Pediatric Transplants

Risk factors and mortality predictors of hepatic veno-occlusive disease after pediatric hematopoietic stem cell transplantation


A cohort of 138 children with 144 hematopoietic stem cell transplantation (HSCT) performed in 1997–2006 were analyzed to evaluate risk factors and mortality predictors of hepatic veno-occlusive disease (VOD). Nineteen patients (13.2%) developed VOD (nine boys, median age 3.5 years) at 1–21 days after HSCT (median 13 days). Age 2 years at transplant (odds ratio (OR)=5.25, P=0.011), BU–CY conditioning (OR=5.16, P=0.001), thalassemia major (OR=3.97, P=0.015), platelet engraftment beyond day +21 (OR=8.67, P=0.025) were univariate risk factors for VOD. The first two remained significant in multivariate regression. Seven patients (36.8%) with VOD died, at a median of 44 days post transplant (range, 30–421 days). The 5-year survival was 62%. All surviving patients had normal liver function on follow-up at 0.5–9 years. Patients with VOD had higher 100-day mortality (16.3 vs 9.6%, P=0.024). Mortality predictors included donors other than autologous or matched sibling (hazard ratio (HR)=23.6, P=0.006), hepatic and cutaneous GVHD (HR=8.15, P=0.038), maximal weight gain >9% (HR=6.81, P=0.023), pleural effusion, intensive care unit admission, peak bilirubin >300 μmol l−1 (HR=13.6, P=0.016), day +21 bilirubin >200 μmol l−1 (HR=33.9, P=0.001), and rise of bilirubin >15 μmol l−1 per day within the first week (HR=19.8, P=0.006). Mortality was substantially higher if >3 predictors were present (HR=33.9, P=0.001). Meticulous monitoring in high-risk patients and early treatment should be considered before VOD progresses beyond salvage.


Hepatic veno-occlusive disease (VOD), also known as sinusoidal obstruction syndrome, is one of the important complications of hematopoietic stem cell transplant (HSCT). Its incidence ranges from 5 to 39% in pediatric patients.1, 2 The cause of hepatic VOD is not entirely clear but a number of risk factors have been reported.3, 4, 5, 6 While many cases of mild VOD resolve spontaneously without complications, moderate-to-severe VOD may result in significant morbidity or even mortality. The management of VOD is mainly supportive, including fluid restriction and diuretics to ameliorate fluid retention and cardio-respiratory support if necessary. Anticoagulants and thrombolytic therapies were tried with only limited success. A newer agent, defibrotide, is increasingly used and has been reported to be useful in several studies.7, 8 Since only a limited number of studies on hepatic VOD after pediatric HSCT are currently available, we would like to evaluate the incidence, management and outcomes of hepatic VOD in children after HSCT in our center, with particular focus on risk factors and mortality predictors of hepatic VOD.

Materials and methods

Study design and participants

This was a retrospective review of all pediatric HSCT performed in Queen Mary Hospital, a University-affiliated quaternary referral center in Hong Kong, over the past 10 years (January 1997 – December 2006). Most of the patients are Chinese, and a few patients are of Southeast Asian ethnicities. The patients' data on demographic and clinical characteristics, progress in the post transplant period including occurrence and treatment of hepatic VOD, and final outcomes were extracted. Apart from centrally stored hospital records, patients' information was retrieved and verified from our Hematology–Oncology–Immunology database through our Departmental Computer Server. Additional relevant clinical information including laboratory results was also retrieved by the Clinical Management System (CMS) of the Hospital Authority Server. All patients were nursed in air-filtered rooms with laminar airflow from the start of conditioning until neutrophil engraftment. Standard protocols were adopted for treatment of specific diseases, infection and GVHD prevention. No prophylaxis for hepatic VOD was used.


Hepatic VOD was defined by the Baltimore criteria9 that includes elevated serum total bilirubin 2 mg dl−1 (34 μmol l−1) before day 21 after HSCT together with two of the following three criteria: (1) tender hepatomegaly; (2) weight gain >5% from baseline and (3) ascites. Resolution of VOD was defined as normalization of serum bilirubin, resolution of ascites and return of body weight to baseline value. The severity of VOD was defined according to established clinical criteria which are mild for VOD that resolved without intervention, moderate for VOD that required treatment and severe for VOD that progressed to multiorgan failure or death.10 Acute GVHD was graded according to the Seattle grading system.11 Neutrophil engraftment was defined as neutrophil count rising from trough to 0.5 × 109 l−1 for 2 consecutive days. Platelet engraftments were defined by two time points, with platelet count consistently and spontaneously rising above 20 × 109 and 50 × 109 l−1 as indicators. Respiratory failure was defined as the need for mechanical ventilation. Renal impairment was defined as an elevated serum creatinine above two times the upper limit of the normal.

Risk factors and mortality predictors of VOD

All of the risk factors and most of the mortality predictors analyzed in the current study were selected a priori. The risk factors included standard transplant variables that were potentially associated with VOD. These were age and gender of the patients, the type of transplant, donor and source of stem cells, underlying diagnosis, viral serostatus before transplant, conditioning and GVHD prophylaxis regimen, nucleated cell dose, timing of neutrophil and platelet engraftment, and presence or absence of GVHD. In addition to the above, some risk factors were selected based on previous reports in the literature, including interval between diagnosis and transplant in malignant diseases, whether total parenteral nutrition had been given within 30 days before transplant, and whether the patient had previous history of pancreatitis. In addition to the above, mortality predictors analyzed included degree of weight gain, peak bilirubin and ALT levels, presence or absence of reverse portal venous blood flow, ascites, pleural effusion, renal impairment, respiratory failure, and requirement of intensive care unit (ICU) admission. Moreover, a few mortality predictors were selected to be analyzed after inspection of the data revealed potential significance. These included whether or not the rate of rise of bilirubin was greater than 15 μmol l−1 per day, and whether or not the bilirubin level was greater than 200 μmol l−1 at day +21 post transplant.

Statistical analyses

The risk factors for hepatic VOD were determined by univariate and multivariate analyses of various clinical and transplant factors among patients with and without hepatic VOD. In univariate analyses, χ2 tests or Fisher's exact tests were used for categorical variables where appropriate and odds ratios (ORs) with 95% CIs were determined. Mann–Whitney U-test was used for testing continuous variables. Logistic regression with forward and backward stepwise variable selection was used for multivariate analysis which included factors found to be significant in univariate analyses. Overall survival was estimated by the Kaplan–Meier method. Survival in patients with or without VOD were compared by log-rank test. The mortality predictors of VOD were determined by univariate and multivariate analyses by Cox regression. P<0.05 was considered statistically significant. All statistical analyses were carried out by the SPSS 11.0 software (SPSS Inc. 2003). This study has been approved by the Institutional Review Board of the Hospital Authority of Hong Kong which complies with the Declaration of Helsinki.


From 1997 to 2006, we had a total of 144 transplants performed in 138 patients. Eighty-two (59.4%) were boys and 56 (40.6%) were girls. The median age at transplant was 6.9 years (range, 0.2–18 years). Allogeneic transplants constituted 70.8% and the remaining were autologous transplants. The types of underlying disease, donor, source of hematopoietic stem cell, conditioning and GVHD prophylaxis regimen are shown in Table 1.

Table 1 Demographic and clinical characteristics of all HSCT recipients

Incidence and characteristics of VOD

Hepatic VOD occurred in 19 of all transplants (13.2%). The cumulative incidence of VOD was 13.4% at day +100. All 19 VOD events occurred in different patients. Nine (47.4%) were boys and 10 (52.6%) were girls. The median age at transplant was 3.5 years (range, 0.4–16.4 years). Nine patients had mild disease (47.4%); four patients had moderate disease (21.1%) and the remaining six patients had severe disease (31.6%). The characteristics of these patients are shown in Table 2. None of the patients with VOD had received abdominal radiotherapy. The median onset of VOD was 13 days post transplant (range, 1–21 days). The median duration of VOD for those who survived was 33 days (range, 8–95 days). The median peak bilirubin level was 102 μmol l−1 (range, 34–920 μmol l−1). All patients had hepatomegaly and substantial weight gain above 5%. The median maximal weight gain was 8.6% (5.1–20.8%). Thirteen patients had ascites (68.4%), eight patients had pleural effusion (42.1%) and one patient had pericardial effusion (5.3%). Renal impairment, respiratory failure and disseminated intravascular coagulation were developed in 5 (26.3%), 4 (21.1%) and 2 (10.5%) patients respectively. Pleural effusion and renal impairment with or without respiratory failure occurred in five of six patients with severe VOD and in all who died of VOD. Fifteen patients had ultrasound scan of the liver performed. Increased periportal echogenicity and splenomegaly were both present in nine patients (47.4%). Reversal of the direction of portal venous blood flow was demonstrated in six patients (40%), at a median of 6 days after onset of VOD (range, 1–29 days). Reversal of portal venous flow was associated with increased severity of VOD (P for trend=0.042), with incidence of 16.7, 50.0 and 80.0% in mild, moderate and severe VOD respectively.

Table 2 Clinical characteristics, treatment and outcomes of patients with hepatic VOD

Risk factors for VOD

There was no gender difference in the incidence of VOD (Table 3). The risk of VOD appeared to decrease with age, as the risk of VOD was 30.4% in patients aged below 2 years, 12.5% in patients aged 2–8 years, and 7.7% in patients above 8 years (P for trend=0.011). Although not statistically significant, allogeneic transplants appeared to have a higher risk of VOD compared to autologous transplants (16.7 vs 4.8%, OR=4.00, 95% CI 0.88–18.2, P=0.06); and all severe VOD occurred in allogeneic transplant recipients.

Table 3 Univariate analyses of risk factors and mortality predictors of hepatic VOD

Patients with thalassemia major had a significantly higher risk of VOD compared to other patients (30.4 vs 9.9%, OR=3.97, 95% CI 1.36–11.6, P=0.015). However, the pretransplant serum ferritin level, Lucarelli risk class, grading of hemosiderosis or presence of fibrosis on liver biopsy were not significantly associated with the risk of VOD (P=0.84, 0.16, 0.33 and 0.72, respectively). Other diagnostic categories were not at increased risk of VOD. None of our 17 transplanted neuroblastoma patients developed VOD despite very intensive pretransplant chemotherapy and the incidence of VOD was not significantly different between patients with solid tumor and other diseases (5.3 vs 16.0%, P=0.09). The incidence of VOD was also similar in leukemic and non-leukemic patients (10.4 vs 14.6%, P=0.86), as well as in patients with or without advanced malignancy (11.8 vs 14%, P=0.71).

On the other hand, the use of BU in the conditioning regimen was associated with a substantially increased risk of VOD (OR=4.49, 95% CI 1.52–13.3, P=0.004). The OR of VOD was even higher when BU was combined with CY (OR=5.16, 95% CI 1.74–15.3, P=0.001). There was no significant difference in the risk of VOD between patients who received BU by the oral route and the i.v. route (20.7 vs 50.0%, P=0.22), or between patients who received BU at a total dose above and below 16 mg kg−1 (25.9 vs 20.0%, P=0.58). Conditioning with fludarabine, melphalan or TBI was not associated with increased risk of VOD (P=0.87, 0.92 and 0.16, respectively). Patients with VOD had a significantly later platelet engraftment to 20 × 109 l−1 (median 33 vs 19 days, P=0.003) and 50 × 109 l−1 (median 45 vs 28 days, P=0.006) compared to patients without VOD. Acute GVHD was not significantly associated with VOD. However, the presence of stage 3–4 hepatic GVHD was associated with the development of severe VOD (OR=9.36, 95% CI 1.46–60.1, P=0.005).

Patients with past CMV or EBV infections as indicated by positive CMV and EBV antibodies before transplant were not significant risk factors for VOD. Chronic hepatitis B also did not appear to increase the risk of VOD, as none of the VOD patients was a carrier of hepatitis B although four patients in the entire cohort (2.9%) were hepatitis B carriers. The role of hepatitis C was uncertain, since only one patient in the entire cohort was a hepatitis C carrier, and had developed VOD post transplant.

In multivariate logistic regression analysis, only two factors were independently associated with increased risk of VOD. These were conditioned with BU and CY (adjusted OR=9.05, 95% CI 2.25–3.64, P=0.002), and decreasing age (adjusted OR for a younger age group=2.43, 95% CI 1.09–5.45, P=0.031). The risks of VOD for different age groups in patients with or without conditioning with BU and CY are shown in Table 4. The risk was the highest in patients aged below 2 years by conditioning with BU and CY (35.3%) and the lowest in patients aged above 8 years without conditioning with BU and CY (2.6%).

Table 4 Risks of hepatic VOD in different groups of patients

Mortality predictors of VOD

Seven of the 19 patients (36.8%) with VOD died at a median of 44 days post transplant (range, 30–421 days) (Table 2). Two of these seven patients died of leukemic relapse rather than VOD. The estimated 5-year survival was 62%. All surviving patients had normal liver function on follow-up at 0.5–9 years after transplant. Although the overall 5-year survival was similar to patients without VOD (63.2%), the 100-day mortality was significantly higher in patients with VOD (16.3 vs 9.6%, P=0.024); and VOD was the primary cause of all these early deaths. The median hospital stay was similar in patients with or without VOD (76.5 vs 85 days, P=0.55). However, patients with VOD were more likely to require admission to ICU (OR=3.82, 95% CI 1.37–10.7, P=0.013) and those who had been admitted to ICU were more likely to die (hazard ratio (HR)=13.6, 95% CI 1.62–114, P=0.016). Certain clinical features were also found to be significant mortality predictors of VOD (Table 3). These included pleural effusion (HR=13.6, 95% CI 1.62–114, P=0.016) and maximal weight gain above 9% (HR=6.81, 95% CI 1.30–35.5, P=0.023). VOD after matched unrelated or haplo-identical transplants also carried a higher risk of death compared to autologous or matched sibling transplants (HR=23.6, 95% CI 2.48–226, P=0.006); and this increased mortality risk was in excess of that for patients without VOD (HR=2.06, 95% CI 1.13–3.76, P=0.019). The risk of death was also greater in the presence of moderate-to-severe hepatic GVHD (HR=8.15, 95% CI 1.13–58.9, P=0.038) or cutaneous GVHD (HR=8.15, 95% CI 1.13–58.9, P=0.038). Again these HRs were greater than in patients without VOD (HR=5.27 for hepatic GVHD and 1.77 for cutaneous GVHD). The time course of bilirubin also predicted mortality (Figure 1). The chance of death was significantly higher if bilirubin at day +21 was above 200 μmol l−1 (HR=33.9, 95% CI 3.87–296, P=0.001), rise of bilirubin was above 15 μmol l−1 per day within the first week of onset of VOD (HR=19.8, 95% CI 2.33–168, P=0.006), or peak bilirubin was above 300 μmol l−1 (HR=13.6, 95% CI 1.62–114, P=0.016).

Figure 1

Serum bilirubin levels in patients with hepatic VOD. VOD=veno-occlusive disease.

The number of mortality predictors present correlated significantly with survival. The chance of death was substantially higher if four or more predictors were present (100 vs 7.7%, HR=33.9, 95% CI 3.87–296, P=0.001). The presence of one additional mortality predictor increased the risk of death by 71% (HR=1.71, 95% CI 1.23–2.36, P=0.001) in Cox regression model. Multivariate analyses showed that rise of bilirubin above 15 μmol l−1 per day (HR=17.9, 95% CI 1.70–188, P=0.016) and transplants other than autologous or matched sibling (HR=24.3, 95% CI 1.20–491, P=0.038) were independent significant mortality predictors.

Management of VOD

All cases of VOD were treated with fluid restriction and 18 patients (94.7%) were also given diuretics. In addition, three patients (15.8%) were treated with recombinant tissue plasminogen activator (t-PA), four patients (21.1%) were treated with defibrotide and two patients (10.5%) were treated with both t-PA and defibrotide. T-PA was given at a dose of 0.2 mg kg−1 per day until resolution of VOD or significant complications occurred. T-PA was started 1–29 days after the onset of VOD. T-PA was started before reversal of portal flow was detected in two patients, and within 1 day of detection of reversal of portal flow in the remaining three patients. One patient who received t-PA developed severe gastrointestinal bleeding necessitating withdrawal of therapy. Three of the five patients (60%) given t-PA eventually died and all three had severe VOD, while the remaining two surviving patients had moderate VOD only. Survival was not significantly different in patients who had or had not received t-PA. Six patients (31.6%) received defibrotide, at a median of 1.5 days after the onset of VOD (range, 0–16 days). Defibrotide was initiated at 10 mg kg−1 per day and stepped up gradually to a maximum of 40 mg kg−1 per day if there was no significant response and no adverse effect. One patient developed severe gastrointestinal bleeding a few days after starting defibrotide. The hemorrhage improved after stopping defibrotide but recurred on its resumption, which necessitated its withdrawal eventually. Four of the six patients (66.7%) who received defibrotide eventually died. Although the survival appeared poorer in patients who had received defibrotide, all four patients who died after receiving defibrotide had severe VOD. Survival did not appear to be related to the timing of initiation of defibrotide.


Hepatic VOD is an important cause of morbidity and mortality in HSCT recipients. Its incidence varied significantly among different centers, which is partly explained by the different definitions used for VOD in different studies. For example, in the series reported by Hasegawa, the incidence of VOD was 27.1% by the Seattle criteria and 14.3% by the Baltimore criteria.6 Our review found a cumulative incidence of 13.4% by the Baltimore criteria, which is within the range reported in the literature. Previous studies involved mainly Caucasians and our studies involved Chinese and Southeast Asian patients. A similar incidence suggests that ethnicity unlikely affects the incidence of VOD. The Baltimore criteria include only early-onset VOD that develops before day +21 post transplant, and might underestimate the real incidence. However, it might be more specific and pick up more severe VOD.12 Measurement of blood levels of plasminogen activator inhibitor-1 may aid the diagnosis of VOD, as maximum plasminogen activator inhibitor-1 plasma level below 120 ng ml−1 has been found to have a strong negative predictive value in the diagnosis of VOD.13

Risk factors

The risk factors for hepatic VOD in children identified in previous studies are shown in Table 5. Our findings concurred with some of these studies. In the current study, diseases with previous intensive chemotherapy such as neuroblastoma, leukemia or advanced malignancy were not found to carry a significantly higher risk of VOD. Since we adopted similar treatment protocols as other centers, the differences may be due to the small number of these patients in our cohort. In contrary, we found that thalassemia major was a strong-risk factor for VOD. This might be related to the pre-existing liver damage caused by iron overload. However, the pretransplant serum ferritin, Lucarelli risk class, grading of hemosiderosis or fibrosis were all not significantly associated with development of VOD in patients with thalassemia major. On the other hand, conditioning regimen in thalassemia major involved the combination of BU and CY, which was another confounding risk factor for VOD. This might be an important potentiating risk factor for VOD development in thalassemic patients, as the OR of developing VOD in thalassemic patients became insignificant but Bu–Cy-conditioning regimen remained significant in multivariate analysis.

Table 5 Summary of risk factors for hepatic VOD in previous pediatric studies and the current study

Busulfan is well known to be associated with post transplant VOD, both in children and in adults. Grochow found that a greater area under curve after BU administration was associated with increased risk of VOD, suggesting a dose–response relationship between BU and VOD.14 This was supported by Meresse's study which found that patients who received BU at a total dose above 16 mg kg−1 had a higher risk of VOD.15 However, we could not replicate his findings in the current study. One of the reasons is the age-related variation in BU metabolism especially in young children and infants. The mechanism for BU hepatotoxicity is not entirely clear, but thought to be related to glutathione depletion at high dose.16 Although i.v. BU seems to be associated with a lower risk of VOD in adults,17 data in children are still preliminary.18, 19 We did not find a significant difference between i.v. and oral BU in the risk of VOD in our pediatric cohort. A further evaluation of i.v. BU is needed. On the other hand, we found that the addition of CY to BU was an additive risk factor for VOD, similar to the study by Barker et al.4 It has been proposed that BU might influence the metabolism of CY, causing an increase in toxic metabolites which are associated with VOD.20, 21 CY itself was also reported to be associated with VOD.1 Nevertheless, using newer drugs with less hepatotoxicity such as treosulfan instead of BU might reduce the incidence of VOD, even when combined with CY.22

We found that the younger the patient, the higher was the risk of VOD. This might be related to poorer hepatic reserves or smaller caliber of hepatic venules which are more readily obstructed even with mild endothelial injury and subendothelial edema. In addition, exposure to BU might be highly variable in children when the same dose per body weight is used. Therefore, some authors have suggested measuring the blood concentration of BU and titrating the dose of BU according to the area under curve.14, 23, 24, 25, 26 Whether this strategy helps to reduce the incidence of VOD in young children requires further evaluation.

We found that platelet non-engraftment at 3 weeks post transplant was associated with VOD, similar to the study by Cesaro et al.3 Whether thrombocytopenia per se predisposes to endothelial dysfunction and VOD or it simply represents the consequence of VOD is not clear. However, delayed platelet engraftment implies a higher degree of myelosuppression induced by the conditioning regimen and it might serve as a marker of hepatic endothelial cell damage induced by BU or other cytotoxic drugs.

On the contrary to the study by Barker et al.,4 we did not find prior exposure to parenteral nutrition or positive CMV serology to be significant risk factors for VOD. These potential risk factors require further evaluation in separate studies with larger sample size or meta-analyses. As none of our patients had a history of pancreatitis, we could not evaluate whether this is an important risk factor for VOD as suggested by Barker et al.4 On the other hand, whether transplants with HLA-mismatched or -unrelated donor really predispose to the development of VOD was not conclusive. Cesaro et al.3 found that autologous transplant was associated with a higher risk of VOD, while others found that HLA-mismatched or -unrelated donor transplants carried a higher risk.4, 5, 6 In our study, all these factors were not significant. Further studies are required to resolve the issue.

Mortality predictors

The current study is in agreement with previous reports that development of VOD post transplant is associated with a higher mortality, especially before 100 days.3, 4, 5 The overall survival rate of 62% in pediatric patients with VOD was within the range of 53–100% reported in the literature.3, 4, 5, 6, 15, 27, 28 There have been few studies on the mortality predictors of VOD in pediatric HSCT recipients. In our study, we found that transplants from donors other than matched sibling or autologous transplants were associated with a higher mortality. Although these transplants are well known to carry a higher risk whether or not the patients develop VOD post transplant, the excess risk associated with these transplants is above the baseline risk for patients without VOD, indicating that patients who receive allogeneic or unrelated donor transplants and develop VOD have a particularly poor outcome.

Patients who had moderate-to-severe cutaneous or hepatic GVHD concurrent with VOD also had a particularly poor survival. It is not surprising that patients with hepatic GVHD and VOD had higher risk of death due to the double insults to the liver. However, it is not entirely clear why patients with severe cutaneous GVHD also had higher mortality. The immunosuppressive drugs used to treat cutaneous GVHD might account for the increased risk of VOD. However, it is also probable that patients with moderate-to-severe cutaneous GVHD also frequently had coexisting hepatic GVHD which increased the risk of VOD.

On the other hand, higher severity of hepatic VOD, as indicated by higher peak bilirubin, higher maximal weight gain, presence of pleural effusion and requirement for ICU admission, was associated with higher mortality. In addition, the time course of bilirubin appeared to have significant prognostic value, as patients with a rapid rise of bilirubin (>15 μmol l−1 per day) in the first week or bilirubin reaching 200 μmol l−1 by day +21 had very high mortality. These features might help clinicians to predict poor outcome and administer treatment early.

In addition, we found that the more the number of mortality predictors present in patients with VOD, the higher was the mortality rate. Patients with VOD and four or more predictors had 100% mortality while patients with VOD and only 1–3 predictors had 92.3% survival. Furthermore, we found that each additional mortality predictor increased the HR by about 70%. Although both Cesaro et al.'s study3 and our study found that patients with reversal of portal venous blood flow tended to have higher mortality, the difference was not statistically significant.

Implications for management

We did not practice routine VOD prophylaxis for our patients at the time, as no prophylactic regimen had demonstrated efficacy consistently in previous trials. Cesaro et al.3 found that children who received heparin prophylaxis had significantly lower incidence of VOD compared to patients who received no prophylaxis or prophylaxis with pentoxifylline or prostaglandin E1. However, the prophylactic regimen given in this study varied in different periods and the conclusion might be subjected to confounding variables during different transplant periods. Routine heparin prophylaxis was also used in the study by Reiss et al.5 after 1995, but this did not reduce VOD significantly. A recent systematic review of controlled trials found that prophylactic ursodeoxycholic acid reduced the incidence of VOD and transplant-related mortality in adults.29 Further evaluation of different prophylactic regimen in prospective clinical trials is essential to determine the risk–benefit ratio of VOD prophylaxis in children. Nevertheless, before such trials are available, we might consider giving prophylaxis in patients with high risk of VOD, such as young children with thalassemia major conditioned with BU–CY.

Apart from standard supportive treatment for VOD, we treated some patients with t-PA and some patients with defibrotide in recent years. In our limited experience, we did not find these therapies to be particularly effective; and adverse effect of significant gastro-intestinal hemorrhage was not infrequent. The lack of effectiveness might be because we used these treatments mainly in patients with moderate-to-severe VOD, implying the disease already progressed beyond the salvageable point. Previous reports on efficacy of t-PA or defibrotide for treatment of VOD were based on cohort studies only. Since no randomized-controlled trial is currently available, the effectiveness of specific treatments of VOD remained uncertain. If treatment is to be considered, it should probably be commenced early so that the pathologic progression can be halted before irreversible changes occur. However, there is an important dilemma that mild VOD might resolve with simple supportive care and therefore does not justify the significant hemorrhagic risk of specific therapies. Consideration of risk factors and mortality predictors might provide useful guide to appropriate therapeutic decision. We found that patients who received allogeneic transplant and patients with moderate-to-severe hepatic GVHD were more likely to develop severe VOD; and patients who had unrelated or mismatched donor transplants, moderate-to-severe hepatic or cutaneous GVHD, or rapid rise of bilirubin were more likely to die from VOD. Therefore, we should consider early treatment of VOD in patients with allogeneic, unrelated or mismatched transplant who have concurrent hepatic or cutaneous GVHD or rapid surge of bilirubin.

Our study has several limitations. First, this is a retrospective review and not a prospective study and is subjected to possible observation and selection biases. Nevertheless, we included all patients treated in our center to minimize selection bias and most relevant data were prospectively captured in the database of our departmental computer server. Second, the small sample size limits the statistical power of detecting significant risk factors and mortality predictors, and the ability to control for confounding variables in multivariate analyses. Further studies are needed to confirm our findings.

In conclusion, hepatic VOD is not an uncommon complication of pediatric HSCT, especially in patients with multiple risk factors. It can be serious and associated with high mortality, particularly in patients with four or more poor prognostic factors. Meticulous monitoring in high-risk patients is important so that VOD can be identified early and treatment instituted promptly before VOD progresses fatally. Randomized-controlled trials by multicenter collaboration on effective prophylaxis and treatment of VOD are required.


  1. 1

    Brugieres L, Hartmann O, Benhamou E, Zafrani ES, Caillaud JM, Patte C 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: 53–58.

    CAS  PubMed  Google Scholar 

  2. 2

    Lapierre V, Mahe C, Auperin A, Stambouli F, Oubouzar N, Tramalloni D et al. Platelet transfusion containing ABO-incompatible plasma and hepatic veno-occlusive disease after hematopoietic transplantation in young children. Transplantation 2005; 80: 314–319.

    Article  PubMed  PubMed Central  Google Scholar 

  3. 3

    Cesaro S, Pillon M, Talenti E, Toffolutti T, Calore E, Tridello G et al. A prospective survey on incidence, risk factors and therapy of hepatic veno-occlusive disease in children after hematopoietic stem cell transplantation. Haematologica 2005; 90: 1396–1404.

    Google Scholar 

  4. 4

    Barker CC, Butzner JD, Anderson RA, Brant R, Sauve RS . Incidence, survival and risk factors for the development of veno-occlusive disease in pediatric hematopoietic stem cell transplant recipients. Bone Marrow Transplant 2003; 32: 79–87.

    CAS  Article  PubMed  Google Scholar 

  5. 5

    Reiss U, Cowan M, McMillan A, Horn B . Hepatic venoocclusive disease in blood and bone marrow transplantation in children and young adults: incidence, risk factors, and outcome in a cohort of 241 patients. J Pediatr Hematol Oncol 2002; 24: 746–750.

    Article  Google Scholar 

  6. 6

    Hasegawa S, Horibe K, Kawabe T, Kato K, Kojima S, Matsuyama T et al. Veno-occlusive disease of the liver after allogeneic bone marrow transplantation in children with hematologic malignancies: incidence, onset time and risk factors. Bone Marrow Transplant 1998; 22: 1191–1197.

    CAS  Article  PubMed  Google Scholar 

  7. 7

    Bulley SR, Strahm B, Doyle J, Dupuis LL . Defibrotide for the treatment of hepatic veno-occlusive disease in children. Pediatr Blood Cancer 2006; 48: 700–704.

    Article  Google Scholar 

  8. 8

    Corbacioglu S, Greil J, Peters C, Wulffraat N, Laws HJ, Dilloo D et al. Defibrotide in the treatment of children with veno-occlusive disease (VOD): a retrospective multicentre study demonstrates therapeutic efficacy upon early intervention. Bone Marrow Transplant 2004; 33: 189–195.

    CAS  Article  Google Scholar 

  9. 9

    Jones RJ, Lee KS, Beschorner WE, Vogel VG, Grochow LB, Braine HG et al. Venoocclusive disease of the liver following bone marrow transplantation. Transplantation 1987; 44: 778–783.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  10. 10

    McDonald GB, Hinds MS, Fisher LD, Schoch HG, Wolford JL, Banaji M 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.

    CAS  Article  Google Scholar 

  11. 11

    Glucksberg H, Storb R, Fefer A, Buckner CD, Neiman PE, Clift RA et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation 1974; 18: 295–304.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. 12

    Blostein MD, Paltiel OB, Thibault A, Rybka WB . A comparison of clinical criteria for the diagnosis of veno-occlusive disease of the liver after bone marrow transplantation. Bone Marrow Transplant 1992; 10: 439–443.

    CAS  PubMed  Google Scholar 

  13. 13

    Pihusch M, Wegner H, Goehring P, Salat C, Pihusch V, Hiller E et al. Diagnosis of hepatic veno-occlusive disease by plasminogen activator inhibitor-1 plasma antigen levels: a prospective analysis in 350 allogeneic hematopoietic stem cell recipients. Transplantation 2005; 80: 1376–1382.

    CAS  Article  PubMed  Google Scholar 

  14. 14

    Grochow LB, Jones RJ, Brundrett RB, Braine HG, Chen TL, Saral R et al. Pharmacokinetics of busulfan: correlation with veno-occlusive disease in patients undergoing bone marrow transplantation. Cancer Chemother Pharmacol 1989; 25: 55–61.

    CAS  Article  Google Scholar 

  15. 15

    Meresse V, Hartmann O, Vassal G, Benhamou E, Valteau-Couanet D, Brugieres L 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: 135–141.

    CAS  PubMed  Google Scholar 

  16. 16

    Gibbs JP, Yang JS, Slattery JT . Comparison of human liver and small intestinal glutathione S-transferase-catalyzed busulfan conjugation in vitro. Drug Metab Dispos 1998; 26: 52–55.

    CAS  PubMed  Google Scholar 

  17. 17

    Lee JH, Choi SJ, Lee JH, Kim SE, Park CJ, Chi HS et al. Decreased incidence of hepatic veno-occlusive disease and fewer hemostatic derangements associated with intravenous busulfan vs oral busulfan in adults conditioned with busulfan+cyclophosphamide for allogeneic bone marrow transplantation. Ann Hematol 2005; 84: 321–330.

    CAS  Article  PubMed  Google Scholar 

  18. 18

    Zwaveling J, den Hartigh J, Lankester AC, Guchelaar HJ, Egeler RM, Bredius RG . Once-daily intravenous busulfan in children prior to stem cell transplantation: study of pharmacokinetics and early clinical outcomes. Anticancer Drugs 2006; 17: 1099–1105.

    CAS  Article  PubMed  Google Scholar 

  19. 19

    Lee MY, Chiou TJ, Bai LY, Hsiao LT, Hung GY, Chang CY et al. Intravenous busulfan as preparative regimen in pediatric patients receiving hematopoietic stem cell transplantation: the preliminary experience in Taiwan. J Chin Med Assoc 2004; 67: 117–122.

    PubMed  Google Scholar 

  20. 20

    Slattery JT, Kalhorn TF, McDonald GB, Lambert K, Buckner CD, Bensinger WI et al. Conditioning regimen-dependent disposition of cyclophosphamide and hydroxycyclophosphamide in human marrow transplantation patients. J Clin Oncol 1996; 14: 1484–1494.

    CAS  Article  Google Scholar 

  21. 21

    Brodsky R, Topolsky D, Crilley P, Bulova S, Brodsky I . Frequency of veno-occlusive disease of the liver in bone marrow transplantation with a modified busulfan/cyclophosphamide preparative regimen. Am J Clin Oncol 1990; 13: 221–225.

    CAS  Article  PubMed  Google Scholar 

  22. 22

    Beelen DW, Trenschel R, Casper J, Freund M, Hilger RA, Scheulen ME et al. Dose-escalated treosulphan in combination with cyclophosphamide as a new preparative regimen for allogeneic haematopoietic stem cell transplantation in patients with an increased risk for regimen-related complications. Bone Marrow Transplant 2005; 35: 233–241.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  23. 23

    Dix SP, Wingard JR, Mullins RE, Jerkunica I, Davidson TG, Gilmore CE et al. Association of busulfan area under the curve with veno-occlusive disease following BMT. Bone Marrow Transplant 1996; 17: 225–230.

    CAS  Google Scholar 

  24. 24

    Yeager AM, Wagner Jr JE, Graham ML, Jones RJ, Santos GW, Grochow LB . Optimization of busulfan dosage in children undergoing bone marrow transplantation: a pharmacokinetic study of dose escalation. Blood 1992; 80: 2425–2428.

    CAS  PubMed  Google Scholar 

  25. 25

    Kletzel M, Jacobsohn D, Duerst R . Pharmacokinetics of a test dose of intravenous busulfan guide dose modifications to achieve an optimal area under the curve of a single daily dose of intravenous busulfan in children undergoing a reduced-intensity conditioning regimen with hematopoietic stem cell transplantation. Biol Blood Marrow Transplant 2006; 12: 472–479.

    CAS  Article  PubMed  Google Scholar 

  26. 26

    Tran H, Petropoulos D, Worth L, Mullen CA, Madden T, Andersson B et al. Pharmacokinetics and individualized dose adjustment of intravenous busulfan in children with advanced hematologic malignancies undergoing allogeneic stem cell transplantation. Biol Blood Marrow Transplant 2004; 10: 805–812.

    CAS  Article  PubMed  Google Scholar 

  27. 27

    Ozkaynak MF, Weinberg K, Kohn D, Sender L, Parkman R, Lenarsky C . Hepatic veno-occlusive disease post-bone marrow transplantation in children conditioned with busulfan and cyclophosphamide: incidence, risk factors, and clinical outcome. Bone Marrow Transplant 1991; 7: 467–474.

    CAS  PubMed  Google Scholar 

  28. 28

    Srivastava A, Poonkuzhali B, Shaji RV, George B, Mathews V, Chandy M et al. Glutathione S-transferase M1 polymorphism: a risk factor for hepatic venoocclusive disease in bone marrow transplantation. Blood 2004; 104: 1574–1577.

    CAS  Article  PubMed  Google Scholar 

  29. 29

    Tay J, Tinmouth A, Fergusson D, Huebsch L, Allan DS . Systematic review of controlled clinical trials on the use of ursodeoxycholic acid for the prevention of hepatic veno-occlusive disease in hematopoietic stem cell transplantation. Biol Blood Marrow Transplant 2007; 13: 206–217.

    CAS  Article  PubMed  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to D K L Cheuk.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cheuk, D., Wang, P., Lee, T. et al. Risk factors and mortality predictors of hepatic veno-occlusive disease after pediatric hematopoietic stem cell transplantation. Bone Marrow Transplant 40, 935–944 (2007).

Download citation


  • children
  • hematopoietic stem cell transplantation
  • veno-occlusive disease
  • risk factor
  • prognosis
  • mortality

Further reading