Genetic correlation and mendelian randomization reveal the impact of sleep traits on urolithiasis risk

Abstract

Urolithiasis is a common and recurrent condition in the urological spectrum. Despite various proposed mechanisms, the causal relationship between sleep traits and the risk of urolithiasis remains unclear. We used publicly available genome-wide association study (GWAS) summary data from the UK Biobank and FinnGen to perform a two-sample Mendelian randomization (MR) analysis and genetic correlation analysis, evaluating the causal relationship and genetic correlation between sleep traits (chronotype, getting up in the morning, sleep duration, nap during the day, and insomnia) and urolithiasis (calculus of the kidney and ureter, and calculus of the lower urinary tract). Additionally, multivariable MR (MVMR) analysis adjusted for body mass index (BMI) and other sleep characteristics was conducted to assess the direct impact of sleep traits on the risk of urinary tract stones. The LD score regression (LDSC) analysis indicated a genetic correlation between insomnia and upper urinary tract stones (r_g = 0.082, P = 0.017, Adjusted P = 0.085), but no significant genetic correlation was found for other sleep traits. Our results indicated no causal relationship between sleep traits and upper urinary tract stones. However, insomnia was significantly associated with a higher risk of lower urinary tract stones (IVW [inverse variance weighted] OR [odds ratio] = 5.91, 95% CI [confidence interval] 1.52–22.98, P = 0.010, Adjusted P = 0.030), while early rising exhibited a protective effect (IVW OR = 0.29, 95% CI 0.11–0.76, P = 0.012, Adjusted P = 0.030). In the MVMR analysis, insomnia consistently showed a similar trend, whereas daytime napping significantly reduced the risk of lower urinary tract stones (OR = 0.28, 95% CI 0.12–0.65, P = 0.003). This study provides MR-based evidence suggesting that insomnia may increase the risk of lower urinary tract stones, while daytime napping may reduce this risk. No causal relationship was found between sleep characteristics and upper urinary tract stones.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-82031-4.

Introduction

Urolithiasis, a prevalent urinary system condition, affects approximately 10% of the global population, exhibiting a notably high recurrence rate, with a recurrence rate of up to a 50% within a five-year period^1–3. Research indicates that the formation of stones is influenced by various factors, such as age, sex, ethnicity, genetics, and the environment. However, the underlying mechanisms driving stone formation remain unclear⁴. Currently, for the pharmacological treatment of stones, different medications can be selected based on their composition. For instance, allopurinol is used to treat uric acid stones, calcium channel blockers are utilized for calcium phosphate stones, and D-penicillamine or tiopronin can be employed for cystine stones. Moreover, most types of stones can be alkalinized using either sodium bicarbonate or potassium citrate⁵. However, once stones progress to the point of requiring surgery, as is often the case when patients initially present with stones, the efficacy of medical dissolution may diminish compared to the severity of the patient’s condition. At this stage, surgical interventions such as shock wave lithotripsy, percutaneous nephrolithotomy and ureteroscopy are widely employed^6,7. Nevertheless, repeated surgeries impose substantial economic burdens on patients and pose significant challenges to public health. Therefore, the prevention and treatment of urolithiasis are urgent issues demanding resolution.

Various factors, such as shift work, time zone changes, and late-night studying, contribute to diverse sleep patterns among individuals. These irregularities in modern lifestyles may disrupt the body’s circadian rhythm, which is considered a physiological stressor that potentially disturbs the homeostasis of bodily systems, including metabolism, the immune system, and gut microbiota, leading to a range of chronic inflammatory health issues^8,9. Research suggests that disruptions in the circadian rhythm could increase renal salt excretion, reduce urine volume, and lower urine pH, thereby promoting the formation of kidney stones¹⁰.

In a cross-sectional study involving 34,190 adult Americans, individuals with normal (7–9 h) and longer sleep durations (> 9 h) had a 17% and 20% lower likelihood, respectively, of developing kidney stones compared to those with shorter sleep durations (< 7 h)¹¹. Another prospective cohort study from China encompassing 512,725 participants identified an increased risk of kidney stones associated with insomnia symptoms (HR 1.11, 95% CI 1.06–1.16) and a short sleep duration (HR 1.13, 95% CI 1.08–1.18)¹². These findings suggest a potential association between sleep traits and urolithiasis, although a detailed causal relationship requires further investigation.

MR analysis is an epidemiological method that utilizes genetic variants as instrumental variables (IVs) for exposure to enhance causal inference^13,14. This design offers two notable advantages over traditional observational studies. First, it minimizes confounding effects, as genetic variations are randomly distributed at conception, independent of environmental or self-selection factors that typically confound the relationship between exposures and outcomes. Second, this method helps mitigate reverse causation since disease occurrence and progression cannot alter germline genotypes^15,16.

Sleep is a multidimensional concept encompassing various traits, such as sleep types, early rising, sleep duration, insomnia, and daytime napping¹⁷. Liu et al. utilized MR to investigate various lifestyle factors, including sleep duration, and found no causal relationship between sleep duration and kidney stones¹⁸. However, their study focused on a single dimension of sleep (duration), neglecting other potentially influential sleep traits, which may independently contribute to urolithiasis risk. Additionally, they provided limited insights into the underlying mechanisms, such as genetic correlations or confounding influences among sleep traits. In contrast, Wang et al. conducted a cross-sectional study showing an association between poor sleep quality and urolithiasis¹⁹. Although this study highlighted important patterns, it relied on subjective self-reports of sleep quality and kidney stone history, which may introduce recall and reporting biases. Moreover, their findings are inherently limited by the inability of cross-sectional designs to establish causal relationships, leaving questions regarding the directionality of the association unanswered. These findings and limitations underscore the need for a comprehensive exploration of the causal links between diverse sleep traits and urolithiasis.

Materials and methods

Study design and data source

We conducted LDSC and two-sample MR to estimate the genetic correlation and causal relationships between the genetically predicted sleep traits and urolithiasis, Fig. 1 presents an overview of the study design. This study including five sleep traits as exposure factors, namely, chronotype (N = 461,658), getting up in the morning (N = 461,658), sleep duration (N = 460,099), insomnia (N = 462,341), and napping during the day (N = 462,400). All data were sourced from the UK Biobank²⁰. Detailed characteristics of these data are presented in S1 Table. To mitigate biases arising from ethnic diversity, only individuals of European ancestry were included in our analysis. For data adjustments, both BMI and other confounders were considered as potential confounding factors. We utilized outcome data from FinnGen version R9 (N = 377,277), sourced from Finnish hospitals and primary health care registers²¹. Within this dataset, we identified 9713 (2.58%) and 1398 (0.37%) individuals with confirmed ICD-10 codes. The specific phenotypic codes used in this study were classified as ‘N14_CALCUKIDUR’ (Kidur) and ‘N14_CALCULOWER’ (Lower). The datasets comprised a total of 376,406 and 368,091 cases, respectively, along with an equal number of 366,693 controls. Further details on these data can be found in S2 Table.

Fig. 1.

The overview study design. IV instrumental variable, GWAS genome wide association study, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, SNP single nucleotide polymorphism, MR mendelian randomization, LDSC linkage disequilibrium score regression, R² R-squared, MR-PRESSO mendelian randomization pleiotropy RESidual Sum and Outlier, IVW inverse variance weighted, F F-statistic.

Genetic instrument selection

In our MR analysis, the IVs used first had to meet three key assumptions: (1) they had to be associated with sleep traits (relevance condition); (2) they had to be unrelated to any confounding factors (exclusion restriction condition); and (3) they could only affect the outcome through their influence on sleep traits (exchangeability condition)²². We employed stringent criteria [P < 5 × 10^–8; linkage disequilibrium (LD) r² < 0.001, clumping window > 10,000 kb; F-statistic > 10] to identify GWAS data for sleep traits, generating genetic instruments and further sieving out strong IVs. The LD r² value was estimated in the European ancestry sample of the 1,000 Genomes Project. The F-statistic was computed using the following formula:

1

where R² represents the exposure variance explained by each instrumental variable separately, and k denotes the number of SNPs. We filtered SNPs with F > 10 to minimize potential bias from weak IVs²³. To ensure that SNPs’ effects on the exposure and outcome corresponded to the same allele, we harmonized the data. Before conducting the MR analysis, we manually screened and excluded IVs associated with confounding factors using PhenoScanner V2 Database (http://www.phenoscanner.medschl.cam.ac.uk/). Detailed information on the IVs is provided in S3–S7 Tables.

Genetic correlation analysis

We employed LDSC to evaluate the genetic correlation (r_g) between sleep traits and urolithiasis. Summary statistics underwent filtration using the HapMap3 reference, excluding non-SNP variants (e.g., indels) and SNPs with ambiguous strand, duplication, or minor allele frequency (MAF) < 0.01. LDSC was utilized to analyze the relationship between test statistics and linkage disequilibrium, aiming to quantify inflation from a genuine polygenic signal or bias. This method enables the evaluation of genetic correlation from GWAS summary statistics and remains unbiased even in cases of sample overlap. We calculated the product of the z-scores of variants from Trait 1 and those from Trait 2, and the genetic covariance was estimated by regressing this product against the LD score. The genetic covariance, normalized by SNP heritability, provides an estimate of genetic correlation^24–26.

Univariate MR analysis

The IVW method was utilized as the primary approach to assess the causal effects between exposure and outcome in MR analysis. Additionally, MR Egger, weighted median, simple mode, and weighted mode methods were considered supplementary and complementary approaches.

Reverse MR analysis

To assess the bidirectional causal relationships between sleep traits and urolithiasis, urolithiasis was designated as the “exposure” while sleep traits were considered the “outcomes” SNPs significantly associated with urolithiasis were selected using the same criteria applied for forward analysis and utilized as IVs. The reverse MR analysis procedure mirrored that of the MR analysis, examining whether sleep traits are causally affected by urolithiasis.

Multivariable MR analysis

MVMR can incorporate multiple exposure factors into a single model, enabling estimation of direct causal associations between exposures and outcomes while mitigating the influence of other factors²⁷. In our two-sample analysis, we observed some SNP overlaps among sleep traits, prompting the utilization of multivariable MR analysis to adjust for confounding factors and assess the direct impact of individual exposures on outcomes, ensuring that these effects were not mediated through other exposures. Weighted linear regression based on the IVW and the MR‒Egger methods was employed to infer causal effects in multivariable MR analysis.

Pleiotropy and sensitivity analysis

In MR‒Egger regression, the intercept describes the average pleiotropic effect of the IV. When the intercept does not significantly differ from zero (P > 0.05), it suggests no horizontal pleiotropy. The MR pleiotropy residual sum and outlier (MR-PRESSO) test aims to identify and correct outliers in IVW linear regression, comprising the MR-PRESSO global test, MR-PRESSO outlier test, and MR-PRESSO distortion test²⁸. Additionally, heterogeneity assessment was conducted using IVW and MR‒Egger methods. To assess the robustness and consistency of the results, a leave-one-out analysis was performed for each SNP. Visualization of results was conducted using scatter plots, funnel plots, and forest plots.

Statistical analysis

All MR analyses were conducted as two-sided tests utilizing the TwoSampleMR, MR-PRESSO, and Mendelian randomization packages. LDSC regression analysis was used to assess the genetic correlation between sleep traits and urolithiasis. Forest plots were generated using the ‘forestploter’ package. Statistical analyses were performed using R software (version 4.3.1).

Significance was assessed using both unadjusted and adjusted p-values. Results were considered statistically significant if p < 0.05 and adjusted p < 0.1, while results with p < 0.05 but adjusted p > 0.1 were deemed nominally significant. Adjusted p-values were derived using the Benjamini-Hochberg (BH) method, which corrects the p-values to control the false discovery rate (FDR). This adjustment ensures that the proportion of false positives is kept under control, enhancing the robustness of the findings in the presence of multiple comparisons.

Results

LDSC regression analysis

We conducted LDSC regression analysis to assess the genetic correlation between various sleep characteristics and urinary stones. The results are summarized in Table 1 and illustrated in Fig. 2. The analysis indicated a positive genetic correlation between insomnia (r_g = 0.082, P = 0.017, adjusted P = 0.085) and upper urinary stones, which met the significance threshold (p < 0.05 and adjusted p < 0.10). For insomnia and lower urinary stones, a genetic correlation of r_g =0.045 was found, but it did not reach statistical significance (P = 0.467, adjusted P = 0.641). Similarly, for getting up in the morning and lower urinary stones, the genetic correlation was r_g = 0.029, with no significant association observed (P = 0.673, adjusted P = 0.673). No significant genetic correlations were found for the other sleep traits with urinary stones (P > 0.05, adjusted P > 0.10 for all).

Table 1.

LDSC analysis results of sleep traits and urolithiasis.

Trait 1	Trait 2	rg	SE	P	Adjusted P
Chronotype	Kidur	0.063	0.034	0.063	0.158
	Lower	0.104	0.063	0.100	0.253
Getting up in morning	Kidur	0.000	0.032	0.993	0.993
	Lower	0.029	0.068	0.673	0.673
Sleep duration	Kidur	-0.005	0.032	0.864	0.993
	Lower	-0.103	0.063	0.101	0.253
Insomnia	Kidur	0.082	0.034	0.017*	0.085*
	Lower	0.045	0.062	0.467	0.641
Nap during day	Kidur	0.049	0.034	0.144	0.240
	Lower	-0.039	0.059	0.513	0.641

LDSC linkage disequilibrium score regression, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, r_g genetic correlation, SE standard error, P p-value, Adjusted P false discovery rate adjusted p-value.

*P<0.05, *Adjusted P<0.10.

Fig. 2.

Genetic correlation between sleep traits and urolithiasis. Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, *P<0.05.

Univariate MR analysis

According to the stringent criteria for strong IVs selection, the chronotype, getting up in morning, sleep duration, insomnia, and nap during day IVs exhibited 74, 67, 55, 32, and 29 independent IVs, respectively (S3–S7 Tables). All IVs had F-statistic values exceeding 10, ranging between 29.84 and 224.46. PhenoScanner V2 database was used to manually exclude SNPs associated with BMI and other potential confounders, and the remaining SNPs were employed in subsequent MR analysis.

The MR analysis results (Table 2; Fig. 3) revealed a positive causal relationship between genetically determined insomnia and an increased risk of lower urinary tract stones (IVW OR = 5.91, 95% CI 1.52–22.98, P = 0.010, Adjusted P = 0.030). In contrast, early rising (getting up in morning) exhibited an inverse causal relationship with lower urinary tract stones (IVW OR = 0.29, 95% CI 0.11–0.76, P = 0.012, Adjusted P = 0.030). Although this causative effect was less evident in the weighted median, simple mode, and weighted mode analyses, the effect estimates were consistent with those observed in the IVW analysis (S8 Table). However, no statistically significant causal associations were found for other sleep traits, including chronotype, sleep duration, and daytime napping for both upper and lower urinary stones, as well as the combined outcome. The P and adjusted P-values for these traits did not reach the significance threshold, indicating a lack of strong evidence for their causal effect on the risk of urolithiasis.

Table 2.

Univariable MR-IVW results of sleep traits on risk of urolithiasis.

Exposure	Outcome	No. of SNPs	OR	95% CI	P	Adjusted P
Chronotype	Kidur	67	0.95	0.73–1.24	0.730	0.908
	Lower	63	0.62	0.32–1.20	0.156	0.195
	Combined	66	0.98	0.76–1.28	0.896	0.896
Getting up in morning	Kidur	63	1.03	0.65–1.62	0.908	0.908
	Lower	62	0.29	0.11–0.76	0.012*	0.030*
	Combined	62	1.05	0.66–1.66	0.845	0.896
Sleep duration	Kidur	46	1.27	0.69–2.36	0.441	0.735
	Lower	52	1.69	0.60–4.73	0.320	0.320
	Combined	51	1.12	0.63–2.01	0.694	0.896
Insomnia	Kidur	30	1.33	0.76–2.32	0.315	0.735
	Lower	29	5.91	1.52–22.98	0.010*	0.030*
	Combined	29	1.29	0.70–2.37	0.418	0.896
Nap during day	Kidur	71	0.80	0.50–1.29	0.360	0.735
	Lower	70	0.39	0.13–1.20	0.102	0.170
	Combined	69	0.80	0.49–1.30	0.362	0.896

MR mendelian randomization, IVW inverse variance weighted, SNP single nucleotide polymorphism, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, OR odds ratio, CI confidence interval, P p-value, Adjusted P false discovery rate adjusted p-value.

*P<0.05, *Adjusted P<0.10.

Fig. 3.

Associations of genetic liability to sleep traits with risk of urolithiasis. SNP single nucleotide polymorphism, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, CI confidence interval, OR odds ratio.

We performed heterogeneity and pleiotropy assessments for variables with preliminary causal relationships using Cochran’s Q test, MR-Egger intercept test, and MR-PRESSO analysis. Cochran’s Q test showed no evidence of heterogeneity for early rising (IVW: Q = 74.03, df = 66, P = 0.233; MR-Egger: Q = 71.75, df = 65, P = 0.264) and insomnia (IVW: Q = 24.98, df = 31, P = 0.769; MR-Egger: Q = 23.41, df = 30, P = 0.798) (Table 3). In the MR-Egger intercept test, no evidence of horizontal pleiotropy was observed for early rising (Intercept = − 0.033, SE = 0.023, P = 0.155) and insomnia (Intercept = 0.026, SE = 0.021, P = 0.220). Additionally, the MR-PRESSO global test did not identify any significant outliers for early rising (P = 0.246) and insomnia (P = 0.770) (Table 4).

Table 3.

Heterogeneity results in univariable MR analysis using Cochran’s Q test.

Exposure/outcome	MR-IVW			MR-Egger
	Q	Q df	Q pval	Q	Q df	Q pval
Chronotype
Kidur	135.98	72	7.86E−06*	135.33	71	6.57E−06*
Lower	101.06	72	1.35E−02*	101.06	71	1.10E−02*
Combined	135.98	72	7.86E−06*	135.33	71	6.57E−06*
Getup in morning
Kidur	130.32	66	4.08E−06*	130.32	65	2.83E−06*
Lower	74.03	66	2.33E−01	71.75	65	2.64E−01
Combined	130.32	66	4.08E−06*	130.32	65	2.83E−06*
Sleep duration
Kidur	114.42	54	3.13E−06*	112.30	53	3.80E−06*
Lower	57.24	54	3.56E−01	52.59	53	4.90E−01
Combined	114.42	54	3.13E−06*	112.30	53	3.80E−06*
Insomnia
Kidur	40.86	31	1.11E−01	36.96	30	1.78E−01
Lower	24.98	31	7.69E−01	23.41	30	7.98E−01
Combined	40.86	31	1.11E−01	36.96	30	1.78E−01
Nap during day
Kidur	98.54	78	5.80E−02	98.48	77	5.00E−02
Lower	66.23	78	8.26E−01	63.31	77	8.69E−01
Combined	98.54	78	5.80E−02	98.48	77	5.00E−02

MR mendelian randomization, IVW inverse variance weighted, Q cochran’s Q statistic, df degrees of freedom, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract.

*Q pval < 0.05.

Table 4.

Investigating directional pleiotropy in univariable MR analysis using the MR-Egger intercept and MR-PRESSO test.

Exposure/outcome	MR-Egger intercept			MR-PRESSO
	Intercept	SE	Pval	Global test Pval	Distortion test Pval
Chronotype
Kidur	− 0.004	0.007	0.562	< 0.001***	0.194
Lower	<− 0.001	0.015	0.999	0.014*	0.761
Combined	− 0.004	0.007	0.562	< 0.001*	0.186
Getup in morning
Kidur	< 0.001	0.012	0.999	< 0.001***	0.003**
Lower	− 0.033	0.023	0.155	0.246	NA
Combined	< 0.001	0.012	0.999	< 0.001***	0.071
Sleep duration
Kidur	− 0.013	0.013	0.322	< 0.001***	0.077
Lower	0.049	0.023	0.036*	0.332	NA
Combined	− 0.013	0.013	0.322	< 0.001***	0.223
Insomnia
Kidur	0.016	0.009	0.085	0.097	NA
Lower	0.026	0.021	0.220	0.770	NA
Combined	0.016	0.009	0.085	0.165	NA
Nap during day
Kidur	− 0.002	0.008	0.825	0.054	NA
Lower	− 0.030	0.018	0.091	0.830	NA
Combined	− 0.002	0.008	0.825	0.027*	NA

MR mendelian randomization, SE standard error, MR-PRESSO MR Pleiotropy RESidual Sum and Outlier, Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract.

*P<0.05, **P<0.01, ***P<0.001.

Overall, the absence of significant heterogeneity or horizontal pleiotropy, as indicated by Cochran’s Q test, MR-Egger intercept, and MR-PRESSO analysis, suggests that our MR estimates are unlikely to be biased by these factors, thereby supporting the robustness of our findings. Furthermore, scatter plots, funnel plots, forest plots, and leave-one-out analysis of the two-sample MR results, as shown in Figs S1–S4, further corroborate the robustness of these findings.

In the reverse MR analysis, we assessed the effects of different sleep traits on urolithiasis. The weighted median analysis showed a statistically significant association between lower urinary stones and “chronotype” (P = 0.018, adjusted P = 0.090), which reached the significance threshold. However, due to the limited power of the weighted median method and the lack of significant results in other models, particularly the IVW model, the causal relationship remains uncertain. Additionally, for lower urinary stones and “getting up in the morning,” the simple mode analysis produced a P value of 0.043, but the adjusted P value of 0.215 did not reach significance. No significant associations were observed for other sleep traits across all models, indicating that most sleep traits do not exhibit a significant reverse causal relationship with urolithiasis. More detailed data can be found in S10 Table.

Multivariable MR analysis

In the multivariable analysis adjusted solely for BMI, a direct causal effect was observed among chronotype, insomnia, daytime napping, and lower urinary tract stones, Specifically, chronotype showed a marginal association with lower urinary tract stones (MVMR-IVW OR = 1.21, 95% CI 0.98–1.49, P = 0.010, Q pval = 0.141), and insomnia was associated with a significantly increased risk (MVMR-IVW OR = 3.29, 95% CI 1.25–8.66, P = 0.016, Q pval = 0.525). In contrast, daytime napping showed a protective effect (MVMR-IVW OR = 0.32, 95% CI 0.14–0.77, P = 0.010, Q pval = 0.607). In the MVMR-Egger analysis, chronotype and daytime napping showed similar results, with chronotype demonstrating an OR of 1.19 (95% CI 0.96–1.47, P = 0.007, Q pval = 0.160, P_inter=0.069) and daytime napping showing an inverse association (MR-Egger OR = 0.31, 95% CI 0.13–0.73, P = 0.007, Q pval = 0.622, P_inter=0.142).

When adjusted for BMI along with other sleep characteristics, the direct causal relationship between insomnia, daytime napping, and lower urinary tract stones persisted. Insomnia continued to show a strong association with increased risk (MVMR-IVW OR = 3.02, 95% CI 1.07–8.53, P = 0.037, Q pval = 0.135), and daytime napping remained protective (MVMR-IVW OR = 0.28, 95% CI 0.12–0.65, P = 0.003, Q pval = 0.135). The MR-Egger method confirmed the associations of daytime napping and lower urinary tract stones with similar estimates (MR-Egger OR = 0.25, 95% CI 0.11–0.58, P = 0.001, Q pval = 0.172, P_inter=0.111).

No significant evidence of heterogeneity was found in the above results, as indicated by Q p-values consistently above the 0.05 threshold in both MVMR-IVW and MR-Egger models, supporting the robustness of the findings. After examining heterogeneity, pleiotropy assessment showed no significant indications of directional pleiotropy, with Pinter values consistently above 0.05. Detailed statistical data for all exposures and outcomes can be found in Supplementary Table S9.

Insomnia consistently exhibited a causal association with an increased risk of lower urinary tract stones in both univariable and multivariable analyses. Notably, while univariable MR displayed an inverse association between early rising and lower urinary tract stones, this significance was no longer observed in the multivariable analysis. In the multivariable MR analysis, chronotype exhibited a positive causal association with lower urinary tract stones when only BMI was adjusted for. However, this association was no longer statistically significant after adjusting for both BMI and other sleep characteristics. Daytime napping consistently demonstrated a robust inverse causal effect in both MVMR1 and MVMR2 analyses.

Our MR analysis did not find evidence of a causal relationship between sleep traits and either upper urinary tract stones or overall urinary stone formation. Detailed results are available in Table 2; Figs. 3 and 4, and Supplementary Table S9.

Fig. 4.

MVMR results of sleep traits on risk of urolithiasis. Kidur calculus of kidney and ureter, Lower calculus of lower urinary tract, MR Mendelian randomization, BMI body mass index, MVMR1 multivariable mendelian randomization analysis adjusting for body mass index, MVMR2 multivariable mendelian randomization analysis adjusting for body mass index and other sleep traits, CI confidence interval, OR odds ratio, IVW inverse variance weighted.

Discussion

Currently, detailed research regarding the potential causal relationships between different sleep traits and urolithiasis is lacking. Leveraging data from the UK Biobank and FinnGen, we utilized MR methods to investigate potential causal associations. We discovered a significant positive correlation between a genetic predisposition to insomnia and the risk of lower urinary tract stones. For other sleep traits, our analysis produced mixed results. Early rising demonstrated a protective effect against lower urinary tract stones in univariate MR analysis, though this significance did not persist in multivariable analysis. Chronotype was found to be a potential detrimental factor in the multivariable MR analysis adjusted solely for BMI. However, this association did not remain significant when adjusting for both BMI and other sleep traits. Similarly, daytime napping consistently showed a protective effect in both univariable and multivariable MR analyses, which could suggest that this habit may lower the risk of lower urinary tract stones. Moreover, the LDSC results showed no genetic correlation between these sleep traits and the outcome, indicating that the causal relationships are independent of shared genetic factors, free from genetic confounding, and potentially driven by environmental or lifestyle influences.

Despite previous observational studies indicating an association between different sleep durations and the risk of kidney stones¹¹, our research did not reveal consistent causal relationships. This could be due to our utilization of urinary tract stone data from the latest version of FinnGen (version R9) for analysis. Additionally, our study primarily aimed to explore the potential impact of various sleep traits on urinary tract stone risk without detailing the effects of different subgroups of sleep duration on the outcome, which might contribute to differing outcomes. A cross-sectional study using data from the West China Natural Population Cohort Study (WCNPCS) involving 34,437 eligible participants showed a significant increase in the risk of urolithiasis associated with sleep disturbance¹⁹. Moreover, an observational study based on a Chinese population suggested that insomnia might partially contribute to kidney stone occurrence¹². LDSC analysis results showed a genetic correlation between insomnia and upper urinary tract stones but no causal relationship between sleep traits and upper urinary tract stones was found in our study. Notably, a previous MR study also did not find a causal relationship between insomnia and kidney stones¹⁸. Thus, for a more in-depth exploration of the relationship between insomnia and urolithiasis, larger-scale studies encompassing diverse populations are warranted in the future. Interestingly, in univariate analysis, early rising was identified as a protective factor against lower urinary tract stones, whereas in multivariable analysis, chronotype and daytime napping were protective factors. This change could be due to the potential mild impact of manually removing confounders in two-sample analyses, with the multivariable MR model robustly estimating the direct exposure-outcome relationship by incorporating BMI. Furthermore, in the analysis of chronotype as the exposure, the results simultaneously including BMI and other sleep characteristics lacked statistical significance compared to the multivariable model including BMI alone, suggesting that protective role of chronotype against lower urinary tract stones might operate through other sleep traits. Both insomnia and daytime napping exhibited consistent results across different multivariable models, where BMI did not seem to mediate their association.

Although the potential mechanisms by which insomnia leads to urolithiasis remain unclear, certain pathways may partially explain this phenomenon. First, prolonged insomnia may lead to emotional changes, even triggering depression²⁹, which is believed to be a risk factor for kidney stones³⁰. Evaluating the adverse effects of insomnia on mental and physical health and its association with stones is crucial for clinical treatment and guidance³¹. Second, during sleep deprivation, substances such as catecholamines³², CRP^33,34, IL-6³⁵, TNF-α³⁶, and IL-1β increase in the body, indicating a heightened inflammatory profile, which might be linked to the occurrence of urolithiasis. Third, insomnia might induce stone formation by decreasing immunity and increasing the risk of infections³⁷. Fourth, research findings suggest that insomnia could alter the gut microbiota, leading to changes in metabolic patterns and potentially triggering metabolic syndrome, where metabolism, BMI, and obesity are common risk factors for stone formation³⁸. Additionally, the kidneys have specific diurnal rhythms for excretion; by disrupting this rhythm, insomnia could result in increased urinary salt excretion, reduced urine volume and decreased urine pH¹⁰, possibly representing another potential mechanism by which insomnia leads to urolithiasis.

The causes of insomnia are diverse. The fast pace of modern life, jet lag, and pressure from work and school often lead to widespread feelings of anxiety and suppression. Additionally, social demands result in 15–33% of the global workforce needing to engage in shift work³⁹, particularly individuals in health care professions⁴⁰. These factors may significantly impact the prevalence of abnormal sleep patterns in society. This study explored a potential positive causal relationship between insomnia and urolithiasis, suggesting that individuals with insomnia may be more prone to developing urinary tract stones than those with normal sleep patterns. Therefore, timely correction of insomnia may hold significant value in preventing urolithiasis. Moreover, our research findings indicate that daytime napping could be a protective factor against urolithiasis. In other words, moderate daytime napping may aid in preventing stone formation, providing an additional perspective on the prevention of stones.

This study has several strengths. First, MR analysis minimizes the impact of confounding factors and reverse causation effects more effectively than traditional observational studies, thereby strengthening causal inferences. Additionally, our use of multivariable MR analysis mitigated potential bias arising from the pleiotropic effects of sleep-related traits on the MR outcomes⁴¹. Second, by exclusively analyzing data from the European population, we minimized potential biases resulting from population stratification. Third, given the scarcity of observational studies on the relationship between sleep characteristics and urolithiasis, our conclusions provide some evidence based on MR that may guide further relevant research. Nevertheless, our study has inherent limitations. While our findings suggest that intervening in insomnia might have a clinically positive impact on urinary tract stones, a comprehensive understanding of the true effects would necessitate large-scale clinical trials. Moreover, the generalizability of our analysis results based on the European population to other ethnicities remains unclear and should be considered cautiously when extrapolating conclusions. Regrettably, we did not find a causal relationship between sleep characteristics and upper urinary tract stones. Subsequent research will explore this relationship further using data from diverse sources and undertake more detailed stratification to elucidate potential connections.

Conclusion

In conclusion, this MR study revealed a positive association between insomnia and lower urinary tract stones, while daytime napping exhibited an inverse trend. This finding underscores the potential impact of insomnia on the risk of urolithiasis, suggesting that addressing sleep issues could hold significant importance in both the prevention and treatment of urinary tract stones. However, this association warrants validation in relevant clinical studies and further exploration of the underlying biological mechanisms.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

BMI

Body mass index

CI

Confidence interval

GWAS

Genome-wide association study

IV

Instrumental variable

IVW

Inverse variance weighted

LDSC

Linkage disequilibrium score regression

MAF

Minor allele frequency

MR

Mendelian randomization

MVMR

Multivariable Mendelian randomization

OR

Odds ratio

SE

Standard error

SNP

Single nucleotide polymorphism

UK Biobank

United Kingdom Biobank

R²

R-squared

FinnGen

Finnish genealogy research project

BH

Benjamini–Hochberg method

FDR

False discovery rate

WCNPCS

West China natural population cohort study

CRP

C-reactive protein

IL-6

Interleukin-6

TNF-α

Tumor necrosis factor-alpha

IL-1β

Interleukin-1 beta

Author contributions

M.Y, Z.Z.and C.L. came up with the concept for the study. J.H. and K.W. formulated the theoretical framework and carried out the computational work. C.L. validated the analytical techniques. All authors engaged in discussions regarding the results and played a role in writing the final manuscript.

Funding

The research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The GWAS summary statistics for sleep traits are publicly available from the OpenGWAS database (https://gwas.mrcieu.ac.uk/datasets). The summary statistics for urolithiasis are from the FinnGen consortium (https://www.finngen.fi/en/). Detailed information can be found in Supplementary Tables S1 and S2. All datasets are publicly accessible, with no additional permissions required.

Ethics statement

Competing interests

The authors declare no competing interests.

Ethical approval

and informed consent were secured from the original GWAS, and no additional ethics statement was necessary for this study.