Abstract
Background
Environmental factors play significant roles in the pathogenesis of inflammatory bowel disease. This study investigates the associations between hard water exposure and the prevalence and incidence of inflammatory bowel diseases, including Crohn’s disease and ulcerative colitis.
Methods
A cross-sectional and cohort study was conducted using 481,468 participants in the UK Biobank. Logistic regression was used to examine associations between hard water exposure and disease prevalence at baseline, while Cox regression assessed associations with incident cases during follow-up.
Results
A total of 481,468 participants were included in the cross-sectional study. Higher levels of hard water exposure were associated with reduced prevalence of Crohn’s disease (OR = 0.88, 95% CI: 0.79–0.99, p = 0.027, and p = 0.039 for the trend) and ulcerative colitis (OR = 0.83, 95% CI: 0.77–0.89, p < 0.001, and p < 0.001 for the trend). 475,895 participants with a mean follow-up of 14.6 years were included in the cohort study. Hard water was significantly associated with a decreased risk of Crohn’s disease (HR = 0.84, 95% CI: 0.73–0.97, p = 0.016, and p = 0.007 for the trend) but not ulcerative colitis. We also observed a stronger inverse association between very hard water and Crohn’s disease among individuals with low genetic risk. (HR = 0.74, 95% CI: 0.57–0.95, p = 0.019).
Conclusion
Hard water is linked to a lower prevalence of both CD and UC, as well as a lower incidence of CD. However, no significant association was observed with the incidence of UC.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12876-025-04230-1.
Keywords: Inflammatory bowel diseases, Hard water, Cohort study, Cross-sectional study
Introduction
Inflammatory bowel diseases (IBD) refer to a group of chronic inflammatory conditions affecting the gastrointestinal tract, encompassing Crohn’s disease (CD) and ulcerative colitis (UC) [1, 2]. Until 2017, the global prevalence of IBD rose to 84.3 per 100,000 people [1]. Notably, the UK recorded the second-highest prevalence, with 449.6 cases per 100,000 population [1].
IBD is driven by a combination of genetic predispositions and environmental influences, although the precise mechanism remains unclear [1–4]. The rapid rise in IBD incidence over recent decades is closely linked to lifestyle westernization and urbanization [5]. Environmental factors include behaviours that increase individual risk (such as diet, medicine use, and smoking) and broader influences that affect populations across regions (such as air pollution and sun exposure) [5–7]. These external factors may impact disease risk and progression by influencing gut microbiota composition and immune system balance [8]. Prior research has explored associations between various environmental exposures and IBD incidence [5, 9].
Among these environmental exposures, water quality, specifically hard water exposure, has received relatively limited attention. Hard water, characterised by high concentrations of dissolved calcium and magnesium, forms as minerals dissolve when water passes through sedimentary rocks [10, 11]. The primary minerals in hard water include calcium, typically found as calcite and dolomite [10]. Hard water is a significant source of calcium and magnesium [11]. These minerals play critical roles in numerous biological and physiological processes, potentially impacting the mechanisms underlying the development of IBD [12]. Several studies have demonstrated protective effects of dietary calcium intake on IBD [12]. A few studies have evaluated the correlation between water supply characteristics and the IBD risk [13, 14]. For example, a study in Norway examined the water supply by assessing its physical characteristics, as well as levels of iron, aluminum, and microbial content [13]. Another study in Iran found that iron exposure through drinking water might offer a protective benefit against the onset of UC [14]. However, the impact of calcium in the water supply on the risk of IBD remains unexplored.
In this research, we aimed to elucidate the associations between hard water and the prevalence and incidence risk of IBD, with a particular focus on potential interactions with genetic susceptibility. We carried out an extensive population-based cross-sectional and cohort study within the UK Biobank, incorporating polygenic risk scores (PRSs) to assess potential interactions with genetic risk.
Methods
Study population
The primary objective of this study is to explore potential associations between hard water and IBD, including both CD and UC. Analyses of the associations with disease prevalence and incidence were pre-specified and conducted using cross-sectional and cohort designs, respectively. This study is based on the UK Biobank, a comprehensive cohort comprising over 500,000 individuals enrolled throughout the UK between 2006 and 2010 [15]. Detailed information about the UK Biobank has been described elsewhere [15].
In the current cross-sectional study, participants with missing genetic information (n = 16,227) and missing hard water exposure information (n = 4478) were excluded. Participants with unclear diagnosis information of IBD were further excluded in IBD subtype analyses (n = 20). For the cohort study, participants with missing genetic information (n = 16,227), missing hard water exposure information (n = 4478), and IBD diagnosis at baseline (n = 5573) were excluded. Participants with unclear diagnosis information of IBD were also excluded from IBD subtype analyses (n = 11, see Fig. 1 for details).
Fig. 1.
Study flow diagram. The flow diagram showed the design of this study
Exposure and outcome measurement
Domestic data of hard water exposure were initially provided by the University of Melbourne [10] and subsequently made available by the UK Biobank. Baseline measurements were based on survey data from 2005. Participants’ residential locations were geocoded using Ordnance Survey Code-Point® data, which assigns British National Grid coordinates to each postcode, rounded to the nearest kilometre. Each participant’s postcode was linked to regional water quality information obtained from local water supply companies. To illustrate regional variation in hard water exposure across the UK, a map of the hard water levels across assessment centres is present in Fig. 2.
Fig. 2.
Water hardness levels across UK Biobank assessment centers. The water hardness levels represent the median water hardness values associated with the participants at each assessment center
Data on domestic hard water levels, encompassing the United States Geological Survey (USGS) classification and the World Health Organisation (WHO) classification, CaCO₃ (mg/L), calcium (mg/L), and magnesium (mg/L) concentrations were included in the analysis. USGS classification categorises hard water levels into four levels according to CaCO₃ concentration: soft water (0–60 mg/L), moderately hard water (> 60–120 mg/L), hard water (> 120–180 mg/L), and very hard water (> 180 mg/L) [10]. In contrast, the WHO classifies water as soft if the CaCO₃ concentration is below 200 mg/L and as hard if it exceeds or is equal to 200 mg/L [10].
Diagnostic information was integrated and provided by UK Biobank, including Read codes from primary care (Category 3000), ICD-9 and ICD-10 codes from hospital inpatient data (Category 2000), ICD-10 codes from death registry data (Fields 40001 and 40002), and self-reported medical conditions (Field 20002). The detailed integration methodology is documented on the UK Biobank website (Resource 593). The primary outcomes of this study were the prevalence and incidence of CD and UC, identified using the relevant International Classification of Diseases codes (ICD-9: 555, 556; ICD-10: K50, K51). In the cohort study, participants who did not have IBD at the start were monitored from the baseline period (2006–2010) until the earliest occurrence of the following: a diagnosis of CD or UC, date of death, date of loss to follow-up, or the final follow-up date.
Covariate assessment
Data on recruitment age, ethnicity, smoking, alcohol consumption, and IPAQ physical activity (International Physical Activity Questionnaire [16]) were collected from a self-completed questionnaire at the assessment centre. The body mass index (BMI) was determined using height and weight data collected during physical evaluations. The Townsend Deprivation Index (TDI) was computed using variables including employment, car ownership, home ownership, and household overcrowding [17]. The Charlson Comorbidity Index (CCI) was constructed using 17 comorbidities, each assigned a specific weight, linked to ICD codes from hospital inpatient data [18] (Supplementary Table S24). Diet quality scores were constructed from touchscreen food frequency questionnaires completed at the assessment center based on the following 10 food items: vegetables, fruits, fish, dairy, whole grains, vegetable oils, refined grains, processed meats, unprocessed red meats, and sugar-sweetened beverages [19]. A higher diet quality score indicates a better diet quality. Information on genetic sex was obtained from genomics data. Continuous covariates, including diet quality scores, BMI, and TDI, were analysed in quantiles, and missing data were categorised as “unknown” for analysis. PRS scores for UC and CD were computed using a Bayesian method applied to meta-analysed GWAS summary data and are accessible via the UK Biobank. Detailed methods for the development of these PRS scores are described elsewhere [18]. Adjustments were also made using the first five principal components derived from ancestry genetic data to control for population stratification. Participants without genetic data or hard water exposure data were excluded from the analysis.
Statistical analyses
In the cross-sectional study, multivariable logistic regression models were used. In the cohort study, Cox proportional hazard regression models were used. Odds ratios (ORs) or hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. Model assumptions, including proportional hazards and linearity of covariates, were tested and satisfied. Model 1 was adjusted for age and sex. Model 2 included additional adjustments for CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, and diet quality scores. Model 3 was additionally adjusted for the initial five principal components of genetic ancestry and PRS scores for UC and CD individually in the subtype analysis.
Subgroup analyses were conducted based on age, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, PRS, and diet quality scores. The interactions between hard water exposure and each variable were assessed using a multiplicative interaction model. Sensitivity analyses were conducted: (1) reanalyzed missing values using multiple imputation (R package mice) [20]; (2) included additional adjustments for C-reactive protein, residential area, waist circumference, educational levels, sun exposure and assessment centers separately; (3) excluded participants diagnosed with IBD within the first year of follow-up; (4) replaced genetic sex by self-reported sex; (5) excluded participants with both UC and CD diagnoses during the follow-up; (6) excluded participants who moved residences within 10 years after inclusion.
The baseline characteristics were evaluated in descriptive analyses. Continuous variables are expressed as mean (SD), whereas categorical variables are reported as counts (percentage). Results of Cox proportional hazard regression models and multivariable logistic regression models were presented as hazard ratios (HRs) or odds ratios (ORs) and 95% confidence intervals (CIs). Hazard ratios represent relative measures of risk. A two-tailed p < 0.05 was considered statistically significant. Given that hard water levels were categorised in two different ways, we applied Bonferroni correction for multiple testing to provide a more stringent reference in our analyses. After correction, p-values < 0.025 were considered statistically significant. P-values were reported as exact values for p ≥ 0.001, and as “p < 0.001” for p < 0.001, in accordance with standard statistical reporting conventions. Statistical analyses were performed using R 4.2.2.
Geographic data visualization and validation
To visualize the regional variations in hard water levels across the UK Biobank assessment centers, we utilized data on CaCO3 concentration derived from the participants at each center. The median CaCO3 concentration was calculated to reflect the overall water environment in which participants at each assessment center reside. Geographic locations for the assessment centers were provided by UK Biobank in the form of grid coordinates, which were subsequently converted into latitude and longitude coordinates for mapping. The map was generated using R 4.2.2.
To validate the accuracy of the water hardness map, we performed the following steps. First, we converted the grid coordinates of each assessment center into the nearest postal codes using an online geocoding tool. Then we retrieved the most recent data of hard water from the official websites of UK water supply companies and created a reference map (Figure S2) that illustrates the latest water hardness levels at the precise locations of each assessment center. To further validate the accuracy, we conducted a paired Wilcoxon signed-rank test between the medium CaCO3 concentration data from each assessment center and the reference data. The results showed no significant difference in the medians of the two datasets (p = 0.467).
Ethical statement
The UK Biobank received ethical approval from the North West-Haydock Research Ethics Committee (REC reference: 16/NW/0274). Informed consent was obtained from all participants at the time of recruitment.
Results
Baseline characteristics of participants in the cross-sectional study
Baseline characteristics of participants in the cross-sectional study were presented according to the IBD diagnosis status (see Table 1). The study included 481,468 individuals, among whom 5,573 were diagnosed with IBD, including 3,750 with UC and 1,803 with CD. The mean age was 56.57 years for non-IBD participants and 57.31 years for those with IBD. Among non-IBD participants, 54.3% were female, compared to 52.1% among IBD participants. Most participants were ethnically white (94.3% for non-IBD and 95.7% for IBD).
Table 1.
Characteristics of participants according to IBD status and subtype in the cross-sectional study
| Non-IBD | IBD | CD | UC | |
|---|---|---|---|---|
| (n = 475895) | (n = 5573) | (n = 1803) | (n = 3750) | |
| Age, mean (SD) | 56.57 (8.08) | 57.31 (7.9) | 56.69 (8.06) | 57.61 (7.81) |
| Genetic sex, n (%) | ||||
| Female | 258383 (54.3) | 2903 (52.1) | 998 (55.4) | 1895 (50.5) |
| Male | 217512 (45.7) | 2670 (47.9) | 805 (44.6) | 1855 (49.5) |
| Ethnicity, n (%) | ||||
| White | 448558 (94.3) | 5334 (95.7) | 1725 (95.7) | 3589 (95.7) |
| Others | 26849 (5.6) | 233 (4.2) | 76 (4.2) | 157 (4.2) |
| Unknown | 488 (0.1) | 6 (0.1) | 2 (0.1) | 4 (0.1) |
| Education levels, n (%) | ||||
| College or university | 154180 (32.4) | 1557 (27.9) | 467 (25.9) | 1086 (29) |
| Others | 321223 (67.5) | 4010 (72.0) | 1334 (74) | 2660 (70.9) |
| Unknown | 492 (0.1) | 6 (0.1) | 2 (0.1) | 4 (0.1) |
| Townsend deprivation index, mean (SD) | −1.3 (3.09) | −1.26 (3.08) | −1.08 (3.19) | −1.35 (3.03) |
| BMI, mean (SD) | 27.44 (4.8) | 27.15 (4.68) | 26.84 (4.8) | 27.3 (4.61) |
| Diet Score, mean (SD) | 49.57 (14.91) | 46.82 (15.4) | 44.76 (15.83) | 47.82 (15.1) |
| Charlson Comorbidity Index, mean (SD) | 0.22 (0.7) | 0.36 (0.92) | 0.38 (0.95) | 0.38 (0.96) |
| IPAQ activity, n (%) | ||||
| High | 150314 (31.6) | 1629 (29.2) | 503 (27.9) | 1120 (29.9) |
| Moderate | 149592 (31.4) | 1674 (30) | 519 (28.8) | 1151 (30.7) |
| Low | 68027 (14.3) | 907 (16.3) | 319 (17.7) | 586 (15.6) |
| Unknown | 107962 (22.7) | 1363 (24.5) | 462 (25.6) | 893 (23.8) |
| Alcohol, n (%) | ||||
| Daily or almost daily | 96867 (20.4) | 1028 (18.4) | 285 (15.8) | 741 (19.8) |
| Three or four times a week | 109971 (23.1) | 1185 (21.3) | 356 (19.7) | 820 (21.9) |
| One to four times a week | 122608 (25.8) | 1389 (24.9) | 449 (24.9) | 939 (25) |
| One to three times a month | 52814 (11.1) | 663 (11.9) | 234 (13) | 427 (11.4) |
| Never or special occasions only | 92590 (19.5) | 1297 (23.3) | 476 (26.4) | 815 (21.7) |
| Unknown | 1045 (0.2) | 11 (0.2) | 3 (0.2) | 8 (0.2) |
| Smoking, n (%) | ||||
| Ever | 282980 (59.5) | 3584 (64.3) | 1186 (65.8) | 2385 (63.6) |
| Never | 190538 (40) | 1967 (35.3) | 610 (33.8) | 1350 (36) |
| Not know | 2377 (0.5) | 22 (0.4) | 7 (0.4) | 15 (0.4) |
| Water hardness (USGS), n (%) | ||||
| 0–60 | 169376 (35.6) | 2163 (38.8) | 695 (38.5) | 1459 (38.9) |
| 60–120 | 94420 (19.9) | 1115 (20) | 347 (19.2) | 762 (20.3) |
| 120–180 | 28236 (5.9) | 342 (6.1) | 107 (5.9) | 232 (6.2) |
| > 180 | 183530 (38.6) | 1951 (35) | 654 (36.3) | 1295 (34.6) |
| Water hardness (WHO), n (%) | ||||
| < 200 | 294657 (62) | 3655 (65.6) | 1163 (64.5) | 2474 (66) |
| ≥ 200 | 180905 (38) | 1916 (34.4) | 640 (35.5) | 1274 (34) |
| CaCO3 concentration, mean (SD) | 146.19 (114.98) | 136.69 (112.12) | 137.64 (113.03) | 135.7 (111.71) |
| Ca concentration, mean (SD) | 51.3 (40.03) | 48.28 (39.15) | 47.99 (39.22) | 48.45 (39.21) |
| Mg concentration, mean (SD) | 4.6 (3.76) | 4.57 (3.8) | 4.55 (3.72) | 4.57 (3.81) |
Abbreviations: BMI Body mass index, IPAQ International Physical Activity Questionnaire, USGS United States Geological Survey, WHO World Health Organization
Hard water was significantly associated with a decreased prevalence of CD and UC
In all three models, we found that very hard water (> 180 mg/L), as classified by the USGS, was significantly associated with a lower prevalence of UC (see Table 2 for details, Model 3: OR = 0.83, 95% CI: 0.77–0.89, p < 0.001, and p < 0.001 for the trend). This result was consistent with the classification by WHO (Model 3: OR = 0.88, 95% CI: 0.82–0.94, p < 0.001). Next, we analysed CaCO₃ concentration as a continuous variable. The results indicated that higher CaCO₃ levels were associated with a lower prevalence of UC (Model 3: OR = 0.93, 95% CI: 0.90–0.96, p < 0.001), with consistent findings observed for Ca2⁺ concentration (Model 3: OR = 0.93, 95% CI: 0.90–0.96, p < 0.001). Very hard water was significantly associated with the prevalence of CD in Model 1 (OR = 0.87, 95% CI: 0.78–0.97, p = 0.011, and p = 0.016 for the trend). When classified according to the WHO classification, a similar trend persisted (OR = 0.90, 95% CI: 0.82–0.99, p = 0.029), although these associations did not remain statistically significant after applying the Bonferroni correction for multiple testing. Results remained significant when analysed for continuous Ca2⁺ concentration in all three models and CaCO₃ concentration in Model 1 and Model 3. When analysing the overall prevalence of IBD, the association with hard water was consistent with the results observed in the separate analyses for UC and CD (Table 3).
Table 2.
Association between water hardness and prevalent UC and CD in the cross-sectional study
| CD | UC | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cases | Model 1 OR (95% CI) | P | Model 2 OR (95% CI) | P | Model 3 OR (95% CI) | P | Cases | Model 1 OR (95% CI) | P | Model 2 OR (95% CI) | P | Model 3 OR (95% CI) | P | ||
| Water hardness (USGS, mg/l) | 0–60 | 695 | ref | ref | ref | 1459 | ref | ref | ref | ||||||
| 60–120 | 347 | 0.90 (0.79–1.02) | 0.096 | 0.91 (0.80–1.04) | 0.158 | 0.91 (0.80–1.04) | 0.151 | 762 | 0.93 (0.86–1.02) | 0.130 | 0.94 (0.86–1.02) | 0.153 | 0.95 (0.87–1.04) | 0.236 | |
| 120–180 | 107 | 0.92 (0.75–1.13) | 0.450 | 0.96 (0.78–1.18) | 0.683 | 0.95 (0.77–1.17) | 0.620 | 232 | 0.96 (0.84–1.11) | 0.583 | 0.97 (0.85–1.12) | 0.694 | 0.98 (0.85–1.13) | 0.797 | |
| > 180 | 654 | 0.87 (0.78–0.97) | 0.011 | 0.90 (0.81–1.00) | 0.058 | 0.88 (0.79–0.99) | 0.027 | 1295 | 0.83 (0.77–0.89) | < 0.001 | 0.84 (0.78–0.91) | < 0.001 | 0.86 (0.80–0.93) | < 0.001 | |
| P for trend | 0.016 | 0.081 | 0.039 | < 0.001 | < 0.001 | < 0.001 | |||||||||
| Water hardness (WHO, mg/l) | < 200 | 1163 | ref | ref | ref | 2474 | ref | ref | ref | ||||||
| ≥ 200 | 640 | 0.90 (0.82–0.99) | 0.029 | 0.92 (0.83–1.02) | 0.098 | 0.90 (0.82–1.00) | 0.048 | 1274 | 0.85 (0.79–0.91) | < 0.001 | 0.86 (0.80–0.92) | < 0.001 | 0.88 (0.82–0.94) | < 0.001 | |
| CaCO3 concentration (per SD) | 0.94 (0.89–0.98) | 0.007 | 0.95 (0.91–1.00) | 0.041 | 0.94 (0.90–0.99) | 0.017 | 0.92 (0.89–0.95) | < 0.001 | 0.92 (0.89–0.96) | < 0.001 | 0.93 (0.90–0.96) | < 0.001 | |||
| Ca concentration (per SD) | 0.92 (0.88–0.97) | 0.001 | 0.93 (0.89–0.98) | 0.008 | 0.93 (0.88–0.97) | 0.002 | 0.93 (0.90–0.97) | < 0.001 | 0.94 (0.91–0.98) | < 0.001 | 0.95 (0.92–0.98) | 0.003 | |||
| Mg concentration (per SD) | 0.99 (0.95–1.04) | 0.814 | 1.00 (0.95–1.05) | 0.985 | 1.00 (0.95–1.05) | 0.918 | 0.99 (0.96–1.03) | 0.624 | 0.99 (0.96–1.03) | 0.626 | 1.00 (0.96–1.03) | 0.895 | |||
Abbreviations: USGS United States Geological Survey, WHO World Health Organization, OR odds ratio, P p-values
Model 1: adjusted for age and sex
Model 2: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI and diet quality scores
Model 3: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, diet quality scores, the first five principal components of ancestry genetics and PRS scores
Bold font for USGS and WHO water hardness groups indicates p < 0.025, while bold font for other variables indicates p < 0.05
Table 3.
Association between water hardness and prevalent IBD in the cross-sectional study
| IBD | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cases | Model 1 OR (95% CI) | P | Model 2 OR (95% CI) | P | Model 3 OR (95% CI) | P | ||
| Water hardness (USGS, mg/l) | 0–60 | 704 | ref | ref | ref | |||
| 60–120 | 353 | 0.92 (0.86–0.99) | 0.030 | 0.93 (0.86–1.00) | 0.051 | 0.94 (0.87–1.01) | 0.076 | |
| 120–180 | 110 | 0.95 (0.85–1.07) | 0.419 | 0.97 (0.87–1.09) | 0.637 | 0.97 (0.87–1.09) | 0.624 | |
| > 180 | 656 | 0.84 (0.79–0.89) | < 0.001 | 0.86 (0.81–0.91) | < 0.001 | 0.86 (0.81–0.92) | < 0.001 | |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||||
| Water hardness (WHO, mg/l) | < 200 | 1181 | ref | ref | ref | |||
| ≥ 200 | 642 | 0.86 (0.81–0.91) | < 0.001 | 0.88 (0.83–0.93) | < 0.001 | 0.88 (0.83–0.93) | < 0.001 | |
| CaCO3 concentration (per SD) | 0.92 (0.90–0.95) | < 0.001 | 0.93 (0.91–0.96) | < 0.001 | 0.93 (0.91–0.96) | < 0.001 | ||
| Ca concentration (per SD) | 0.93 (0.90–0.95) | < 0.001 | 0.94 (0.91–0.97) | < 0.001 | 0.94 (0.91–0.97) | < 0.001 | ||
| Mg concentration (per SD) | 0.99 (0.96–1.02) | 0.534 | 0.99 (0.97–1.02) | 0.639 | 1.00 (0.97–1.02) | 0.830 | ||
Abbreviations: USGS United States Geological Survey, WHO World Health Organization, OR odds ratio, P p-values
Model 1: adjusted for age and sex
Model 2: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI and diet quality scores
Model 3: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, diet quality scores, the first five principal components of ancestry genetics
Bold font for USGS and WHO water hardness groups indicates p < 0.025, while bold font for other variables indicates p < 0.05
Baseline characteristics of participants in the cohort study
In the cohort study, a total of 475,895 individuals were included, among whom 3,212 participants were diagnosed with IBD (1,035 CD and 2,166 UC) during the mean follow-up of 14.56 years. Approximately 80.5% of participants did not change their residential address within 10 years after baseline. The mean age was 56.56 years for non-IBD participants and 57.27 years for those with IBD. A slightly higher proportion of male participants were diagnosed with IBD (50.1%). Detailed characteristics were shown in Table 4.
Table 4.
Characteristics of participants according to IBD status and subtype in the cohort study
| Non-IBD | IBD | CD | UC | |
|---|---|---|---|---|
| (n = 472683) | (n = 3212) | (n = 1035) | (n = 2166) | |
| Age, mean (SD) | 56.56 (8.08) | 57.27 (8.01) | 56.97 (8.25) | 57.42 (7.88) |
| Genetic sex, n (%) | ||||
| Female | 256779 (54.3) | 1598 (49.9) | 561 (54.2) | 1037 (47.9) |
| Male | 215904 (45.7) | 1603 (50.1) | 474 (45.8) | 1129 (52.1) |
| Ethnicity, n (%) | ||||
| White | 445543 (94.3) | 3005 (93.9) | 976 (94.3) | 2029 (93.7) |
| Others | 26656 (5.6) | 192 (6) | 58 (5.6) | 134 (6.2) |
| Unknown | 484 (0.1) | 4 (0.1) | 1 (0.1) | 3 (0.1) |
| Education levels, n (%) | ||||
| College or university | 153372 (32.4) | 807 (25.2) | 266 (25.7) | 541 (25) |
| Others | 318823 (67.4) | 2390 (74.7) | 768 (74.2) | 1622 (74.9) |
| Unknown | 488 (0.1) | 4 (0.1) | 1 (0.1) | 3 (0.1) |
| Townsend deprivation index, mean (SD) | −1.31 (3.09) | −0.89 (3.25) | −0.71 (3.26) | −0.97 (3.25) |
| BMI, mean (SD) | 27.43 (4.8) | 27.98 (5.06) | 28.08 (5.22) | 27.93 (4.97) |
| Diet Score, mean (SD) | 49.59 (14.91) | 47.89 (15.27) | 47.18 (15.22) | 48.22 (15.29) |
| Charlson Comorbidity Index, mean (SD) | 0.22 (0.73) | 0.33 (0.86) | 0.33 (0.81) | 0.34 (0.89) |
| IPAQ activity, n (%) | ||||
| High | 149363 (31.6) | 947 (29.6) | 287 (27.7) | 660 (30.5) |
| Moderate | 148657 (31.4) | 931 (29.1) | 320 (30.9) | 611 (28.2) |
| Low | 67547 (14.3) | 479 (15) | 156 (15.1) | 323 (14.9) |
| Unknown | 107116 (22.7) | 844 (26.4) | 272 (26.3) | 572 (26.4) |
| Alcohol, n (%) | ||||
| Daily or almost daily | 96205 (20.4) | 661 (20.6) | 200 (19.3) | 461 (21.3) |
| Three or four times a week | 109344 (23.1) | 625 (19.5) | 195 (18.8) | 430 (19.9) |
| One to four times a week | 121809 (25.8) | 794 (24.8) | 242 (23.4) | 552 (25.5) |
| One to three times a month | 52468 (11.1) | 346 (10.8) | 120 (11.6) | 226 (10.4) |
| Never or special occasions only | 91821 (19.4) | 766 (23.9) | 276 (26.7) | 490 (22.6) |
| Unknown | 1036 (0.2) | 9 (0.3) | 2 (0.2) | 7 (0.3) |
| Smoking, n (%) | ||||
| Ever | 280777 (59.4) | 2194 (68.5) | 705 (68.1) | 1489 (68.7) |
| Never | 189545 (40.1) | 991 (31) | 328 (31.7) | 663 (30.6) |
| Unknown | 2361 (0.5) | 16 (0.5) | 2 (0.2) | 14 (0.6) |
| Water hardness (USGS), n (%) | ||||
| 0–60 | 168183 (35.6) | 1191 (37.2) | 415 (40.2) | 776 (35.8) |
| 60–120 | 93755 (19.8) | 661 (20.7) | 206 (19.9) | 455 (21) |
| 120–180 | 28052 (5.9) | 183 (5.7) | 37 (3.6) | 146 (6.7) |
| > 180 | 182363 (38.6) | 1163 (36.4) | 375 (36.3) | 788 (36.4) |
| Water hardness (WHO), n (%) | ||||
| < 200 | 292598 (61.9) | 2052 (64.2) | 664 (64.3) | 1388 (64.1) |
| ≥ 200 | 179755 (38.1) | 1146 (35.8) | 369 (35.7) | 777 (35.9) |
| CaCO3 concentration, mean (SD) | 145.68 (115.08) | 139.8 (112.89) | 137 (113.74) | 141.09 (112.48) |
| Ca concentration, mean (SD) | 51.1 (40.08) | 51.02 (40.32) | 49.33 (40.09) | 51.79 (40.42) |
| Mg concentration, mean (SD) | 4.58 (3.76) | 4.46 (3.45) | 4.35 (3.32) | 4.51 (3.52) |
Abbreviations: BMI Body mass index, IPAQ International Physical Activity Questionnaire, USGS United States Geological Survey, WHO World Health Organization
Hard water was significantly associated with a decreased risk of CD but not UC
In the cohort study, hard water (120–180 mg/l) and very hard water (> 180 mg/l), as classified by USGS, were found to be significantly associated with lower incidence of CD in all three models (see Table 5 for details, Model 3: HR = 0.84, 95% CI: 0.73–0.97, p = 0.016, and p = 0.007 for the trend). When analysing CaCO₃ concentration as a continuous variable, each SD increase in concentration was also significantly associated with a lower incidence of CD in three models (Model 3: HR = 0.93, 95% CI: 0.87–0.99, p = 0.017). However, when analysed according to the WHO classification, the results were not significant (see Table 5). Similarly, analysing Ca2⁺ concentration as a continuous variable did not show a significant association (see Table 5). While analyzing the overall incidence of IBD, hard water classified by WHO (HR = 0.92, 95% CI: 0.85–0.99, p = 0.021), and continuous CaCO₃ concentration (HR = 0.96, 95% CI: 0.92–0.99, p = 0.014) all showed significant associations (see Table 6). However, when UC was analysed separately, these associations were no longer significant (see Table 5).
Table 5.
Association between water hardness and incident UC and CD in the cohort study
| CD | UC | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cases | Model 1 HR (95% CI) | P | Model 2 HR (95% CI) | P | Model 3 HR (95% CI) | P | Cases | Model 1 HR (95% CI) | P | Model 2 HR (95% CI) | P | Model 3 HR (95% CI) | P | ||
| Water hardness (USGS, mg/l) | 0–60 | 415 | ref | ref | ref | 776 | ref | ref | ref | ||||||
| 60–120 | 206 | 0.89 (0.75–1.05) | 0.169 | 0.91 (0.77–1.07) | 0.252 | 0.90 (0.76–1.07) | 0.241 | 455 | 1.05 (0.93–1.18) | 0.430 | 1.06 (0.94–1.19) | 0.356 | 1.07 (0.95–1.20) | 0.260 | |
| 120–180 | 37 | 0.53 (0.38–0.75) | < 0.001 | 0.56 (0.40–0.78) | < 0.001 | 0.55 (0.39–0.77) | < 0.001 | 146 | 1.13 (0.95–1.35) | 0.172 | 1.15 (0.97–1.38) | 0.113 | 1.16 (0.97–1.39) | 0.094 | |
| > 180 | 375 | 0.84 (0.73–0.96) | 0.014 | 0.85 (0.74–0.98) | 0.024 | 0.84 (0.73–0.97) | 0.016 | 788 | 0.95 (0.86–1.04) | 0.275 | 0.95 (0.85–1.04) | 0.270 | 0.96 (0.87–1.07) | 0.469 | |
| P for trend | 0.007 | 0.012 | 0.007 | 0.258 | 0.262 | 0.452 | |||||||||
| Water hardness (WHO, mg/l) | < 200 | 664 | ref | ref | ref | 1388 | ref | ref | ref | ||||||
| ≥ 200 | 369 | 0.91 (0.8–1.03) | 0.152 | 0.91 (0.8–1.04) | 0.167 | 0.90 (0.79–1.03) | 0.126 | 777 | 0.92 (0.84–1.01) | 0.070 | 0.92 (0.84–1.00) | 0.054 | 0.93 (0.85–1.02) | 0.106 | |
| CaCO3 concentration (per SD) | 0.93 (0.87–0.99) | 0.020 | 0.93 (0.87–0.99) | 0.027 | 0.93 (0.87–0.99) | 0.017 | 0.97 (0.93–1.01) | 0.130 | 0.97 (0.93–1.01) | 0.118 | 0.97 (0.93–1.02) | 0.223 | |||
| Ca concentration (per SD) | 0.96 (0.9–1.02) | 0.174 | 0.95 (0.89–1.02) | 0.140 | 0.95 (0.89–1.01) | 0.104 | 1.03 (0.98–1.07) | 0.207 | 1.02 (0.98–1.07) | 0.310 | 1.03 (0.98–1.07) | 0.233 | |||
| Mg concentration (per SD) | 0.93 (0.87–1) | 0.056 | 0.95 (0.88–1.01) | 0.119 | 0.94 (0.88–1.01) | 0.110 | 0.98 (0.94–1.03) | 0.473 | 0.99 (0.95–1.04) | 0.693 | 1.00 (0.95–1.04) | 0.848 | |||
Abbreviations: USGS United States Geological Survey, WHO World Health Organization, HR hazards ratio, P p-values
Model 1: adjusted for age and sex
Model 2: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI and diet quality scores
Model 3: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, diet quality scores, the first five principal components of ancestry genetics and PRS scores
Bold font for USGS and WHO water hardness groups indicates p < 0.025, while bold font for other variables indicates p < 0.05
Table 6.
Association between water hardness and incident IBD in the cohort study
| IBD | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cases | Model 1 HR (95% CI) | P | Model 2 HR (95% CI) | P | Model 3 HR (95% CI) | P | ||
| Water hardness (USGS, mg/l) | 0–60 | 1193 | ref | ref | ref | |||
| 60–120 | 665 | 1.00 (0.91–1.1) | 0.949 | 1.01 (0.92–1.11) | 0.861 | 1.02 (0.92–1.12) | 0.757 | |
| 120–180 | 184 | 0.93 (0.79–1.08 | 0.330 | 0.95 (0.81–1.11) | 0.528 | 0.95 (0.81–1.11) | 0.525 | |
| > 180 | 1167 | 0.91 (0.84–0.99) | 0.022 | 0.91 (0.84–0.99) | 0.029 | 0.92 (0.84–1.00) | 0.040 | |
| P for trend | 0.014 | 0.019 | 0.025 | |||||
| Water hardness (WHO, mg/l) | < 200 | 2059 | ref | ref | ref | |||
| ≥ 200 | 1150 | 0.92 (0.85–0.99) | 0.020 | 0.91 (0.85–0.98) | 0.017 | 0.92 (0.85–0.99) | 0.021 | |
| CaCO3 concentration (per SD) | 0.95 (0.92–0.99) | 0.010 | 0.95 (0.92–0.99) | 0.010 | 0.96 (0.92–0.99) | 0.014 | ||
| Ca concentration (per SD) | 1.01 (0.97–1.04) | 0.748 | 1.00 (0.96–1.04) | 0.966 | 1.00 (0.96–1.04) | 0.950 | ||
| Mg concentration (per SD) | 0.97 (0.93–1.01) | 0.108 | 0.98 (0.94–1.02) | 0.254 | 0.98 (0.94–1.02) | 0.327 | ||
Abbreviations: USGS United States Geological Survey, WHO World Health Organization, HR hazards ratio, P p-values
Model 1: adjusted for age and sex
Model 2: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI and diet quality scores
Model 3: adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, diet quality scores, the first five principal components of ancestry genetics
Bold font for USGS and WHO water hardness groups indicates p < 0.025, while bold font for other variables indicates p < 0.05
A stronger protective association of hard water with CD incidence was observed among individuals with low genetic risk
We further examined the association between hard water and incident CD stratified by genetic risk. Among individuals with low genetic risk, significant inverse associations were observed for both hard water (HR = 0.35, 95% CI: 0.17–0.70, p = 0.003) and very hard water (HR = 0.74, 95% CI: 0.57–0.95, p = 0.019), based on the USGS classification (see Table 7). In contrast, the association in the high genetic risk group was weaker and not statistically significant after Bonferroni correction (HR = 0.66, 95% CI: 0.45–0.97, p = 0.035). No significant interaction was found between hard water and genetic risk, indicating that their effects on CD incidence were independent.
Table 7.
Association between water hardness and the risk of CD by genetic susceptibility
| Low genetic risk | High genetic risk | |||||
|---|---|---|---|---|---|---|
| P for interaction | HR (95% CI) | P | HR (95% CI) | P | ||
| Water hardness (USGS, mg/l) | 0–60 | ref | ref | ref | ||
| 60–120 | 0.250 | 0.87 (0.65–1.16) | 0.342 | 0.92 (0.75–1.14) | 0.211 | |
| 120–180 | 0.124 | 0.35 (0.17–0.70) | 0.003 | 0.66 (0.45–0.97) | 0.035 | |
| > 180 | 0.794 | 0.74 (0.57–0.95) | 0.019 | 0.89 (0.75–1.07) | 0.457 | |
| P for trend | 0.008 | 0.162 | ||||
| Water hardness (WHO, mg/l) | < 200 | ref | ref | ref | ||
| ≥ 200 | 0.427 | 0.83 (0.66–1.05) | 0.127 | 0.94 (0.80–1.10) | 0.446 | |
| CaCO3 concentration (per SD) | 0.348 | 0.88 (0.79–0.99) | 0.032 | 0.95 (0.88–1.02) | 0.171 | |
| Ca concentration (per SD) | 0.272 | 0.90 (0.80–1.01) | 0.085 | 0.97 (0.90–1.05) | 0.462 | |
| Mg concentration (per SD) | 0.655 | 0.02 (0.81–1.05) | 0.226 | 0.95 (0.88–1.04) | 0.285 | |
Abbreviations: USGS United States Geological Survey, WHO World Health Organization, HR hazards ratio, P p-values
Adjusted for age, sex, CCI scores, BMI, IPAQ physical activity, alcohol consumption, smoking status, ethnicity, TDI, diet quality scores, the first five principal components of ancestry genetics
Bold font for USGS and WHO water hardness groups indicates p < 0.025, while bold font for other variables indicates p < 0.05
Subgroup analyses, sensitivity analyses, and additional analyses
In the subgroup analysis, the association between hard water and the risk of CD and UC was further examined, stratifying by several factors (Supplementary Table S1-S10). Overall, stronger associations between hard water and the risk of CD were observed in the low-to-moderate IPAQ physical activity group (Supplementary Table S5), the group with alcohol consumption below weekly frequency (Supplementary Table S6), and among smokers (Supplementary Table S7). Although hard water showed no effect on the overall risk of UC, subgroup analysis revealed a possible protective effect among males (Supplementary Table S1), individuals with low-to-moderate physical activity (Supplementary Table S5), and non-smokers (Supplementary Table S7). To further investigate the inconsistency between the cross-sectional and cohort results for UC, we performed stratified analyses of UC prevalence by disease duration. The findings indicated that hard water was significantly associated with UC in participants with a longer disease duration (Supplementary Table S23). Sensitivity analyses yielded consistent results across multiple conditions, including multiple imputation of missing data, additional adjustments for potential confounders (e.g., C-reactive protein, sun exposure), exclusion of early IBD cases, replacement of genetic sex with self-reported sex, and restriction to participants who did not change residence within 10 years after inclusion (Supplementary Table S11-S20). When further adjusted assessment centers, hard water (120–180 mg/l) remained significantly associated with the incidence of CD (Supplementary Table S21). To account for the competing risk of mortality, a Fine-Gray subdistribution hazard model was applied for CD incidence, treating death as a competing event. The findings were consistent with the main analysis, with a significant p for trend (Supplementary Table S22).
Discussion
A large-scale population-based cross-sectional and cohort study using data from 481,468 UK Biobank participants was conducted. Our findings indicate that hard water exposure is associated with a reduced prevalence of IBD. Additionally, we report the association between hard water and a decreased incidence of CD in participants without IBD at baseline. However, hard water did not influence UC incidence.
The observed discrepancy between UC prevalence and UC incidence may be attributed to disease duration effects and potential survival bias. Our stratified analysis showed that the association between hard water and UC prevalence was significant among participants with longer disease duration but not in those with shorter disease duration. This suggests that hard water may influence disease progression or survival rather than the initial risk of developing UC.
Although previous research has shown that calcium intake may have protective benefits for IBD, the relationship between hard water and IBD remains unexplored. A previous case-controlled study reported that increased dietary calcium intake was correlated with a reduced prevalence of UC [21]. Another cohort study found that individuals who consumed milk exhibited a lower risk of developing CD [12]. However, the underlying mechanisms have not been fully understood.
Calcium is the most prevalent cation in the human body and acts as a ubiquitous second messenger, influencing various cellular functions such as differentiation, apoptosis, and proliferation [22]. Calcium is also closely related to barrier permeability and immune response [23–25]. The regulatory role of calcium ions at the cellular level in IBD has been well-documented, and dietary calcium intake has also been shown to alleviate colitis in mouse models [26]. Experimental studies have demonstrated that dietary calcium supplementation can alleviate colitis severity, potentially via suppression of the pro-inflammatory cytokine TNF-α and its upstream regulators [26]. Moreover, vitamin D has been shown to act synergistically with dietary calcium to reduce mucosal inflammation, further implicating the calcium-VD-TNF-α axis in intestinal immune modulation [26, 27].
Meanwhile, calcium supplementation may exert beneficial effects in IBD by reshaping the gut microbiota [28]. Experimental evidence supports a prebiotic-like role of calcium in shaping gut microbial composition by selectively enriching beneficial bacterial taxa such as Bifidobacterium spp. and Lactobacillus spp. These changes were associated with improved metabolic profiles, including lower plasma lipopolysaccharide [28].
The inconsistent results observed between UC and CD in our study may be partly attributed to differences in how these disease phenotypes respond to calcium-mediated pathways. While CD tends to involve transmural inflammation and systemic immune activation, UC is largely confined to mucosal layers and may be more dependent on local microenvironment [29]. Previous studies have demonstrated distinct differences between UC and CD in terms of microbiota composition, intestinal permeability, immune response pathways, and nutrient absorption, which may partly explain their heterogeneous responses to environmental factors such as calcium intake [30].
The possible protective effect of hard water on the risk of CD was significantly stronger in participants with lower genetic susceptibility to CD. The pathological processes of IBD are triggered by both genetic and environmental factors [4, 31]. Previous studies have identified numerous potential loci and their interactions with environmental factors [31]. Nutrition intake potentially interacts with susceptible gene loci, regulating the course of IBD through mechanisms such as gut microbiota and immune response [4]. Researchers also found that the impact of dietary iron intake on IBD was significantly modified by a coding variant in the FcγRIIA gene, providing evidence for an environmental metal-gene interaction in the progression of IBD [32]. However, the potential interaction effect between calcium and IBD susceptible gene variants remains unclear and warrants further investigation.
Water is a fundamental necessity for life. The associations between water quality and health-related outcomes carry significant public health implications [33]. Earlier research has examined the impact of hard water on different health conditions [11, 34–37]. A large-scale study found that hard water has a protective effect against several types of cancer, including gastrointestinal cancers such as esophageal, stomach, and colorectal cancers [11]. Several studies have focused on the relationship between water quality and IBD, which evaluated several metal ions and other quality indices [13, 14]. A study on UK tap water indicated that the effects of water composition on health may be mediated through modulation of the gut microbiota [38]. However, direct evidence on the association between hard water and IBD risk remains scarce. Compared to these prior studies, our analysis provides more specific insights into the potential role of calcium and magnesium levels in drinking water, adding novel evidence to this underexplored area.
In our study, although some associations reached statistical significance, the observed effect sizes were modest. Such modest associations are common in large-scale epidemiological studies of complex, multifactorial diseases like IBD, where genetic, environmental, and lifestyle factors interact to influence disease development. These small effect sizes may also reflect limitations in exposure assessment and potential residual confounding. Moreover, regional variation in hard water levels across the UK may overlap with other geographical patterns previously associated with IBD risk [39]. For instance, regional differences in the distribution of IBD prevalence across the UK have been reported in earlier studies, potentially reflecting differences in environmental exposures [40]. Further research using precise geographic data could better elucidate the influence of correlated environmental exposures. Nevertheless, even small effects may have meaningful implications at the population level, particularly when the exposure is widespread and potentially modifiable through public health interventions.
There are several strengths of our study. Firstly, our study is a population-based cross-sectional and cohort study with a large sample size and long-term follow-up. The combination of cross-sectional and longitudinal analyses provides more comprehensive evidence to support the conclusions. Secondly, to our knowledge, this is the first study to evaluate the associations between hard water and the prevalence and incidence of IBD. We adjusted potential confounding variables such as genetic risk, alcohol consumption, and smoking status and provided robust statistical evidence. Thirdly, we have considered the influence of genetic factors on IBD risk. The interaction effect between genetics and hard water was further analysed. We examined how water quality impacts groups with varying levels of genetic risk differently, providing evidence for the combined influence of environmental and genetic factors in the pathogenesis of IBD.
However, there are still limitations to our study. Firstly, hard water exposure was assigned based on residential postcode at baseline, which may not reflect individual-level consumption patterns or temporal changes during the follow-up. In addition, we lacked data on bottled water consumption and the use of household water filtration systems, both of which could influence actual exposure. These unmeasured behaviours may contribute to exposure misclassification. However, such misclassification is likely to be non-differential and would typically bias results toward the null, meaning the true associations may be underestimated. Secondly, because the participants in this study were primarily of White British descent, the generalizability of the conclusions to other ethnic groups is unclear. Thirdly, the participants in this study were aged 40–69, which may not fully reflect the population of individuals with an early onset of disease. Additionally, the absence of data on disease severity limited our ability to assess whether the associations varied by disease activity or progression, which represents an important limitation. Besides, although we adjusted for assessment centers in the sensitivity analyses, we were unable to include precise residential latitude due to data limitations, which may confound the association between hard water and IBD, given known geographical gradients in environmental exposures [9, 41]. While assessment centers may serve as a rough proxy for geographic location, they cannot fully substitute for individual-level geospatial data, nor do they adequately capture the distribution of latitude-related environmental exposures. The direction and magnitude of this potential bias are difficult to quantify, but we acknowledge that this limitation may affect the accuracy of our causal inferences. Future studies with detailed geographic data are needed to better control environmental gradients and clarify the role of environmental determinants in IBD. We also acknowledge the possibility that some findings may be due to chance, given the observational nature of the study.
In conclusion, we observed associations between hard water and IBD risk. Specifically, hard water exposure was associated with a reduced prevalence of both CD and UC and a lower incidence of CD, but not with UC incidence. These findings underscore the potential role of hard water as an environmental determinant of IBD, adding to the understanding of its multifactorial etiology. As hard water is a common and modifiable exposure, these findings may inform public health efforts to address environmental contributors to IBD. Clinicians could incorporate water hardness information into risk assessments and preventive counseling for patients at higher risk of IBD, while policymakers might consider this evidence in the formulation of water quality regulations. Integrating such evidence into prevention strategies could ultimately contribute to reducing the incidence and prevalence of IBD at the population level. As with all observational studies, potential limitations such as residual confounding and measurement constraints should be considered when interpreting these findings. Further research with refined approaches is warranted to clarify the contribution of environmental exposures, including water hardness, to IBD risk.
Supplementary Information
Acknowledgements
This study was conducted under the application number 176969 of the UK Biobank. We sincerely acknowledge the contributions of the participants and the UK Biobank team.
Abbreviations
- BMI
Body mass index
- CCI
Charlson Comorbidity Index
- CD
Crohn’s disease
- GWAS
Genome-wide association study
- HR
Hazard ratios
- IBD
Inflammatory bowel diseases
- ICD
International Classification of Diseases
- IPAQ
International Physical Activity Questionnaire
- OR
Odds ratios
- PRS
Polygenic risk scores
- TDI
Townsend Deprivation Index
- UC
Ulcerative colitis
- USGS
United States Geological Survey
- WHO
World Health Organisation
- 95% CI
95% Confidence intervals
Authors’ contributions
K.Y. conducted data processing, performed data analysis, edited tables, and wrote the original draft. T.Z. and Z.J. contributed to methodology, data processing, and manuscript editing. W.J. and Z.L. were involved in methodology and manuscript editing. Z.S. provided supervision, secured funding, conceptualized the study, and managed project administration. All authors reviewed the manuscript.
Funding
This study was funded by the High-Level Talent Science and Technology Innovation Leading Talent Program of Zhejiang Province, reference number: 2023R5239.
Data availability
All data analysed during this study are included in this article and its supplementary information files.
Declarations
Ethics approval and consent to participate
The UK Biobank received ethical approval from the North West-Haydock Research Ethics Committee (REC reference: 16/NW/0274). Informed consent was obtained from all participants at the time of recruitment.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Zhao M, Gönczi L, Lakatos PL, Burisch J. The burden of inflammatory bowel disease in Europe in 2020. J Crohns Colitis. 2021;15:1573–87. 10.1093/ecco-jcc/jjab029. [DOI] [PubMed] [Google Scholar]
- 2.Adolph TE, Meyer M, Schwärzler J, Mayr L, Grabherr F, Tilg H. The metabolic nature of inflammatory bowel diseases. Nat Rev Gastroenterol Hepatol. 2022;19:753–67. 10.1038/s41575-022-00658-y. [DOI] [PubMed] [Google Scholar]
- 3.Agrawal M, Allin KH, Petralia F, Colombel J-F, Jess T. Multiomics to elucidate inflammatory bowel disease risk factors and pathways. Nat Rev Gastroenterol Hepatol. 2022;19:399–409. 10.1038/s41575-022-00593-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Noble AJ, Nowak JK, Adams AT, Uhlig HH, Satsangi J. Defining interactions between the genome, epigenome, and the environment in inflammatory bowel disease: progress and prospects. Gastroenterology. 2023;165:44-60.e2. 10.1053/j.gastro.2023.03.238. [DOI] [PubMed] [Google Scholar]
- 5.Ananthakrishnan AN, Bernstein CN, Iliopoulos D, Macpherson A, Neurath MF, Ali RAR, et al. Environmental triggers in IBD: a review of progress and evidence. Nat Rev Gastroenterol Hepatol. 2018;15:39–49. 10.1038/nrgastro.2017.136. [DOI] [PubMed] [Google Scholar]
- 6.Jairath V, Feagan BG. Global burden of inflammatory bowel disease. Lancet Gastroenterol Hepatol. 2020;5:2–3. 10.1016/S2468-1253(19)30358-9. [DOI] [PubMed] [Google Scholar]
- 7.Zhao M, Feng R, Ben-Horin S, Zhuang X, Tian Z, Li X, et al. Systematic review with meta-analysis: environmental and dietary differences of inflammatory bowel disease in Eastern and Western populations. Aliment Pharmacol Ther. 2022;55:266–76. 10.1111/apt.16703. [DOI] [PubMed] [Google Scholar]
- 8.Caruso R, Lo BC, Núñez G. Host–microbiota interactions in inflammatory bowel disease. Nat Rev Immunol. 2020;20:411–26. 10.1038/s41577-019-0268-7. [DOI] [PubMed] [Google Scholar]
- 9.Yang AZ, Jostins-Dean L. Environmental variables and genome-environment interactions predicting IBD diagnosis in large UK cohort. Sci Rep. 2022;12:10890. 10.1038/s41598-022-13222-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lopez DJ, Singh A, Waidyatillake NT, Su JC, Bui DS, Dharmage SC, et al. The association between domestic hard water and eczema in adults from the UK biobank cohort study. Br J Dermatol. 2022;187:704–12. 10.1111/bjd.21771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yang H, Wang Q, Zhang S, Zhang J, Zhang Y, Feng J. Association of domestic water hardness with all-cause and cause-specific cancers: evidence from 447,996 UK Biobank participants. Environ Health Perspect. 2024;132:067008. 10.1289/EHP13606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Opstelten JL, Leenders M, Dik VK, Chan SSM, van Schaik FDM, Khaw K-T, et al. Dairy products, dietary calcium, and risk of inflammatory bowel disease: results from a European prospective cohort investigation. Inflamm Bowel Dis. 2016;22:1403–11. 10.1097/MIB.0000000000000798. [DOI] [PubMed] [Google Scholar]
- 13.Aamodt G, Bukholm G, Jahnsen J, Moum B, Vatn MH, the IBSEN Study Group. The association between water supply and inflammatory bowel disease based on a 1990–1993 cohort study in southeastern Norway. Am J Epidemiol. 2008;168:1065–72. 10.1093/aje/kwn218. [DOI] [PubMed] [Google Scholar]
- 14.Bagherzadeh F, Karami Horestani M, Sadeghi M, Ahmadi A, Bahreini R, Fadaei A, et al. Influence of metal ions concentration in drinking water in the development of ulcerative colitis. Int J Environ Sci Technol. 2022;19:3539–46. 10.1007/s13762-021-03498-6. [Google Scholar]
- 15.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12:e1001779. 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Craig CL, Marshall AL, Sjöström M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35:1381–95. 10.1249/01.MSS.0000078924.61453.FB. [DOI] [PubMed] [Google Scholar]
- 17.Guillaume E, Pornet C, Dejardin O, Launay L, Lillini R, Vercelli M, et al. Development of a cross-cultural deprivation index in five European countries. J Epidemiol Community Health. 2015;70:493. 10.1136/jech-2015-205729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Choi SW, Mak TSH, O’Reilly PF. A guide to performing Polygenic Risk Score analyses. Nat Protoc. 2020;15:2759–72. 10.1038/s41596-020-0353-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhuang P, Liu X, Li Y, Wan X, Wu Y, Wu F, et al. Effect of diet quality and genetic predisposition on hemoglobin A1c and type 2 diabetes risk: gene-diet interaction analysis of 357,419 individuals. Diabetes Care. 2021;44:2470–9. 10.2337/dc21-1051. [DOI] [PubMed] [Google Scholar]
- 20.van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45: 1–67. 10.18637/jss.v045.i03. [Google Scholar]
- 21.Hajhashemy Z, Saneei P, Keshteli AH, Daghaghzadeh H, Tavakkoli H, Adibi P, et al. A population based case–control study of association between dietary calcium intake and ulcerative colitis in adults. Sci Rep. 2022;12:7913. 10.1038/s41598-022-11597-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Berridge MJ, Lipp P, Bootman MD. The versatility and universality of calcium signalling. Nat Rev Mol Cell Biol. 2000;1:11–21. 10.1038/35036035. [DOI] [PubMed] [Google Scholar]
- 23.Dalal PJ, Muller WA, Sullivan DP. Endothelial cell calcium signaling during barrier function and inflammation. Am J Pathol. 2020;190:535–42. 10.1016/j.ajpath.2019.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Schepens MAA, Bruggencate SJM, Schonewille AJ, Brummer R-JM, Meer R, Bovee-Oudenhoven IMJ. The protective effect of supplemental calcium on colonic permeability depends on a calcium phosphate-induced increase in luminal buffering capacity. Br J Nutr. 2012;107:950–6. 10.1017/S0007114511003977. [DOI] [PubMed] [Google Scholar]
- 25.Owen JL, Cheng SX, Ge Y, Sahay B, Mohamadzadeh M. The role of the calcium-sensing receptor in gastrointestinal inflammation. Semin Cell Dev Biol. 2016;49:44–51. 10.1016/j.semcdb.2015.10.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhu Y, Mahon BD, Froicu M, Cantorna MT. Calcium and 1 alpha,25-dihydroxyvitamin D3 target the TNF-alpha pathway to suppress experimental inflammatory bowel disease. Eur J Immunol. 2005;35:217–24. 10.1002/eji.200425491. [DOI] [PubMed] [Google Scholar]
- 27.Ananthakrishnan AN, Khalili H, Higuchi LM, Bao Y, Korzenik JR, Giovannucci EL, et al. Higher predicted vitamin D status is associated with reduced risk of Crohn’s disease. Gastroenterology. 2012;142:482–9. 10.1053/j.gastro.2011.11.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chaplin A, Parra P, Laraichi S, Serra F, Palou A. Calcium supplementation modulates gut microbiota in a prebiotic manner in dietary obese mice. Mol Nutr Food Res. 2016;60:468–80. 10.1002/mnfr.201500480. [DOI] [PubMed] [Google Scholar]
- 29.Kostic AD, Xavier RJ, Gevers D. The microbiome in inflammatory bowel diseases: current status and the future ahead. Gastroenterology. 2014;146:1489–99. 10.1053/j.gastro.2014.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sartor RB. Mechanisms of disease: pathogenesis of Crohn’s disease and ulcerative colitis. Nat Clin Pract Gastroenterol Hepatol. 2006;3:390–407. 10.1038/ncpgasthep0528. [DOI] [PubMed] [Google Scholar]
- 31.Graham DB, Xavier RJ. Pathway paradigms revealed from the genetics of inflammatory bowel disease. Nature. 2020;578:527–39. 10.1038/s41586-020-2025-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Khalili H, de Silva PS, Ananthakrishnan AN, Lochhead P, Joshi A, Garber JJ, et al. Dietary iron and heme iron consumption, genetic susceptibility, and risk of Crohnʼs disease and ulcerative colitis. Inflamm Bowel Dis. 2017;23:1088–95. 10.1097/MIB.0000000000001161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sengupta P. Potential health impacts of hard water. Int J Prev Med. 2013;4:866–75. [PMC free article] [PubMed] [Google Scholar]
- 34.Jabbar-Lopez ZK, Craven J, Logan K, Greenblatt D, Marrs T, Radulovic S, et al. Longitudinal analysis of the effect of water hardness on atopic eczema: evidence for gene–environment interaction. Br J Dermatol. 2020;183:285–93. 10.1111/bjd.18597. [DOI] [PubMed] [Google Scholar]
- 35.Moghimani M, Rahati S, Sobhani SR, Hashemi M, Afshari A. Evaluation of the relationship between water hardness and incidence of cardiovascular diseases in Mashhad, Iran during 2018–2020: a cross-sectional study. Health Sci Rep. 2025;8:e70376. 10.1002/hsr2.70376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Zhang J, Luo H, Wu H, Qian Y, Tang Z, Wang J, et al. The association between domestic water hardness and kidney stone disease: a prospective cohort study from the UK Biobank. Int J Surg. 2025;111:1957–67. 10.1097/JS9.0000000000002198. [DOI] [PubMed] [Google Scholar]
- 37.Mao R, Zhang T, Zhong Y, Meng X, Yi C, Wang F, et al. Associations between domestic hard water exposure and incident psoriasis in adults: insights from the UK Biobank cohort study. J Autoimmun. 2025;151:103373. 10.1016/j.jaut.2025.103373. [DOI] [PubMed] [Google Scholar]
- 38.Bowyer RCE, Schillereff DN, Jackson MA, Le Roy C, Wells PM, Spector TD, et al. Associations between UK tap water and gut microbiota composition suggest the gut microbiome as a potential mediator of health differences linked to water quality. Sci Total Environ. 2020;739:139697. 10.1016/j.scitotenv.2020.139697. [DOI] [PubMed] [Google Scholar]
- 39.Piovani D, Danese S, Peyrin-Biroulet L, Nikolopoulos GK, Lytras T, Bonovas S. Environmental risk factors for inflammatory bowel diseases: an umbrella review of meta-analyses. Gastroenterology. 2019;157:647-659.e4. 10.1053/j.gastro.2019.04.016. [DOI] [PubMed] [Google Scholar]
- 40.Pasvol TJ, Horsfall L, Bloom S, Segal AW, Sabin C, Field N, et al. The burden of inflammatory bowel disease in Europe in 2020. BMJ Open. 2020;10:e036584. 10.1136/bmjopen-2019-036584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Piovani D, Danese S, Peyrin-Biroulet L, Bonovas S. Environmental, nutritional, and socioeconomic determinants of IBD incidence: a global ecological study. J Crohns Colitis. 2020;14:323–31. 10.1093/ecco-jcc/jjz150. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data analysed during this study are included in this article and its supplementary information files.