Afifty-three-year-old, right-hand-dominant man presented to the emergency
department of our institution with a two-week history of progressive pain in
the right shoulder, fever, and chills. In the week before admission, he first
was seen by a local chiropractor and then was seen in the emergency room of
another hospital. In both instances, the patient was diagnosed with a muscle
strain and was managed with pain medication and muscle relaxants. The shoulder
pain gradually increased in intensity over the two-week period before
admission, ultimately escalating to the point at which the patient was unable
to move the shoulder girdle without severe pain. Pain and the sensation of
fullness were localized to the posterior aspect of the shoulder in the
scapular and axillary regions. The patient denied having sustained any blunt
or penetrating trauma to the shoulder region, and he had undergone no previous
operations on the right shoulder. He had had no contact with any infectious
individual and had no history of skin infection or dermal compromise, systemic
illnesses or immunocompromise, recent infection, intravenous drug usage, or
previous aspirations or injections about the shoulder.
During the initial evaluation in our emergency department, the oral
temperature was 100.9°F (37.8°C). Visual inspection of the right
shoulder girdle demonstrated moderate erythema and edema of the scapular and
axillary regions. The patient had scapular winging. There was severe
tenderness along the swollen medial and inferior borders of the scapula as
well as in the adjacent axillary region. The patient reported no pain during
palpation of the anterior glenohumeral joint region, the subacromial region,
the acromioclavicular joint, or the sternoclavicular joint. Range of motion
was severely limited secondary to pain: active forward elevation to 70°,
abduction to 45°, and external rotation to 10° were all associated
with extreme discomfort. Motor and sensory examinations revealed that the
axillary, radial, ulnar, and median nerve functions were intact. Brisk radial
and ulnar pulses were palpable at the wrist.
The serum white blood-cell count on the day of admission was
22,000/mm3 (22.0 × 109/L) (normal, 4000 to
11,000/mm3 [4.0 to 11.0 × 109/L]), with 79%
neutrophils, 2% lymphocytes, 8% monocytes, and 0% eosinophils and basophils.
The platelet count was 571,000/mm3 (571.0 × 109/L)
(normal, 140,000 to 400,000/mm3 [140.0 to 400.0 ×
109/L]). The Westergren erythrocyte sedimentation rate was 105
mm/hr (normal, 0 to 20 mm/hr). The C-reactive protein level was 14.24 µg/dL
(142 µg/L) (normal, 0 to 0.4 µg/dL [0 to 4 µg/L]). Cultures of blood
specimens were obtained at the time of admission and ultimately were found to
be negative. Urinalysis and a chest radiograph did not identify a source of
infection. Liver-function tests were normal, and hepatitis screening tests
were negative for hepatitis B (HbsAg) and hepatitis C (HCV Ab). The test for
the human immunodeficiency virus was negative.
Multiplanar magnetic resonance images of the area about the right shoulder
girdle, including T1-weighted images, STIR (short-tau-inversion-recovery)
images, and T1-weighted images with fat saturation following intravenous
injection of gadolinium, demonstrated a multiloculated fluid collection
between the subscapularis muscle and the posterior portion of the chest wall
(Figs. 1 and
2). The rotator cuff muscles,
including the subscapularis muscle, demonstrated no abnormalities. The walls
of the lesion were thickened and irregular, with multiple septations that were
enhanced with the gadolinium. There was moderate surrounding edema. The
abscess extended inferiorly from the area anterior to the scapula, along the
midaxillary and posterior axillary lines, to the level of the diaphragm. The
pleura, hilum, and mediastinum of the lung appeared to be normal. Aspiration
of fluid from the lesion was not performed before the operation.
The patient was taken to the operating room for urgent open drainage of the
abscess on the evening of admission. After induction of general anesthesia,
the patient was placed in the lateral decubitus position on a beanbag. Routine
preparation and draping was performed. A curvilinear incision was made over
the medial border of the scapula and was extended inferiorly. The inferior
border of the trapezius muscle was retracted superiorly to expose the lower
half of the medial scapular border. Electrocautery was used to release the
rhomboid musculature from the medial scapular edge and to expose the
underlying space. Upon entrance into the scapulothoracic space, a massive
amount of purulent fluid was immediately encountered. Approximately 500 mL of
purulent material was evacuated. Intraoperative gram-staining of the specimen
yielded gram-positive cocci in clusters.
Multiple areas of loculation were released by gently sweeping an index
finger through the area. A moderate degree of necrotic tissue was found to be
forming the septations and lining the edges of the abscess and was removed
with a rongeur. The space was then irrigated with pulsatile lavage with use of
9 L of saline solution with cefazolin antibiotic (2 g of cefazolin per 3-L
bag). Two large Hemovac drains were placed to evacuate fluid from the
scapulothoracic space. The rhomboid musculature was reapproximated to the
medial scapular border with absorbable PDS (polydioxanone) monofilament
suture. Prolene monofilament suture (Ethicon, Somerville, New Jersey) was used
for skin closure. A shoulder sling (DonJoy UltraSling; dj Orthopedics, Vista,
California) was used postoperatively, with the arm positioned at the side to
allow unstressed wound-healing.
The patient was initially given intravenous vancomycin antibiotic therapy
until the culture sensitivities were available. Staphylococcus aureus
that was sensitive to oxacillin grew on final culture. Anaerobic and fungal
cultures were ultimately negative. When sensitivities were reported on the
third postoperative day, the intravenous antibiotic therapy was changed to
nafcillin (2 g every four hours), in accordance with the recommendation of our
infectious disease consultant. The two Hemovac drains were removed on the
third postoperative day.
The patient was hospitalized for ten days postoperatively for the
administration of intravenous antibiotic therapy. When the patient was
discharged on the tenth postoperative day, the white blood-cell count was
9100/mm3 (9.1 × 109/L), the erythrocyte
sedimentation rate was 55 mm/hr, and the C-reactive protein level was 1.03
µg/dL (10 µg/L). The patient was discharged to home with instructions to
complete a regimen of oral dicloxacillin (500 mg four times per day) over the
next eleven days (for a total of twenty-one days of antibiotic therapy). The
shoulder was immobilized in a sling for three weeks, after which the patient
was permitted to perform gentle pendulum exercises of the shoulder; at six
weeks postoperatively, the patient was advanced to full passive motion.
Strengthening exercises were started at eight weeks, after the patient had
achieved a full painless arc of passive motion of the shoulder. At three
months, the patient had painless active forward elevation to 180°,
abduction to 170°, external rotation to 80°, internal rotation to T7,
and grade-5 (of 5) strength of the supraspinatus, infraspinatus, and
subscapularis. By six months, the white blood-cell count had returned to
4900/mm3 (4.9 × 109/L); the erythrocyte
sedimentation rate, to 1.0 mm/hr; and the C-reactive protein level, to <0.3
µg/dL (<3 µg/L).