Stronger El Niño conditions were associated with higher odds of reported diarrhea among children under five, while stronger La Niña conditions were associated with lower odds. The overall links were modest, but the El Niño pattern was larger in some rural, lower-income and less-protected water, sanitation and hygiene (WASH) settings.
The study reports odds ratios, a comparison measure in which a value above 1 indicates higher odds and a value below 1 indicates lower odds. Each standard-deviation increase means one typical step on the study's exposure scale. For El Niño, that step was associated with an odds ratio of 1.032, with a 95% confidence interval from 1.025 to 1.040. For La Niña, it was 0.977, with a 95% confidence interval from 0.971 to 0.984.
How the analysis was built
The researchers set out to examine whether El Niño-Southern Oscillation, or ENSO, was associated with diarrhea in children under five and whether water, sanitation and hygiene conditions changed the relationship. The analysis used 1,340,334 children from 103 DHS surveys in 53 low- and middle-income countries, covering 2005 to 2019.
In the surveys, women of reproductive age were asked whether the child had diarrhea in the two weeks before the interview. Overall, 176,811 children were reported to have diarrhea, a prevalence of 13.19%.
ENSO exposure was treated as a continuous measure rather than a simple exposed-or-unexposed label. The researchers used three-month running means of the E and C indices, then built location-specific teleconnections, or local climate links, from temperature and precipitation patterns over the previous 12 months.
To account for differences between children and places, the model adjusted for child sex and age, maternal education, residence, household wealth, seasonal patterns and survey year. It also allowed for clustering by survey location and country, and the authors said a linear model fit better than nonlinear alternatives they tested.
The pattern was not the same everywhere
The strongest El Niño associations appeared where resources and services were less protective. In rural areas, the odds ratio was 1.042 (95% CI 1.033 to 1.052); in low-income households, it was 1.047 (95% CI 1.035 to 1.060). The association was also stronger where toilet facilities or drinking-water sources were unimproved.
Geography and season also shaped the El Niño pattern. The association was strongest in Africa, with an odds ratio of 1.058 (95% CI 1.049 to 1.066), and in drier-teleconnection regions, at 1.046 (95% CI 1.033 to 1.060). It peaked in spring, at 1.101 (95% CI 1.079 to 1.124), while summer showed the lowest level; the seasonal contrast met the reported P<0.05 threshold.
La Niña showed a different pattern. Associations were stronger in Africa and middle-income countries, and the strongest seasonal estimate came in summer, with an odds ratio of 0.915 (95% CI 0.899 to 0.932). The association was also stronger in wetter-teleconnection regions, at 0.932 (95% CI 0.923 to 0.942); cooler and warmer regions did not differ significantly (P=0.274).
The La Niña analysis also found effect-size differences by improved drinking-water source, improved toilet facility and water at the handwashing site. Sex and age did not show significant subgroup differences.
What the study cannot establish
The findings are associations, not proof that ENSO causes diarrhea. Because the study used cross-sectional survey data, it cannot show that a change in El Niño or La Niña exposure would change an individual child's outcome.
Some uncertainty also comes from how the data were collected. The diarrhea report could be affected by maternal recall and outcome misclassification. ENSO exposure was assigned at an administrative-unit level because individual residential addresses were unavailable, and displaced cluster coordinates may have added measurement error.
The authors reported that the main associations held up when they changed the ENSO indices, teleconnection coefficients, exposure months and adjustment for teleconnection direction. That robustness does not answer whether the underlying pathway involved drought-related water scarcity, flood contamination or another factor, a distinction the available water measures could not resolve.
The study also notes that the ENSO associations were relatively small compared with poor WASH, undernutrition and lack of rotavirus vaccination. Its result is therefore a climate-related association to investigate further, not an estimate of preventable cases or a demonstrated intervention benefit.
Paper data and sources
Original title: El Niño–Southern Oscillation and Diarrhea Risk among Young Children in Low- and Middle-Income Countries
Authors: Li G, Lu B, Geng M et al.
Journal/Repository: Environmental health perspectives
Status: Peer-reviewed
First online: 2026-10-09
DOI: Not available
Original paper · Full text