Secondary Logo

Journal Logo


Subacromial Space Measurement

A Reliable Method Indicating Fatty Infiltration in Patients with Rheumatoid Arthritis

van de Sande, M A. J, MD; Stoel, B C, PhD; Rozing, P M, MD, PhD

Author Information
Clinical Orthopaedics and Related Research®: October 2006 - Volume 451 - Issue - p 73-79
doi: 10.1097/01.blo.0000229294.06475.41
  • Free


The severity of rotator cuff abnormalities (fatty infiltration of the rotator cuff muscles and rotator cuff tears) has a negative effect on the outcome of shoulder surgery.7,11,15,16,19,26 Fatty infiltration and rotator cuff tears lead to a greater risk of surgical failure and are associated with inferior functional results after tendon repair and shoulder arthroplasty.7,11,15 Some authors suggest fatty infiltration does not diminish after rotator cuff repair and it seems irreversible.15,28 This implies the importance of timely diagnosis and treatment.1

Magnetic resonance arthrography (MRA) is considered the gold standard in diagnosing rotator cuff tears; however, tears also can be diagnosed using musculoskeletal magnetic resonance imaging (MRI), ultrasound, or computed tomography arthrography (CTA).8,20 These diagnostic tools, including MRA, all are relatively expensive, time consuming, depend on the radiologist's experience, or require invasive techniques, such as contrast injection or radiation, which makes them less practical in screening and followup studies. Simple measurement of subacromial space narrowing based on plain shoulder radiographs to determine supraspinatus tendon ruptures have been described.3,12,31 Proximal migration of the humeral head has been indicative of a rotator cuff tear, but not diagnostic.4,5,9,18,23 Goutallier et al presumed a relationship between fatty infiltration of the rotator cuff muscles and proximal migration of the humeral head.15 This was supported by Nove-Josserand et al, who reported a strong correlation between decreased subacromial space and fatty infiltration or the presence of rotator cuff tears.25 However, using an absolute measure for the acromiohumeral interval and a visual score for the amount of fatty infiltration24 made it difficult to present a reliable cut-off value indicating rotator cuff disease. One study showed a relative measure (dividing the distance from the center of the humeral head to the undersurface of the acromion by the radius of the humeral head18) proved much more accurate and reliable in measuring the subacromial space, compared with the absolute measure of the acromion-humeral interval.29 Assessing fatty infiltration has been described using a visual score according to Goutallier et al.14 However, their score is only moderately reproducible and requires experienced observers, and separately evaluating each muscle is less reliable.15,24 We introduced a quantitative technique using CT images to calculate the mean muscle density as a measure of fatty infiltration.30 This technique produced a greater interobserver reliability compared with the score of Goutallier et al.

Our first hypothesis was that fatty infiltration of the rotator cuff muscles causes a decrease of the acromiohumeral interval, more so than other patient-related factors such as age, shoulder complaints, or the presence of a rotator cuff tear. The secondary hypothesis was that fatty infiltration of the infraspinatus and teres minor muscles (shoulder depression) primarily is responsible for proximal migration of the humeral head. The third hypothesis was that a cut-off-point for proximal migration can be used to screen for rotator cuff abnormalities such as fatty infiltration.


In patients with rheumatoid arthritis and shoulder complaints we obtained an anteroposterior (AP) radiograph, CT scan, and ultrasound of both shoulders were done to assess the relationship between proximal migration, fatty infiltration, and the presence of a rotator cuff tear. A power analysis suggested a sample size of 59 shoulders would achieve 81% power using an F test to detect a slope of 0.07 when the standard deviation of the amount of fatty infiltration is 0.60, the standard deviation of proximal migration is 0.12, and the significance level is 0.05. The actual standard deviations for both measurements were much smaller (0.45 and 0.07), therefore fewer shoulders were needed to achieve an equal power.

Between January 2003 and July 2004, we included 29 consecutive patients with rheumatoid arthritis (54 shoulders). Patients were included after their treating physician ordered bilateral AP radiographs to assess shoulder complaints. Final inclusion was based on the following criteria: (1) clinically diagnosed with rheumatoid arthritis (RA) according to the American Rheumatism Association criteria 1987 as having rotator cuff disease (eg, rotator cuff tear and fatty infiltration) seen commonly in shoulders with RA20; (2) older than 50 years (the age limit was chosen to impose the smallest risk from radiation exposure [effective dose, 1.6 mSv] [European Union (EU) guidelines]); (3) shoulder symptoms in at least one shoulder; and (4) no previous shoulder trauma or surgery. The study had prior institutional review board approval. All patients were informed and provided signed informed consent. There were six men and 23 women with an average age of 63 years (range, 50-81 years). The mean Constant and Murley score for pain and function was 22 points (95% confidence interval [CI], 9-35 points).2 Forty-four of the 54 shoulders were symptomatic (objective pain and loss of function) with a mean Constant and Murley score of 20. Ten shoulders were asymptomatic (mean Constant and Murley score for pain and function, 28 points). The mean interval between the diagnosis of RA and the CT scan was 14 years (range, 1-40 years).

Standard protocol AP radiographs were taken of all patients who were in the supine position and slightly turned to the image side (20°), with the arm in external rotation and palm facing forward.23 The film focus distance was measured at 115 cm, and a15° craniocaudal tilt was used to project the undersurface of the acromion perpendicular. This created a true AP projection 90° toward the shoulder (Fig 1). All radiographs were taken in a clinical setting in the presence of the principal investigator (MS) who controlled for image quality and positioning.29

Fig 1
Fig 1:
The proximal migration measurement using standard AP radiographs shows moderate proximal migration of the humeral head. A = undersurface acromion, C = humeral head center; R = humeral head radius

Proximal migration was measured using the upward migration index.18 The distance between the center of the humeral head to the undersurface of the acromion was divided by the radius of the humeral head (Fig 1). A manual circle fit was used to determine the humeral head center. An upward migration index of 1.0 indicated severe proximal migration, and an upward migration index greater than 1.26 was considered normal.18 Previous research comparing CT images with AP radiographs suggested measuring the subacromial space on AP radiographs (controlled for positioning, scaling, and individual differences by using the upward migration index) was an accurate indicator for proximal migration.29 All radiographs also were scored (WRO) for progression of rheumatoid disease using the score described by Larsen et al.21

Patients' shoulders were scanned with a 16-slice CT scanner (Aquilion, Toshiba, Tokyo, Japan) using a standard protocol and calibration technique. The scanning parameters were 120 kVp, 125 mAs, 250-mm field of view, and a detector pitch of 15. A reconstruction filter (FC12) and raster artifact suppression (RASP) were used to produce a 512 × 512-matrix slice thickness of 1 mm (slice overlap, 0.5 mm). Consecutive multiplanar reconstruction images were calculated with a 0.5-mm thickness in the parasagittal plane parallel to the glenohumeral joint space. The parasagittal images were evaluated from the most lateral section on which the spine of the scapula was in contact with the body of the scapula (Fig 2).10,30

Fig 2
Fig 2:
A parasagittal CT scan shows the regions of interest for the supraspinatus, infraspinatus/teres minor, and subscapularis muscles (SSp, ISp/TMI, SSc). The supraspinatus and infraspinatus muscles show Grade 2 fatty infiltration according to Goutallier et al. White arrows; TMa = teres major

The mean muscle density, a quantitative measure for fatty infiltration, was calculated using software developed by our division of image processing.27 The muscles were outlined manually while carefully excluding pixels containing subcutaneous and intermuscular fat (Fig 2). The teres minor and infraspinatus were analyzed together, as separating these muscles is difficult and unreliable.32 All pixels containing bone tissue were excluded automatically from the segmentation by applying a threshold value of 200 HU.

A histogram was constructed from all pixels in the outlined region of interest to calculate the mean muscle density of the rotator cuff muscles. The mean muscle density was defined as the mean CT number in one outlined rotator cuff muscle. To correct for individual muscle/fat content, the mean muscle density was divided by the patient's body mass index (BMI).13 Because no correlation was found between BMI and proximal shoulder migration (correlation coefficient, 0.01), proximal migration was not standardized for BMI.

Two blinded observers rated the rotator cuff muscles for fatty infiltration using the Goutallier score; conflicts in agreement had to be resolved (WRO and PMR).14 This score qualitatively rates the rotator cuff muscles for fatty infiltration using a score from zero to four (0, no fat; 1, some strands of fat; 2, less fat than muscle; 3, as much fat as muscle; and 4, more fat than muscle). Mean muscle density for all rotator cuff muscles was subdivided for shoulder complaints.

All shoulders were screened for rotator cuff disease by an experienced musculoskeletal radiologist (JTL) using ultrasound (Table 1). All rotator cuff muscles were screened for the presence of tendinitis, a small tear, or a massive tear using standard ultrasonic methods.8

Ultrasound Assessment for Rotator Cuff Abnormalities

A general linear model of univariate analysis was used to assess the effects of progression of RA, the presence of shoulder complaints, gender, and age on the subacromial space. Linear regression analysis and Pearson's and Spearman's correlations were used to evaluate the relation between proximal migration, the presence of a rotator cuff tear, and fatty infiltration. Analysis of variance (ANOVA) was used to assess the differences in the mean muscle density between the different patient groups. To adjust for muscle density and duration of RA and other variables, the partial correlation coefficient between the upward migration index and mean muscle density controlled for age, gender, progression of rheumatoid disease, and rotator cuff disease diagnosed by ultrasound was calculated. A Student's t test was used for detecting differences among the groups. All analyses were performed using SPSS for Windows (Version 11.05, SPSS Inc, Chicago, IL). Significance was set at p < 0.05.


Symptomatic shoulders had smaller (p = 0.0003) mean muscle density when compared with asymptomatic shoulders (Table 2). The mean muscle density of the rotator cuff muscles also was smaller (p < 0.001) when a rotator cuff tear in the supraspinatus and/or infraspinatus was present (Table 2).

Mean Muscle Density of the Rotator Cuff

Increase of fatty infiltration of the rotator cuff muscles was related to a decreased acromiohumeral interval as the upward migration index showed a correlation (r = 0.86; p < 0.0001) with the mean muscle density of the rotator cuff muscles (Table 3). Fatty infiltration of the rotator cuff muscles accounted for 73% (r2 = 0.73) of the variation in proximal migration of the humeral head. Rotator cuff abnormalities diagnosed by ultrasound (tendinitis, small tear, large tear) showed a weak correlation (r2 = 0.12; p = 0.01) with proximal migration of the humeral head. (Tables 1 and 3). A rotator cuff tear in the infraspinatus tendon did show a weak (r = −0.4; p = 0.003) correlation with proximal migration (Table 3).

Correlation Coefficients with Upward Migration Index

The presence of a rotator cuff tear correlated (r = 0.63) with the mean muscle density of the rotator cuff muscles (Table 4). There also was a probable relationship (r = −0.25) between the duration of RA and proximal migration (Table 3). The partial correlation coefficient between the upward migration index and mean muscle density controlled for age, gender, progression of rheumatoid disease, and rotator cuff disease remained (r = 0.77; p < 0.0001) (Table 5). A general linear model revealed no confounders for the relation between fatty infiltration and the upward migration index.

Pearson's Correlation Coefficients
Partial Correlation Coefficients

The mean muscle density of the infraspinatus/teres minor muscles correlated (r = 0.86) with the upward migration index and the mean supraspinatus density (r = 0.78). The mean muscle density of the infraspinatus/teres minor muscles was the best predictor (standardized coefficient beta, 0.85; p < 0.001; F = 137.3; p < 0.001) for proximal migration. Introducing both parameters, the partial correlation coefficient (r) for the infraspinatus/teres minor muscles decreased from 0.85 to 0.62, and for the supraspinatus to 0.37; beta decreased to 0.61 for the infraspinatus/teres minor muscles and to 0.31 for the supraspinatus. This showed a stronger (r = 0.62) relationship between the mean muscle density of the infraspinatus/teres minor muscles with proximal migration of the humeral head compared with the supraspinatus (r = 0.37). Fatty infiltration may weaken the depressing muscle force generated by the infraspinatus/teres minor muscles, which allows for proximal migration of the humeral head. Suspected loss of interposition of the supraspinatus also may play an important role, yet severe proximal migration was found only in patients with supraspinatus tears and fatty infiltration of the infraspinatus and teres minor muscles (Fig 3).

Fig 3
Fig 3:
A scatterplot shows the upward migration index as a function of the mean muscle density of the rotator cuff muscles. The linear regression line (mean muscle density = 1.08 + 0.14 * upward migration index; p < 0.0001) indicates a very strong relation between both measures. The mean muscle densities of the rotator cuff muscles for the three proposed groups are: high risk group = 0.81; medium risk group = 1.74; and small risk group = 2.05 (p < 0.01).□ No rotator cuff tear▾Rotator cuff tear- Preliminary cut-off points-95% confidence interval… Preliminary cut-off pointsSevere fatty infiltration (Goutallier Grades 3-4)▪Mild fatty infiltration (Goutallier Grade 2)No fatty infiltration (Goutallier Grades 0-1)

We propose three preliminary groups to discriminate between the different degrees of fatty infiltration. A large amount of fatty infiltration was indicated by an upward migration index less than 1.25, a medium amount by an upward migration index of 1.25 to 1.35, and a small amount or no fatty infiltration by an upward migration index greater than 1.35. A subdivision was made in the Goutallier score (0-1/2/3-4), resulting in three severities of fatty infiltration.15 These three groups fit in our proposed groups (Fig 3). The mean muscle density of the rotator cuff muscles differed (p < 0.01) among the three groups (Fig 3).


The relationship between decreased subacromial space and rotator cuff abnormality has been assumed.15 We sought to determine whether measuring the subacromial space on a standard AP radiograph is a reliable indicator for rotator cuff abnormality, and to present a cut-off value for proximal migration indicating a high suspicion of rotator cuff disease.

A standard protocol should be used to acquire the AP radiographs, and caution must be used in measuring upward migration when suboptimal AP radiographs are used. All radiographs should be fluoroscopically controlled for positioning to allow for easier correction and control. Also a relative measure is advocated to assess the subacromial space. We observed a greater accuracy in measuring the subacromial space using a relative measure (upward migration index) compared with the absolute acromion-humeral distance.29 The mean absolute difference between the upward migration index measured on AP radiographs and CT images was only 0.06 (standard deviation [SD], 0.07). The mean difference for the acromion-humeral distance on CT images and AP radiographs was 2.6 mm (SD, 2.1).29 Therefore, we think radiographs can be used reliably to screen for rotator cuff disease, but MRI, CTA, or ultra-sound should be used to assess the disease in greater detail.

Patient positioning (eg, upright-supine) might influence the subacromial space, and therefore, should be kept constant. We found no studies comparing supine and upright positioning in radiograph measurements. To keep this variable constant, all patients were scanned while in the supine position and all radiographs were taken with the patients in the same position.

Dinnes et al reported the pooled sensitivity of ultra-sound in diagnosing full and partial-thickness tears was 0.87 and 0.67, respectively. Results for the more invasive MRA were 0.95 and 0.93, respectively.6 Therefore, the relationship between rotator cuff tears and fatty infiltration may be less precise than we would have hoped. Conversely, we think the correlation between the mean muscle density of the rotator cuff muscles and upward migration index, although not totally independent of the presence of rotator cuff tears, is strong enough to explain a major part of the proximal migration.

The choice for patients with RA in assessment of rotator cuff tears, fatty infiltration, and proximal migration, allows for a wide distribution of abnormalities in a relatively small group of patients. However, the etiology of rotator cuff tears and fatty infiltration in this patient group might differ from another patient group with rotator cuff tears or impingement.17 Therefore generalization of our results to these patient groups might be restricted.

Our results were in concordance with results in other studies of rotator cuff abnormalities and proximal migration.15,25,31 Authors of these studies reported multitendon tears and/or severe fatty infiltration of the rotator cuff muscles were associated with decreased subacromial space. Proximal migration of the humeral head was compared with fatty infiltration or the presence of rotator cuff tears in just one study.25 The results for subacromial space (8.6 versus 8.4 mm) and fatty infiltration measurement (Mean Goutallier score, 2) in patients with rotator cuff tears were identical to results in our patient group.25,29 Nove-Josserand et al also described a clear relationship between subacromial space measurement and fatty infiltration of the infraspinatus, and a relation between rotator cuff tears and fatty infiltration. However, they did not report a reliable value that could be used to determine the critical degree of proximal migration indicating rotator cuff disease. Our study showed fatty infiltration of the infraspinatus and teres minor muscles is the most important factor associated with proximal migration. We also presented two cut-off-values indicating healthy and abnormal rotator cuff muscles.

The proposed lower cut-off point discriminating between severe and mild fatty infiltration coincides with the mean upward migration index in healthy (unaffected) shoulders (upward migration index = 1.26) described by Hirooka et al18 and Lehtinen et al.22 We think a second cut-off point discriminating between healthy and mildly affected rotator cuff muscles will make it easier to screen for rotator cuff disease using the upward migration index. A third cut-off point eventually may be introduced dividing the mildly affected shoulders into two groups.

When treating rotator cuff tears or placing shoulder prostheses, measuring the proximal migration of the humeral head using the upward migration index provides a reliable screening method indicating fatty infiltration of the rotator cuff. It provides valuable information for surgical planning and functional prognosis. Additional research regarding clinical implications of fatty infiltration is needed to evaluate and improve its diagnostic value.


We thank Wim R. Obermann, MD, PhD, (WRO), and John G. S. Tjong a Lieng, MD, (JTL), Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands, for radio-graph, CT, and ultrasound evaluations.


1. Codman EA. The Shoulder. 2nd ed. Boston, MA: Thomas Todd; 1934.
2. Constant CR, Murley AH. A clinical method of functional assessment of the shoulder. Clin Orthop Relat Res. 1987;214:160-164.
3. Cotton RE, Rideout DF. Tears of the humeral rotator cuff: a radiological and pathological necropsy survey. J Bone Joint Surg Br. 1964;46:314-328.
4. Cruess RL. Rheumatoid arthritis of the shoulder. Orthop Clin North Am. 1980;11:333-342.
5. Cuomo F, Greller MJ, Zuckerman JD. The rheumatoid shoulder. Rheum Dis Clin North Am. 1998;24:67-82.
6. Dinnes J, Loveman E, McIntyre L, Waugh N. The effectiveness of diagnostic tests for the assessment of shoulder pain due to soft tissue disorders: a systematic review. Health Technol Assess. 2003;7: 1-166.
7. Edwards TB, Boulahia A, Kempf JF, Boileau P, Nemoz C, Walch G. The influence of rotator cuff disease on the results of shoulder arthroplasty for primary osteoarthritis: results of a multicenter study. J Bone Joint Surg Am. 2002;84:2240-2248.
8. Farin PU, Kaukanen E, Jaroma H, Vaatainen U, Miettinen H, Soimakallio S. Site and size of rotator-cuff tear: findings at ultrasound, double-contrast arthrography, and computed tomography arthrography with surgical correlation. Invest Radiol. 1996;31:387-394.
9. Flatow EL, Soslowsky LJ, Ticker JB, Pawluk RJ, Hepler M, Ark J, Mow VC, Bigliani LU. Excursion of the rotator cuff under the acromion: patterns of subacromial contact. Am J Sports Med. 1994;22:779-788.
10. Fuchs B, Weishaupt D, Zanetti M, Hodler J, Gerber C. Fatty degeneration of the muscles of the rotator cuff: assessment by computed tomography versus magnetic resonance imaging. J Shoulder Elbow Surg. 1999;8:599-605.
11. Godeneche A, Boileau P, Favard L, Le Huec JC, Levigne C, Nove-Josserand L, Walch G, Edwards TB. Prosthetic replacement in the treatment of osteoarthritis of the shoulder: early results of 268 cases. J Shoulder Elbow Surg. 2002;11:11-18.
12. Golding FC. The shoulder: the forgotten joint. Br J Radiol. 1962;35:149-158.
13. Goodpaster BH, Carlson CL, Visser M, Kelley DE, Scherzinger A,Harris TB. Stamm E, Newman AB. Attenuation of the skeletal muscle and strength in the elderly: the Health ABC Study. J Appl Physiol. 2001;90:2157-2165.
14. Goutallier D, Postel JM, Bernageau J, Lavau L, Voisin MC. Fatty muscle degeneration in cuff ruptures. pre-and postoperative evaluation by CT scan. Clin Orthop Relat Res. 1994;304:78-83.
15. Goutallier D, Postel JM, Gleyze P, Leguilloux P, Van Driessche S. Influence of cuff muscle fatty degeneration on anatomic and functional outcomes after simple suture of full-thickness tears. J Shoulder Elbow Surg. 2003;12:550-554.
16. Goutallier D, Postel JM, Lavau L, Bernageau J. Impact of fatty degeneration of the supraspinatus and infraspinatus muscles on the prognosis of surgical repair of the rotator cuff. Rev Chir Orthop Reparatrice Appar Mot. 1999;85:668-676.
17. Hashimoto T, Nobuhara K, Hamada T. Pathologic evidence of degeneration as a primary cause of rotator cuff tear. Clin Orthop Relat Res. 2003;415:111-120.
18. Hirooka A, Wakitani S, Yoneda M, Ochi T. Shoulder destruction in rheumatoid arthritis: classification and prognostic signs in 83 patients followed 5-23 years. Acta Orthop Scand. 1996;67:258-263.
19. Jost B, Pfirrmann CW, Gerber C, Switzerland Z. Clinical outcome after structural failure of rotator cuff repairs. J Bone Joint Surg Am. 2000;82:304-314.
20. Kieft GJ, Dijkmans BA, Bloem JL, Kroon HM. Magnetic resonance imaging of the shoulder in patients with rheumatoid arthritis. Ann Rheum Dis. 1990;49:7-11.
21. Larsen A, Dale K, Eek M. Radiographic evaluation of rheumatoid arthritis and related conditions by standard reference films. Acta Radiol Diagn (Stockh). 1977;18:481-491.
22. Lehtinen JT, Belt EA, Kauppi MJ, Kaarela K, Kuusela PP, Kautiainen HJ, Lehto MU. Bone destruction, upward migration, and medialisation of rheumatoid shoulder: a 15 year follow up study. Ann Rheum Dis. 2001;60:322-326.
23. Lehtinen JT, Belt EA, Lyback CO, Kauppi MJ, Kaarela K, Kautiainen HJ, Lehto MU. Subacromial space in the rheumatoid shoulder: a radiographic 15-year follow-up study of 148 shoulders. J Shoulder Elbow Surg. 2000;9:183-187.
24. Lesage P, Maynou C, Elhage R, Boutry N, Herent S, Mestdagh H. Reproducibility of CT scan evaluation of muscular fatty degeneration: intra-and interobserver analysis of 56 shoulders presenting with a ruptured rotator cuff muscles. Rev Chir Orthop Reparatrice Appar Mot. 2002;88:359-364.
25. Nove-Josserand L, Edwards TB, O'Connor DP, Walch G. The acromiohumeral and coracohumeral intervals are abnormal in rotator cuff tears with muscular fatty degeneration. Clin Orthop Relat Res. 2005;433:90-96.
26. Rozing PM, Brand R. Rotator cuff repair during shoulder arthroplasty in rheumatoid arthritis. J Arthroplasty. 1998;13:311-319.
27. Stoel BC, Vrooman HA, Stolk J, Reiber JH. Sources of error in lung densitometry with CT. Invest Radiol. 1999;34:303-309.
28. Uhthoff HK, Matsumoto F, Trudel G, Himori K. Early reattachment does not reverse atrophy and fat accumulation of the supraspinatus: an experimental study in rabbits. J Orthop Res. 2003;21:386-392.
29. Van De Sande MA, Rozing PM. Proximal migration can be measured accurately on standardized anteroposterior shoulder radio-graphs. Clin Orthop Relat Res. 2006;443:260-265.
30. Van De Sande MA, Stoel BC, Obermann WR, Lieng JG, Rozing PM. Quantitative assessment of fatty degeneration in rotator cuff muscles determined with computed tomography. Invest Radiol. 2005;40:313-319.
31. Weiner DS, Macnab I. Superior migration of the humeral head. A radiological aid in the diagnosis of tears of the rotator cuff. J Bone Joint Surg Br. 1970;52:524-527.
32. Zanetti M, Gerber C, Hodler J. Quantitative assessment of the muscles of the rotator cuff with magnetic resonance imaging. Invest Radiol. 1998;33:163-170.
© 2006 Lippincott Williams & Wilkins, Inc.