Intussusception after administration of the rhesus tetravalent rotavirus vaccine (Rotashield): the search for a pathogenic mechanism.

OBJECTIVES

The rhesus tetravalent rotavirus vaccine (RRV) was withdrawn from the routine program for childhood immunization in the United States because of the rare and unexpected occurrence of intussusception in the 2-week period after administration of the first dose.

METHODS

To search for the pathogenesis of this association, we compared the pathology of surgical specimens from infants who had surgical reduction of their intussusceptions within 2 weeks of receiving the vaccine (case patients; n = 8) with the pathology of specimens from children who had surgery > 2 weeks after immunization (n = 6) or who had never been immunized (n = 26). Tissue was examined for evidence of the vaccine strain of rotavirus by reverse transcriptase-polymerase chain reaction (RT-PCR), in situ hybridization, and immunohistochemical staining.

RESULTS

RRV was identified by RT-PCR in tissue samples from 7 of the 8 case patients and in 2 of the 6 children who received the vaccine at a more distant time (29 and 58 days before surgery), but it was not identified in samples from any of the nonvaccinated children. No evidence of rotavirus tissue involvement was detected in any of the children by in situ hybridization or immunohistochemical staining. Pathologic evidence (for example, inclusion bodies, smudge cells) of adenovirus infection was present in 35% of the 37 specimens examined by routine staining and immunohistochemistry.

CONCLUSIONS

The fact that RRV was detected by RT-PCR but not by either of the other assays could be explained by RRV being present in the lumen of the gut but not in the tissues of appendix, ileum, or Peyer's patches. The Peyer's patches were not hyperplastic, and we could not establish the pathogenic mechanism for this association.