Exploring the relationship between polio type 2 serum neutralizing antibodies and intestinal immunity using data from two randomized controlled trials of new bOPV-IPV immunization schedules

Figures

• 

  Fig. 1. (A) Model-estimated median titer of shed type 2 poliovirus as a function of neutralizing antibody titers pre-mOPV2 challenge and nominal day post-challenge of stool sample collection. Lines indicate regression fit to observed data ignoring censoring at NAb LLOQ (2.5 log2) and ULOQ (10.5 log2). Observed data points are jittered to convey density. Samples negative for viral shedding are assigned a value of 0. (B) The same as panel A, except data (circles) are not jittered, and values at NAb LLOQ/ULOQ are replaced with one illustrative realization form the censored-data simulation (triangles). Median regression line indicates model fit predictions aggregated across censored data simulations.

• 

  Fig. 2. (A) Model-estimated median shedding index endpoint (SIE) of shed poliovirus (type 2) as a function of neutralizing antibody titers pre-mOPV2 challenge. Line indicates regression fit to observed data ignoring censoring at NAb LLOQ (2.5 log2) and ULOQ (10.5 log2). Observed data points are jittered to convey density. Red boxes and accompanying error bars and numbers indicate the point estimates of median SIE for each discrete level of NAb titer, along with corresponding bootstrap-based 95% confidence intervals, and the number of subjects available for analysis, respectively. Also shown are the medians and accompanying bootstrap-based 95% confidence intervals for both SIE (vertical error bars) and NAb titers (horizontal error bars) for each regimen from the combined studies; symbol size indicates sample size. The IPV001 arm given three doses of tOPV is shown (lower right) to provide context, and was not included in model-fitting procedures. (B) The same as panel A, except data (circles) are not jittered, and values at NAb LLOQ/ULOQ are replaced with one illustrative realization from the censored-data simulation (triangles). Median regression line indicates model fit predictions…

• 

  Fig. 3. Model-estimated median shedding index endpoint (SIE) of shed poliovirus (type 2) by country as a function of neutralizing antibody titers pre-mOPV2 challenge and nominal day post-challenge of stool sample collection for the subset analysis of IPV001 participants only. Faint line indicates regression fit to observed data ignoring censoring at NAb LLOQ (2.5 log2) and ULOQ (10.5 log2); bold line indicates aggregated median regression fit predictions from simulation method to account for censoring. Observed data points are jittered to convey density. Figure is limited to observed range of data due to poor performance of the spline fit at the low-sample-size extremes of the NAb range.