Chest Wall Hematoma After Percutaneous Liver Biopsy
A 51-year-old obese woman with a medical history of systemic lupus erythematosus (SLE) with antiphospholipid syndrome, rheumatic mitral valve disease with porcine mitral valve replacement, transient ischemic attacks, and a seizure disorder after an episode of SLE cerebritis was referred to our gastroenterology service for evaluation of hepatosplenomegaly.
She had an ultrasound-guided liver biopsy (2 cores with 18-gauge needle) done by an experienced interventional radiologist after ensuring that she was not on antiplatelet agents or nonsteroidal anti-inflammatory drugs for 7 days and that her platelet counts (120K) and international normalized ratio (1.1) were acceptable. During the observation period she complained of right-sided pain over the liver that came on in spasms and was associated with deep inspiration. Between these spasms of pain she was pain-free. An ultrasound was repeated immediately after the biopsy and was normal. Her vital signs were stable throughout, and her physical exam was benign except for mild tenderness over the biopsy site. After 4 hours of further observation during which she still complained of similar spasms of chest wall pain (considerably ameliorated by Percocet [Endo Pharmaceuticals, Chadds Ford, PA]), she was discharged.
One week later she was still complaining of right upper quadrant pain, increased by deep inspiration, as well as dyspnea. An ultrasound was repeated, which was again normal. A chest x-ray suggested a right lower lobe infiltrate and a subpulmonic or loculated right pleural effusion. This was therefore followed up with computer-aided tomography (CT) of the thorax. The CT scan showed a large, loculated, elliptical, high-density fluid collection along the right lateral aspect of the liver that was considered to be a hematoma of the chest wall compressing the liver (Figure A, arrow). A very small amount of fluid layering was also seen in the right costophrenic angle. This hematoma was managed conservatively with analgesics and slowly resorbed, leaving the patient free of pain.
Laparoscopic observations suggest that some bleeding occurs after all liver biopsies.1 However, severe bleeding presenting with a drop of hemoglobin >2 g/dL or sufficient tachycardia or hypotension to require transfusions or surgical/radiologic intervention occurs in less than 1 of 2500 cases in patients with diffuse liver disease.1 The occurrence of chest wall hematomas is also documented in the literature, and in the series by Terjung et al2 it was as high as 2.8% of all patients who had a postbiopsy ultrasound, done routinely or for clinical indications. Such hematomas, which are “confined” to tissue planes, might not cause bleeding sufficient to affect the vital signs or drop the hematocrit. They might, however, cause significant pain. The use of ultrasound to detect such hematomas might be insensitive as a result of patient habitus, operator inexperience, and the fact that fresh hematomas are often isoechogenic to surrounding tissue.3 Findings on ultrasound often do not correlate with clinical symptoms such as local pain and might depend on how long after the biopsy the exam is done.3 This might suggest that such hematomas might be underdiagnosed by ultrasound.
The use of CT is strongly recommended in cases in which imaging is required after liver biopsy for symptoms or signs suggestive of a significant complication, especially persistent pain. In our case, 2 postbiopsy ultrasounds separated by a week and performed by different operators were read by 3 different radiologists as normal. Both ultrasounds “missed” the large chest wall hematoma that was easily seen on a subsequent CT scan.