Katja Fall

Supplementary Table 8 reports stratified associations between the Anti-Inflammatory Diet Index (AIDI) and bladder cancer risk by baseline smoking status (never vs ever smokers). Hazard ratios (HRs) and 95% confidence intervals are presented for AIDI quartiles (Q2–Q4 vs Q1) for baseline AIDI (1998) and for AIDI modelled as a repeated measure (1998 and 2009; cumulative-average method), including p-values for trend within each smoking stratum. Fully adjusted Model 3 covariates are specified, and p-values for interaction between AIDI trend and smoking group are provided for both exposure specifications.

Associations between only pro-inflammatory components of AIDI and bladder cancer risk, overall and by disease stage (1998–2020).

Supplementary Table 6 presents sensitivity analyses examining whether additional adjustment for alcohol intake and/or leisure-time physical activity influences the association between the Anti-Inflammatory Diet Index (AIDI) and bladder cancer risk. Hazard ratios (HRs) and 95% confidence intervals are shown for quartiles of AIDI (Q2–Q4 vs Q1) and p-values for trend, using Model 3 for baseline AIDI (1998) and for AIDI modelled as a repeated measure (1998 and 2009; cumulative-average method). Results are reported for the full sample and for complete-case subsets for alcohol, physical activity, and both combined, with models additionally adjusted for total alcohol intake (g/day) and/or leisure-time exercise (hours/week category) as indicated.

Supplementary Table 2 reports hazard ratios (HRs) and 95% confidence intervals for bladder cancer risk across quartiles of the Anti-Inflammatory Diet Index (AIDI) among Swedish women (1998–2020), overall and stratified by tumour stage (non-muscle invasive and muscle invasive). Results are shown for baseline AIDI (1998) and for AIDI modelled as a repeated measure (1998 and 2009; cumulative-average method), with tests for linear trend. Estimates are presented from three progressively adjusted Cox models: Model 1 (age and sex), Model 2 (additionally smoking pack-years, BMI, education, employment status, and mean-centred energy intake), and Model 3 (additionally diabetes, hypertension, and family history of cancer). Case counts and person-years are provided for each AIDI category, and stage-specific analyses include cases diagnosed from 2004 onwards.

Supplementary Table 3 reports hazard ratios (HRs) and 95% confidence intervals for bladder cancer risk across quartiles of the Anti-Inflammatory Diet Index (AIDI) among Swedish men (1998–2020), overall and stratified by tumour stage (non-muscle invasive and muscle invasive). Results are shown for baseline AIDI (1998) and for AIDI modelled as a repeated measure (1998 and 2009; cumulative-average method), with tests for linear trend. Estimates are presented from three progressively adjusted Cox models: Model 1 (age and sex), Model 2 (additionally smoking pack-years, BMI, education, employment status, and mean-centred energy intake), and Model 3 (additionally diabetes, hypertension, and family history of cancer). Case counts and person-years are provided for each AIDI category, and stage-specific analyses include cases diagnosed from 2004 onwards.

Descriptive baseline characteristics of Swedish men and women by quartiles of the AIDI (maximum score = 13).

Supplementary Table 7 shows lagged sensitivity analyses evaluating the association between the Anti-Inflammatory Diet Index (AIDI) and bladder cancer risk after excluding cases diagnosed early in follow-up (no lag, 2-year lag, and 3-year lag). For each lag period, the table reports the number of excluded cases, remaining cases and person-years, hazard ratios (HRs) with 95% confidence intervals for AIDI quartiles (Q2–Q4 vs Q1), and p-values for trend, using fully adjusted Model 3 for baseline AIDI (1998) and for AIDI modelled as a repeated measure (1998 and 2009; cumulative-average method).

Associations between only anti-inflammatory components of AIDI and bladder cancer risk, overall and by disease stage (1998–2020).

Associations between AIDI and bladder cancer risk by stage (1998–2020).

Supplementary Table 1 lists the food frequency questionnaire (FFQ)-derived components of the Anti-Inflammatory Diet Index (AIDI), indicates whether each component is classified as anti-inflammatory or pro-inflammatory, specifies the intake cut-offs used to assign a score of 1, and defines the binary scoring rule (1 if the criterion is met; otherwise 0).