Association between Kidney Stones and CKD: A Bidirectional Mendelian Randomization Study

Abstract

Key Points:

1. Common kidney stones are unlikely to be an independent and direct cause of CKD in the general population.

2. CKD may protect against kidney stones because of changes in key urinary factors critical for stone formation.

Background: Kidney stones and CKD are common disorders with a substantial interaction. Although observational studies have suggested a potential for enhanced CKD risk after prior kidney stones, the exact relationship remains ambiguous.

Methods: Shared comorbidities between two diseases were identified using unbiased screening. Genome-wide association study summary statistics were obtained from the UK Biobank (UKBB), FinnGen, and CKDGen, followed by genetic association analyses across various traits. Bidirectional Mendelian randomization (MR) analyses were performed to define causal links, complemented by multivariable MR that included the shared comorbidities including hypertension, diabetes, and obesity. Observational analyses were undertaken using cohorts from the Mayo Clinic and a UKBB subset.

Results: Despite identifying a total of 123 conditions as shared comorbidities, there was no significant genetic correlation between kidney stones and CKD. Unadjusted MR analysis revealed no significant association between kidney stones and CKD risk (UKBB [exposure]/FinnGen [outcome]: odds ratio [OR]=0.97, 95% confidence interval [CI], 0.88 to 1.06; FinnGen/UKBB: OR=1.17, 95% CI, 0.98 to 1.39). Kidney stones did significantly associate with a higher urinary albumin-creatinine ratio (β=0.014, 95% CI, 0.002 to –0.025), but this association disappeared in the multivariable MR model (β=0.009, 95% CI, −0.003 to 0.020). Furthermore, in a cross-sectional analysis limited to the UKBB cohort, a robust regression model did not detect an independent association between kidney stones and urinary albumin-creatinine ratio (β=0.16, 95% CI, −0.04 to 0.35) or eGFR (β=0.10, 95% CI, −0.07 to 0.28). Conversely, CKD associated with a diminished risk of kidney stones in multivariable MR models (UKBB/FinnGen: OR=0.77, 95% CI, 0.69 to 0.87; FinnGen/UKBB: OR=0.73, 95% CI, 0.66 to 0.81). Furthermore, in the Mayo Clinic cohort with available urinary biochemistries, lower eGFR was associated with lower urinary calcium excretion and urinary calcium oxalate/phosphate supersaturation.

Conclusions: In this study, kidney stones were not independently associated with CKD. Conversely, CKD was associated with a lower risk of calcium kidney stones likely via changes in key urinary traits, including lower calcium excretion.

Graphical abstract

Figure 1

Overall workflow of the study. GWAS, genome-wide association study; MR, Mendelian randomization; UACR, urinary albumin-creatinine ratio; UKBB, UK Biobank.

Figure 2

Identification of shared comorbidities between kidney stones and CKD. (A) Venn plot showing the count of unique and shared comorbidities of kidney stones and CKD. (B) Forest plot showing the effect sizes of a subset of shared comorbidities on kidney stones and CKD. CI, confidence interval; OR, odds ratio.

Figure 3

GWAS results and genetic correlation for kidney stones and CKD. Miami plot showing the GWAS findings for kidney stones and CKD in (A) UKBB and (C) FinnGen. Genetic correlations across traits in (B) UKBB and (D) FinnGen, with correlation coefficients and ellipse visualizations. In the UKBB GWAS, CKD was defined as N18.3–N18.5, whereas FinnGen included all N18 codes.

Figure 4

MR analyses of kidney stones' influence on CKD risk and UACR levels. (A) Unadjusted and (B) multivariable MR analyses using dataset pairs UKBB (exposure) and FinnGen (outcome) and vice versa. (C) Unadjusted MR analysis using dataset pairs FinnGen (exposure) and CKDGen (outcome). (D) Single covariate screening for contributing factors of UACR using MR analysis in the same dataset pair. Estimates were derived using IVW methods. In the UKBB GWAS, CKD was defined as N18.3–N18.5, whereas FinnGen included all N18 codes. UACR levels were assessed as log-transformed values. IVW, inverse variance weighted.

Figure 5

MR analyses of CKD's influence on kidney stone risk. (A) Unadjusted and (B) multivariable MR analyses using dataset pairs UKBB (exposure) and FinnGen (outcome) and vice versa. Estimates were derived using IVW methods. In the UKBB GWASs, CKD was defined as N18.3–N18.5 with eGFR <60 ml/min per 1.73 m², whereas FinnGen included all N18 codes.

Figure 6

Differential analysis of kidney stone–related urinary parameters in patients with and without CKD in the Mayo cohort. (A) Urinary 24-h calcium, (B) urinary calcium concentration, (C) urinary calcium oxalate supersaturation, (D) urinary 24-h magnesium, (E) urinary 24-h oxalate. CKD was defined as eGFR <60 ml/min per 1.73 m². The boxplot shows data distribution by displaying the median, quartiles, and range. Outliers are marked as individual points to highlight significant deviations.