Lifestyle factors, genetic susceptibility, and risk of incident diverticulitis: an integrated analysis of four prospective cohort studies and electronic health records-linked biobank

Abstract

Background

Both lifestyle factors and genetic predisposition contribute to development of diverticulitis.

Objective

To examine whether lifestyle modification can reduce the genetic risk of diverticulitis.

Design

We derived an overall healthy lifestyle score for diverticulitis based on smoking, body mass index [BMI], physical activity, fiber, and red meat among 179,564 participants in three prospective cohorts-the Nurses’ Health Study (NHS), NHSII, and the Health Professionals Follow-Up Study (HPFS). The association between the healthy lifestyle score and incident diverticulitis was confirmed among 30,750 participants in the Southern Community Cohort Study (SCCS). We assessed genetic risk using a polygenic risk score (PRS) among 36,077 individuals with genotype data available. We further validated our findings in the Mass General Brigham Biobank (MGBB).

Results

A healthy lifestyle score was associated with a decreased risk of diverticulitis. Compared to a score of 0, the multivariable-adjusted hazard ratio (HR) for a score of 5 was 0.50 (95% CI, 0.44-0.57; P-trend <.0001). This association was consistent across the SCCS in both non-Hispanic Black and White populations. Each unit increase in the healthy lifestyle score was associated with a reduced diverticulitis risk similarly across genetic risk categories, with HRs of 0.89 (95% CI, 0.83-0.95) for low, 0.86 (0.81-0.92) for mid, and 0.87 (0.83-0.91) for high genetic risk. In the MGBB cohort, a higher BMI was associated with an increased diverticulitis risk across genetic risk categories.

Conclusion

Maintaining a healthy lifestyle was associated with a reduced risk of developing diverticulitis, regardless of population differences and genetic susceptibilities.

INTRODUCTION

Diverticulitis, inflammation of diverticula that are small sacs forming through weak areas of the colon wall, is the sixth most common gastrointestinal indication for hospitalization¹ and is a major reason for emergent colectomy. Patients with diverticulitis suffer from acute symptoms of pain and fever, and some may develop other complications, including peritonitis, obstruction, fistula, or abscess, chronic gastrointestinal complaints, and recurrent episodes^{2 3}.

Both genetic and lifestyle factors have been identified as critical drivers for diverticulitis and more broadly, diverticular disease. Earlier twin studies estimated that 40-53% of the heritability of diverticular disease can be attributed to genetic factors^{4 5}. The candidate gene approach identified the TNFSF15 gene encoding a cytokine of the tumor necrosis factor (TNF) family associated with diverticulitis⁶. Multiple genome-wide association studies (GWAS) have been conducted and associated 150 genetic variants with diverticular disease, pointing to a role for altered structure of the colon, gut motility, gastrointestinal mucus, and ionic homeostasis that may all contribute to the development of this condition^7-11. Polygenic risk scores (PRS) combining the influences of specific risk alleles have demonstrated predictive ability for diverticulitis overall and its severity and recurrences^11-13.

Accumulating evidence has also linked lifestyle factors to the development of diverticulitis, including smoking, obesity, lack of physical activity, and a Western dietary pattern of high intake of red and processed meat and low dietary fiber^{14 15}. Adherence to a healthy lifestyle incorporating the individual components could potentially reduce the incidence of diverticulitis by up to 50%¹⁶.

However, with the need for datasets containing both genetic inflammation and comprehensive lifestyle data, it remains largely unknown the extent to which genetic risk of diverticulitis can be mitigated by lifestyle changes, or any potential interplay between genetic predisposition and lifestyle in influencing diverticulitis risk. In the current study, we analyzed genetic and detailed lifestyle data from participants in three prospective cohorts, Nurses’ Health Study (NHS), NHSII, and the Health Professionals Follow-up Study (HPFS). We separately validated the healthy lifestyle score in the Southern Community Cohort Study (SCCS), a cohort with two-thirds non-Hispanic Black individuals. Gene-lifestyle associations were further validated in Mass General Brigham Biobank (MGBB).

METHODS

Study population

A schema of study population is shown in Supplemental Figure 1. The NHS, NHSII, and HPFS are three ongoing prospective cohort studies of US health professionals. The NHS recruited 121,700 female registered nurses aged 30 to 55 in 1976, while the NHSII recruited 116,429 younger female nurses between 25 and 42 years old in 1989¹⁷. The HPFS enrolled 51,520 male health professionals aged 40 to 75 in 1986¹⁸. The participants were followed up every two years by questionnaires querying about health status, medication use, and lifestyle behaviors. Detailed dietary information was obtained through validated food-frequency questionnaires (FFQs) every four years^{19 20}. A subset of NHS, NHSII, and HPFS participants of European ancestry had genomic data available from previous GWAS^21-23.

The SCCS is a prospective cohort study recruiting 84,507 men and women aged 40 to 79 across 12 US southeastern states between March 2002 and September 2009²⁴. Two-thirds of the SCCS cohort were self-reported non-Hispanic Black participants; over half had annual household income <$15,000. Baseline interviews were conducted to collect information on sociodemographics, lifestyle, and medical history. Habitual dietary intake was assessed using a validated FFQ²⁵.

The Mass General Brigham Biobank (MGBB) is a comprehensive repository that houses stored biospecimens linked to clinical information from electronic health records (EHRs) and lifestyle surveys for more than 150,000 consented patients within the MGB healthcare network in Boston, MA. Genomic data is available for over 65,000 participants²⁶.

The study was approved by the Institutional Review Boards of Mass General Brigham and Harvard T.H. Chan School of Public Health (2014P000837; 2023P001404).

Genotyping and polygenic risk score

The GWAS samples of NHS, NHSII, and HPFS were genotyped in SNP arrays including Illumina HumanHap, Affymetrix, OmniExpress, OncoArray, HumanCoreExome2, and Global Screening Array (GSA). The MGBB samples were genotyped on three arrays offered by Illumina, including the Multi-Ethnic Genotyping Array (MEGA), Multi-Ethnic Global (MEG) BeadChip, and the GSA. For both datasets, we used data imputed using the TopMed imputation server based on the GRCh38 human reference genome.

A polygenic risk score (PRS) for diverticulitis was calculated using The Polygenic Score Catalog Calculator (pgsc_calc) with weights derived based on meta-analyzed summary statistics of GWAS for diverticular disease in UK Biobank and Million Veteran Program²⁷. This included all SNPs, not just the significant ones. In brief, a genome-wide association meta-analysis was performed using METAL and was comprised of two independent GWASs for diverticular disease (phecode 562), one in the Million Veteran Program cohort²⁸ and the other performed by the PanUKBB initiative (https://pan.ukbb.broadinstitute.org/phenotypes). The Department of Veteran Affairs Million Veteran Program is a longitudinal cohort of over 1 million veterans and collects precision and genomic information, as previously described in detail elsewhere²⁹. The PanUKBB initiative undertook pan-ancestry genetic analysis of the UK biobank, a publicly available database that recruited approximately 500,000 volunteers and has been described elsewhere³⁰. Polygenic risk weights were calculated using PRS-CSx based on the combined European and African population summary statistics from the GWAS meta-analysis. PRS-CSx is a Bayesian polygenic modeling framework³¹ and in this study produced a weight file with > 1.2 million SNPs.

Lifestyle factors and healthy lifestyle score

We derived the healthy lifestyle score by considering five major lifestyle factors that have been associated with diverticulitis¹⁵, including smoking, body mass index (BMI), physical activity, dietary fiber intake, and total red meat intake (red and processed meat). A summary of the current evidence on the five lifestyle risk factors for diverticulitis is shown in Supplemental Table 1. Participants received a score of 1 for each component if they met the low-risk lifestyle criteria and a score of 0 otherwise. Specifically, those who have never smoked were considered low-risk, as both past and current smoking have been linked to an increased risk of diverticulitis^{32 33}. Individuals with a BMI of less than 25 kg/m² were classified as low-risk since being overweight or obese was associated with a greater risk³⁴. The thresholds for low-risk levels of physical activity (top 40%), fiber intake (top 40%), and total red meat intake (lowest 40%) were determined based on cohort- and questionnaire cycle-specific distributions, consistent with lifestyle score definitions for major chronic diseases and guideline-recommended levels^{16 35}. The overall healthy lifestyle score was calculated by summing up the scores for each individual component (ranging from 0 to 5).

Ascertainment of diverticulitis

Participants of NHS and NHSII were asked if they had diverticulitis requiring antibiotic therapy or hospitalization on biennial questionnaires (2008, 2012, 2014, 2016, and 2020 for NHS; 2015, 2017, and 2019 for NHSII). Those who reported having diverticulitis were subsequently asked the year of the episode, dating back to 1990 for NHS and 2003 for NHSII. For HPFS, participants were asked biennially, starting in 1990 until 2020, if they were newly diagnosed with diverticulitis. If they reported diverticulitis, supplementary questionnaires were sent for the date of diagnosis, presenting symptoms, methods of diagnosis, and relevant treatments. Participants were defined as having diverticulitis if they had abdominal pain related to diverticular disease along with one of the following: 1) treated with antibiotic; surgery, or hospitalization; 2) complications such as perforation, abscess, fistula, or obstruction; 3) presentation with fever, requiring medical therapy, or evaluated with computed tomography. The accuracy of ascertaining diverticulitis within the cohorts has been validated^{33 34 36}, with 84%-92% of self-reported cases being confirmed through medical records review.

In the SCCS cohort, diagnosis of diverticulitis was identified through linkage to the Centers for Medicare and Medicaid Services (CMS) and defined as any inpatient, emergency room or outpatient encounter with a colonic diverticulitis diagnosis based on the International Classification of Diseases (ICD) codes (ICD9: 562.11, 562.13; ICD10: K57.2, K57.32, K57.33, K57.4, K57.52, K57.53, K57.8 (K57.80, K57.81), K57.92, K57.93) listed as the primary diagnosis.

In the MGBB dataset, we used the same ICD codes (ICD9: 562.11, 562.13; ICD10: K57.2, K57.32, K57.33, K57.4, K57.52, K57.53, K57.8 (K57.80, K57.81), K57.92, K57.93) to identify diverticulitis cases among participants with genomic data and matched each case with five controls according to age, sex, and the last encounter year. We assessed the validity of the identification of patients with diverticulitis using ICD codes by reviewing the medical records of a randomly selected sample of 106 patients, with 91.5% of diverticulitis cases being confirmed.

Statistical analysis

We evaluated the associations of lifestyle factors and the overall healthy lifestyle score with incident diverticulitis in a combined cohort of NHS, NHSII, and HPFS participants. We excluded participants with a history of diverticulitis, cancer, or inflammatory bowel disease, as well as those with missing lifestyle information at baseline. Person-time was calculated from baseline until the date of diagnosis of diverticulitis, death, last follow-up questionnaire, or the end of the study period (1990-2020 for NHS, 2003-2019 for NHSII, and 1986-2020 for HPFS), whichever came first. We used time-varying Cox proportional hazards regression models stratified by age, questionnaire cycle, and cohort and adjusted for other demographical and covariates, including alcohol intake, total calorie intake, regular medication use (aspirin, non-steroidal anti-inflammatory drug, or acetaminophen), physical examination in the past two years, and menopausal hormone therapy (only for women). Lifestyle factors and covariates were updated by using the most recent information prior to the questionnaire cycle of interest to account for changes over time.

Among participants with genomic data, we examined the association of healthy lifestyle score and PRS individually with diverticulitis, as well as the association between healthy lifestyle score and diverticulitis across PRS categories. We assessed multiplicative interaction using the Wald test by including a product term between continuous measures of lifestyle score and PRS. We also estimated additive interaction between continuous measures of PRS and unhealthy lifestyle score (evaluated in reverse from healthy lifestyle score so effect estimate was positive) by calculating the relative excess risk due to interactions (RERIs), with RERI >0 indicating the presence of significant additive interaction between PRS and unhealthy lifestyle³⁷. We also calculated population-attributable risk (PAR) for lifestyle score overall and within each PRS category.

In the SCCS, we excluded any participants with prevalent diverticulitis or missing lifestyle information and restricted the remaining SCCS participants to those individuals aged ≥ 65 years at cohort enrollment, or persons < 65 years at enrollment who: a) reported being covered by Medicaid (provides medical benefits to low-income adults and uninsured persons) or Medicare (the primary health insurance program for persons aged ≥ 65 or those with disability under age 65) on the baseline questionnaire; or b) did not report Medicare or Medicaid on the baseline questionnaire but had a CMS claim within one year of being enrolled in SCCS. A total of 30,750 participants were included in the analysis of lifestyle and incident diverticulitis using Cox proportional hazard models. Participants were censored on the date of death, diagnosis of diverticulitis, loss to follow-up, or the end of claims data linkage in 2021. We similarly examined the association between healthy lifestyle score and incident diverticulitis as well as potential effect modification by race, with significance evaluated using the Wald test of a product term of healthy lifestyle score and race.

We conducted a case-control study within the MGBB dataset, matching each diverticulitis case with five controls according to age, sex, and year of last encounter. Logistic regression models were used to evaluate the associations of PRS and diverticulitis. To examine potential interaction with lifestyle, we examined BMI in relation to PRS since data on other lifestyle factors were not consistently collected in this cohort.

Patient Involvement

No patients were involved in setting the research question or the outcome measures, nor were they involved in the design and implementation of the study. There are no plans to involve patients in dissemination.

RESULTS

In a combined cohort of 179,564 participants from NHS, NHSII, and HPFS, we documented a total of 10,299 incident diverticulitis cases over an average follow-up of 20 years. Individual lifestyle factors were each significantly associated with the incidence of diverticulitis (Table 1). For example, compared to participants with a BMI < 25 kg/m², the multivariable-adjusted hazard ratio (HR) was 1.32 (95% confidence interval [CI], 1.26-1.38) for those who were overweight and 1.44 (95% CI, 1.37-1.52) for those who were obese. Both past smokers (HR, 1.17; 95% CI, 1.12-1.22) and current smokers (HR, 1.13; 95% CI, 1.04-1.23) had an increased risk of diverticulitis when compared to those who never smoked. Higher levels of physical activity were associated with a reduced risk, with an HR of 0.84 (95% CI, 0.79-0.90) comparing participants in the highest to the lowest quintile. Finally, a greater intake of fiber was associated with a reduced risk (extreme-quintile HR, 0.86; 95% CI, 0.80-0.92), whereas a greater intake of red meat was associated with an increased risk (extreme-quintile HR, 1.09; 95% CI, 1.01-1.17). When all five lifestyle factors were considered together, a higher score indicating a healthier lifestyle was linearly linked to a lower risk of diverticulitis (HR per 1-point increase: 0.88; 95% CI, 0.86-0.89). Participants with a score of 5 had an HR of 0.50 (95% CI, 0.44-0.57) when compared to those with a score of 0.

Table 1.

Multivariable-adjusted HR (95% CI) of diverticulitis according to lifestyle risk factors in NHS, NHSII, and HPFS

	Pooled	NHS 1990-2020 (n=56,158)	NHSII 2003-2019 (n=79,055)	HPFS 1986-2020 (n=44,351)
Cases/Person-years	10,309/3,690,578	5,832/1,440,078	3,316/1,197,075	1,151/1,051,628
BMI, kg/m2
<25	1 [ref]	1 [ref]	1 [ref]	1 [ref]
25.0-29.9	1.32 (1.26-1.38)	1.29 (1.22-1.37)	1.48 (1.36-1.61)	1.13 (0.99-1.29)
≥30	1.44 (1.37-1.52)	1.41 (1.31-1.51)	1.61 (1.46-1.76)	1.24 (1.03-1.50)
Smoking status
Never smoker	1 [ref]	1 [ref]	1 [ref]	1 [ref]
Past smoker	1.17 (1.12-1.22)	1.16 (1.10-1.22)	1.20 (1.12-1.30)	1.13 (1.00-1.28)
Current smoker	1.13 (1.04-1.23)	1.09 (0.97-1.21)	1.18 (1.02-1.36)	1.31 (1.04-1.67)
Physical activity
Q1 (lowest)	1 [ref]	1 [ref]	1 [ref]	1 [ref]
Q2	0.96 (0.91-1.02)	0.96 (0.89-1.04)	0.95 (0.86-1.05)	1.01 (0.84-1.20)
Q3	0.96 (0.90-1.02)	0.97 (0.90-1.04)	0.92 (0.83-1.02)	0.98 (0.82-1.18)
Q4	0.92 (0.87-0.98)	0.91 (0.84-0.99)	0.91 (0.82-1.02)	0.97 (0.81-1.17)
Q5 (highest)	0.84 (0.79-0.90)	0.82 (0.75-0.89)	0.84 (0.74-0.94)	0.94 (0.78-1.14)
Fiber intake
Q1 (lowest)	1 [ref]	1 [ref]	1 [ref]	1 [ref]
Q2	0.94 (0.88-1.00)	0.97 (0.90-1.05)	0.93 (0.83-1.03)	0.83 (0.70-0.99)
Q3	0.95 (0.89-1.01)	1.00 (0.92-1.08)	0.92 (0.82-1.02)	0.85 (0.71-1.01)
Q4	0.91 (0.95-0.97)	0.93 (0.86-1.02)	0.92 (0.92-1.03)	0.80 (0.66-0.96)
Q5 (highest)	0.86 (0.80-0.92)	0.93 (0.85-1.01)	0.84 (0.75-0.95)	0.64 (0.51-0.79)
Total red meat intake (unprocessed and processed)
Q1 (lowest)	1 [ref]	1 [ref]	1 [ref]	1 [ref]
Q2	1.09 (1.02-1.17)	1.07 (0.98-1.16)	1.08 (0.96-1.21)	1.28 (1.04-1.57)
Q3	1.13 (1.06-1.21)	1.09 (1.00-1.19)	1.14 (1.01-1.28)	1.39 (1.13-1.71)
Q4	1.10 (1.03-1.18)	1.06 (0.97-1.16)	1.15 (1.02-1.29)	1.21 (0.97-1.50)
Q5 (highest)	1.09 (1.01-1.17)	1.00 (0.91-1.10)	1.16 (1.02-1.31)	1.42 (1.13-1.79)
Healthy lifestyle score, out of 5 (components: BMI<25 kg/m2, never smoker, PA Q4-5, fiber intake Q4-5, and red meat intake Q1-2)
0 (unhealthy)	1 [ref]	1 [ref]	1 [ref]	1 [ref]
1	0.86 (0.81-0.92)	0.85 (0.78-0.93)	0.86 (0.76-0.98)	0.76 (0.63-0.91)
2	0.79 (0.74-0.84)	0.78 (0.72-0.85)	0.76 (0.67-0.86)	0.69 (0.57-0.83)
3	0.70 (0.65-0.75)	0.73 (0.66-0.80)	0.60 (0.53-0.69)	0.58 (0.48-0.71)
4	0.56 (0.51-0.61)	0.57 (0.51-0.64)	0.49 (0.42-0.60)	0.46 (0.36-0.60)
5 (healthy)	0.50 (0.44-0.57)	0.49 (0.41-0.60)	0.42 (0.33-0.53)	0.54 (0.37-0.78)
Per 1 increase	0.88 (0.86-0.89)	0.89 (0.87-0.90)	0.84 (0.81-0.86)	0.86 (0.82-0.90)
P for trend	<.0001	<.0001	<.0001	<.0001

Models were adjusted for alcohol intake, total calorie intake, regular aspirin, non-steroidal anti-inflammatory drug, and acetaminophen use, physical examination in the past two years, and menopausal hormone therapy (only for NHS and NHSII), and stratified by age, questionnaire cycle, and cohort (in the pooled analyses).

Abbreviations: BMI, body mass index; CI, confidence interval; HPFS, Health Professionals Follow-up Study; HR, hazard ratio; NHS, Nurses’ Health Study.

The association between the healthy lifestyle score and incident diverticulitis was consistently observed across NHS, NHSII, and HPFS (Table 1). In the SCCS cohort, a total of 2,183 incident diverticulitis cases were ascertained during an average follow-up period of 11.9 years. A healthy lifestyle score similarly showed a significant association with diverticulitis (Figure 1). Those with a healthy lifestyle score of 3-5 had a substantially lower risk (HR: 0.69; 95% CI: 0.58-0.83) compared to those with a score of 0. No notable differences were found comparing the White (HR, 0.60; 95% CI, 0.44-0.82) and Black (HR, 0.78; 95% CI, 0.61-0.98; P-interaction = 0.33) populations.

Figure 1. Associations between healthy lifestyle score and risk of incident diverticulitis among 30,750 participants in the Southern Community Cohort Study.

Results from Cox proportional hazards regression models adjusted for age at enrollment, sex, race, educational attainment, annual household income, marital status, enrollment source, and alcohol consumption. Missing data was imputed for race, educational attainment, annual household income, marital status, and alcohol consumption using multiple imputation with 5 imputations. Healthy lifestyle score was derived by summing up scores for individual lifestyle factors. Participants got a score of 1 if they meet the low-risk lifestyle criteria (body mass index <25 kg/m², never smoker, physically active (≥150 minutes of moderate or ≥75 vigorous leisure-time physical activity per week), fiber intake in the highest quintile, red/processed meat intake in the lowest quintile) and 0 otherwise. Scores 3, 4, 5 were combined due to small numbers in each group. No statistically significant interaction was observed between a healthy lifestyle score and race on the risk of diverticulitis, based on the Wald test of a product term of healthy lifestyle score and race (P for interaction = 0.33).

Among the subset of NHS, NHSII, and HPFS population with available genomic data, there were no notable differences in lifestyle factors across PRS categories (Supplemental Table 2). PRS was significantly associated with incident diverticulitis (Table 2). The adjusted HR was 1.58 (95% CI, 1.52-1.65) for each SD increase in PRS. The link between PRS and diverticulitis appeared to be stronger among younger participants. The HR for diverticulitis with each SD increase in PRS was 1.81 (95% CI, 1.67-1.96) for those younger than 60 years, 1.61 (95% CI, 1.50-1.72) for those aged 60-69 years, and 1.42 (95% CI, 1.33-1.52) for those over 70 years (P-interaction < .0001).

Table 2.

Association of PRS with diverticulitis risk and its interaction with age in NHS, NHSII, and HPFS

	All participants	Age subgroups			P for interaction b
		<60 years	60-69 years	≥70 years
Cases/Person-years	2357/808408	651/291530	832/252267	874/264611	-
Age-adjusted HR (95% CI) per 1 SD increase	1.59 (1.54-1.66)	1.83 (1.69-1.98)	1.62 (1.51-1.73)	1.42 (1.33-1.52)	<.0001
Multivariable-adjusted HR (95% CI) per 1 SD increase a	1.58 (1.52-1.65)	1.81 (1.67-1.96)	1.61 (1.50-1.72)	1.42 (1.33-1.52)	<.0001

Cox proportional hazards regression models were stratified by age, questionnaire cycle, and cohort.

^aMultivariable-adjusted models were adjusted for body mass index, aspirin use, non-steroidal anti-inflammatory drug use, acetaminophen use, physical examination, smoking status, physical activity, menopausal hormone use, fiber, red meat, alcohol consumption, total calorie intake.

^bP value for interaction was assessed using the Wald test by including a product term between continuous age and continuous PRS.

Abbreviations: CI, confidence interval; HR, hazard ratio; PRS, polygenic risk score; SD, standard deviation.

The association between the healthy lifestyle score and diverticulitis was similar across categories of PRS (Table 3). For instance, compared to participants with a score of 0, those with a score of 4-5 had an HR (95% CI) of 0.63 (0.42-0.94) for individuals in the lowest tertile of PRS, 0.48 (0.36-0.64) in the middle tertile, and 0.50 (0.39-0.63) in the highest tertile (P for multiplicative interaction = 0.67). Population-attributable risk analysis showed that adopting a healthy lifestyle could prevent 23% to 42% of diverticulitis cases across PRS categories.

Table 3.

Association of healthy lifestyle score with incident diverticulitis according to PRS tertile in NHS, NHSII, and HPFS

	All participants	Low PRS	Middle PRS	High PRS	P for multiplicative interaction b	P for additive interaction c
Cases/Person-years	2357/808408	470/276342	726/272013	1161/26005	-
Healthy lifestyle score, multivariable-adjusted HR (95% CI) a
0 (unhealthy)	1 [ref]	1 [ref]	1 [ref]	1 [ref]	.67	<.0001
1	0.81 (0.71-0.93)	1.24 (0.88-1.74)	0.74 (0.58-0.94)	0.71 (0.59-0.86)
2	0.68 (0.59-0.78)	0.88 (0.62-1.25)	0.67 (0.52-0.85)	0.62 (0.51-0.75)
3	0.66 (0.57-0.76)	0.99 (0.70-1.41)	0.58 (0.45-0.75)	0.60 (0.49-0.74)
4,5 (healthy)	0.50 (0.43-0.59)	0.63 (0.42-0.94)	0.48 (0.36-0.64)	0.50 (0.39-0.63)
Per 1 increase	0.87 (0.84-0.90)	0.89 (0.83-0.95)	0.86 (0.81-0.92)	0.87 (0.83-0.91)
P for trend	<.0001	.001	<.0001	<.0001
PAR of healthy lifestyle score d	26% (18%-34%)	42% (37%-46%)	31% (13%-47%)	23% (11%-35%)

^aModels were adjusted for alcohol intake, total calorie intake, regular aspirin, non-steroidal anti-inflammatory drug, and acetaminophen use, physical examination in the past two years, and menopausal hormone therapy, and were stratified by age, questionnaire cycle, and cohort.

^bP value for multiplicative interaction was assessed using the Wald test by including a product term between continuous healthy lifestyle score and continuous PRS.

^cP value for additive interaction was assessed between continuous unhealthy lifestyle score (evaluated in reverse from healthy lifestyle score so effect estimate was positive) and continuous PRS. Relative excess risk due to interaction (RERI) was 0.09 (95% CI, 0.06-0.13).

^dPAR was calculated with healthy lifestyle score of 4 or 5 as reference, indicating the percentage of diverticulitis that can be prevented if all individuals were in this group.

Abbreviations: CI, confidence interval; HR, hazard ratio; PAR, population attributable risk; PRS, polygenic risk score.

We also examined the combined impact of PRS and lifestyle on the risk of diverticulitis. Individuals in the highest PRS group with a healthy lifestyle score of 0 or 1 were 4.8-fold more likely to develop diverticulitis compared to those in the lowest PRS group with a score of 4 or 5 (Supplemental Figure 2). There was a statistically significant additive interaction between lifestyle score and PRS when both were evaluated as continuous variables (RERI: 0.09; 95% CI, 0.06-0.13; P for additive interaction < .0001), indicating that with each 1-point decrease in healthy lifestyle score and each SD increase in PRS, the hazard ratio of developing diverticulitis was 0.09 higher than it would be without the interaction.

In our validation MGBB cohort, an association between PRS and diverticulitis risk was confirmed. We found that for each SD increase in PRS, the odds of developing diverticulitis increased by 1.67 times (95% CI, 1.60-1.75). A higher BMI was associated with an increased risk of diverticulitis regardless of the genetic risk level (Table 4). Individuals in the highest PRS tertile with a BMI ≥ 30 kg/m² were 5.55 times (95% CI, 4.51-6.82) more likely to develop diverticulitis compared to those in the lowest PRS tertile with a BMI < 25 kg/m². Furthermore, a similar additive interaction between BMI and PRS on diverticulitis risk was also observed (RERI, 0.13; 95% CI, 0.09-0.17; P for additive interaction < .0001).

Table 4.

Risk of diverticulitis according to joint categories of BMI and PRS in MGBB

	BMI			P for additive interactiona
PRS	<25 kg/m2	25-29.9 kg/m2	≥30 kg/m2
Low	1 [ref]	1.27 (1.00-1.61)	1.73 (1.37-2.18)	<.0001
Middle	1.46 (1.14-1.87)	2.38 (1.92-2.96)	2.96 (2.38-3.68)
High	3.12 (2.50-3.89)	4.36 (3.53-5.36)	5.55 (4.51-6.82)

^aP value for additive interaction was assessed between continuous BMI and continuous PRS. Relative excess risk due to interaction (RERI) was 0.13 (95% CI, 0.09-0.17).

Abbreviations: BMI, body mass index; CI, confidence interval; MGBB, Mass General Brigham Biobank; PRS, polygenic risk score; RERI, relative excess risk due to interaction

DISCUSSION

In the current study, we showed that a healthy lifestyle score, comprising five components, including no smoking, normal BMI, adequate physical activity, high fiber intake, and low red meat intake, was associated with a lower risk of diverticulitis across multiple cohorts, including a cohort comprised of predominantly non-Hispanic Black participants. Conversely, individuals with a higher PRS were shown to be at increased risk of developing diverticulitis. However, regardless of the genetic risk, adherence to a healthy lifestyle was associated with a significantly decreased risk of the disease in both prospective cohort studies and an EHR-linked biobank study.

Our results support the benefits of adherence to a healthy lifestyle in reducing the risk of developing diverticulitis among various racial groups. Existing evidence on the associations of lifestyle, including obesity^{36 38}, physical activity^{38 39}, smoking^{32 33}, and diet^{40 41}, with the risk of incident diverticulitis was largely from prospective cohorts of predominantly non-Hispanic White populations, whereas the role lifestyle plays in diverticulitis risk among other minority groups remains poorly understood⁴². The risk of diverticulitis or other diverticular disease may vary by race or ethnicity. For example, non-Hispanic Black individuals were reported to have the highest prevalence of proximal diverticulosis among those who underwent colonoscopies⁴³, while analyses using National Inpatient or Emergency Department Visits data have found that the prevalence of diverticular bleeding was highest in Blacks whereas that of diverticulitis was highest in Whites^{44 45}. By leveraging data from the SCCS cohort including 65% non-Hispanic Black individuals, we were able to enhance the generalizability of our findings and offer insight into the impact of lifestyle on diverticulitis risk in diverse populations.

While the association between overall lifestyle score and diverticulitis was similar across cohorts, we observed differences for individual lifestyle factors. For example, the inverse association of physical activity with diverticulitis in women (NHS/NHSII) was linear. In men (HPFS), however, only vigorous physical activity showed a significant association³⁹. Overweight and obesity appeared to be a more important risk factor in women, whereas diet showed a stronger association with diverticulitis in men. These differences may reflect variation in the relative contribution of potential mechanisms, such as metabolic dysfunction or inflammation, in the development of diverticulitis across populations.

It should be noted that existing GWASs have focused on diverticular disease as the primary outcome due to limitations in the diagnostic coding within biobank studies, making it challenging to distinguish diverticulitis from diverticulosis. Despite this, recent studies have demonstrated that utilizing a PRS combining the subtle effects of individual genetic variants associated with diverticular disease can successfully predict the risk of diverticulitis^11-13. A higher PRS was associated with a greater risk of diverticulitis as well as severe diverticulitis and recurrent diverticulitis^{12 13}. We herein validated the association between PRS and incident diverticulitis in both population-based cohort and electronic health record-linked biobank. Additionally, the association on the relative scale appeared to be more pronounced in individuals under the age of 60, indicating that PRS might be more predictive for diverticulitis in younger populations.

There was no statistically significant multiplicative interaction between the healthy lifestyle score and PRS in relation to diverticulitis, suggesting that the impact of genetic or lifestyle factors on diverticulitis does not depend on each other. These results suggest that most individuals, regardless of genetic risk, could achieve similar relative benefits in reducing the risk of diverticulitis by adopting a healthy lifestyle. However, we observed a modest yet significant interaction on the additive scale, indicating the combined impact of genetic susceptibility and lifestyle resulted in a slightly greater risk than the sum of the individual effects when considered independently. This may suggest that the absolute risk of diverticulitis associated with a poor lifestyle varies according to genetic risk, possibly depending on the baseline risk at the population level. These findings have direct implications for both clinical practice and public health.

The notable strengths of this study include the incorporation of well-established prospective cohort studies including diverse racial and ethnic populations leading to greater generalizability, detailed and repeated collection of lifestyle factors to minimize the potential for measurement error, the utilization of PRS based on results from the latest and largest GWAS, as well as the validation of genetic-lifestyle associations with diverticulitis in an external biobank study. Nevertheless, it should be acknowledged that we were unable to examine the interaction between PRS-lifestyle derived from the NHS, NHSII and HPFS cohorts with the full healthy lifestyle score in the MGB Biobank given the available data on lifestyle. Further, the ascertainment of diverticulitis was based on different approaches (i.e., self-reports, ICD codes) in different cohorts. However, the comparative validity of these methods has been demonstrated to be high. Moreover, variation in the ascertainment of diverticulitis within each individual cohort would be expected to underestimate the consistency of the associations that we observed between PRS, lifestyle score, and diverticulitis across the cohorts.

In conclusion, our data provide consistent evidence from multiple datasets indicating that adherence to a healthy lifestyle is linked to a reduced risk of developing diverticulitis, irrespective of one’s genetic predisposition.

What is already known on this topic

Prospective cohort studies have linked lifestyle factors to risk of diverticulitis and diverticular disease. Genome-wide association studies have identified up to 150 genetic variants associated with diverticular disease. However, no study has examined whether lifestyle changes can mitigate the genetic risk of diverticulitis or how genetic predisposition and lifestyle factors interact to influence disease risk.

What this study adds

We provide consistent evidence from four prospective cohort studies that a healthy lifestyle score incorporating obesity, physical activity, smoking, dietary fiber, and red meat consumption was associated with a lower risk of diverticulitis across diverse populations. Moreover, regardless of genetic risk, adherence to a healthy lifestyle was associated with a reduced risk of diverticulitis in both prospective cohort studies and an electronic health records-linked biobank study.

How this study might affect research, practice or policy

Maintaining a healthy lifestyle was associated with a reduced risk of developing diverticulitis, regardless of population differences and genetic susceptibilities.