Adherence to the EAT-Lancet Diet, Genetic Predisposition, and Risk of Kidney Stone Disease

Eating closer to the EAT-Lancet planetary health diet linked to 19% lower kidney stone incidence; genetics barely changed the math

Journal: The American Journal of Clinical Nutrition | Published: 2026-03-14 | Type: Prospective cohort study | PMID: 41833906 Authors: Xu Kun, Bai Song et al. (Department of Urology, Shengjing Hospital of China Medical University, Shenyang, China) Funding/COI: Funding not listed; authors report no conflicts of interest

Summary

A prospective analysis of 197,759 UK Biobank participants found that closer adherence to the EAT-Lancet planetary health diet was associated with a 19% lower incidence of kidney stone disease over roughly 11.6 years of follow-up. BMI explained about 15% of that association — suggesting diet protects partly through weight, and mostly through other mechanisms. A polygenic risk score built from 20 SNPs didn't meaningfully interact with diet score, meaning the dietary association held regardless of inherited genetic risk.

Claims

• Highest vs. lowest EAT-Lancet Diet Index (ELD-I): HR 0.81 (95% CI 0.69–0.95) for kidney stone disease
• 1,708 incident KSD cases across 2,286,927 person-years of follow-up
• BMI mediated 14.94% of the ELD-I → KSD association
• No significant interaction between polygenic risk score (20 SNPs) and ELD-I
• Joint analysis (highest ELD-I + lower genetic risk vs. lowest ELD-I + higher genetic risk): HR 0.61 (95% CI 0.49–0.76)

Study Quality

Nearly 200,000 participants and over 2 million person-years is legitimate scale. Cox proportional hazards regression with multivariable adjustment is appropriate for the design, and incident-only case capture avoids prevalent-case bias. The polygenic risk score using 20 SNPs is modest but not unreasonable for a genome-wide significant composite. Mediation analysis was conducted to partition the BMI pathway from residual diet effects.

The critical methodological weakness is the dietary exposure: ELD-I was calculated from at least one 24-hour dietary recall. Single recalls are notoriously noisy — they capture what someone ate on one day, not habitual intake. That measurement error biases associations toward null, so the observed HR may be conservative, but it also means the exposure variable is fundamentally weak. UK Biobank participants also skew healthier and more educated than the general population (the healthy-volunteer problem), limiting generalizability.

Red Flags

• Dietary recall floor: At minimum one 24-hour recall is the weakest acceptable dietary exposure. Even multiple recalls carry high within-person variability; effect sizes derived from them should be read cautiously.
• No stone composition data: Calcium oxalate, uric acid, and struvite stones have distinct dietary drivers. Lumping all KSD together hides clinically meaningful heterogeneity.
• Funding undisclosed: Not a disqualifier, but absence of disclosure is a gap.
• Mediation precision overstated: Reporting BMI mediation at "14.94%" implies a precision that observational mediation estimates don't support — uncertainty intervals on mediation proportions are typically wide and often unreported.
• Mechanism unexplored: The EAT-Lancet diet was designed for planetary sustainability, not urologic health. Why it might reduce KSD beyond BMI is not established here.
• UK Biobank selection bias: Results may not generalize to less healthy or less educated populations.

Strengths

• Very large prospective cohort with long follow-up
• Incident cases only — dietary data collected before stone events
• Polygenic risk score analysis adds a layer most dietary cohort studies omit
• Mediation analysis attempts mechanistic parsing rather than stopping at association
• No conflicts of interest reported

Verdict

The scale is real and the signal is statistically significant, but the dietary measurement is doing a lot of heavy lifting on a shaky foundation — one 24-hour recall is a crude proxy for habitual eating patterns. A 19% hazard reduction is modest, BMI accounts for only a slice of it, and the mechanism is unexplained. The genetics layer adds novelty without being over-interpreted, which is a methodological credit. File this as hypothesis-generating evidence that the EAT-Lancet pattern may have downstream urologic effects — not as proof that it prevents kidney stones.