Gallstone Disease Increases the Risk of Kidney Stone Disease: A Cross-Sectional and Longitudinal Cohort Study

People with gallstones had 80% higher odds of developing kidney stones in a 121,000-person Taiwanese cohort

Journal: International Journal of Medical Sciences | Published: 2026-04-08 | Type: Journal Article | PMID: 42080068 Authors: Lin Yu-Ting et al., Kaohsiung Medical University / MacKay Memorial Hospital, Taiwan Funding/COI: Funding not disclosed; authors declare no competing interests

Summary

Using over 121,000 participants from the Taiwan Biobank, this study found that gallstone disease is an independent risk factor for kidney stone disease — even after adjusting for age, BMI, metabolic comorbidities, and lifestyle factors. The longitudinal arm followed 25,208 stone-naive participants and confirmed the cross-sectional signal: gallstones preceded kidney stones at a statistically significant rate. The proposed mechanism is shared metabolic dysfunction, though the paper does not establish it.

Claims

• In the cross-sectional cohort (n=121,550), KSD prevalence was 11% in those with gallstones vs. 6% without (p<0.001); adjusted OR 1.51 (95% CI: 1.38–1.65)
• In the longitudinal cohort (n=25,208, no prior KSD), incident KSD was 5% in the gallstone group vs. 2% without (p<0.001); adjusted HR 1.80 (95% CI: 1.36–2.38)
• Unadjusted cross-sectional OR was 1.92 — adjustment for metabolic confounders pulled it down to 1.51, suggesting shared metabolic pathways explain part but not all of the association
• Subgroup analysis showed no significant interaction by age; sex interaction was significant in the longitudinal analysis (stronger association in women), though subgroup ORs in the cross-sectional arm (men 1.49, women 1.53) did not differ meaningfully

Study Quality

The cross-sectional sample is large enough to detect modest associations with high precision, and the Taiwan Biobank's community-based recruitment avoids the referral bias typical of clinical datasets. The longitudinal design adds temporal evidence — participants were stone-free at baseline, and GSD status was fixed there to avoid immortal time bias, which the authors explicitly note and address. Cox proportional hazards modeling with comprehensive covariate adjustment (BMI, waist circumference, fasting glucose, triglycerides, uric acid, comorbidities, smoking, physical activity) is appropriate for this design.

That said, both GSD and KSD are self-reported, not confirmed by imaging or medical records. The Taiwan Biobank enrolls only cancer-free adults aged 30–70, which limits generalizability. The longitudinal cohort (25,208) is only about 20% of the cross-sectional cohort — participants needed regular follow-up to qualify, introducing potential selection bias toward healthier or more health-engaged individuals.

Red Flags

• Self-report only: Neither gallstones nor kidney stones were confirmed by ultrasound, CT, or medical records — recall bias and under-diagnosis are real concerns, especially for asymptomatic stones
• Funding not disclosed: No source is listed; the Taiwan Biobank itself has governmental funding, but paper-specific support is unspecified
• Longitudinal dropout: Only ~21% of cross-sectional participants had follow-up data, and the selection criteria for this subset are not fully detailed
• Mechanism is asserted, not shown: The paper invokes "shared metabolic pathways" throughout but does not analyze stone composition, bile acid profiles, or any mechanistic marker
• Sex interaction inconsistency: Significant sex interaction in longitudinal analysis is not reflected in the cross-sectional subgroups — the authors do not reconcile this
• GSD prevalence is low (4.6%): The gallstone-exposed group in the longitudinal arm is only 1,144 people, leaving the HR estimate with wide confidence intervals (1.36–2.38)

Strengths

• One of the largest population-based studies on this association in Asia
• Dual design (cross-sectional + longitudinal) allows both prevalence and incidence inference
• Explicit immortal time bias mitigation via fixed baseline exposure definition
• Comprehensive confounder adjustment including laboratory values (glucose, triglycerides, uric acid)
• Findings replicate prior results from European, American, and other Asian cohorts, strengthening external validity
• Community-based sample reduces referral and ascertainment bias common in clinical registry studies

Verdict

Solid epidemiology with an honest limitation profile. The size is the paper's main asset — 121,000 people is enough to take a 51% adjusted odds ratio seriously. The longitudinal arm is underpowered by comparison (CI 1.36–2.38 on the HR), but directionally consistent. The self-report problem is real and unresolved: both conditions can be asymptomatic, which means misclassification goes both ways and likely attenuates the true association. This paper establishes an epidemiological link worth investigating further; it does not explain why it exists. Clinicians and researchers who work with metabolic stone disease will find it useful as a cross-population replication. For mechanistic insight, look elsewhere.