Geographic variation in pneumococcal vaccine efficacy estimated from dynamic modeling of epidemiological data post-PCV7

Abstract

Although mean efficacy of multivalent pneumococcus vaccines has been intensively studied, variance in vaccine efficacy (VE) has been overlooked. Different net individual protection across settings can be driven by environmental conditions, local serotype and clonal composition, as well as by socio-demographic and genetic host factors. Understanding efficacy variation has implications for population-level effectiveness and other eco-evolutionary feedbacks. Here I show that realized VE can vary across epidemiological settings, by applying a multi-site-one-model approach to data post-vaccination. I analyse serotype prevalence dynamics following PCV7, in asymptomatic carriage in children attending day care in Portugal, Norway, France, Greece, Hungary and Hong-Kong. Model fitting to each dataset provides site-specific estimates for vaccine efficacy against acquisition, and pneumococcal transmission parameters. According to this model, variable serotype replacement across sites can be explained through variable PCV7 efficacy, ranging from 40% in Norway to 10% in Hong-Kong. While the details of how this effect is achieved remain to be determined, here I report three factors negatively associated with the VE readout, including initial prevalence of serotype 19F, daily mean temperature, and the Gini index. The study warrants more attention on local modulators of vaccine performance and calls for predictive frameworks within and across populations.

Figure 1

Model flow diagram in the absence of vaccination. S, susceptible, I _V, single carriers of any vaccine serotype, I _N, single carriers of any non-vaccine serotyoe, I _VV double carriers of any two vaccine serotypes, I _NN, double carriers of any two non-vaccine serotypes, I _VN, double carriers of a vaccine and non-vaccine serotype. With vaccination, the structure of the model is doubled to account for vaccination status of each epidemiological class. The parameters are defined in the text. See Methods 1–3 for details, and Supplementary Material S1 for full ODE equation formulation.

Figure 2

Transmission intensity vs. overall pathogen prevalence in a SIS framework. The mean R ₀’s estimated from our datasets are matched with the expected overall prevalence of pneumococcal carriage in the pre-vaccine era. The bars reflect the 95% confidence interval for binomial proportions over 300 simulations with same R ₀, taking into account the mean sample size in each setting (Table 1). The line depicts the theoretical nonlinear relationship expected by the model (Eq. (1)). The shaded regions represent the 95% quantile of simulations under different sample sizes: n = 100 (light blue), n = 250 (darker blue) and n = 500 (darkest shade).

Figure 3

Posterior distributions for heptavalent vaccine efficacy against PCV7 serotypes, estimated in a setting-specific manner. Parameter inference is based on dynamic model fitting to cross-sectional data in day care settings from different countries, accounting for different transmission intensities, initial serotype distribution, and vaccine coverage. Red lines indicate normal distribution fits. Numerical results are detailed in Table 3.

Figure 4

Observed and fitted prevalence of asymptomatic carriage of pneumococcus in day care centres by country. The model-predicted 95% credible envelopes for temporal trajectories of VT prevalence (left panels) and overall prevalence (right panels) after vaccination are superimposed on observed data in different day care settings at specific time-points. (a,b) Portugal. (c,d) Norway. (e,f) France. (g,h) Greece. (i,j) Hungary. (k,l) Hong Kong. The credible envelopes were computed using 500 simulations with different parameter combinations sampled from the joint posterior, and here account only for the uncertainty around parameter values and initial conditions. For the credible envelopes accounting for sampling uncertainty throughout time (with mean sample size as in Table 1) see Fig. S6.

Figure 5

Possible factors for variation in vaccine efficacy against acquisition. (a) Prevalence of serotype 19F pre-PCV7 and model-based vaccine efficacy estimates. The decreasing trend suggests that high initial frequency of a serotype for which vaccine protection may be inferior (e.g. 19F) could influence model estimates of the net realized vaccine efficacy against PCV7 serotypes when aggregated together, (R ² = 47%, (R ²-adj = 11%, GLM fit: ρ = −0.0003, p = 0.10 n.s.). (b) Temperature across countries correlates with vaccine efficacy values estimated by the model (R ² = 92%, R ²-adj = 87%, GLM fit: ρ = −0.019, p < 0.005 significant), suggesting a possible environmental factor in immunomodulation and vaccination responses. (c) Income inequality, as a proxy for transmission heterogeneity, is associated negatively to realized vaccine efficacy extracted by the model (R ² = 68%, R ²-adj = 46%, GLM fit: ρ = −0.013, p < 0.05 significant). More details are provided in the Supplementary Tables S4 and S5. The fitted regression lines are displayed in Fig. S7.