Effects of Shigella diarrhea with and without antibiotic treatment on linear growth: an individual patient data meta-analysis of five multisite studies among children in low-resource settings

Background Quantifying the effect of Shigella diarrhea with and without antibiotic treatment on linear growth faltering is critical to understanding the potential impact of Shigella vaccines. Methods Using individual-level data from five multisite studies, we estimated the effect of Shigella diarrhea on length/height-for-age z-score (HAZ) 60-90 days after the episode compared to diarrhea episodes with no etiology identified and non-diarrheal controls. Effects were estimated under treatment with and without antibiotics using augmented inverse probability weighted estimators with ensemble machine learning. Findings Among 26,752 diarrhea episodes, 5,503 (20.6%) were attributed to Shigella and of those, 2,567 (46.6%) were treated with guideline recommended antibiotics. Compared to other diarrhea episodes, Shigella diarrhea without treatment with guideline recommended antibiotics was associated with small reductions in HAZ (HAZ difference: -0.03, 95% CI: -0.05, -0.01), but not when treated with guideline recommended antibiotics (HAZ difference: -0.01, 95% CI: -0.02, 0.01). Compared to non-diarrheal controls, Shigella diarrhea was associated with decrements in HAZ in the following 60-90 days regardless of antibiotic treatment (HAZ difference overall: -0.08, 95% CI: -0.10, -0.07). A larger impact of Shigella diarrhea was observed among younger children. Interpretation Shigella diarrhea was associated with short-term decrements in height. Treatment with guideline recommended antibiotics prevented the impact of Shigella on linear growth compared to other diarrhea episodes but not compared to non-diarrheal controls. These results suggest that improved recognition and treatment or prevention of Shigella could improve child growth.