Herpes Zoster in Healthy Children Immunized With Varicella Vaccine

Herpes zoster rarely develops in healthy children who received varicella vaccine. The incidence of zoster among them was reported to be 14 cases per 100 000 person years.1 Furthermore zoster in children immunized with varicella vaccine has not always been caused by the varicella Oka vaccine strain.1 We present two cases of pediatric herpes zoster that developed 3 to 7 years after vaccination against varicella in which causative viruses were revealed by PCR analysis to be wild strains of varicella-zoster virus (VZV). Case reports. Case 1. A healthy 1-year-old girl received varicella vaccine (Lot BVZ0034, 2.0 × 104 plaque-forming units/dose; Biken, Osaka, Japan) on November 27, 1993. On February 25, 1997, at 4 years of age, she developed an exanthema on the left side of her neck which extended to the anterior part of her left ear on the next day. She consulted a dermatologist and was diagnosed as having herpes zoster on February 28. On March 3 her mouth was noted to be abnormally distorted when she sang a song. Because she complained of a strange sensation in her left ear, she visited our hospital. Redness and vesicles were seen in the skin distribution of the third trigeminal branch and the second and third cervical nerve regions. No vesicle was found in the left external auditory canal. The left corner of her mouth drooped and the left nasolabial sulcus disappeared (Ramsay-Hunt syndrome). Lagophthalmus, however, was not observed. She received acyclovir intravenously. On March 4 she could not easily chew her food and complained of slight taste abnormality. She was treated with prednisolone (15 mg/day). All vesicles in the zoster lesions formed crusts on March 10 (14th day of the disease). Her facial palsy gradually improved and she recovered without hearing disturbance. Attempted virus isolation from vesicles in the zoster lesion was unsuccessful, and crusts were collected from zoster lesion to extract VZV DNA. Case 2. A healthy 2-year-old boy received varicella vaccine (Lot BVZ018, 1.8 × 104 plaque-forming units/dose; Biken) on July 3, 1990. On December 4 he contracted mild varicella with 7 to 8 vesicles in total and without fever. On September 15, 1997 (at 9 years of age), redness and exanthema appeared in the right side of his back and right upper abdomen (T9 to 10 dermatomes). On September 22 he was diagnosed as having herpes zoster by a dermatologist. Because the boy complained of nausea and headache, he was admitted to our hospital and received acyclovir intravenously (750 mg/day). No meningeal sign was present. Zoster lesions began to crust on September 24 and healed uneventfully. Virus isolation was not performed, but crusts were collected to analyze VZV DNA. Virologic examinations. To know whether genomic DNAs extracted from crusts formed in the zoster lesions of the patients were identical with the DNA of the Oka vaccine strain, we determined whether the DNA contained the Pst I restriction site in gene 38 or not. The gene 38 to 43 region was amplified by PCR, and amplification products were analyzed by restriction fragment length polymorphism analysis described previously.2 The PCR products were precipitated in ethanol and digested with Pst I for 2 h at 37°C. The digested PCR products were separated by 1.4% agarose gel electrophoresis and was visualized by ethidium bromide staining. Results. Lack of the Pst I restriction site in gene 38 has been identified as a marker for grouping VZV isolates.2–4 The Oka vaccine strain lacks the Pst I site. The cleavage patterns of DNAs obtained from Cases 1 and 2 were identical (Fig. 1) and different from that of the varicella Oka vaccine strain. Both DNAs conserved the Pst I site. The results showed that the VZV strains causing zoster in both cases conformed to wild strains.Fig. 1: Restriction fragment length polymorphism analysis of the gene 38 to 43 region digested with PstI. Lane 1, DNA from Oka vaccine strain; Lanes 2 and 3, DNAs from crust of Cases 1 and 2, respectively. The digests were analyzed in a 1.4% agarose gel. The restriction enzyme site that separates the band indicated with a closed arrowhead into the two bands indicated with two open arrowheads was absent from the Oka vaccine strain marked with a minus sign at the bottom (Lane 1). DNAs from two patients contained the PstI site (Lanes 2 and 3).Discussion. Most young zoster patients contracted varicella in early infancy.5 Some children having zoster with no history of clinical varicella were thought to have been infected with VZV in utero.6 On the other hand in Japan about 20% of children who received varicella vaccine contracted natural varicella 2 to 3 years after immunization (breakthrough varicella).7 The cause of herpes zoster in Case 1 may be explained by the assumption that she had mild, unrecognized infection with a VZV wild strain in her first year of life or that she was inapparently infected with VZV wild strain after varicella vaccination, like the adult case that Hammershlag et al.8 reported. The latter seems to be a more likely explanation in Case 1 than the former, because her mother did not have varicella during pregnancy and her brothers did not contract varicella during her first year of life. The zoster lesion of Case 2 was thought to be produced by the VZV wild strain which caused mild breakthrough varicella after vaccination and remained silently in nerve ganglions thereafter. If analysis of viral DNA had not been performed, Case 1 would have been mistakenly accepted as a zoster patient affected by the varicella vaccine strain, because her parents denied exposure to varicella. Plotkin et al.9 reported two cases of herpes zoster in otherwise healthy children after vaccination against varicella. However, they did not perform analysis of VZV DNA; therefore the zoster in these children could have been caused by VZV wild strain. In follow-up studies of children immunized with varicella vaccine, viral DNA analysis should be performed before zoster lesions in vaccinees is attributed to the vaccine strain.