Volume 8, Issue 10 , Pages 826-829, October 2010
Stephen Caldwell, MD, Patrick G. Northup, MD
published online 28 June 2010.
Liver biopsy remains a cornerstone of diagnosis and disease staging and an important measure of natural history and therapeutic outcomes in many forms of liver disease. The major limitations of biopsy are the risk of procedure-related complications especially bleeding and adequacy of the sample to ensure accurate histologic interpretation. In the current issue of Clinical Gastroenterology and Hepatology, Seeff et al report liver biopsy complication rates in 1 of the largest cohorts of patients with histologically advanced disease.1
The study involved 2740 percutaneous biopsies performed over 7 years at 10 centers as part of the Hepatitis C Antiviral Long-Term Treatment against Cirrhosis (HALT-C) trial. The study assessed the efficacy of low dose maintenance peginterferon alfa-2a therapy in patients with histologically advanced hepatitis C—related fibrosis undergoing biopsy at baseline (n = 1187), 1.5 years (n = 852), and 3.5 years (n = 701) of treatment. Eighty percent of the biopsies were performed with ultrasound guidance, 40% used an aspiration needle, and 60% used a cutting needle. Single passes were reported in 40% and 2 passes in 60%. Complication rates were determined prospectively with no difference between baseline and follow-up biopsies.
Overall, serious adverse events (AE) occurred in 29 of the biopsies with no deaths. Severe bleeding was the most common cause seen in 16 (0.6%) of the biopsies including hemoperitoneum in 8, subcapsular hematoma in 4, hemobilia in 3, and hemothorax in 1. About one fourth of the adverse events were due to severe pain and one fourth were due to other causes including punctured gallbladder in 2, pneumothorax in 1, syncope in 1, and hypotension in 2. Because of its higher frequency and potential severity, the authors focused on bleeding risk assessment. No difference in those who bled versus those who did not was seen between 2 passes versus a single pass, use of a cutting needle versus an aspiration needle, or performance by a fellow as opposed to an attending physician. Factors that were statistically different between those with bleeding and those without were lower albumin, presence of varices, platelets less than 60,000 and INR ≥ 1.3 although it should be noted that among the 8 patients with international normalized ratio (INR) > 1.5, none had bleeding. Moreover, while these variables were statistically significant, the clinical differences were not clinically dramatic as is evident in Tables 6 and 7 of the report. The relationship to varices and albumin levels suggests that the variation in risk may be due to changes in hepatic vasculature in more advanced disease rather than changes in hemostatic mechanisms.
Although several other factors limit interpretation of the data including preset study inclusion criteria (based in part on conventional coagulation parameters) and the retrospective acquisition of procedure details by questionnaire as discussed by the authors, the study raises the interesting and challenging issue of bleeding risk assessment with percutaneous liver biopsy. Reassuringly, the mortality was zero although the incidence of nonfatal severe bleeding was higher in this cohort of advanced fibrosis stage patients, 1 in 200, compared with about 1 in 500 biopsies from a compilation of past reports of patients of all stages.2
Is the INR useful to predict bleeding in a cirrhosis patient? Despite the implications of the current study, the answer to this is clearly no. Although this test unfortunately continues to hold on to some adherents as a valid indicator of bleeding risk in cirrhosis patients, the practice is unfounded both physiologically (see below) and methodologically as the conventional INR in cirrhosis patients is not normalized to the various thromboplastin reagents used in the test.3 This problem is very evident by the marked interlaboratory variation consistently reported now in a number of studies represented in Figure 1.4, 5, 6, 7, 8 Thus an INR of 1.2 in 1 clinical reference laboratory can be an INR of 2 on the identical sample in another clinical reference laboratory. The problem lies with variation in activity of the thromboplastin and normalization of the test to warfarin-treated patients (international sensitivity index; ISI) rather than to liver disease patients.9 In the present study, liver disease—corrected INR (termed INRLiver) was not utilized so the degree of variation between the 10 involved centers is unknown. Even ignoring this, it is noteworthy that 6 of the patients who had bleeding had INR <1 while none of the 8 patients with INR >1.5 had bleeding.
Variation between different clinical reference laboratories in INR values on identical samples from patients with cirrhosis. High and low values for identical sample sets are shown. The variation is thought to result from differences in thromboplastin reagents. The values straddle the 1.5 cutoff in 8 of 19 cases. Aside from the substantial physiological limitations of this test (based on abnormalities of the hemostatic system in cirrhosis; see text), this degree of interlaboratory variation renders nonsensical the application of a universal clinical cutoff value as a measure of bleeding risk in cirrhosis patients although it remains a valid indicator in warfarin therapy for which the test was developed..
Clearly, the problem with INR as an index of bleeding risk in cirrhosis is more profound than simply marked interlaboratory variation. Because the test is derived from the prothrombin time, it is a reflection only of procoagulant factor activity and fails to account for the full breadth of both anti- and pro-hemostatic disturbances which are now known to exist in cirrhotic patients. It is especially notable that in stable cirrhosis patients the magnitude of procoagulant thrombin generation can be on the order of congenital protein C deficiency.10 This results from a relative resistance to endothelial-derived thrombomodulin (a protein C cofactor) due itself to decreased protein C levels and increased factor VIII activity. Hypercoagulability in cirrhosis is also supported by increased von Willebrand factor and decreased antithrombin III the latter of which can also lead to heparin resistance in this group of patients.11 Clinically, the presence of a hypercoagulable state is evident in the prevalence of deep vein thrombosis in this group.12, 13
On the other hand, bleeding is clearly a problem in these patients especially during episodes of decompensation requiring hospitalization. In a recent survey from our inpatient service, we observed 34 bleeding events (12 variceal and 22 nonvariceal) over a 42-day period about three fourths of which warranted transfusion of human blood products (Shah N, unpublished survey). Nonvariceal sources included mucosal bleeding, puncture wound bleeding, severe epistaxis, severe limb hematomas, and contusions and severe bleeding after dental extractions. Thus the issue is not an absence of bleeding risk but rather it is how to measure the risk and how to respond to a poorly balanced pro- and anti-hemostatic system. Other more discrete variables also contribute significantly to impaired hemostasis in this setting. These include hyperfibrinolysis seen in about 10% of cirrhosis patients and which should be suspected with delayed bleeding following a procedure,14, 15 renal insufficiency with uremic changes in platelets, and associated volume expansion which engorges the portal system,16 infection and associated release of endogenous heparinoids, and the very poorly characterized condition of dysfibrinogenemia which occurs in cirrhosis.16, 17 These complicated interactions are not measured by conventional coagulation indices such as INR. On the other hand, although INR clearly is limited as a bleeding risk measure, it remains unknown whether or not it is useful as a marker of dose effect of hemostatic agents used when cirrhosis patients bleed and it remains as a useful indicator of prognosis albeit with some unresolved problems arising from interlaboratory variation.
Are platelet levels useful to predict bleeding? The data are clearer with regard to platelet levels. While qualitative platelet changes which may be either pro- or anti-coagulant are largely unmeasured, convincing data have been published that show platelet levels of around 56,000/mL are associated with adequate in vitro thrombin production.18 Similar to the conclusions in the present study, a prior report also indicated a greater bleeding risk in liver biopsy when platelet levels were <60,000/mL.19 On the other hand it should be recalled that laparoscopically measured liver biopsy bleeding was unrelated to conventional coagulation parameters including platelet count in another study.20 Similarly, in the present study, 11 of the bleeding episodes occurred in patients with platelet levels >60,000/mL and 3 of them had platelet levels >150,000/mL.
Postbiopsy bleeding occurs in several forms ranging from arterial spurting to portal venous oozing to internalized hemobilia. Of these, arterial spurting is the most dramatic and usually becomes evident early in the postbiopsy period with associated hemodynamic changes. Although the timing of the bleeding episode after the biopsy and details of the management were not available, a number of the patients underwent interventional vascular imaging and embolization consistent with inadvertent small arterial puncture and bleeding. Such small vessels are not easily seen by conventional ultrasound and the incidence of arterial bleeding may well represent roughly the same frequency of incidental transgression of these structures. Histologic analysis of the specimens was not available to determine the presence or absence of these structures in those who bled versus those who did not bleed but it is our experience (thankfully infrequent) that such structures are evident on the histopathology slides when arterial bleeding is encountered (Figure 2). Because of changes in the hepatic vascular bed as early cirrhosis transitions to advanced cirrhosis with atrophy, it is possible that the likelihood of such an event increases with more advanced disease. Such a relationship is suggested by the association of bleeding with varices and lower albumin in the present study. As the authors likely have these biopsies in repository, it would be helpful to review these with attention to the presence or absence of vascular structures vis-à-vis the history of bleeding.
(A) A 48-year-old male with human immunodeficiency virus (HIV) in remission and possible drug-induced liver disease underwent ultrasound-guided biopsy with 16-gauge automated needle with a single pass. Prebiopsy INR = 0.9 and platelets 71,000/mL. Developed pain and decreased blood pressure postbiopsy. Computerized tomography (CT) showed site of bleeding (arrow). (B) Same patient underwent arteriography showing arterial extravasation (arrow). (C) Same patient status post coiling of the bleeding site. (D) Biopsy sample shows vessel wall of very small artery which likely represents the site of bleeding (H&E 200×). (E) Closer view of (D) (H&E 400×)..
The role of ultrasound (US) guidance either as real time imaging or prebiopsy site marking was not directly addressed as a potential risk modifier in the present study. This is probably because the majority (80%) of the biopsies were described as ultrasound-guided which likely limited the comparison. Whether or not the patients with bleeding were equally distributed in the US-guided group versus those without image guidance would be interesting but is not reported. However, given the size of the small arteriolar vessels that may be the source of severe hemorrhage (Figure 2) and the limitations of ultrasound resolution, it seems very unlikely that ultrasound guidance will have an effect on the risk of severe hemorrhage. This assessment is supported by several prior observational studies and a controlled trial of US-guided versus nonguided biopsy.21, 22, 23 In the latter study, a substantial difference in the complication rate (including bleeding) between US-guided and non-US-guided biopsy was only reported as a difference which included both severe pain and significant bleeding combined. Based on these studies, the use of US guidance is likely to have its greatest impact on the incidence of postbiopsy pain and more significantly on organ puncture including gall bladder puncture (seen in 2 patients in the present study) and pneumothorax (seen in 1) which is more likely to be reduced with US guidance. In the present study, it would be helpful to know the relative incidence of these events in those with US guidance versus those without.
Although other parameters may eventually supplant biopsy as a measure of the natural history of liver diseases and/or as an indicator of treatment response, at this time biopsy remains a key part of the evaluation in these patients. Acquisition of an adequate sample while ensuring safe performance remains paramount in this situation. Current data supports aiming for platelet levels in the range of at least 55,000/mL–60,000/mL and taking into account the possibility of coexisting conditions such as hyperfibrinolysis which might be evident from a history of easy bleeding for example. For the reasons discussed above, the INR warrants much less emphasis and certainly no data exist to support a specific cutoff. Transvenous biopsy may at first appear to offer a safer approach but a close review of the literature reveals a complication rate similar to that of percutaneous biopsy albeit the patient population may be selected with higher inherent risk.2 Most clearly, better measures of the hemostatic system balance in cirrhosis and the possible role of other prophylactic measures such as emerging procoagulants need clinical investigation to complement recent laboratory advances in this field.
1.Seeff LB, Everson GT, Morgan TR, et al. HALT-C TRIAL GROUP Complication rate of percutaneous liver biopsies among persons with advanced chronic liver disease in the HALT-C Trial. Clin Gastroenterol Hepatol. 2010;8:877–883
2.Rockey DC, Caldwell SH, Goodman ZD, et al. Liver biopsy. Hepatology. 2009;49:1017–1044
3.Malloy PC, Grassi CJ, Kundu S, et al. Consensus guidelines for periprocedural management of coagulation status and hemostasis risk in percutaneous image-guided interventions. J Vasc Interv Radiol. 2009;20:S240–S249
4.Tripodi A, Chantarangkul V, Primignani M, et al. The international normalized ratio calibrated for cirrhosis (INR(liver)) normalizes prothrombin time results for model for end-stage liver disease calculation. Hepatology. 2007;46:520–527
5.Bellest L, Eschwege V, Poupon R, et al. A modified international normalized ratio as an effective way of prothrombin time standardization in hepatology. Hepatology. 2007;46:528–534
6.Lisman T, van Leeuwen Y, Adelmeijer J, et al. Interlaboratory variability in assessment of the model of end-stage liver disease score. Liver Int. 2008;28:1344–1351
7.Trotter JF, Brimhall B, Arjal R, et al. Specific laboratory methodologies achieve higher model for endstage liver disease (MELD) scores for patients listed for liver transplantation. Liver Transpl. 2004;10:995–1000
8.Trotter JF, Olson J, Lefkowitz J, et al. Changes in international normalized ratio (INR) and model for endstage liver disease (MELD) based on selection of clinical laboratory. Am J Transplant. 2007;7:1624–1628
9.Porte RJ, Lisman T, Tripodi A, et al. The International Normalized Ratio (INR) in the MELD score: problems and solutions. Am J Transplant. 2010;10:1349–1353
10.Tripodi A, Primignani M, Chantarangkul V, et al. An imbalance of pro- vs anti-coagulation factors in plasma from patients with cirrhosis. Gastroenterology. 2009;137:2105–2111
11.Lisman T, Bongers TN, Adelmeijer J, et al. Elevated levels of von Willebrand factor in cirrhosis support platelet adhesion despite reduced functional capacity. Hepatology. 2006;44:53–61
12.Northup PG, McMahon MM, Ruhl AP, et al. Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism. Am J Gastroenterol. 2006;101:1524–1528quiz 1680
13.Sogaard KK, Horvath-Puho E, Gronbaek H, et al. Risk of venous thromboembolism in patients with liver disease: a nationwide population-based case-control study. Am J Gastroenterol. 2009;104:96–101
14.Ferguson JW, Helmy A, Ludlam C, et al. Hyperfibrinolysis in alcoholic cirrhosis: relative plasminogen activator inhibitor type 1 deficiency. Thromb Res. 2008;121:675–680
15.Gunawan B, Runyon B. The efficacy and safety of epsilon-aminocaproic acid treatment in patients with cirrhosis and hyperfibrinolysis. Aliment Pharmacol Ther. 2006;23:115–120
16.Smalberg JH, Leebeek FW. Superimposed coagulopathic conditions in cirrhosis: infection and endogenous heparinoids, renal failure, and endothelial dysfunction. Clin Liver Dis. 2009;13:33–42
17.Cunningham MT, Brandt JT, Laposata M, et al. Laboratory diagnosis of dysfibrinogenemia. Arch Pathol Lab Med. 2002;126:499–505
18.Tripodi A, Primignani M, Chantarangkul V, et al. Thrombin generation in patients with cirrhosis: the role of platelets. Hepatology. 2006;44:440–445
19.Sharma P, McDonald GB, Banaji M. The risk of bleeding after percutaneous liver biopsy: relation to platelet count. J Clin Gastroenterol. 1982;4:451–453
20.Ewe K. Bleeding after liver biopsy does not correlate with indices of peripheral coagulation. Dig Dis Sci. 1981;26:388–393
21.Lindor KD, Bru C, Jorgensen RA, et al. The role of ultrasonography and automatic-needle biopsy in outpatient percutaneous liver biopsy. Hepatology. 1996;23:1079–1083
22.Manolakopoulos S, Triantos C, Bethanis S, et al. Ultrasound-guided liver biopsy in real life: comparison of same-day prebiopsy versus real-time ultrasound approach. J Gastroenterol Hepatol. 2007;22:1490–1493
23.Stone MA, Mayberry JF. An audit of ultrasound guided liver biopsies: a need for evidence-based practice. Hepatogastroenterology. 1996;43:432–434
Conflicts of interest The authors disclose the following: Dr Caldwell consults for and has research support from CL Behring and consults for Bioengineering Inc, Charlottesville, VA. Dr Northup discloses no conflicts.
© 2010 AGA Institute. Published by Elsevier Inc. All rights reserved.