Varicella in the neonatal ICU due to the Varicella vaccine Oka strain

1 Infant case presentation

A 3000-gram infant born at 36-weeks gestation was admitted to the Neonatal Intensive Care Unit (NICU) on day of life (DOL) 0 for management of a prenatally diagnosed myelomeningocele and Chiari II malformation, and placement of a ventriculoperitoneal (VP) shunt for hydrocephalus on DOL 1. On DOL 24, the VP shunt was removed due to infection with Staphylococcus epidermidis. On DOL 36, the infant developed vesicular and pustular lesions on an erythematous base, beginning on the buttocks and perineum and spreading to the trunk, head, and extremities; the lesions were too numerous to count. Samples obtained from unroofed vesicles were positive for varicella zoster virus (VZV) by PCR (Solana HSV/VZV PCR Assay, Quidel, San Diego, CA). The infant received intravenous acyclovir 10 mg/kg/dose every 8 hours for seven days and remained afebrile without respiratory distress or neurologic findings. The lesions crusted over by DOL 42 and the infant was discharged home from the NICU on DOL 44.

A fluid sample from the lesions was sent to Merck & Co., Inc. to be sequenced using PCR-based amplification. On DOL 47, treating providers were notified that the VZV genotype was the vaccine strain, vOka [1]. A report to the Vaccine Adverse Event Reporting System (VAERS) was filed by the providers; no additional information was requested.

2 Maternal history

The infant’s mother was a 40-year-old, HIV-negative, primiparous woman without a history of varicella infection and non-immune to varicella as documented by prenatal serologic testing. Prior to discharge from the hospital on postpartum day 4, the mother received Varivax (Merck & Co., Inc., Kenilworth, NJ) on the post-partum unit as per standard practice. No cutaneous lesions were noted by the mother following vaccination. The mother visited and held her infant in the NICU, but did not directly breastfeed. When the infant was diagnosed with chickenpox, NICU providers learned of the recent vaccination after discussion with the mother and review of her medical record. Thus, the maternal infectious period was considered to be DOL 11–25 and the infant’s incubation period was considered to be DOL 18–46.

3 Infection prevention and control measures in the NICU

Numerous infection prevention and control (IP&C) measures were implemented for the index case, staff, potentially exposed patients and visitors (Table 1). Staff who had contact with the infected infant were identified; none had primary chickenpox or zoster during the month prior to the index case and all were documented immune to varicella. As per routine, the infant was in an open crib in the NICU, which potentially exposed 18 other infants in the same large open unit during the infectious period, defined as 24 hours prior to the onset of lesions until the infant’s transfer to an Airborne Infection Isolation Room. No new admissions were permitted into the exposed cohort. No secondary cases of VZV occurred in exposed neonates.

Due to the large number of lesions, an immunologic evaluation of the infant was performed at 2 months of age. Absolute lymphocyte count was 5960 cells/μl with normal numbers and percentages of T, B and NK cells including CD4⁺ and CD8⁺ T cell subpopulations. Immunodeficiency genetic testing by next-generation sequencing (INVITAE, San Francisco, CA, 407 gene panel) revealed 10 variants of uncertain significance; 9 were single alleles for autosomal recessive conditions and one was an autosomal dominant gene (TCF3), associated with agammaglobulinemia, but predicted to be a benign variant [2]. Research level genomics searching for novel immunodeficiency genes was not performed given the normal immune profile and uncomplicated varicella infection.

The infant was seen in the outpatient clinic for well-child visits. Citing their experience with the varicella vaccine, parents initially delayed routine 2-month vaccinations due to concern for adverse effects, but the infant was appropriately vaccinated by 6 months of age. On DOL 109, the infant underwent re-insertion of the VP shunt. Cerebrospinal fluid obtained during insertion was PCR-negative for VZV.

5 Discussion

To our knowledge, this is the third reported case of transmission of the varicella vaccine strain vOka from a mother to her infant and the first report that occurred in an infant hospitalized in a NICU. This case has implications for vOka transmission, the importance of varicella vaccination to prevent complications in pregnant women and their neonates, management of varicella in infected and exposed infants, and patient flow for the NICU and the obstetrical services.

Although rare due to vaccination and herd immunity, non-immune pregnant women and their fetuses/ neonates may experience substantial morbidity and mortality from wild-type varicella zoster virus (VZV) infection [3]. Pregnant women who develop chickenpox are at increased risk of varicella pneumonia [4]. Maternal chickenpox during the first and second trimester is associated with miscarriage, fetal demise, premature birth, and rarely the congenital varicella syndrome (CVS). CVS can affect the limbs (e.g., bone and muscle hypoplasia), skin (e.g., scarring), eyes (e.g., cataracts, chorioretinitis), and central nervous system (e.g., seizures and microcephaly); affected children often die during infancy [3]. Furthermore, maternal infection 5 days prior to delivery through 2 days after delivery can result in severe disseminated disease including pneumonia, hepatitis with associated coagulopathy, and encephalopathy; neonatal mortality has decreased to approximately 7% thanks to varicella zoster immune globulin (VZIG) post-exposure prophylaxis [3]. Seemingly unaffected infants born to mothers who have chickenpox during pregnancy may develop herpes zoster early in life due to latent VZV infection from maternal viremia and transplacental transmission [7].

The Center for Disease Control (CDC) and American College of Obstetrics & Gynecology (ACOG) recommend administration of live-attenuated vaccines in the post-partum period to non-immune patients [5]. The American Academy of Pediatrics (AAP) also endorses the benefits of maternal vaccination, including transmission of antibodies via breast milk [6]. Post-partum vaccination of non-immune adults is clearly desirable, as the introduction of the VZV vaccine in the general population has greatly reduced varicella.

Previous reports describing vOka transmission from vaccinated individuals are relatively rare, and support transmission of disease from skin lesions. Immunocompetent individuals infected with the vOka strain generally have mild symptoms, mostly limited to a rash typically consisting of 60 or fewer lesions [7]. We speculate that our patient’s mother may not have noticed cutaneous lesions or that transmission occurred from respiratory secretions, as known to occur with wild-type VZV. Two additional case reports describe maternal to infant transmission of the vOka strain. One documented transmission from a mother without identified lesions to her otherwise healthy full-term infant who developed 50–60 lesions; the mother was vaccinated on the infants DOL 3 and the infant developed lesions on DOL 25 [8]. A second report described vOka in full term infant on DOL 24 whose mother was vaccinated on infant’s DOL 1 and develop lesions 20 days later [9]. Consistent with these previous reports, the infant in the current case report developed lesions 32 days after maternal vaccination and also remained clinically well likely due to the decreased virulence of the attenuated vOka strain and no identifiable immunodeficiency. The risk of zoster following vOka vaccine, while substantially lower than the risk of zoster from wild type VZV, can occur [7].

However, due to their concern of their infant’s vOka infection, the parents initially deferred standard vaccinations despite physician counseling. This case highlights the importance of anticipatory guidance for this rare, but potential adverse effect of vaccination in the postpartum period. Mothers who receive varicella vaccination post-partum, particularly mothers of infants in the NICU, should be told that cutaneous lesions can develop after vOka vaccination and to inform their provider if this occurs so that mitigating measures can be taken, i.e., covering the lesions and/or limiting contact with their hospitalized infant. Healthcare workers in the NICU should also be informed when a parent receives a live viral vaccine.

The IP&C strategies implemented in response to this case impacted patient flow for both the NICU and obstetrical services. The infant was in an open crib in a section of the NICU consisting of 2-3 bed pods during the infectious window. As the mode of acquisition, strain of varicella (wild-type VZV versus vOka), and clinical impact of vOka in premature infants were unknown at the time of diagnosis, post-exposure prophylaxis (PEP) was provided to eligible infants (Table 1). Additionally, the decision to provide PEP was influenced by limited airborne isolation capacity in the NICU, exacerbated by the need to isolate infants born to SARS-CoV-2 infected mothers. The prolonged incubation period of 28 days following administration of PEP further restricted admissions to the NICU and required diversion of some high-risk obstetrical patients.

In summary, ensuring immunity to varicella prior to pregnancy and administering varicella vaccine post-partum to non-immune pregnant women are highly desirable to prevent morbidity and mortality from varicella in pregnant women and their fetuses/ neonates. This case report describes the IP&C strategies utilized in the NICU setting after identification of vaccine-strain associated varicella in a hospitalized infant. Anticipatory guidance for mothers vaccinated in the postpartum period and support of parents of an infected infant are recommended.

Funding

Irene Frantzis MBBS is supported by the training grant “Training in Pediatric Infectious Diseases” (National Institute of Allergy and Infectious Diseases T32AI007531, Principal Investigator Lisa Saiman).