Supplementary Appendices for 'Maternal BMI mediates the impact of crop-related agricultural work during pregnancy on infant length in rural Pakistan: a mediation analysis of cross-sectional data'

Appendix 1. Sampling.

Appendix 2. Baseline survey questionnaire.

Appendix 3. Hypothesized models of pathways related to maternal BMI represented on a directed acyclic graph (DAG).

Appendix 4. Sample flow chart.

Background

Stunted growth in early infancy is a public health problem in low-and-middle income countries. Evidence suggests heavy agricultural work during pregnancy is inversely associated with maternal body mass index (BMI) and infant birth weight in low- and middle-income countries; but pathways linking agricultural work to length-for-age Z-scores (LAZ) in early infancy have not been examined. This study aimed to investigate the relationship between agricultural work during pregnancy, post-natal maternal BMI and LAZ among young infants in rural Pakistan; and explored whether maternal BMI mediated the relationship between agricultural work and infant LAZ.

Methods

A cross-sectional survey was conducted from December 2015 to January 2016 in rural Sindh, Pakistan. Mother-infant dyads were recruited via systematic random cluster sampling at 2–12 weeks’ post-partum (n = 1161). Anthropometric measurements (maternal and infant height/length and weight) and questionnaire data were collected. Multivariable linear regression and structural-equation based mediation analyses were used to examine associations of agricultural work during pregnancy with maternal BMI and infant LAZ.

Results

During pregnancy, women reported engaging in livestock-related work (57.0%), crop-related work (42.7%), and cotton harvesting (28.4%). All three forms of agricultural work were negatively associated with maternal BMI (β = − 0.67 [− 1.06; − 0.28], β = − 0.97 [− 1.51; − 0.48]; and β = − 0.87 [− 1.33; − 0.45], respectively). Maternal engagement in cotton harvesting alone was negatively associated with infant LAZ after controlling for confounding factors. The total negative effect of cotton harvesting on infant LAZ was − 0.35 [− 0.53; − 0.16]. The indirect effect of maternal BMI on infant LAZ was − 0.06 [− 0.08; − 0.03], revealing that 16% (− 0.06/− 0.35) of the relationship between cotton harvesting and infant LAZ, after adjustment, was mediated via maternal BMI.

Conclusion

These results underscore a need to reduce labour-intensive agricultural workload demands during pregnancy, especially in cotton harvesting, to reduce risks of negative maternal energy balance and poor growth outcomes in early infancy.