1699

COPYRIGHT © 2006 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED

The Demographic and

Morphological Features
of Rotator Cuff Disease
A COMPARISON OF ASYMPTOMATIC AND SYMPTOMATIC SHOULDERS
BY KEN YAMAGUCHI, MD, KONSTANTINOS DITSIOS, MD, WILLIAM D. MIDDLETON, MD,
CHARLES F. HILDEBOLT, PHD, LEESA M. GALATZ, MD, AND SHARLENE A. TEEFEY, MD
Investigation performed at the Department of Orthopaedic Surgery and the Department of Diagnostic Radiology,
MIR Institute of Radiology, Barnes-Jewish Hospital, Washington University School of Medicine, St. Louis, Missouri

Background: Very little comparative information is available regarding the demographic and morphological character-
istics of asymptomatic and symptomatic rotator cuff tears. This information is important to provide insight into the
natural history of rotator cuff disease and to identify which factors may be important in the development of pain. The
purpose of the present study was to compare the morphological characteristics and prevalences of asymptomatic
and symptomatic rotator cuff disease in patients who presented with unilateral shoulder pain.
Methods: Five hundred and eighty-eight consecutive patients in whom a standardized ultrasonographic study had
been performed by an experienced radiologist for the assessment of unilateral shoulder pain were evaluated with re-
gard to the presence and size of rotator cuff tears in each shoulder. The demographic factors that were analyzed in-
cluded age, gender, side, and cuff thickness. All of these factors were evaluated with regard to their correlation with
the presence of pain.
Results: Of the 588 consecutive patients who met the inclusion criteria, 212 had an intact rotator cuff bilaterally,
199 had a unilateral rotator cuff tear (either partial or full thickness), and 177 had a bilateral tear (either partial or
full thickness). The presence of rotator cuff disease was highly correlated with age. The average age was 48.7 years
for patients with no rotator cuff tear, 58.7 years for those with a unilateral tear, and 67.8 years for those with a bilat-
eral tear. Logistic regression analysis indicated a 50% likelihood of a bilateral tear after the age of sixty-six years (p <
0.01). In patients with a bilateral rotator cuff tear in whom one tear was symptomatic and the other tear was asymp-
tomatic, the symptomatic tear was significantly larger (p < 0.01). The average size of a symptomatic tear was 30%
greater than that of an asymptomatic tear. Overall, patients who presented with a full-thickness symptomatic tear had
a 35.5% prevalence of a full-thickness tear on the contralateral side.
Conclusions: There is a high correlation between the onset of rotator cuff tears (either partial or full thickness) and
increasing age. Bilateral rotator cuff disease, either symptomatic or asymptomatic, is common in patients who
present with unilateral symptomatic disease. As the size of a tear appears to be an important factor in the develop-
ment of symptoms, we recommend surveillance at yearly intervals for patients with known rotator cuff tears that are
treated nonoperatively.

tial population at risk for the development of pain1-4. Given

R
otator cuff disease ranks among the most prevalent of
musculoskeletal disorders, with as many as 17 million their clinical importance, surprisingly little comparative infor-
individuals in the United States at risk for disability1,2. mation is available regarding asymptomatic and symptomatic
Although it appears that a vast majority of individuals with rotator cuff tears5-7. Such information may lend insight into
rotator cuff tears are asymptomatic, they represent a substan- the natural history of rotator cuff disease and into which fac-
tors may be important for the development of symptoms6.
The use of shoulder ultrasound has been a primary im-
A commentary is available with the electronic versions of this article,
on our web site (www.jbjs.org) and on our quarterly CD-ROM (call our aging modality at our institution since 19898-10. The technique
subscription department, at 781-449-9780, to order the CD-ROM). has been validated to be highly accurate for both the detection
 1700
THE JOURNAL OF BONE & JOINT SURGER Y · JBJS.ORG T H E D E M O G R A P H I C A N D MO R P H O L O G I C A L
VO L U M E 88-A · N U M B E R 8 · A U G U S T 2006 FE A T U RE S OF ROT ATOR CU F F DI S E AS E

and the characterization of rotator cuff tears11. Additionally, dardized fashion, with partial and full-thickness tears defined
according to protocol, bilateral shoulder examinations have as previously described11. A high-resolution linear-array trans-
been performed routinely. This presented a unique opportu- ducer with the center of frequency ranging from 7 to 10 MHz
nity to compare the findings of a large number of bilateral was used11. The images included longitudinal and transverse
shoulder examinations in patients who presented with unilat- images of the rotator cuff.
eral pain. The purpose of the present study was to compare As a matter of protocol, patients were asked if one or
the morphological and demographic findings associated with both shoulders were considered to be painful, and this in-
asymptomatic and symptomatic rotator cuff disease in pa- formation was recorded. Additionally, patients were asked
tients who presented with unilateral shoulder pain. whether or not they had had a previous operation. The size of
a rotator cuff tear was recorded at the level of the anatomic
Materials and Methods neck in the transverse dimension. Tears that measured ≤15
Clinical mm were considered to involve only the supraspinatus or one
he records for 588 consecutive patients who had had an tendon. Tears that measured >15 mm but ≤30 mm were con-
T ultrasonographic evaluation of both shoulders between
June 1996 and July 2001 for the evaluation of unilateral shoul-
sidered to involve both the supraspinatus and the infraspina-
tus tendon. Tears that measured >30 mm in the transverse
der pain were retrospectively studied. All studies had been dimension were considered to include the supraspinatus, the
performed in real time with use of an ATL HDI 5,000 Scanner infraspinatus, and part of the teres minor and thus were con-
(Advanced Technologies Laboratory, Bothell, Washington). sidered massive.
According to protocol, all patients had undergone standard-
ized bilateral ultrasonography of the shoulder. A full-thickness Data Analysis
rotator cuff tear was recorded when the rotator cuff could not Data on the age of the patient; the thickness of the rotator cuff
be visualized because of complete avulsion and retraction un- (when intact); the presence or absence of an intact rotator
der the acromion or when a focal defect in the rotator cuff was cuff, a partial-thickness tear, or a full-thickness tear; the size of
created by a variable degree of retraction of the torn tendon the tear; and pain were all tested for statistical associations.
edges. If no tear was visualized, the deltoid muscle was com- Specifically, we examined the thickness of the rotator cuff as a
pressed against the cuff with a transducer in an attempt to function of gender, age, and pain. Additionally, we tested the
separate the torn edges at the side of a nonretracted tear. A size of the tear as a function of age and symptoms.
partial-thickness tear was recorded when there was minimal
flattening of the bursal side of the rotator cuff or when a dis- Statistical Analysis
tinct hypoechoic or mixed hypoechoic defect was visualized in For the data on intact rotator cuffs, associations between cuff
both the longitudinal and transverse planes. thickness, gender, age, and pain were tested with the T test, the
The subjects who were involved in the study included paired T test, Fisher’s exact test, regression analysis, and multi-
patients who (1) had undergone an ultrasound examination variate analysis of variance. The quality of variances was tested
of both shoulders for the investigation of unilateral shoulder with the O’Brien, Brown-Forsythe, Levene, and Vartlett tests.
pain and (2) had complete data for both shoulders regarding Testing was performed with use of JMP statistical software
the presence, absence, size, and location of any rotator cuff (SAS Institute, Cary, North Carolina).
abnormality. The exclusion criteria were (1) the presence of For the data on partial or full-thickness rotator cuff
bilateral symptoms, (2) a previous surgical procedure on ei- tears, the associations among the presence of cuff tears, the
ther shoulder, (3) inflammatory arthropathy, and (4) preceding size of cuff tears, pain, and age were tested for significance
trauma. with the T test, the paired T test-Wilcoxon signed-rank test,
The review of medical records was approved by our lo- regression analysis, logistic regression analysis, and multivari-
cal institutional review board. Of the 588 consecutive patients ate analysis of variance. The quality of the variances was tested
who met the inclusion criteria, 212 had an intact rotator cuff with the O’Brien, Brown-Forsythe, Levene, and Vartlett tests.
bilaterally, 199 had a unilateral rotator cuff tear (either partial Tests for the normality of distributions were performed with
or full thickness), and 177 had a bilateral tear (either partial or the Shapiro-Wilk W test. Testing was performed again with
full thickness). use of JMP statistical software (SAS Institute).
In addition to the 588 patients who were included in the
study, there were fifty-two patients who were excluded be- Results
cause the records were incomplete. The majority of these Clinical
records did not state whether the contralateral shoulder was he overall average age (and standard deviation) of the 376
painful or nonpainful. Sixty-six patients were eliminated be-
cause of the presence of bilateral shoulder pain.
T patients with a rotator cuff tear was 62.8 ± 11.3 years. In
this group, significantly more patients (244 [65%] of 376) had
pain on the right side (p < 0.01).
Ultrasonography
The high-resolution ultrasound examinations were performed Unilateral Tears
by two experienced radiologists (S.A.T. and W.D.M.) in a stan- One hundred and ninety-nine patients presented with a uni-
 1701
THE JOURNAL OF BONE & JOINT SURGER Y · JBJS.ORG T H E D E M O G R A P H I C A N D MO R P H O L O G I C A L
VO L U M E 88-A · N U M B E R 8 · A U G U S T 2006 FE A T U RE S OF ROT ATOR CU F F DI S E AS E

Fig. 1
Illustration depicting the prevalence of contralateral rotator cuff disease among patients who
presented with a symptomatic unilateral full-thickness rotator cuff tear. Such patients had a
35.5% chance of having a full-thickness tear on the contralateral, nonpainful side. Only 43.7% of
such patients had normal findings on the contralateral side.

lateral tear. Ninety-seven of these 199 patients had a partial- tomatic) side, a 20.8% (thirty-eight in 183) chance of having a
thickness tear, and 102 had a full-thickness tear. Of the ninety- partial-thickness tear on the contralateral side, and only a
seven partial-thickness tears, ninety-three were painful and 43.7% (eighty in 183) chance of having normal findings on
four were not painful. Of the 102 full-thickness tears, 101 were the contralateral side (Fig. 1).
painful and one was not painful. If a patient presented with a partial-thickness tear on the
painful side, he or she had a 4.3% (six in 140) chance of hav-
Bilateral Tears ing a full-thickness tear on the contralateral (asymptomatic)
One hundred and seventy-seven patients with unilateral pain side, a 29.3% (forty-one in 140) chance of having a partial-
presented with a bilateral tear. Eighty-two of these 177 pa- thickness tear on the contralateral side, and a 66.4% (93 in
tients had a bilateral full-thickness rotator cuff tear. The re- 140) chance of having normal findings on the contralateral
maining ninety-five patients in this group had at least one side.
partial-thickness tear. Among these ninety-five patients, forty- In contrast, if a patient presented with normal findings
one had a bilateral partial-thickness tear, six had a partial- on the painful side, there was very little chance of rotator cuff
thickness tear in the painful shoulder and a full-thickness tear disease on the contralateral (asymptomatic) side. Specifically,
in the nonpainful shoulder, and forty-eight had a full-thickness such a patient had only a 0.5% (one in 217) chance of having a
tear in the painful shoulder and a partial-thickness tear in the full-thickness tear on the contralateral side, a 1.8% (four in
nonpainful shoulder. 217) chance of having a partial-thickness tear on the con-
tralateral side, and a 97.7% (212 in 217) chance of having nor-
Prevalence of Rotator Cuff Tears mal findings on the contralateral side (Fig. 2).
The prevalence of rotator cuff disease increased with age.
There was an approximately ten-year difference between the Correlation with Tear Size
three subgroups of patients: the average age was 48.7 years for The median tear width for asymptomatic shoulders was 15.0
patients with no rotator cuff tear, 58.7 years for those with a mm (indicating involvement of only the supraspinatus ten-
unilateral tear, and 67.8 years for those with a bilateral tear. don). The corresponding mean values were 17.3 ± 10.0 mm
Logistic regression analysis showed a 50% likelihood of a bi- for nonpainful shoulders and 18.8 ± 10.6 mm for painful
lateral tear after the age of sixty-six years (p < 0.01). shoulders (indicating involvement of both the supraspinatus
If a patient presented with a full-thickness tear on the and infraspinatus tendons). Age was not significantly re-
painful side, he or she had a 35.5% (sixty-five in 183) chance lated to the size of the tear in shoulders with or without pain
of having a full-thickness tear on the contralateral (asymp- (r2 < 0.01, p = 0.37). Tear width measurements were available
 1702
THE JOURNAL OF BONE & JOINT SURGER Y · JBJS.ORG T H E D E M O G R A P H I C A N D MO R P H O L O G I C A L
VO L U M E 88-A · N U M B E R 8 · A U G U S T 2006 FE A T U RE S OF ROT ATOR CU F F DI S E AS E

Fig. 2
Illustration depicting the prevalence of contralateral rotator cuff disease among patients who
presented with a symptomatic shoulder without a rotator cuff tear. Such patients had a 1.8%
chance of having a partial-thickness tear on the contralateral, nonpainful side and only a 0.5%
chance of having a full-thickness tear on the contralateral, nonpainful side.

for eighty-two patients who had a bilateral tear. For shoul- Discussion
ders with no pain, the mean and median tear widths were o date, the epidemiology of symptomatic and asymptom-
17.4 ± 10.0 mm and 15.0 mm, respectively. For shoulders
with pain, the mean and median tear widths were 22.7 ± 10.0
T atic rotator cuff tears has been poorly characterized. Most
previous studies on the prevalence of rotator cuff disease were
mm and 23 mm, respectively (indicating involvement of both performed on cadavers, and very little information was
the supraspinatus and infraspinatus tendons). The mean dif- available regarding the presence or absence of pain or other
ference in size between painful and nonpainful shoulders in demographic factors such as age12-19. Given these limitations,
patients with bilateral tears was 5.4 mm (95% confidence in- previous studies of cadavers have demonstrated a potentially
terval, 7.4 to 3.3 mm) (p < 0.01). Age did not significantly large prevalence of rotator cuff tears, ranging from 9.4%
affect this relationship. (eighteen of 192) in the study by Keyes18 to 39% (thirty-nine
of 100) in the study by DePalma et al.13, with an increasing fre-
Shoulders with an Intact Rotator Cuff quency in older individuals12-19. Two prospective studies, one
Two hundred and twelve patients (including 112 female pa- performed with magnetic resonance imaging and the other
tients and 100 male patients) had an intact rotator cuff. The performed with ultrasonography, investigated the prevalence
average age of the female patients was significantly greater of rotator cuff tears in individuals without pain and with ap-
than that of the male patients (50.6 compared with 47.2 parently normal function1,2. Sher et al. reported that thirteen
years; p = 0.03). There was not a difference between male (28%) of forty-six patients who were more than sixty years old
and female patients with regard to the distribution of which had a full-thickness tear2. Milgrom et al., in a study conducted
side was painful. The right side was painful in fifty-nine with ultrasonography, reported that fifteen (65%) of twenty-
(53%) of the 112 female patients and in fifty-four (54%) of three patients who were more than seventy years old had a ro-
the 100 male patients (p = 0.89). In male patients, the mean tator cuff tear1. Correlations with gender, cuff thickness, and
rotator cuff thickness was 4.7 ± 0.8 mm in painful shoulders cuff tear size were unavailable in those studies. Most impor-
and 4.6 ± 0.9 mm in nonpainful shoulders. This small differ- tantly, no direct comparison between painful and nonpainful
ence was not significant (p = 0.34). Multivariate analysis of shoulders was performed.
variance did not identify significant relationships between Despite the relative paucity of demographic information
gender, age, and cuff thickness between shoulders with and in previous studies, it is apparent that a substantial proportion
without pain. of individuals over the age of sixty years have rotator cuff tears.
 1703
THE JOURNAL OF BONE & JOINT SURGER Y · JBJS.ORG T H E D E M O G R A P H I C A N D MO R P H O L O G I C A L
VO L U M E 88-A · N U M B E R 8 · A U G U S T 2006 FE A T U RE S OF ROT ATOR CU F F DI S E AS E

The clinical importance of rotator cuff tears in these apparently functional abilities, which is an important consideration in
asymptomatic individuals is unknown6,7. However, this issue is determining the association of symptoms. The prevalences
quite important because individuals with asymptomatic tears that we found also were specific to a population of patients
represent a very large at-risk population for whom secondary with a contralateral symptomatic shoulder and cannot be gen-
prevention intervention (early detection and treatment of dis- eralized to patients with rotator cuff disease in whom both
ease before it becomes symptomatic)6,7 may be helpful. shoulders are nonpainful. However, the results were very rele-
A previous study from our institution involving a small vant to a substantial at-risk population, namely, patients who
cohort of patients suggested that a large percentage of individu- present with pain and have contralateral nonpainful disease.
als with previously asymptomatic tears are at risk for the devel- The strengths of the present study include the large number of
opment of new-onset pain when seen for the evaluation of patients and the accuracy of the surveillance tool that was
contralateral shoulder pain6. In that study, twenty-three of forty- used (ultrasonography). However, although ultrasonography
five previously asymptomatic individuals became symptomatic is quite accurate for the detection and characterization of full-
over a mean of 2.8 years after the onset of pain in the contralat- thickness tears, it should be noted that it is less so for partial-
eral shoulder. The development of symptoms was associated thickness tears. This lack of accuracy for determining exact
with a significant (p < 0.05) increase in pain and a decrease in size and location precluded further specific analysis of these
the ability to perform activities of daily living. In the present variables.
study, the finding of a strong (50%) likelihood of a bilateral tear Our data demonstrated that specific age-groups are at
after the age of sixty-six years is consistent with an intrinsic eti- a higher risk for the presence of full-thickness rotator cuff
ology for rotator cuff tears associated with natural aging. disease. We believe that patients who undergo nonoperative
The high rate of bilateral disease can be important dur- treatment of symptomatic or asymptomatic full-thickness
ing the workup of an individual who presents with unilateral rotator cuff tears should be monitored for tear size progres-
shoulder pain. Our data showed that 35.5% of the patients sion over time. In our practice, we generally do this by asking
who presented with a full-thickness tear on the painful side a patient to return in six months for repeat ultrasonography.
had a full-thickness tear on the nonpainful, contralateral side. If that study does not show tear size progression, then repeat
This information is particularly important in the context of ultrasound studies are suggested to the patient on a yearly
our previous study, which suggested that a large percentage of basis. The patient is also warned that any new increase in
these tears may be associated with the development of pain pain should prompt a return for a follow-up evaluation, in-
in the future. In contrast, if a patient had either a partial- cluding ultrasonography, to verify that the tear has not got-
thickness tear or a normal rotator cuff on the painful side, the ten larger. Follow-up for such patients is important because
rate of a full-thickness tear on the contralateral, asymptomatic the development of new symptoms may indicate enlargement
side was low (4.3% and 0.5%, respectively). of the tear. „
In our previous longitudinal study of asymptomatic
tears, nine of twenty-three patients had an increase in tear size Ken Yamaguchi, MD
Konstantinos Ditsios, MD
over time6. Although significance could not be detected with
William D. Middleton, MD
the small numbers available, there appeared to be a trend to- Charles F. Hildebolt, PhD
ward a correlation between tear size progression and the de- Leesa M. Galatz, MD
velopment of new symptoms. The results of the present study Sharlene A. Teefey, MD
appear to substantiate this finding. In a direct comparison of Shoulder and Elbow Service, Departments of Orthopaedic Surgery (K.Y.,
the eighty-two patients with a bilateral tear, with a painful tear K.D., and L.M.G.) and Diagnostic Radiology, MIR Institute of Radiology
on one side and a nonpainful tear on the other side, there was (W.D.M., C.F.H., and S.A.T.), Barnes-Jewish Hospital, Washington Uni-
versity School of Medicine, One Barnes Hospital Plaza, Suite 11300 West
a significant difference in tear width associated with symp- Pavilion, St. Louis, MO 63110
toms. On the average, painful shoulders had a 30% larger ro-
tator cuff tear than nonpainful shoulders did. In support of their research for or preparation of this manuscript, one or
The present study had several limitations that are im- more of the authors received grants or outside funding from the Ortho-
portant to recognize. This was a retrospective study, and paedic Research and Education Foundation CDA. None of the authors
therefore a careful prospective analysis of symptoms was not received payments or other benefits or a commitment or agreement to
performed. Although the radiologists asked the patients if provide such benefits from a commercial entity. No commercial entity
paid or directed, or agreed to pay or direct, any benefits to any research
they considered the shoulders to be painful or nonpainful, the fund, foundation, educational institution, or other charitable or non-
answers were highly subjective and were not carefully quanti- profit organization with which the authors are affiliated or associated.
tated with use of visual analog pain scales. In addition, the
presence or absence of pain was not further correlated with doi:10.2106/JBJS.E.00835

References
1. Milgrom C, Schaffler M, Gilbert S, van Holsbeeck M. Rotator-cuff changes in 2. Sher JS, Uribe JW, Posada A, Murphy BJ, Zlatkin MB. Abnormal findings on
asymptomatic adults. The effect of age, hand dominance and gender. J Bone magnetic resonance images of asymptomatic shoulders. J Bone Joint Surg Am.
Joint Surg Br. 1995;77:296-8. 1995;77:10-5.
 1704
THE JOURNAL OF BONE & JOINT SURGER Y · JBJS.ORG T H E D E M O G R A P H I C A N D MO R P H O L O G I C A L
VO L U M E 88-A · N U M B E R 8 · A U G U S T 2006 FE A T U RE S OF ROT ATOR CU F F DI S E AS E

3. Miniaci A, Dowdy PA, Willits KR, Vellet AD. Magnetic resonance imaging evalua- 11. Teefey SA, Hasan SA, Middleton WD, Patel M, Wright RW, Yamaguchi K.
tion of the rotator cuff tendons in the asymptomatic shoulder. Am J Sports Med. Ultrasonography of the rotator cuff. A comparison of ultrasonographic and ar-
1995;23:142-5. throscopic findings in one hundred consecutive cases. J Bone Joint Surg Am.
2000;82:498-504.
4. Neumann CH, Holt RG, Steinbach LS, Jahnke AH Jr, Petersen SA. MR imaging
of the shoulder: appearance of the supraspinatus tendon in asymptomatic volun- 12. Codman EA, Akerson IB. The pathology associated with rupture of the supra-
teers. AJR Am J Roentgenol. 1992;158:1281-7. spinatus tendon. Ann Surg. 1931;93:348-59.
5. Yamaguchi K, Sher JS, Andersen WK, Garretson R, Uribe JW, Hechtman K, Nev- 13. DePalma AF, Callery G, Bennett GA. Variational anatomy in degenerative
iaser RJ. Glenohumeral motion in patients with rotator cuff tears: a comparison of lesions of the shoulder joint. Instr Course Lect. 1949;6:255-81.
asymptomatic and symptomatic shoulders. J Shoulder Elbow Surg. 2000;9:6-11.
14. DePalma AF, White JB, Callery G. Degenerative lesions of the shoulder joint at
6. Yamaguchi K, Tetro AM, Blam O, Evanoff BA, Teefey SA, Middleton WD. Natural various age groups which are compatible with good function. Instr Course Lect.
history of asymptomatic rotator cuff tears: a longitudinal analysis of asymptom- 1950;7:168-80.
atic tears detected sonographically. J Shoulder Elbow Surg. 2001;10:199-203.
15. Fukuda H, Mikasa M, Yamanaka K. Incomplete thickness rotator cuff
7. Yamaguchi K, Lashgari C. Natural history of rotator cuff disorders and non sur- tears diagnosed by subacromial bursography. Clin Orthop Relat Res. 1987;
gical treatment. In: Norris TR, editor. Orthopaedic knowledge update. Shoulder 223:51-8.
and elbow 2. 2nd ed. Rosemont, IL: American Academy of Orthopaedic Surgeons;
2002. p 155-62. 16. Fukuda H, Hamada K, Yamanaka K. Pathology and pathogenesis of bursal-
side rotator cuff tears viewed from en bloc histologic sections. Clin Orthop Relat
8. Middleton WD, Edelstein G, Reinus WR, Melson GL, Murphy WA. Ultra- Res. 1990;254:75-80.
sonography of the rotator cuff: technique and normal anatomy. J Ultrasound
Med. 1984;3:549-51. 17. Grant JCB, Smith GC. Age incidence of rupture of the supraspinatus tendon
[abstract]. Anat Rec. 1948;100:666.
9. Middleton WD, Reinus WR, Totty WG, Melson CL, Murphy WA. Ultrasonographic
evaluation of the rotator cuff and biceps tendon. J Bone Joint Surg Am. 1986; 18. Keyes EL. Anatomic observations on senile changes in the shoulder. J Bone
68:440-50. Joint Surg Am. 1935;17:953-60.

10. Middleton WD. Ultrasonography of the shoulder. Radiol Clin North Am. 19. Meyer AW. Further evidences of attrition in the human body. Am J Anat. 1924;
1992;30:927-40. 34:241-67.