Long-term risk of irritable bowel syndrome associated with social isolation and loneliness: a large-scale prospective cohort study with mediation analysis

Abstract

Background:

Social isolation and loneliness are common among older adults and have been linked to metabolic disorders, psychological conditions, and unhealthy lifestyle behaviors. However, their associations with incident irritable bowel syndrome (IBS) remain unclear.

Objective:

To investigate the associations of social isolation and loneliness with incident IBS risk, and the mediating roles of anxiety or depression.

Design:

Population-based prospective cohort study.

Methods:

Participants without IBS and with available social isolation and loneliness data at recruitment were included (N = 394,458). Social isolation was assessed using household size, social contact frequency, and weekly social activities, while loneliness was measured by self-reported feelings of loneliness and willingness to confide. The primary outcome was incident IBS. Cox regression with sequential mediation analysis was conducted to estimate the effect.

Results:

During a median follow-up of 14.5 years, 8307 (2.1%) incident IBS cases were identified. Overall, 209,081 (53.0%), 149,729 (38.0%), and 35,648 (9.0%) participants were classified as least, moderately, and most isolated, respectively. The 14-year cumulative incidence of IBS increased from 2.0% (95% confidence interval (CI): 2.0–2.1) in the least isolated group to 2.2% (95% CI: 2.2–2.3) and 2.6% (95% CI: 2.4–2.8) in the moderately and most isolated groups (population-attributable fraction (PAF) = 3.6%). Compared with the least isolated group, most isolated individuals had a 20.0% higher IBS risk (hazard ratio (HR) = 1.20, 95% CI: 1.11–1.29), with 24.7% of the effect mediated by depression or anxiety. The 14-year cumulative incidence of IBS was 3.1% (95% CI: 2.9–3.4) in the lonely group versus 2.1% (95% CI: 2.1–2.1) in the non-lonely group (PAF = 2.1%). Compared with non-lonely individuals (95.2%), lonely individuals (4.8%) showed a 46.0% increased IBS risk (HR = 1.46, 95% CI: 1.34–1.59), with 26.5% of the effect mediated by depression or anxiety. In addition, one increment in social isolation score (HR = 1.08, 95% CI: 1.04–1.11) and loneliness score (HR = 1.27, 95% CI: 1.23–1.32) were associated with 8.0% and 27.0% higher IBS risk, respectively. Particularly, most isolated and lonely individuals exhibited a 60.0% greater IBS risk versus their non-lonely and least isolated counterparts (HR = 1.60, 95% CI: 1.35–1.89).

Conclusion:

Both social isolation and loneliness are associated with increased IBS risk, partially mediated by depression or anxiety. Multi-level psychosocial and environmental interventions may help reduce IBS burden.

Keywords

cohort studies irritable bowel syndrome loneliness mediation analysis social isolation

Introduction

Irritable bowel syndrome (IBS) is a prevalent gut-brain interaction disorder, characterized by recurrent abdominal pain or discomfort and altered bowel habits. It affects approximately 10% of the global population, severely impairs health-related quality of life, and imposes substantial economic burdens on both individuals and society.¹ Given the incomplete understanding of IBS pathophysiology and the lack of effective treatments,² identifying modifiable risk factors is crucial for improving prevention and management strategies. Among these, social determinants of health are increasingly recognized as important contributors to IBS risk.³

Social isolation, defined as the objective absence of social connections, and loneliness, characterized by the subjective feeling of disconnection, are common among older adults.⁴ Meta-analyses indicate that around 25% of older adults experience social isolation, while loneliness affects 5.2%–21.3% of this population.^5,6 Both conditions have been linked to heightened risks of metabolic and psychological disorders, as well as unhealthy behaviors such as tobacco use, physical inactivity, and poor diet, all of which may contribute to the development of IBS.^7
–9 A Norwegian twin study revealed a positive correlation between IBS risk and low social support (e.g., a lack of comfort, help, or understanding), with shared genetic factors largely accounting for this relationship.¹⁰ Furthermore, experimental studies suggest that social disconnection and loneliness may adversely impact intestinal health through chronic stress-induced dysregulation of the gut-brain axis, increased proinflammatory cytokine activity, gut dysbiosis, impaired gut barrier integrity, and visceral hypersensitivity.^11
–14 However, to date, no studies have investigated the risk of incident IBS associated with social isolation and loneliness, nor have they explored the potential mediating effects of anxiety or depression in these associations.

To address these knowledge gaps, we aimed to examine the independent and joint associations of social isolation and loneliness with the risk of incident IBS, as well as the mediating roles of depression or anxiety, in a large population-based prospective cohort study.

Methods

Study population

The UK Biobank (UKB) is a large, population-based cohort study that enrolled over 500,000 participants aged 37–73 years from England, Scotland, and Wales between 2006 and 2010. At recruitment, all participants provided written informed consent, completed baseline questionnaires, and underwent a range of physical and biochemical tests. Ethical approval was obtained from the North West Multi-centre Research Ethics Committee (21/NW/0157).

Participants free of IBS with available social isolation and loneliness data at recruitment were included. Those who withdrew (N = 34) or had a diagnosis of cancer (N = 43,519), inflammatory bowel disease (N = 4843), or coeliac disease (N = 2669) at enrollment were excluded. All disease diagnoses were determined via the International Classification of Diseases, 10th Revision (ICD-10; Table S1). Finally, 394,458 participants were included (Figure S1).

Assessment of social isolation

Social isolation at baseline was assessed using a touchscreen questionnaire comprising three items adapted from the Berkman-Syme Social Network Index (SNI).¹⁵ As a formative index, the SNI has a sound conceptual basis and established predictive validity for multiple health outcomes in population-based studies.^16,17 Although internal consistency statistics are rarely reported, related social network measures show moderate test-retest reliability, supporting their use as a population-level indicator of social integration.¹⁸ The questionnaire included the following items: (i) “Including yourself, how many people are living together in your household?” Responses were scored as 0 for “More than one person” and 1 for “Only one person”; (ii) “How often do you visit friends or family, or have them visit you?” Frequent interactions (“Almost daily,” “2–4 times a week,” “Once a week,” or “Once a month”) were scored as 0, whereas infrequent interactions (“Once every few months,” “Never or almost never,” or “No friends or family outside the household”) were scored as 1; (iii) “Which of the following do you participate in once a week or more often: sports club, gym, pub, religious group, adult education class, or other?” Participation in any of these activities was scored as 0, while selecting “None of the above” was scored as 1. Further details are provided in Supplemental Method 1.

An overall social isolation score was calculated by summing the scores of the three items (range: 0–3), with higher scores reflecting greater social isolation. Participants were then categorized into three groups based on their scores: least isolated (score = 0), moderately isolated (score = 1), and most isolated (score = 2 or 3).

Assessment of loneliness

Loneliness was evaluated using two questions adapted from the revised University of California, Los Angeles Loneliness Scale, a widely used instrument with established reliability and validity.¹⁹ Shortened versions also demonstrate good reliability (Cronbach’s α ≈ 0.87–0.90) and high item-total correlations (generally >0.69), supporting their use in population-based studies.^20,21 The questionnaire included the following items: (i) “Do you often feel lonely?” (0 points for “No,” 1 point for “Yes”); and (ii) “How often are you able to confide in someone close to you?” (0 points for responses such as “Almost daily,” “2–4 times a week,” “Once a week,” “Once a month,” or “Once every few months,” 1 point for “Never or almost never”).

A total loneliness score was obtained by summing the responses to the two items (range: 0–2), with higher scores indicating greater loneliness. Participants were then categorized into two groups according to their scores: no loneliness (score = 0 or 1) and loneliness (score = 2).

Ascertainment of IBS

The primary outcome was incident IBS (ICD-10 code K58), with a censoring date of September 30, 2023. Incident IBS was identified through self-reports or linkage to primary care and/or hospital admission records across the UK. Additionally, Rome III criteria via the digestive healthcare questionnaire (DHQ) during 2017–2018 were also used to ascertain incident IBS in sensitivity analysis (see Supplemental Method 2 for details).

Ascertainment of depression and anxiety

Depression and anxiety were ascertained at baseline using ICD-10 codes (F32 for depression; F40 and F41 for anxiety). Both the exposures (social isolation/loneliness) and the proposed mediators (depression/anxiety) were assessed at baseline, whereas incident IBS was ascertained during follow-up.

Covariates

Baseline covariates were selected based on previous epidemiological evidence and data availability^22,23: age (continuous), sex (male, female), Townsend Deprivation Index (TDI; quartiles), education level (non-university, university), ethnicity (White, non-White), smoking status (never, previous, current), alcohol drinking (never, previous, current), healthy diet (yes, no), physical activity (low, moderate, high), body mass index (BMI; <18.5, 18.5–24.9, 25.0–29.9, ⩾30 kg/m²), and type 2 diabetes mellitus (T2DM; yes, no). A healthy diet was defined as adherence to at least four of the seven commonly consumed food groups (i.e., fruits, vegetables, fish, processed meats, unprocessed red meats, whole grains, and refined grains), with further details provided in Supplemental Method 3.²⁴ Physical activity levels were categorized according to the International Physical Activity Questionnaire.

Statistical analysis

The 14-year cumulative incidence of IBS was estimated using the Kaplan–Meier method. Absolute risk differences were calculated by subtracting the 14-year cumulative incidence in the reference group from that in the exposed group. For social isolation, comparisons were made between the least isolated group and the moderately or most isolated groups, whereas loneliness was analyzed by comparing non-lonely and lonely participants. The population-attributable fraction (PAF) was estimated from the prevalence of the exposure in the study population and the corresponding adjusted hazard ratio (HR) using the standard formula PAF = [P_e(HR − 1)]/[1 + P_e(HR − 1)], where P_e denotes the prevalence of the exposure. Given the low incidence of IBS in this cohort, the adjusted HR was used as an approximation of the relative risk.

Cox proportional hazards models were employed to investigate the risk of incident IBS associated with social isolation, loneliness, and their joint effects. The proportional hazards assumption was tested using Schoenfeld residuals, and no violations were found (all p > 0.05). The follow-up period started from the date of enrollment to the date of first IBS diagnosis, death, loss to follow-up, or the end of the study (September 30, 2023), whichever came first. In addition to the univariable analysis, two multivariable models were applied: Model 1 adjusted for age and sex, while Model 2 additionally adjusted for TDI, education level, ethnicity, smoking status, alcohol drinking, healthy diet, physical activity, BMI, and T2DM. Because the proportion of missing values was very small for most covariates (<1%), missing indicator categories were created to retain participants in the analyses and minimize loss of sample size. Given the low level of missingness, this approach was considered unlikely to materially affect the results.

Mediation analysis was performed using the SAS CAUSALMED procedure to estimate HRs for the natural direct (HR_NDE) and indirect effect (HR_NIE) of social isolation, loneliness, and their joint effects on incident IBS, as well as the proportion of the association mediated by depression/anxiety (binary; coded as 1 if either depression or anxiety was present, and 0 otherwise) via adjusted Model 2. The causal mediation analysis was conducted under the counterfactual framework, which assumes no unmeasured confounding of the exposure-outcome, exposure-mediator, and mediator-outcome relationships, and no mediator-outcome confounder affected by the exposure. To make the assumed causal structure explicit, a directed acyclic graph illustrating the hypothesized relationships among the exposure, mediator, outcome, and potential confounders was presented in Figure S2.

Subgroup analyses were conducted, stratified by age (<60, ⩾60 years), sex (male, female), education level (university, non-university), TDI (<−2.18, ⩾−2.18), smoking status (never, previous/current), drinking status (never/previous, current), healthy diet (no, yes), and BMI (<25, ⩾25 kg/m²). Potential effect modifications were examined by including a cross-product interaction term as an additional independent variable in Model 2.

Several sensitivity analyses were performed to test the robustness of our findings, including: (i) excluding participants diagnosed with IBS within 1 or 2 years after recruitment to minimize reverse causation; (ii) to account for the potential underdiagnosis of IBS, additionally considering those who fulfilled Rome III criteria via DHQ in 2017–2018 but had never self-reported an IBS diagnosis as incident IBS²⁵; (iii) to further rule out the possible misclassification bias of IBS diagnosis, excluding those who fulfilled Rome III criteria via DHQ without an incident ICD-10 diagnosis (i.e., considering these IBS cases as prevalent cases at baseline)²⁵; (iv) considering the different sources of IBS diagnosis codes in the cohort (i.e., death registry, primary care, hospital admissions, and self-reports), restricting IBS cases to those with IBS codes from at least two independent data sources to ensure diagnostic accuracy, while those with only one source of IBS code were considered non-IBS in the analysis; (v) further excluding individuals with only one source of IBS code from the analysis to account for potential misclassification bias; (vi) applying the Fine-Gray competing risk model, with death and loss to follow-up considered as competing events; and (vii) excluding participants with depression/anxiety at baseline and treating newly developed depression/anxiety during follow-up and prior to IBS diagnosis as the mediator.

All statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA) and R version 4.3.0 (Institute for Statistics and Mathematics, Vienna, Austria), with a two-tailed p value <0.05 considered statistically significant.

The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.²⁶

Results

Baseline characteristics

Overall, 209,081 (53.0%), 149,729 (38.0%), and 35,648 (9.0%) participants were classified as least, moderately, and most isolated, respectively (Table 1). Compared with participants in the least or moderately isolated groups, those in the most isolated group were more likely to be male, have lower educational attainment, greater socioeconomic deprivation, unhealthier lifestyles (i.e., more current smoking, poorer diet, and less physical activity), and a higher prevalence of obesity, depression, anxiety, and T2DM. Regarding loneliness, 375,689 (95.2%) participants were classified as non-lonely, and 18,769 (4.8%) as lonely, with the lonely group similarly exhibiting less favorable characteristics.

Table 1.

Baseline characteristics according to social isolation and loneliness status in the UK Biobank cohort.

Characteristics	Total (N = 394,458)	Social isolation				Loneliness
Characteristics	Total (N = 394,458)	Least isolated (N = 209,081)	Moderately isolated (N = 149,729)	Most isolated (N = 35,648)	p Value	Non-lonely (N = 375,689)	Lonely (N = 18,769)	p Value
Age (years)^a	56.21 (8.10)	56.25 (8.12)	56.11 (8.13)	56.33 (7.82)	<0.001	56.23 (8.11)	55.68 (7.94)	<0.001
Sex					<0.001			<0.001
Male	187,908 (47.6)	99,637 (47.7)	70,157 (46.9)	18,114 (50.8)		177,821 (47.3)	10,087 (53.7)
Female	206,550 (52.4)	109,444 (52.3)	79,572 (53.1)	17,534 (49.2)		197,868 (52.7)	8682 (46.3)
Ethnicity					<0.001			<0.001
Non-White	19,586 (5.0)	9508 (4.5)	7644 (5.1)	2434 (6.8)		18,374 (4.9)	1212 (6.5)
White	373,741 (94.7)	199,053 (95.2)	141,649 (94.6)	33,039 (92.7)		356,275 (94.8)	17,466 (93.1)
Unknown	1131 (0.3)	520 (0.2)	436 (0.3)	175 (0.5)		1040 (0.3)	91 (0.5)
Education level					<0.001			<0.001
Non-university	259,445 (65.8)	133,937 (64.1)	101,060 (67.5)	24,448 (68.6)		245,416 (65.3)	14,029 (74.7)
University	132,062 (33.5)	73,779 (35.3)	47,445 (31.7)	10,838 (30.4)		127,533 (33.9)	4529 (24.1)
Unknown	2951 (0.7)	1365 (0.7)	1224 (0.8)	362 (1.0)		2740 (0.7)	211 (1.1)
TDI					<0.001			<0.001
Mean (SD)	−1.35 (3.06)	−1.80 (2.80)	−1.07 (3.15)	0.05 (3.50)	<0.001	−1.40 (3.03)	−0.38 (3.47)	<0.001
Q1 (⩽−3.67)	98,521 (25.0)	59,565 (28.5)	33,618 (22.5)	5338 (15.0)		95,050 (25.3)	3471 (18.5)
Q2 (−3.67 to −2.18)	98,471 (25.0)	56,804 (27.2)	35,149 (23.5)	6518 (18.3)		94,610 (25.2)	3861 (20.6)
Q3 (−2.18 to 0.44)	98,476 (25.0)	51,366 (24.6)	38,352 (25.6)	8758 (24.6)		93,974 (25.0)	4502 (24.0)
Q4 (>0.44)	98,486 (25.0)	41,134 (19.7)	42,386 (28.3)	14,966 (42.0)		91,586 (24.4)	6900 (36.8)
Unknown	504 (0.1)	212 (0.1)	224 (0.1)	68 (0.2)		469 (0.1)	35 (0.2)
Smoking status					<0.001			<0.001
Never	216,693 (54.9)	119,106 (57.0)	80,386 (53.7)	17,201 (48.3)		207,650 (55.3)	9043 (48.2)
Previous	135,198 (34.3)	72,663 (34.8)	50,934 (34.0)	11,601 (32.5)		128,908 (34.3)	6290 (33.5)
Current	41,378 (10.5)	16,777 (8.0)	17,907 (12.0)	6694 (18.8)		38,025 (10.1)	3353 (17.9)
Unknown	1189 (0.3)	535 (0.3)	502 (0.3)	152 (0.4)		1106 (0.3)	83 (0.4)
Alcohol drinking					<0.001			<0.001
Never	15,760 (4.0)	7355 (3.5)	6355 (4.2)	2050 (5.8)		14,796 (3.9)	964 (5.1)
Previous	13,312 (3.4)	4966 (2.4)	5803 (3.9)	2543 (7.1)		12,229 (3.3)	1083 (5.8)
Current	365,110 (92.6)	196,662 (94.1)	137,454 (91.8)	30,994 (86.9)		348,423 (92.7)	16,687 (88.9)
Unknown	276 (0.1)	98 (0.0)	117 (0.1)	61 (0.2)		241 (0.1)	35 (0.2)
Healthy diet (n, %)^b	287,631 (72.9)	156,047 (74.6)	107,280 (71.6)	24,304 (68.2)	<0.001	275,964 (73.5)	11,667 (62.2)	<0.001
Healthy diet score^a,b	4.46 (1.63)	4.53 (1.60)	4.41 (1.65)	4.29 (1.70)	<0.001	4.48 (1.62)	4.02 (1.73)	<0.001
IPAQ					<0.001			<0.001
Low	57,169 (14.5)	25,231 (12.1)	24,907 (16.6)	7031 (19.7)		53,652 (14.3)	3517 (18.7)
Moderate	127,320 (32.3)	68,114 (32.6)	47,980 (32.0)	11,226 (31.5)		122,068 (32.5)	5252 (28.0)
High	128,996 (32.7)	76,246 (36.5)	44,016 (29.4)	8734 (24.5)		123,656 (32.9)	5340 (28.5)
Unknown	80,973 (20.5)	39,490 (18.9)	32,826 (21.9)	8657 (24.3)		76,313 (20.3)	4660 (24.8)
BMI (kg/m²)					<0.001			<0.001
Mean (SD)	27.45 (4.77)	27.25 (4.50)	27.62 (4.95)	27.90 (5.41)	<0.001	27.39 (4.72)	28.49 (5.48)	<0.001
<18.5	1831 (0.5)	784 (0.4)	751 (0.5)	296 (0.8)		1706 (0.5)	125 (0.7)
18.5–24.9	126,664 (32.1)	68,335 (32.7)	47,285 (31.6)	11,044 (31.0)		121,703 (32.4)	4961 (26.4)
25.0–29.9	168,060 (42.6)	92,254 (44.1)	62,108 (41.5)	13,698 (38.4)		160,651 (42.8)	7409 (39.5)
⩾30.0	95,995 (24.3)	46,881 (22.4)	38,778 (25.9)	10,336 (29.0)		89,869 (23.9)	6126 (32.6)
Unknown	1908 (0.5)	827 (0.4)	807 (0.5)	274 (0.8)		1760 (0.5)	148 (0.8)
Depression (n, %)	30,090 (7.6)	13,226 (6.3)	12,602 (8.4)	4262 (12.0)	<0.001	26,805 (7.1)	3285 (17.5)	<0.001
Anxiety (n, %)	13,596 (3.4)	6331 (3.0)	5645 (3.8)	1620 (4.5)	<0.001	12,409 (3.3)	1187 (6.3)	<0.001
T2DM (n, %)	9888 (2.5)	4306 (2.1)	4180 (2.8)	1402 (3.9)	<0.001	9019 (2.4)	869 (4.6)	<0.001

Categorical variables are presented as frequencies and percentages.

Data are presented as the mean ± standard deviation.

A healthy diet was based on adherence to at least four of seven commonly eaten food groups following recommendations on dietary priorities for cardiometabolic health.

BMI, body mass index; IPAQ, International Physical Activity Questionnaire; SD, standard deviation; T2DM, type 2 diabetes mellitus; TDI, Townsend Deprivation Index.

Risk of incident IBS associated with social isolation

During a median follow-up of 14.5 years, 8307 (2.1%) participants developed IBS. The 14-year cumulative incidence of IBS increased from 2.0% (95% confidence interval (CI): 2.0–2.1) in the least isolated group to 2.2% (95% CI: 2.2–2.3) and 2.6% (95% CI: 2.4–2.8) in the moderately and most isolated groups, corresponding to absolute risk differences of 0.2 and 0.6 percentage points, respectively (Figure S3). The PAF was estimated to be approximately 3.6%. Compared with the least isolated group, participants who were moderately isolated (HR = 1.06, 95% CI: 1.01–1.11) and most isolated (HR = 1.20, 95% CI: 1.11–1.29) had a 6.0% and 20.0% increased risk of incident IBS, demonstrating a significant dose-response relationship (p_trend < 0.001; Table 2). In addition, one increment in the social isolation score was associated with an 8.0% elevated IBS risk (HR = 1.08, 95% CI: 1.04–1.11).

Table 2.

Risk of incident IBS associated with social isolation.

Social isolation status	Events	Participants	Person-years	HR (95% CI)
Social isolation status	Events	Participants	Person-years	Unadjusted	Model 1	Model 2
Social isolation
Least isolated	4148	209,081	2,951,740	Reference	Reference	Reference
Moderately isolated	3275	149,729	2,093,541	1.11 (1.06–1.16)	1.10 (1.06–1.16)	1.06 (1.01–1.11)
Most isolated	884	35,648	487,089	1.29 (1.20–1.38)	1.31 (1.22–1.41)	1.20 (1.11–1.29)
p for trend				<0.001	<0.001	<0.001
Social isolation score
Per 1 increase	8307	394,458	5,532,370	1.12 (1.08–1.15)	1.13 (1.09–1.16)	1.08 (1.04–1.11)
Items of social isolation
Live alone
No	6615	322,609	4,543,111	Reference	Reference	Reference
Yes	1692	71,849	989,259	1.17 (1.11–1.24)	1.14 (1.08–1.21)	1.11 (1.05–1.17)
Contact with family/friends
More	7613	361,059	5,070,295	Reference	Reference	Reference
Little	694	33,399	462,075	1.00 (0.92–1.08)	1.10 (1.02–1.19)	1.08 (1.00–1.16)
Engage in weekly leisure or social activities
Yes	5594	275,962	3,880,305	Reference	Reference	Reference
No	2713	118,496	1,652,065	1.14 (1.09–1.19)	1.13 (1.08–1.19)	1.06 (1.01–1.11)

Model 1: age and sex were adjusted; Model 2: Townsend Deprivation Index, education level, ethnicity, smoking status, alcohol drinking, healthy diet, physical activity, body mass index, and type 2 diabetes mellitus were additionally adjusted. p for trend was calculated by assigning values of 0, 1, and 2 to the least, moderately, and most isolated groups, respectively, in the models.

CI, confidence interval; HR, hazard ratio; IBS, irritable bowel syndrome.

Regarding individual items of social isolation, living alone showed an 11.0% higher IBS risk compared with not living alone (HR = 1.11, 95% CI: 1.05–1.17). Not engaging in weekly leisure or social activities was also related to an increased IBS risk versus those who participated in such activities (HR = 1.06, 95% CI: 1.01–1.11).

Risk of incident IBS associated with loneliness

The 14-year cumulative incidence of IBS was 3.1% (95% CI: 2.9–3.4) in the lonely group versus 2.1% (95% CI: 2.1–2.1) in the non-lonely group, corresponding to an absolute risk difference of 1.0 percentage point (Figure S4). The PAF was estimated to be approximately 2.1%. Participants with loneliness had a 46.0% increased risk of incident IBS versus those without loneliness (HR = 1.46, 95% CI: 1.34–1.59; Table 3). Each 1-point increment in the loneliness score was associated with a 27.0% higher IBS risk (HR = 1.27, 95% CI: 1.23–1.32).

Table 3.

Risk of incident IBS associated with loneliness.

Loneliness status	Events	Participants	Person-years	HR (95% CI)
Loneliness status	Events	Participants	Person-years	Unadjusted	Model 1	Model 2
Loneliness
Non-lonely	7736	375,689	5,273,921	Reference	Reference	Reference
Lonely	571	18,769	258,449	1.51 (1.38–1.64)	1.58 (1.45–1.72)	1.46 (1.34–1.59)
Loneliness score
Per 1 increase	8307	394,458	5,532,370	1.30 (1.26–1.35)	1.32 (1.28–1.37)	1.27 (1.23–1.32)
Items of loneliness
Feeling lonely
No	6060	323,332	4,547,619	Reference	Reference	Reference
Yes	2247	71,126	984,751	1.71 (1.63–1.79)	1.60 (1.53–1.68)	1.53 (1.45–1.60)
Able to confide
Usually	7088	335,975	4,719,235	Reference	Reference	Reference
Never/almost never	1219	58,483	813,135	1.00 (0.94–1.06)	1.13 (1.06–1.20)	1.08 (1.01–1.14)

CI, confidence interval; HR, hazard ratio; IBS, irritable bowel syndrome.

For individual items of loneliness, feeling lonely versus not feeling lonely (HR = 1.53, 95% CI: 1.45–1.60), and never/almost never being able to confide in someone versus usually being able to confide (HR = 1.08, 95% CI: 1.01–1.14), were both associated with a higher IBS risk.

Joint effects of social isolation and loneliness on risk of incident IBS

The combined effects of social isolation and loneliness were associated with a further increase in IBS risk, with HRs ranging from 1.05 to 1.60, indicating a significant dose-response relationship (p_trend < 0.001; Figure 1). Particularly, individuals who were both lonely and most isolated exhibited a 60.0% greater IBS risk compared to their non-lonely and least isolated counterparts (HR = 1.60, 95% CI: 1.35–1.89).

Figure 1.

Joint associations of social isolation and loneliness with the risk of incident IBS.

Mediation effect of depression or anxiety

Mediation analysis indicated that the association between social isolation (most vs least isolated) and incident IBS was partially mediated by depression/anxiety (24.7%), with an HR_NDE of 1.14 (95% CI: 1.05–1.22) and an HR_NIE of 1.04 (95% CI: 1.03–1.05; Figure 2(a)). For loneliness (lonely vs non-lonely), the HR_NDE and HR_NIE were separately 1.33 (95% CI: 1.21–1.44) and 1.09 (95% CI: 1.08–1.10), indicating a mediation proportion of 26.5% by depression/anxiety (Figure 2(b)). When examining the joint effects of social isolation and loneliness (lonely and most isolated vs non-lonely and least isolated), the HR_NDE and HR_NIE were separately 1.38 (95% CI: 1.14–1.61) and 1.13 (95% CI: 1.11–1.15), with 32.1% of the association explained by depression/anxiety (Figure 2(c)). Additionally, mediation effects of depression/anxiety were identified for items of social isolation and loneliness, with mediation proportions of 36.3% for living alone and 25.5% for feeling lonely (Figure S5).

Figure 2.

Mediation effect of depression and/or anxiety in the association between social isolation, loneliness, and incident IBS. (a) Social isolation (most vs least isolated). (b) Loneliness (lonely vs non-lonely). (c) Social isolation and loneliness (most isolated and lonely vs least isolated and non-lonely).

Subgroup analysis

A greater IBS risk associated with being most socially isolated was generally observed across age, sex, educational level, TDI, smoking status, alcohol drinking, healthy diet, and BMI subgroups, with a significantly higher IBS risk in never/previous drinkers (p_interaction = 0.036) or those with BMI ⩾25 kg/m² (p_interaction = 0.010; Figure S6 and Tables S2–S9). For single items of social isolation, greater IBS risk was detected in those living alone aged ⩾60 years (p_interaction = 0.049), those who had little contact with family/friends with a university degree (p_interaction = 0.011), and those who were not engaged in weekly social activities with a BMI ⩾25 kg/m² (p_interaction < 0.001). Similarly, increased IBS risk associated with loneliness was generally identified across all subgroups, with a significantly higher risk among individuals with TDI <−2.18 (p_interaction = 0.045).

Sensitivity analysis

In sensitivity analyses, the results for both social isolation and loneliness were consistent with the primary findings across various definitions of incident IBS, competing risk models, and analyses excluding participants with depression/anxiety at baseline (Tables S10–S17). Mediation analyses using newly developed depression/anxiety during follow-up before IBS diagnosis as the mediator also yielded similar results, further supporting the robustness of our findings (Figure S7).

Discussion

Main findings

In this large-scale prospective cohort study of approximately 0.4 million adults, we found that individuals who were moderately or most socially isolated had a 6.0% and 20.0% higher risk of developing IBS, respectively, compared with their least isolated counterparts, indicating a significant dose-response relationship. Loneliness was associated with a 46.0% higher IBS risk compared with those without loneliness. Notably, the coexistence of both social isolation and loneliness was associated with a further increase in IBS risk. Furthermore, these associations were partially mediated by depression and anxiety, with mediation proportions ranging from 24.7% to 32.1%.

Comparisons with previous studies

To our knowledge, no previous studies have investigated the associations of social isolation and loneliness with the risk of developing IBS. However, evidence from a Norwegian twin study indicated that insufficient social support, particularly a lack of comfort, help, or understanding, was associated with a higher IBS risk, which was largely attributable to shared genetic factors.¹⁰ In addition, social isolation and loneliness have been linked to a range of adverse health outcomes, such as cardiovascular disease, metabolic disorders, inflammatory bowel disease, depression, and anxiety.^7
–9,27 Together, these findings lend indirect support to our results and highlight the potential importance of psychosocial factors in the development of IBS.

Potential mechanisms

Although the exact mechanisms remain incompletely understood, social isolation and loneliness may increase susceptibility to IBS via dysregulation of the brain-gut-microbiota axis (Figure S8). First, social isolation and loneliness may act as chronic psychosocial stressors that activate the hypothalamic-pituitary-adrenal axis and the autonomic nervous system, resulting in neuroendocrine and neurotransmitter dysregulation that may disrupt brain-gut signaling, alter gastrointestinal motility, and increase visceral sensitivity.^28
–30 Second, chronic stress may induce persistent low-grade inflammation and immune activation, which can sensitize enteric nerves and exacerbate abdominal pain and bloating.^12,31 Third, chronic stress may impair intestinal barrier function, increase intestinal permeability, and thereby facilitate mucosal immune responses and symptom aggravation.¹⁴ Fourth, social isolation and loneliness may also influence gut microbiota composition through stress-related pathways and lifestyle changes, potentially resulting in reduced microbial diversity, altered microbial metabolites, and disrupted neuroimmune signaling along the brain-gut axis.^13,32

Mediation effect of depression/anxiety

Depression/anxiety may represent key psychological pathways linking social isolation and loneliness to the development of IBS. Previous structural equation modeling studies have suggested a bidirectional interplay in which social isolation heightens loneliness, thereby worsening depressive and anxiety symptoms, while these mental health conditions may in turn reinforce feelings of loneliness and social withdrawal.⁸ Longitudinal studies have further shown that chronic loneliness was associated with an approximately 2.8-fold higher risk of depression.³³ Similarly, sustained social isolation was associated with a 5.7-fold higher likelihood of loneliness and a 1.6-fold higher risk of depression compared with non-isolated individuals.³³ Notably, depression/anxiety not only frequently co-occur with IBS but may also contribute to its development.^34
–36 A prospective cohort study reported that individuals with depression/anxiety were three times more likely to develop IBS, a finding supported by Mendelian randomization evidence.^35,36 Depression/anxiety associated with social isolation and loneliness may further exacerbate dysfunction of the brain-gut-microbiota axis.^37,38 In our study, depression/anxiety mediated 24.7%–32.1% of the increased IBS risk linked to social isolation and loneliness. These findings support the possibility that psychological well-being may be involved in the pathway linking social disconnection to IBS.

Implications for practice and future research

Our findings suggest that social isolation and loneliness may be relevant psychosocial factors associated with IBS risk. Although the absolute increases in individual-level risk were small, these associations may still be important from a public health perspective because social disconnection is common and potentially modifiable. The estimated PAFs suggest that reducing or eliminating social isolation and loneliness could potentially prevent about 3.6% and 2.1% of incident IBS cases, respectively, at the population level. In addition, our mediation analyses indicated that depression and anxiety partially mediated these associations, suggesting that psychological distress may represent one potential pathway linking social disconnection with IBS development. However, whether interventions targeting social isolation, loneliness, or related psychological distress can reduce IBS risk remains to be established. Further longitudinal studies, mechanistic investigations, and intervention trials are needed to clarify causality and to determine whether psychosocial or community-based strategies can improve gastrointestinal and mental health outcomes.

Strengths and limitations

To the best of our knowledge, this is the first large-scale prospective cohort study to comprehensively examine the associations of social isolation, loneliness, and their individual items with the risk of incident IBS. In addition, we are the first to elucidate the joint effects of social isolation and loneliness on IBS risk, as well as the mediating roles of depression/anxiety in these associations. Furthermore, we conducted a series of subgroup and sensitivity analyses to ensure the robustness and reliability of our findings.

However, several limitations should be acknowledged. (i) Social isolation and loneliness were assessed using abbreviated measures rather than full scales. Although these simplified instruments have been widely used in large population-based studies and have demonstrated acceptable validity,⁷ they may not fully capture the complexity of psychosocial experiences, potentially leading to measurement error and exposure misclassification. (ii) Information on social isolation and loneliness was collected only at baseline, which prevented us from capturing potential changes in these exposures during the long follow-up period. Such changes may lead to exposure misclassification and regression dilution bias, potentially attenuating the observed associations. Although repeat assessments are available for a subset of participants in the UKB, they were not available for the entire cohort and occurred at different time points. Future studies incorporating repeated measurements and time-varying analyses may help better characterize the dynamic nature of social relationships and their long-term health effects. (iii) The exposure and mediator were both assessed at baseline, so their temporal sequence could not be firmly established. In addition, IBS symptoms may precede a formal diagnosis and influence psychological status. Although sensitivity analyses excluding participants with baseline depression/anxiety and using newly developed depression/anxiety during follow-up before IBS diagnosis as the mediator yielded similar results, the mediation findings should still be interpreted with caution. (iv) Incident IBS may have been underascertained, as some symptomatic individuals may not seek medical care or receive a formal diagnosis, and ICD-based ascertainment may preferentially capture more severe cases. Hence, differential healthcare-seeking behavior related to social isolation and loneliness cannot be excluded. (v) The reliability of an IBS diagnosis based on a single IBS code may raise concerns. However, we conducted additional sensitivity analyses incorporating DHQ data with Rome III criteria and IBS coding information derived from multiple independent sources, all of which yielded consistent results. (vi) Because systematic Rome-based assessments for disorders of gut-brain interaction (DGBI) were unavailable in the UKB, some participants classified as free of IBS at baseline may have had other DGBI. (vii) Despite carefully adjusting for numerous potential confounders, residual confounding cannot be completely ruled out. (viii) The predominantly White and relatively healthy middle-aged to older composition of the UKB cohort may limit the generalizability of our findings. Future studies in younger and non-Western populations are needed, with consideration of potential cultural differences in loneliness and social isolation.

Conclusion

In this large-scale prospective cohort study, both social isolation and loneliness were independently and jointly associated with an increased risk of incident IBS, with depression or anxiety partially mediating these associations. These findings may have implications for the prevention and management of IBS at the population level, given the high prevalence of social isolation and loneliness. Future studies in ethnically diverse populations are needed to confirm these findings and further elucidate the underlying mechanisms.

Supplemental Material

sj-docx-1-tag-10.1177_17562848261446458 – Supplemental material for Long-term risk of irritable bowel syndrome associated with social isolation and loneliness: a large-scale prospective cohort study with mediation analysis

Supplemental material, sj-docx-1-tag-10.1177_17562848261446458 for Long-term risk of irritable bowel syndrome associated with social isolation and loneliness: a large-scale prospective cohort study with mediation analysis by Yesheng Zhou, Si Liu, Qian Zhang, Shutian Zhang and Shanshan Wu in Therapeutic Advances in Gastroenterology

Supplemental Material

sj-docx-2-tag-10.1177_17562848261446458 – Supplemental material for Long-term risk of irritable bowel syndrome associated with social isolation and loneliness: a large-scale prospective cohort study with mediation analysis

Supplemental material, sj-docx-2-tag-10.1177_17562848261446458 for Long-term risk of irritable bowel syndrome associated with social isolation and loneliness: a large-scale prospective cohort study with mediation analysis by Yesheng Zhou, Si Liu, Qian Zhang, Shutian Zhang and Shanshan Wu in Therapeutic Advances in Gastroenterology

Footnotes

Acknowledgements

This research has been conducted using the UK Biobank Resource under application number 74444.

Declarations

ORCID iD

Shanshan Wu

Supplemental material

Supplemental material for this article is available online.

References

Sperber

Bangdiwala

Drossman

, et al. Worldwide prevalence and burden of functional gastrointestinal disorders, results of Rome Foundation Global Study. Gastroenterology 2021; 160: 99–114.e3.

Wang

Thakur

Shapiro

Non-pharmaceutical treatments for irritable bowel syndrome. BMJ 2024; 387: e075777.

Grover

Kolla

Pamarthy

, et al. Psychological, physical, and sleep comorbidities and functional impairment in irritable bowel syndrome: results from a national survey of U.S. adults. PLoS One 2021; 16: e0245323.

Gerlach

Solway

Malani

PN.

Social isolation and loneliness in older adults. JAMA 2024; 331: 2058.

Teo

Cheng

, et al. Global prevalence of social isolation among community-dwelling older adults: a systematic review and meta-analysis. Arch Gerontol Geriatr 2023; 107: 104904.

Surkalim

Luo

Eres

, et al. The prevalence of loneliness across 113 countries: systematic review and meta-analysis. BMJ 2022; 376: e067068.

Song

Zhu

Shi

, et al. Social isolation, loneliness, and incident type 2 diabetes mellitus: results from two large prospective cohorts in Europe and East Asia and Mendelian randomization. EClinicalMedicine 2023; 64: 102236.

Santini

Jose

York Cornwell

, et al. Social disconnectedness, perceived isolation, and symptoms of depression and anxiety among older Americans (NSHAP): a longitudinal mediation analysis. Lancet Public Health 2020; 5: e62–e70.

Richard

Rohrmann

Vandeleur

, et al. Loneliness is adversely associated with physical and mental health and lifestyle factors: results from a Swiss national survey. PLoS One 2017; 12: e0181442.

10.

Kutschke

Harris

Bengtson

MB.

How are perceptions of social strain and low support related to irritable bowel syndrome?—a Norwegian twin study. Neurogastroenterol Motil 2021; 33: e14007.

11.

Donovan

Mackey

Platt

, et al. Social isolation alters behavior, the gut-immune-brain axis, and neurochemical circuits in male and female prairie voles. Neurobiol Stress 2020; 13: 100278.

12.

Smith

Gavey

, et al. The association between loneliness, social isolation and inflammation: a systematic review and meta-analysis. Neurosci Biobehav Rev 2020; 112: 519–541.

13.

Wang

Ullah

Deng

, et al. Social isolation induces intestinal barrier disorder and imbalances gut microbiota in mice. Neurosci Lett 2024; 826: 137714.

14.

Yang

Sun

Wen

, et al. Chronic stress-induced serotonin impairs intestinal epithelial cell mitochondrial biogenesis via the AMPK-PGC-1alpha axis. Int J Biol Sci 2024; 20: 4476–4495.

15.

Berkman

Syme

SL.

Social networks, host resistance, and mortality: a nine-year follow-up study of Alameda County residents. Am J Epidemiol 1979; 109: 186–204.

16.

Pohl

Cochrane

Schepp

, et al. Measuring social isolation in the national health and aging trends study. Res Gerontol Nurs 2017; 10: 277–287.

17.

Laugesen

Baggesen

Schmidt

SAJ

, et al. Social isolation and all-cause mortality: a population-based cohort study in Denmark. Sci Rep 2018; 8: 4731.

18.

Chan

SCY

Wong

Huang

, et al. The psychometric properties of the Lubben Social Network Scale (LSNS-6) and its associations with well-being indicators in Hong Kong older adults. Australas J Ageing 2023; 42: 683–689.

19.

Hughes

Waite

Hawkley

, et al. A short scale for measuring loneliness in large surveys: results from two population-based studies. Res Aging 2004; 26: 655–672.

20.

Ali

Al-Dossary

Laranjeira

, et al. Loneliness among dementia caregivers: evaluation of the psychometric properties and cutoff score of the three-item UCLA loneliness scale. Front Psychiatry 2025; 16: 1526569.

21.

Bahadori

Asadollahi

Ghajari

, et al. Psychometric properties and cross-cultural adaptation of the UCLA loneliness scale, 6-item among older Iranian people. BMC Psychol 2025; 13: 592.

22.

Ford

Sperber

Corsetti

, et al. Irritable bowel syndrome. Lancet 2020; 396: 1675–1688.

23.

Sudlow

Gallacher

Allen

, 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.

24.

Mozaffarian

Dietary and policy priorities for cardiovascular disease, diabetes, and obesity: a comprehensive review. Circulation 2016; 133: 187–225.

25.

UK Biobank. Digestive health web-based questionnaire, https://biobank.ndph.ox.ac.uk/ukb/ukb/docs/digestive_health.pdf (2019, accessed 17 March 2025).

26.

Von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577.

27.

Zhao

Zhang

, et al. Social isolation, loneliness, and incident inflammatory bowel disease: results from a large prospective cohorts and Mendelian randomization. J Crohns Colitis 2024; 18: i2014–i2015.

28.

Brandt

Liu

Heim

, et al. The effects of social isolation stress and discrimination on mental health. Transl Psychiatry 2022; 12: 398.

29.

Gros

Mesonero

, et al. Neurotransmitter dysfunction in irritable bowel syndrome: emerging approaches for management. J Clin Med 2021; 10: 3429.

30.

Guzel

Mirowska-Guzel

The role of serotonin neurotransmission in gastrointestinal tract and pharmacotherapy. Molecules 2022; 27: 1680.

31.

Schneider

Blank

Alvarez

, et al. The enteric nervous system relays psychological stress to intestinal inflammation. Cell 2023; 186: 2823–2838.e20.

32.

Margolis

Cryan

Mayer

EA.

The microbiota-gut-brain axis: from motility to mood. Gastroenterology 2021; 160: 1486–1501.

33.

Zhang

Kuang

Xin

, et al. Loneliness, social isolation, depression and anxiety among the elderly in Shanghai: findings from a longitudinal study. Arch Gerontol Geriatr 2023; 110: 104980.

34.

Zamani

Alizadeh-Tabari

Zamani

Systematic review with meta-analysis: the prevalence of anxiety and depression in patients with irritable bowel syndrome. Aliment Pharmacol Ther 2019; 50: 132–143.

35.

Zhang

Wellens

Zhao

, et al. Psychological resilience and mental wellbeing mitigate the risk of irritable bowel syndrome. Am J Gastroenterol. Epub ahead of print November 2025. DOI: 10.14309/ajg.0000000000003833.

36.

Wang

Peng

, et al. Association of major depressive disorder and increased risk of irritable bowel syndrome: a population-based cohort study and a two-sample Mendelian randomization study in the UK biobank. J Affect Disord 2024; 345: 419–426.

37.

Sonali

Ray

Ahmed Tousif

, et al. Mechanistic insights into the link between gut dysbiosis and major depression: an extensive review. Cells 2022; 11: 1362.

38.

Toader

Dobrin

Costea

, et al. Mind, mood and microbiota-gut-brain axis in psychiatric disorders. Int J Mol Sci 2024; 25: 3340.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

8.84 MB

0.00 MB

0.03 MB