J Shoulder Elbow Surg (2018) 27, 572–576

[Link]/locate/ymse

Operative versus nonoperative treatment for the

management of full-thickness rotator cuff tears: a
systematic review and meta-analysis
Christine C. Piper, MDa,*, Alice J. Hughes, MDa, Yan Ma, PhDb, Haijun Wang, PhDb,
Andrew S. Neviaser, MDa

a
Department of Orthopaedic Surgery, The George Washington University Hospital Medical Faculty Associates, Washington,
DC, USA
b
Department of Epidemiology and Biostatistics, Milken Institute School of Public Health, The George Washington University,
Washington, DC, USA

Background: Rotator cuff disease is the most common pathology of the shoulder, responsible for ap-
proximately 70% of clinic visits for shoulder pain. However, no consensus exists on the optimal treatment.
The aim of this study was to analyze level I and II research comparing operative versus nonoperative man-
agement of full-thickness rotator cuff tears.
Methods: A literature search was performed, in accordance with the Preferred Reporting Items for Sys-
tematic Reviews and Meta-Analyses (PRISMA) statement, to identify level I and II studies comparing
operative versus nonoperative treatment of rotator cuff tears. Two independent researchers reviewed a total
of 1013 articles. Three studies qualified for inclusion. These included 269 patients with 1-year follow-up.
The mean age ranged from 59 to 65 years. Clinical outcome measures included the Constant score and
visual analog scale (VAS) score for pain. Meta-analysis, using both fixed- and random-effects models,
was performed on pooled results to determine overall significance.
Results: Statistically significant differences favoring surgery were found in both Constant and VAS scores
after 1 year, with mean differences of 5.64 (95% confidence interval, 2.06 to 9.21; P = .002) and −1.08
(95% confidence interval, −1.56 to −0.59; P < .0001), respectively.
Conclusion: There was a statistically significant improvement in outcomes for patients managed operatively
compared with those managed nonoperatively. The differences in both Constant and VAS scores were small
and did not meet the minimal difference considered clinically significant. Larger studies with longer follow-
up are required to determine whether clinical differences between these treatments become evident over time.
Level of evidence: Level II; Meta-Analysis
© 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
Keywords: Rotator cuff tear; atraumatic tears; rotator cuff disease; operative treatment; nonoperative
treatment; conservative management; rotator cuff repair; arthroscopy

This study was exempt from institutional review board review.

Rotator cuff disease is the most common etiology of
*Reprint requests: Christine C. Piper, MD, Department of Orthopaedic
Surgery, The George Washington University Hospital Medical Faculty
shoulder pain, responsible for up to 70% of all shoulder-
Associates, 2300 M Street NW, Fifth Floor, Washington, DC 20037, USA. related visits to physicians.18,21 Rotator cuff tearing is present
E-mail address: pipercc@[Link] (C.C. Piper). in 20% to 54% of persons aged between 60 and 80 years.1,15

1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
[Link]
Meta-analysis of treatment of rotator cuff tears 573

Despite this wide prevalence, controversy exists over the Two independent reviewers (C.C.P. and A.J.H.) screened all ar-
optimal treatment. Physical therapy is widely used for ticles eligible for inclusion. The inclusion criteria were as follows:
atraumatic tears, and several studies have demonstrated its randomized controlled trial, full-thickness rotator cuff tear, and age
reliable and durable success.6,11,25 Treatment with physical 18 years or old. The exclusion criteria included any history of rotator
cuff surgery and a follow-up period of less than 1 year.
therapy does not result in healing of the torn rotator cuff,
Meta-analyses were performed comparing outcomes after op-
however, and natural history studies have raised concerns
erative versus nonoperative treatment of rotator cuff tears. Differences
about tear progression and irreversible fatty infiltration in Constant scores and pain scores (as rated by a visual analog scale
worsening over time.7,22,24 [VAS]) before and after intervention (surgery or physical therapy)
Operative treatment is also a successful treatment option. were selected outcomes measured because they were included in
The widespread use of arthroscopy has corresponded to a sig- all studies. Pooled mean differences were calculated using fixed-
nificant increase in rotator cuff repair procedures in recent and random-effects models.5 We tested the significance of hetero-
decades.4 Operatively treated patients return to work sooner geneity between studies using the Q test and I2 statistic.3,13 Fixed-
and incur less cost burden when compared with patients treated effects models were chosen if the Q test was not significant and I2
nonoperatively.17 Successful outcomes following rotator cuff was low (<20%). Otherwise, random-effects models were applied.
repair do not diminish with midterm and long-term follow-up.8 Forest plots were used for presentation of the mean differences in
outcomes and confidence intervals from individual studies along with
Several randomized controlled trials have compared op-
the pooled mean difference and test for homogeneity.
erative and nonoperative treatment of full-thickness rotator
cuff tears; the results have been mixed. The aim of this study
was to analyze level I and II comparisons of operative versus
Results
nonoperative management of atraumatic rotator cuff tears
through meta-analysis.
The initial database search yielded 1472 abstracts. After
removal of duplicates, 1013 articles remained for review, of
Materials and methods which 5 met the criteria for inclusion in the study. Of these
5 articles, 2 were excluded because they were follow-up studies
This systematic review was conducted in accordance with the
on articles already chosen for review; these patient popula-
Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) statement. 19 One independent reviewer tions could not be considered separately from their original
systematically searched MEDLINE (via Ovid), PubMed (National studies for statistical review purposes and were excluded
Library of Medicine), Scopus (Elsevier, Amsterdam, Nether- (Fig. 1). Thus, 3 studies with a total of 269 patients with 1-year
lands), and the Cochrane Controlled Trials Register (John Wiley & follow-up were included.12,15,20 Patient demographic data and
Sons, Hoboken, NJ, USA) from inception to October 2016. The da- study characteristics are displayed in Table I. All studies had
tabase search was limited to level I and II studies, English- similar follow-up intervals and a minimum of 12 months’
language studies, and human studies. The search strategy applied follow-up.
a combination of MeSH (Medical Subject Headings) and keyword One study included 3 subgroups of patients for analysis,
searches using the following search terms: “rotator cuff injury”; 1 of which underwent physical therapy and subacromial de-
“rotator cuff”; “rotator cuff tear”; “non traumatic tears”; “rotator cuff
compression without rotator cuff repair.15 This cohort of
rupture”; “rotator cuff disease” and “surgical procedures, opera-
patients (57 patients) was excluded. This same study used a
tive”; “general surgery”; “surgery”; “operative treatment”; “non
operative treatment”; “conservative management”; “rotator cuff subscale of the Constant score (scale of 0-15) to measure pain
repair”; “orthopedic procedures”; “surgical procedures, opera- instead of a VAS (scale of 0-10). These pain scores were
tive”; “operative surgical procedures”; “impingement syndrome”; plotted on a graph from which no numerical data could ac-
and “arthroscopy.” The references of selected articles were also re- curately be extracted. This study was not included in the
viewed, when applicable, to identify additional studies. analysis of VAS scores. The Q test was not significant, and

Table I Study characteristics

Kukkonen et al15 Moosmayer et al20 Heerspink et al12
Total patients 110* 103 56
Sex, n
Men 50 73 35
Women 60 30 21
Treatment, n
Nonoperative 55 51 31
Operative 55 52 25
Average age, yr 65 60 60
Follow-up, mo 3, 6, and 12 6 and 12 12
* One-third of patients (n = 57) received physical therapy and subacromial decompression without rotator cuff repair and were excluded from this study.
574 C.C. Piper et al.

Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. Modified from Moher D, Liberati A, Tetzlaff
J, Altman DG, Prisma Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med
2009;6:e1000097. [Link]

Figure 2 Forest plot for Constant scores. SD, standard deviation; MD, mean difference; CI, confidence interval.

I2 equaled 0 for both outcomes, implying low between- scores at 1-year follow-up as compared with the nonoperative
study heterogeneity. Therefore, our results were derived from cohort, with a mean difference in VAS score of −1.08 (95%
mean differences calculated using fixed-effects models (Figs. 2 confidence interval, −1.56 to −0.59; P < .0001).
and 3).
A greater improvement in Constant score was found in op-
erative patients relative to patients treated nonoperatively, and Discussion
this was statistically significant. The mean difference between
operatively treated patients and nonoperatively treated pa- The combined analysis of the 3 randomized controlled trials
tients was 5.64 (95% confidence interval, 2.06-9.21; P = .002). showed a statistically significant benefit in both Constant and
Patients treated operatively had significantly decreased pain VAS scores for the operative cohort compared with the
Meta-analysis of treatment of rotator cuff tears 575

Figure 3 Forest plot for visual analog scale (VAS) scores. SD, standard deviation; MD, mean difference; CI, confidence interval.

nonoperative cohort. However, both values were below the and subscapularis tears in addition to tears of the supraspi-
minimal clinically important differences of 10.4 and 1.4 for natus. Whereas both Moosmayer et al20 and Kukkonen et al15
the Constant and VAS scores, respectively.16,23 The clinical were clear in their exclusion of massive cuff tears, maximum
significance of this statistical advantage is therefore limited. tear size was not explicitly limited by Heerspink et al. Tear
Many studies have evaluated the success of surgical and non- size has consistently been shown to be inversely proportion-
surgical management of small- and medium-sized rotator cuff al to treatment success of tendon repair.10 Although appropriate
tears, but high-level direct comparisons are few. In the short randomization should ensure that this variable is equally dis-
term, there is not a significant clinical advantage to opera- tributed between treatment groups within a single trial, the
tive treatment. Whether longer-term follow-up will demonstrate effect on the intertrial comparison is unknown. A subset of
clinical superiority of either treatment cannot be deter- traumatic rotator cuff tears was also included in the study by
mined from the current data. Two-year follow-up of the patients Moosmayer et al. These patients were not subdivided from
studied by Kukkonen et al15 showed clinical equivalence per- the atraumatic cuff tear group in the analysis and may have
sisted between all treatment groups. Five-year follow-up influenced the findings that more strongly favored surgery than
reported by Moosmayer et al20 showed 37% of patients treated the other trials.
without surgery had tear progression on ultrasound imaging, Lack of uniformity in treatment modalities among the 3
and this correlated with worsening clinical outcome scores. studies was another limitation. Heerspink et al12 used a
Of 52 patients, 12 opted to undergo later surgical repair, but standardized physical therapy protocol, but the other studies
clinically significant differences between the 2 treatment groups did not. Heerspink et al also offered a maximum of 3
were not present at later time points. corticosteroid injections to the control group in addition to
Tear progression remains a concern with nonoperative treat- formal physical therapy sessions. The duration of physical
ment because spontaneous healing of a torn rotator cuff tendon therapy was not controlled and was left to the patient’s
is not thought to occur.14 Tear size may increase in most pa- discretion. Surgical technique also lacked uniformity between
tients over a period of 2 or more years.20,24 Whether the studies. Two of the three studies performed cuff repair via
progression of rotator cuff tearing is a sufficient risk to warrant an open or mini-open approach, whereas the other study
prophylactic tendon repair remains unknown. Although rotator used arthroscopy. These techniques have been shown to be
cuff repair may halt further tear progression, fatty infiltra- equivalent for most tear sizes and are unlikely to substan-
tion, and muscle atrophy, even the most successful repair tially alter outcomes.2 Supplemental procedures such as
cannot reverse pre-existing degenerative changes within the acromioplasty and biceps procedures were left to the discre-
rotator cuff.10 Dunn et al6 showed that patient expectations tion of the surgeon, and their effect on clinical outcomes
are the most significant factor influencing the success of remains a matter of controversy.9
nonoperative treatment, and patients who choose to undergo
surgery after initial nonoperative treatment typically do so in
the first 12 weeks. This would suggest that tear progression
and its clinical sequela, which are unlikely to change during
Conclusion
such a short time, do not affect a patient’s decision for surgery.
On the basis of the results of this meta-analysis, there is
Structural factors such as tear size and retraction were, in fact,
a statistically significant advantage in both objective and
not predictors for choosing surgery in this cohort.
subjective outcomes for patients treated with surgery for
The limitations of this meta-analysis are directly related
a full-thickness rotator cuff tear. However, this statistical
to the inherent characteristics of the 3 studies included. Al-
advantage is not a clinically significant one. The most re-
though these were well-designed trials, there were differences
markable finding of this study is the paucity of high-
among them that made their grouping imperfect. There was
level evidence available to guide treatment of full-
variation in the types of rotator cuff tears included. Kukkonen
thickness rotator cuff tears. Both operative managementand
et al15 limited their inclusion criteria to isolated supraspina-
nonoperative management reliably improve func-
tus tears, whereas Heerspink et al12 included both infraspinatus
576 C.C. Piper et al.

correlate with poor functional outcome. Am J Sports Med 2007;35:719-

tional outcome and pain scores and are reasonable initial 28. [Link]
treatment choices. Longer-term follow-up and larger studies 11. Goldberg BA, Nowinski RJ, Matsen FA III. Outcome of nonoperative
are needed to determine whether the risk of tear progres- management of full-thickness rotator cuff tears. Clin Orthop Relat Res
sion and irreversible degeneration of the rotator cuff are 2001;(382):99-107.
12. Heerspink FO, van Raay JJ, Koorevaar RC, van Eerden PJ, Westerbeek
sufficiently high to warrant early tendon repair. RE, van’t Riet E, et al. Comparing surgical repair with conservative
treatment for degenerative rotator cuff tears: a randomized controlled
trial. J Shoulder Elbow Surg 2015;24:1274-81. [Link]
10.1016/[Link].2015.05.040
Acknowledgment 13. Higgins J, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses [journal article as teaching resource,
The authors acknowledge Gisela Butera, a reference li- deposited by John Flynn]. BMJ 2003;327:557-60. [Link]
10.1136/bmj.327.7414.557
brarian at The George Washington University Himmelfarb 14. Keener JD, Galatz LM, Teefey SA, Middleton WD, Steger-May K,
Health Sciences Library, for her contributions to the lit- Stobbs-Cucchi G, et al. A prospective evaluation of survivorship of
erature search in this study. asymptomatic degenerative rotator cuff tears. J Bone Joint Surg Am
2015;97:89-98. [Link]
15. Kukkonen J, Joukainen A, Lehtinen J, Mattila KT, Tuominen EK, Kauko
T, et al. Treatment of nontraumatic rotator cuff tears. A randomized
controlled trial with two years of clinical and imaging follow-up. J Bone
Disclaimer Joint Surg Am 2015;97:1729-37. [Link]
JBJS.N.01051
The authors, their immediate families, and any research 16. Kukkonen J, Kauko T, Vahlberg T, Joukainen A, Äärimaa V.
foundations with which they are affiliated have not re- Investigating minimal clinically important difference for Constant score
ceived any financial payments or other benefits from any in patients undergoing rotator cuff surgery. J Shoulder Elbow Surg
commercial entity related to the subject of this article. 2013;22:1650-5. [Link]
17. Mather RC, Koenig L, Acevedo D, Dall TM, Gallo P, Romeo A, et al.
The societal and economic value of rotator cuff repair. J Bone Joint Surg
Am 2013;95:1993-2000. [Link]
References 18. Mitchell C, Adebajo A, Hay E, Carr A. Shoulder pain: diagnosis and
management in primary care. BMJ 2005;331:1124-8. [Link]
1. Bartolozzi A, Andreychik D, Ahmad S. Determinants of outcome in the 10.1136/bmj.331.7525.1124
treatment of rotator cuff disease. Clin Orthop Relat Res 1994;(308):90-7. 19. Moher D, Liberati A, Tetzlaff J, Altman DG, Prisma Group. Preferred
2. Bishop J, Klepps S, Lo IK, Bird J, Gladstone JN, Flatow EL. Cuff reporting items for systematic reviews and meta-analyses: the PRISMA
integrity after arthroscopic versus open rotator cuff repair: a prospective statement. PLoS Med 2009;6:e1000097. [Link]
study. J Shoulder Elbow Surg 2006;15:290-9. [Link] pmed.1000097
10.1016/[Link].2005.09.017 20. Moosmayer S, Lund G, Seljom U, Svege I, Hennig T, Tariq R, et al.
3. Cochran WG. The combination of estimates from different experiments. Comparison between surgery and physiotherapy in the treatment of small
Biometrics 1954;10:101-29. and medium-sized tears of the rotator cuff: a randomised controlled study
4. Colvin AC, Egorova N, Harrison AK, Moskowitz A, Flatow EL. National of 103 patients with one-year follow-up. Bone Joint J 2010;92:83-91.
trends in rotator cuff repair. J Bone Joint Surg Am 2012;94:227-33. [Link]
[Link] 21. Rees JL. The pathogenesis and surgical treatment of tears of the rotator
5. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin cuff. Bone Joint J 2008;90:827-32. [Link]
Trials 1986;7:177-88. 90B7.19874
6. Dunn WR, Kuhn JE, Sanders R, An Q, Baumgarten KM, Bishop JY, 22. Safran O, Schroeder J, Bloom R, Weil Y, Milgrom C. Natural history
et al. 2013 Neer Award: predictors of failure of nonoperative treatment of nonoperatively treated symptomatic rotator cuff tears in patients 60
of chronic, symptomatic, full-thickness rotator cuff tears. J Shoulder years old or younger. Am J Sports Med 2011;39:710-4.
Elbow Surg 2016;25:1303-11. [Link] [Link]
7. Fucentese SF, Andreas L, Pfirrmann CW, Gerber C, Jost B. Evolution 23. Tashjian RZ, Deloach J, Porucznik CA, Powell AP. Minimal clinically
of nonoperatively treated symptomatic isolated full-thickness important differences (MCID) and patient acceptable symptomatic state
supraspinatus tears. J Bone Joint Surg Am 2012;94:801-8. (PASS) for visual analog scales (VAS) measuring pain in patients treated
[Link] for rotator cuff disease. J Shoulder Elbow Surg 2009;18:927-32.
8. Galatz LM, Griggs S, Cameron BD, Iannotti JP. Prospective longitudinal [Link]
analysis of postoperative shoulder function. J Bone Joint Surg Am 24. Yamaguchi K, Tetro AM, Blam O, Evanoff BA, Teefey SA, Middleton
2001;83:1052-6. WD. Natural history of asymptomatic rotator cuff tears: a longitudinal
9. Gartsman GM, O’Connor DP. Arthroscopic rotator cuff repair with and analysis of asymptomatic tears detected sonographically. J Shoulder
without arthroscopic subacromial decompression: a prospective, Elbow Surg 2001;10:199-203.
randomized study of one-year outcomes. J Shoulder Elbow Surg 25. Zingg PO, Jost B, Sukthankar A, Buhler M, Pfirrmann CW, Gerber C.
2004;13:424-6. [Link] Clinical and structural outcomes of nonoperative management of massive
10. Gladstone JN, Bishop JY, Lo IK, Flatow EL. Fatty infiltration and rotator cuff tears. J Bone Joint Surg Am 2007;89:1928-34.
atrophy of the rotator cuff do not improve after rotator cuff repair and [Link]