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

Can Antibiotics Prevent Stunting?

In low-resource settings, Shigella bacteria are a leading cause of diarrhea in children under two years old. This is a window of development that is exceptionally sensitive to nutritional and environmental shocks. While the immediate danger of Shigella is often framed in terms of acute mortality, there is a quieter threat: growth faltering. This refers to the slowing of linear growth—the physical lengthening of a child. Researchers often measure this using height-for-age z-scores (HAZ). These scores track how a child’s height compares to a global standard.

Researchers have long suspected that repeated enteric (intestinal) infections contribute to stunting. However, the specific role of Shigella has remained difficult to isolate. A critical question remains. Does treating these infections with antibiotics mitigate the impact on a child's growth trajectory? This study investigates whether antibiotic intervention can reduce the developmental delays associated with Shigella.

The hidden cost of enteric infection

The central question of this research is whether Shigella-attributed diarrhea is associated with short-term decrements in linear growth. The authors also ask if guideline-recommended antibiotics can mitigate this impact. Understanding this has been difficult due to two major methodological hurdles.

First, many studies struggle with the fact that antibiotic treatment is a "mediator." This means the infection causes the need for antibiotics, which then affects the growth outcome. If a researcher simply compares treated children to untreated children, they may accidentally measure disease severity rather than the drug's effect. Second, previous studies used inconsistent comparison groups. Some compared Shigella to other types of diarrhea. Others compared infected children to healthy ones. Without a unified way to handle these variables, the field could not clearly define the value of preventing Shigella via vaccination versus treating it with medicine.

A massive scale of individual data

To address these hurdles, the authors performed an individual patient data (IPD) meta-analysis. They pooled data from five multisite studies involving 29,417 children. This approach is much more powerful than a standard meta-analysis. It uses the raw, granular data for every single diarrhea episode recorded.

The investigators used a sophisticated statistical framework. They used augmented inverse probability weighted (AIPW) estimators combined with ensemble machine learning (a tool called SuperLearner). This "doubly robust" method builds flexible models. These models account for complex confounding factors (variables that distort the relationship between an exposure and an outcome). These factors include a child's initial height, socioeconomic status, and the presence of other pathogens like rotavirus. The method also treats antibiotic use as a mediator. This helps the model disentangle the direct effect of the Shigella infection from the indirect effect caused by the subsequent medical treatment.

Measuring the growth gap

The study analyzed 26,752 diarrhea episodes. It identified 5,503 as being caused by Shigella. The researchers focused on the HAZ measured 60–90 days after the infection. This timeframe captures the short-term developmental impact.

The findings reveal a nuanced picture. When comparing Shigella episodes to other types of diarrhea where the cause was unknown, untreated Shigella was associated with a small but significant reduction in height .

The HAZ difference was -0.03 (95% CI: -0.05, -0.01). This represents a minor deviation from the expected growth curve. However, when children received guideline-recommended antibiotics, this reduction was mitigated. There was no significant difference in HAZ compared to other diarrhea episodes .

Yet, this mitigation is relative. When the researchers compared Shigella cases to entirely non-diarrheal controls, Shigella was associated with a decrement in HAZ regardless of treatment .

The overall HAZ difference compared to healthy controls was -0.08 (95% CI: -0.10, -0.07). This suggests that while antibiotics may help a child recover better than they would have otherwise, they do not return the child to the growth trajectory of a healthy, uninfected peer.

Vulnerability in the first year

The impact of Shigella is not distributed evenly. The data indicates a clear trend. The younger the child, the more pronounced the growth decrement. As shown in [Figure 2B], the negative effects on HAZ are most notable among infants. For children aged 0–11 months, the study found a HAZ decrement of -0.10 (95% CI: -0.17, -0.03) compared to non-diarrheal controls.

The observed statistical effect also varies by time. In the MAL-ED study data, the HAZ difference for moderate-to-severe diarrhea was larger at nine months following the episode than at one month .

This was followed by "catch-up" growth. This temporal pattern suggests that the measurable gap in height relative to healthy controls may widen before eventually narrowing.

Furthermore, the effectiveness of antibiotic treatment is threatened by biology. In the EFGH study, the authors found that when Shigella isolates were non-susceptible to the antibiotics being administered, the decline in HAZ was significantly larger .

The HAZ difference for non-susceptible strains was -0.05 (95% CI: -0.07, -0.03). This is a larger reduction than the -0.02 (95% CI: -0.04, 0.00) seen in susceptible strains.

Implications for prevention and policy

These results shift the focus of global health strategy. The core implication is that while antibiotic treatment is a vital tool, it may be an incomplete solution for preventing growth faltering. If antibiotics do not fully restore a child to the growth trajectory of a healthy peer, clinical management alone may be insufficient to combat stunting.

This provides a scientific rationale for the development of Shigella vaccines. A vaccine would prevent the infection from occurring. This bypasses the cascade of inflammation and nutritional loss. Even successful antibiotic treatment fails to fully erase these effects. Moreover, the heightened vulnerability of infants suggests that vaccine schedules should prioritize early administration. Targeting the first year of life could maximize the benefit for linear growth.

The study does not claim to solve the problem of antibiotic resistance. It identifies resistance as a critical barrier to growth protection. Future research must determine if the "catch-up growth" observed in some cohorts is sufficient. We still do not know if these short-term HAZ decrements represent a permanent loss of developmental potential.

Figures from the paper