Modified diagnostic criteria, grading classification and newly elucidated pathophysiology of hepatic SOS/VOD after haematopoietic cell transplantation

Summary

Sinusoidal obstruction syndrome (SOS), previously known as hepatic veno-occlusive disease (VOD) remains a multi-organ system complication following haematopoietic cell transplantation (HCT). When SOS/VOD is accompanied by multi-organ dysfunction, overall mortality rates remain greater than 80%. However, the definitions related to the diagnosis and grading of SOS/VOD after HCT are almost 25 years old and require new and contemporary modifications. Importantly, the pathophysiology of SOS/VOD including the contribution of dysregulated inflammatory and coagulation cascades and the critical importance of liver and vascular derived endothelial dysfunction have been elucidated. Here we summarise new information on pathogenesis of SOS/VOD, identify modifiable and unmodifiable risk factors for disease development, propose novel, contemporary, and panel opinion -based diagnostic criteria and an innovative organ-based method of SOS/VOD grading classification, and review current approaches for prophylaxis and treatment of SOS/VOD. This review will hopefully illuminate pathways responsible for drug induced liver injury and manifestations of disease, sharpen awareness of risk for disease development, and enhance the timely and correct diagnosis of SOS/VOD post HCT.

Introduction

Hepatic veno-occlusive disease (VOD), now termed sinusoidal obstructive syndrome (SOS), has been recognised in the veterinary literature for some time (Willmot and Robertson 1920). Twelve cases of hepatic SOS/VOD, termed Chiari’s syndrome due to senecio poisoning, were initially reported in humans by Selzer et al (Selzer and Parker 1951). Hepatic SOS/VOD was later described in association with conventional chemotherapeutic agents including azathioprine, 6-thioguanine, cytarabine, high-dose liver irradiation, and liver radio-embolisation (Bras et al, 1954). Subsequently, Lazarus et al reported a high incidence of hepatic SOS/VOD in a prospective series of 29 refractory malignancy patients given high-dose mitomycin C (60-90 mg/m² intraveneously) and autologous bone marrow transplantation (Lazarus et al, 1982).

The incidence of SOS/VOD post HCT, varies depending on several factors including: the type of transplant (allogeneic> autologous), patient age (paediatric > adult), conditioning regimen intensity (myeloablative (MAC) vs. reduced intensity (RIC)), the presence of pre-existing risk factors, and importantly, the clinical criteria used to make the clinical diagnosis. The incidence of SOS/VOD reported in the contemporary literature ranges from 10 to 20% after allogeneic HCT using MA conditioning (and has high as 60% in a subset of paediatric patients), to 0 to 10% following RIC allo-HCT, 0-1% following non-myeloablative conditioning (NMA) AlloHCT, and as low as 5% following autologous HCT (Carreras et al, 2011; Coppell et al, 2010; Corbacioglu et al, 2018; Corbacioglu et al, 2012; Mohty et al, 2015; Roeker et al, 2019; Yakushijin et al, 2016). Signs and symptoms of SOS/VOD typically develop within 3 weeks of stem cell transfusion, although in approximately 20% of cases manifestations of disease occur later (Carreras et al, 2007; Lee et al, 1999). With improved BMT technologies, including healthier patients coming to HCT, pharmacokinetic monitoring and increased use of NMA/RIC conditioning, the number of reported cases of SOS/VOD has waned in recent years (indeed many early career BMT faculty may have only seen a case or two). However, when associated with multi-organ dysfunction, mortality rates associated with SOS/VOD remain unacceptable and has high as 85% (Coppell, et al 2010; Richardson et al, 2016). In this context, having the tools necessary to make a prompt and accurate diagnosis and to identify advanced stages of disease progression (specifically renal and pulmonary toxicity) and enhanced understanding of the pathophysiology is critical to optimise patient care (Kernan et al, 2018).

The traditionally used definitions and grading systems for VOD following HCT are decades old, lack inclusion of contemporary signs and symptoms, and diagnostic procedures and can result in delays in diagnosis and less precise reporting of toxicity grading. More recently the pathophysiology of SOS/VOD, including the contribution of dysregulated inflammatory and coagulation cascades and the critical importance of liver and vascular derived endothelial dysfunction has been elucidated facilitating more rational drug development for this lethal syndrome post HCT. In the following review, we summarise the recent developments in the pathophysiology of SOS/VOD, propose novel, contemporary and consensus based revised diagnostic criteria and innovative and more rationale grading classification to facilitate a more timely and correct diagnosis.

Pathogenesis of SOS

Inflammation and endothelial cell (EC) activation

SOS/VOD is one of the most severe multi-organ complications following HCT. It is believed that SOS/VOD primarily involves injury to micro-vascular, hepatic ECs ultimately culminating in obstruction of sinusoids in zone 3 of the hepatic acinus (Bearman 1995; Carreras 2012; Richardson et al, 2013a; Richardson et al, 2013b). Resting vascular derived ECs have several critical functions including maintenance of blood fluidity, regulation of flow and vessel permeability, and quiescence of circulating leukocytes (Ley et al, 2007; Vestweber 2015). When vascular derived ECs are injured, they become activated and more adhesive. Expression and upregulation of adhesion molecules along with the release of soluble proteins (cytokines and chemokines) initiate inflammatory pathways that propagate EC damage (Cooke et al, 2008). A critical pathway controlling vascular derived EC integrity involves the receptor tyrosine kinaseTie-2 found that is primarily expressed on the surface of ECs and its ligands Angiopoietin (Ang)-1 and Ang-2. The two proteins have been identified as an agonist/antagonist pair that regulates early vascular development and endothelial barrier integrity during physiologic homeostasis and disease (Fiedler et al, 2006). In brief, Ang-1 activates Tie-2, resulting in receptor phosphorylation and subsequent signal transduction that promotes EC survival and vessel stability. Ang-1 may also have anti-inflammatory properties by down-regulating surface-adhesion molecules such as VCAM-1 and E-selectin. By contrast, Ang-2 is a known antagonist of Tie-2 that competitively binds the receptor and interferes with agonistic Ang-1:Tie-2 receptor:ligand functions and thus enhances vessel permeability. Moreover, Ang-2 release can be driven by inflammatory stimuli such as tumor necrosis factor-alpha to regulate the expression of adhesion molecules on the EC surface (Fiedler, et al 2006). Accordingly, Ang-2 is one of several proteins that are elevated in the serum of patients affected by SOS/VOD (Akil et al, 2015). The passage of leukocytes across damaged vascular endothelium and into inflamed tissues is also tightly controlled. This process is regulated, in part, by the expression of adhesion molecules on vascular derived EC surfaces and ultimately by the integrity of the endothelial barrier itself (Ley, et al 2007; Vestweber 2015).

Sinusoidal narrowing and occlusion during SOS

The initial damage to hepatocytes and vascular derived sinusoidal endothelium during the development of SOS/VOD results in a cascade of events involving vascular derived EC activation, sinusoidal narrowing, and micro-vessel occlusion (Fig 1A) (Carreras 2012; Coppell et al, 2003; Richardson, et al 2013a; Richardson, et al 2013b). Activated ECs release heparinase, which serves to break down extracellular matrix proteins. Subsequent disruption of cytoskeletal architecture contributes to the loss of inter-cellular tight junctions; when the balance between contractile forces within ECs is disrupted, ECs become rounded creating intercellular gaps that increase vascular permeability and allow fluid and other blood elements to leak into the extravascular space of Disse and surrounding tissues, resulting in narrowing of the sinusoids and slowing of flow through the vascular lumen (Fig 1B).

Fig 1.

Pathophysiology of SOS/VOD following HCT. Inflammation engendered by toxic metabolites from BMT conditioning and cytokine release results in activation of and damage to the endothelial cell lining of the hepatic sinusoids (A). Activated ECs upregulate surface adhesion molecules and release of heparinase, which contributes to the breakdown of EC scaffolding. Damage ECs round up resulting in gap formation, leakage and sinusoidal narrowing (B). Inflammation is accompanied by alterations of coagulation pathways. Increased expression of vWF along with the release of tissue factor and PAI-1 leads to platelet aggregation and a pro-thrombotic, hypo-fibrinolytic state. Ultimately, dissection of and embolisation of ECs along with Fibrin depo and clot formation culminate in sinusoidal blockage (1C).

Inflammation contributing to and engendered by EC damage also triggers the release of multiple factors that regulate coagulation and fibrinolysis. Two such proteins are von Willebrand’s factor, which stimulates platelet aggregation and tissue factor, which promotes activation of other clotting proteins (Cooke, et al 2008; Coppell, et al 2003). Furthermore, plasminogen activator inhibitor-1 (PAI-1), an inhibitor of fibrinolysis, is released (Richardson et al, 1996 (abstract); Salat et al, 1997). All of these events contribute to the development of a prothrombotic and hypofibrinolytic state within the hepatic sinusoids. Fibrin deposition and platelet aggregation result in microvascular clot formation, which in conjunction with EC embolisation and sinusoidal narrowing ultimately results in further blockage and sinusoidal obstruction (Fig 1C). Disruption in anterograde sinusoidal flow may be associated with portal hypertension, a reduction of hepatic venous outflow and further hepatocyte cell death. This cascade of events culminates in the development of symptoms that characterise SOS/VOD, and in fact, occur well before clinical (painful hepatomegaly, ascites, weight gain) and laboratory (hyperbilirubinemia and liver enzyme elevation) findings of SOS/VOD are present. In the worst case scenario, SOS/VOD may rapidly progress to reversal of portal venous outflow and hepatorenal syndrome culminating in SOS/VOD with renal or pulmonary dysfunction and multi-organ failure (Fig 2) (Carreras, et al 2011; Coppell, et al 2010).

Fig 2:

The complex and multifactorial pathophysiology of SOS/VOD following HCT. The initial damage to hepatocytes and sinusoidal endothelium during the development of SOS/VOD results in an inflammatory cascade involving EC damage, sinusoidal narrowing, and micro-vessel occlusion / sinusoidal blockage. These events culminate in the development of symptoms that characterise SOS/VOD, and in fact, occur well before clinical and laboratory findings of SOS/VOD are present. SOS/VOD may rapidly progress to reversal of portal venous outflow and hepatorenal syndrome culminating in SOS/VOD with renal or pulmonary dysfunction and multi-organ failure and in the worst case scenario, death of the patient.

MOD, multi-organ dysfunction; MODS, multiple organ dysfunction syndrome.

Liver sinusoidal endothelial cells (LSECs), hepatic stellate cells (HSCs) and hepatobiliary transport systems

More recently, alterations in LSECs, HSCs and hepatobiliary transport systems have thought to also contribute to the pathophysiology of SOS/VOD. LSECs have significantly higher endocytotic activity compared to other vascular derived endothelial cells and can rapidly endocytose extracellular and cellular components including toxic metabolites via clearance of macromolecules and nanoparticles by clathrin-mediated endocytosis (DeLeve and Maretti-Mira 2017; Sorensen et al, 2015). Furthermore, LSECs maintain HSCs quiescence, thereby promoting inhibition of intrahepatic vasoconstriction and fibrosis. HSCs reside in the Space of Disse and are considered resident mesenchymal cells that possess features of fibroblasts (Higashi et al, 2017). Drug induced damage to the hepatic tissue microenvironment induces HSC activation resulting in LSECs becoming capillarised with increased lost fenestrations that promotes vasoconstriction and angiogenesis (Poisson et al, 2017). Endothelin-1, a significant inducer of vasoconstriction, is secreted by HSCs and also potentially lead to portal hypertension (Figure S1) (Li et al, 2012). Lastly, damage to hepatobiliary export pump proteins, such as ABCC2 (multidrug resistance associated protein 2) secondary to drug induced liver injury (DILI) maybe associated with cholestasis associated with HCT related SOS/VOD (Zollner et al, 2014). Targeted therapeutic approaches to these additional pathophysiological processes and pathways should be pursued in the future as alternative therapy for SOS/VOD (DeLeve and Maretti-Mira 2017; Higashi, et al 2017; Li, et al 2012; Poisson, et al 2017; Sorensen, et al 2015; Stapelbroek et al, 2010; van der Schoor et al, 2015; Zollner, et al 2014).

Risk factors associated with the development of SOS/VOD following HCT

Risk factors for the development of hepatic SOS/VOD may be viewed as “unmodifiable” versus “modifiable” and include patient/disease-, hepatic-, and transplant-related factors (Bearman 1995; Cesaro et al, 2005; Cheuk et al, 2007; Corbacioglu et al, 2019; McDonald et al, 1984; Mohty, et al 2015). Reports examining such features often utilise odds ratio data to characterise the degree of risk each factor associated with this multi-organ syndrome (Table I). Most importantly, prior and/or current liver dysfunction, prior intense chemo and/or radiotherapy and abnormal busulfan pharmacokinetics have been major risk factors in the past for increasing the risk of VOD post HCT (Corbacioglu, et al 2019; Dalle and Giralt 2016; Faraci et al, 2019; Hwang et al, 2016; Yakushijin, et al 2016). Importantly, administering high dose busulfan prior to high dose cyclophosphamide prior to HCT significantly increase the levels of cyclophosphamide metabolites including of hydroxycyclophosphamide and carboxyethylphosphoramide, which is significantly associated with an increase the risk of SOS/VOD. Whereas, reversing the order, and administering high dose cyclophosphamide prior to the high dose busulfan significantly lowers the risk of SOS/VOD post HCT (McCune et al, 2007; Rezvani et al, 2013). Busulfan pharmacokinetics (PK) monitoring has become standard practice in patients undergoing myeloablative conditioning utilizing intravenous busulfan prior to HCT. Conceivably PK monitoring by optimizing busulfan dosing, may reduce the incidence of VOD/SOS. However, Strouse et al reported recently that busulfan PK monitoring was independently associated with a higher incidence of VOD/SOS (Strouse et al, 2018). While the reasons for this surprising finding remains to be determined, one hypothesis is that the implementation of busulfan PK may ultimately result in higher exposure to the durg as reported by Weil and colleagues wherein nearly all patients ultimately had a dose adjustment (up) after levels were obtained (Weil et al, 2017). Importantly, Strouse and colleagues, continue to recommend busulfan PK monitoring.

Table I.

Risk factors associated with increased risk (3-10) and (>10 fold) of SOS/VOD following HCT.

	3-10 Times Greater Risk	>10 Times Greater Risk
Pre-transplantation Factors Unmodifiable	Increased recipient age Increased serum transaminase Hyper-ferritinemia (>1,000 ng/mL) Pre-existing liver disease Thalassemia; CML or immunodeficiency diagnosis CMV seropositivity Use acyclovir pre-transplantation Use parenteral alimentation Genetic factors (GSTM1 polymorphism) Sepsis syndrome	Serum bilirubin >1.5 mg/dL (>26 μmol/L) before transplantation Prior Mylotarg® (gemtuzumab ozogamicin) Prior Besponsa™ (inotuzumab ozogamicin) Prior norethisterone
Transplantation Related Factors Modifiable	Conditioning: High-dose (myeloablative) regimens Busulfan-containing (versus non-busulfan) prior to cyclophosphamide TBI + cyclophosphamide Fludarabine-containing Transplantation: Allogeneic (versus autologous) transplantation GVHD prophylaxis: Sirolimus + methotrexate + tacrolimus Methotrexate + cyclosporine Cyclosporine-containing

Dalle et al 2016(Dalle and Giralt 2016)

Key recent unmodifiable risk factors include the use of the anti-leukemic immunoconjugates gemtuzumab ozogamicin (anti-CD33; Mylotarg®) and inotuzumab ozogamicin (anti-CD22; Besponsa™), during induction and salvage treatment, which have been shown to significantly increase the risk of SOS/VOD following HCT (Magwood-Golston et al, 2016; Wadleigh et al, 2003). Kantarjian et al (Kantarjian et al, 2017) recently reported the hepatic adverse event profile of inotuzumab ozogamicin given to adult relapsed acute lymphoblastic leukemia patients in the course of the open-label, randomised, phase 3 INO-VATE study. Hepatic SOS/VOD developed in 22 (13%) of inotuzumab-treated patients while none in the standard of care cohort. HCT performed after inotuzumab exposure was associated with hepatic SOS/VOD in 17/77 (22%) of patients compared to only 3% (one of 32) if transplants were undertaken after standard therapy (McDonald et al, 2019). Recent studies suggest that antibody drug conjugates such as inotuzumab-ozogamicin are taken up by LSECs thereby inducing DILI and that prior DILI is the critical risk factor associated with an increased risk of SOS/VOD following subsequent HCT (McDonald et al, 1993).

Several features unique to children are also associated with an increased incidence of hepatic SOS/VOD after HCT. These factors include: age (very young, i.e. < 2 years); low body weight; and specific underlying disorders (haemophagocytic lympho-histiocytosis; osteopetrosis; neuroblastoma; Wilms tumor; rhabdomyosarcoma; juvenile myelomonocytic chronic leukemia; and haemoglobinopathies ([sickle cell anemia; thalassemia]) (Cecen et al, 2007; Cesaro et al, 2011; McPherson et al, 2011; Mohty, et al 2015). Younger age as a risk factor may in part be explained by the differential in HCT drug pharmacokinetics in younger age recipients. Finally, antecedent inflammation and damage to the liver by prior cytotoxic therapy (chemotherapy and or irradiation), infection (viral, fungal) and iron deposition (transfusional) also significantly increase the risk of developing SOS/VOD (Table I).

Modifiable factors

Some risk factors for SOS/VOD can be considered modifiable”. A large body of data suggest that elevated serum ferritin pre-HCT, as an indicator of iron overload, is associated with increased non-relapse mortality and decreased overall survival (Armand et al, 2014; Artz et al, 2016). In this liver iron deposition secondary to repeated infusions of red blood cells (pRBCs) can result in hepatocellular inflammation and fibrosis and likely contributes to an increased risk of SOS/VOD in many HCT candidates. Hence, efforts to aggressively chelate iron stores pre-HCT may have significant merit particularly when the urgency to proceed to BMT is less as is often the case in patients with transfusion-dependent haemoglobinopathies (thalassemia and sickle cell anemia) and MDS. The impact of such approaches on the incidence and severity of VOD should be studies prospectively. Additionally, the approach to HCT can be modified in an effort to avoid hepatic SOS/VOD. For example, the use of RIC regimens is associated with a lower incidence of hepatic SOS/VOD, although the risk of disease development still remains (George et al, 2007; Kim et al, 2007; Mohty, et al 2015; Schetelig et al, 2002; Shimoni et al, 2007; Tsirigotis et al, 2014; Van Besien et al, 2003). With increasing use of RIC/NMA pre-HCT conditioning, the risk of VOD continues to be diminished and is soon evolving into a risk-benefit discussion. Transplant-related factors associated with an elevated but less than 3 times risk for subsequent hepatic SOS/VOD after HCT are shown in Table I.

New revised diagnostic criteria of SOS/VOD

The development of diagnostic criteria is critical to allow clinicians to identify signs and symptoms of a disease entity, make the correct diagnosis in a prompt and confident manner, and ultimately initiate therapy when appropriate (Chao 2014). Two long-standing criteria have been established for the diagnosis of SOS/VOD. The modified Seattle criteria (McDonald, et al 1993) define hepatic SOS/VOD by the development of two or more of the following events within 20 days of HCT and without an obvious, alternative medical explanation: serum total bilirubin concentration greater than 2 mg/dL (>34.2 micromoles/L), hepatomegaly or right upper quadrant pain and sudden weight gain due to fluid accumulation (>2 percent of baseline body weight). The Baltimore criteria (Jones et al, 1987) initially proposed by Jones et al, define hepatic SOS/VOD by a serum bilirubin >2 mg/dL within 21 days of HCT plus at least two of the following: hepatomegaly, ascites or weight gain >5 percent from pre-HCT weight, without an alternative medical explanation. Established over 25 years ago, these criteria provided the platforms from which data related to the risks, incidence, severity and outcomes of SOS/VOD were generated. Over time, collective knowledge of patients who had SOS/VOD but whose signs and symptoms of disease did not fit nicely within either criteria have become a prevailing issue. Specifically, major limitations of these established diagnostic criteria include: 1) the time constraint of 21 days post HCT, 2) the development of anicteric SOS/VOD (addressed in part by the modified Seattle criteria), 3) recent clinical descriptions that are different from those described 25 years ago and 4) the development of newer imaging capabilities which may be more sensitive and specific indicators of SOS/VOD.

The European Society for Blood and Marrow Transplantation (EBMT) has also proposed a revised differential diagnostic criteria for both children and adults (Corbacioglu, et al 2018; Mohty et al, 2016). We have elected to streamline the current and past diagnostic criteria in a contemporary and versatile definition incorporating more signs and symptoms and advances in the diagnostic procedures including liver ultrasonsography and Doppler studies with the primary goals of increasing awareness of this haematological disease and thereby allowing for earlier detection and more rapid institution of therapy to improved survival (Herbetko et al, 1992; Nishida et al, 2018; Sharafuddin et al, 1997; Teefey et al, 1995). We propose the following revised SOS/VOD diagnostic criteria that includes two of the following criteria following HCT probably or directly secondary to VOD/SOS and not other etiologies: elevated bilirubin (≥2mg/dL) (≥34.2 micromoles/L or upper institutional limits), unexpected weight gain , excessive (≥5% compared to baseline weight pre-HCT), excessive platelet transfusions consistent with refractory thrombocytopenia post HCT size, hepatomegaly for age or increased liver size compared to pre-HCT, right upper quadrant pain (which may be difficult to assess in infants and small children), ascites confirmed by physical exam and/or imaging studies and/or reversal of portal venous flow (hepatofugal flow) by doppler ultrasound of the liver (which may be a late finding). In addition, any one of the following criteria after HCT would also be defined as SOS/VOD including hepatic biopsy consistent with SOS/VOD and/or unexplained elevated portal venous wedge pressure by direct measurements, which may be difficult to standardize from institution to institution (Table II). While we do not recommend either a liver biopsy and/or direct portal wedge pressure measurements, should the clinician have reason to perform such tests, and they are diagnostic, they should be utilized as diagnostic criteria. Further revising will be required in the future as new diagnostic procedures become identified such as recent hepatic ultrasound shear ultra-wave elastography (Reddivalla et al, 2018) .

Table II.

Revised diagnostic criteria for SOS/VOD following HCT.*

Any two of the following criteria following HCT
• Elevated bilirubin (≥2mg/dL) (≥34.2 micromoles/L) or greater than upper institutional limits **
• Unexpected weight gain (≥5% compared to baseline weight pre-HCT)
• Excessive platelet transfusions consistent with refractory thrombocytopenia post HCT
• Hepatomegaly for age or increase size over pre HCT
• Right upper quadrant pain
• Ascites confirmed by physical exam and/or imaging studies
• Reversal of portal venous flow (hepatofugal flow) by doppler ultrasound
OR
Any one of the following criteria following HCT
• Hepatic biopsy consistent with SOS/VOD
OR
• Unexplained elevated portal venous wedge pressure
• *Probably or definitely secondary to VOD/SOS and not other etiologies.
• **Patients with an already elevated bilirubin prior to HCT conditioning, this criteria should not be utilized in the diagnostic criteria.

It is important to emphasise that the proposed diagnostic criteria underscore our belief that the biology of SOS/VOD is not inherently different in paediatric patients compared to adults and as such separate criteria for the two populations of patients are not necessary. We do however acknowledge and consider certain manifestations of disease (including anicteric VOD and refractory thrombocytopenia) that are considered somewhat paediatric-specific and the use of contemporary radiographic tests including ultrasonography that can have diagnostic utility / implications. We must also underscore that the listed criteria (like elevated bilirubin and reversal of flow on ultrasound) are not required but if present can certainly contribute to making the correct/accurate diagnosis of VOD.

The role of ultrasound

The role of ultrasound in making the diagnosis and following disease progression has been investigated in several studies (Herbetko, et al 1992; Nishida, et al 2018; Reddivalla, et al 2018; Sharafuddin, et al 1997; Teefey, et al 1995). Most abnormalities noted on ultrasonography often serve to confirm clinical findings SOS/VOD, particularly those which may be challenging to assess in overweight or young HCT recipients, and may assist in ruling in or ruling out other entities included in the differential diagnoses. By contrast, a decrease or reversal of the portal venous flow, generally believed to be a late finding, is considered more specific for SOS/VOD. However, some reversal of flow may be a late finding, clinicians should not wait to make a diagnosis and/or initiate therapy if there is no presence of reversal of portal flow. The accuracy of ultrasound findings, including isolated alterations changes in portal venous velocity in diagnosing SOS/VOD has been evaluated in several studies (Herbetko, et al 1992; Lassau et al, 2002; Lassau et al, 1997; Mahgerefteh et al, 2011; McCarville et al, 2001; Nishida, et al 2018; Reddivalla, et al 2018; Sharafuddin, et al 1997; Teefey, et al 1995; Trenker et al, 2018). It remains to be determined whether newer US technologies including shear wave elastography (Reddivalla, et al 2018) and contrast enhancement (Trenker, et al 2018) may increase sensitivity, specificity and / or predictive value of diagnosing SOS/VOD.

SOS/VOD: Old and revised new grading classification

The original SOS/VOD grading classification was developed by McDonald et al in 1993 (McDonald, et al 1993). Mild disease was defined as patients with no apparent adverse effects from their VOD- associated liver disease, did not require diuretic or analgesic medications, and had complete reversal of signs, symptoms and laboratory abnormalities secondary to VOD (McDonald, et al 1993). Patients with moderate VOD required treatment of fluid retention or analgesics for an enlarged liver, but eventually had compete resolution of signs, symptoms and laboratory alterations secondary to VOD (McDonald, et al 1993). Severe VOD was defined as patients with VOD who experienced signs, symptoms, and laboratory abnormalities secondary to VOD that did not return to normal by day ⁺100 post HCT or died, in part secondary to VOD, whichever occurred first (Table III) (McDonald, et al 1993).

Table III.

McDonald grading system of SOS/VOD.

MILD	No adverse effects of liver disease No treatment of fluid retention No analgesic treatment of enlarged liver Complete reversal of signs, symptoms, symptoms and abnormal laboratory values secondary to VOD
MODERATE (One or More)	Adverse effects of liver disease secondary to VOD Treatment of fluid retention (Diuretic, fluid restriction and/or Na+ restriction)          AND Complete resolution of signs, symptoms and abnormal laboratory values secondary to VOD
SEVERE (Both)	Adverse effects of liver disease secondary to VOD Signs, symptoms and abnormal laboratory values secondary to VOD that did not all normalise by Day +100          OR Died in part secondary to VOD of the liver

McDonald GB et al, Ann Intern Med 1993(McDonald, et al 1993)

Bearman and colleagues further developed a grading scale of no, mild, moderate or severe toxicity as defined by patient weight gain or bilirubin concentration, respectively (Bearman et al, 1993). Bearman et al utilised this grading system to further prognosticate the risk of developing severe VOD post HCT based on the McDonald grading criteria (Bearman model) (Bearman, et al 1993).

More recently, the EBMT published SOS/VOD severity criteria (grading classification) in both adult and paediatric HCT recipients, respectively (Corbacioglu, et al 2018; Mohty, et al 2016). The adult criteria utilised time of onset, bilirubin concentration, kinetics of hyperbilirubinemia, degree of transaminase elevation, weight gain and renal function; these features thus defined mild, moderate, severe and very severe SOS/VOD with multi-organ dysfunction/failure as grading severity (Mohty, et al 2016). The paediatric criteria utilised the same adult categories and added ascites, coagulation abnormalities, pulmonary dysfunction and central nervous system injury (Corbacioglu, et al 2018). However, precise systemic organ toxicity associated SOS/VOD following HCT is not captured by these recent grading characteristics.

The most common organ dysfunction /complications that accompany SOS/VOD include: liver (bilirubin, transaminase, portal hypertension), renal (creatinine elevation, glomerular filtration rate, oliguria), weight gain/fluid retention, ascites (symptoms/intervention), pulmonary (hypoxia, dyspnea), central nervous system (encephalopathy), and cardiac (failure) (Table IV). Most of the past SOS/VOD grading criteria have lacked uniformity, universal acceptance, inclusivity of all systemic organs/complications affected by/resulting from past and current SOS/VOD, were not validated in multiple clinical databases and to some extent, are over 25 years old. The current National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE v5.0) SOS grading system has no grade 1 but only grade II-V classification and only includes serum bilirubin levels, medical interventions, coagulopathy and life threatening criteria (Fig 3). However, the parameters outlined in the NCI CTCAE v5.0 grading severity are uniformly accepted, validated, include signs and symptoms that characterise dysfunction of multiple organs, and are updated regularly (now version 5.0). We therefore envision an opportunity to utilise the NCI CTCAE v5.0 criteria in a new grading classification of SOS/VOD. As with other disorders, such as tumor lysis syndrome (TLS) and cytokine release syndrome (CRS), which involves multiple organs, new grading classification system were established utilising the CTCAE grading criteria (Cairo and Bishop 2004; Lee et al, 2019).

Table IV.

New proposed SOS/VOD multi-organ grading systems

Probably or Directly Attributed to SOS/VOD
Hepatic (Bilirubin) Hepatic (ALT/AST) Hepatic (Portal Hypertension) Renal (Creatinine) Weight Gain (Recent change) (Compared to Baseline) Pulmonary (Hypoxia) Central Nervous System (Encephalopathy) Cardiac (Failure) Ascites (Symptoms/Intervention)

ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Fig 3.

Current CTCAE v5.0 SOS grading severity.

In this context, we propose a new and revised SOS/VOD grading system based on the maximum organ/complications grade probably or directly related to SOS/VOD following HCT based on our new definition of SOS/VOD (Table II). We propose to grade the severity of SOS/VOD post HCT on CTCAE v5.0 criteria on liver (bilirubin, transaminase/portal hypertension), recent weight gain (compared to baseline), ascites (symptoms/interventions), cardiac (failure) and central nervous system (encephalopathy) (Table V). The proposed advantages of this novel SOS/VOD grading classification are outlined in supplemental Table SI.

Table V.

Sinusoidal obstruction syndrome (SOS) post HCT revised grading classification^*

	Grade 0	Grade I	Grade II	Grade III	Grade IV	Grade V
Hepatic
Bilirubin	No SOS	>ULN - 1.5 x ULN if baseline was normal; > 1.0 - 1.5 x baseline if baseline was abnormal	>1.5 - 3.0 x ULN if baseline was normal; >1.5 - 3.0 x baseline if baseline was abnormal	>3.0 - 10.0 x ULN if baseline was normal; >3.0 - 10.0 x baseline if baseline was abnormal	>10.0 x ULN if baseline was normal; >10.0 x baseline if baseline was abnormal	Death
Transaminase	No SOS	>ULN - 3.0 x ULN if baseline was normal; 1.5 - 3.0 x baseline if baseline was abnormal	>3.0 - 5.0 x ULN if baseline was normal; >3.0 - 5.0 x baseline if baseline was abnormal	>5.0 - 20.0 x ULN if baseline was normal; >5.0 - 20.0 x baseline if baseline was abnormal	>20.0 x ULN if baseline was normal; >20.0 x baseline if baseline was abnormal	Death
Portal Hypertension	No SOS	--	Decreased portal vein flow	Reversal/retrograde portal vein flow; associated with varices and/or ascites	Life-threatening consequences; urgent intervention indicated	Death
Fluid Retention
Weight Gain	No SOS	5 - <10% from baseline	10 - <20% from baseline	>=20% from baseline	--	--
Ascites	No SOS	Asymptomatic; clinical or diagnostic observations only; intervention not indicated	Symptomatic; medical intervention indicated	Severe symptoms; invasive intervention indicated	Life-threatening consequences; urgent operative intervention indicated	Death
Renal
Creatinine	No SOS	>ULN - 1.5 x ULN	>1.5 - 3.0 x baseline; >1.5 – 3.0 x ULN	>3.0 x baseline; >3.0 - 6.0 x ULN	>6.0 x ULN	Death
Pulmonary
Hypoxia	No SOS		Decreased oxygen saturation with exercise (e.g., pulse oximeter <88%); intermittent supplemental oxygen	Decreased oxygen saturation at rest (e.g., pulse oximeter <=55 mm Hg)	Life-threatening airway compromise; urgent intervention indicated (e.g., tracheotomy or intubation)	Death
Cardiac
Failure	No SOS	Asymptomatic with laboratory (e.g., BNatriuretic Peptide) or cardiac imaging abnormalities	Symptoms with moderate activity or exertion	Symptoms at rest or with minimal activity or exertion; hospitalisation; new onset of symptoms	Life-threatening consequences; urgent intervention indicated (e.g., continuous IV therapy or mechanical hemodynamic support)	Death
Central Nervous System
Encephalopathy	No SOS	Mild symptoms	Moderate symptoms; limiting instrumental ADL	Severe symptoms; limiting self-care ADL	Life-threatening consequences; urgent intervention indicated	Death

^*Maximal grade will be any organ system listed above and the maximal grade assigned probably or definitely secondary to SOS/VOD.

ADL, activities of daily living.

Prognosis and late effects of SOS/VOD following HCT

Patients with mild-moderate VOD and no evidence of associated renal or pulmonary dysfunction generally do well and have favorable survival rates (McDonald, et al 1993; Norvell 2015). The prognosis for patients with severe disease (accompanied by multi-organ dysfunction) is much less favorable but has improved in recent years with enhanced supportive care and the implementation of novel therapeutic agents (Ando et al, 2010; Carreras, et al 2011; Coppell, et al 2010; Dalle and Giralt 2016; Mohty, et al 2015; Nieder et al, 2011; Richardson et al, 2009; Richardson et al, 1998; Richardson et al, 2002; Richardson et al, 2017; Richardson et al, 2010; Strouse et al, 2016).

Prophylaxis and treatment of SOS/VOD following HCT

Ursodeoxycholic acid

Hydrophobic bile acids are toxic to liver parenchymal cells whereas hydrophilic bile acids are nontoxic. Ursodeoxycholic acid has been shown to provide liver protection by modulating the release and expression of inflammatory cytokines (Carreras, et al 2011), and exerting immunomodulatory effects (Carreras, et al 2011; Carreras and Diaz-Ricart 2011) (Fuste et al, 2004).

In randomised, controlled trials and historically controlled studies, ursodeoxycholic acid resulted in a reduction in SOS/VOD incidence and mortality (Essell et al, 1998; Essell et al, 1992). These results were further supported by a recent meta-analysis, which included data from 612 patients enrolled on four randomised clinical trials (Cheuk et al, 2015; Ruutu et al, 2014).

Heparin

The use of heparin in SOS/VOD prophylaxis remains very controversial as small sized randomised trials showed mixed results (Attal et al, 1992; Imran et al, 2006; Or et al, 1996).

Defibrotide

Defibrotide was found to be effective as prophylaxis for high-risk paediatric SOS/VOD patients, and demonstrated a favorable safety profile (Corbacioglu, et al 2012). The study population consisted of 356 patients at high risk of developing SOS/VOD after HCT using a myeloablative preparative regimen due to one or more pre-existing risk factors for SOS/VOD. Patients were randomised to receive defibrotide intravenously at 25 mg/kg per day given on the first day of conditioning until day 30 post HCT (n = 180) or no additional therapy (n = 176) (Corbacioglu, et al 2012). Patients who were assigned to, and ultimately received, defibrotide prophylaxis had a significantly lower cumulative incidence of SOS/VOD at 30 days post-HCT (12 vs. 20%). The adverse events rates were similar between groups. Several other studies in both adults and children, have reported significant reductions in the incidence of SOS/VOD following the prophylactic use of defibrotide (Richardson et al, 2018).

Treatment of SOS/VOD following HCT

Supportive care

One of the goals of supportive care should be to minimise extracellular fluid overload without worsening intravascular volume and renal perfusion / function in the context of maintaining weight within-2 percent to 5 percent of baseline. Hence, strict attention to fluid intake, salt load, urine output and daily weights remain the mainstay of supportive care. Some patients may develop ascites with or without pleural effusion, or pulmonary infiltrates in the absence of ascites, and become hypoxemic. Patients may require serial small volume paracentesis (to maintain renal perfusion and adequate lung volumes in children) for ascites that is associated with discomfort or pulmonary compromise. Haemodialysis or haemofiltration may be necessary for patients who experience renal failure or require more invasive methods to control fluid balance during the evolution of SOS/VOD.

Defibrotide

Defibrotide is approved by the US Food and Drug Administration and European Medicines Agency for the treatment of adults and children with SOS/VOD following HCT with renal or pulmonary dysfunction and “severe” SOS/VOD, respectively. When used in this population, about 30 to 40 percent of patients are expected to be alive beyond day +100 (Richardson, et al 2018).

An open-label, prospective, multicenter, randomised dose-finding phase II study compared 2 doses of defibrotide (25 mg and 40 mg/kg/day) given intravenously and found no significant difference in efficacy between the doses (Richardson, et al 2010). Rates of complete response (the primary endpoint), and day +100 survival post-HCT were 46% and 42%, respectively. Based on these efficacy and safety findings, defibrotide 25 mg/kg/day was selected as the dosage for future studies. Initiation of a randomised study for patients with multi-organ dysfunction had been rejected on ethical grounds due to the life-threatening nature of SOS/VOD in this scenario, the lack of effective alternative agents, and suggested impact on survival from several prior defibrotide studies. In a recent treatment trial, the efficacy and safety of defibrotide at 25 mg/kg/day in SOS/VOD with multi-organ dysfunction were compared with those of patients from a historic control group who met the study’s inclusion criteria (supplemental Table SII) (Richardson, et al 2016). The trial confirmed significantly higher day +100 survival rates for patients receiving defibrotide vs. best supportive care (38% vs. 25%, with a propensity-adjusted between-group estimated difference of 23%; P = 0.01). A large international multicenter compassionate use program was conducted between December 1998 and March 2009 in patients with SOS/VOD, with or without multi-organ dysfunction (supplemental Table SII) (Corbacioglu et al, 2016). The overall Kaplan–Meier estimated day +100 survival was 54% overall, and 58% for the 25 mg/kg/day dose group (the approved dose), and 65% in paediatric and 46% in adult patients, respectively). Patients without multi-organ dysfunction had an estimated survival of 65% compared with 40% among those with multi-organ dysfunction (Richardson, et al 2016).

The results from the historically controlled treatment trial are further supported by those of an observational database study using registry data from the Center for International Blood and Marrow Transplant; 8,341 patients who underwent allogeneic HCT between 2008 and 2011 were studied (supplemental Table SII) (Strouse, et al 2016). Within this subset, 4.5% (n = 376) patients developed SOS/VOD. Among these patients with SOS/VOD, 27%% (n = 101) had concomitant multi-organ dysfunction and 73% (n = 275) did not.

Liver transplantation and Transjugular intrahepatic porto-systemic shunt (TIPS)

Lastly, orthotopic liver transplantation and transjugluar, intrahepatic, porto-systemic shunts have been successfully performed in small numbers of patients with SOS/VOD (Dalle and Giralt 2016; Dignan et al, 2013; Fried et al, 1996; Kim et al, 2002; Richardson and Guinan 2001; Smith et al, 1996).

Conclusion

In summary, this review highlights our recent elucidation of the pathophysiology of SOS/VOD following HCT, contemporary risk factors (modifiable and unmodifiable) associated with SOS/VOD, current strategies for the prophylaxis and treatment of SOS/VOD once manifestations of severe disease develop, and risks of late effects in survivors of SOS/VOD following HCT. Importantly, we propose contemporary and more versatile diagnostic criteria for SOS/VOD facilitating a more comprehensive method of diagnosis, along with a novel, innovative, and practical severity grading classification of SOS/VOD that is based on specific organ toxicity included in NCI CTCAE v5.0 criteria. These recommendations should facilitate early identification of risk factors, a more timely and accurate diagnosis, contemporary grading classification, and innovative methods of treatment of SOS/VOD, thereby hopefully improving survival in this lethal disorder. Recent studies have suggested that early intervention with defibrotide in patients with SOS/VOD following HCT significantly improves overall survival (Kernan, et al 2018). Future studies incorporating our increased understanding of the pathophysiology (Fig 1 and Figure S1) and these new diagnostic criteria and grading classifications will also facilitate future safety and efficacy studies of new investigations in the prophylaxis and treatment of SOS/VOD following HCT. Future studies are warranted to validate the above mentioned revised VOD/SOS diagnostic criteria and grading classifications.

Supplementary Material

Supp info

Figure SI* Extracellular regulation of HSC activation.

Table SI. New proposed SOS/VOD grading classification advantages.

Table SII. Summary of treatment options for VOD following HCT.