Objectives To describe postexposure prophylaxis (PEP) against varicella zoster virus (VZV) in children being treated for malignancy in the UK and Ireland: the population at risk, frequency of exposure, clinical practice and attitudes among healthcare providers.
Design An observational study in three parts: (1) a retrospective survey of serostatus at diagnosis of malignancy, (2) collation of varicella zoster immune globulin (VZIG) dispensing data over a 3-year period and (3) an online survey of paediatric oncologists' clinical practice and beliefs in relation to VZV disease and its prevention.
Setting UK and Ireland.
Participants Children diagnosed with malignancy in 2009 (serostatus survey) or receiving VZIG between April 2006 and March 2009 (VZIG dispensing study). Paediatric oncologists and haematologists working in tertiary paediatric oncology centres and related shared care units in the UK and Ireland (physician survey).
Results Of 1500 children diagnosed with malignancy each year, at least 24% are VZV seronegative. Few centres make efforts to prevent household exposure by vaccinating VZV-susceptible family members. Exposures to VZV result in the administration of PEP to approximately 250 children with cancer annually: half receive an intramuscular injection of VZIG while the remainder receive a course of oral aciclovir. The choice of PEP is made by doctors. There is no consensus among paediatric oncologists as to which is the better option, reflecting the lack of a secure evidence base.
Conclusions A randomised controlled trial to compare the effectiveness and acceptability of VZIG and aciclovir as PEP against varicella is both desirable and feasible.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Primary varicella zoster virus (VZV) infection usually follows a benign clinical course and significant complications are uncommon in healthy children.1 By contrast, infection in the immunocompromised patient can result in significant morbidity and even mortality; in children with cancer, untreated varicella has a case death rate of 7%.2 3 Despite modern antiviral therapy, the disease continues to cause hospitalisation (including the need for intensive care) and chemotherapy interruption among this group, as well as later zoster.4,–,6 Consequently, VZV exposure requires active management among susceptible children receiving treatment for cancer, particularly in countries such as the UK where universal varicella vaccination has not been implemented. The suggested elements of this process include (1) identification of VZV-susceptible children by history and serostatus at diagnosis, (2) counselling of seronegative patients regarding the need to avoid VZV exposure as far as possible, (3) active varicella vaccination of other VZV-susceptible individuals within the household and healthcare team and (4) timely delivery of postexposure prophylaxis (PEP) to VZV-susceptible patients in the event of exposure7,–,9 (reviewed in ref. 10).
What is already known on this topic
▶ Chickenpox can produce severe or fatal illness in varicella zoster virus-naive children who are receiving chemotherapy.
▶ Two major methods of postexposure prophylaxis are in widespread use in the UK, both reflecting national guidance.
What this study adds
▶ A quarter of children diagnosed with cancer in the UK and Ireland are varicella zoster virus (VZV) seronegative but measures to reduce exposure are inconsistently applied.
▶ Exposures to VZV result in the need for prophylaxis in ∼250 children with cancer annually; half receive varicella zoster immune globulin, half oral aciclovir.
▶ Doctors select post-exposure prophylaxis on behalf of patients, but there is no consensus among paediatric oncologists as to which is the better option.
PEP is conventionally delivered in the form of varicella zoster immune globulin (VZIG), a blood product obtained by pooling immunoglobulin from VZV-hyperimmune individuals and delivered by intramuscular injection as soon as possible after exposure. VZIG and similar preparations have been shown to be effective in preventing varicella, especially severe disease, in North American and European studies.3 11,–,13 However, some trials suggested a high rate (50%) of breakthrough disease even when VZIG was administered early after exposure.13 One study revealed that 20 of 41 children with varicella undergoing acute lymphoblastic leukaemia (ALL) treatment had received VZIG as PEP, including 6 of 8 patients with severe varicella (one of whom died).5 Furthermore, the scarcity and cost of VZIG require its supply to be rigorously managed (eg, through the Health Protection Agency (HPA) in England and Wales), resulting in inconvenience and delays for patients.
Use of the antiviral agent aciclovir as an alternative means of PEP was first proposed in Japan, where VZIG is not licensed. Small observational studies in healthy children reported varicella rates of 0–77% after PEP with aciclovir.14,–,16 The drug was also reported to prevent or ameliorate varicella following household exposure of three children with ALL.17 Following favourable experiences during a period of VZIG shortage, paediatric oncologists in the UK were encouraged to consider aciclovir as an alternative to VZIG in widely consulted national guidance issued by the Royal College of Paediatrics and Child Health (RCPCH).9 Anecdotally, clinicians and families report many advantages and few concerns over the effectiveness of antiviral therapy as PEP. A more recent descriptive study from Japan reported the successful use of aciclovir in protecting nosocomially exposed children from VZV.18 There are, however, no controlled studies on the use of aciclovir as PEP in immunocompromised children. As a result, opinion is divided among paediatric oncologists and both forms of PEP are in widespread use in the UK and Ireland.19 The current study was conducted to review current practice and to explore the feasibility of a randomised controlled trial (RCT) to compare VZIG and aciclovir in this setting.
Serostatus of children with malignancy
It is standard practice to assess VZV serostatus at diagnosis of malignancy in childhood.9 We approached members of the CCLG (Children's Cancer and Leukaemia Group) at each of the 21 paediatric oncology principal treatment centres (PTCs) in the UK and Ireland to request the serostatus and age of all children diagnosed with malignancy in the calendar year 2009. Serostatus was ascertained retrospectively from computerised laboratory results. No personal identifiers were collected.
Accurate data are available on the dispensation of VZIG in England and Wales because its supply is strictly regulated by the HPA and meticulous records are kept. Although an annual summary report is prepared by the HPA detailing VZIG usage, malignancies other than leukaemia are not separately coded. We therefore hand-searched original paper returns from dispensing laboratories, to extract any further information that identified patients in the ‘immunocompromised’ category as children with solid tumours and lymphoma. Repeat VZIG prescriptions to individual patients were identified on the basis of birth date and initials. Records covering the period April 2006 to March 2009 were examined. Data were collated and analysed using Microsoft Excel.
The aims of this part of the study were to describe accurately how clinicians view VZV disease and what approach they take to PEP. Consultant paediatric oncologists on the CCLG's Data Centre mailing list were invited by email to complete a 15-question online survey via the SurveyMonkey website (see online supplementary appendix 1). Consultants at both PTCs and their related shared care centres (Paediatric Oncology Shared Care Units (POSCUs)) were invited to participate. Clinicians were asked to identify the centre in which they work so that a geographical picture of clinical practice could be built up.
Advice from the National Research Ethics Service was that ethical review for this service evaluation was not required under NHS research governance arrangements.
Data were returned from 12 of 21 PTCs (table 1). The total number of new malignancies among these centres in 2009 was 1152, with a mean age of 6.1 years (median age, 5.0). VZV serostatus was available for 68% of patients (>80% of those with haematological malignancies).
Sixty-three per cent of patients for whom results were available were seropositive for VZV (n=493, 43% of all patients) and 37% were either VZV seronegative (n=272, 24% of all patients) or equivocal (n=15, 1% of all patients). The mean age of seropositive patients was 7.7 years compared with 3.1 years for those who were seronegative (p<0.0001, Student t test). The mean ages for those with equivocal serology (4.1 years) and those for whom no result was available (6.4 years) were intermediate. The latter group probably represents a mixture of children whose serostatus had been measured elsewhere and those in whom it was not assessed, either by accident or by design (perhaps because of positive varicella history and/or low anticipated risk of severe varicella during treatment).
VZIG dispensation and exposure rates
We ascertained VZIG dispensations to children with cancer in England and Wales as fully as possible by hand-searching quarterly returns from all dispensing HPA laboratories. Over the period April 2006 to March 2009, 285 patients received 346 doses of VZIG: 237 patients received one dose, 40 patients received two doses, 6 patients received three doses and 1 patient each received five and six doses. Regarding diagnosis, haematological malignancies were present in 68% of patients (table 2). Of the patients who received multiple doses, >90% had leukaemia (other diagnoses: Ewing's sarcoma (1), brain tumour (2) and unknown (1)).
The number of doses dispensed per month varied between 2 and 24 with some seasonal variation (figure 1), resulting in an average rate of 115 doses per year. Assuming that VZIG administration is a reliable surrogate, this provides an estimate of VZV exposures among children cared for at centres in England and Wales that use VZIG as PEP. The physician survey (see below) identified eight centres in England and Wales that use VZIG exclusively and one further centre that uses either VZIG or aciclovir. From centre-specific data at the National Registry of Childhood Tumours, we estimated the proportion of children with cancer cared for at these centres to be 42–48% of the whole. Assuming that exposure rates do not vary systematically by centre or mode of PEP, the rate of VZV exposures across the entire UK and Ireland children's cancer cohort can then be estimated to be 240–273 per annum.
Two hundred and sixty-seven physicians caring for children with cancer and leukaemia were contacted by email via the CCLG Data Centre mailing list and invited to complete an online survey hosted by SurveyMonkey (see online supplementary appendix 1). Complete responses were obtained from 107 individuals and partial responses from another 10, for a total response rate of 44%. Responses from POSCUs were very patchy and indicated that they would take their lead from PTCs. The following analysis therefore relates predominantly to responses from physicians in PTCs (n= 67); these included at least one respondent from each of the 21 centres.
Perception of VZV disease
Ninety-nine per cent of PTC respondents thought VZV was a problem in paediatric oncology practice; 49% considered it a major problem and 50% a minor problem. Free text comments in response to this question emphasised (1) the potential for severe morbidity/mortality, treatment interruption and time in hospital; (2) inconvenience, anxiety and discomfort associated with PEP for child and family; (3) the logistic challenges of providing PEP. When asked about VZV disease or exposure in their own patients within the last 3 months, 56% of PTC respondents knew of at least one case of varicella, 67% at least one case of zoster and 59% had at least one patient who had required PEP. Only 40% reported that their centre recorded VZV disease prospectively.
Prevention of exposure in seronegative patients
Twenty per cent of PTC respondents reported that counselling of seronegative families was conducted by a designated individual within the team, but more commonly (54%) this was the responsibility of whoever next saw the patient after a seronegative result was ascertained. Regarding written advice for these families, 56% reported that this is currently provided, but there was overwhelming (95%) interest in obtaining suitable materials from this study group in the future. This suggests the need for good, standardised written materials for families. Current advice from the Department of Health (DH) and RCPCH recommends varicella vaccination of healthy seronegative contacts within the household of individuals immunocompromised individuals susceptible to VZV.8 9 However, only 14% of PTC respondents reported attempts to identify such individuals while only 5% recommended their vaccination ‘every time’. Few respondents rechecked serology in patients who were seropositive at the start of therapy, but 22% did so upon exposure (consistent with DH policy8). A few were starting to recheck serology regularly in all patients.
Choice of PEP
The majority of respondents (52%) recorded VZIG as their personal preference while 40% preferred aciclovir. There was some discordance with unit policy, which was slightly more likely to recommend VZIG (58%) compared with aciclovir (33%). When responses were analysed by centre, 11 could be assigned to a ‘VZIG only’, 7 to an ‘aciclovir only’ and 3 to an ‘either/both’ unit policy. Most respondents ‘rarely’ (33%) or ‘never’ (20%) discussed the choice of PEP with families, while only 13% ‘always’ discussed it. We asked respondents to rate various factors for their importance in influencing the choice of PEP. The following were ranked ‘very important’ by at least half of the respondents: unit policy (79%), published evidence (66%), national guidelines (50%) and own experience (50%). Among those individuals working in centres where aciclovir is recommended as a single-agent PEP, 14 (67%) of 21 stated that published evidence was ‘very important’ in influencing their choice of PEP.
Use of alternative forms of PEP and attitudes to an RCT
Most respondents could imagine using alternative forms of PEP, particularly as part of an RCT (71%) or if there was a contraindication to their preferred agent (65%). Forty-nine per cent would change their practice if evidence of non-inferiority were provided by an RCT (62% of those in centres that currently use VZIG but only 31% of those using aciclovir). In free text, 15 respondents stated that such a trial ought to show no more than a 10% difference in efficacy, 8 wanted a 5% cut-off and 6 did not want to be drawn. Regarding attitudes to randomisation in such a trial, responses were disparate depending on whether respondents currently use aciclovir or VZIG, with the latter more likely to have ‘no problem’ (60%) or ‘some misgivings’ (35%) while aciclovir users were more likely to have ‘some misgivings’ (52%) or feel ‘very uncomfortable’ (14%). This split was borne out by the free text comments, of which there were many (24 from PTC physicians and 11 among respondents from POSCUs). Among current aciclovir users, several comments intimated reluctance to submit patients to a painful injection, perhaps summarised by the following comment: ‘We have worked hard to take the pain out of cancer care—when poss—and this is just putting it back in’. Other respondents expressed reservations about the logistic challenges of rechecking serology and providing PEP to patients living far from the PTC.
This study confirms that VZV exposure and disease are frequent and significant clinical problems in paediatric oncology practice in the UK and Ireland. We show that at least a quarter of children with malignancy are VZV seronegative at diagnosis. The rate of susceptibility among the cohort of children on therapy is likely to be higher because treatment is longer for certain malignancies that are common in younger children (eg, ALL). Furthermore, we strongly suspect that susceptible children are included among the large group for whom serostatus could not be ascertained. Despite the known risks of varicella in children with malignancy, measures to prevent this are inconsistently applied. In particular, DH advice to identify and vaccinate VZV-susceptible family members is rarely implemented, indicating a lost opportunity to prevent the most infectious type of exposure with a simple intervention. The approach to PEP is highly polarised among centres, with the result that patients are given VZIG or aciclovir in equal measure, but few are offered any choice. Both camps believe they are following national guidelines and assert published evidence as an important influence on their choice of therapy. Most oncologists appear reluctant to consider the alternative form of PEP unless in the context of an RCT. Paediatric oncologists who currently use aciclovir as PEP are less willing to submit patients to such a study, citing issues with VZIG's acceptability and logistic challenges in timely delivery.
Owing to our study design, we were unable to express the rate of VZV exposures in terms of person years of therapy. Furthermore, no epidemiological data were collected regarding VZV disease and the effectiveness of PEP. Alternative study models that could have provided this information were rejected on the grounds of size and complexity, but these are important issues that we would wish to address within a future study. Previous surveys of varicella during ALL therapy have suggested a cumulative risk of varicella of around 15%4 5 and a 30% risk of VZV exposure20; however, there are no recent data on VZV exposure and disease in the UK and Ireland cohort.
Our study highlights an important issue of clinical governance in relation to VZV PEP in the UK. The lack of published evidence for the effectiveness of aciclovir represents a major barrier to the extension of patient choice. Were the two treatments equally effective, involvement of patients and families in decision-making on PEP could be promoted. Informal soundings from patient families through the National Alliance of Childhood Cancer Parent Organisations (NACCPO) indicate that this would be welcomed (eg, see online supplementary appendix 2). Furthermore, there could be a strong health economic argument for the use of antiviral therapy as opposed to VZIG. Unfortunately, in contrast to VZIG, there has been no attempt to collect prospective data on the effectiveness of aciclovir as PEP. We are of the opinion that only a well-conducted RCT could provide data of sufficient quality to support a change in practice. The next step will be to assess the willingness of patients and families to participate in a randomised trial, and it is proposed to conduct a pilot study at selected oncology centres. The implications of a successful study would reach well beyond paediatric oncology practice in the UK.
JB, JC, PH, JB, RS, SM, SP and SH contributed on behalf of the Supportive Care Group of the Children's Cancer and Leukaemia Group. We are extremely grateful to all those who submitted data and responded to the online questionnaires, to CCLG and NACCPO.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.