<scp>BSH</scp> Guideline: management of thrombotic and haemostatic issues in paediatric malignancy

Abstract

The guideline was compiled according to the British Society for Haematology (BSH) process at www.b-s-h.org.uk/guidelines. The scope of the guideline was agreed by the Writing Group, which consisted of UK-based consultant paediatric haematologists with a special interest in haemostasis and thrombosis or haematological malignancy. A literature search was performed covering relevant articles published up until July 2016 (details in Appendix S1) The Writing Group reviewed the literature search and drafted the manuscript, which was subsequently revised, by consensus, by members of the Haemostasis and Thrombosis Task Force of the BSH and the BSH Guidelines Committee. The guideline was then reviewed by a sounding board of the BSH, with particular emphasis to its application to the UK setting. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and strength of recommendations (http://www.gradeworkinggroup.org). The incidence of VTE in children with cancer varies widely in reported studies, due to differences in patient populations, treatment protocols and methods used to identify and define a thrombotic event. Studies using radiological screening to diagnose asymptomatic thromboses have identified rates of up to 40% (Mitchell et al, 2003a; Albisetti et al, 2013), although the significance of many of these asymptomatic events is unclear. The incidence of symptomatic VTE varies according to cancer subtype, being more common in acute lymphoblastic leukaemia (ALL) (5% from a meta-analysis published in 2006, range 3–14%), sarcoma (7–13%) and lymphoma (2–17%) and less common in children with other malignancies (Caruso et al, 2006; Nowak-Göttl et al, 2009; Piovesan et al, 2014; Van Ommen & Chan, 2014). Many factors are relevant to the risk of VTE in these groups of children (see Table 1), of which congenital thrombophilia and the use of central venous lines (CVLs) will be discussed in more detail. The evidence for congenital prothrombotic defects increasing the risk of VTE in childhood cancer is conflicting and most studies have been carried out in the setting of ALL. The largest prospective study, performed in 301 children with ALL treated on the Berln-Frankfürt-Münster (BFM) 90/95 protocol, identified a strong association between prothrombotic defects and the incidence of symptomatic VTE during treatment (Nowak-Göttl et al, 1999). In this study, patients were screened for the presence of three genetic mutations (MTHFR C677T, F2 G20210A and F5 G1691A), deficiencies of protein C, protein S and antithrombin and raised levels of lipoprotein (a). Of those children with no identified prothrombotic defects, 2·2% had a thrombotic event, compared to 39·5% with one prothrombotic defect and 80% with two defects. In contrast, in two other prospective studies of childhood ALL, none of the children with prothrombotic defects on thrombophilia screening had a thrombotic event during the study periods (Mauz-Körholz et al, 2000; Mitchell et al, 2003a). Of note, the overall incidence of symptomatic VTE was much lower in these two studies than that in the earlier study (Nowak-Göttl et al, 1999). Similarly, in a prospective study of 67 children with malignancies other than ALL, no effect of underlying thrombophilia was observed, but only 4 thrombotic events in total were reported in the study (Wermes et al, 1999). In summary, there is some evidence for congenital prothrombotic defects providing an additional VTE risk in children with ALL, but this is most apparent when the overall incidence of VTE is relatively high due to the nature of the treatment protocol. It is not clear if this effect applies to protocols for ALL in which the overall VTE risk is lower, or to patients with other malignancies. Importantly, there are insufficient data to support general changes in management for those with an identified prothrombotic defect. Therefore, routine thrombophilia screening cannot currently be recommended. In the very rare case of a known family history of antithrombin deficiency, there may be an argument for screening, given the particularly high risk of VTE in such kindreds. However, even if congenital antithrombin deficiency is identified, there is no clear evidence for how this should influence management (see sections on replacement therapy and prophylactic anticoagulation). Central venous lines are used for the administration of chemotherapy and blood products and for venous sampling. CVL-associated thrombosis is a frequent complication. The patient-related and disease-related factors in Table 1 contribute to the risk of occurrence, with rates of line-associated thrombosis highest in ALL, where it accounts for up to half of the episodes of VTE (Fratino et al, 2001, 2005; Molinari et al, 2001; Elihu & Gollin, 2007; Pinon et al, 2009; Ociepa et al, 2010). Different centres vary in respect to their use of internal versus external devices, site of vascular access, and timing of CVL insertion in relation to induction therapy for ALL, with no consensus in terms of the effect of these factors on thrombotic risk. For guidance on these issues, see section on reducing the risk of VTE. Venous thromboembolism risk assessment and prevention is a routine part of the care of adults in hospital in the UK. It is also recommended that ambulatory adults with cancer should be assessed for thrombosis risk, and scoring systems have been developed to assist with this (Watson et al, 2015). However, similar assessment tools are not in routine use for children. One study has validated a simple model to predict VTE in children with ALL treated in 3 separate protocols in a multicentre setting, using a combination of risk factors (congenital thrombophilia, presence of CVL and type of induction therapy) (Mitchell et al, 2010). A high-risk group could be identified in which the rate of symptomatic VTE was 64·7%, as compared to 2·5% in the low risk group (P < 0·001). This scoring system would be of less value in identifying a high-risk group among patients treated on the same ALL protocol and is obviously not applicable to children with other malignancies. It does however highlight the potential usefulness of risk assessment in identifying patients at high risk of VTE who might be candidates for interventional studies. Children with cancer should be encouraged to reduce their risk of VTE by maintenance of good hydration, early mobilisation whenever possible and removal of CVLs when no longer required. The use of compression stockings or other mechanical measures in this patient group has not been studied, and there is no evidence for their use in adult patients except in the post-surgical setting. In addition, compression stockings are impractical in young children and may cause harm if they are poorly fitted. Therefore, their use should only be considered in post-surgical adolescent patients. Given the multiple risk factors for VTE in adolescents with malignancy, combined oral contraceptives should be discontinued and alternative therapy for menstrual suppression should be discussed with the patient (see section on menstrual suppression). Amongst children requiring central venous access, the only data regarding peripherally inserted central catheters (PICCs) showed a higher incidence of thrombosis when compared with external tunnelled lines in a non-randomised trial (Jaffray et al, 2015). This finding was in keeping with a meta-analysis of adult studies that identified a higher rate of thrombosis in PICC lines, particularly in adults who were critically ill or had malignancy (Chopra et al, 2013), although more recent data in adults have shown a lower risk of thrombosis compared with traditional CVLs in a population with acute leukaemia (Cerchione et al, 2016). Thrombosis occurs less frequently in association with internal lines (ports) than external tunnelled lines, such as Hickman or Broviac catheters (McClean et al, 2005; White et al, 2012; Mangum et al, 2013). A multicentre randomised controlled trial (RCT) in adults with malignancy, comparing complication rates between external tunnelled lines, ports and PICCs for the administration of chemotherapy, is currently ongoing [Cancer and Venous Access (CAVA); ISRCTN 44504648]. Sub-analysis of data from the PROTEKT (PROphylaxis of ThromboEmbolism in Kids Trial) study found that the highest rate of line-associated thrombosis occurred in relation to femoral venous access (Male et al, 2005). Although this study did not consider children with malignancy separately, they would be expected to have a relatively higher risk, particularly in the presence of venous compression from an abdominal or pelvic mass. Avoidance of femoral venous access may not always be possible (e.g. in an individual with a mediastinal mass precluding the use of an upper limb vein) although an alternative site of access should be sought as soon as practicable. Studies comparing rates of line-associated thrombosis between jugular and subclavian vein access have shown conflicting results (Male et al, 2003, 2005; Pinon et al, 2009; Ribiero et al, 2012). One study compared insertion technique (percutaneous versus venous cut-down) and identified a higher rate of thrombosis with the former (Male et al, 2003). There is currently no formal guidance for the optimal timing of CVL placement in relation to induction therapy for ALL, although some centres choose to avoid inserting a CVL until the end of induction wherever possible. This practice be in of the for in children. Studies comparing thrombotic rates between lines early and those no (McClean et al, 2005; et al, et al, 2015). of these studies the nature of data in type and site of used and differences in of early versus The use of as treatment for ALL is a risk for VTE. The effect is by a in levels of and S most as as levels of (Mitchell et al, et al, 2000; et al, This has to to reduce the risk by replacement in the of or antithrombin levels of and as as levels of and of are also known to be by to a more In most studies have shown to be poorly in levels of antithrombin other and in reducing of data an effect on VTE rates are on of practice in centres and to et al, 2009; et al, 2013). the use of prophylactic during the of effect have been in children or In addition, there are that use in this setting might to of the and could reduce leukaemia rates et al, et al, 2013). be shown to of in patients with ALL during periods of et al, et al, et al, Nowak-Göttl et al, et al, However, published studies to have to a clear in the incidence of VTE in this group of children (Mitchell et al, et al, 2013). A study in adults with ALL did a lower rate of VTE in those who but to the study et al, has been the of the of studies of prophylactic in children with have been performed with in children with ALL. were and data have not a in VTE incidence et al, studies have a of using prophylactic in of children with ALL identified to be at high risk of but these were not randomised et al, Mitchell et al, 2010). has also been in combination with et al, In this study of children with ALL, levels and and from the of until antithrombin levels The rate of symptomatic VTE in the two groups the study was and This was (P but the study was randomised In data support the potential of using these two in combination et al, The only of prophylactic in children with cancer was using with a of et al, an the study was as the incidence of VTE was higher in the group oral are a alternative to particularly in the setting of paediatric ALL, as they of antithrombin and be In data on from children with ALL and studies on are et al, 2006; et al, 2013). studies have to be published in this patient although there is an ongoing study In summary, there is no evidence for the use of replacement therapy or prophylactic in VTE in children with There is some evidence for the of combined with in children with ALL treated with but this There is also a that may have a in VTE in high risk which is with the known in adult cancer patients (Watson et al, 2015). paediatric studies are in the it would be to consider using prophylactic in the particularly high-risk group of adolescent patients with additional risk factors for VTE et al, In addition, in the very rare of known congenital antithrombin deficiency, with prophylactic should be considered during periods of additional risk, for the of in children & It may also be to such patients with during periods when antithrombin levels are by although the evidence for this is The to a with a thrombosis is by the presence of the risk of or and the to reduce the risk of et al, 2015). treatment of symptomatic VTE in children with 3 of et al, et al, 2012; et al, 2013). There are no randomised studies comparing CVL versus removal in from adult cancer patients that if a patient is with the there is risk of or of thrombosis et al, Therefore, if the CVL is in a good and evidence does not removal if is the is to be it is to of et al, et al, 2012). this it may avoid in the optimal of is not although it has been that children with events and of thrombosis CVL removal could be with a of This is currently being in children with VTE in the study although it should be that this study children with 3 of the of guidance prophylactic is given a This has not been to it influence the risk of of note, has not been shown to be in reducing risk of thrombosis et al, et al, Therefore, the to is not clear and should the presence of other ongoing risk is an risk for in children. In the which children with had a haematological malignancy and had et al, 2014). in this study was with in a high of There are no comparing versus no in children with is on from adult data and studies of management in children in a of studies have reported the effect of on and the risk of in association with from the BSH and the for children with The presence of only an of venous is not as a to but is where there is more et al, et al, 2012). recommendations are on the reported of in this setting and of an in thrombosis and in those not et al, 2007; et al, 2010). to the of these two the study reported in those not although there was no to other et al, 2014). venous thrombosis is most in children treatment for ALL, in which the timing is to the administration of et al, et al, 2016). data on in these where there may be additional regarding the risk of are relatively In the study, children were There was one to and no due to et al, 2016). In the Society for Haematology and study, children but was reported or to the et al, 2015). Although these data are in terms of there does not to be evidence to that the management of children with ALL should from other children with Although less common than thrombosis and VTE at other does with in children with cancer et al, 2014). There are data on which to treatment recommendations and is the same as for other VTE Guidelines for adults with cancer although there is no evidence for this compared to 3 (Watson et al, 2015). of treatment 3 should be on an individual of risk factors at ongoing risk cancer and risk. high rates of asymptomatic thrombosis to have been reported in two studies in children with cancer additional as part of trial (Mitchell et al, 2003a; Albisetti et al, 2013). on asymptomatic VTE at other in children is One study as an finding in of a paediatric although not of these events were reported et al, 2010). In a meta-analysis of adult studies, where has been for other than the of was reported at 2·5% and was highest in patients with cancer et al, 2010). have been reported for VTE and lower limb vein thrombosis in adults with cancer routine abdominal and pelvic et al, 2007; et al, et al, 2010). The history of VTE in adults with cancer is not but there are some data to an risk of thrombosis and (Watson et al, 2015). the of this it is recommended that VTE in adults is treated in the same as symptomatic In the of other VTE in children should be in the same although this should other risk particularly the risk of Given the reported high rates of thrombosis to a an alternative in this particularly where the of the event is is to with and only if of the occurs or the patient subsequently However, is in childhood cancer with asymptomatic thrombosis et al, et al, and may treatment if the thrombosis is to be a recent event. In adults with is recommended on overall et al, 2015). data not for is to be the for children as This to the for in therapy to the risk of and potential or which might to in it is recommended that therapy is in children. levels should be for a of Given the of antithrombin particularly in children the use of an antithrombin is to be more an antithrombin may not in such patients in In children with an rate less than of and may be required. In patients at high risk of treatment with may be more given its and of In an may be considered if is (e.g. due to or but it should be as soon as it to occurs during the should be and with it should be that may not be if is being and blood may be if is The same when VTE in and young except that higher of or are to levels (see Table maintenance using are currently in children and in may in terms of of administration and The use of antithrombin replacement during therapy has been but clear evidence to support its use is Although is a risk during the management of VTE in patients with antithrombin deficiency, in practice this does not to be a it is by of then antithrombin could be However, care should be as it to a in It to be by a in and will be required. management of therapy on the risk, the of the and the risk of VTE during on published should be for at to an et al, 2006; & 2007; et al, 2014). be an there are no regarding However, it is at prophylactic to the For those on to an with should be considered in those at high risk, recent VTE. studies that the risk of VTE during a is in the the event, but is for up to 3 et al, & an may be in patients with high risk lower in is for a There are no data to define a for of therapy in those being treated for VTE. that a is for of et al, 2006; et al, although a lower of is used by some centres in patients with no other risk factors for et al, In or if thrombosis is more recent thrombosis the the should be using For those with a lower risk of an alternative is to of when the is between and and of therapy when the is This is recommended in a of data on or et al, et al, and has been to paediatric practice et al, In should be and levels 1 to reduce the risk of in patients on The as to this is in patients treatment for VTE that occurred in relation to given that the to the prothrombotic and that the is not of an risk of However, the combination of low and particularly in a patient with risk of would be a Therefore, consensus is to levels and the with use if therapy is required. is a of chemotherapy for ALL and early due to is with an et al, Therefore, in children with the risk of with has to be the to leukaemia from the UK and the have been published this children treated on the protocol had a VTE and of were who had had a The to the of antithrombin and there were no VTE et al, 2010). The children who had VTE on were to and 4 had VTE et al, It is if were at the of and not were to although and for are to be The timing of cannot be from published data but the of or of in therapy is It has been that be considered have to or to may additional in should be with prophylactic of if the of paediatric studies have been to the of prophylactic practice is on which routine for or if the are or the patient is et al, 2001; et al, 2016). on adult studies at a when was used as an and there were in for A that there was insufficient evidence to from practice but there was no evidence that a prophylactic et al, 2012). A study found that prophylactic and the incidence of there was no to the in those patients who et al, 2013). Recommendations regarding for a of be found in the for and guideline for blood and the BSH guideline for and children et al, 2016). on the for A recent of UK paediatric centres found a range of with the using a of but a of to be et al, There are data from studies that lower are to be et al, 2000; Van et al, et al, 2007; & et al, but this to be recommendations be to lower the in children with cancer be by the underlying its treatment an underlying congenital A and should be performed on patients at of haemostasis may be if these are or the history a that the of particular with malignancy but is most in patients with particularly acute leukaemia It cause and should be as the BSH treatment of the underlying et al, is a most with et al, but also in sarcoma et al, The are by low and low and The use of is recommended for or but the may be until the underlying is et al, should be considered if the at was or the of the patient in the might have to a (e.g. or very should also a and should not be to be due to in a with However, in the of these are not recommended to or in patients with ALL in the as there is no evidence for an risk of or for in results in this setting. the of the is with a prothrombotic even levels are replacement is recommended to if the is et al, replacement is also if is with low levels and there may be value in for levels the range if is is more to levels than et al, Nowak-Göttl et al, is not currently for replacement in in but should be considered if to be or levels are to support with et al, 2014). The use of prophylactic in patients with haematological malignancies and is currently being in a the and of in patients with haematological malignancies with ISRCTN In patients who are oral and are frequently used in although use should be if is In the of ongoing of may also be of there are some of being in patients with which was to blood et al, However, there are the risk of venous and thrombosis with its use in this setting and should be assessment should be of the for menstrual risk at risk of VTE and the for The combined oral should be discontinued at in of the VTE risk. treatment for adolescent at risk of (e.g. and The factors should influence the of cancer risk, potential on and individual has been used for menstrual suppression but should be in those with at high risk of is to and studies have shown a association between to higher and VTE et al, et al, is not to may be in patients at high risk of VTE 2012). is as a to for using a in the risk of with their However, more recent studies conflicting results and data on longer to a on there is insufficient evidence to for this et al, 2009; et al, et al, et al, 2014; et al, et al, et al, et al, 2016). are it should be that they cause an of a should be considered with the & of cause of the and and in with In of their use should be reviewed in patients with ALL, they should be discontinued in the maintenance of therapy when is It should also be that is in the of therapy of and relevant if chemotherapy is and early is In this or are more for menstrual suppression may the use of prophylactic measures to (e.g. may be in this but should be if is et al, and could also be should be sought in of to these protocols for should be where The to for in the literature The would also to the BSH Haemostasis and Thrombosis Task the BSH and the BSH Guidelines for their support in this of the reviewed the literature search and to the and of the The BSH the during the of this have a of to the BSH and Task Force which may be reviewed on of the Writing Group will the Writing Group if evidence that would the strength of the recommendations in this or it The will be and from the BSH if it recommendations are an will be published on the BSH The is not for the or of by the than should be to the for the