The incidence of infectious complications associated with liver biopsy (LB) is low, with reported rates of 0% to 3.15% of percutaneous biopsy procedures. Transient bacteremia, cholangitis, and liver abscesses have been described. Some studies have documented a higher incidence of biopsy-related infections in patients with a bile duct obstruction or a Roux-en-Y loop biliary bypass in comparison with patients with a choledochocholedochostomy (CDC),[3, 4] but this finding has not been supported by other authors. Thus, the published results to date are inconsistent with respect to the existence of a high-risk population within this group of patients. Furthermore, very few available studies have addressed complications related to an increasingly more widely used technique, transjugular liver biopsy (TJLB), in addition to complications related to percutaneous liver biopsy (PLB). Thus, there is currently insufficient scientific evidence to support or rule out the use of antibiotic prophylaxis in this setting to prevent septic complications. The aims of this study were to identify risk factors for the development of infectious complications related to LB in recipients of orthotopic liver transplantation (OLT) over the course of the last 10 years and to investigate the incidence of this event with both TJLB and PLB procedures.
Data from published studies regarding risk factors for liver biopsy (LB)–related infectious complications in liver transplant recipients are inconsistent. We carried out a retrospective cohort study analyzing consecutive LBs for orthotopic liver transplant patients at a tertiary hospital (2001-2011): there were 667 LB procedures (575 percutaneous procedures and 92 transjugular procedures) in 286 liver transplant recipients. There were 20 complications in 19 patients (overall incidence = 3.0%): 10 were infectious complications (8 cases of bacteremia and 2 cases of peritonitis). The causal microorganisms were mainly Pseudomonas aeruginosa (4 patients) and Enterobacteriaceae (4 patients). All complications occurred with biopsies performed in patients hospitalized for more than 48 hours (381 biopsies for 201 patients); hence, only this group was included in the risk factor analysis. The variables associated with the development of infectious complications after LB were the presence of impaired biliary drainage at the time of biopsy (40% versus 15.1%, P = 0.03) and low albumin levels (2.4 versus 3.1 g/dL, P = 0.01). In conclusion, according to our experience, infectious complications secondary to LB in liver transplant recipients are related to hospitalization at the time of biopsy, particularly in the presence of impaired biliary drainage and low albumin levels. Liver Transpl 20:552–556, 2014. © 2014 AASLD.
PATIENTS AND METHODS
We carried out a historical cohort study of all consecutive LBs performed in OLT recipients between January 2001 and December 2011 at Vall d'Hebron Hospital in Barcelona, Spain. Our institution is a 1000-bed university hospital at which 1078 adult liver transplants have been performed since 1988. Biopsies were identified through the Department of Hepatobiliopancreatic Surgery and Transplantation registry at our center. Data were collected from the medical records of each patient. Biopsies performed in patients younger than 18 years and those performed with techniques other than PLB or TJLB were excluded.
Study Variables and Data Collection
The following information was recorded: demographic, clinical, laboratory, histological, and microbiological data and data related to biopsy procedures and complications. Complications included severe bleeding, infection, pneumothorax, hemobilia, and other serious events. Patients were followed for 2 weeks after the procedure or until death during this period. All patients received immunosuppressive therapy according to our protocol: an calcineurin inhibitor plus steroids for patients without hepatitis C virus and an anticalcineurinic plus mycophenolate for patients with hepatitis C virus. The study was approved by our institutional review committee.
A bleeding complication following biopsy was defined as evident clinical signs of bleeding supported by a change in the arterial blood pressure (mean blood pressure < 70 mm Hg or decrease > 40 mm Hg), radiological evidence of intraperitoneal bleeding, and/or a >4% drop in the hematocrit requiring a blood transfusion and/or an angiographic or surgical intervention.
An infectious complication was diagnosed on the basis of positive blood or sterile body fluid cultures with clinical signs/symptoms of sepsis and/or radiological findings (eg, a liver abscess) after the procedure when other foci of infection had been ruled out. For the purposes of this study, impaired biliary drainage was defined as any situation involving abnormal bile elimination, such as papilla of Vater stenosis or sphincter of Oddi dysfunction, the presence of sand or gallstones in the biliary tree, bile duct stenosis, or evidence on LB of hepatic cholestasis that was suggestive of obstructive biliary disease.
Continuous variables are expressed as medians and interquartile ranges, and qualitative variables are expressed as numbers and percentages. Comparisons were performed with the chi-square test for categorical variables or with the Mann-Whitney U test for continuous variables. All statistical tests were 2-tailed, and the threshold of statistical significance was P < 0.05. Statistical analyses were carried out with SPSS 20.0 (SPSS, Chicago, IL).
Six hundred sixty-seven LBs (575 PLBs and 92 TJLBs), performed between 2001 and 2011 in 286 OLT recipients, were analyzed. A single biopsy was carried out for 87 patients (30.4%), and 2 to 9 biopsies were carried out for the remaining 199 patients (69.6%).
The demographic and clinical characteristics of patients undergoing LB and the recorded infectious complications are shown in Table 1. No statistically significant differences were found between patients with infectious complications and patients without infectious complications except for the hospitalization status at the time of the biopsy procedure.
|Cases (n = 667)||Infectious Complication||P Value|
|Yes (n = 10)||No (n = 657)|
|Male sex [n (%)]||451 (67.6)||5 (50)||446 (68)||0.23|
|Age (years)a||58 (48-64)||63 (58-68)||58 (48-64)||0.07|
|Comorbidity [n (%)]|
|Diabetes mellitus||164 (24.6)||0||164 (25)||0.07|
|Human immunodeficiency virus||55 (8.2)||0||55 (8.4)||0.34|
|Reason for transplantation [n (%)]|
|Hepatitis B virus||36 (5.4)||0||36 (5.5)||—|
|Hepatitis C virus||524 (78.6)||8 (80)||516 (78.5)||0.91|
|Alcohol||153 (22.9)||2 (20)||151 (23)||0.82|
|Complete OLT [n (%)]||651 (94)||10 (100)||641 (97.6)||—|
|Previous OLT [n (%)]||26 (4)||1 (10)||25 (3.8)||—|
|CDC [n (%)]||624 (93)||8 (80)||616 (93.8)||—|
|Prior biopsy (n)a||1 (0-2)||0 (0-2)||1 (0-2)||0.84|
|Inpatient status [n (%)]||381 (57)||10 (100)||371 (56.5)||0.01|
Overall, the median time interval between transplantation and biopsy was 5 months (interquartile range = 2-15 months); 86.2% were percutaneous procedures, and 57% were performed during hospitalization.
There were 20 complications in 19 patients (overall incidence = 3.0%): 10 bleeding complications (1.5%; 5 cases of hematoma, 3 cases of hemothorax, and 2 cases of hemobilia) and 10 infectious complications (1.5%). One patient presented with 2 complications in 2 separate biopsies: polymicrobial bacterial peritonitis and psoas hematoma requiring transfusion support. Nineteen of the 20 complications presented within the first 5 days after the procedure. There was only 1 case of bacteremia due to Pseudomonas aeruginosa that developed 10 days after TJLB.
The mortality rate for patients who presented with any complication was 15% (3 of 20 patients); death was related to massive hemothorax (1 case) or sepsis with positive blood cultures for P. aeruginosa (1 case) or Bacteroides uniforme (1 case).
As for the infectious complications, there were 8 patients with bacteremia and 2 cases of peritonitis. The microorganisms most frequently isolated were P. aeruginosa and Enterobacteriaceae (Table 2). Five of the 10 patients were receiving an antibiotic treatment at the time of biopsy other than cotrimoxazole for Pneumocystis jirovecii prophylaxis. In 3 of the 5 patients who were receiving antibiotic treatment, the causal microorganism was susceptible to the antibiotic administered.
|Complication||Sample||Previous Antibiotic||Isolated Microorganism||Antibiogram||Outcome|
|Sepsis||Blood culture||Amoxicillin/clavulanic acid||E. coli||—||Favorable|
|Sepsis||Blood culture||None||P. aeruginosa||Susceptible to amikacin, intermediately susceptible to cefepime, and resistant to the other antibiotics||Favorable|
|Sepsis||Blood culture||None||P. aeruginosa||Susceptible||Died|
|Sepsis||Blood culture||Piperacillin-tazobactam||P. aeruginosa||Susceptible||Favorable|
|Sepsis||Blood culture||None||P. aeruginosa||Susceptible to amikacin, imipenem, meropenem, and fosfomycin||Favorable|
|Sepsis||Blood culture||Amoxicillin/clavulanic acid||S. aureus||Resistant to cloxacillin||Favorable|
|Sepsis||Blood culture||Ampicillin + ceftriaxone||B. uniforme||Resistant to penicillin||Favorable|
|Sepsis||Blood culture||None||B. uniforme||Susceptible||Died|
|Bacterial peritonitis||Ascitic fluid culture||Vancomycin + piperacillin-tazobactam||Polymicrobial floraa||—||Favorable|
|Bacterial peritonitis||Ascitic fluid culture||None||E. coli||Resistant to ampicillin, ciprofloxacin, and cotrimoxazole||Favorable|
We identified 2 other infectious complications that were not included in the analysis. The first corresponded to a case of Escherichia coli bacteremia occurring 3 weeks after the biopsy. The second was a patient who presented with signs of meningoencephalitis 2 days after LB; Listeria monocytogenes was isolated from both cerebrospinal fluid and blood cultures. This case was not considered a biopsy-related complication because of the absence of LB-related Listeria sepsis cases reported in the literature, although cases have been described in relation to early after liver transplantation.[5, 6]
Because all infectious complications occurred in hospitalized patients, the risk factor analysis was performed only for this population; hence, the final analysis included 381 biopsies for 201 patients.
As for the inpatient analytical data, patients who developed an infectious complication had higher direct bilirubin levels (8.45 versus 2.87 mg/dL, P = 0.09). Moreover, there was a statistically significant difference related to albumin levels, which were lower in this group of patients (2.4 versus 3.1 g/dL, P = 0.01; Table 3). As for procedure-related factors (Table 4), we found that patients with infectious complications were receiving antibiotic treatment at the time of biopsy more frequently than those without complications (50% versus 24.5%, P = 0.06). On the other hand, there was a statistically significant higher rate of infectious complications in patients who presented with impaired biliary drainage on biopsy (40% versus 15.1%, P = 0.03).
|Infectious Complication||P Value|
|Yes (n = 10)||No (n = 371)|
|Leukocytes (×109/L)||6.5 (4.6-7.8)||5.4 (3.9-8.1)||0.61|
|Platelets (×109/L)||107.0 (84.0-137.7)||133.0 (86.0-180.0)||0.26|
|Prothrombin time (%)||77 (51-93)||83 (67-95)||0.28|
|Creatinine (mg/dL)||1.2 (0.87-1.85)||1.15 (0.93-1.6)||0.87|
|Direct bilirubin (mg/dL)||8.45 (1.87-13.2)||2.87 (0.70-8.1)||0.09|
|Albumin (g/dL)||2.4 (2.16-3.2)||3.1 (2.6-3.6)||0.01|
|Infectious Complication||P Value|
|Yes (n = 10)||No (n = 371)|
|Time from OLT to LB (months)a||1 (1-2.75)||2 (1-8)||0.08|
|TJLB [n (%)]||4 (40)||83 (22.4)||0.19|
|Steroid bolus treatment for acute rejection 3 months before biopsy [n (%)]||5 (50)||130 (35.0)||0.39|
|Antibiotic treatment at time of LB [n (%)]||5 (50)||91 (24.5)||0.06|
|Cotrimoxazole [n (%)]||3 (30)||190 (51.2)||0.18|
|Impaired biliary drainage [n (%)]||4 (40)||56 (15.1)||0.03|
In this study, we analyzed 667 LB procedures performed for patients who had undergone OLT over the course of the last 10 years. There were 20 complications (10 bleeding complications and 10 infectious complications) in 19 patients, and this yielded an overall incidence of 3.0%, which is higher than the reported rates of 0.6% to 2.7%[3, 4, 7]; the overall procedure-related mortality rate was 0.45%. The infectious complication rate in our study was 1.5%, which is also higher than the rate of 0.6% reported in previous studies.[3, 4] The hospitalized patients likely had a more severe and more complex clinical status, which could justify the higher incidence of complications in our series in comparison with others.
The most common infectious complication was bacteremia (8 of 10 cases). Four of the 8 cases were due to P. aeruginosa. The nosocomial nature of our episodes could justify the fact that the infectious etiologies differed from the results of previous studies, in which isolates corresponded to gastrointestinal flora[1, 3] or half the cases were due to skin flora such as Staphylococcus species.
As for microorganism susceptibility in our series, 2 of the 8 cases were due to multiresistant strains: 1 case of P. aeruginosa resistant to all antibiotics except amikacin and 1 case of methicillin-resistant Staphylococcus aureus. It is particularly noteworthy that 5 of the 10 patients who developed an infectious complication were receiving antibiotic treatment according to other indications on the basis of clinicians' criteria at the time of LB. Moreover, 3 of these 5 patients presented with an infection caused by a microorganism that was susceptible to the antibiotic indicated.
The main objective of our study was to identify risk factors for the development of LB-related infectious complications in OLT recipients in order to establish adequate preventive measures. The main risk factor identified in our study was hospitalization at the time of the procedure; all infectious complications affected inpatients. Because of this fact, we considered only this population for the analysis. In 1991, Bubak et al. estimated a higher incidence of infectious complications after LB in OLT recipients who had undergone Roux-en-Y choledochojejunostomy (CCJ) versus those who had undergone CDC, and they concluded that patients with a CCJ could be candidates for antibiotic prophylaxis before biopsy. Nonetheless, other similar studies have reported no significant differences in the infectious complication rates between patients with a CCJ and patients with a CDC,1,8,9 and this is in keeping with our findings (2.3% for CCJ patients versus 7.4% for CDC patients, P50.15).
In our inpatient population, we found statistically significant differences in albumin levels, with lower values in patients who developed an infectious complication. These results are consistent with similar studies in which a low albumin level (<3 g/dL) was a risk factor for the development of early cholangitis after transhepatic percutaneous drainage in patients with biliary tree obstructions.[10, 11]
We also found that patients whose biopsy samples presented with signs of bile duct obstruction had a higher infection rate than patients whose samples did not have these features. This was also supported by higher bilirubin levels in those patients who developed infectious complications. These findings agree with those reported by Larson et al., who found that the main risk factor for the development of an infectious complication was the presence of biliary abnormalities; a Roux-en-Y anastomosis was a prerequisite for the event.
In our opinion, the present results reflect an increased risk of complications for the patients facing the procedure under poorer clinical and analytical conditions, regardless of whether they were receiving adequate antibiotic treatment according to the susceptibility pattern of the isolated microorganism.
The limitations of our study are those inherent to a retrospective design and the low number of cases, which precluded a multivariate analysis of risk factors. Its interest resides in the fact that it provides recent information about the incidence of LB-related complications in the OLT population, and unlike other studies published to date, it includes results for both percutaneous and transjugular biopsy procedures.
In conclusion, the results of this study do not support routine antibiotic prophylaxis for OLT patients undergoing LB because antibiotic use did not show a protective effect against the development of infections. Nevertheless, in select liver transplant recipients in poor clinical condition (expressed as low albumin levels) who have proven or suspected impaired biliary drainage and are undergoing LB during their hospital stay, targeted prophylaxis (depending on local epidemiology) may be offered to decrease the severity of sepsis.
orthotopic liver transplantation
percutaneous liver biopsy
transjugular liver biopsy