Objective Routine varicella zoster vaccination for children aged 18 months began in Australia from November 2005. The aim of this study was to compare the current incidence and outcomes of congenital and neonatal varicella in Australia with similarly collected data from 1995 to 1997.
Methods Active national prospective surveillance was carried out for congenital and neonatal varicella using the Australian Paediatric Surveillance Unit (APSU) for 3.5 years from June 2006. Around 1300 clinicians reported monthly according to predefined case criteria.
Results During the study period the mean monthly return rate of APSU report cards was 93.7%. Two cases of congenital varicella (0.19 per 100 000 live births per annum) and 16 cases of neonatal varicella (2.0 per 100 000 live births per annum) were identified. During 2008 and 2009 no cases of congenital varicella were reported; neonatal varicella rates declined to 0.7 per 100 000 live births per annum, a significant trend (p=0.005) and a reduction of over 85% compared with rates during 1995–1997 (the prevaccination era) and the first year of the current surveillance study. Eleven of 16 neonatal cases followed prenatal maternal infection; seven of the 11 infections were acquired from children, four of whom were living in the same household. Ten (62.5%) infants with neonatal varicella were admitted to hospital, one of whom developed varicella pneumonitis requiring ventilatory support, but none died. Only one infecting contact had been vaccinated.
Conclusions There has been an apparent reduction of congenital varicella and a significant reduction of neonatal varicella in Australia following the introduction of universal varicella vaccination in 2005.
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.
Varicella-zoster virus (VZV) causes a highly contagious primary infection (chickenpox). Congenital infection, following intrauterine infection, may cause spontaneous abortion, premature delivery or stillbirth.1 Features of congenital varicella syndrome (CVS) include limb hypoplasia, cicatricial skin lesions in a dermatomal distribution, neurological abnormalities and structural damage to the eye.2 Neonatal varicella is caused either by intrauterine or early postnatal VZV infection. Perinatal maternal varicella infection manifesting within 5 days before and 2 days after delivery is associated with more severe or disseminated disease and a mortality of up to 20%.3 4
What is already known on this topic
▶ Universal varicella vaccination will decrease exposure of pregnant women to varicella infection from their own or other children.
▶ Congenital and neonatal varicella infection is associated with significant morbidity and mortality.
What this study adds
▶ Universal varicella vaccination is associated with reductions in the incidence of both congenital and neonatal varicella.
▶ With good management, including antivirals and zoster immune globulin, serious outcomes of neonatal varicella may be prevented.
In Australia, a significant proportion of women of childbearing age (8.9%, 95% CI 6.4% to 12.0%) are not immune to VZV.5 Women born overseas, especially in South Asia, may have lower rates of childhood exposure, and therefore, immunity.6 Infection within the first 20 weeks of gestation is associated with a risk for CVS of up to 2%, but CVS may also occur following infection later in pregnancy.1 7 Published data on CVS and neonatal varicella are limited. A surveillance study by the Australian Paediatric Surveillance Unit (APSU) from 1995 to 1997 identified six cases of congenital varicella, equivalent to 0.8 per 100 000 live births per year (95% CI 0.3 to 1.8 per 100 000) with an average of 2.3 cases per year. During this period, 44 cases of neonatal varicella were also identified by APSU, equivalent to 5.8 per 100 000 live births per year (95% CI 4.3 to 7.8 per 100 000).8
The VZV vaccine has been recommended for routine use in the USA since 1996 and Canada since 1999. VZV vaccine has been shown to prevent varicella infection in about 85% of immunised children, with 97% protection against moderate and severe disease.9 In Europe, only Greece and Germany have introduced universal varicella vaccination nationally while it is also used in parts of Italy (Sicily) and Spain (autonomous community of Madrid).10 VZV vaccination was publicly funded as part of the National Immunisation Programme (NIP) in Australia from November 2005 for all children aged 18 months and as a catch-up vaccine for children aged 10–13 years of age who have no history of either varicella infection or varicella vaccination.11
It is likely that the introduction of varicella vaccination in the NIP has decreased exposure of pregnant women to infection from their own or other children. However, break-through varicella infection occurs in a small proportion of children who receive one dose of vaccine, and a two-dose varicella schedule is likely to be required to decrease varicella infection in the population.12 The aim of this study was to compare the incidence and outcomes (eg, disease severity) of congenital and neonatal varicella in Australia between June 2006 and November 2009 with similarly collected data from 1995 to 1997, in the prevaccination era.
Active national surveillance for CVS and neonatal varicella from June 2006 until November 2009 was performed using APSU methods as previously described.8 13 Approximately 1300 paediatricians and other child health specialists were sent a monthly report card (82% by email) and asked to indicate whether or not they had seen an infant matching the case criteria for CVS or neonatal varicella, as part of the 23 conditions under surveillance during this time period. When a case was reported, clinicians were asked to complete a de-identified questionnaire requesting detailed information on demographics, diagnosis, history, treatment and short-term outcomes. Cases were reviewed by a group of expert investigators to determine whether or not they met the case definition.
The yearly incidence of CVS and neonatal varicella was calculated with 95% CI using a standard formula and χ2 for trend was performed on the neonatal varicella data using Epi Info 3.5.1. A p value of ≤0.05 was considered as significant.
CVS was defined as any termination, stillbirth, newborn infant or child up to the age of 2 years with definite or suspected CVS, with or without defects, and who met at least one of the following criteria:
A. Cicatricial skin lesions in a dermatomal distribution and/or pox-like skin scars and/or limb hypoplasia.
B. Development of herpes zoster in the first year of life.
Cases that did not come to a paediatrician's attention, for example those ending in therapeutic abortion, are not routinely reported and could not be included in this study.
Neonatal varicella was defined as any infant with neonatal varicella based on history, clinical and/or laboratory findings (eg, culture, PCR, IgM or immunofluorescence positive or seroconversion) in the first month of life but without features of CVS.
This study was approved by the Royal Alexandra Hospital for Children (the Children's Hospital at Westmead) Human Research Ethics Committee.
The mean monthly return rate of APSU report cards was 96% in 2006, 95% in 2007, 93% in 2008 and 91% in 2009 (average 93.7%).
Congenital varicella syndrome
There were 11 reports of CVS and 11 completed questionnaires were returned. Of these, five cases were duplicate reports and four reports did not meet the case criteria (one outside age range, two cases of cytomegalovirus reported by mistake as varicella, one intrauterine varicella infection that did not meet the case definition for CVS), leaving two cases of CVS, one in 2006 and one in 2007. These two cases followed antenatal maternal infection at 12 and 20 weeks (table 1). The incidence of CVS was 0.19 per 100 000 live births compared with 0.8 per 100 000 (six cases) in 1995–1997, a non-significant reduction (table 2 and figure 1). However, during 2008 and 2009 there were no cases.
There were 25 reports of possible neonatal varicella and 22 (91.7%) completed questionnaires were returned. Of the 22, four cases were duplicate reports and two cases did not meet the case criteria, leaving 16 confirmed cases of neonatal varicella. The estimated incidence of 2.05 per 100 000 live births per year was significantly lower (p=0.00002) during this study period compared with the incidence reported in 1995–1997 of 5.8 per 100 000 live births per year (table 2 and figure 1). Furthermore, during 2008 and 2009 there were only two cases per year of neonatal varicella reported, giving an overall rate of 0.7 per 100 000 live births per year, a reduction of over 85% compared with both prevaccination surveillance and the first year of the programme. Also, there was a significant downward trend in neonatal varicella (p=0.005) over this period of observation from 2006.
All 16 infants with neonatal varicella recovered without any reported sequelae (table 3). All had a history of exposure to varicella infection. In 11 (68.7%) perinatal maternal varicella infection was reported and in the other five there was postnatal exposure. Six neonates (37.5%) developed varicella following maternal exposure within the high-risk period (4 days before to 2 days after delivery). In seven of the 16 cases perinatal maternal varicella infection was acquired from children, four of whom were living in the same household as the infected mother. Only one of the child contacts was immunised against varicella. Maternal ethnicity/country of birth information was available for 10 cases of which seven were born in Australia and the rest were born overseas (one each in the USA, Northern Ireland and India). Ten infants (62.5%) required hospital admission (median stay 6 days). One developed varicella pneumonitis requiring intensive care admission for ventilatory support. There were no deaths.
This study is the first to document the impact of universal vaccination on rates of CVS and neonatal varicella infection. The varicella vaccine was licensed in Australia for children aged from 12 months in 1999 and was scheduled and funded on the NIP from November 2005 for all children aged 18 months. By 2008–2009 both CVS and neonatal varicella had decreased (by 100% and >85% respectively) compared with rates in both prevaccination surveillance and the first year of the current surveillance study (2005–2006; figure 1). The impact was most marked after 2006, which likely reflects increasing vaccination uptake. In 2006, 55.6% of infants aged 2 years had received varicella vaccination, increasing to 81.8% in 2009 (B Hull, personal communication). The reduction in incidence of both CVS and neonatal varicella probably reflects inclusion of varicella vaccination in the NIP, with increased herd immunity and reduced transmission, thereby indirectly protecting infants who themselves are not eligible for vaccination.
Countries considering the routine use of varicella vaccine need to consider both direct and indirect effects. A recent UK study estimated birth incidence rates of nearly 0.06 per 100 000 live births per year for congenital varicella and 0.16 for neonatal varicella, based on the incidence of varicella infection in nulliparous women in 2002–2003.14 However, the prevaccination study from Australia8 suggests that the estimated rates of CVS and neonatal varicella in the UK may be significantly underestimated.
Recent Australian studies of varicella-related hospitalisations also show an effect of universal vaccination. In Victoria there was a 7% (95% CI 5% to 9%) annual decline in hospitalisation rates from 2000 to 2007. This decline was most prominent among children aged under 5 years, in whom admission rates declined by 12% per year (95% CI 9% to 16%).15 Presentations of varicella to primary care in Australia have also decreased significantly since the introduction of vaccination, from 2.01 per 1000 encounters in 1998–1999 to 0.58 per 1000 in 2008–2009. In the first year of the universal vaccination programme there was only a small reduction in presentations compared with the previous year (1.31 in 2005 and 1.14 in 2006). However, there was a greater decline in subsequent years to 0.64 in 2007 and 0.58 in 2008, which is consistent with this study's findings of little impact in 2006 but a larger fall thereafter in both CVS and neonatal varicella rates.16 In comparison, in South Australia, where varicella is a notifiable disease, a reduction in varicella notifications has not been observed.17 This might be explained by differences in case definition (allowing notification of chickenpox or shingles simply as ‘varicella’) and also by the high uptake of varicella vaccine in South Australian children before it was included in the NIP (48% of children up to the age of 4 years were reported by parents as vaccinated for varicella by mid 2004).18
Congenital varicella infection is now very uncommon in Australia.8 Almost 10% of Australian women of childbearing age (15–44 years) are susceptible to varicella infection,5 which is usually acquired from their own or others' children as data suggest.19 It has been recommended that varicella vaccination should be considered in women of childbearing age without a history of varicella infection (or in whom screening indicates susceptibility) prior to pregnancy, in order to prevent CVS.8 This may be particularly important in women born in countries without routine vaccination against VZV.
The use of zoster immune globulin (ZIG) protects against congenital varicella when given shortly following exposure.1 It is primarily given to prevent severe maternal varicella. Mothers of the two infants with CVS received ZIG following development of clinical illness; however ZIG is most effective if given prior to the development of the illness. Both infants survived (table 1) whereas there was one termination of pregnancy at 21 weeks and one neonatal death at 10 days in the previous survey, 1995–1997.8
Neonatal varicella may be a serious illness associated with high mortality and morbidity partly because of relative immunologic immaturity.20 There were no deaths among the 16 cases reported and no known sequelae. Like the previous study, the neonatal varicella cases found in this study were not as severe as expected. This may reflect use of ZIG (37.5% (six) given prophylactically and 25% (four) given after the onset of clinical varicella infection) and/or antiviral treatment, one or both of which was given to over half of the infants. However, one third of the infants did not receive ZIG and of those who received it 40% (four) did not get it prophylactically as recommended. One neonate who received ZIG developed varicella pneumonitis, requiring ventilatory support. The use of ZIG is recommended to prevent varicella in exposed neonates in the USA, UK and Australia.21,–,23
The incidence figures given here are likely to be underestimates. Despite the authors' best efforts there were three (8.3%, two in 2007 and one in 2009) notified cases of possible neonatal varicella for which completed questionnaires were not returned and without detailed data these could not be included in analyses. However, the number of potential but unconfirmed cases (due to missing clinical information) is very similar to the previous APSU study (6.7% in 1995–1997) and this study used the same surveillance methodology. The mean return rate of APSU report cards during the study period was very high (93.7%). The incidence estimates are therefore likely to be comparable between the time periods and not subject to systematic bias.
Despite no reports of CVS in the last 2 years of the surveillance, the reduction in CVS rate is not statistically significant compared with the 1990s prevaccination study. This is likely to be due to small numbers.
This study reports reductions in the annual incidence of congenital and neonatal varicella in Australia following the introduction of the varicella vaccine to the NIP in 2005. Countries considering the routine use of varicella vaccine need to consider both direct and indirect effects (including clinical and economic benefits beyond the neonatal period) of universal varicella vaccination. Ongoing surveillance is necessary to understand the disease burden and changing pattern of both varicella and zoster infections following the introduction of universal vaccination.
The activities of the APSU are supported by the Australian Government Department of Health and Ageing; NHMRC Enabling grant no. 402784; NHMRC Practitioner Fellowship no. 457084 (EJE); the Faculty of Medicine, University of Sydney; and the Royal Australasian College of Physicians. The authors would also like to thank all the participating clinicians who reported the cases described in this paper.
Funding The activities of the APSU are supported by the Australian Government Department of Health and Ageing; NHMRC Enabling grant no. 402784; NHMRC Practitioner Fellowship no. 457084 (EJE); the Faculty of Medicine, University of Sydney; and the Royal Australasian College of Physicians.
Competing interests HM has participated as a board member of Global Advisory Boards for Merck and GlaxoSmithKline and has participated as an Investigator in industry-sponsored clinical vaccine trials. JB has served on advisory and/or data safety monitoring committees for CSL Vaccines and GSK for which the Murdoch Childrens Research Institute (MCRI) receives payments into an educational fund. RB has received financial support by pharma, including CSL, Sanofi, GSK, Novartis, Roche and Wyeth to conduct research and present at scientific meetings; if fees were offered, these were placed in a university research account. The other authors declare that they have no conflict of interest in relation to this work.
Ethics approval This study was conducted with the approval of the Royal Alexandra Hospital for Children (The Children's Hospital at Westmead) Human Research Ethics Committee, Westmead, New South Wales, Australia.
Provenance and peer review Not commissioned; externally peer reviewed.