Outcome of Patients Hospitalized for Complications after Outpatient Liver Biopsy

Methods

Between 1 April 1989 and 1 April 1991, 430 liver biopsies were done in the division of gastroenterology. Only 25 of these were inpatient biopsies; the remaining 405 were outpatient biopsies. All biopsies were done for the evaluation of chronic liver diseases (Table 1). The most frequent indications were the diagnosis and staging of primary biliary cirrhosis (40%), chronic active hepatitis (27%), and primary sclerosing cholangitis (14%). Alcoholic liver disease accounted for only 5% of the biopsies. Hemochromatosis, α₁-antitrypsin deficiency, and other chronic diseases each accounted for a small proportion of the biopsies. It has been the policy at the Mayo Clinic over the past few years to have all mass lesions biopsied under imaging guidance in the department of radiology. This is because biopsies of mass lesions seem to carry a higher risk for bleeding complications than those done for diagnosis of chronic liver diseases, and this risk is directly related to the number of passes required to obtain diagnostic tissue [4]. Hence, the approximately 300 biopsies of mass lesions were not included in our analysis. Postorthotopic liver transplant biopsies were also not included because they (approximately 200 per year) are done using a different protocol. Standard criteria for outpatient liver biopsy include the requirement that patients have a hemoglobin greater than 100 g/L (10.0 g/dL), a platelet count greater than 50 × 10⁹/L, and a prothrombin time less than 14.0 seconds (with control values of 10.2 to 12.3 seconds).

The biopsies were done by trainees in gastroenterology (postgraduate years 4 to 6) or a gastroenterology staff physician. A Tru-Cut (Baxter, Valencia, California) needle was used in 92% of the biopsies and a Jamshidi or Klatskin needle in the rest. More than one pass was generally done if diagnostic tissue was not obtained on the first pass. After the procedure, the patient was asked to lie flat, and the vital signs were checked every 15 minutes for 3 hours. Patients were then allowed to leave the outpatient area if no problems occurred, but they were required to remain in the Rochester area for at least 24 hours.

Reasons considered to justify admission were the development of fever; orthostatic hypotension; peritoneal signs; localized, persistent nonperitoneal pain; or lightheadedness, even without demonstrable change in vital signs. If complications occurred, patients were immediately hospitalized. An admission hemoglobin and 24-hour hemoglobin were obtained from all hospitalized patients. Further imaging after biopsy was obtained in cases of persistent pain, persistent or severe orthostatic hypotension, or decreasing hemoglobin levels.

Results

Thirteen patients were admitted with complications, which represented 3.2% of all patients undergoing outpatient liver biopsy during this period (Table 2). All complications were noted during the first 3 hours after biopsy. In patients admitted for complications, all biopsies had been done using a Tru-Cut needle, with an average of 1.2 passes. However, the number of passes was not statistically related to the occurrence of hemorrhage. Ten (77%) of those admitted had their liver biopsies done by fellows, who did 86% of the biopsies. Biopsies in the remaining three patients were done by staff gastroenterologists. The prothrombin time was normal in all patients, and the platelet counts were greater than 140 × 10⁹/L in all patients except one, who had a platelet count of 65 × 10⁹/L.

Five patients (38%) were admitted for persistent pain only (see Table 2). Of these patients, only one had a decrease in hemoglobin greater than 10 g/L (1 g/dL) during a 24-hour period. Four of the five patients had an imaging study of the liver (computed tomography [one patient] or ultrasound [three patients]) after biopsy, including the patient with a decrease in hemoglobin. Two complications, a small subcapsular hematoma and an echogenic gallbladder consistent with hemobilia were found in two patients who showed no decrease in hemoglobin concentration. The remainder of the imaging studies were negative. No one in this group received blood transfusions. Two patients, including the one with the subcapsular hematoma, received narcotics (each received meperidine hydrochloride, 100 mg total). No additional management, such as chest tube placement or surgery, was required.

Five patients (38%) were admitted for orthostatic hypotension and one patient was admitted for hypotension and pain (8%) (see Table 2). Five of these six patients had a decrease in hemoglobin of more than 10 g/L [1 g/dL] in a 24-hour period. The one patient was simply observed overnight. Three patients (including the patient with hypotension and pain) had a follow-up computed tomographic scan, which revealed an area of bleeding. One patient had a hemothorax, one patient had a subcapsular hematoma, and one patient had both a subcapsular hematoma and a hemothorax. The patient with the hemothorax and the patient with the subcapsular hematoma received 2 and 4 units of blood, respectively. No invasive management, such as chest tube placement or surgery, was required.

One patient was admitted for peritoneal signs but did not have hypotension or a decrease in hemoglobin concentration (see Table 2). Abdominal computed tomographic scan after liver biopsy showed no evidence of bleeding. His pain was presumed secondary to post-biopsy bile peritonitis. One other patient was admitted for lightheadedness and an increase in blood pressure. This patient, observed overnight, was free of complications and discharged.

The average length of the hospital stay was 1.5 days. All 13 patients did well and required no further intervention.

Discussion

The effectiveness of liver biopsy in altering treatment and disease outcome was once arguable because of limited treatment modalities for chronic liver disease. During the period of our study, the importance of liver biopsy, not just for diagnosis, but for therapeutic decisions and for monitoring has grown. This is reflected in expanded treatment options such as ursodeoxycholic acid for primary biliary cirrhosis and primary sclerosing cholangitis, the use of α-interferon for viral disease, and liver transplantation.

A skillfully done percutaneous liver biopsy has a relatively low rate of complications. Average mortality rates of 0.01% to 0.1% and morbidity rates of 0.1% have been reported [5-8]. The reported rate of major complications, including hemorrhage, bile leak, hemothorax, and perforation of abdominal viscera, for our institution was 3.2% [4]. Delayed complications, such as bleeding noted after 24 hours, are extremely rare. In one study of 68 276 liver biopsies, such complications were detected in only four patients, none of whom required treatment [8]. This finding suggests that a 24-hour observation period after biopsy should be sufficient.

During the past 10 years, most liver biopsies have been done on an outpatient basis, in conformance with recommendations outlined by the American Gastroenterological Society [1]. In the absence of a large randomized study comparing the safety of inpatient and outpatient biopsies, the low rate of severe complications and the rarity of delayed complications make outpatient liver biopsy a valid consideration if complications are easily recognizable, occur early, and are limited.

Of the 405 patients undergoing outpatient biopsy during the study period, only 13 (3.2%) required admission. Most patients (92%) were admitted for pain or hypotension. The most worrisome complication was hemorrhage. A decrease of more than 10 g/L (1 g/dL) in hemoglobin was noted in 6 of 13 patients, and 5 bleeding sites were radiologically identified (see Table 1). However, when bleeding occurred after liver biopsy, it was noted early, required no invasive management, and led to no deaths. All complications became manifest within the 3-hour period of close observation after liver biopsy, a finding that is consistent with previous studies [8, 9]. Transfusions, narcotics, or invasive management was rarely needed.

It has been our policy recently to require that those undergoing outpatient liver biopsy be able to return within 30 minutes to the site where the procedure is done. It has also been requested that during the subsequent 24 hours, a reliable individual remain with the patient and be able to provide care and transportation to medical services. Patients undergoing outpatient liver biopsy have also been highly selected according to defined criteria, including acceptable prothrombin time, platelet count, and hemoglobin level, and the ability to cooperate. If there is any evidence of bleeding, bowel leak, pneumothorax, or other organ puncture, patients are immediately hospitalized. Assessment of vital signs and patient monitoring after biopsy are done by the outpatient endoscopy staff. A special nursing unit has not been established, and, hence, no additional unusual outpatient costs have been added to the procedure.

Our study suggests that in carefully selected patients, when the outlined criteria are used, outpatient liver biopsy can be done safely and might lead to considerable reallocation of resources.