repair and Regression analysis of each outcome score demonstrated
nonsurgical treatment groups at all time points (Fig. 2), with that higher baseline FCIs corresponded to worse baseline
mean WORCnorm scores (and 95% CI) of 81.4 (76.9, 85.9) and outcome scores (WORCnorm, ASES, and pain VAS) (Table V).
68.8 (63.7, 74.0), respectively, at the final time point (>24 Additionally, male sex was predictive of higher ASES scores at
months) (Table III). All outcome scores were significantly better baseline. In the longitudinal analysis, better outcome scores were
in the surgical repair group compared with the nonsurgical independently associated with younger age, shorter symptom
treatment group (p < 0.001). The WORCnorm, ASES, and pain duration, and surgical repair.
VAS scores had a medium effect size, ranging from 0.52 to 0.56,
while the SANE score had a small effect size of 0.43. Discussion
The change in the WORCnorm score from baseline to final
follow-up differed between the surgical repair and nonsurgical
treatment groups by a mean (and 95% CI) of 19.6 (14.0, 28.9)
P atients with rotator cuff repair had better patient-reported
outcome scores and a greater improvement in outcomes
compared with those who underwent nonsurgical treatment in
(Table IV). The change in the WORCnorm, ASES, SANE, and a study in which patients were propensity score matched for
pain VAS scores differed significantly between the treatment age, sex, FCI, symptom duration, tear size, injury mechanism,
groups (p £ 0.001), with surgical repair resulting in greater and atrophy. Both the surgical repair group and the nonsurgical
improvement compared with nonoperative treatment. The change treatment group demonstrated improved functional outcomes
in the WORCnorm score had a large effect size, and the change in and reduced pain compared with baseline values, but the final
the ASES, SANE, and pain VAS scores had a medium effect size. outcome scores and the magnitude of the improvement were

TABLE V Results of Regression Analysis Evaluating Effect of Study Variables on Each Outcome Score

WORCnorm ASES SANE Pain VAS

B (95% CI) P Value B (95% CI) P Value B (95% CI) P Value B (95% CI) P Value

Baseline
Male sex 7.0 (21.0, 15.0) 0.87 7.6 (0.1, 15.2) 0.048 22.1 (211.9, 7.8) 0.680 21.7 (210.9, 7.4) 0.712
FCI 24.8 (27.7, 21.8) 0.002 24.5 (27.3, 21.7) 0.002 21.9 (25.6, 1.7) 0.302 4.4 (1.0, 7.8) 0.011
Longitudinal
Time · age 0.1 (0.1, 0.2) 0.002 0.1 (0.1, 0.2) 0.001 0.1 (0.1, 0.2) 0.002 20.1 (20.2, 20.1) 0.043
Time · symptom duration 3.4 (0.4, 6.4) 0.028 3.2 (0.4, 6.0) 0.028 4.2 (0.4, 7.9) 0.030 23.1 (26.5, 0.4) 0.083
Time · surgery 3.7 (2.5, 5.0) <0.001 2.7 (1.5, 3.8) <0.001 4.0 (2.5, 5.5) <0.001 23.9 (25.3, 22.5) <0.001
 1780
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

greater in the surgical repair group. Additionally, our longitu- higher reported pain level than our nonsurgical treatment
dinal regression analysis showed that younger age, shorter group at baseline, which may suggest that these patients were
symptom duration, and surgical repair were independently more likely to choose surgery after an informed conversation
predictive of improved outcomes over time. with a surgeon. We used propensity score analysis to approx-
Previous studies comparing operative and nonopera- imate a randomized controlled trial to minimize the effects of
tive management of rotator cuff tears have had conflicting confounding bias while maintaining a realistic method of
results. Lambers Heerspink et al. found that patients who had treatment allocation.
undergone surgical repair had significant improvements in This study was a retrospective review of a prospectively
visual analog pain and disability scores at a 1-year follow-up, collected patient cohort, which makes it susceptible to the
but no difference in functional outcomes measured with the limitations of a retrospective study. Although the shoulder
Constant-Murley score, when compared with nonoperatively registry captures many important variables, it is not all-
treated patients17. Moosmayer et al. found that, at 1-year encompassing. Items that were not included in the registry
follow-up, patients who had been treated with surgical repair included physical examination findings, biceps tendon
had significant improvements in functional outcomes mea- pathology, rotator cuff retraction, and psychosocial factors.
sured with the ASES, improved patient satisfaction, and Although we evaluated patient-reported outcome scores,
improved pain VAS scores, but no difference in SF-36 quality- there may be a difference in the physical examination find-
of-life scores, when compared with nonoperatively treated ings that individual patients consider acceptable. Biceps
patients18. Kukkonen et al. found no difference in 2-year tendon involvement, rotator cuff retraction, and psychoso-
Constant scores, pain VAS scores, or binary patient satisfac- cial factors could offer additional comparisons in future
tion scores between surgical repair and nonoperatively treated studies51. Although we evaluated rotator cuff atrophy, we did
groups34. Piper et al. performed a meta-analysis of random- not isolate traumatic rotator cuff tears, which may differ
ized controlled trials of operatively and nonoperatively trea- from chronic tears with regard to outcomes. Authors of future
ted full-thickness rotator cuff tears50. They found statistically studies should consider incorporating patient satisfaction. It
significant but not clinically important improvement in out- is possible that patients could be highly satisfied with their
comes for the operatively treated patients compared with the treatment and yet have inferior patient-reported outcome
nonoperatively treated patients. scores and vice versa.
This study was designed with the aim of improving The method of surgical repair was not standardized at the
on the methodology of previous studies. The power analysis in inception of the registry. Although our surgeons use similar
this study ensured that our sample size was large enough to arthroscopic techniques, we cannot guarantee that all proce-
perform a statistical comparison of surgical repair and nonop- dures were comparable with regard to the number of sutures
erative treatment of rotator cuff tears. Authors of previous and anchors. It is unclear if this influenced our surgical out-
studies did not perform a power analysis and sample size cal- comes. Furthermore, an MRI was not acquired at the time of
culation17. We excluded patients with concomitant diagnoses final follow-up as part of the registry, which precludes us from
thought to independently affect patient outcomes. In previous making statements regarding the status of the rotator cuff
studies, 38% to 47% of included patients had known gleno- repair at the time of final follow-up.
humeral osteoarthritis, which introduces a potential con-
founding factor affecting patient outcomes34. In our study, we Conclusions
incorporated patient comorbidities by using the FCI in the Patients with full-thickness rotator cuff tears reported
analysis. In previous studies34, the investigators did not con- improvement in pain and functional outcome scores after
sider patient comorbidity when comparing outcome scores nonoperative treatment or surgical repair. However, patients
between patients. This is problematic as patients with many who were offered and elected to undergo rotator cuff repair
comorbidities may have lower outcome scores than healthy reported greater improvement in outcome scores and
patients34. Propensity score matching allowed us to evaluate reduced pain compared with those who chose nonoperative
all rotator cuff tear sizes and all degrees of atrophy by bal- treatment.
ancing these variables in the 2 treatment groups. In previous
studies, the analysis was limited to specific tear sizes and Appendix
degrees of muscle atrophy, which limits their applicability to Supporting material provided by the authors is posted
patients with similar rotator cuff tears18. with the online version of this article as a data supplement
Our study was not a randomized controlled trial, which at jbjs.org (http://links.lww.com/JBJS/F449). n
may be a limitation. Treatment allocation was based on surgeon
equipoise and shared decision-making, which is how surgical
decisions are made. Selection bias in choosing surgical candi-
dates is appropriate and important to achieve optimal patient Austin J. Ramme, MD, PhD1
outcomes. The aim of randomized controlled trials is to re- Christopher B. Robbins, PhD1
move selection bias, but this is not realistic. It should be noted Karan A. Patel, MD1
that our surgical repair group had lower functional scores and a James E. Carpenter, MD1
 1781
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

Asheesh Bedi, MD1 Email address for B.S. Miller: [email protected]

Joel J. Gagnier, ND, MSc, PhD1,2
Bruce S. Miller, MD, MS1 ORCID iD for A.J. Ramme: 0000-0002-6081-1089
ORCID iD for C.B. Robbins: 0000-0001-9263-3865
1Department of Orthopaedic Surgery, The University of Michigan, ORCID iD for K.A. Patel: 0000-0003-1107-5094
Ann Arbor, Michigan ORCID iD for J.E. Carpenter: 0000-0001-9310-8018
ORCID iD for A. Bedi: 0000-0001-8926-7139
2Department of Epidemiology, School of Public Health, The University of ORCID iD for J.J. Gagnier: 0000-0002-3162-3935
Michigan, Ann Arbor, Michigan ORCID iD for B.S. Miller: 0000-0002-8695-5988

References
1. Kahlenberg CA, Dare DM, Dines JS. Further research is needed to define the 19. Ruotolo C, Nottage WM. Surgical and nonsurgical management of rotator cuff
benefits of non-operative rotator cuff treatment. HSS J. 2016 Oct;12(3):291-4. Epub tears. Arthroscopy. 2002 May-Jun;18(5):527-31.
2016 Feb 29. 20. Boorman RS, More KD, Hollinshead RM, Wiley JP, Mohtadi NG, Lo IKY, Brett KR.
2. Seida JC, Schouten JR, Mousavi SS, Tjosvold L, Vandermeer B, Milne A, Bond K, What happens to patients when we do not repair their cuff tears? Five-year rotator
Hartling L, LeBlanc C, Sheps DM. Comparative effectiveness of nonoperative and cuff quality-of-life index outcomes following nonoperative treatment of patients with
operative treatments for rotator cuff tears. AHRQ Comparative Effectiveness full-thickness rotator cuff tears. J Shoulder Elbow Surg. 2018 Mar;27(3):444-8.
Reviews. 2010 Jul. 21. Agout C, Berhouet J, Spiry C, Bonnevialle N, Joudet T, Favard L; French Arthro-
3. Pegreffi F, Paladini P, Campi F, Porcellini G. Conservative management of rotator scopic Society. Functional outcomes after non-operative treatment of irreparable
cuff tear. Sports Med Arthrosc Rev. 2011 Dec;19(4):348-53. massive rotator cuff tears: Prospective multicenter study in 68 patients. Orthop
4. Makhni EC, Steinhaus ME, Morrow ZS, Jobin CM, Verma NN, Cole BJ, Bach BR Jr. Traumatol Surg Res. 2018 Dec;104(8S):S189-92. Epub 2018 Aug 2.
Outcomes assessment in rotator cuff pathology: what are we measuring? J Shoulder 22. Bokor DJ, Hawkins RJ, Huckell GH, Angelo RL, Schickendantz MS. Results of
Elbow Surg. 2015 Dec;24(12):2008-15. Epub 2015 Oct 21. nonoperative management of full-thickness tears of the rotator cuff. Clin Orthop
5. Longo UG, Franceschi F, Berton A, Maffulli N, Droena V. Conservative treatment Relat Res. 1993 Sep;(294):103-10.
and rotator cuff tear progression. Med Sport Sci. 2012;57:90-9. Epub 2011 Oct 4. 23. Goldberg BA, Nowinski RJ, Matsen FA 3rd. Outcome of nonoperative management
6. Mantone JK, Burkhead WZ Jr, Noonan J Jr. Nonoperative treatment of rotator cuff of full-thickness rotator cuff tears. Clin Orthop Relat Res. 2001 Jan;(382):99-107.
tears. Orthop Clin North Am. 2000 Apr;31(2):295-311. 24. Zingg PO, Jost B, Sukthankar A, Buhler M, Pfirrmann CW, Gerber C. Clinical and
7. Varkey DT, Patterson BM, Creighton RA, Spang JT, Kamath GV. Initial medical structural outcomes of nonoperative management of massive rotator cuff tears. J
management of rotator cuff tears: a demographic analysis of surgical and nonsur- Bone Joint Surg Am. 2007 Sep;89(9):1928-34.
gical treatment in the United States Medicare population. J Shoulder Elbow Surg. 25. Tashjian RZ, Deloach J, Green A, Porucznik CA, Powell AP. Minimal clinically
2016 Dec;25(12):e378-85. Epub 2016 Aug 2. important differences in ASES and simple shoulder test scores after nonoperative
8. Reider B. To cut … or not? Am J Sports Med. 2015 Oct;43(10):2365-7. treatment of rotator cuff disease. J Bone Joint Surg Am. 2010 Feb;92(2):296-303.
9. Cowan JB, Bedi A, Carpenter JE, Robbins CB, Gagnier JJ, Miller BS. Evaluation of 26. Fucentese SF, von Roll AL, Pfirrmann CW, Gerber C, Jost B. Evolution of non-
American Academy of Orthopaedic Surgeons Appropriate Use Criteria for the man- operatively treated symptomatic isolated full-thickness supraspinatus tears. J Bone
agement of full-thickness rotator cuff tears. J Shoulder Elbow Surg. 2016 Jul;25(7): Joint Surg Am. 2012 May 2;94(9):801-8.
1100-6. Epub 2016 Feb 26. 27. Khatri C, Ahmed I, Parsons H, Smith NA, Lawrence TM, Modi CS, Drew SJ,
10. Baumer TG, Chan D, Mende V, Dischler J, Zauel R, van Holsbeeck M, Siegal DS, Bhabra G, Parsons NR, Underwood M, Metcalfe AJ. The natural history of full-
Divine G, Moutzouros V, Bey MJ. Effects of rotator cuff pathology and physical thickness rotator cuff tears in randomized controlled trials: a systematic review and
therapy on in vivo shoulder motion and clinical outcomes in patients with a symp- meta-analysis. Am J Sports Med. 2019 Jun;47(7):1734-43. Epub 2018 Jul 2.
tomatic full-thickness rotator cuff tear. Orthop J Sports Med. 2016 Sep 26;4(9): 28. Collin PG, Gain S, Nguyen Huu F, Lädermann A. Is rehabilitation effective in
2325967116666506. massive rotator cuff tears? Orthop Traumatol Surg Res. 2015 Jun;101(4)(Suppl):
11. Aleem AW, Brophy RH. Outcomes of rotator cuff surgery: what does the evidence S203-5. Epub 2015 Apr 15.
tell us? Clin Sports Med. 2012 Oct;31(4):665-74. 29. Kuhn JE, Dunn WR, Sanders R, An Q, Baumgarten KM, Bishop JY, Brophy RH,
12. Watson S, Allen B, Robbins C, Bedi A, Gagnier JJ, Miller B. Does the rotator cuff Carey JL, Holloway BG, Jones GL, Ma CB, Marx RG, McCarty EC, Poddar SK, Smith
tear pattern influence clinical outcomes after surgical repair? Orthop J Sports Med. MV, Spencer EE, Vidal AF, Wolf BR, Wright RW; MOON Shoulder Group. Effectiveness
2018 Mar 27;6(3):2325967118763107. of physical therapy in treating atraumatic full-thickness rotator cuff tears: a multi-
13. Brophy RH, Dunn WR, Kuhn JE; MOON Shoulder Group. Shoulder activity level is center prospective cohort study. J Shoulder Elbow Surg. 2013 Oct;22(10):1371-9.
not associated with the severity of symptomatic, atraumatic rotator cuff tears in Epub 2013 Mar 27.
patients electing nonoperative treatment. Am J Sports Med. 2014 May;42(5): 30. Maman E, Harris C, White L, Tomlinson G, Shashank M, Boynton E. Outcome of
1150-4. Epub 2014 Mar 21. nonoperative treatment of symptomatic rotator cuff tears monitored by magnetic
14. Unruh KP, Kuhn JE, Sanders R, An Q, Baumgarten KM, Bishop JY, Brophy RH, resonance imaging. J Bone Joint Surg Am. 2009 Aug;91(8):1898-906.
Carey JL, Holloway BG, Jones GL, Ma BC, Marx RG, McCarty EC, Poddar SK, Smith 31. Hsu J, Keener JD. Natural history of rotator cuff disease and implications on
MV, Spencer EE, Vidal AF, Wolf BR, Wright RW, Dunn WR; MOON Shoulder Group. The management. Oper Tech Orthop. 2015 Mar 1;25(1):2-9.
duration of symptoms does not correlate with rotator cuff tear severity or other patient- 32. Eljabu W, Klinger HM, von Knoch M. The natural history of rotator cuff tears: a
related features: a cross-sectional study of patients with atraumatic, full-thickness systematic review. Arch Orthop Trauma Surg. 2015 Aug;135(8):1055-61. Epub
rotator cuff tears. J Shoulder Elbow Surg. 2014 Jul;23(7):1052-8. Epub 2014 Jan 8. 2015 May 6.
15. Kweon C, Gagnier JJ, Robbins CB, Bedi A, Carpenter JE, Miller BS. Surgical 33. Tashjian RZ. Epidemiology, natural history, and indications for treatment of
versus nonsurgical management of rotator cuff tears: predictors of treatment allo- rotator cuff tears. Clin Sports Med. 2012 Oct;31(4):589-604. Epub 2012 Aug 30.
cation. Am J Sports Med. 2015 Oct;43(10):2368-72. Epub 2015 Aug 12. 34. Kukkonen J, Joukainen A, Lehtinen J, Mattila KT, Tuominen EK, Kauko T,
16. Dunn WR, Kuhn JE, Sanders R, An Q, Baumgarten KM, Bishop JY, Brophy RH, Äärimaa V. Treatment of nontraumatic rotator cuff tears: a randomized controlled
Carey JL, Harrell F, Holloway BG, Jones GL, Ma CB, Marx RG, McCarty EC, Poddar SK, trial with two years of clinical and imaging follow-up. J Bone Joint Surg Am. 2015 Nov
Smith MV, Spencer EE, Vidal AF, Wolf BR, Wright RW; MOON Shoulder Group. 2013 4;97(21):1729-37.
Neer Award: predictors of failure of nonoperative treatment of chronic, symptomatic, 35. Austin PC. An introduction to propensity score methods for reducing the effects
full-thickness rotator cuff tears. J Shoulder Elbow Surg. 2016 Aug;25(8):1303-11. of confounding in observational studies. Multivariate Behav Res. 2011 May;46(3):
17. Lambers Heerspink FO, van Raay JJ, Koorevaar RC, van Eerden PJ, Westerbeek 399-424. Epub 2011 Jun 8.
RE, van ’t Riet E, van den Akker-Scheek I, Diercks RL. Comparing surgical repair with 36. Gagnier JJ, Allen B, Watson S, Robbins CB, Bedi A, Carpenter JE, Miller BS. Do
conservative treatment for degenerative rotator cuff tears: a randomized controlled medical comorbidities affect outcomes in patients with rotator cuff tears? Orthop J
trial. J Shoulder Elbow Surg. 2015 Aug;24(8):1274-81. Sports Med. 2017 Aug 21;5(8):2325967117723834.
18. Moosmayer S, Lund G, Seljom U, Svege I, Hennig T, Tariq R, Smith HJ. Com- 37. Groll DL, To T, Bombardier C, Wright JG. The development of a comorbidity index
parison between surgery and physiotherapy in the treatment of small and medium- with physical function as the outcome. J Clin Epidemiol. 2005 Jun;58(6):595-602.
sized tears of the rotator cuff: a randomised controlled study of 103 patients with 38. de Witte PB, Henseler JF, Nagels J, Vliet Vlieland TP, Nelissen RG. The Western
one-year follow-up. J Bone Joint Surg Br. 2010 Jan;92(1):83-91. Ontario rotator cuff index in rotator cuff disease patients: a comprehensive reliability
 1782
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
SU RGIC AL VE RSUS NON SU RGIC AL MANAGEME NT OF R O TAT O R C U F F
V O LU M E 101 -A N U M B E R 19 O C T O B E R 2, 2 019
d d
TEARS

and responsiveness validation study. Am J Sports Med. 2012 Jul;40(7):1611-9. 46. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power
Epub 2012 May 10. analysis program for the social, behavioral, and biomedical sciences. Behav Res
39. Kirkley A, Alvarez C, Griffin S. The development and evaluation of a disease- Methods. 2007 May;39(2):175-91.
specific quality-of-life questionnaire for disorders of the rotator cuff: The Western 47. Gagnier JJ, Robbins C, Bedi A, Carpenter JE, Miller BS. Establishing minimally
Ontario Rotator Cuff Index. Clin J Sport Med. 2003 Mar;13(2):84-92. important differences for the American Shoulder and Elbow Surgeons score and the
40. Michener LA, McClure PW, Sennett BJ. American Shoulder and Elbow Surgeons Western Ontario Rotator Cuff Index in patients with full-thickness rotator cuff tears. J
Standardized Shoulder Assessment Form, patient self-report section: reliability, Shoulder Elbow Surg. 2018 May;27(5):e160-6. Epub 2018 Jan 4.
validity, and responsiveness. J Shoulder Elbow Surg. 2002 Nov-Dec;11(6):587-94. 48. Kirkley A, Griffin S, Dainty K. Scoring systems for the functional assessment of
41. Cunningham G, Lädermann A, Denard PJ, Kherad O, Burkhart SS. Correlation the shoulder. Arthroscopy. 2003 Dec;19(10):1109-20.
between American Shoulder and Elbow Surgeons and Single Assessment Numerical 49. Werner BC, Chang B, Nguyen JT, Dines DM, Gulotta LV. What change in Amer-
Evaluation Score After Rotator Cuff or SLAP Repair. Arthroscopy. 2015 Sep;31(9): ican Shoulder and Elbow Surgeons Score represents a clinically important change
1688-92. Epub 2015 Apr 22. after shoulder arthroplasty? Clin Orthop Relat Res. 2016 Dec;474(12):2672-81.
42. Sindhu BS, Shechtman O, Tuckey L. Validity, reliability, and responsiveness of a Epub 2016 Jul 8.
digital version of the visual analog scale. J Hand Ther. 2011 Oct-Dec;24(4):356-63, 50. Piper CC, Hughes AJ, Ma Y, Wang H, Neviaser AS. Operative versus nonoperative
quiz 364. Epub 2011 Aug 6. treatment for the management of full-thickness rotator cuff tears: a systematic review
43. Huskisson EC. Measurement of pain. Lancet. 1974 Nov 9;2(7889):1127-31. and meta-analysis. J Shoulder Elbow Surg. 2018 Mar;27(3):572-6. Epub 2017 Nov 21.
44. Zanetti M, Gerber C, Hodler J. Quantitative assessment of the muscles of the 51. Ravindra A, Barlow JD, Jones GL, Bishop JY. A prospective evaluation of pre-
rotator cuff with magnetic resonance imaging. Invest Radiol. 1998 Mar;33(3):163-70. dictors of pain after arthroscopic rotator cuff repair: psychosocial factors have a
45. Sawilowsky S. New effect size rules of thumb. J Mod Appl Stat Methods. 2009; stronger association than structural factors. J Shoulder Elbow Surg. 2018 Oct;
8(2):597-9. 27(10):1824-9. Epub 2018 Aug 16.