Joint Associations of Humid-Heat Exposure, Social Vulnerability, and Habitual Numbing–Spicy Diet with Cardiovascular Disease: A Prospective Cohort Study

Background: Social, behavioural, and environmental determinants of cardiovascular disease (CVD) frequently co-occur, but their joint associations are poorly understood. We examined whether social vulnerability and habitual numbing–spicy food consumption modified the association between pre-baseline residential humid heat and CVD. Methods: The Chongqing sub-cohort of the China Multi-Ethnic Cohort enrolled 23,308 adults aged 30–79 years in 2018–2019 and followed them through September 2025. The incidence analysis included 21,734 individuals and 641 CVD events. The mortality analysis included 22,923 individuals and 82 CVD deaths. Median follow-up was 6.84 years (145,349 and 154,559 person-years, respectively). Social vulnerability was measured using a 16-item psychosocial risk score (PsRS). Diet was classified into four categories for the incidence analysis and as any versus no numbing–spicy consumption for the mortality analysis. Humid-heat exposure was characterised by total humid-heat days, heatwave frequency, and longest heatwave duration, derived from daily Heat Index values in the three complete years before baseline. Cox models estimated the main associations and tested two-way and three-way interactions. For each heat metric, 36 exposure combinations were defined by 3 PsRS groups × 4 dietary categories × 3 heat-exposure groups in the incidence analysis. Results: Consumption of both numbing and spicy foods was associated with lower CVD incidence (HR 0.75, 95% confidence interval [CI] 0.61–0.93). The three-way interaction for heatwave frequency was statistically significant before (p = 0.0433) but not after (p = 0.1299) Holm adjustment across the three heat metrics. No corresponding interaction was detected for total HTH days (p = 0.8514) or longest heatwave duration (p = 0.4190). In exploratory analyses, high social vulnerability was associated with CVD mortality (HR range 2.73–2.78, all p < 0.05). Neither diet nor humid heat showed a statistically significant association with CVD mortality. Conclusions: The three-way interaction involving heatwave frequency, social vulnerability, and diet was exploratory and did not remain statistically significant after correction for multiple testing. These findings warrant replication. The dietary associations should not be interpreted as evidence of a causal protective effect.