Fruit, nuts, eggs, and milk tracked with taller stature across four decades of Chinese dietary change

China’s long food transition has produced a striking parallel trend: as diets became more varied, children and adolescents grew taller. In a population-based observational analysis spanning 1985 to 2023, the clearest height-associated pattern centered on higher consumption of fruits, nuts, eggs, and milk, a combination the investigators labeled FUEL. Across the most recent period, that pattern showed the strongest and most consistent association with greater height, while also carrying relatively favorable environmental and economic profiles.

The study drew on height data from about 1.7 million Chinese students aged 7 to 18 years and provincial food consumption data from national statistical sources. Rather than treating diet as a set of isolated exposures, the analysis examined how foods moved together over time and how those shifting combinations lined up with growth outcomes. That approach reflects a basic reality of pediatric nutrition: linear meters of calories do not explain stature on their own. Protein quality, mineral density, and food diversity matter during the years when skeletal growth depends on both substrate and hormonal signaling.

Height rose as food consumption became more diverse

From the 1980s through the 2020s, China’s food supply became more diverse, and average height increased alongside it. Eggs and seafood maintained stable associations with height across the study periods, while milk and fruits became more important after 2000. Nine food consumption patterns were identified across three socioeconomic transition phases: 1985–1999, 2000–2010, and 2011–2023. The most recent period produced the most compelling signal for the FUEL pattern, with a correlation of ρ = 0.518 (95% CI 0.453 to 0.580; p < 0.001) with greater height.

One sentence captures the central finding plainly: more of the right foods tracked with more growth.

The association is biologically plausible. Fruits contribute vitamin C and other micronutrients relevant to collagen synthesis and tissue growth. Nuts bring energy, unsaturated fats, and minerals. Eggs and milk supply high-quality protein, calcium, and other nutrients needed for linear skeletal development. During childhood and adolescence, those inputs support bone mineralization and the endocrine environment that permits height gain to continue at a healthy pace.

Protein, calcium, and zinc aligned with growth; carbohydrate density moved in the opposite direction

The nutrient-level analysis sharpened the picture. Protein density, calcium density, and zinc density were each positively associated with height across all three periods, while carbohydrate density showed an inverse association. The food patterns linked to greater height also tended to have better nutritional profiles, especially higher calcium density and higher calcium-to-phosphorus ratios. In practical terms, the foods that appeared most favorable for growth were not simply energy-rich. They were nutrient-dense in ways that matter for the developing skeleton.

This distinction matters in pediatric nutrition policy. A diet can be abundant in calories and still fall short on the mineral and protein composition needed for optimal growth. Calcium and phosphorus interact tightly in bone metabolism, and a more favorable ratio can help explain why milk-containing patterns appeared repeatedly among the food combinations associated with taller stature. Zinc also plays a recognized role in growth and skeletal maturation, making its repeated positive association unsurprising in a cohort at ages 7 to 18 years.

The Food Consumption–Height score offered a population-level signal

Investigators went a step further and built a Food Consumption–Height (FCH) score to quantify the association between food consumption and growth outcomes. The score was derived using machine-learning models and Shapley Additive Explanations weights, then tested in an independent observational sample from Guangzhou. Higher FCH scores were positively associated with height, with β = 0.075 (95% CI 0.069 to 0.082; p < 0.001). In the Guangzhou sample of 3,917 participants aged 10 to 16 years, the association was directionally consistent.

The score is not a clinical diagnostic tool, and it is not a substitute for individual dietary assessment. Its value lies elsewhere: it offers a data-driven way to compare food consumption patterns at the population level and to judge whether a food system is supporting growth in ways that are compatible with health and sustainability. Public health researchers often need tools that can summarize complex dietary environments without flattening them into one nutrient or one food group. The FCH score was built for that purpose.

Sustainability gains accompanied the growth-associated pattern

The most interesting feature of the FUEL pattern may be that it was not just associated with taller stature; it also carried lower environmental and economic burdens relative to less favorable patterns. Food systems research has long wrestled with the assumption that better nutrition must automatically cost more or produce greater ecological strain. This analysis complicates that view. A diet pattern centered on fruits, nuts, eggs, and milk aligned with height gains while remaining comparatively moderate in environmental and economic demand.

That finding matters in a country navigating nutritional modernization at scale. China’s food choices have changed rapidly across four decades, and the study suggests that dietary diversification need not be separated from sustainability goals. The challenge is not merely to increase food availability, but to steer the composition of that food supply toward patterns that support growth during childhood and adolescence without imposing unnecessary environmental costs.

Observational strength, but no causal proof

The study’s size and long time horizon are its main strengths. It combines repeated provincial food consumption data with a very large student height dataset and adjusts for provincial socioeconomic and environmental factors including per capita gross domestic product, educational attainment, sunshine duration, and temperature. The authors also tested alternative analytic approaches, including elastic net regression, weighted quantile sum regression, and principal component analysis, which helps show the stability of the main dietary patterns.

Still, the limitations are substantial and should not be softened. This was an observational study, so causal inference is limited. The main analyses relied on population-level food consumption data rather than individual dietary records, which means the results describe broad ecological associations rather than direct person-level exposure effects. The Guangzhou validation was also observational. Even with a striking signal, the study cannot prove that eating more fruit, nuts, eggs, and milk makes a child taller; it shows that places and periods with those consumption patterns also tended to have taller students.

What clinicians and policy teams should take from the data

The practical message is not that one food group can stand in for a growth strategy. It is that nutritional quality appears to matter as much as, and likely more than, total food volume. For children and adolescents, the combination of protein, calcium, zinc, and other micronutrients embedded in a diversified diet may help explain why certain consumption patterns track with greater height. The findings also give nutrition planners a rare combination to work with: a pattern associated with growth that does not appear to demand the heaviest environmental or economic cost.

Longitudinal follow-up and randomized trials are still needed before anyone can translate these observations into firm dietary prescriptions. For now, the strongest conclusion is narrower: across four decades of Chinese food transition, the pattern most closely aligned with taller height was also the one built around nutritionally dense foods, not empty calories.


Reporting by Dr. A.R. Athira, MBBS for Medicine World Foundation, with additional data provided by PLOS Medicine.