Multiple observational studies show an association between fetal and post-natal aflatoxin exposure and reduced linear growth. However, the effects of confounding factors such as socio-economic status, food insecurity and nutrient deficiencies due to monotonous diets have not been ruled out. This trial will quantify the causal role of infant aflatoxin ingestion on post-natal growth by performing a cluster randomized trial in children 6-18 months of age in the Dodoma Region of Tanzania.

All health facilities in one district in Dodoma will be randomized to the control or intervention arm. Infants will be recruited into the study over one year to account for seasonal variation in AF exposure. Both arms will receive infant and young child feeding education, a thermos flask and plastic measuring scoops. The intervention arm will receive a low-aflatoxin pre-blended porridge flour containing maize and groundnut (ratio 4:1 respectively) and low-aflatoxin groundnut flour, whereas in the control arm the same porridge mix will be promoted through education, but acquired by the household. The primary outcome is length-for-age Z-score at 18 months.

The objective of this study is to determine the effect of a very low AF complementary feeding intervention on LAZ using a cluster randomized design (CRT), while promoting a nutritionally adequate diet to all infants between 6-18 months of age. The hypothesis is that ingestion of AF can reduce infant length and that a reduction of AF exposure will result in improved length for age z-scores (LAZ). The primary outcome is LAZ, which will be measured at 18 months.

The unit of randomization is government-run health facilities (health centers, dispensaries and hospital, 52 clusters in total) in the Kongwa District. Mothers and infants will be recruited into the trial based on 42 day EPI visit attendance, which has very high (>95%) coverage in Kongwa District. Recruitment of infants will be performed for one complete calendar year capture variability in exposure by season.

Critical to casual inference is the intervention's ability to create a contrast of AF consumption between the control and intervention groups, without creating differential macro- or micro-nutrient intake or differences in feeding and care practices that could affect stunting between arms. To reduce the risk of introducing these biases, the investigators designed the intervention to include: 1) education to improve infant feeding and care practices in both arms, and 2) behavior change communication on the use of blended infant porridge flours in both arms. Participants in the intervention group will receive low-aflatoxin blended infant porridge flour and groundnut flour, made in accordance with Tanzanian food and mycotoxin regulations. Those in the control group will not receive any flour, but will be advised to feed their infants a blend similar in ingredients and ratio to what the intervention group will receive. Both groups will receive a thermos flask to store porridge and a small plastic measuring scoop to measure porridge flour for preparation of the porridge.

The sample size was calculated using a one-sided test of independent sample means, with a standard deviation of 1.2 Z, type I error of 0.05, power of 0.90, design effect of 2.0 and assuming a coefficient of variation of .144 for varying cluster size, based on previous year's data for EPI attendance at 42 days. Given these parameters our total sample size is calculated to be 2,322 (1,161 infants per cluster). Conservatively estimating a 20% loss to follow-up and infant mortality, the total number of infants is 2,787, or 54 infants recruited per health cluster annually or 4.5 infants per cluster per month. Rounding up, 5 infants per cluster each month will be recruited for a total of 3,120 potential infants, recognizing that in approximately 6 of the health facilities, it is unlikely that all 60 infants because of the size of the population served by the facility.

The trial will be conducted in the Kongwa District of Dodoma, Tanzania, where the investigators have performed formative research (Protocol Identification#: 1703007043) and confirmed AF contamination in local foods, primarily groundnuts, and that infants are exposed to aflatoxin. The frequency and level of exposure is similar in range to the West African observational studies. Kongwa District is a good location to perform this study, as exposure is high enough to be suspected of contributing to stunting, but low enough as to not cause aflatoxicosis.

NOTE: Data collection was stopped between April 9 and June 8, 2020 due to the SARS-CoV-2 outbreak, in accordance with the guidance of the Tanzanian National Institute for Medical Research. The delivery of the intervention continued during this time.

Inclusion Criteria:

1. Babies >6 weeks old and <4 months old, who seek Expanded Program on Immunization (EPI) from a randomized health facility and reside in Kongwa District

Exclusion Criteria, assessed at recruitment and again at the 6-month visit:

1. Babies with disabilities that preclude normal feeding and swallowing
2. Refusal to consent to assigned intervention
3. An infant who has shown signs of a potential groundnut allergy (assessed the first time mother reports groundnut consumption)
4. An infant who is a twin
5. If the mother plans to travel for more than 2 months at or after the randomized intervention begins
6. If the mother is below 16 years of age

De-identified participant data that underlie the results related to the primary outcome will be made publicly available in a public data repository following the publication of the primary outcome paper, ideally 12 months following the end of data collection. Data that underlie the results for all secondary outcomes will also be placed in a public repository following the publication of these papers.

We estimate that data related to the primary outcome will be available 12 months following the end of data collection. We estimate that data related to the secondary outcomes will be available 12-36 months following the end of data collection. Once made public all data will be available for any purpose with no end date .

There are no criteria restrictions, as data will be placed in a public repository.