Varicella-zoster virus (VZV) can be fatal or induce severe complications in children with acute lymphoblastic leukemia (ALL). Untreated VZV infection has a mortality rate of 7%–10% in this group of patients.1 In about 30% of children who do not receive antiviral therapy, there is dissemination to lungs, liver and heart.2 The routine implementation of prompt prophylaxis with immunoglobulins and acyclovir, in particular at exposure to VZV, has significantly reduced the incidence of severe complications of VZV infection.3–5 However, in countries where the VZV vaccine is not part of the national vaccination program, the risk of VZV exposure still compromise the daily lives of many children with ALL during their 2–3 years of antileukemic treatment. Their families have to be alert to potential infection at all times, and are not always aware of an exposure, because the classic varicella presentation with a vesiculopapular rash is not always present.6,7 To minimize the risk of VZV exposure, Danish preschool children who have not had VZV are isolated from normal day care during their cancer treatment. Until recently it has been internationally approved practice to vaccinate children with ALL against VZV, and several studies have shown that the vaccine can be used safely and effectively in these patients.8–11 However, a recent study argued that the risks associated with vaccination may be greater than those associated with acquisition of the infection itself.12 The main concerns are of one reported case of death related to VZV vaccination and the possible increased risk of relapse of ALL due to discontinuation of maintenance therapy before, during and after vaccination.13 In most studies, maintenance therapy has been interrupted for 14 days around the time of the first vaccination to ensure a proper immune response to the vaccine. Only a few and relatively small studies describe vaccination of children with ALL without interruption of chemotherapy. In these studies, the children achieved a sufficient immune response toward the vaccine.14–16 Since 1997, VZV-seronegative children have been vaccinated against VZV without concomitant interruption of maintenance therapy at the Copenhagen University Hospital, Rigshospitalet, Denmark. In this retrospective analysis, we describe the mortality and morbidity related to VZV vaccination and infection among seronegative children with ALL.
MATERIALS AND METHODS
The analysis included all 223 children diagnosed with ALL from 1996 to 2011 at Copenhagen University Hospital. Of these, 78 (35%) were VZV seronegative at the time of diagnosis. Five children were excluded: 3 had moved to another part of the country and their files were not retrievable and 2 died within 2 weeks of diagnosis. Thus, altogether 21 VZV-seronegative girls and 52 boys, aged 0–16 years at the time of diagnosis (median: 3, range 11/12–16 5/12) were included in the analysis. Thirty patients had standard risk ALL, 29 had intermediate risk, 12 had high risk, 1 had infant ALL and 1 was diagnosed in Italy. Risk grouping criteria have been reported previously.17 The children were treated according to the ALL protocols 1992 (n = 25), 2000 (n = 28), 2008 (n = 18) and interfant 99 of the Nordic Society of Paediatric Haematology and Oncology and in 1 case mainly by the Italian protocol Italian Association of Pediatric Hematology and Oncology.18,19 All patients received maintenance therapy with 6-mercaptopurine (75 mg/m2/day) and methotrexate (MTX; 20 mg/m2/week) targeted to a white blood cell count of 1.5–3.5 x 109/L (ALL92/ALL2000) or 1.5–3.0 × 109/L (ALL2008). Furthermore, depending on the protocol, patients received high-dose MTX (HDM; 5 g/m2 with leukovorin rescue and intrathecal MTX), vincristin (2.0 mg/m2 × 1) with dexamethasone, or prednisolone for 5–7 days at alternating intervals. Seven children experienced relapse of ALL at a median time of 2 years and 9 months from diagnosis. Four of these children were immunized within a median time of 19 months (range: 17–45 months) before the time of recurrence, and only 1 of these had maintenance therapy discontinued for 3 days related to a VZV vaccine-induced rash.
The children fulfilled the following criteria at the time of VZV vaccination: (1) seronegative at the time of diagnosis, (2) in complete remission, (3) age ≥ 1.0 year, (4) lymphocyte count ≥ 0.6 × 109/L at time of vaccination (although ≥0.8 × 109/L during the first 6 months of 1997) and (5) receiving MTX/6-mercaptopurine maintenance therapy. The children were vaccinated with the live attenuated vaccine Varilrix (GSK) without discontinuation of chemotherapy before, during or after the vaccination (except from the routine criteria for pausing due to, eg, fever or cytopenia). If the children were seronegative after the first vaccination, they were vaccinated once more. In 2007, international guidelines changed and all children were vaccinated twice.20 However, the children who developed a rash after the first vaccination did not receive a second vaccination irrespective of the presence of VZV antibodies. From 1997 to 2011, the vaccinated children were only treated with acyclovir, if they developed a vaccine-induced rash after vaccination. To make sure they did not transmit VZV to seronegative children, they were isolated from the ward from day 7 to 40 after vaccination. In contrast, those children vaccinated between 2011 and 2013 (n = 11) received acyclovir prophylaxis (valacyclovir 20 mg/kg/dose twice a day) from day 7 to 40 after the time of vaccination to avoid this isolation. Day 7 was chosen to ensure that the primary viremia was completed.
The following data were retrieved from hospital files: dates of vaccine-induced rash, late chickenpox, reactivation of VZV and herpes zoster; lymphocyte count at the day of vaccination; number of days of acyclovir prophylaxis and treatment; number of days of interruption of maintenance therapy; number of days spent isolated from the ward; length in days of hospital admissions and administration of chemotherapy around the time of vaccination (HDM from 3 weeks before to 2 weeks after vaccination and dexamethasone from 2 weeks before until 40 days after vaccination). The number of chickenpox lesions was registered and divided into 3 groups: a small number (0–49), a moderate number (50–99) and multiple (>100).
The laboratory testing for the characterization of the varicella type as either vaccine or wild-type has not been consistent during the analysis period. Therefore, we defined chickenpox as vaccine-induced VZV if developed within 60 days of vaccination, whereas reactivations were defined as repeat incidences of rash within 60 days of vaccination. Chickenpox developed more than 60 days after vaccination was defined as late chickenpox. Day 60 was chosen to ensure that the rash was not vaccine induced because previous studies have shown that immunocompromised children are at risk of developing vaccine-induced rash for longer than healthy children.15 From 1997 to 2006, the parents of the vaccinated children were given a questionnaire at the day of vaccination. They were asked to register incidences of rash, fever, medical contacts, local reactions and other adverse effects associated with the vaccination. As part of this analysis, a questionnaire was sent to all parents, unless their child had died (n = 69), and 53 received a follow-up call to clarify their response. The children’s VZV titer was measured in 38 (84%) of the 45 immunized children after the first vaccination. The remaining 7 children all developed a vaccine-induced rash and were considered seropositive. Titers were not measured routinely after the second vaccination, and all children were regarded as being VZV protected irrespective of their antibody response. A follow-up titer was measured in both immunized (33/45) and nonimmunized children (13/29) during 2012 and 2013 to analyze their long-term seroconversion. There was a median follow-up time of 4 years and 6 months (2/12–15 7/12) from time of vaccination, and a median follow-up time of 7 years and 9 months (8/12–16 2/12) from time of late chickenpox. Seven children received a bone marrow transplant within a median time of 3 years and 3 months (range: 7/12–9 9/12 years) from diagnosis and were excluded from the follow-up blood tests. All vaccinated children were kept in isolation until their vaccine-induced rash had cleared or until they had had 2 vaccinations.
Frequencies were compared using Fisher exact test. Quantitative data are presented with medians and ranges. The association between explanatory variables (HDM, dexamethasone, acyclovir and lymphocyte count) and the risk of rash, as well as of getting late VZV, was assessed using survival analysis: that is, for each patient time-to-rash was determined, censoring patients at 60 days after vaccination (corresponding to the patients classified as no longer being at risk of developing vaccine-induced rash). Those patients who did not appear to develop a rash after the first vaccine received a second, that is, multiple survival times occur. The data were analyzed using an extension of the Cox regression model to multiple events and time-dependent explanatory variables. The model was stratified on vaccination (first or second) and the 2 baseline hazards were assumed identical (considering the time point of vaccination as time 0 for both vaccinations) using a Prentice-Williams-Peterson gap time model.21 Treatments given after vaccination (acyclovir and/or dexamethasone) were treated as time-dependent explanatory variables. The effects of the explanatory variables were further assumed equal for both vaccinations. Univariable and multivariable analyses were performed and, as the dataset is relatively small, a backward elimination procedure was applied to identify the most important explanatory variables. Analysis of time to late VZV was performed using years since diagnosis as underlying timescale and delayed entry at the time of first vaccination. Survival estimates were calculated by the Kaplan–Meier method and univariable Cox regression analyses with rash (yes/no), number of vaccinations (1 or 2) and acyclovir treatment (yes/no) as time-dependent explanatory variables were performed. The statistical analyses were performed using SAS version 9.4. P values less than 0.05 were considered significant.
Altogether 45 children were vaccinated. They received their first vaccination within a median time of 37 weeks (range: 3–102 weeks) from the beginning of maintenance therapy. Fifteen of these (33%) were vaccinated twice with a median time of 11 weeks (range: 5–22 weeks) between the 2 vaccinations. No child experienced serious adverse events such as local inflammation, VZV pneumonitis, visceral dissemination, secondary bacterial infection or death due to VZV vaccination.
Nine children, including 6 children still on maintenance therapy, developed late chickenpox despite their vaccination within a median time of 49 weeks (range: 11–86 weeks) from vaccination (Table 1). Thus, 78% (65; 92) of all vaccinated children were regarded as being protected by the vaccine. In the group vaccinated before 2011 that did not receive acyclovir prophylaxis, 86% were protected (73; 100). We found an association between acyclovir prophylaxis after vaccination and the risk of developing late chickenpox [hazard ratio (HR) 5.40 (1.43, 20.41), P = 0.01]. Furthermore, our data indicated that a vaccine-induced rash reduced the risk of late chickenpox [HR 0.08, (0.01, 0.66), P = 0.02; Table 1].
Twenty-five of all vaccinated children (56%) developed a vaccine-induced rash within a median time of 27 days (16–43), and of these, the majority (80%) developed fever at a median duration of 2 days (range: 1–5 days). Among the children that did not receive acyclovir prophylaxis, 65% (22/34) developed a rash (Table 2). Sixteen (64%) of all children with a rash had a small number of lesions, 5 (20%) had a moderate number of lesions and 4 (16%) had multiple lesions. Acyclovir prophylaxis was associated with a reduced risk of rash (P = 0.01; Table 2).
Four children experienced reactivation of their vaccine-induced rash at day 49–57 after vaccination of whom 2 cases were verified as vaccine type. Four children developed herpes zoster at day 64, 70, 78 and 442 from vaccination.
Children treated with dexamethasone within 15 days before and up to 40 days after vaccination tended to have a higher risk of rash (P = 0.01). We found no association between HDM 3 weeks before vaccination and a vaccine-induced rash (P = 0.05; Table 2). The vaccinated children with a rash had a median lymphocyte count of 0.96 × 109/L (range: 0.5–1.4 × 109/L) at the time of first vaccination and 1.13 × 109/L (range: 1.06–1.19 × 109/L) at the time of second vaccination. Three children had a lymphocyte count lower than 0.6 × 109/L, but were not excluded from the analysis (Table 3). A high lymphocyte count was associated with a reduced risk of vaccine-induced rash in the unadjusted analysis, but the association disappeared when adjusted for either dexamethasone (P = 0.10) or acyclovir (P = 0.11). No child had their maintenance therapy discontinued because of vaccination, but 6 children (13%) had an interruption of chemotherapy for other reasons around the time of vaccination, with a median duration of 6.5 days (range: 1–15 days). Fifteen children (33%) experienced discontinuation of maintenance therapy because of vaccine-induced rash for a median time of 4 days (range: 1–9 days).
More children with a rash were found to be seropositive after the first vaccination compared with children with no rash [14/16 (88%) vs. 6/22 (27%), P = 0.01; Table 3]. Additionally, more children not receiving acyclovir prophylaxis seroconverted after the first vaccination, compared with children receiving prophylaxis [8/28 (71%) vs. 0/10 (0%), P = 0.01]. Long-term seroconversion was found in 52% (17/33) of the children.
A total of 28 children were not vaccinated because of low lymphocyte count, poor general condition or their treating physician being unaware of the possibility of vaccination. Sixteen of these (57%) were exposed to VZV during maintenance therapy with a median of 2 exposures (range: 1–6), regardless of attempt to secure persistent isolation from exposed nonimmune healthy children. Despite the strict isolation regimen, 4 children developed clinically verified chickenpox during maintenance therapy and another 15 children developed chickenpox after completion of maintenance therapy. None of the children died of late chickenpox.
During a period of 15 years, 45 children with ALL have been vaccinated against VZV without concomitant interruption of chemotherapy. Despite most children being vaccinated early in maintenance therapy, rather than 3 months after cessation of maintenance therapy or a year from the time of remission as suggested in some guidelines,22,23 none of these children died because of VZV vaccination, and no child developed serious complications such as pneumonitis, visceral dissemination or secondary bacterial infection. Sixty-five percent of children not receiving acyclovir prophylaxis during vaccination developed a vaccine-induced rash, which is higher than previously reported in studies with suspension of maintenance therapy.8,24 The majority experienced only a mild rash and were all treated with acyclovir. There may be 2 reasons for the high number of vaccine-induced rashes in this cohort. Firstly, the children had a low lymphocyte count at the time of vaccination, because maintenance therapy was not discontinued, and our results indicate that a higher lymphocyte count might be protective against a vaccine-induced rash. We speculate that the low lymphocyte count gives the live attenuated vaccine a better opportunity to cause skin manifestations before being controlled by the immune system. Secondly, a high proportion of our cohort received dexamethasone around the time of vaccination, which was associated with vaccine-induced rash. Steroid treatment is a known risk factor for developing a vaccine-induced rash.25,26 Therefore, the timing of vaccination in relation to dexamethasone treatment is of great importance. In a recent study by Caniza et al,12 the concern surrounding VZV vaccine is centered on the death of a child related to the vaccine, as well as the interruption of chemotherapy that most ALL study groups advocate around the time of vaccination, which potentially increases the risk of recurrence of ALL. Although we did not completely avoid withholding chemotherapy, the length of time was considerably less. In previous studies, the VZV vaccine has been shown to be effective in children with ALL by protecting 85% of them against developing clinically verified chickenpox upon exposure and close to 100% protection against severe chickenpox.8 We found similar protection in our children not receiving acyclovir prophylaxis, with 86% experiencing long-term protection and 100% avoiding severe varicella. Those children who developed a vaccine-induced rash seemed to have better protection from the vaccine in terms of lower incidence of late chickenpox, which is in agreement with findings in other studies.9 It indicates that children developing clinical signs of chickenpox amount a better immune response. Until 2011, all children were isolated in their ward room if they needed admittance to the pediatric oncology department during the vaccination period because of the risk of exposing other immunosuppressed seronegative children. Acyclovir prophylaxis was introduced in 2011 to avoid this isolation. However, this prophylaxis seems to have resulted in significantly fewer children developing a vaccine-induced rash and hence being protected against late chickenpox. The negative effect of acyclovir prophylaxis on vaccine efficacy has not previously been reported.
Only 52% of all vaccinated children had long-term VZV seroconversion. However, only 9 children developed late chickenpox despite the high exposure risk to VZV. This indicates that cell-mediated immunity plays an important role regarding immunity in VZV, as described in other studies.27–29 One of the main concerns for parents of seronegative children with ALL in Denmark is the extensive isolation from an active social life, because the varicella vaccine is not part of the national vaccination program. It is impossible to avoid exposure to VZV completely, because there is a long incubation period and often the parents are not aware that their child has been exposed. In this cohort, 57% of the nonimmunized children were exposed to VZV, in spite of their isolation regimen, and 4 actually developed chickenpox during maintenance therapy. Accordingly, the majority of these nonimmunized children could have benefitted from VZV vaccination and lived a close to normal social life during maintenance therapy.
The weakness of our analysis is its relatively small size and retrospective design. However, it is still larger than previous reports of VZV vaccination without discontinuation of maintenance therapy.14–16
This analysis indicates that VZV vaccination without interruption of chemotherapy is feasible and justified in seronegative children with ALL in countries where VZV vaccination is not part of the national vaccination program. However, it is of great importance that the vaccination is administered by specialists and all adverse effects registered, because the vaccine is used off label.
Acyclovir should not be used as prophylaxis to avoid vaccine-induced rash, because it increases the risk of vaccine failure, although acyclovir treatment should be initiated as soon as the child shows signs of a rash.
The authors thank the staff of the Department of Paediatric Haematology and Oncology at Copenhagen University Hospital.
1. Feldman S, Hughes WT, Daniel CB. Varicella in children with cancer: seventy-seven cases. Pediatrics
2. Feldman S, Lott L. Varicella in children with cancer: impact of antiviral therapy and prophylaxis. Pediatrics
3. Shinjoh M, Takahashi T. Varicella zoster exposure on paediatric wards between 2000 and 2007: safe and effective post-exposure prophylaxis with oral aciclovir. J Hosp Infect. 2009;72(2):163–168.
4. Kumagai T, Kamada M, Igarashi C, et al. Varicella-zoster virus-specific cellular immunity in subjects given acyclovir after household chickenpox exposure. J Infect Dis. 1999;180:834–837.
5. Zaia JA, Levin MJ, Preblud SR, et al. Evaluation of varicella-zoster immune globulin: protection of immunosuppressed children after household exposure to varicella. J Infect Dis. 1983;147:737–743.
6. Rowland P, Wald ER, Mirro JR Jr., et al. Progressive varicella presenting with pain and minimal skin involvement in children with acute lymphoblastic leukemia
. J Clin Oncol. 1995;13:1697–1703.
7. Matsuzaki A, Suminoe A, Koga Y, et al. Fatal visceral varicella-zoster virus infection without skin involvement in a child with acute lymphoblastic leukemia
. Pediatr Hematol Oncol. 2008;25:237–242.
8. LaRussa P, Steinberg S, Gershon AA. Varicella vaccine for immunocompromised children: results of collaborative studies in the United States and Canada. J Infect Dis. 1996;174(suppl 3):S320–S323.
9. Gershon AA, Steinberg SP, Gelb L. Live attenuated varicella vaccine use in immunocompromised children and adults. Pediatrics
. 1986;78(4, pt 2):757–762.
10. Brunell PA, Shehab Z, Geiser C, et al. Administration of live varicella vaccine to children with leukaemia. Lancet. 1982;2:1069–1073.
11. Takahashi M, Kamiya H, Baba K, et al. Clinical experience with Oka live varicella vaccine in Japan. Postgrad Med J. 1985;61(suppl 4):61–67.
12. Caniza MA, Hunger SP, Schrauder A, et al; Members of International Study Group of Childhood ALL (“Ponte di Legno Working Group”). The controversy of varicella vaccination
in children with acute lymphoblastic leukemia
. Pediatr Blood Cancer. 2012;58:12–16.
13. Schrauder A, Henke-Gendo C, Seidemann K, et al. Varicella vaccination
in a child with acute lymphoblastic leukaemia. Lancet. 2007;369:1232.
14. Ha K, Baba K, Ikeda T, et al. Application of live varicella vaccine to children with acute leukemia or other malignancies without suspension of anticancer therapy. Pediatrics
15. Arbeter AM, Granowetter L, Starr SE, et al. Immunization of children with acute lymphoblastic leukemia
with live attenuated varicella vaccine without complete suspension of chemotherapy. Pediatrics
16. Sato Y, Miyano T, Kawauchi K, et al. Use of live varicella vaccine in children with acute leukemia and malignant lymphoma. Biken J. 1984;27:111–113.
17. Schmiegelow K, Forestier E, Hellebostad M, et al; Nordic Society of Paediatric Haematology and Oncology. Long-term results of NOPHO ALL-92 and ALL-2000 studies of childhood acute lymphoblastic leukemia
. Leukemia. 2010;24:345–354.
18. Frandsen TL, Heyman M, Abrahamsson J, et al. Complying with the European Clinical Trials directive while surviving the administrative pressure – an alternative approach to toxicity registration in a cancer trial. Eur J Cancer. 2014;50:251–259.
19. Toft N, Birgens H, Abrahamsson J, et al. Risk group assignment differs for children and adults 1-45 yr with acute lymphoblastic leukemia
treated by the NOPHO ALL-2008 protocol. Eur J Haematol. 2013;90:404–412.
20. Committee on Infectious Diseases. Prevention of varicella: recommendations for use of varicella vaccines in children, including a recommendation for a routine 2-dose varicella immunization schedule. Pediatrics
21. Prentice RL, Williams BJ, Peterson AV. On the regression analysis of multivariate failure time data. Biometrika. 1981;68:373–379.
24. Gershon AA, Steinberg SP. Persistence of immunity to varicella in children with leukemia immunized with live attenuated varicella vaccine. N Engl J Med. 1989;320:892–897.
25. Hill G, Chauvenet AR, Lovato J, et al. Recent steroid therapy increases severity of varicella infections in children with acute lymphoblastic leukemia
26. Lydick E, Kuter BJ, Zajac BA, et al. Association of steroid therapy with vaccine-associated rashes in children with acute lymphocytic leukaemia who received Oka/Merck varicella vaccine. NIAID Varicella Vaccine Collaborative Study Group. Vaccine. 1989;7:549–553.
27. Gershon AA, Steinberg SP. Cellular and humoral immune responses to varicella-zoster virus in immunocompromised patients during and after varicella-zoster infections. Infect Immun. 1979;25:170–174.
28. Arvin AM, Koropchak CM, Williams BR, et al. Early immune response in healthy and immunocompromised subjects with primary varicella-zoster virus infection. J Infect Dis. 1986;154:422–429.
29. Weinberg A, Levin MJ. VZV T cell-mediated immunity. Curr Top Microbiol Immunol. 2010;342:341–357.
Keywords:Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
varicella zoster virus; vaccination; pediatrics; acute lymphoblastic leukemia; childhood leukemia