SEPTIC SHOULDER PRESENTING AS BICEPS MUSCLE ABSCESS

Jeffrey C. Allard, MD

Allan Naimark, MD(1,2)

Victor W. Lee, MD(1)

Sidney Pollack, MD(1)

1. From the Department of Radiology at Boston University and

Boston City Hospital, 818 Harrison Ave., Boston, Massachusetts, 02118.

2. Reprint requests to Dr. A. Naimark, Radiology Dept.,

University Hospital, 75 E. Newton St., Boston, Massachusetts, 02118.

Septic arthritis of the shoulder is increasing in incidence, accounting for 3.4% of all septic joints before 1958 and 14.2% since, probably because of increased usage of corticosteroid injections (1). Delay in diagnosis is common, with osteomyelitis as a complication in about half of cases (2). We report a patient with septic arthritis of the shoulder who presented with a confusing clinical picture, including an abscess in the peribicipital region. The pathway of spread from the shoulder joint to the biceps muscle is explained by an understanding of the weak points in the shoulder capsule when the joint is distended. Arthrography and gallium (Ga) scanning proved most useful in making the diagnosis.

CASE PRESENTATION

A 32 year old alcoholic was admitted with a diagnosis of left arm cellulitis and possible thrombophlebitis. Two weeks prior to admission (PTA) he had symptoms of cough, fever, chills, and myalgia. One week PTA, after a minor fall, he presented to the emergency department with left arm pain. Shoulder radiographs demonstrated no fracture and he was diagnosed as having deltoid contusion. Five days PTA a visit to orthopedic clinic led to the consideration of frozen shoulder syndrome or rotator cuff tear. Physiotherapy was started 2 days PTA.

On admission the temperature was 102.4 F and the pulse 112. The left shoulder was stiff and in a "droopy" position. The entire left arm was swollen and erythematous. The white-cell count was 24,600 with a left shift and the erythrocyte sedimentation rate was 75 mm. per hour. The patient was treated with oxacillin and heparin. Upper extremity venography was negative. Plain films showed erosions of the humeral head (fig 1).

After 13 days of antibiotics and only minimal improvement, 12 cc. of sanguopurulent fluid was aspirated from a mass in the upper anterior arm. The following day, computed tomography (CT) was done (fig 2). Five cc. of identical fluid was aspirated from the shoulder joint, after which arthrography was performed, demonstrating immediate extravasation around the biceps with only an 8 cc. injection (fig 3). Fluid analysis revealed white-cell count of 42,000 with 74% leucocytes, decreased viscosity, and rare gram positive cocci. Bone and Ga scans showed evidence of contiguous infection extending from the shoulder joint into the upper arm and probable osteomyelitis of the humeral head (fig 4).


DISCUSSION

Unlike the knee, shoulder pathology rarely presents with extra-articular manifestations. Baker's cysts, for instance, can extend from the knee joint into the calf and when ruptured are well known to be difficult to distinguish from thrombophlebitis or cellulitis. Although synovial cysts are less common in the shoulder, they have been reported to occur as a result of local steroid injections, often in association with chronic rotator cuff tears (3). Chronic inflammatory conditions can distend the joint capsule and weaken the synovium, resulting in synovial cysts or joint capsule rupture. Some processes known to cause synovial cysts include rheumatoid arthritis, juvenile rheumatoid arthritis, Reiter's arthritis, septic arthritis, gout, osteoarthritis, pigmented villonodular synovitis, osteochondromatosis, and neuropathic arthropathy (3).

Overdistention of the shoulder joint during arthrography can result in leakage from the subscapularis recess or into the biceps sheath and is not considered pathological, as these are normal weak areas of the shoulder joint (fig 5). Joint capsule rupture has occured after physiotherapy (4), as it may have in our patient. Joint rupture or cyst formation in the shoulder usually occurs in the axilla, resulting from extensions or rupture of the axillary or subscapularis recesses of the normal joint. Clinically significant extension of shoulder joint pathology into the biceps region is rarely noted. Two patients, both with juvenile rheumatoid arthritis, have been reported with primary shoulder pathology presenting as masses in the anterior arm (5,6). Arthrographic communication with the shoulder joint was shown in these cases, presumably by way of the bursa surrounding the long head of the biceps.

Although gradual joint distention can be asymptomatic until the mass effects of the resulting cyst are felt, rupture of a cyst or joint capsule can have a dramatic and confusing clinical presentation. As with our patient, previous case reports have mentioned the diagnostic confusion of joint rupture with cellulitis, soft tissue abscess, or thrombophlebitis (4,7). In those cases, however, the capsule ruptured anteroinferiorly into the axilla and not in relation to the biceps sheath.

Diagnostic confusion with rotator cuff tear can occur because febrile responses in septic shoulder can be absent or unimpressive, and both conditions are associated. In one series of six septic joints, four had rotator cuff tears demonstrable by arthrography (8). Possible explanations for this association include the frequent usage of intraarticular steroid injections for complaints of shoulder pain or that the inflammatory process may erode through the rotator cuff. The amount of contrast injected in our case was suboptimal for the demonstration of possible rotator cuff tear as contrast flowed preferentially into the abscess cavity.

Initial plain films in septic arthritis are often normal, but a large percentage develop abnormalities (8). Late findings in six patients with infected shoulder joints included demineralization in five, subluxation in four, and joint space narrowing in three. Humeral head erosions were reported in 5 of 24 infected joints (2,8). Of six patients studied with arthrography, all had synovial irregularity and four had contrast extravasation into extraarticular cavities. In none of these cases was extravasation into the biceps region.

Our patient illustrates both the difficulty and the delay in diagnosis of this disease. Inability to move the shoulder in the absence of fracture led to the misdiagnoses of frozen shoulder and rotator cuff tear. The swelling and erythema of the upper arm resulted in confusion with cellulitis and thrombophlebitis. Finally, without an obvious explanation for our patient's lack of improvement, a bicipital abscess and its presumed precursor, the septic shoulder, were discovered. As the contour of this patient's upper arm cavity was ill-defined and irregular, it is unlikely that a dissecting synovial cyst was responsible for the resulting abscess, unless it had ruptured. We postulate that the pathway and pathophysiology, nevertheless, are similar except for the rate of development being more rapid in an extravasating abscess compared to a cyst. Although the exact communication could not be demonstrated on arthrography, we presume it to be along the route of the biceps tendon and its surrounding bursa, because this is a known potential pathway between the shoulder joint and the upper arm.

Diagnostically, nuclear medicine studies would have been very helpful earlier in the workup, although diagnostic aspiration and arthrography would also be essential. Combined imaging with TcMDP and Ga is very useful for elucidating the site and nature of inflammatory processes. In our case the Ga ,in particular, showed joint infection communicating with the upper arm abscess. In addition to its high sensitivity, scintigraphy is also superior in determining extent of involvement in multifocal disease. The finding of a cold defect in the upper arm corresponding to the humeral shaft together with the normal uptake of TcMDP in this area demonstrates that the infection originated from the joint and spread into the soft tissues, rather than from a primary focus of osteomyelitis. The relative contribution of CT was questionable as it did not demonstrate the soft tissue abscess. Perhaps magnetic resonance, with its improved soft tissue imaging features, would have been helpful.

REFERENCES

1. Kelly PJ, Coventry MB, Martin WJ. Bacterial arthritis of the shoulder. Mayo Clin Proc 1965;40:695-99

2. Gelberman RH, Menon J, Austerlitz MS et al. Pyogenic arthritis of the shoulder in adults. J Bone Jt Surg 1980;62:550-3

3. Nance EP, Jones TB, Kaye, JJ. Dissecting synovial cysts of the shoulder: a complication of chronic rotator cuff tears. AJR 1982; 138:739-41

4. DeJager JP, Fleming A. Shoulder joint rupture and pseudothrombosis in rheumatoid arthritis. Ann Rheumatic Dis 1984;43:503-4

5. Barbaric ZL, Young LW. Synovial cysts in juvenile rheumatoid arthritis. AJR 1972;116:655-60

6. Costello PB, Kennedy AC, Green FA. Shoulder joint rupture in juvenile arthritis producing bicipital masses and a hemorrhagic sign. J Rheumatology 1980;7:563-6

7. Lane PWF, Dyer NH, Hawkins, CF. Synovial rupture of shoulder joint. Br Med J 1972;1:356

8. Masters R, Weisman MH, Ambuster TG, Resnick D et al. Septic arthritis of the glenohumeral joint. Unique clinical and radiographic features and a favorable outcome. Arthritis and Rheumatology 1977; 20:1500-6

FIGURE LEGENDS

1. Internal rotation view demonstrates erosion (arrow).

2. CT shows an erosion of the lateral humeral head. Axial sections (not shown) at the level of the abscess were unremarkable.

3. Shoulder arthrogram with eight cc. of contrast demonstrates an irregular cavity in the upper arm that communicates with the shoulder joint. Insufficient contrast makes bicipital sheath connection impossible to visualize.

4a. Tc99m-methylene diphosphonate (TcMDP) scan shows increased activity in the region of the shoulder joint and humeral head (arrow).

4b. Gallium 67 (Ga) scan shows a large area of activity extending from the shoulder into the upper arm with a central linear cold defect corresponding to the humeral shaft. The discordance between the larger degree of activity in the shoulder on Ga compared to TcMDP indicates significant soft tissue (joint) inflammation. Cold Ga defect with normal TcMDP in the humeral shaft indicates the origin of the infection was from the shoulder and not from the humeral shaft (osteomyelitis).

5. Example of a normal arthrogram with extravasation into the biceps sheath (arrow).{courtesy of Dr. A. Newberg, Boston, Ma.}