Determining the dose-response relationship for pneumococcal conjugate vaccines: a nested analysis of the fractional dose PCV trial

BACKGROUND: The relationship between polysaccharide dose and immune response to pneumococcal conjugate vaccines (PCV) has never been established. An individually randomized controlled clinical trial was conducted in Kenyan infants to assess whether immune responses after fractional doses of PCV10 (Synflorix, GlaxoSmithKline plc.) or PCV13 (Prevnar13, Pfizer Inc.) were non-inferior to full-dose schedules (ClinicalTrials.org: NCT03489018; Pan African Clinical Trial Registry: PACTR202104717648755). We analysed these data to describe the serotype-specific dose-response relationships and evaluate factors associated with the immune response. METHODS: We analysed data from participants who received PCV doses at 6 and 14 weeks of age, with immunogenicity assessed at 18 weeks of age. Participants received 20 %, 40 % or full doses of PCV10 or PCV13. We used mixed-effects linear regression models to estimate the dose-response relationships between polysaccharide dose and log-transformed IgG concentrations and to determine factors associated with immune response. We estimated the minimum dose required to obtain a high (≥95 %) probability of an immunological response above the level considered to be associated with protection using logistic regression. RESULTS: 1342 infants were included. The polysaccharide dose, product, child's ethnicity, sex, maternal age, and interval between second PCV dose and immunogenicity sample were independently associated with IgG concentrations elicited by the primary schedule. For PCV13, the relationship between dose and log(IgG) was best described by a quadratic function for serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 19A and 23F, whereas a log-dose model was the best fit for serotypes 14, 18C and 19F. For PCV10, a linear relationship was the best fit for serotypes 1, 6B, 7F, 9V and 23F and log-dose model was the best fit for 4, 5, 14, 18C and 19F. CONCLUSION: Our findings suggest that available PCV10/13 could be reformulated to optimize protective responses.