(P07-012-25) Spatial Trends in Effects of Multiple Micronutrient Supplementation on Low Birth Weight and Infant Wasting in Rural Bangladesh

Objectives: A cluster-randomized trial in NW Bangladesh (JiVitA-3) found antenatal multiple micronutrient (MMS) vs iron-folic acid (IFA) supplements reduced low birth weight (LBW) by 12%. The 435 sq km study area exhibited a West to East gradient in rurality and poverty with increasing proximity to the Teesta River in the East. We examined if MMS effects on LBW and infant wasting also exhibited a geospatial trend.

Methods: We included 9,701 singletons weighed within 72 hours of birth and with household (HH) geospatial data.   Outcomes included LBW ( < 2500 g) and a thin mid-upper arm circumference (MUAC) at birth ( < 9 cm), 1 mo ( < 11 cm) and at 3, 6 & 12 mo ( < 11.5 cm) of age.  Using R, we estimated continuous intensity surfaces of the concentration of low and normal weight births (NBW), and MUAC below and above cut-offs, sexes combined, across the study area, calculating ratios of the surfaces to identify hotspots (higher concentration areas) of LBW and thin MUAC relative to normal.  Geographic gradients of intensity ratios were stratified into five classes, using ArcGIS, and mapped to visualize variation in LBW and wasting by MUAC.  Clustering of LBW was assessed by Moran’s I.  We used GEE logistic regression to estimate relative odds of LBW and wasting in MMS vs IFA groups, adjusted for design effect, gestational age at birth (for MUAC), HH socioeconomic status and food security, maternal education and parity. 

Results: Risks of LBW and wasting exhibited strong spatial variation moving West to East, tracking increasing poverty. Moran’s I revealed spatial clustering of LBW and all wasting by MUAC outcomes: for LBW, I = 0.022; for infant wasting, I = 0.047 at birth, 0.054 at 1 mo, 0.036 at 3 mo, 0.046 at 6 mo, and 0.051 at 1 year (all p < 0.001). Risk ratios (RR) with 95% CI reflected protective effects of MMS for LBW (0.88; 0.83-0.93) and infant wasting at birth (0.89; 0.81-0.97), 1 mo (0.65; 0.48-0.84) and 3 mo (0.86; 0.75-0.97), but not at 6 (0.85; 0.64-1.12) and 12 (1.23; 0.95-1.44) mo of age. RRs of MMS vs IFA did not reveal spatial trends.

Conclusions: There was strong spatial variation in intensity of LBW and infant wasting, tracking a gradient in poverty, but protective effects of MMS against LBW and early infant wasting were comparable across the study region, not modified by geography.

Funding Sources: Gates Foundation