Histo-Blood Group Antigen Phenotype Determines Susceptibility to Genotype-Specific Rotavirus Infections and Impacts Measures of Rotavirus Vaccine Efficacy

Abstract

Background: Lewis and secretor histo-blood group antigens (HBGAs) have been associated with decreased susceptibility to P[8] genotype rotavirus (RV) infections. Efficacy of vaccines containing attenuated P[8] strains is decreased in low-income countries. Host phenotype might impact vaccine efficacy (VE) by altering susceptibility to vaccination or RV diarrhea (RVD). We performed a substudy in a monovalent RV vaccine (RV1) efficacy trial in Bangladesh to determine the impact of Lewis and secretor status on risk of RVD and VE.

Methods: In infants randomized to receive RV1 or no RV1 at 10 and 17 weeks with 1 year of complete active diarrheal surveillance, we performed Lewis and secretor phenotyping and genotyped the infecting strain of each episode of RVD.

Results: A vaccine containing P[8] RV protected secretors and nonsecretors similarly. However, unvaccinated nonsecretors had a reduced risk of RVD (relative risk, 0.53 [95% confidence interval, .36-.79]) mediated by complete protection from P[4] but not P[8] RVs. This effect reduced VE in nonsecretors to 31.7%, compared to 56.2% among secretors, and decreased VE for the overall cohort.

Conclusions: Host HBGA status may impact VE estimates by altering susceptibility to RV in unvaccinated children; future trials should therefore account for HBGA status.

Clinical trials registration: NCT01375647.

Figure 1.

Cumulative incidence of rotavirus diarrhea (RVD) in year 1 of life among unvaccinated infants according to secretor/Lewis phenotype. The distribution pattern of RVD incidence when comparing all groups together significantly differed according to phenotype. P value by Mantel–Cox log-rank test. Abbreviations: Le⁺, Lewis-positive; Le^–, Lewis-negative; RVD, rotavirus diarrhea; Se, secretor; se, nonsecretor.

Figure 2.

Cumulative incidence of rotavirus diarrhea (RVD) according to Lewis phenotype, secretor status, and vaccination status. Solid lines indicated unvaccinated infants; dashed lines indicate vaccinated infants. A, Lewis phenotype had no detectable effect modification on vaccine effect (P = .86), and was not associated with risk of RVD from week 18 to week 52 of life, irrespective of vaccination. B, Secretor status had a significant effect on RVD from week 18 to week 52 of life among unvaccinated infants but not among vaccinated infants. P values by Mantel–Cox log-rank test. Abbreviations: CI, confidence interval; HR, hazard ratio; Le⁺, Lewis-positive; Le^–, Lewis-negative; RVD, rotavirus diarrhea; Se, secretor; se, nonsecretor.