Lifestyle factors and the risk of gallstones: results from the national health and nutrition examination survey 2018–2020 and mendelian randomization analysis

Abstract Objectives This study aimed to investigate the relationship between lifestyle and gallstones. Materials and Methods We performed an observational study using the 2018–2020 National Health and Nutrition Examination Survey (NHANES). Univariate and multivariate-adjusted logistic regression analyses were performed to assess the correlations between lifestyle factors and gallstone risk. Second, Mendelian randomization (MR) was applied to decrease the causal relationship between lifestyle factors and gallstones. Results This observational study enrolled 11,970 individuals. The risk of gallstones was found to increase with increased sitting time (odds ratio (OR) 1.03, 95% CI 1.00–1.05, p = 0.02). In contrast, the risk of gallstones was found to decrease with recreational activity (OR 0.50, 95% CI 0.29–0.87, p = 0.02). The results of the MR also showed that time spent watching television (OR 1.646; 95% CI 1.161–2.333, p = 0.005) and physical activity (OR 0.953, 95% CI 0.924–0.988, p = 0.003) remained independently causally associated with gallstones. Conclusions Prolonged sitting increases the risk of gallstones, whereas recreational activity reduces the risk. These findings need to be verified in further prospective cohort studies with larger sample sizes and longer follow-up periods.


Introduction
Gallstone disease is a common global condition that causes more hospitalizations than other digestive diseases. About 5-25% of adults in the Western world have gallstone disease [1]. In 2015, there were 1.5 million medical visits for gallstone-related conditions in the United States [2]. With the development of the economy and the progress of society, people's lifestyles are becoming increasingly colorful. Lifestyle factors included the time to fall asleep, sleep time, travel patterns, smoking, drinking, and sedentary time. According to the available data, the risk factors for gallstones include age, female sex, pregnancy, lack of physical activity, obesity, over-nutrition, factors related to metabolic syndrome, and abnormal mutations of the hepatic cholesterol transporter ABCG8 [3]. It means that lifestyle may affect the formation of gallstones. Sleep quality also affects leptin, insulin, and other hormones and metabolic variables, thus increasing the possibility of obesity [4]. In other words, sleep disorders may also be a risk factor for gallstone disease. Sleep can be evaluated from a number of perspectives, including duration, time to fall asleep, regularity, timing, and presence of sleep disorders [5]. However, few studies have analyzed the correlation and causation between lifestyle factors and gallstones. Therefore, this study aimed to analyze the relationship between lifestyle factors and gallstones using data from patients with gallstones in the National Health and Nutrition Examination Survey (NHANES).

Study design and participants in NHANES
The NHANES is a large, cross-sectional study designed to assess the health and nutritional status of children and adults in the United States. This investigation began in 1960 and consisted mainly of interviews and medical examinations (https://www.cdc.gov/nchs/nhanes/). In this study, 14,498 adult participants were selected from the National Health and NHANES database between 2017 and 2020. A total of 1457, 196, 19, 84, three, and 769 people were excluded due to lacking gallstone and fatty liver data, sleep data, marriage and education information, exercise data, smoking data, and physical examination indicator data, respectively. Therefore, a total of 11,970 people were included in this study (Supplementary Figure 1). The data were weighted according to the NHANES database prompts and represent a population of 213,752,934 in the United States. This study was approved by the Institutional Review Board of the National Center for Health Statistics, and all participants provided written informed consent.

Definition of variables in NHANES
Sociodemographic and lifestyle information, including age, sex, race, gallstones, education, marital status, smoking, sitting time, physical activity, work status, and sleep status, were collected using professional and standardized family questionnaires.
Trained medical staff performed laboratory examinations at a mobile examination center.
Gallstones were defined by the question, 'Have a doctor or other health professional ever told you that you had gallstones?' Nonalcoholic fatty liver disease (NAFLD) cases were ascertained based on the controlled attenuation parameter (CAP) [6]. Each person's body mass index (BMI) was calculated as their weight in kilograms divided by the square of their height in meters. The presence of diabetes mellitus (DM) was determined based on the patient's medical history and blood glucose value [7]. The BARD scoring system, which includes the BMI, diabetes, and AST/ALT ratio, was used. In this scoring system, a BMI !28 kg/m 2 was scored as one point, AST/ALT (AAR) !0.8 as two points, and those with type 2 diabetes as one point [8]. The body roundness index (BRI) score was defined as the waistline/BMI [9]. Ethnicity was divided into four groups: Black, White, Mexican American, and others. Education was divided into college graduates or higher, high school graduates or higher, and others. Marital status was classified as never married, married, living with a partner, divorced, or separated. According to the question, 'Have you smoked at least 100 cigarettes in your life?' and 'Do you smoke now?', participants were divided into never smoked, former smokers, and current smokers groups. Sitting time (hours/day) was defined according to the question, 'How much time do you usually spend sitting in a day?'. Walking or cycling was defined according to the question, 'In a normal week, do you walk or ride a bike to and from different places for at least ten minutes continuously?'. Work activity was defined according to the question, 'Does your work include moderate or vigorous activities?'. Recreational activity was defined according to the question, 'Do you engage in moderate-intensity or vigorous-intensity recreation in your daily routine?'. Trouble sleeping was defined according to the question, 'Have you ever told a doctor or other health professional that you have trouble sleeping?'. Sleep and bedtime were recorded based on the questionnaire results.
Study design and sample source in the mendelian randomization (MR) analysis Three assumptions are essential for the validity of the MR: (1) the instrument variable (genetic variant) is associated with the risk factor, (2) the instrument variable is not associated with confounders, and (3) the instrument variable influences the outcome only through the risk factor. In this study, we used a two-sample MR design to explore whether lifestyle had a causal effect on the occurrence of gallstones. Univariate MR analysis was designed to study the correlation between a single exposure and gallbladder stones, while multivariate MR analysis was designed to compare the independent effects of multiple exposures on gallbladder stones. Supplementary Table 1 provides an overview of the sample information.

Instrumental variable in MR
Single nucleotide polymorphisms (SNPs) (p < 5 Â 10-8) were significantly correlated with exposure as instrumental variables. Subsequently, we removed the chain imbalance reaction with thresholds of r2 < 0.001 and Kb >10,000 [10] and extracted the effect estimates of selected instrumental variables from the 'finn-b-K11_CHOLELITH' dataset. To address the MR second hypothesis, we removed SNPs associated with gallstones and confounding factors through the PhenoScannerV2 database (http://www.phenoscanner. medschl.cam.ac.uk/) [11]. Palindromic SNPs refer to SNPs whose alleles correspond to nucleotides that pair with each other in a DNA molecule and intermediate allele frequencies denote the allele frequencies between 0.01 and 0.30 [12].
The selection of instrumental variables should remove palindromic SNPs. To satisfy the first hypothesis, the R2 was used to explain the proportion of trait variance. The R2 was calculated using: is the effect size, SE is the standard error, N is the sample size, and MAF is the minimum allele frequency for each SNP. We then calculated an F statistic to assess the total strength of the selected SNPs in explaining phenotypic variation as follows: F ¼ b 2 /SE 2 . F > 10 indicates that the instrumental variable SNPs are highly effective and can reduce any potential bias, while F 10 indicates that the SNP is a weak instrumental variable [13].

MR analysis
For the univariate MR analysis, we used three MR methods to analyze the causal effect of exposure on outcomes, with the inverse variance weighted (IVW) method as the primary MR analysis and the MR Egger and weighted media methods as supplements [14,15]. We further assessed the independent causal effects of lifestyle on gallstone outcomes using multivariate IVW. The effect sizes for the MR analysis are shown as odds ratio (OR) values and 95% confidence intervals. Pleiotropy tests using MR Egger and MR Pleiotropy RESidualSum and Outlier (MR-PRESSO) showed that pleiotropy was not observed at p > 0.05. MR Egger and IVW in Cochran's Q statistic were used for the heterogeneity analysis, and there was no significant heterogeneity when p > 0.05. In addition, we used the 'leave-one-out' sensitivity analysis to demonstrate that the causal effect of exposure on gallstone outcomes was not affected by individual SNPs. Statistical power calculations for MR were performed using the mRnd website (https://shiny.cnsgenomic.com).

Statistical analysis
All analyses were performed accounting for the complex survey design using the appropriate subsample weights, strata, and primary sampling units as per the recommendations of NHANES (CDC/NCHS, 2018). The mean (standard error) was used to describe the normal or approximately normal distribution of the measurement data. Survey-weighted univariate and multivariate logistic regression analyses were used to analyze possible risk factors for gallstones. All analyses were conducted by the 'nhanesR' and 'survey' packages in R (version 4.2.1, www.r-proiect.org).
As for MR analyses, all analyses were conducted by 'devtools', 'TwoSampleMR', 'LDlinkR', and 'MRPRESSO' packages in R, and a p-value less than 0.05 was considered statistically significant evidence of causation.

Patients in NHANES
A total of 11970 cases were included in this study, including 5921 (49.47%) males and 6049 (50.53%) females. There were 1214 cases of gallstones, and the incidence of gallstones was 10.67%; patients with gallstones were divided into the stone group, and others were divided into the control group. The gallstone group was older (46.95 vs. 56.13, p < 0.001) and tended to sit for longer (348.34 vs. 374.40, p < 0.001). There were no statistically significant differences in sleep duration between the two groups. The BMI, BRI, and BARD values in the gallstone group were higher than those in the control group (29.22 vs. 32.93, p < 0.001; 5.44 vs. 6.96, p < 0.001; 1.67 vs. 1.94, p < 0.001) ( Table 1).

Univariate logistic regression analysis of the gallstone risk factors in NHANES
As shown in Table 2, the univariate analysis showed that compared to married individuals, age, sitting time, BMI, BRI, and BARD were risk factors for gallstones (p < 0.001). Additionally, NAFLD, trouble sleeping, and DM were risk factors for gallstones (p < 0.001). Walking or cycling was a protective factor (OR ¼ 0.62, p ¼ 0.002). White, Mexican, and other Americans had 1.930-fold, 1.488-fold, and 1.399-fold higher risks of gallstones, respectively, than Black Americans (p < 0.05). Compared with the subjects who had never married, those who were married or living with a partner or had divorced or separated had a higher risk of gallstones (p < 0.001). The risk of gallstones was 0.76 times higher in subjects who undertook moderate-and vigorous-intensity activity than in those with no activity. Subjects who undertook vigorous recreational activity and moderate and vigorous recreational activities had a 0.300-fold and 0.501-fold higher risk of gallstones, respectively, than those without recreational activity. Moreover, former smoking was found to be a risk factor for gallstones compared with never smoking. However, sleep duration, bedtime, and educational attainment were not factors that contributed to gallstones.
Multivariate logistic regression analysis of the gallstone risk factors in NHANES As shown in Table 3, the multivariate analysis demonstrated that age was an independent risk factor for gallstones (p < 0.001). In females, an average increase of 1 h in sitting time was associated with a 1.03-fold increase in gallstone risk (p ¼ 0.02). For females, the risk of gallstones was 1.45 times higher in subjects with NAFLD than in those without NAFLD (p < 0.001), and the risk of gallstones was 1.62 times higher in subjects who had trouble sleeping than in those without trouble sleeping. Moreover, White Americans, Mexican Americans, and other Americans had a higher risk of developing gallstones than Black Americans. For females, divorced or separated subjects had a 0.63-fold higher risk of gallstones than those who had never married. For females, the risk of gallstones was 1.82 times higher in subjects with diabetes than in those without diabetes (p < 0.001), and the risk of gallstones was 1.42-fold and 1.54-fold higher in subjects who formerly and now smoked, respectively, than in those who never smoked (p < 0.05). For males, the risk of gallstones was 0.26 times higher in subjects who undertook vigorous recreational activity than in those with no recreational activity (p < 0.001). However, the risk of gallstones was 1.73 times higher in subjects with moderate recreational activity than in those with no recreational activity (p ¼ 0.02). For males, the risk of gallstones was 0.39 times higher in subjects with moderate work activity than in those with no work activity (p < 0.001), but for females, the risk of gallstones was 1.43-fold and 1.50-fold higher in subjects with moderate work activity and both moderate and vigorous work than in those with no work activity (p < 0.05).

Univariate MR analysis
Seven lifestyle factors had six or more non-palindromic SNPs with P-values less than 5 Â 10 À 8. The F values ranged from 169.8 to 2492.8 (Supplementary Table 2). As shown in Table 4, the amount of time spent watching television (p < 0.0001; OR ¼ 1.324; 95%CI: 1.693-2.166) and amount smoked per day (p ¼ 0.041; OR ¼ 3.577; 95%CI: 1.071-11.949) were positively associated with a higher incidence of gallstones, while physical activity (p ¼ 0.033; OR ¼ 1.163; 95%CI: 1.013-1.337) was negatively correlated with gallstones, and no significant pleiotropy and heterogeneity were observed (Table 5). Insomnia, time spent driving, time spent using a computer, and smoking were not associated with gallstones (Table 4).

Multivariate MR analysis
In the multivariate MR analysis, the time spent watching television, smoking per day, physical activity, insomnia, time spent driving, time spent using computers, and ever smoked corrected each other. The time spent watching television (p ¼ 0.005; OR ¼ 1.646; 95%CI: 1.161-2.333) and physical activity (p ¼ 0.003; OR ¼ 0.953; 95%CI: 0.924-0.988) still maintained a causal relationship with gallstones ( Figure 1).

Discussion
In this study, we conducted observational and MR analyses of nationally representative data from NHANES to 2017-2020 to explore the association between lifestyle factors and the risk of gallstones. These results indicate that different lifestyles are differently associated with gallstone risk. Smoking, diabetes, sedentary behavior, and other factors were positively correlated with gallstone formation, while physical activity was inversely correlated with gallstone formation. The MR analysis showed an independent causal relationship between the time spent watching television, physical activity, and gallstone risk. The amount of time spent watching television was a risk factor for gallstone formation, while physical activity was a protective factor for gallstone formation. Combining the results of the NHANES database and MR analyses for people at a high risk of gallstones, especially females, reducing sedentary time, enhancing physical exercise, and improving sleep quality will reduce the risk of gallstones.
It is well known that gallstones are a chronic disease that increases in prevalence with age. This is consistent with the results of this study, which showed that age was a risk factor for gallstones. There are two possible reasons for this finding. First, with increasing age, the body's metabolism gradually slows. Another possibility is a decline in physical activity among middle-aged and older adults. People of different races in the United States also have different eating habits and lifestyles. The prevalence has steadily increased with the introduction of a high-calorie, high-carbohydrate, and lowfiber diet (excess nutrition) and with increasing rates of physical inactivity [16,17]. Most Blacks in the United States have relatively poor working conditions, which means that White Americans have better diets than the rest of the population; this is likely to contribute to the higher prevalence of gallstones. Studies have shown that the incidence of symptomatic or previous gallstones is 50% in females and 20% in males [18,19]. Estrogen significantly increases the incidence of cholesterol stones [20]. Cirillo et al. found an association between estrogen therapy and the risk of gallstones and cholecystectomy in two randomized, double-blind, placebo-controlled trials [21]. Progesterone is a steroid hormone that promotes gallstone formation by impairing gallbladder contraction and reducing the emptying rate [22].
Some studies have suggested a causal relationship between obesity and gallstones, but there was no strong association between waist circumference and risk factors for gallstones [23]. However, in the analysis of this paper, after correcting for other confounding factors, BRI and BMI were found to be risk factors for gallstones, and the OR of BRI was much greater than that of BMI. The higher risk of gallstones in obese people is due to altered metabolic factors, liver secretion of supersaturated bile, dyslipidemia, intestinal and gallbladder motility, gallbladder stasis, decreased bile acid secretion, cholesterol crystallization, and precipitation, and supersaturated gallbladder bile [24]. People with high BRI values have higher levels of visceral fat, so we believe that BRI should be used as an indicator of risk in our daily lives. NAFLD is defined as limited fat accumulation in the liver in the absence of excessive alcohol consumption and secondary causes of liver disease. It has been linked to obesity, hyperlipidemia, and type 2 diabetes [25]. Gallstones were strongly correlated with the above factors [26]. In addition, insulin resistance and metabolic syndrome were important features of the pathogenesis of gallstones [27]. There was a common pathogenesis between gallstones and NAFLD, as well as many common risk factors, so there may be some associations between gallstones and NAFLD, which may be a risk factor for gallstones.
Smoking may increase the risk of gallstones through various biological mechanisms, and the incineration of cigarettes produces reactive oxygen species, which may activate the signaling cascade within epithelial cells, leading to inflammatory  activation, such as tumor necrosis factor [28]. Smoking may also influence gallstone formation by promoting endothelial dysfunction and thrombosis, as well as by influencing the immune response and microbial composition [29,30]. Previous studies have shown that diabetes is a risk factor for gallstone formation, which was consistent with our findings [23]. There were several mechanisms of association between diabetes and gallstones; for example, insulin bile directly promoted gallstone formation [31], and a lack of motivation in people with diabetes may be another predisposing factor for gallstone formation [32]. It has been reported previously that sedentary behavior leads to an increased risk of gallstones; however, that study used a random effects model to pool and analyze articles published over a period of time [33]. This approach has limitations in terms of publication bias and statistical heterogeneity. Furthermore, as the data were taken from different articles and may be from the same database, the results of the study may be biased. Our study used the NHANES database, a more current and realistic dataset. In addition, Zhang's study only showed an association between sedentary behavior and gallstones, not a causal relationship [33]. In contrast, we also performed an MR analysis, and the results showed that the time spent watching television and gallstones maintained an independent causal relationship and were independent risk factors for the formation of gallstones. Physical activity also maintained an independent causal relationship with gallstones and was a protective factor against gallstone formation. Physical activity may affect gallstones through the following mechanisms: physical activity may increase glucose tolerance [34,35], physical activity may reduce biliary stasis due to gallbladder motility stimulation [36,37], or physical activity may increase the release of cholecystokinin and colonic transit [38].
Despite the high quality of this study, some limitations must be acknowledged. First, our study was representative of the data used in the NHANES database, which is only representative of the baseline United States population. Therefore, this may limit the generalizability of our findings to other national populations. Second, although cholesterol stones account for more than 80% of all subtypes of gallstones, including cholesterol, pigmented, and mixed gallstones, our study did not examine the relationship between various subtypes of gallstones based on lifestyle factors. Moreover, we could not eliminate residual confounding factors due to the observational nature of the study, although we controlled for many known confounders. The results of this study may also be influenced by recall and detection bias, wherein not all patients with asymptomatic gallstones or gallstones with few symptoms undergo imaging diagnostics. The current study also included a largely heterogeneous patient population, which should be considered when reviewing our findings.

Ethical approval
Publicly available datasets were analyzed in this study. This data can be found at: https://www.cdc.gov/nchs/nhanes/index.htm.