Background
Inflammatory bowel disease (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), belongs to a group of chronic diseases occurring in the gastrointestinal tract with increasing incidence across the world [
1]. The number of patients suffering from IBD worldwide is expected to grow exponentially over the next few decades, which will pose a huge challenge to healthcare systems [
2]. In addition to the characteristic symptoms of gastric discomfort and diarrhea, extraintestinal symptoms frequently occur in patients with IBD [
3]. Studies have reported that IBD is related to various extraintestinal manifestations, containing ophthalmologic, genitourinary, dermatologic, hematologic, pulmonary, cardiovascular, neurologic, pancreatic and hepatobiliary systems [
4]. However, CD and UC can affect the development of extraintestinal complications to varying degrees [
5].
The incidence of kidney stone disease (KSD) increases across sex, race and age, affecting approximately 15% of the population [
6]. For the majority of the population, KSD is caused by a multifactorial etiology involving genetic and environmental factors. Genetic approaches studying KSD have uncovered that the following play significant roles in the formation of renal stones: channels and transporters; ions, protons and amino acids; the metabolic pathways for cysteine, vitamin D, oxalate, uric acid and purines; and the calcium-sensitive receptor signaling pathway [
7]. Extraintestinal manifestations are quite common in patients with IBD, and kidney stone disease has been reported with increasing frequency in these patients and with a higher risk of CD than UC [
8,
9]. As for its relevant pathogenesis, hypercalcemia and osteoporosis due to long-term corticosteroid exposure may contribute to the development of calcium-containing calculus [
10]. Furthermore, the gut microbiota residing in the human gastrointestinal tract has been reported to be involved in the formation of renal calculus [
11], making the interaction between IBD and the gut microbiota a key risk factor [
12]. In a large cohort study, the risk of urolithiasis was related to anti-TNF therapy and surgery. Small bowel resection or ileostomy will alter intestinal absorption in patients diagnosed with IBD [
13]. Interestingly, a clinical trial showed kidney stones were discovered in none of the UC patients but in some of the CD patients. Meanwhile, hyperoxaluria occurred in 36% of patients with CD but was absent in those with UC, suggesting the occurrence of renal calculi might be related to the patients with CD [
14]. Nevertheless, the genetic correlation and causal relationship between IBD and KSD remain unclear at present.
Traditional epidemiological approaches have some limitations that affect causal estimates. In observational studies, the causal link between IBD and KSD is more likely to be biased by the likelihood of potential confounders such as age [
15], oxaliplatin [
16], obesity [
17,
18] and lipid metabolism [
19,
20]. In addition to the drawbacks mentioned in observational studies, randomized controlled trials (RCTs) are identically difficult to directly investigate the etiology of diseases due to the strict control of experimental conditions, the high standards of experimental design and implementation and medical ethical considerations. Two-sample Mendelian randomization (MR) is a method that employs genetic variants as instrument variables (IVs) of exposure. It is widely applied to study the causal relationship between potential risk elements and health outcomes in observational data [
21]. Besides, it can not only avoid the insurmountable problems present in epidemiological studies but also generate more reliable evidence concerning which interventions are supposed to produce health benefits [
22]. In the current study, a two-sample bidirectional and multivariable MR analysis was performed to infer the causal association between IBD (including CD and UC) and KSD.
Discussion
The two-sample bidirectional MR study determined that genetically predicted CD was causally related to KSD, lending support to the findings of epidemiological studies [
4,
16,
42,
43]. Whereas our findings indicated that UC was not causally associated with the formation of urinary calculi, somewhat contradicting a few published studies. In the reverse MR analyses, there was no causal evidence supporting the idea that KSD could increase the risk of IBD (including CD and UC). Univariable MR analysis has indicated that IBD as a whole may play a vital role in developing renal calculus. Nevertheless, MVMR analysis refuted the results after adjusting potential confounders including HDL-C, LDL-C and TG, hinting that dyslipidemia might act as a mediator to promote the formation of kidney stones. Relevant studies have shown that patients with IBD have high amounts of inflammatory cytokines in their blood. These inflammatory cytokines may lead to a reduction in lipoprotein lipase enzyme activity, resulting in a typical lipoprotein profile with low levels of HDL-C and elevated levels of LDL-C and TG [
44]. However, changes in the patient’s lipid profile may indicate abnormalities in urine physicochemistry and stone risk. For instance, low HDL and high TG levels were linked to reduced urine pH. Non-HDL has a strong relationship with uric acid and urinary sodium. Uric acid stones were more prevalent in patients with high TG levels [
32].
Nephrolithiasis is a systemic metabolic disorder, and its main pathogenic factors include metabolic abnormalities, urinary tract infections, and drug factors. IBD (CD and UC) is a kind of idiopathic inflammatory bowel disease characterized by considerable clinical and genetic heterogeneity. Studies have demonstrated the following reasons accounting for the relationship between IBD and KSD: (i) patients with IBD exposed to surgery could lead to supersaturation for calcium oxalate and uric acid [
8] and TNF-alpha inhibitor treatment per se drives the increased risk of stone formation [
13]; (ii) malabsorption of bile salts and fatty acids caused by recurrent inflammation of the intestine can increase the solubility of oxalate [
45]; (iii) chronic intestinal inflammation could lead to fluid losses, bicarbonate losses, and reduced magnesium absorption, which further promote the formation of urinary calculus [
46,
47].
According to previous studies, 4–23% of patients with IBD suffered from renal and urinary tract complications, with nephrolithiasis being the most common form of renal manifestation [
48]. Based on a meta-analysis including 1624 individuals, urinary complications might occur in up to 22% of patients suffering from IBD, and calcium oxalate is more frequent in patients with CD than with UC [
4]. Likewise, the latest retrospective analysis containing 1874 patients diagnosed with IBD reported that renal involvement may be observed in approximately 6% of patients suffering from IBD. Patients with CD seem to be more susceptible than those with UC. It also found renal manifestations were associated with surgical resection history and disease activity in CD patients, whereas no such link was identified in patients with UC [
49]. These studies all demonstrated that CD patients presented a higher risk of KSD compared with UC patients. The different effects of CD and UC on KSD might be interpreted as follows: CD belongs to a kind of systemic disorder with a prolonged premorbid stage, whereas UC is frequently confined to the distal colonic tract and characterized by acute mucosal lesions. According to anatomical position, CD occurs primarily in the small intestine and frequently affects the terminal ileum [
50], which directly leads to hyperoxaluria in patients with ileal dysfunction [
45]. Moreover, extensive resection of the lesions may result in estrogen deficiency [
51], which may increase the risk of calculus recurrence by elevating calcium oxalate saturation and urinary calcium [
52].
In this MR analysis, the causal impact of UC on KSD was not identified in our outcomes, contradicting most previous studies regarding UC as one of the causes of nephrolithiasis [
4,
42]. But in a comparative study, McConnell et al. reported similar findings to ours: renal calculi and hyperoxaluria were found in none of the patients with UC [
14]. At the same time, in a recent MR analysis, CD has been shown to sharply increase the risk of urolithiasis, while UC failed to produce this effect [
53]. The reasons for the difference between many epidemiological studies and MR analysis are as follows: First, some interference factors may exert an influence on the results of epidemiological observational studies. For instance, the nutritional status and medication status of different groups of people will affect the incidence of kidney stones. Second, MR is an approach using genetic data as a bridge to explore causal associations between exposure and outcome, which is rarely affected by causal inversion and confounding factors. Therefore, it is necessary to implement more epidemiological studies to precisely assess the relationship between UC and KSD. Lastly, publicly available GWAS summary statistics about the effect of disease activity on KSD have not been issued. Thus, it’s difficult to judge whether the MR results were biased by the stage of the disease.
This study had several advantages. Our MR analysis was the first study focusing on this topic. In this research, all the individuals were of European descent, which guaranteed the homology of the population. Meanwhile, two independent populations were utilized to examine these connections, and the consistent findings guaranteed the stability of our results. In addition, various MR methods were used to lend support for exploring the causal impacts of genetically predicted IBD, CD and UC on KSD. Following selection criteria, we screened eligible SNPs as IVs to infer the causal evidence between the risk of interest and outcome after removing LD, outlier SNPs, and pleiotropic genetic variants. To avoid the influence of many weak IVs on our MR analysis, F statistics were calculated and much greater than 10, suggesting a small likelihood of bias caused by weak IVs. With many weak IVs, the MR-RAPS method was also performed to provide a robust inference for the MR results. Since we included two groups of exposure GWAS summary data, the causal estimates might differ from each other, and thus, the combined effect size was obtained using meta-analysis, which augmented the causal inference in terms of IBD (including CD and UC) with the risk of KSD. Lastly, we applied MVMR analyses to adjust confounding factors for investigating the direct causal effect of IBD on KSD and reverse MR to examine the effect of KSD on IBD.
Several limitations existed in this study. First, the exposure and outcome datasets included in the MR analysis referred to data from patients of European descent, which confined the generalizability to other ethnicities. Further investigation is needed to check our conclusions with those of other ancestors. Second, Previous studies have shown minimal renal dysfunction in IBD is related to disease activity but not with 5-ASA use [
54], indicating disease activity may play a critical role in extra-intestinal manifestations. In this study, the effect of IBD activity on KSD was unable to be examined with MR analysis due to the lack of GWAS data related to IBD activity. It may not be the disease itself but the IBD activity that is the key promoter of the formation of kidney calculi. Third, our results might be misled on account of the relatively small samples of KSD. Therefore, it must expand to contain data with a larger sample size of nephrolithiasis to study the effect of IBD on KSD in the future. Fourth, although we demonstrated that CD could raise the risk of kidney stone disease, the specific signaling pathways remain in need of more studies. Fifth, overlapping participants are not supposed to be contained in both exposure and outcome datasets applied in two-sample MR analyses. In this research, although the degree of overlap could not be estimated, strong instruments can minimize the bias from population overlap (F statistic>10) [
55]. Sixth, it remains uncertain to which degree the risk of KSD is associated with the treatment of IBD or disease severity, and further inclusion in the population for analysis is recommended. Lastly, the important limitation is unobserved pleiotropy, which means the risk of renal calculus might be affected by genetic instruments through other pathways but through IBD, CD and UC, despite the study design being less susceptible to confounders than observational research.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.