Varicella-zoster virus (chickenpox) infection in pregnancy
Dr RF Lamont and Dr R Romero, Perinatology Research Branch, NICHD/NIH/DHHS, Wayne State University/Hutzel Women’s Hospital, 3990 John R, Box 4, Detroit, MI 48201, USA. Emails email@example.com, firstname.lastname@example.org
Please cite this paper as: Lamont R, Sobel J, Carrington D, Mazaki-Tovi S, Kusanovic J, Vaisbuch E, Romero R. Varicella-zoster virus (chickenpox) infection in pregnancy. BJOG 2011; DOI: 10.1111/j.1471-0528.2011.02983.x.
Congenital varicella syndrome, maternal varicella-zoster virus pneumonia and neonatal varicella infection are associated with serious fetomaternal morbidity and, not infrequently, mortality. Vaccination against varicella-zoster virus can prevent the disease, and outbreak control limits the exposure of pregnant women to the infectious agent. Maternal varicella-zoster immunoglobulin administration before rash development, with or without antiviral medication, can modify the progression of the disease.
Varicella-zoster virus (VZV) is a highly contagious infectious agent and chickenpox is a common childhood illness. Accordingly, contact between a pregnant woman and a contagious individual is not uncommon. In temperate climes, 90% of women of childbearing age will be immune to the disease, but this is not the case among migrant women from tropical climes. Although the complications of chickenpox are rare, the potential for significant fetomaternal morbidity and even mortality cannot be ignored. This being the case, obstetricians should be aware of the potential for serious adverse sequelae, the steps needed to implement a programme for the management of exposure incidents and the appropriate use of vaccination prophylaxis and intervention with varicella-zoster immunoglobulin (VZIG) and/or antiviral therapy.
The incidence of chickenpox varies between temperate and tropical climes. In temperate climes, the disease occurs most commonly in late winter and early spring. Prior to the introduction of childhood vaccination, by the age of 15, 90% of individuals in temperate climes would have had a primary infection,1,2 and hence be seropositive, compared with only 25–80% of individuals in tropical countries.3,4 The incidence of chickenpox is not known precisely as it is not a reportable disease. Best estimates suggest an incidence of 2–3 per 1000 in the UK5 and between 1.6 and 4.6 per 1000 in the USA among 15–44-year-old individuals during the 1990s.6 In both of these countries, the incidence appears to be increasing,7–9 which may be a result of increasing immigration8,10 of susceptible individuals. If this is the case, the rate would be expected to decrease with the uptake of vaccination programmes.11
VZV is a DNA virus of the herpes family and is highly contagious. Humans are the only source and the virus enters the host through the conjunctivae and mucous membranes of the nasopharynx.12 At the end of the second viraemic phase, nonspecific prodromal symptoms, such as headache, fever and malaise, occur. This is followed by pruritis and a maculopapular rash, which becomes vesicular before crusting usually about 5 days later. The sufferer is contagious from 2 days prior to the development of the rash until crusting of the vesicles occurs. Primary infection generally provides lifelong immunity, but symptomatic re-infections have been reported. 13.3% of individuals with varicella have reported a previous episode of chickenpox,13,14 and subclinical re-infections have been detected serologically.15 The reason for re-infection may be a failure to develop, maintain or reactivate immune memory cells at the time of infection, or a high inoculum from close contacts which overwhelms the immune system.16
Chickenpox in pregnancy
Chickenpox is not associated with first-trimester spontaneous abortion.17–20 Before 24 weeks of gestation, vertical transmission to the fetus has been detected clinically/serologically and by polymerase chain reaction (PCR) in approximately 24 and 8% of cases of virologically confirmed maternal chickenpox, respectively. Intrauterine growth restriction occurs in approximately 23% of cases21 and low birthweight is virtually universal.22 In a case–control study, nonexposed controls had a spontaneous preterm birth rate of 5.6%, compared with 14.3% for women with chickenpox in pregnancy (P = 0.05).18 The highest rate of mortality and morbidity associated with chicken pox in pregnancy occurs in the presence of congenital varicella syndrome (CVS), maternal varicella pneumonia and neonatal varicella.
Congenital varicella syndrome
CVS was first described in 1947; since then, at least 130 cases have been reported in the literature,20 most of which have been reported in the last 15 years. CVS is associated with a mortality rate of 30% in the first few months of life and a 15% risk of developing herpes zoster (HZ) between the second and 41st month of life. However, in spite of a poor initial prognosis, a good long-term outcome for survivors can occur.23
Incidence of CVS according to gestational age of acquisition of chickenpox
Primary VZV infection in the first two trimesters of pregnancy results in intrauterine infection in up to 25% of cases,20 and congenital anomalies described in CVS can be expected in approximately 12% of infected fetuses.24 Maternal chickenpox in the first 20 weeks of pregnancy is associated with an incidence of CVS of 0.91% (13 cases of CVS in 1423 live births).25 Although cohort studies and case reports have recorded CVS following maternal chickenpox between 20 and 28 weeks of gestation, no cases of CVS have been reported following maternal chickenpox after 28 weeks of gestation.25 It has been calculated that the annual number of cases of CVS in the USA, UK, Germany and Canada are 41, seven, seven and four, respectively.6
Clinical features of congenital varicella infection
The clinical features of CVS are multi-system, but some tissues and organs are selectively damaged.22 Skin lesions occur in approximately 70% of cases and limb hypoplasia in 46–72%.6,21 Neurological abnormalities, such as microcephaly, cortical atrophy, hydrocephaly and mental retardation, occur in 48–62% of cases. Eye disorders, such as microphthalmia, chorioretinitis and cataracts, occur in 44–52% of cases.26–28 Muscle hypoplasia, developmental delay and abnormalities of the gastrointestinal and genitourinary tracts and the cardiovascular system occur in 7–24% of cases.6,13,18 Survivors may have long-term learning difficulties and developmental problems; however, case–control studies do not suggest long-term neurodevelopmental disorders in asymptomatic children.29
Mechanism of CVS
The mechanism of CVS is thought to be the reactivation of VZV in utero,30 akin to the mechanism of HZ development, rather than primary infection. The short period of latency between primary infection and reactivation may be a result of immature fetal cell-mediated immunity.31 The evidence for reactivation stems from the dermatomal pattern of skin lesions, similar to HZ,22 the segmental maldevelopment of the musculoskeletal system and the segmental dysfunction of the somatic and autonomic nervous systems.30
Diagnosis of CVS
The diagnosis of CVS can be confirmed by a record of maternal chickenpox in pregnancy, together with congenital skin lesions in a dermatomal distribution, with or without the presence of neurological signs, eye defects, limb deformities and neonatal seizures.32 Retrospective evidence of maternal VZV immunoglobulin G (IgG) seroconversion during pregnancy is also helpful. Proof of intrauterine infections, irrespective of whether or not CVS develops, can be deduced from the detection of VZV DNA in the fetus or neonate, the presence of specific IgM in fetal or cord blood, the persistence of specific IgG beyond 7 months of age and the development of HZ during infancy.30,33 Nearly 20% of infants with intrauterine acquisition of VZV infection develop neonatal or infantile HZ, which is usually uncomplicated.21 Although fetal serological detection of specific IgM is useful in confirming evidence of intrauterine infection, in general, serology has a low sensitivity and is therefore unreliable, and is not recommended for the diagnosis of CVS caused by maternal chickenpox.34 Similarly, viral isolation is often unsuccessful and VZV DNA detection by PCR is much more reliable for the demonstration of fetal infection.35 As VZV and herpes simplex virus (HSV) share common surface antigens, there is serological cross-reactivity,13 but no cross-protection has been demonstrated. Accordingly, any rise in heterologous antibody titres may be a result of cross-reactivity, but may also signify simultaneous infection. Coxsackie B and HSV-2 can cause similar congenital lesions,36–38 and one case has been reported in which there was cutaneous scarring and limb hypoplasia, but serology and PCR revealed HSV-2 infection rather than VZV.36 Cases which present with rare abnormalities and subclinical maternal infection may require confirmation with molecular virological methods to establish a causal relationship between maternal infection and congenital abnormalities.39
Prenatal diagnosis of CVS
PCR of amniotic fluid for VZV DNA is now the method of choice for the determination of whether or not the fetus has been infected. Prenatal diagnosis relies on the identification of a combination of signs on a detailed ultrasound examination, such as limb deformities, microcephaly, hydrocephaly, polyhydramnios, soft tissue calcification and intrauterine growth restriction.6,40–42 As the clinical features are a result of reactivation of VZV, time is needed for this damage to manifest itself. Accordingly, no less than 5 weeks from the time of maternal rash should be permitted before the first detailed scan is performed, as scans at 4 weeks have resulted in missed diagnoses.43
Can maternal HZ cause CVS?
As VZV remains latent in the sensory root ganglia, and the enervation for the uterus arises from T10 to L4, theoretically, intrauterine shedding of virus from HZ is a possibility. However, in 301 cases of HZ in the first and second trimesters, no cases of CVS have been reported,6,15,18,34,44 although there has been one case report of a child born with typical findings of CVS from a mother who had disseminated HZ at 12 weeks of gestation.30 There have been no reports of clinical or serological evidence of VZV in infants whose mothers developed perinatal HZ,45 and therefore VZIG is not indicated.
Maternal varicella pneumonia
The average incidence of chickenpox in pregnancy has been calculated to be 0.7–3 per 1000 pregnancies.46,47 Maternal pneumonia complicates about 10–20% of cases of chickenpox in pregnancy, resulting in a higher mortality/morbidity than in nonpregnant adults.48–50 Pregnant women with VZV pneumonia should be hospitalised for monitoring and to initiate antiviral therapy, because up to 40% of women may need mechanical ventilation.51 Mortality in severe cases (those who require mechanical ventilation) in the pre-antiviral era was 20–45%,52–54 and is currently estimated to be 3–14%.54–57 Between 1985 and 2002, in the confidential enquiries into maternal deaths in the UK, there were nine indirect and one late maternal death associated with maternal VZV pneumonia.58
The risk for pneumonia also increases with increasing gestational age. Although this has been associated with relative maternal immunosuppression,57 it still remains unproven and may be purely mechanical, with increasing splinting of the diaphragm as the gravid uterus occupies more space.59
In the era before neonatal intensive care, VZIG and antivirals, the mortality rate for neonatal chickenpox was 31%,51,54,60–62 and is still 7% in the modern era.45 Maternal chickenpox in late pregnancy may result in neonatal chickenpox before passive immunity from mother to baby can be conferred,54 and the cell-mediated immune response of the neonate is unlikely to be sufficient to prevent the haematogenous spread of VZV.63 Neonatal chickenpox may occur by transplacental transmission, ascending infection or via the neonatal respiratory tract. The incubation period of intrauterine transmitted VZV from the beginning of maternal rash to the outset of neonatal rash varies. Accordingly, neonatal chickenpox in the first 10–12 days of life is caused by intrauterine transmission, whereas, after this time, it is by postnatal infection. When maternal chickenpox occurs 1–4 weeks before delivery, up to 50% of neonates will be infected,45 and 23% of these will develop clinical chickenpox despite high titres of passively acquired antibodies. The mortality from neonatal chickenpox is low.64 The exception is babies born less than 28 weeks of gestation or less than 1000 g, who are at increased risk of severe chickenpox,2 because they have less protection from maternally transmitted antibodies.65,66 Passively acquired antibodies can be detected in all babies whose mothers developed VZV rash more than 7 days before delivery.
Prevention of chickenpox in pregnancy
In women who reach childbearing age without natural immunity or vaccination as part of a childhood immunisation programme, chickenpox in pregnancy can be avoided by vaccination. Although two vaccines are licensed for use in the UK,58 they are not included in the standard childhood immunisation programmes nor routinely recommended for nonimmune adult women apart from healthcare workers. VZV vaccine has been shown to be effective in preventing infection following exposure, and is most effective when given within 3 days of exposure.67,68
The Varivax vaccine is a live attenuated vaccine; therefore, some advise avoidance of pregnancy for 1 month69–71 or 3 months58 after vaccination, although no birth defects related to inadvertent vaccine exposure have been reported.49,72 There has been one case report of a VZV-susceptible pregnant woman who, following vaccination of her 1-year-old child, developed chickenpox. Transmission was confirmed using PCR. A therapeutic termination of pregnancy was performed, but no virus was isolated from fetal tissue.73 Vaccinees who develop chickenpox less than 42 days after vaccination are likely to represent wild virus infection,74 but the disease is mild, infectivity is low and there is little or no risk of complications.75 Breastfeeding is safe following postnatal vaccination and, after VZV vaccination, breast milk samples have failed to show any VZV DNA.76,77
Management of exposure incidents
An essential part of the prevention strategy to avoid or reduce the incidence of chickenpox in pregnancy and the cost of management of an outbreak requires an organised approach to the management of exposure incidents. Screening should be carried out pre-pregnancy if there is an opportunity to do so (e.g. family planning/infertility clinics).30 Screening should also be carried out in early pregnancy so that those who are uncertain78 can be tested, and those who are susceptible can be counselled about the risks, instructed on the procedure should contact occur and co-opted into a protocol for the management of exposure incidents. All healthcare workers who deal with pregnant women should be screened and vaccinated, or identified as susceptible, to permit redeployment to nonpatient areas.79 An evaluation of the economic and clinical outcomes of a programme of routine antenatal screening and postpartum vaccination of seronegative women found that a selective serotesting strategy prevented nearly one-half of the VZV cases in their cohort. This is particularly relevant to those areas with high immigrant populations from tropical climes, where immunity is much less likely and immune status is much less likely to be known.10 However, this evaluation was based on an analytical cost-effectiveness model following a hypothetical cohort of over four million women over a 20-year period.
Management of chickenpox in pregnancy
Antivirals for use in VZV infections
Acyclovir is a synthetic nucleoside analogue of guanine which is highly specific for cells infected by VZV or HSV. When phosphorylated by viral thymidine kinase in cells infected with VZV, there is inhibition of viral DNA polymerase which stops the replication of human herpes viruses. Oral acyclovir has low bioavailability and must be given in frequent doses to achieve therapeutic levels.49 Further bioavailability data suggest that the physiological changes of pregnancy do not alter the maternal pharmacokinetics from those of nonpregnant women.80,81 Valacyclovir and famciclovir are prodrugs of acyclovir and penciclovir, respectively. As prodrugs, they have a longer half-life and better oral absorption and bioavailability; therefore, because of their less frequent administration, they are a better choice for oral therapy with improved compliance.81,82
Antiviral therapy, either alone or in combination with VZIG, has been recommended in the management of chickenpox in pregnancy.83,84 Antiviral prophylaxis is best given on the seventh day after exposure.85 All pregnant women with established chickenpox should receive oral acyclovir, 800 mg five times daily, or valacyclovir, 1 g three times daily, both for 7 days.49 Compared with placebo, this reduces the duration of fever and symptoms of chickenpox in immunocompetent adults if commenced within 24 hours of rash development.86 If given within 24 hours and up to 72 hours of the development of rash, acyclovir is effective in reducing the fetomaternal mortality and morbidity associated with VZV infection,69 particularly if used intravenously.59,87,88 Intravenous acyclovir in severe pregnancy complications, such as pneumonia, is preferred to oral treatment because of bioavailability, especially in the second half of pregnancy. The dose is usually 10–15 mg/kg body weight intravenously every 8 hours for 5–10 days for VZV pneumonia, and should be started within 24–72 hours of rash development. There is no evidence of fetal benefits with respect to CVS or chickenpox, but acyclovir crosses the placenta and can be found in amniotic fluid, umbilical cord blood and other fetal tissues,22 although it does not appear to accumulate in the fetus.80 Acyclovir may inhibit viral replication during maternal viraemia, which may inhibit the transplacental transmission of VZV.89,90
Neonates showing signs of chickenpox and those with chickenpox showing evidence of neurological or ophthalmic complications have been reported to benefit from the use of acyclovir intravenously.91–93
Registries of neonates exposed to acyclovir in utero have found no significant risk of teratogenesis from the use of acyclovir in pregnancy, but theoretical risks exist with use in the first trimester.58,94,95 Although there is a potential for complications of in utero exposure,96 small studies of valacyclovir use in late pregnancy have found no clinical or laboratory evidence of toxicity in infants followed up to 1 month97 or 6 months98 of age.
Susceptible pregnant women with significant VZV exposure should be offered VZIG to prevent or attenuate maternal disease.99,100 Significant VZV exposure is defined differently in different guidelines, but reflects the proximity and duration of contact and the potential for droplet and vesicular fluid contact with the conjunctivae and nasopharyngeal mucous membranes.49,58 A history of chickenpox negates the need for serological testing. With no history of chickenpox, serology should be checked if time permits; otherwise, VZIG should be given.58 The main indication for VZIG is to modify disease and prevent maternal morbidity.90,101
VZIG should be given to susceptible women within 72 hours, but can be given up to 96 hours after exposure to the virus.58,90 Beyond 96 hours, VZIG has not been evaluated,69 but some recommend VZIG for up to 10 days after exposure.58,102,103 This may be because a more concentrated Ig formation is available in some countries.104 VZIG is ineffective and should not be given once clinical illness has developed.59,105
It is not known whether VZIG prevents viraemia or CVS, but this is unlikely to be tested bearing in mind the numbers required to test the hypothesis and the ethics of randomisation of care. The Royal College of Obstetricians and Gynaecologists’ guidelines58 point out that VZIG is derived from non-UK donors with high VZV antibody titres, but that no cases of blood-borne infections have been reported. As VZIG is in scarce supply and is expensive, treatment should be optimised rather than liberal, and availability should be checked before a patient is offered the choice.
The optimal dose of VZIG is unclear and the calculation of unit dosage differs internationally; however, in the USA, VZIG is recommended in a dose of 125 units/10 kg to a maximum of 625 units69 (equivalent to a 50-kg women receiving 125 units/10 kg). Alternatively, 1 mg/kg body weight can be administered intravenously.106 Whether 625 units is sufficient for women weighing more than 50 kg is not clear.102 VZIG may also prolong incubation and this should be considered when arranging surveillance, monitoring, isolation and follow-up, where many suggest adding a week to standard operating procedures relative to those who do not receive VZIG. Intravenous administration appears to demonstrate benefit over intramuscular administration with more rapid achievement of optimal serum levels.107 The duration of action of VZIG is unknown, but is likely to be at least one half-life of IgG (3 weeks). Accordingly, subsequent exposure within 3 weeks after a dose of VZIG may require additional doses.69
Management of perinatal infections
Primary maternal infection with VZV around the time of delivery poses important problems.108 Following maternal chickenpox around term, elective delivery may be delayed by 5–7 days to facilitate passive immunity of the neonate, but experience with this practice is limited.109,110 Theoretically, epidural rather than spinal anaesthesia may be safer because the dura mater is not penetrated and a site which is free of cutaneous lesions should be chosen for needle placement.111 A neonatal ophthalmic examination should be performed together with serological testing of the neonate for IgM at birth and IgG at 7 months of age.
VZIG is recommended for neonates whose mothers develop VZV rash from 5 days before delivery up to 2 days after delivery.91 Neonates born before or after this time probably do not need passive immunisation because they are not at risk of severe neonatal chickenpox.24,112 Although VZIG may not prevent infection, it may reduce the severity of neonatal infection,45 but it is of no benefit once signs of chickenpox become evident.113,114 Monitoring of the neonate should be prolonged to 28 days because VZIG may prolong the incubation period. VZIG is also recommended for the nonimmune neonate who is exposed to VZV or HZ from an index subject other than the mother in the first 7 days of life.
If signs of neonatal infection develop despite VZIG, the neonate should be treated with acyclovir, and there are anecdotal reports of the benefit of a combination of VZIG and acyclovir in maternal VZV exposure near term or in exposed neonates to prevent neonatal varicella.83,84
Maternal HZ peripartum does not require any action because the neonate will have passive immunity. This does not apply to babies born before 28 weeks or those less than 1000 g birth weight because they may not have developed passive immunity.115
Chickenpox is a common childhood illness but, if it develops in pregnancy, is associated with serious adverse sequelae, such as CVS, maternal VZV pneumonia and neonatal varicella infection, which may lead to fetomaternal morbidity and mortality. Vaccination against VZV is available, but is not currently included in standard childhood immunisation programmes nor routinely recommended for nonimmune adult women in the UK. Prevention strategies should also include plans for the management of exposure incidents. When chickenpox occurs in pregnancy, antiviral therapy, either alone or in combination with VZIG, has been recommended for management. The use of antivirals decreases the risk of mortality and morbidity from chickenpox, but these will still occur. VZIG reduces the incidence and severity of chickenpox, but does not eliminate it completely, and is of no benefit once the signs of chickenpox become evident. The scenario of a pregnant woman with a history of contact with an index subject with chickenpox, either arriving at a hospital public area, or telephoning for advice, merits each obstetric unit having a written protocol to reduce unnecessary costs and, at the same time, offering the best available protection for those most susceptible to adverse sequelae.
Disclosure of interest
Contribution to authorship
All authors contributed to this work.
Details of ethics approval
This research was supported in part by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Department of Health and Human Services.