The data were sourced from NHANES (2007–2018), a nationwide survey representative of the U.S. population. Data were collected through household interviews and mobile examination centers, covering dietary intake, health conditions, and lifestyle information.

The source population comprised participants from the NHANES database, which aims to assess the health and nutritional status of individuals across various age groups and ethnicities in the United States.

Based on inclusion and exclusion criteria, a total of 19,841 eligible participants were included in the final analysis: exclusions included pregnant women, individuals with missing data (e.g., poverty income ratio, alcohol consumption, and BMI), and those with unclear KS status.

The dependent variable in this study is the presence of kidney stones (KS), determined based on participants’ responses to the question, “Has a doctor ever diagnosed you with kidney stones?” A “yes” response indicates the presence of KS, while a “no” response indicates its absence.

The DI-GM score is based on the intake of 14 specific foods or nutrients. Avocado, broccoli, chickpeas, coffee, cranberries, fermented dairy products, dietary fiber, green tea, soy, and whole grains are classified as beneficial components, while red meat, processed meat, refined grains, and high-fat diets (≥40% of energy from fat) are considered detrimental components (9). Dietary components are categorized based on sex-specific medians: for beneficial components, intake equal to or above the median scores 1 point, while intake below the median scores 0 points; for detrimental components, intake below the median scores 1 point, while intake equal to or above the median scores 0 points. The total score is the sum of all component scores, ranging from 0 to 13, with higher scores indicating dietary patterns more favorable to gut microbiota health (16).

To control for confounding effects, we included a range of covariates in the analysis. Demographic covariates included age (as a continuous variable in years), gender (male or female), and race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Mexican American, and other racial/ethnic groups). Socioeconomic covariates included the poverty income ratio (PIR, a continuous variable), educational level (categorized as less than high school, high school or GED, and above high school), and marital status (married/living with partner, single, or never married). Lifestyle factors included smoking status (categorized as never smoker, former smoker, and current smoker) and alcohol consumption (categorized as never drinker, low drinker, and heavy drinker). Health-related covariates included hypertension status (yes or no), body mass index (BMI, as a continuous variable in kg/m²), serum calcium level (as a continuous variable in mg/dL), and estimated glomerular filtration rate (eGFR, as a continuous variable in mL/min/1.73 m²). Additionally, diabetes status was included as a clinical covariate, categorized as diabetic, prediabetic, or non-diabetic. These covariates were selected based on their established association with kidney stone formation and gut microbiota health and were included in the multivariable models to minimize confounding bias.

In this study, diabetes was defined by the following criteria: fasting plasma glucose (FPG) ≥7.0 mmol/L (126 mg/dL) or HbA1c ≥6.5% (24). Additionally, the response to the question “Has a doctor ever diagnosed you with diabetes?” was used to confirm diabetes status, with “yes” indicating diabetes and “no” indicating non-diabetes. In Table 1, we categorized diabetes status into three groups: “diabetic,” “prediabetic,” and “non-diabetic.” For statistical analyses in other figures, the “prediabetic” and “non-diabetic” groups were combined to facilitate the assessment of the relationship between diabetes, DI-GM index, and KS.

NHANES employed a stratified, multistage probability sampling design to ensure data representativeness and accuracy. The final sample size comprised 19,841 participants.

Data collection was conducted through standardized questionnaires (household interviews) and physical examinations at mobile examination centers. Dietary data were collected using 24-h dietary recall and were used to calculate DI-GM scores.

Data were collected by NHANES staff and subjected to quality control procedures to ensure accuracy. All measurements were calibrated and validated using standardized protocols.

Statistical analyses in this study followed CDC guidelines. Continuous variables are expressed as mean ± standard deviation and categorical variables as percentages. Normal data were analyzed by ANOVA, non-normal data by the Kruskal–Wallis test, and categorical data by chi-square test. A multivariable logistic regression model evaluated the association between DI-GM and KS, providing odds ratio (OR) with 95% confidence intervals (95% CI). Model 1 had no covariate adjustments; Model 2 adjusted for sex, age, and race; Model 3 further adjusted for education, marital status, smoking, drinking, hypertension, PIR, calcium, eGFR, and BMI. Subgroup analyses used multivariable regression, and mediation analysis assessed diabetes as a mediator for DI-GM and KS. A p-value <0.05 was considered significant. Analyses were conducted using Empower version 4.2 (X&Y Solutions, Inc., Boston, MA, United States) and R version 3.4.3 (R Foundation) (see Figure 1).

All NHANES data were approved by the NCHS Research Ethics Review Board, and all participants signed informed consent forms. The data are publicly available and do not involve personal privacy or additional ethical approval.

Of the 19,841 participants analyzed, 1,912 (9.64%) were identified with KS. A distinct profile emerged for those with KS compared to those without. Males were more frequently affected (55.23%) than females, with significant racial differences observed, as the prevalence of KS was higher among non-Hispanic Black individuals. Marital status was also associated with KS presence, with a larger proportion of affected participants being married or living with a partner. Regarding health conditions, those with KS had higher rates of diabetes mellitus (25.99%) and hypertension (49.79%). Lifestyle factors also varied, with a higher proportion of KS patients being current smokers (20.55%) and heavy drinkers (59.78%). Clinically, participants with KS were older on average (55.12 ± 16.24 years) and had a higher mean BMI (30.57 ± 6.87 kg/m²) compared to those without KS. Moreover, mean serum calcium levels were lower, while estimated glomerular filtration rates (eGFR) were also reduced among participants with KS (93.74 ± 16.27 mL/min/1.73 m²). Analysis of dietary indices related to gut microbiota revealed a slightly lower mean score in the “beneficial to gut microbiota” category among those with KS (2.23 ± 1.41), though no significant differences were found in the “Unbeneficial to gut microbiota” category (see Table 1).

Table 2 revealed a statistically significant association between the DI-GM and KS prevalence, even after adjusting for various confounding factors. Participants with a DI-GM score of 6 or higher had a 17% lower prevalence of KS compared to those with a score of 0–3 (Model 3: OR = 0.83, 95% CI: 0.72–0.95, p = 0.007), with a significant trend of decreasing KS risk as DI-GM scores increased (p-trend <0.001). The “beneficial to gut microbiota” component of DI-GM was also independently associated with a lower KS risk (Model 3: OR = 0.94, 95% CI: 0.90–0.97, p < 0.001). Additionally, the analysis indicated that diabetes was associated with a higher prevalence of KS, with diabetic individuals having a 46% higher likelihood of developing KS compared to non-diabetic participants (Model 3: OR = 1.46, 95% CI: 1.29–1.65, p < 0.001).

Table 3 shows that both the “beneficial to gut microbiota” and “unfavorable to gut microbiota” components of the DI-GM were significantly associated with diabetes prevalence. Specifically, the “beneficial to gut microbiota” component showed a borderline association with reduced diabetes risk after full adjustment for confounding factors (Model 3: OR = 0.97, 95% CI: 0.94–1.00, p = 0.0285). The “unfavorable to gut microbiota” component demonstrated a stronger inverse association, where a higher score was associated with a significantly lower diabetes prevalence (Model 3: OR = 0.90, 95% CI: 0.86–0.93, p < 0.0001).

The results of the RCS demonstrated a nonlinear correlation between DI-GM and the prevalence of KS (Figure 2).

Figure 3 presents a subgroup analysis of the association between DI-GM scores and KS prevalence across various demographic and health-related factors. The analysis shows that the protective association between DI-GM and KS is consistent across most subgroups, including age, sex, race, education, marital status, BMI, hypertension status, serum calcium levels, and eGFR categories (p for interaction >0.05). Drinking status is the only factor that significantly modified the association (p = 0.004). Among heavy drinkers, a higher DI-GM score was associated with a reduced risk of KS (OR = 0.92, 95% CI: 0.88–0.95), while no significant association was observed between never-drinkers or moderate-drinkers.

The outcome of the mediation analysis shows that diabetes partially mediates the relationship between DI-GM and KS prevalence. In Model 3, the total effect of DI-GM on KS was significant (Estimate = −0.009489, p < 0.0001). The mediation effect through diabetes was also significant (Estimate = −0.000903, p = 0.0020), accounting for 9.27% of the total effect. The direct effect of DI-GM remained significant after adjusting for diabetes (Estimate = −0.008633, p < 0.0001) (see Table 4 and Figure 4).

This study contributes to a growing body of research exploring the role of gut microbiota in KS pathogenesis. Prior studies have highlighted the role of specific microbial taxa, such as Oxalobacter formigenes, in degrading oxalates, reducing urinary oxalate excretion, and lowering the risk of calcium oxalate stones (25). Our findings align with this mechanism but extend the scope by demonstrating the impact of overall dietary patterns on gut microbial functionality. Furthermore, while previous research has identified reduced urinary citrate as a contributing factor to KS formation, this study links gut microbiota dysbiosis, such as an increased abundance of Escherichia-Shigella, to lower citrate levels, suggesting a novel pathway connecting microbiota composition with lithogenesis (26). Unlike studies that focus on single microbial or dietary factors, this research underscores the cumulative impact of diet-induced microbial shifts on KS risk (27, 28).

The protective effect of DI-GM appeared more pronounced in subgroups such as females, non-Hispanic Black individuals, and heavy drinkers. Among heavy drinkers, the strong inverse association suggests that dietary components beneficial to gut microbiota may mitigate the adverse metabolic effects of alcohol on the kidneys. Alcohol consumption is known to disrupt gut microbiota balance, increasing intestinal permeability and systemic inflammation (29, 30). Diets rich in fiber and polyphenols, reflected in higher DI-GM scores, may counteract these disruptions by enhancing microbial diversity and short-chain fatty acid (SCFA) production, which have anti-inflammatory and renal-protective effects (31, 32). However, the absence of significant effects in non-drinkers and moderate drinkers highlights the complexity of interactions between alcohol, diet, and gut microbiota. These findings align with prior studies on alcohol’s effects on gut health and suggest the need for further investigation into subgroup-specific dietary interventions (33).

Gut microbiota impacts KS formation through several pathways. First, oxalate metabolism remains a central mechanism, as oxalate-degrading bacteria, such as Oxalobacter formigenes, reduce hyperoxaluria by breaking down dietary oxalate in the gut and limiting its intestinal absorption, which is a key risk factor for calcium oxalate stones (34, 35). Loss of these bacteria due to dysbiosis or antibiotic use is associated with increased urinary oxalate excretion and higher KS risk (36, 37). Second, microbial metabolites, such as short-chain fatty acids (SCFAs), including butyrate and acetate, not only improve gut barrier function but also modulate systemic inflammation and oxidative stress, both of which are critical in maintaining renal health and reducing stone risk (31, 38). Moreover, SCFAs help regulate metabolic pathways related to citrate excretion and urinary pH, creating a less favorable environment for stone formation (39). Lastly, gut dysbiosis promotes systemic inflammation by increasing gut permeability and endotoxin levels, such as lipopolysaccharides (LPS), impairing renal function and exacerbating metabolic conditions (22, 40). These mechanisms highlight the potential of dietary patterns that enhance gut microbial diversity and functionality to protect against KS.

The mediation analysis revealed that diabetes accounted for 9.27% of the total effect of DI-GM on KS risk. While this proportion is modest, it underscores the interconnected roles of diet, gut microbiota, and metabolic health. Diabetes exacerbates lithogenesis through hyperglycemia-induced increases in urinary calcium and reduced citrate excretion (41, 42). These mechanisms create a urinary environment conducive to stone formation, with higher calcium concentrations promoting calcium oxalate crystal aggregation and reduced citrate levels decreasing the inhibition of such crystallization. Furthermore, diabetic patients often exhibit changes in urinary pH due to impaired renal ammoniagenesis, which may further contribute to stone formation by facilitating uric acid stone precipitation (43).

In addition to metabolic alterations, diabetic patients have distinct gut microbial profiles characterized by reduced microbial diversity and an increased abundance of pro-inflammatory taxa, which exacerbate systemic inflammation and oxidative stress (44, 45). This gut dysbiosis has downstream effects on metabolic pathways relevant to kidney stone formation, such as increased intestinal permeability and the translocation of microbial endotoxins like lipopolysaccharides (LPS) into circulation. LPS triggers systemic inflammation, which has been linked to renal tubular damage, altered calcium handling, and increased oxalate excretion, thereby enhancing lithogenesis. Moreover, inflammation may directly disrupt gut-kidney axis interactions, highlighting the multifaceted role of gut microbiota in kidney stone risk (46, 47).

These findings align with prior studies demonstrating that high-fiber, low-sugar diets support beneficial bacterial taxa, such as Bifidobacterium and Lactobacillus, which improve insulin sensitivity and reduce urinary calcium excretion (21, 48). Additionally, such dietary patterns promote the production of short-chain fatty acids (SCFAs), such as butyrate, which not only improve gut barrier integrity but also have anti-inflammatory properties that may mitigate the systemic effects of diabetes on kidney stone formation. These dietary influences extend beyond diabetes itself to directly impact kidney stone risk through gut microbiota-mediated pathways (49, 50).

The direct effect of DI-GM on kidney stones, which constitutes a larger proportion of the total effect, highlights the importance of dietary quality in directly modulating gut microbiota composition and kidney stone risk, independent of diabetes.

This study’s strengths include its use of a large, nationally representative dataset and robust statistical analyses, which enhance its reliability and generalizability within the U.S. population. However, several limitations must be acknowledged. First, while the DI-GM is a validated tool for assessing the impact of diet on gut microbiota, it provides an indirect estimation rather than a direct measurement of microbial composition, highlighting the need for future studies to incorporate microbiome sequencing technologies. Second, the NHANES dataset lacks detailed information on specific kidney stone types (e.g., calcium oxalate, uric acid stones), comprehensive antibiotic use data (e.g., frequency, duration, and type), and urinary factors critical to kidney stone risks, such as oxalate and citrate levels. These urinary metabolites play essential roles in kidney stone formation: oxalate is a key component of calcium oxalate stones, while citrate acts as an inhibitor of crystal aggregation. The absence of such data limits the ability to fully assess the mechanistic pathways linking diet, gut microbiota, and kidney stone risk. Antibiotic use, in particular, is known to disrupt gut microbial taxa, including oxalate-degrading bacteria, potentially influencing stone risk. Lastly, while diabetes was included as a mediating variable, future research should explore additional mediators, such as systemic inflammation, oxalate metabolism, and other urinary factors, to better understand the complex pathways linking diet, gut microbiota, and kidney stones. Future studies should also examine diverse populations and leverage multi-omics approaches to provide a more nuanced understanding of these relationships.

Higher DI-GM scores, reflecting diets rich in fiber, whole grains, fruits, vegetables, and fermented foods, are associated with a reduced risk of KS, partially mediated by diabetes. These findings highlight the importance of dietary strategies targeting gut microbiota health in KS prevention. Specifically, increasing the intake of fiber-rich foods (e.g., legumes, whole grains, leafy greens), reducing added sugars and processed foods, and incorporating probiotics or fermented foods (e.g., yogurt, kefir, sauerkraut) may enhance gut microbial diversity and functionality (51). For high-risk populations, such as individuals with diabetes or recurrent KS, targeted interventions like personalized nutrition plans and supplementation with oxalate-degrading probiotics (Oxalobacter formigenes) warrant further exploration (52). Future research should focus on developing practical dietary recommendations and microbial therapies to address KS risk effectively.