Background: Vaccine-strain herpes zoster (HZ) can occur after varicella vaccination. This study determined the number and proportion of HZ cases caused by vaccine-strain varicella zoster virus (VZV), assessed the positive predictive value of provider diagnosis of HZ, and computed HZ incidence rates in vaccinated and unvaccinated children.
Methods: We used electronic medical records to identify all office visits with an HZ diagnosis for children aged <18 years in a managed care plan. Providers collected skin specimens and completed a questionnaire. Specimens were tested by polymerase chain reaction to identify wild-type or vaccine-strain VZV.
Results: From May 2005 to September 2009, we enrolled 322 subjects. VZV was detected in 82% of specimens (84% wild-type, 15% vaccine-strain, 1% possible vaccine-wild-type recombinant). Among the 118 vaccinated subjects, VZV was detected in 70% (52% wild-type). The positive predictive value for provider diagnosis of "definite HZ" was 93% for unvaccinated and 79% for vaccinated children. The incidence of laboratory-confirmed HZ was 48 per 100,000 person-years in vaccinated children (both wild-type and vaccine-strain) and 230 per 100,000 person-years in unvaccinated children (wild-type only).
Conclusions: HZ incidence in vaccinated children was 79% lower than in unvaccinated children. Among vaccinated children, half of HZ cases were due to wild-type VZV.
Keywords: children; epidemiology; herpes zoster; incidence; shingles; varicella vaccine; varicella zoster virus.