Introduction Observational studies have documented an association between aflatoxin (AF) exposure and reduced linear growth in infants and young children. Our objective was to assess the effectiveness of reducing AF exposure on child linear growth and serum AF levels in rural areas in Eastern Kenya.

Methods A cluster randomised controlled design was used (28 intervention and 28 control clusters). The intervention arm received a swapping (contaminated maize was replaced with safe maize) and a stockist intervention (households were encouraged to purchase from a stockist supplied with clean maize). Women in the fifth to final month of pregnancy were invited to enrol in the study. Outcomes were child length-for-age Z-score (LAZ), the prevalence of stunting and child serum AFB₁-lysine adduct level 24 (endline, primary outcomes) and 11 to 19 months (midline, secondary outcomes) after trial commencement, respectively. The trial was registered with socialscienceregistry.org.

Results Of the 1230 unborn children enrolled in the study, 881 (72%) were included in the LAZ and 798 (65%) in the serum AFB₁ analysis. The intervention significantly reduced endline ln serum AFB₁-lysine adduct levels (intervention effect—0.273, 95% CI −0.547 to 0.001; one-sided p=0.025), but had no effect on endline LAZ or stunting (mean LAZ at endline was −1.64). At midline, the intervention increased LAZ by 0.16 (95% CI −0.009 to 0.33; one-sided p=0.032) and reduced stunting by seven percentage points (95% CI −0.125 to −0.007; one-sided p=0.015), but had no impact on serum AFB₁ levels.

Conclusion Improving access to AF-free maize substantially reduced endline serum AF, but had no effect on child linear growth. The midline analysis suggests that AF may affect linear growth at younger ages.

Trial registration number AEARCTR-0000105.