- Email: [email protected]
Safety and Utility of Transjugular Liver Biopsy in Hematopoietic Stem Cell Transplant Recipients
CLINICAL STUDY
Safety and Utility of Transjugular Liver Biopsy in Hematopoietic Stem Cell Transplant Recipients Bela Kis, MD, PhD, Vishwan Pamarthi, MD, Chieh-Min Fan, MD, Dmitry Rabkin, MD, PhD, and Richard A. Baum, MD
ABSTRACT Purpose: Hematopoietic stem cell transplant (HSCT) recipients are at high risk in the setting of percutaneous liver biopsy as a result of comorbid coagulopathy and ascites, and are commonly referred to undergo transjugular liver biopsy. The present study was performed to assess the safety and utility of transjugular liver biopsy in HSCT recipients and to analyze the correlation between corrected hepatic sinusoidal pressure gradient (CHSPG) and pathologic diagnoses. Materials and Methods: Data from reports of transjugular liver biopsy procedures, pathology reports, and laboratory values of 141 consecutive HSCT recipients who underwent transjugular liver biopsy with pressure measurement between January 2005 and August 2011 in a single institution were retrospectively reviewed and analyzed. Results: A total of 166 biopsy procedures were performed in 141 patients. Technical success rate was 98.8%. Biopsy was diagnostic in 95.7% of patients. There were three major complications (1.8%), including one death. CHSPG in patients with venoocclusive disease (VOD) was significantly higher (Po .001) than in those without VOD (16.2 mm Hg ⫾ 9.2 vs 5.6 mm Hg ⫾ 3.7). A CHSPG of 10 mm Hg or higher was 90.8% specific and 77.3% sensitive for VOD. Conclusions: The present data show that transjugular liver biopsy is a relatively safe procedure that provides important information for the clinical management of patients with HSCT. Measurement of CHSPG during the procedure can support the diagnosis of VOD.
ABBREVIATIONS ALP = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BUN = blood urea nitrogen, CHSPG = corrected hepatic sinusoidal pressure gradient, GVHD = graft-versus-host disease, HSCT = hematopoietic stem cell transplantation, INR = International Normalized Ratio, PT = prothrombin time, VOD = venoocclusive disease
Hematopoietic stem cell transplantation (HSCT) therapy is primarily used to treat hematologic malignant disorders, but its indication is expanding with the development of novel techniques that have led to improved outcomes. Worldwide, 50,417 HSCT procedures were performed in 2006 (1), and the number of patients receiving transplants
From the Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02115. Received June 2, 2012; final revision received September 10, 2012; accepted September 13, 2012. Address correspondence to B.K.; E-mail: [email protected] From the SIR 2012 Annual Meeting. B.K. was supported by the 2011/12 Radiological Society of North America Research Fellow Grant. Statistical consultation was received from a Harvard Catalyst biostatistician. None of the authors have identified a conflict of interest. & SIR, 2013 J Vasc Interv Radiol 2013; 24:85–89 http://dx.doi.org/10.1016/j.jvir.2012.09.011
from unrelated donors is expected to double in the next 5 years (2). Graft-versus-host disease (GVHD) of the liver and hepatic venoocclusive disease (VOD) are unique complications of HSCT and are among the leading causes of morbidity and mortality in HSCT recipients. The morbidity rate related to hepatic involvement can be as high as 80% and is responsible for 4%–14% of deaths (3). The treatment of this high-morbidity patient population is complicated by the uncertainty of the diagnosis based on the clinical, laboratory, and imaging findings (4). Therefore, liver biopsy is often performed as part of diagnostic workup. A liver biopsy specimen can be obtained with a percutaneous or transjugular approach. However, many patients with HSCT are at risk for bleeding secondary to thrombocytopenia and/or coagulopathy, and many have ascites, which is a relative contraindication to percutaneous biopsy. Consequently, patients with HSCT are often referred for transjugular liver biopsy. An additional benefit of transjugular liver biopsy is the ability to measure the corrected hepatic sinusoidal pressure gradient (CHSPG) during the procedure.
86
’
Safety/Utility of Transjugular Liver Biopsy in HSC Transplants
Despite these advantages, transjugular liver biopsy in HSCT recipients is regarded as a high-risk procedure based on small-scale published studies that reported very high complication rates for transjugular liver biopsy in HSCT recipients (5–7). Oshrine et al (5) reported an 80% major complication rate (four of five patients) of transjugular liver biopsy in pediatric HSCT recipients. Shulman et al (6) and Chahal et al (7) reported 31% (nine of 29 patients) and 11% (three of 27 patients) major complication rates, respectively, in adult HSCT recipients. However, as the number of patients undergoing HSCT increases (2), requests for transjugular liver biopsy are expected to increase as well (8). There is a need in the scientific literature to better evaluate the risks of transjugular liver biopsy in HSCT recipients. The purposes of the present study were to assess the safety and utility of transjugular liver biopsy in a larger patient population to estimate the potential risks and benefits of the procedure and to analyze the correlation between CHSPG and the pathologic diagnosis obtained through biopsy.
MATERIALS AND METHODS This study was approved by the institutional review board of our hospital. Medical records of consecutive patients who underwent transjugular liver biopsy with venous pressure measurements between January 2005 and August 2011 in our hospital were retrospectively reviewed and analyzed. This included a total of 166 transjugular liver biopsy procedures performed in 141 HSCT recipients. The HSCT recipients who underwent transjugular liver biopsy included 87 men and 54 women with a mean age of 45.3 years ⫾ 13.2 (standard deviation). Twenty-two patients underwent more than one biopsy procedure; 19 patients underwent two procedures each and three patients underwent three each. One to 10 biopsy samples were obtained per patient (mean, 4.4 ⫾ 1.7 among 164 patients; there was one failed biopsy and one procedure for which the number of samples was not reported). The biopsies were performed by a single team of interventional radiologists in a digital fluoroscopy suite by using the following technique. The right internal jugular vein was accessed under sonographic guidance and a 10F 40-cm vascular sheath was advanced into the inferior vena cava. Hepatic venography was performed through a 5-F multipurpose angled catheter to determine anatomy and patency of the selected hepatic vein and the optimal positioning for biopsy. The multipurpose angled catheter was exchanged for an 11.5-mm occlusion balloon catheter, and wedged and free hepatic venous pressures were measured (9,10). The CHSPG was calculated by subtracting free hepatic venous pressure from wedged hepatic venous pressure. The balloon catheter was exchanged for a 14gauge transjugular guiding cannula from a Liver Access and Biopsy set (Cook, Bloomington, Indiana), which was positioned in the selected hepatic vein. A 19-gauge core biopsy
Kis et al
’
JVIR
needle was then delivered through the cannula into the right hepatic lobe, and biopsy specimens were obtained under fluoroscopic guidance. Biopsy samples were immediately placed in formalin and sent for histologic analysis. Samples were also sent in saline solution for microbiology cultures if requested by the clinical team. The number of samples obtained was at the discretion of the interventional radiologist and depended on the number of studies ordered by the clinical team and on the size of the obtained tissue samples. All data presented in the present study were collected from retrospective review of patients’ computerized medical records. The records included charts, reports of the transjugular liver biopsy procedures, pathology reports of the biopsy specimens, and laboratory values. The analyzed laboratory values were hematocrit level, platelet count, International Normalized Ratio (INR), prothrombin time (PT), aspartate aminotransferase (AST) level, alanine aminotransferase (ALT) level, alkaline phosphatase (ALP) level, total bilirubin level, creatinine level, and blood urea nitrogen (BUN) level. Complications were defined as adverse events that occurred as a direct result of the liver biopsy. These were further classified into minor and major complications according to the Society of Interventional Radiology classification system (11). The need for transfusion of red blood cells within 24 hours after biopsy was considered a major complication. In general, platelet counts of more than 50/mL and INR values lower than 2 were preferred in patients who underwent transjugular liver biopsy. However, the interpretation of these guidelines varied on a case-by-case basis. Platelet transfusions were administered to 57 patients before the procedure, in whom the pretransfusion platelet count was 24,900/mL ⫾ 12,700 and increased to 69,400/mL ⫾ 22,700. The final preprocedural platelet count among all patients was 114,800/mL ⫾ 89,000 (n ¼ 165; one patient had no data available). The preprocedural INR was 1.3 ⫾ 0.4 (n ¼ 159; seven patients had no data available) and PT was 37.0 seconds ⫾ 13.7 (n ¼ 150; 16 patients had no data available). None of the HSCT recipients in the study had an INR greater than 2, and none received fresh frozen plasma transfusion before the procedure. The 24-hour postprocedural hematocrit level of 30.6% ⫾ 5.2 (n ¼ 149; 17 patients had no data available) was not significantly different compared with the preprocedural value of 31.3% ⫾ 5.5 (N ¼ 166). The patients had elevated liver function test results (ALT, 378.9 U/L ⫾ 737.7; AST, 331.3 U/L ⫾ 1,011.4; ALP, 342.9 U/L ⫾ 338.5; total bilirubin, 7.8 mg/dL ⫾ 9.4; all measurements recorded in all 166 patients). Preprocedural creatinine and BUN measurements were 1.1 mg/dL ⫾ 1.0 and 33.9 mg/dL ⫾ 28.2, respectively (N ¼ 166). Statistical analysis was performed with SigmaStat (version 2.03) statistical software (SPSS, Chicago, Illinois). Data are presented as means ⫾ standard deviation. A P value lower than .05 was considered to be statistically significant. When data in two groups passed the normality and equal variance tests, differences between the two
Volume 24
’
Number 1
’
January
’
2013
groups were assessed by Student t test. When a Student t test could not be performed, the two groups were compared with a Mann–Whitneyrank-sum test. When data in multiple groups passed the normality and equal variance tests, differences between multiple groups were assessed by one-way analysis of variance followed by Tukey comparison tests. When one-way analysis of variance could not be performed, groups were compared with Kruskal–Wallis one-way analysis of variance on ranks, followed by Dunn pairwise multiple comparison procedures. To study the relationship of the histopathologic diagnosis of VOD to the range of CHSPG values, a series of 2 2 contingency tables was constructed by using a Fisher exact test. For each of the pressure gradient values, the sensitivity, specificity, and positive and negative predictive values for finding VOD were calculated. Logistic regression analysis curves of specificity and sensitivity of CHSPG for VOD were constructed by MedCalc (version 12.2.1) statistical software (MedCalc, Mariakerke, Belgium).
87
Figure 1. CHSPG in HSCT recipients with different histopathologic diagnoses. ‘‘Others‘‘ indicates patients with a histopathologic diagnosis other than GVHD or VOD (*P o .05).
RESULTS During the study period, 166 transjugular liver biopsy procedures were performed in 141 patients who underwent HSCT. The technical success was 98.8% (164 of 166; one sample had insufficient tissue for diagnosis and one biopsy failed as a result of patient agitation, although pressure measurements were performed). There were three major complications (1.8%): one patient developed arteriobiliary fistula, which required embolization; one patient required blood transfusion within 24 hours after the procedure as a result of decreasing hematocrit level; and one patient experienced subcapsular and intraperitoneal hemorrhage leading to death within 36 hours. The overall mortality rate was 0.6%. Histologic analysis of the biopsy samples led to a diagnosis in 95.7% of cases (158 of 165; there was one nondiagnostic sample and six samples with nonspecific findings per pathology report). The histologic diagnosis was GVHD in 88 patients, VOD in 31 patients, coexistence of VOD with GVHD in 13 patients, other pathologic process (eg, lymphoma, steatohepatitis, drug toxicity, extramedullary hematopoiesis) in 27 patients, and nonspecific in six patients. Based on the histopathologic diagnosis, the patients were divided into four groups for further analysis of CHSPG data and laboratory values: patients with GVHD, patients with VOD, patients with coexistence of VOD with GVHD, and patients with another histologic diagnosis. The final group included patients with nonspecific histologic findings. The CHSPG was significantly higher (P o .001) in the VOD and VOD/GVHD groups compared with the other groups (Fig 1). When patients were divided into two groups—patients with VOD (n ¼ 44) and patients without VOD (n ¼ 121), the mean CHSPG in the VOD group (16.2 ⫾ 9.2) was significantly higher (P o .001) than in the non-VOD group (5.6 ⫾ 3.7).
Figure 2. Sensitivity and specificity curves of CHSPG for VOD. The solid and dashed thick lines represent sensitivity and specificity, respectively. The thin lines paralleling the respective thick lines represent 95% confidence intervals.
Logistic regression analysis curves of specificity and sensitivity of CHSPG for VOD are shown in Figure 2. Figure 2 demonstrates that a CHSPG of 10 mm Hg or higher provides 90.8% specificity and 77.3% sensitivity for the diagnosis of VOD. Specificity, sensitivity, and positive/ negative predictive values for the 9–11 mm Hg range of CHSPG are shown in Table 1. Comparison of the laboratory values of the GVHD and VOD groups are shown in Table 2. Patients with VOD had significantly higher INR, PT, AST, creatinine, and BUN values, whereas patients with GVHD had significantly higher ALP levels. There was no statistically significant difference in total bilirubin and ALT levels between the two groups.
88
’
Safety/Utility of Transjugular Liver Biopsy in HSC Transplants
DISCUSSION The number of patients receiving allogenic HSCT is increasing worldwide, and hepatic complications occur in a significant proportion of HSCT recipients (1,2). Despite human leukocyte antigen identity between a patient and donor, approximately 40% of recipients of human leukocyte antigen–identical grafts develop GVHD, and the liver is involved in 50% of these cases (2). Histologically, hepatic GVHD presents as lymphocytic infiltration of the portal areas, pericholangitis, endothelial injury, and bile duct destruction (12). VOD is characterized clinically by the triad of hepatomegaly, ascites/weight gain, and jaundice (13). VOD develops secondary to high-dose cytoreductive therapy in susceptible patients, resulting in endothelial injury in sinusoids and small hepatic venules with consequent activation of the coagulation cascade. This leads to clot formation and obliteration of the hepatic venules (13). The clinical manifestations of liver diseases caused by GVHD or VOD can mimic other hepatic complications of HSCT, such as viral hepatitis (herpes simplex virus, adenovirus, varicella virus, hepatitis B and C viruses), fungal infection, leukemic infiltration, drug toxicity, and cholestasis secondary to septicemia, total parenteral nutrition, or medications (4,6). GVHD and VOD require specific treatment, and early treatment improves the probability of treatment success (2,8,13). Therefore, timely and accurate diagnosis is essential. When the diagnosis is unclear based on the clinical picture, liver biopsy is required. Even with Table 1 . Specificity, Sensitivity, and Positive and Negative Predictive Values for Venoocclusive Disease based on CHSPG CHSPG (%) Statistic
4 8 mm Hg
4 9 mm Hg
Sensitivity
88.36
79.55
4 10 mm Hg 77.27
Specificity PPV
84.17 66.7
89.17 72.9
90.83 75.6
NPV
94.4
92.2
91.6
CHSPG ¼ corrected hepatic sinusoidal pressure gradient, NPV ¼ negative predictive values, PPV ¼ positive predictive values.
Kis et al
’
JVIR
a typical clinical presentation, biopsy may be necessary. According to the study of Esposito et al (14), 66% (14 of 21) of HSCT recipients had a different pathologic diagnosis than what was clinically suspected. Shulman et al (6) reported that liver histologic findings provided additional information that was beneficial for HSCT recipient management in 88% of cases (53 of 60 patients). Most of the time, HSCT recipients are referred for transjugular liver biopsy as a result of thrombocytopenia and/or coagulopathy. In addition, ascites commonly presents in HSCT recipients, which may defer percutaneous biopsy. Previous small-scale studies (5–7) of transjugular liver biopsy in HSCT recipients (the largest included 29 patients) reported major complication rates as low as 11% and as high as 80%. Our present retrospective analysis of 166 transjugular liver biopsies in 141 HSCT recipients revealed a 1.8% major complication rate and a 0.6% mortality rate. Reported complication rates of transjugular liver biopsies involving more than 100 patients range between 2.4% and 6% (15–17). In the present study, the technical success rate of transjugular liver biopsy in HSCT recipients was 98.8%. Biopsy samples led to histologic diagnosis in 95.7% of cases, which highlights the clinical utility of the procedure. Our results are comparable to transjugular liver biopsy data in the general patient population, which were reported to show a technical success rate of 96.8% and a histologic diagnosis rate of 96.1% in a metaanalysis of 64 series reporting 7,649 transjugular liver biopsies (18). The most frequent histologic diagnosis in our HSCT patient population was GVHD, followed by VOD, and this corresponds to data described in the literature (7). The distinction between GVHD and VOD is very important because GVHD requires high-dose immunosuppressive therapy (2), whereas the treatment of VOD primarily includes supportive care, fibrinolysis, and anticoagulation (13). The measurement of CHSPG has been reported in small series to discriminate between VOD and other liver pathologic processes in HSCT recipients (6). Our data in a relatively large patient population demonstrate again that measurement of CHSPG during the transjugular liver biopsy procedure can support the diagnosis of VOD.
Table 2 . Laboratory Values of Patients with GVHD and VOD Parameter
GVHD
VOD
P Value
1.16 ⫾ 0.25
1.73 ⫾ 0.55
o .05
0.9–1.1
34.03 ⫾ 15.56
43.47 ⫾ 8.26
o .05
23.8–36.6
ALT, U/L AST, U/L
409.86 ⫾ 445.29 212.81 ⫾ 277.70
322.39 ⫾ 441.15 483.84 ⫾ 586.31
.35 o .05
10–50 10–50
ALP, U/L
428.23 ⫾ 410.99
193.48 ⫾ 129.03
o .05
35–130
8.50 ⫾ 10.21 1.00 ⫾ 0.60
7.04 ⫾ 7.77 2.08 ⫾ 1.58
.47 o .05
0.0–1.0 0.5–1.2
27.82 ⫾ 26.84
48.81 ⫾ 25.53
o .05
6–23
INR PT, s
Total bilirubin, mg/dL Creatinine, mg/dL BUN, mg/dL
Normal Range
Values presented as mean ⫾ standard deviation. ALP ¼ alkaline phosphatase, ALT ¼ alanine aminotransferase, AST ¼ aspartate aminotransferase, BUN ¼ blood urea nitrogen, GVHD ¼ graft-versus-host disease, INR ¼ International Normalized Ratio, PT ¼ prothrombin time, VOD ¼ venoocclusive disease.
Volume 24
’
Number 1
’
January
’
2013
A CHSPG of 10 mm Hg or higher is 90.8% specific and 77.3% sensitive for VOD. Therefore, measurement of CHSPG can lead to initiation of specific treatment days before the final histologic diagnosis and potentially reduce morbidity and mortality in HSCT recipients. In addition, according to published data, CHSPG may be helpful in determining the prognosis of patients with VOD (6). Analysis of laboratory values for HSCT recipients who underwent transjugular liver biopsy at our institution showed that the patients had abnormal liver function test results, as expected. The average red blood cell count and platelet count were lower than normal range, and the INR and PT values were slightly increased. These results are not surprising in this patient population. There were a few interesting findings when the laboratory values of patients with GVHD were compared with those of patients with VOD. Patients with GVHD had significantly higher ALP levels. This finding corresponds to liver histologic findings in GVHD, which show lymphocytic infiltration of portal areas, pericholangitis, and bile duct destruction. In the liver, ALP is found in the epithelial cells lining the biliary tract (19), and it is well known that plasma ALP level is increased in pathologic conditions involving the biliary tree (20). Another interesting finding was the significantly increased creatinine and BUN levels in patients with VOD compared with those with GVHD. Increased risks of acute renal failure have been observed in patients with VOD, whereas GVHD was not associated with renal impairment (21). The relationship between VOD and renal impairment is not clear; decreased hepatic clearance of nephrotoxic endotoxins absorbed from the gut may play a role (22). The main deficiency of the present study is its retrospective nature. In addition, we report a single institution’s experience, and patient selection was at the preference of the stem cell transplant clinical team, which may have skewed our results. Despite these limitations, the value of our investigation lies in the size of our patient population, which, to our knowledge, is nearly five times the size of the next largest series. In summary, the present study supports the relative safety of transjugular liver biopsy in HSCT recipients. The measurement of CHSPG during the transjugular liver biopsy procedure can provide important diagnostic information days before the availability of the pathologic diagnosis, and may accelerate the initiation of diseasespecific therapy. This additional information makes transjugular liver biopsy the preferred route even in patients who otherwise would be eligible for percutaneous biopsy, particularly when VOD is a consideration in the differential diagnosis.
89
REFERENCES 1. Gratwohl A, Baldomero H, Aljurf M, et al. Hematopoietic stem cell transplantation: a global perspective. JAMA 2010; 303:1617–1624. 2. Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet 2009; 373:1550–1561. 3. Locasciulli A, Alberti A, de Bock R, et al. Impact of liver disease and hepatitis infections on allogeneic bone marrow transplantation in Europe: a survey from the European Bone Marrow Transplantation (EBMT) Group—Infectious Diseases Working Party. Bone Marrow Transplant 1994; 14:833–837. 4. Carreras E, Granena A, Navasa M, et al. On the reliability of clinical criteria for the diagnosis of hepatic veno-occlusive disease. Ann Hematol 1993; 66:77–80. 5. Oshrine B, Lehmann LE, Duncan CN. Safety and utility of liver biopsy after pediatric hematopoietic stem cell transplantation. J Pediatr Hematol Oncol 2011; 33:e92–e97. 6. Shulman HM, Gooley T, Dudley MD, et al. Utility of transvenous liver biopsies and wedged hepatic venous pressure measurements in sixty marrow transplant recipients. Transplantation 1995; 59:1015–1022. 7. Chahal P, Levy C, Litzow MR, Lindor KD. Utility of liver biopsy in bone marrow transplant patients. J Gastroenterol Hepatol 2008; 23:222–225. 8. Mahgerefteh SY, Sosna J, Bogot N, Shapira MY, Pappo O, Bloom AI. Radiologic imaging and intervention for gastrointestinal and hepatic complications of hematopoietic stem cell transplantation. Radiology 2011; 258:660–671. 9. Smith TP, Kim CY, Smith AD, Janas G, et al. Hepatic venous pressure measurements: comparison of end-hole and balloon catheter methods. J Vasc Interv Radiol 2012; 23:219–226. 10. Maleux G, Willems E, Fieuws S, et al. Prospective study comparing different indirect methods to measure portal pressure portal pressure. J Vasc Interv Radiol 2011; 22:1553–1558. 11. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14(suppl):S199–S202. 12. Fujii N, Takenaka K, Shinagawa K, et al. Hepatic graft-versus-host disease presenting as an acute hepatitis after allogeneic peripheral blood stem cell transplantation. Bone Marrow Transplant 2001; 27:1007–1010. 13. Senzolo M, Germani G, Cholongitas E, Burra P, Burroughs AK. Venoocclusive disease: update on clinical management. World J Gastroenterol 2007; 13:3918–3924. 14. Esposito AA, Nicolini A, Meregaglia D, Sangiovanni A, Biondetti P. Role of transjugular liver biopsy in the diagnostic and therapeutic management of patients with severe liver disease. Radiol Med 2008; 113:1008–1017. 15. Bull HJ, Gilmore IT, Bradley RD, Marigold JH, Thompson RP. Experience with transjugular liver biopsy. Gut 1983; 24:1057–1060. 16. Choh J, Dolmatch B, Safadi R, et al. Transjugular core liver biopsy with a 19-gauge spring-loaded cutting needle. Cardiovasc Intervent Radiol 1998; 21:88–90. 17. Smith TP, Presson TL, Heneghan MA, Ryan JM. Transjugular biopsy of the liver in pediatric and adult patients using an 18-gauge automated core biopsy needle: a retrospective review of 410 consecutive procedures. AJR Am J Roentgenol 2003; 180:167–172. 18. Kalambokis G, Manousou P, Vibhakorn S, et al. Transjugular liver biopsy—indications, adequacy, quality of specimens, and complications— a systematic review. J Hepatol 2007; 47:284–294. 19. Jacoby F, Martin BF. The relationship of bile alkaline phosphatase to histochemically detectable alkaline phosphatase in the biliary tract; including reference to the histology of the gall-bladder epithelium. J Anat 1951; 85:391–399. 20. Popper H, Schaffner F. Pathophysiology of cholestasis. Hum Pathol 1970; 1:1–24. 21. Hingorani SR, Guthrie K, Batchelder A, et al. Acute renal failure after myeloablative hematopoietic cell transplant: incidence and risk factors. Kidney Int 2005; 67:272–277. 22. McDonald GB, Hinds MS, Fisher LD, et al. Veno-occlusive disease of the liver and multiorgan failure after bone marrow transplantation: a cohort study of 355 patients. Ann Intern Med 1993; 118:255–267.
Safety and Utility of Transjugular Liver Biopsy in Hematopoietic Stem Cell Transplant Recipients Bela Kis, MD, PhD, Vishwan Pamarthi, MD, Chieh-Min Fan, MD, Dmitry Rabkin, MD, PhD, and Richard A. Baum, MD
ABSTRACT Purpose: Hematopoietic stem cell transplant (HSCT) recipients are at high risk in the setting of percutaneous liver biopsy as a result of comorbid coagulopathy and ascites, and are commonly referred to undergo transjugular liver biopsy. The present study was performed to assess the safety and utility of transjugular liver biopsy in HSCT recipients and to analyze the correlation between corrected hepatic sinusoidal pressure gradient (CHSPG) and pathologic diagnoses. Materials and Methods: Data from reports of transjugular liver biopsy procedures, pathology reports, and laboratory values of 141 consecutive HSCT recipients who underwent transjugular liver biopsy with pressure measurement between January 2005 and August 2011 in a single institution were retrospectively reviewed and analyzed. Results: A total of 166 biopsy procedures were performed in 141 patients. Technical success rate was 98.8%. Biopsy was diagnostic in 95.7% of patients. There were three major complications (1.8%), including one death. CHSPG in patients with venoocclusive disease (VOD) was significantly higher (Po .001) than in those without VOD (16.2 mm Hg ⫾ 9.2 vs 5.6 mm Hg ⫾ 3.7). A CHSPG of 10 mm Hg or higher was 90.8% specific and 77.3% sensitive for VOD. Conclusions: The present data show that transjugular liver biopsy is a relatively safe procedure that provides important information for the clinical management of patients with HSCT. Measurement of CHSPG during the procedure can support the diagnosis of VOD.
ABBREVIATIONS ALP = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BUN = blood urea nitrogen, CHSPG = corrected hepatic sinusoidal pressure gradient, GVHD = graft-versus-host disease, HSCT = hematopoietic stem cell transplantation, INR = International Normalized Ratio, PT = prothrombin time, VOD = venoocclusive disease
Hematopoietic stem cell transplantation (HSCT) therapy is primarily used to treat hematologic malignant disorders, but its indication is expanding with the development of novel techniques that have led to improved outcomes. Worldwide, 50,417 HSCT procedures were performed in 2006 (1), and the number of patients receiving transplants
From the Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02115. Received June 2, 2012; final revision received September 10, 2012; accepted September 13, 2012. Address correspondence to B.K.; E-mail: [email protected] From the SIR 2012 Annual Meeting. B.K. was supported by the 2011/12 Radiological Society of North America Research Fellow Grant. Statistical consultation was received from a Harvard Catalyst biostatistician. None of the authors have identified a conflict of interest. & SIR, 2013 J Vasc Interv Radiol 2013; 24:85–89 http://dx.doi.org/10.1016/j.jvir.2012.09.011
from unrelated donors is expected to double in the next 5 years (2). Graft-versus-host disease (GVHD) of the liver and hepatic venoocclusive disease (VOD) are unique complications of HSCT and are among the leading causes of morbidity and mortality in HSCT recipients. The morbidity rate related to hepatic involvement can be as high as 80% and is responsible for 4%–14% of deaths (3). The treatment of this high-morbidity patient population is complicated by the uncertainty of the diagnosis based on the clinical, laboratory, and imaging findings (4). Therefore, liver biopsy is often performed as part of diagnostic workup. A liver biopsy specimen can be obtained with a percutaneous or transjugular approach. However, many patients with HSCT are at risk for bleeding secondary to thrombocytopenia and/or coagulopathy, and many have ascites, which is a relative contraindication to percutaneous biopsy. Consequently, patients with HSCT are often referred for transjugular liver biopsy. An additional benefit of transjugular liver biopsy is the ability to measure the corrected hepatic sinusoidal pressure gradient (CHSPG) during the procedure.
86
’
Safety/Utility of Transjugular Liver Biopsy in HSC Transplants
Despite these advantages, transjugular liver biopsy in HSCT recipients is regarded as a high-risk procedure based on small-scale published studies that reported very high complication rates for transjugular liver biopsy in HSCT recipients (5–7). Oshrine et al (5) reported an 80% major complication rate (four of five patients) of transjugular liver biopsy in pediatric HSCT recipients. Shulman et al (6) and Chahal et al (7) reported 31% (nine of 29 patients) and 11% (three of 27 patients) major complication rates, respectively, in adult HSCT recipients. However, as the number of patients undergoing HSCT increases (2), requests for transjugular liver biopsy are expected to increase as well (8). There is a need in the scientific literature to better evaluate the risks of transjugular liver biopsy in HSCT recipients. The purposes of the present study were to assess the safety and utility of transjugular liver biopsy in a larger patient population to estimate the potential risks and benefits of the procedure and to analyze the correlation between CHSPG and the pathologic diagnosis obtained through biopsy.
MATERIALS AND METHODS This study was approved by the institutional review board of our hospital. Medical records of consecutive patients who underwent transjugular liver biopsy with venous pressure measurements between January 2005 and August 2011 in our hospital were retrospectively reviewed and analyzed. This included a total of 166 transjugular liver biopsy procedures performed in 141 HSCT recipients. The HSCT recipients who underwent transjugular liver biopsy included 87 men and 54 women with a mean age of 45.3 years ⫾ 13.2 (standard deviation). Twenty-two patients underwent more than one biopsy procedure; 19 patients underwent two procedures each and three patients underwent three each. One to 10 biopsy samples were obtained per patient (mean, 4.4 ⫾ 1.7 among 164 patients; there was one failed biopsy and one procedure for which the number of samples was not reported). The biopsies were performed by a single team of interventional radiologists in a digital fluoroscopy suite by using the following technique. The right internal jugular vein was accessed under sonographic guidance and a 10F 40-cm vascular sheath was advanced into the inferior vena cava. Hepatic venography was performed through a 5-F multipurpose angled catheter to determine anatomy and patency of the selected hepatic vein and the optimal positioning for biopsy. The multipurpose angled catheter was exchanged for an 11.5-mm occlusion balloon catheter, and wedged and free hepatic venous pressures were measured (9,10). The CHSPG was calculated by subtracting free hepatic venous pressure from wedged hepatic venous pressure. The balloon catheter was exchanged for a 14gauge transjugular guiding cannula from a Liver Access and Biopsy set (Cook, Bloomington, Indiana), which was positioned in the selected hepatic vein. A 19-gauge core biopsy
Kis et al
’
JVIR
needle was then delivered through the cannula into the right hepatic lobe, and biopsy specimens were obtained under fluoroscopic guidance. Biopsy samples were immediately placed in formalin and sent for histologic analysis. Samples were also sent in saline solution for microbiology cultures if requested by the clinical team. The number of samples obtained was at the discretion of the interventional radiologist and depended on the number of studies ordered by the clinical team and on the size of the obtained tissue samples. All data presented in the present study were collected from retrospective review of patients’ computerized medical records. The records included charts, reports of the transjugular liver biopsy procedures, pathology reports of the biopsy specimens, and laboratory values. The analyzed laboratory values were hematocrit level, platelet count, International Normalized Ratio (INR), prothrombin time (PT), aspartate aminotransferase (AST) level, alanine aminotransferase (ALT) level, alkaline phosphatase (ALP) level, total bilirubin level, creatinine level, and blood urea nitrogen (BUN) level. Complications were defined as adverse events that occurred as a direct result of the liver biopsy. These were further classified into minor and major complications according to the Society of Interventional Radiology classification system (11). The need for transfusion of red blood cells within 24 hours after biopsy was considered a major complication. In general, platelet counts of more than 50/mL and INR values lower than 2 were preferred in patients who underwent transjugular liver biopsy. However, the interpretation of these guidelines varied on a case-by-case basis. Platelet transfusions were administered to 57 patients before the procedure, in whom the pretransfusion platelet count was 24,900/mL ⫾ 12,700 and increased to 69,400/mL ⫾ 22,700. The final preprocedural platelet count among all patients was 114,800/mL ⫾ 89,000 (n ¼ 165; one patient had no data available). The preprocedural INR was 1.3 ⫾ 0.4 (n ¼ 159; seven patients had no data available) and PT was 37.0 seconds ⫾ 13.7 (n ¼ 150; 16 patients had no data available). None of the HSCT recipients in the study had an INR greater than 2, and none received fresh frozen plasma transfusion before the procedure. The 24-hour postprocedural hematocrit level of 30.6% ⫾ 5.2 (n ¼ 149; 17 patients had no data available) was not significantly different compared with the preprocedural value of 31.3% ⫾ 5.5 (N ¼ 166). The patients had elevated liver function test results (ALT, 378.9 U/L ⫾ 737.7; AST, 331.3 U/L ⫾ 1,011.4; ALP, 342.9 U/L ⫾ 338.5; total bilirubin, 7.8 mg/dL ⫾ 9.4; all measurements recorded in all 166 patients). Preprocedural creatinine and BUN measurements were 1.1 mg/dL ⫾ 1.0 and 33.9 mg/dL ⫾ 28.2, respectively (N ¼ 166). Statistical analysis was performed with SigmaStat (version 2.03) statistical software (SPSS, Chicago, Illinois). Data are presented as means ⫾ standard deviation. A P value lower than .05 was considered to be statistically significant. When data in two groups passed the normality and equal variance tests, differences between the two
Volume 24
’
Number 1
’
January
’
2013
groups were assessed by Student t test. When a Student t test could not be performed, the two groups were compared with a Mann–Whitneyrank-sum test. When data in multiple groups passed the normality and equal variance tests, differences between multiple groups were assessed by one-way analysis of variance followed by Tukey comparison tests. When one-way analysis of variance could not be performed, groups were compared with Kruskal–Wallis one-way analysis of variance on ranks, followed by Dunn pairwise multiple comparison procedures. To study the relationship of the histopathologic diagnosis of VOD to the range of CHSPG values, a series of 2 2 contingency tables was constructed by using a Fisher exact test. For each of the pressure gradient values, the sensitivity, specificity, and positive and negative predictive values for finding VOD were calculated. Logistic regression analysis curves of specificity and sensitivity of CHSPG for VOD were constructed by MedCalc (version 12.2.1) statistical software (MedCalc, Mariakerke, Belgium).
87
Figure 1. CHSPG in HSCT recipients with different histopathologic diagnoses. ‘‘Others‘‘ indicates patients with a histopathologic diagnosis other than GVHD or VOD (*P o .05).
RESULTS During the study period, 166 transjugular liver biopsy procedures were performed in 141 patients who underwent HSCT. The technical success was 98.8% (164 of 166; one sample had insufficient tissue for diagnosis and one biopsy failed as a result of patient agitation, although pressure measurements were performed). There were three major complications (1.8%): one patient developed arteriobiliary fistula, which required embolization; one patient required blood transfusion within 24 hours after the procedure as a result of decreasing hematocrit level; and one patient experienced subcapsular and intraperitoneal hemorrhage leading to death within 36 hours. The overall mortality rate was 0.6%. Histologic analysis of the biopsy samples led to a diagnosis in 95.7% of cases (158 of 165; there was one nondiagnostic sample and six samples with nonspecific findings per pathology report). The histologic diagnosis was GVHD in 88 patients, VOD in 31 patients, coexistence of VOD with GVHD in 13 patients, other pathologic process (eg, lymphoma, steatohepatitis, drug toxicity, extramedullary hematopoiesis) in 27 patients, and nonspecific in six patients. Based on the histopathologic diagnosis, the patients were divided into four groups for further analysis of CHSPG data and laboratory values: patients with GVHD, patients with VOD, patients with coexistence of VOD with GVHD, and patients with another histologic diagnosis. The final group included patients with nonspecific histologic findings. The CHSPG was significantly higher (P o .001) in the VOD and VOD/GVHD groups compared with the other groups (Fig 1). When patients were divided into two groups—patients with VOD (n ¼ 44) and patients without VOD (n ¼ 121), the mean CHSPG in the VOD group (16.2 ⫾ 9.2) was significantly higher (P o .001) than in the non-VOD group (5.6 ⫾ 3.7).
Figure 2. Sensitivity and specificity curves of CHSPG for VOD. The solid and dashed thick lines represent sensitivity and specificity, respectively. The thin lines paralleling the respective thick lines represent 95% confidence intervals.
Logistic regression analysis curves of specificity and sensitivity of CHSPG for VOD are shown in Figure 2. Figure 2 demonstrates that a CHSPG of 10 mm Hg or higher provides 90.8% specificity and 77.3% sensitivity for the diagnosis of VOD. Specificity, sensitivity, and positive/ negative predictive values for the 9–11 mm Hg range of CHSPG are shown in Table 1. Comparison of the laboratory values of the GVHD and VOD groups are shown in Table 2. Patients with VOD had significantly higher INR, PT, AST, creatinine, and BUN values, whereas patients with GVHD had significantly higher ALP levels. There was no statistically significant difference in total bilirubin and ALT levels between the two groups.
88
’
Safety/Utility of Transjugular Liver Biopsy in HSC Transplants
DISCUSSION The number of patients receiving allogenic HSCT is increasing worldwide, and hepatic complications occur in a significant proportion of HSCT recipients (1,2). Despite human leukocyte antigen identity between a patient and donor, approximately 40% of recipients of human leukocyte antigen–identical grafts develop GVHD, and the liver is involved in 50% of these cases (2). Histologically, hepatic GVHD presents as lymphocytic infiltration of the portal areas, pericholangitis, endothelial injury, and bile duct destruction (12). VOD is characterized clinically by the triad of hepatomegaly, ascites/weight gain, and jaundice (13). VOD develops secondary to high-dose cytoreductive therapy in susceptible patients, resulting in endothelial injury in sinusoids and small hepatic venules with consequent activation of the coagulation cascade. This leads to clot formation and obliteration of the hepatic venules (13). The clinical manifestations of liver diseases caused by GVHD or VOD can mimic other hepatic complications of HSCT, such as viral hepatitis (herpes simplex virus, adenovirus, varicella virus, hepatitis B and C viruses), fungal infection, leukemic infiltration, drug toxicity, and cholestasis secondary to septicemia, total parenteral nutrition, or medications (4,6). GVHD and VOD require specific treatment, and early treatment improves the probability of treatment success (2,8,13). Therefore, timely and accurate diagnosis is essential. When the diagnosis is unclear based on the clinical picture, liver biopsy is required. Even with Table 1 . Specificity, Sensitivity, and Positive and Negative Predictive Values for Venoocclusive Disease based on CHSPG CHSPG (%) Statistic
4 8 mm Hg
4 9 mm Hg
Sensitivity
88.36
79.55
4 10 mm Hg 77.27
Specificity PPV
84.17 66.7
89.17 72.9
90.83 75.6
NPV
94.4
92.2
91.6
CHSPG ¼ corrected hepatic sinusoidal pressure gradient, NPV ¼ negative predictive values, PPV ¼ positive predictive values.
Kis et al
’
JVIR
a typical clinical presentation, biopsy may be necessary. According to the study of Esposito et al (14), 66% (14 of 21) of HSCT recipients had a different pathologic diagnosis than what was clinically suspected. Shulman et al (6) reported that liver histologic findings provided additional information that was beneficial for HSCT recipient management in 88% of cases (53 of 60 patients). Most of the time, HSCT recipients are referred for transjugular liver biopsy as a result of thrombocytopenia and/or coagulopathy. In addition, ascites commonly presents in HSCT recipients, which may defer percutaneous biopsy. Previous small-scale studies (5–7) of transjugular liver biopsy in HSCT recipients (the largest included 29 patients) reported major complication rates as low as 11% and as high as 80%. Our present retrospective analysis of 166 transjugular liver biopsies in 141 HSCT recipients revealed a 1.8% major complication rate and a 0.6% mortality rate. Reported complication rates of transjugular liver biopsies involving more than 100 patients range between 2.4% and 6% (15–17). In the present study, the technical success rate of transjugular liver biopsy in HSCT recipients was 98.8%. Biopsy samples led to histologic diagnosis in 95.7% of cases, which highlights the clinical utility of the procedure. Our results are comparable to transjugular liver biopsy data in the general patient population, which were reported to show a technical success rate of 96.8% and a histologic diagnosis rate of 96.1% in a metaanalysis of 64 series reporting 7,649 transjugular liver biopsies (18). The most frequent histologic diagnosis in our HSCT patient population was GVHD, followed by VOD, and this corresponds to data described in the literature (7). The distinction between GVHD and VOD is very important because GVHD requires high-dose immunosuppressive therapy (2), whereas the treatment of VOD primarily includes supportive care, fibrinolysis, and anticoagulation (13). The measurement of CHSPG has been reported in small series to discriminate between VOD and other liver pathologic processes in HSCT recipients (6). Our data in a relatively large patient population demonstrate again that measurement of CHSPG during the transjugular liver biopsy procedure can support the diagnosis of VOD.
Table 2 . Laboratory Values of Patients with GVHD and VOD Parameter
GVHD
VOD
P Value
1.16 ⫾ 0.25
1.73 ⫾ 0.55
o .05
0.9–1.1
34.03 ⫾ 15.56
43.47 ⫾ 8.26
o .05
23.8–36.6
ALT, U/L AST, U/L
409.86 ⫾ 445.29 212.81 ⫾ 277.70
322.39 ⫾ 441.15 483.84 ⫾ 586.31
.35 o .05
10–50 10–50
ALP, U/L
428.23 ⫾ 410.99
193.48 ⫾ 129.03
o .05
35–130
8.50 ⫾ 10.21 1.00 ⫾ 0.60
7.04 ⫾ 7.77 2.08 ⫾ 1.58
.47 o .05
0.0–1.0 0.5–1.2
27.82 ⫾ 26.84
48.81 ⫾ 25.53
o .05
6–23
INR PT, s
Total bilirubin, mg/dL Creatinine, mg/dL BUN, mg/dL
Normal Range
Values presented as mean ⫾ standard deviation. ALP ¼ alkaline phosphatase, ALT ¼ alanine aminotransferase, AST ¼ aspartate aminotransferase, BUN ¼ blood urea nitrogen, GVHD ¼ graft-versus-host disease, INR ¼ International Normalized Ratio, PT ¼ prothrombin time, VOD ¼ venoocclusive disease.
Volume 24
’
Number 1
’
January
’
2013
A CHSPG of 10 mm Hg or higher is 90.8% specific and 77.3% sensitive for VOD. Therefore, measurement of CHSPG can lead to initiation of specific treatment days before the final histologic diagnosis and potentially reduce morbidity and mortality in HSCT recipients. In addition, according to published data, CHSPG may be helpful in determining the prognosis of patients with VOD (6). Analysis of laboratory values for HSCT recipients who underwent transjugular liver biopsy at our institution showed that the patients had abnormal liver function test results, as expected. The average red blood cell count and platelet count were lower than normal range, and the INR and PT values were slightly increased. These results are not surprising in this patient population. There were a few interesting findings when the laboratory values of patients with GVHD were compared with those of patients with VOD. Patients with GVHD had significantly higher ALP levels. This finding corresponds to liver histologic findings in GVHD, which show lymphocytic infiltration of portal areas, pericholangitis, and bile duct destruction. In the liver, ALP is found in the epithelial cells lining the biliary tract (19), and it is well known that plasma ALP level is increased in pathologic conditions involving the biliary tree (20). Another interesting finding was the significantly increased creatinine and BUN levels in patients with VOD compared with those with GVHD. Increased risks of acute renal failure have been observed in patients with VOD, whereas GVHD was not associated with renal impairment (21). The relationship between VOD and renal impairment is not clear; decreased hepatic clearance of nephrotoxic endotoxins absorbed from the gut may play a role (22). The main deficiency of the present study is its retrospective nature. In addition, we report a single institution’s experience, and patient selection was at the preference of the stem cell transplant clinical team, which may have skewed our results. Despite these limitations, the value of our investigation lies in the size of our patient population, which, to our knowledge, is nearly five times the size of the next largest series. In summary, the present study supports the relative safety of transjugular liver biopsy in HSCT recipients. The measurement of CHSPG during the transjugular liver biopsy procedure can provide important diagnostic information days before the availability of the pathologic diagnosis, and may accelerate the initiation of diseasespecific therapy. This additional information makes transjugular liver biopsy the preferred route even in patients who otherwise would be eligible for percutaneous biopsy, particularly when VOD is a consideration in the differential diagnosis.
89
REFERENCES 1. Gratwohl A, Baldomero H, Aljurf M, et al. Hematopoietic stem cell transplantation: a global perspective. JAMA 2010; 303:1617–1624. 2. Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet 2009; 373:1550–1561. 3. Locasciulli A, Alberti A, de Bock R, et al. Impact of liver disease and hepatitis infections on allogeneic bone marrow transplantation in Europe: a survey from the European Bone Marrow Transplantation (EBMT) Group—Infectious Diseases Working Party. Bone Marrow Transplant 1994; 14:833–837. 4. Carreras E, Granena A, Navasa M, et al. On the reliability of clinical criteria for the diagnosis of hepatic veno-occlusive disease. Ann Hematol 1993; 66:77–80. 5. Oshrine B, Lehmann LE, Duncan CN. Safety and utility of liver biopsy after pediatric hematopoietic stem cell transplantation. J Pediatr Hematol Oncol 2011; 33:e92–e97. 6. Shulman HM, Gooley T, Dudley MD, et al. Utility of transvenous liver biopsies and wedged hepatic venous pressure measurements in sixty marrow transplant recipients. Transplantation 1995; 59:1015–1022. 7. Chahal P, Levy C, Litzow MR, Lindor KD. Utility of liver biopsy in bone marrow transplant patients. J Gastroenterol Hepatol 2008; 23:222–225. 8. Mahgerefteh SY, Sosna J, Bogot N, Shapira MY, Pappo O, Bloom AI. Radiologic imaging and intervention for gastrointestinal and hepatic complications of hematopoietic stem cell transplantation. Radiology 2011; 258:660–671. 9. Smith TP, Kim CY, Smith AD, Janas G, et al. Hepatic venous pressure measurements: comparison of end-hole and balloon catheter methods. J Vasc Interv Radiol 2012; 23:219–226. 10. Maleux G, Willems E, Fieuws S, et al. Prospective study comparing different indirect methods to measure portal pressure portal pressure. J Vasc Interv Radiol 2011; 22:1553–1558. 11. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14(suppl):S199–S202. 12. Fujii N, Takenaka K, Shinagawa K, et al. Hepatic graft-versus-host disease presenting as an acute hepatitis after allogeneic peripheral blood stem cell transplantation. Bone Marrow Transplant 2001; 27:1007–1010. 13. Senzolo M, Germani G, Cholongitas E, Burra P, Burroughs AK. Venoocclusive disease: update on clinical management. World J Gastroenterol 2007; 13:3918–3924. 14. Esposito AA, Nicolini A, Meregaglia D, Sangiovanni A, Biondetti P. Role of transjugular liver biopsy in the diagnostic and therapeutic management of patients with severe liver disease. Radiol Med 2008; 113:1008–1017. 15. Bull HJ, Gilmore IT, Bradley RD, Marigold JH, Thompson RP. Experience with transjugular liver biopsy. Gut 1983; 24:1057–1060. 16. Choh J, Dolmatch B, Safadi R, et al. Transjugular core liver biopsy with a 19-gauge spring-loaded cutting needle. Cardiovasc Intervent Radiol 1998; 21:88–90. 17. Smith TP, Presson TL, Heneghan MA, Ryan JM. Transjugular biopsy of the liver in pediatric and adult patients using an 18-gauge automated core biopsy needle: a retrospective review of 410 consecutive procedures. AJR Am J Roentgenol 2003; 180:167–172. 18. Kalambokis G, Manousou P, Vibhakorn S, et al. Transjugular liver biopsy—indications, adequacy, quality of specimens, and complications— a systematic review. J Hepatol 2007; 47:284–294. 19. Jacoby F, Martin BF. The relationship of bile alkaline phosphatase to histochemically detectable alkaline phosphatase in the biliary tract; including reference to the histology of the gall-bladder epithelium. J Anat 1951; 85:391–399. 20. Popper H, Schaffner F. Pathophysiology of cholestasis. Hum Pathol 1970; 1:1–24. 21. Hingorani SR, Guthrie K, Batchelder A, et al. Acute renal failure after myeloablative hematopoietic cell transplant: incidence and risk factors. Kidney Int 2005; 67:272–277. 22. McDonald GB, Hinds MS, Fisher LD, et al. Veno-occlusive disease of the liver and multiorgan failure after bone marrow transplantation: a cohort study of 355 patients. Ann Intern Med 1993; 118:255–267.