Meta-effect of insect resistant maize on fumonisin B1 in grain estimated by variance-weighted and replication-weighted analyses

Abstract

Fumonisins are a class of mycotoxins occurring in foods and feeds at concentrations that may be of concern to human and animal health. Meta-analysis was used to predict the probability of fumonisin B1 (FB1) exposure reduction that may be achieved in the maize grain supply regionally over time through use of insect resistant (Bt) varieties and how analytical outcomes may be influenced by data richness. All analyses showed the positive effect of Bt maize to reduce FB1 concentrations in grain relative to comparable non-Bt maize, but the effect size was smaller for variance-weighted data than for the less statistically robust data sets. Variance weighted data indicated a 14 to 18% decrease in FB1 concentrations in Bt maize relative to a comparable non-genetically engineered variety, whereas a replication weighted approach indicated a 67% decrease. Understanding the effects of data selection criteria on outcomes of meta-analyses are useful in policy considerations related to the benefits of wide-area adoption of fumonisin mitigation practices, such as for economic analysis of the benefit of widespread Bt adoption, prediction of both short and long-term exposures to FB1 in feed rations and understanding variation of FB1 presence in maize sourced for food products.