The Role of Calcium, 25-Hydroxyvitamin D, and Parathyroid Hormone in Irritable Bowel Syndrome: A Bidirectional Two-Sample Mendelian Randomization Study

Several observational studies have indicated the potential associations among calcium, vitamin D (Vit-D), and irritable bowel syndrome (IBS). However, the causal relationship deduced from these studies is subject to residual confounding factors and reverse causation. Therefore, we aimed to explore the bidirectional causal effects among serum calcium, Vit-D, PTH, and IBS at the genetic level by a two-sample Mendelian randomization (MR) analysis of the datasets from IEU OpenGWAS database. Sensitivity analyses were performed to evaluate the robustness. The estimates were presented as odds ratios (ORs) with their 95% confidence intervals (CIs). The results of the inverse variance weighted method did not reveal any causal relationship between the genetically predisposed calcium (OR = 0.92, 95% CI: 0.80–1.06, p = 0.25) and Vit-D (OR = 0.99, 95% CI: 0.83–1.19, p = 0.94) level and the risk of IBS. The bidirectional analysis demonstrated that genetic predisposition to IBS was associated with a decreased level of PTH (beta: −0.19, 95%CI: −0.34 to −0.04, p = 0.01). In conclusion, the present study indicates no causal relationship between the serum calcium and Vit-D concentrations and the risk of IBS. The potential mechanisms via which IBS affects serum PTH need to be further investigated.


Introduction
Irritable bowel syndrome (IBS) is one of the most common gastrointestinal diseases, and affects approximately 5-10% of the global population, which exerts an immense impact on the patient's quality of life, society, and economy [1]. The most complained symptoms include abdominal pain/discomfort and diarrhea/constipation. The pathogenesis of IBS is complex and recent studies bring the consensus that IBS mainly results from the disorder of gut-brain interactions [2]. Furthermore, epidemiological studies suggest that genetics, diet, gut microbiota dysbiosis, gut infection, and psychological factors are all risk factors for IBS, which can exert effects on IBS via disrupting the bidirectional interactions of the gut-brain axis [3,4]. Considering these factors, the common therapeutics for IBS include dietary exclusion, probiotics/fecal microbiota transplant, antibiotics, psychotropic medications, and symptom-relieving drugs (e.g., antispasmodics, antidiarrheal agents, and laxative) [5]. However, all the treatments have limited therapeutic effectiveness. Therefore, there is still

Study Design
To investigate the causality between exposures and disease, we conducted two-sample MR analysis that used genetic variants as instrumental variables to explore the causal effects of risk factors on outcomes. Compared with observational studies, MR can avoid reverse causation and reduce confounding factors. The graphical flow of the experimental design is shown in Figure 1. The validity of MR analysis relies on three assumptions: (1) there is strong association between the IVs and the exposure; (2) each IV is not associated with confounding variables; (3) each IV is only associated with the outcome through the exposure; there are no alternative pathways for the association.

Data Sources and Study Population
The data of our study were obtained from the IEU OpenGWAS database (https://gwas.mrcieu.ac.uk/, accessed on 10 October 2022), a database of 244,879,032,980 genetic associations from 42,334 GWAS summary datasets, for query or download.
The datasets utilized in our study are shown in Table 1. For the dataset of Vit-D, summary statistics of serum 25-hydroxyvitamin D levels were from a GWAS of the EBI database with a sample size of 496,946 (ebi-a-GCST90000618) [18]. For the dataset of calcium, we used summary statistics from a GWAS of UK Biobank from Neale lab with a sample size of 315,153 (ukb-d-30680_irnt). For the dataset of PTH, the complete GWAS summary data on protein levels as described by Sun et al. (2018) was used (prot-a-2431) [19], and the sample size was 3301. For the dataset of IBS, summary statistics from FinnGen biobank analysis including 4605 patients of IBS and 182,423 controls were used (finn-b-K11_IBS) [20]. All cases were defined by the code M13 in the International Classification of Diseases-Tenth Revision (ICD-10).
All the above data samples were of European ethnicity. In all original studies, ethical approval and consent to participate were obtained.

Selection of Instrumental Variables
Firstly, the summary-level data above for Vit-D and calcium were screened by the genome-wide significance (p < 5 × 10 −8 ) to select the SNPs genetically associated with the traits. To avoid inaccurate results due to too few SNPs, the significance threshold of PTH data was relaxed to 5 × 10 −6 . Secondly, we utilized the linkage disequilibrium clumping to exclude some undesirable SNPs (r 2 > 0.001). Thirdly, we harmonized the respective exposure and outcome datasets using effect allele frequencies, while removing palindromic The validity of MR analysis relies on three assumptions: (1) there is strong association between the IVs and the exposure; (2) each IV is not associated with confounding variables; (3) each IV is only associated with the outcome through the exposure; there are no alternative pathways for the association.

Data Sources and Study Population
The data of our study were obtained from the IEU OpenGWAS database (https://gwas.mrcieu.ac.uk/, accessed on 10 October 2022), a database of 244,879,032,980 genetic associations from 42,334 GWAS summary datasets, for query or download.
The datasets utilized in our study are shown in Table 1. For the dataset of Vit-D, summary statistics of serum 25-hydroxyvitamin D levels were from a GWAS of the EBI database with a sample size of 496,946 (ebi-a-GCST90000618) [18]. For the dataset of calcium, we used summary statistics from a GWAS of UK Biobank from Neale lab with a sample size of 315,153 (ukb-d-30680_irnt). For the dataset of PTH, the complete GWAS summary data on protein levels as described by Sun et al. (2018) was used (prot-a-2431) [19], and the sample size was 3301. For the dataset of IBS, summary statistics from FinnGen biobank analysis including 4605 patients of IBS and 182,423 controls were used (finn-b-K11_IBS) [20]. All cases were defined by the code M13 in the International Classification of Diseases-Tenth Revision (ICD-10). All the above data samples were of European ethnicity. In all original studies, ethical approval and consent to participate were obtained.

Selection of Instrumental Variables
Firstly, the summary-level data above for Vit-D and calcium were screened by the genome-wide significance (p < 5 × 10 −8 ) to select the SNPs genetically associated with the traits. To avoid inaccurate results due to too few SNPs, the significance threshold of PTH data was relaxed to 5 × 10 −6 . Secondly, we utilized the linkage disequilibrium clumping to exclude some undesirable SNPs (r 2 > 0.001). Thirdly, we harmonized the respective exposure and outcome datasets using effect allele frequencies, while removing palindromic SNPs with intermediate allele frequencies. Lastly, according to the third assumption of MR that genetic variation cannot be associated with any possible confounding factor, we used PhenoScanner V2 [21] (a database of human genotype-phenotype associations) to search the SNPs and exclude those associated (p < 1 × 10 −5 ) with confounding factors such as drinking [22], smoking [23], depression, and anxiety [24].
The IV exposure strength of genetic instruments was assessed from the F statistic using an approximation. If F > 10, there is sufficient strength to avoid a problem of weak instrument bias in the two-sample model. The F statistics were computed by the admittedly reliable formula F = R 2 (N − 2)/(1 − R 2 ). R 2 and N refer to the cumulative explained variance of selected SNPs and sample size separately [25]. R 2 was calculated using the formula R 2 = 2 × MAF × (1 − MAF) × Beta 2 [26].

Statistical Analyses
Multiple approaches were used in our study. We utilized the method of inverse variance weighted (IVW) as the primary analysis for its efficiency to estimate the causal effect. The weighted median was used as auxiliary method when the heterogeneity was significant, and the MR-Egger regression method was used to assess the pleiotropy by intercept test. According to the assumption of MR analysis, the instrumental variable must be only associated with the outcome through the risk factor; thus, if there are other pathways via which the outcome is influenced by genetic variants, bias will occur, and the horizontal pleiotropy may increase the false positive rate. Therefore, the pleiotropy should be evaluated using the method of MR-Egger and MR-PRESSO. The former can evaluate the potential pleiotropy in the IVW model, and the latter is used for testing horizontal pleiotropy via identifying and removing outlying instrumental variables (NbDistribution = 3000, SignifThreshold = 0.05). The leave-one-out sensitivity analysis was performed to evaluate the robustness of the study findings. The estimates were presented as odds ratios (ORs) with their 95% confidence intervals (CIs) per one standard deviation (SD) increase in the exposures. The statistical analyses above were performed in R 4.1.3 with R package of "TwoSampleMR" (version 0.5.6) and "MRPRESSO".

Instrumental Variables
In the analysis investigating the effect of Vit-D and calcium on IBS risk, 110 and 186 SNPs were screened, respectively, as potential instrumental variables (p < 5 × 10 −8 ). As for PTH, 15 SNPs were screened for instrumental variables (p < 5 × 10 −6 ). After linkage disequilibrium clumping and the removal of palindromic SNPs and confounders, 102, 176, and 14 SNPs could be used in the analyses as the instrumental variables of Vit-D, calcium, and PTH, respectively. The F statistics demonstrated that there was no bias due to weak instruments (F > 10, Tables A1-A4).

Main Analyses and Sensitivity Analyses
As shown in Figure 2, genetically predicted risk of IBS was not associated with the levels of vitamin D (p = 0.938, OR = 0.99, 95% CI: 0.83-1.19), calcium (p = 0.248, OR = 0.92, 95% CI: 0.80-1.06), and parathyroid hormone (p = 0.427, OR = 1.04, 95% CI: 0.94-1.15) using the IVW method. As shown in Table 2, Cochran's Q statistics demonstrated no heterogeneity based on genetically predicted SNPs of Vit-D, calcium, and PTH (p > 0.05). The MR Egger intercept test showed no evidence of directional pleiotropy (p > 0.05). The results of the leave-one-out method demonstrated that the removal of SNP did not fundamentally affect the results, which indicated that the results were actually robust.  As shown in Figure 3, genetically predicted levels of PTH were associated with the risk of IBS (p = 0.012, Beta = −0.188) while IBS was not the risk factor of Vit-D (p = 0.898, Beta = −0.001) and calcium (p = 0.432, Beta = −0.006) using the IVW method. As shown in Table 2, Cochran's Q statistics demonstrated no heterogeneity based on genetically predicted SNPs of Vit-D, calcium, and PTH (p > 0.05). The MR Egger intercept test showed no evidence of directional pleiotropy (p > 0.05). The results of the leave-one-out method demonstrated that the removal of SNP did not fundamentally affect the results, which indicated that the results were actually robust.  As shown in Figure 3, genetically predicted levels of PTH were associated with the risk of IBS (p = 0.012, Beta = −0.188) while IBS was not the risk factor of Vit-D (p = 0.898, Beta = −0.001) and calcium (p = 0.432, Beta = −0.006) using the IVW method. As shown in Table 2, Cochran's Q statistics demonstrated no heterogeneity based on genetically predicted SNPs of Vit-D, calcium, and PTH (p > 0.05). The MR Egger intercept test showed no evidence of directional pleiotropy (p > 0.05). The results of the leave-one-out method demonstrated that the removal of SNP did not fundamentally affect the results, which indicated that the results were actually robust.

Discussion
To the best of our knowledge, this is the first two-sample MR study to generally clarify the causal relationship among calcium, Vit-D, PTH, and IBS. Despite employing the latest large sample size and strong instruments, our MR results did not indicate the significantly causal associations of genetically predicted calcium, Vit-D, and PTH with the risk of IBS.
Researchers have been devoted to exploring the role of micronutrients in the pathogenesis and treatment of IBS [27,28]. A systematic review including 12 interventional and 14 observational studies showed that IBS patients generally had lower levels of Vit-D, vitamin B2, calcium, and iron compared with control subjects. Meanwhile, studies also found that exclusion diets were associated with deficiencies of the aforementioned micronutrients [7]. As the major circulating form of Vit-D, 25-hydroxyvitamin D is used as indicator of Vit-D status. 25-Hydroxyvitamin D is critical to regulate calcium metabolism and a series of pathological and physiological processes in intestinal homeostasis [29]. The various effects of Vit-D supplementation on IBS patients were reported in several randomized controlled trials and systematic reviews. Jalili et al. conducted a randomized, doubleblind, placebo-controlled clinical trial to assess the impact of Vit-D supplementation on symptoms severity and quality of life (QOL) in IBS patients and found that, compared to the placebo group, Vit-D therapy could markedly improve the symptoms and QOL of IBS patients [14]. Similarly, a systematic review and meta-analysis including four randomized, placebo-controlled trials showed that Vit-D supplementation was remarkably superior to placebo in improving the symptom severity (WMD: −84.21, 95% CI: −111.38 to −57.05, I 2 = 73.2%; WMD: −28.29, 95% CI: −49.95 to −6.62, I 2 = 46.6%, respectively) and QOL (WMD: 14.98, 95% CI: 12.06 to 17.90, I 2 = 0.0%; WMD: 6.55, 95% CI: −2.23 to 15.33, I 2 = 82.7%, respectively) of IBS patients [30]. However, the other randomized, double-blind, placebocontrolled study by Williams et al. demonstrated that there were no improvements in the IBS symptom severity and QOL between the trial (Vit-D supplementation) and placebo

Discussion
To the best of our knowledge, this is the first two-sample MR study to generally clarify the causal relationship among calcium, Vit-D, PTH, and IBS. Despite employing the latest large sample size and strong instruments, our MR results did not indicate the significantly causal associations of genetically predicted calcium, Vit-D, and PTH with the risk of IBS.
Researchers have been devoted to exploring the role of micronutrients in the pathogenesis and treatment of IBS [27,28]. A systematic review including 12 interventional and 14 observational studies showed that IBS patients generally had lower levels of Vit-D, vitamin B2, calcium, and iron compared with control subjects. Meanwhile, studies also found that exclusion diets were associated with deficiencies of the aforementioned micronutrients [7]. As the major circulating form of Vit-D, 25-hydroxyvitamin D is used as indicator of Vit-D status. 25-Hydroxyvitamin D is critical to regulate calcium metabolism and a series of pathological and physiological processes in intestinal homeostasis [29]. The various effects of Vit-D supplementation on IBS patients were reported in several randomized controlled trials and systematic reviews. Jalili et al. conducted a randomized, double-blind, placebo-controlled clinical trial to assess the impact of Vit-D supplementation on symptoms severity and quality of life (QOL) in IBS patients and found that, compared to the placebo group, Vit-D therapy could markedly improve the symptoms and QOL of IBS patients [14]. Similarly, a systematic review and meta-analysis including four randomized, placebo-controlled trials showed that Vit-D supplementation was remarkably superior to placebo in improving the symptom severity (WMD: −84.21, 95% CI: −111.38 to −57.05, I 2 = 73.2%; WMD: −28.29, 95% CI: −49.95 to −6.62, I 2 = 46.6%, respectively) and QOL (WMD: 14.98, 95% CI: 12.06 to 17.90, I 2 = 0.0%; WMD: 6.55, 95% CI: −2.23 to 15.33, I 2 = 82.7%, respectively) of IBS patients [30]. However, the other randomized, double-blind, placebo-controlled study by Williams et al. demonstrated that there were no improvements in the IBS symptom severity and QOL between the trial (Vit-D supplementation) and placebo groups [15]. Moreover, a systematic review and meta-analysis based on six randomized controlled trials including 616 participants indicated that Vit-D supplementation led to no significant improvements in symptom severity and QOL of IBS subjects in contrast to placebo [31]. Considering that Vit-D contributes to the regulation of the gut microbiome, immune system, inflammatory processes, and the intestinal mucosal barrier, the present interventional trials on IBS mainly focused on Vit-D supplementation. Few studies evaluated the effects of calcium supplementation on IBS symptom severity and QOL. In contrast to studies that reported the relationship among Vit-D, calcium, and IBS, our study suggested no causal association among Vit-D, calcium, and IBS. The contradictory findings might be explained by several factors: trial participants with different ages, races, sexes, and vitamin D statuses, sample size, intervention duration, intervention diet, and placebo effects.
In addition, our bidirectional two-sample MR analysis identified that IBS was associated with a lower level of PTH, although there was no causal effect of PTH on IBS. The main function of PTH is to increase the concentration of serum calcium and decrease the concentration of serum phosphorus by impacting its primary target organs of bone and kidney, so as to regulate the homeostasis of calcium and phosphorus in vivo. It was noteworthy that recent studies suggested an increased risk of osteoporosis and osteoporotic fracture for IBS patients. A systematic review and meta-analysis including four cohorts and one cross-sectional study with 526,633 participants indicated that IBS patients had a remarkably higher risk of osteoporosis than the non-IBS subjects (pooled RR: 1.95, 95%CI: 1.04-3.64, I 2 = 100%) [32]. Moreover, even though not statistically significant, IBS subjects had an increased risk of osteoporotic fracture (pooled RR: 1.58, 95%CI: 0.95-2.62, I 2 = 99%). The possible mechanisms for the association between IBS and osteoporosis comprise chronic inflammation, abnormal activation of the hypothalamic-pituitary-adrenal (HPA) axis, smoking, and malnutrition. When suffering from osteoporosis, the secretion of PTH was reduced to inhibit the activity of osteoclasts, thus impeding the progression of osteoporosis, which might be the plausible explanation for the relationship between IBS and reduced level of PTH.
To our knowledge, this is the first study to elucidate the causal correlation among calcium, Vit-D, PTH, and IBS from the perspective of genetic variants using a bidirectional two-sample MR approach. This method could greatly circumvent the possible impacts of reverse causation and residual confounding factors, such as incomplete adjustment for confounders, the absence of high-quality evidence, and relatively small sample sizes of trials. Additionally, we performed several sensitivity analyses to strengthen the robustness of our results.
However, there are some limitations associated with this study. Firstly, the study mainly analyzed the European participants enrolled in the GWAS biobank; hence, the results could not precisely reflect the fact of patients from other regions and races. Secondly, we failed to accomplish the sex-specific, IBS subtype-specific, age-specific, and race-specific analyses due to a lack of data. Lastly, MR analysis possesses some inherent shortcomings, making it impossible to eliminate the effects of confounding factors and horizontal pleiotropy.
In conclusion, the present study provides no evidence that calcium, Vit-D, and PTH are causally associated with IBS, and it suggests a lower concentration of PTH in IBS subjects. Our findings may reduce possible expenses and research interests in elucidating the effects of calcium, Vit-D, and PTH on IBS. More importantly, further research is needed to investigate the causal relationship between micronutrients and IBS.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Ethical approval and consent to participate were not needed for this current study because it is a secondary analysis of previously published data. In all original studies related to this study, ethical approval and consent to participate were obtained.

Data Availability Statement:
This study is based on the public database, and all related-datasets are available at https://gwas.mrcieu.ac.uk/ (accessed on 10 October 2022).

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.