Hepatic veno-occlusive disease (VOD) is a serious early complication of intensive myeloablative chemotherapy or radiation therapy before hematopoietic stem cell transplantation for malignancies. Histologically, VOD leads to nonthrombotic obstruction of centrolobular veins by subendothelial connective tissue and centrolobular hepatocyte necrosis (1
The incidence of VOD is relatively high in adults (54%) (2 3 2 4 5 6 2 7 8
Doppler ultrasonography (DUS) is a noninvasive method that can be used at the bedside in a sterile room. It can be repeated as often as necessary, does not require any preparation, does not give rise to complications, and is very well accepted by patients. US can be used to study organ morphology (especially the liver) and can be used to visualize vessels and measure hemodynamic parameters of arterial or venous flow (flow rates and resistance) when coupled with Doppler imaging. The method is suitable for studying VOD-induced morphologic and hemodynamic changes (9 9
MATERIALS AND METHODS
Patients
Since 1992, 71 patients who required hematopoietic stem cell transplantation after intensive myeloablative therapy that contained busulfan, which is known to be a major risk factor for VOD, have been evaluated with hepatic US and Doppler before transplantation and then at least once, at the bedside, during hospitalization after transplantation. Patients who developed VOD from 1993 to 1998 were included in this study. Patients were hospitalized in laminar-flow rooms in the transplantation unit in the pediatrics department. The preparative regimen consisted of 600 mg/m2 busulfan os over 4 days combined with another alkylating agent. All patients underwent autologous transplantation with either bone marrow or peripheral blood hematopoietic stem cells.
Methods
DUS was performed using a mobile Hitachi EUB 450 or a Toshiba Ecocee device equipped with a 3.5-MHz probe. The average duration of each examination was approximately 20 min. Seven US and seven Doppler criteria were recorded at each examination.
Morphological US criteria were as follows: (1) liver size, the vertical axis of the liver was measured on the preaortic, right midclavicular, and prerenal lines (normal values <7 cm, 8 cm, and 11 cm, respectively). An increase in at least 2 of 3 measurements >1 cm relative to the baseline measurements was considered as enlargement; (2) spleen size, >1 cm relative to the baseline measurement of the greatest axis (normal value <11 cm); (3) gallbladder wall thickening, a value >6 mm was used as a criterion for marked wall thickening; (4) portal vein diameter, a diameter >8 mm was considered enlarged; (5) the diameter of the hepatic vein (HV) measured 2 cm from the inferior vena cava. A diameter <3 mm for the 3 HVs was considered as markedly reduced; (6) ascites; and (7) visualization of the PUV. A hypoechoic lumen in the hyperechoic ligamentum teres indicated recanalization. One point was assigned to each criterion and a morphological US score of 0 to 7 was obtained when criteria were summed.
The seven Doppler criteria were as follows: (1) portal flow, a maximal velocity <10 cm/sec (normal mean value >11 cm/sec), or hepatofugal flow; (2) flow demodulation, i.e., the disappearance of normal velocity variations with breathing; (3) a substantial decrease in spectral density that reflects a decline in the amount of red cells in portal flow; (4) a congestion index (cross-sectional area of vein over average velocity) >0.1 (normal mean value 0.070±0.09); (5) flow in the PUV, flow (hepatofugal) recorded with pulsed Doppler in the recanalized vein (color Doppler was not available with the mobile sonograph used for the examinations in the sterile rooms); (6) flow in the HVs, monophasic flow or no flow recorded with pulsed Doppler. Flow in the HV is normally triphasic and modulated by cardiac contractions; and (7) arterial hepatic flow, a resistive index (RI, systolic velocity minus diastolic velocity over systolic velocity) >0.75, which reflects a marked increase in resistance (normal mean value 0.64±0.06).
A point was attributed for each criterion, and a Doppler score of 0 to 7 was obtained when criteria were summed. The sum of US criteria and Doppler criteria yielded a total score of 0 to 14.
It was not possible to determine US and Doppler scores for some patients because one or several criteria were missing (pregraft examination 57 US scores, 45 Doppler scores, and 42 DUS scores; postgraft examination 64 US scores, 46 Doppler scores, and 42 US-Doppler scores).
The diagnosis of VOD was based on clinical criteria and on biological data (elevated bilirubin, a decrease in factor VII, and low posttransfusion platelet recovery). The bilirubin level was measured in each patient when the posttransplantation DUS examination was performed. We compared the DUS scores and the bilirubin values to determine which were the most predictive of severity at the time of the clinical diagnosis. The VOD grade was determined retrospectively on the basis of outcome.
Statistical Analysis
The chi-squared test and the exact test were used to compare morphological and Doppler criteria between the three VOD grades. Grade group scores were compared by the Kruskal-Wallis test. The multivariate analysis of the US and Doppler criteria for the severity of VOD was performed by polytomous logistic regression with the proportional odds assumption, which is in accordance with the data.
RESULTS
Seventy-one patients (52 males and 19 females) with a median age of 4 years (12 months–20 years) were included. All patients received treatment for malignant disease (33 for neuroblastoma, 23 for medulloblastoma, 13 Ewing’s for family tumors, 1 for ependymoma, and 1 for Wilms’ tumor). VOD was clinically suspected at a median time of 16 days posttransplantation (range 5–34 days).
Seventeen patients (24%) had grade 1 disease, 34 patients (48%) had grade 2 disease, and 20 patients (28%) had grade 3 disease. Two of the 71 (3%) patients died of hepatic failure. Ninety-three percent of these patients had low posttransfusional platelet recovery. The median value of factor VII was 64, 66, and 57 in the three groups, respectively. Bilirubin was elevated (>34 μmol/L) in 24%, 26%, and 65% of patients in the grade 1, 2, and 3 groups, respectively. With a threshold value of 100 μmol/L at the time of the diagnosis of VOD, the specificity and the positive predictive value (PPV) of bilirubin for grade 3 disease was 100% (4 patients) but the sensitivity was only 20% and the negative predictive value (NPV) was 76%. With a threshold value of 50 μmol/L, the specificity was 92%, the PPV was 64%, the sensitivity was still low at 35%, and the NPV was 78% (Fig. 1 ).
Figure 1: Bilirubin level at diagnosis of VOD according to VOD grade.
Sixty-four of the 71 patients had US and Doppler examinations before the graft (Table 1 ). Three of the seven US criteria (gallbladder wall >6 mm, ascites, and visualization of the PUV) and three of the seven Doppler criteria (portal flow <10 cm/sec, flow recorded in PUV, and a decrease in portal flow density) were never observed during the pregraft examinations. Conversely, two criteria were relatively frequently above normal before the graft (portal vein diameter >8 mm in 7 patients and hepatic artery RI >0.75 in 22 patients).
Table 1: Table 1. Results of the 64 pregraft US and Doppler examinations and of the 71 examinations after transplantation (in the first week of clinical suspicion of VOD)
All patients had US and Doppler examinations after the transplantation. Examinations were performed from 0 to 8 days after the onset of the first clinical and biological signs and suspicion of the diagnosis of VOD (7 to 35 days after the transplantation, median 21 days).
Compared to the pregraft examination, all US and Doppler criteria had increased. The median US, Doppler, and total DUS scores had increased after the transplantation from 0 to 4.5 (range 1–7), from 1 to 3.5 (range 0–7), and from 1 to 9 (range 1–13), respectively (Table 1 ).
In the univariate analysis, three US and three Doppler criteria were correlated with the severity of VOD (Table 2 ). Among the US criteria, ascites was by far the most significant criterion. It was observed in 100% of patients with grade 3 VOD. Although hepatomegaly was frequently observed, whatever the grade of VOD, splenomegaly was significantly more frequent among patients with grade 3 (90%) and 2 (82%) VOD than among patients with grade 1 (5%) disease. Visualization of the PUV was more frequently observed as the grade increased: 18% in grade 1, 32% in grade 2, and 60% in grade 3. All Doppler criteria were more frequently observed in patients with grade 3 VOD. Decreased portal flow velocity or reversed flow, lowered spectral density, and flow recorded in the PUV were the three criteria significantly associated with the severity of VOD. The multivariate analysis (Table 3 ) of US criteria showed that splenomegaly (P =0.003) and ascites (P =0.006) were independently associated with the VOD grade. Among the Doppler criteria, flow recorded in the PUV was the only criterion independently associated with the severity of VOD. The multivariate analysis of pooled US and Doppler criteria showed that the flow recorded in the PUV was the only criterion significantly associated with the VOD grade (P =0.0001).
Table 2: Table 2. US-DOPPLER examination at diagnosis of VOD (in the first week), according to VOD grade
Table 3: Table 3. Multivariate analysis using polytomous logistic regression:odds ratio (OR) and 95% confidence intervals (CI)
Two scores were calculated: the total score using all US and Doppler criteria and the reduced score using only the two independent US predictive factors (splenomegaly and ascites) and the one independent Doppler criterion (flow recorded in the PUV). The threshold chosen as predictive of grade 3 VOD was ≥9 for the total score and ≥2 for the reduced score. With these thresholds, the sensitivity and the NPV of the two scores were 100%, but the specificity and the PPV of the reduced score was only 49% and 44% while the specificity and the PPV of the total score was 69% and 59% (Figs. 2 and 3 ). Among the 22 patients with a total DUS score ≥9, 13 patients (59%) developed grade 3 VOD, 8 patients (36%) developed grade 2 VOD, and 1 patient developed grade 1 VOD. None of the 20 patients with a score <9 developed grade 3 VOD (Figs. 2 and 3 ). The total US score, the total Doppler score, and the total DUS scores were also strongly correlated with the severity of VOD (Table 2 ).
Figure 2: Doppler ultrasonography score according to VOD grade. Total score included (7 US Criteria and 7 Doppler criteria.)
Figure 3: Reduced score, 2 US criteria (splenomegaly and ascites) and 1 Doppler criterion (flow recorded in PUV).
Among the four patients whose bilirubin level was >100 μmol/L, the DUS score was 13, 11, and 9 at the same time in three of the patients. It was not possible to calculate the score for the remaining patient because one criterion was missing. With the 13 criteria available, the total score for this patient was 5.
The two patients who died of hepatic failure had grade 3 disease, a bilirubin level of 172 and 134 μmol/L, respectively, and a Doppler-US score of 9 and 13, respectively, at the time of the diagnosis.
DISCUSSION
VOD is a frequent complication of hematopoietic stem cell transplantation, with an incidence that ranges from 21% to 54% and a mortality rate as high as 47% in adults (2,3,6,10,11 10 12 13 9,14,15 16–20
In a previous study of 100 patients (65 adults and 35 children), we showed that seven US and seven Doppler criteria were all significantly linked to the diagnosis of VOD (9 21–26
The aim of the present study was to determine the value of the US and Doppler criteria for the early prediction of the severity of VOD in children. Although the initial pretransplantation examination was not predictive, it was important because it provided baseline morphologic information mainly on the size of the liver and spleen and on the diameter of the HV. Because hemodynamic criteria are less important at this stage, the duration of the pregraft examination can be shortened because it will focus exclusively on morphologic criteria that are useful for the follow-up.
The postgraft DUS examinations were performed during the first days after the onset of dissociated biological and clinical signs that lead to suspicion of VOD (vague abdominal pain, weight gain, and low posttransfusion platelet recovery).
The DUS criteria we chose to assess were based on our previous study and not on the threshold of other studies. In a case report, Kriegshauser et al. (21 22
In previous studies, Doppler was used to evaluate the consequences of VOD on the portal vein, HVs, and hepatic artery flow. Brown et al. (23 24 25 26 27
We tested seven Doppler criteria in children. Flow in the PUV was never recorded in patients with grade 1 VOD. In the multivariate analysis of Doppler criteria and Doppler-US criteria, this criterion, which indicates major portal hypertension, appeared to be critical in establishing the VOD grade. In fact, it appeared to be the only Doppler criterion significantly correlated with the VOD grade.
According to Shulman et al.(28 8 6
A high HVPG is known to be a poor prognostic factor. Carreras et al. (29 6
The serum bilirubin level is also considered a prognostic factor, but data differ between children and adults. With regard to the prognosis of VOD, a bilirubin level that exceeds 150 μmol/L is generally used as a severity threshold. In the series described by Jones et al.(30 31
The results of our univariate analysis showed that three US criteria and three Doppler criteria were associated with the grade of VOD and that the three scores calculated at the time of the first clinical and biological signs of the disease had a predictive value for the severity of VOD. In the multivariate analysis, only two US criteria and one Doppler criterion were independently associated with grade 3. Calculating the Doppler-US score with the 14 criteria and not simply with the independent criterion appears to be of greater interest. No patient with a score <9 developed grade 3 VOD. If the reduced score with the three criteria is used to select all patients who develop grade 3 VOD, 2 is the score that must be chosen but then it is not discriminating. No patient with a score <9 at the time of the first clinical and biological signs of the disease developed grade 3 disease, whereas 25% of the patients who developed grade 3 VOD had a normal bilirubin level at the same time. The probability of developing grade 3 VOD was 59% (95% CI, 36–79%) if the Doppler-US score was greater than or equal to 9. Comparatively, the PPV of developing grade 3 VOD was 64% for a bilirubin level >50 μmol/L and 100% for a value >100 μmol/L. However, there were, respectively, 64% and 80% of false negatives among patients with grade 3 VOD with these thresholds of bilirubin and 0% with this threshold of Doppler-US score.
It is, thus, possible to predict the risk of high-grade VOD early using the Doppler-US score and then to envisage curative treatment during the first phase of the disease. Moreover, Doppler-US is a simple technique for the follow-up of patients. The status of the morphologic and Doppler criteria on repeated examination is the first objective sign of progression or of the beginning of regression of the disease. Doppler-US permits early evaluation of curative treatment, which is always dangerous for these patients. In fact, there is currently no curative therapy that has been fully approved, but different treatments have been attempted such as portocaval shunt (32 33 34
Follow-up of patients who received prophylactic treatment against VOD, in particular, ursodeoxycholic acid (35
The results of our study showed that the early postgraft Doppler-US performed at the time of clinical suspicion of VOD could predict the grade of VOD. The Doppler-US scores were highly instrumental in establishing the grade of VOD as early as the first clinical and biological signs. This noninvasive technique can predict the severity of VOD at the time of the diagnosis with more sensitivity than the bilirubin level and the HVPG. Furthermore, Doppler-US examinations can be repeated as often as necessary to monitor both the course and outcome of VOD and can contribute to decision making on treatment and on the efficacy of treatment.
Acknowledgment.
We thank Lorna Saint-Ange for editing.
REFERENCES
1. Shulman HM, Gown AM, Nugent DJ. Hepatic veno-occlusive disease after bone marrow transplantation. Am J Pathol 1987; 127: 549.
2. 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.
3. Meresse V, Hartmann O, Vassal G, et al. Risk factors for hepatic veno-occlusive disease after high-dose busulfan-containing regimens followed by autologous bone marrow transplantation: a study in 136 children. Bone Marrow Transplant 1992; 10: 135.
4. Bearman SI. The syndrome of hepatic veno-occlusive disease after marrow transplantation. Blood 1995; 85: 3005.
5. DeLeve L, McCuskey R, Wang X, et al. Characterisation of a reproducible rat model of hepatic veno-occlusive disease. Hepatology 1999; 29: 1779.
6. McDonald GB, Sharma P, Matthews DE, Shulman HM. Veno-occlusive disease of the liver after bone marrow transplantation: diagnosis, incidence and predisposing factors. Hepatology 1984; 4: 116.
7. Shulman HM, McDonald GB. Transvenous liver biopsies and pressure measurements in bone marrow transplant recipients. Hepatology 1992; 16: 148.
8. Carreras E, Granena A, Navasa M, et al. Transjugular liver biopsy in BMT. Bone Marrow Transplant 1993; 11: 21.
9. Lassau N, Leclere J, Auperin A, et al. Hepatic veno-occlusive disease after myeloablative treatment and bone marrow transplantation: value of Gray-Scale and Doppler US in 100 patients. Radiology 1997; 204: 545.
10. Brugieres L, Hartmann O, Benhamou E, et al. Veno-occlusive disease of the liver following high-dose chemotherapy and autologous bone marrow transplantation in children with solid tumors: incidence, clinical course and outcome. Bone Marrow Transplant 1988; 3: 53.
11. Bearman SI, Anderson GL, Mori M, Hinds MS, Shulman HM, McDonald GB. Veno-occlusive disease of the liver: development of a model for predicting fatal outcome after marrow transplantation. J Clin Oncol 1993; 11: 1729.
12. Atkinson K, Biggs J, Noble G, Ashby M, Concannon A, Dodds A. Preparative regimens for marrow transplantation containing busulfan are associated with hemorrhagic cystitis and hepatic veno-occlusive disease but a short duration of leucopenia and little oro-pharyngeal mucositis. Bone Marrow Transplant 1987; 2: 385.
13. Locasciulli A, Bacigalupo A, Albert A, et al. Predictability before transplant of hepatic complications following allogeneic bone marrow transplantation. Transplantation 1989; 48: 68.
14. Matsuishi E, Anzai K, Dohmen K, et al. Sonographic diagnosis of venoocclusive disease of the liver and Danazol therapy of autoimmune thrombocytopenia in an autologous marrow transplant patient. J Clin Oncol 1990; 20: 188.
15. Nelson RC, Lovett K, Chezmar JL, et al. Comparison of pulsed Doppler sonography and angiography in patients with portal hypertension. Am J Roentgenol 1987; 149: 77.
16. Patriquin H, Lafortune M, Burns P, Dauzat M. Duplex Doppler examination in portal hypertension: technic and anatomy. AJR 1987; 149: 71.
17. Moriyasu F, Nishida O, Ban N, et al. “Congestion Index” of the portal vein. AJR 1986; 146: 735.
18. Lafortune M, Constantin A, Breton G, Legare AG, Lavoie P. The recanalized umbilical vein in portal hypertension: a myth. AJR 1985; 144: 549.
19. Mostbeck GH, Wittich GR, Herold C, et al. Hemodynamic significance of the paraumbilical vein in portal hypertension: assessment with duplex US. Radiology 1989; 170: 339.
20. Gibson RN, Gibson PR, Donlan JD, Clunie DA. Identification of a patent paraumbilical vein by using Doppler sonography: importance in the diagnosis of portal hypertension. AJR 1989; 153: 513.
21. Kriegshauser JS, Charboneau JW, Letendre L. Hepatic venoocclusive disease after bone-marrow transplantation: diagnosis with duplex sonography. Am J Roentgenol 1988; 150: 289.
22. Schiavetti A, Matrunola M, Varasso G, Padula A, Castello MA. Ultrasound in the management of hepatic veno-occlusive disease in three children treated with dactinomycin and vincristine. Pediatr Hematol Oncol 1996; 13: 521.
23. Brown BP, Abu-Yousef M, LaBrecque D, Gingrich R. Doppler sonography: a noninvasive method for evaluation of hepatic veno-occlusive disease. Am J Roentgenol 1990; 154: 721.
24. Herbetko J, Grigg AP, Buckley AR, Phillips GL. Veno-occlusive liver disease after bone marrow transplantation: findings at duplex sonography. Am J Roentgenol 1992; 158: 1001.
25. Teefey SA, Brink JA, Borson RA, Middleton WD. Diagnosis of venoocclusive disease of the liver after bone marrow transplantation: value of duplex sonography. Am J Roentgenol 1995; 164: 1397.
26. Hommeyer SC, Teefey SA, Jacobson AF, et al. Veno-occlusive disease of the liver: prospective study of US evaluation. Radiology 1992; 184: 683.
27. McCarville MB, Hoffer FA, Howard SC, Goloubeva O, Kauffman WM. Hepatic veno-occlusive disease in children undergoing bone-marrow transplantation: usefulness of sonographics findings. Pediatr Radiol 2001; 31: 102.
28. Shulman HM, Hinterberger W. Hepatic veno-occlusive disease, liver toxicity syndrome after bone marrow transplantation. Bone Marrow Transplant 1992; 10: 197.
29. Carreras E, Granena A, Rozman C. Hepatic veno-occlusive disease after bone marrow transplant. Blood Rev 1993; 7: 43.
30. Jones RJ, Lee KSK, Beschorner WE, et al. Veno-occlusive disease of the liver following bone marrow transplantation. Transplantation 1987; 44: 778.
31. Mac Donald GB, Hinds M, Fisher L, Schoch HG. Liver toxicity following cytoreductive therapy for bone marrow transplantation: risk factors, incidence, and outcome. Hepatology 1991; 14: 459.
32. Murray JA, LaBrecque DR, Gingrich RD, Pringle KC, Mitros FA. Successful treatment of hepatic veno-occlusive disease in a bone marrow transplant patient with side-to-side portacaval shunt. Gastroenterol 1987; 92: 1073.
33. Ibrahim A, Pico JL, Maraninchi D, et al. Hepatic veno-occlusive disease following bone marrow transplantation treated by prostaglandin E1. Bone Marrow Transplant 1991; 7: 53.
34. Bearman SI, Shuhart MC, Hinds MS, McDonald GB. Recombinant human tissue plasminogen activator for the treatment of established severe veno-occlusive disease of the liver after bone marrow transplantation. Blood 1992; 80: 2458.
35. Essel JH, Schroeder MT, Harman GS, et al. Ursodiol prophylaxis against hepatic complications of allogenic bone marrow transplantation: a randomized, double blind, placebo controlled trial. Ann Intern Med 1998; 128: 975.
