Background Both diet and air pollution are associated with mortality risks. However, no epidemiological study has examined the potential interaction between diet and air pollution on mortality. We assessed their interaction on an additive scale.

Methods We analysed the data on daily concentrations of ambient air pollutants (PM₁₀, NO₂, SO₂ and O₃) and a total of 23 484 deaths in 1998 in Hong Kong. A standardised questionnaire was used in all four death registries to collect food frequency data from proxy respondents while waiting for the registration to be completed. We fitted a linear odds ratio model and estimated excess relative risk due to the interaction (ERRI) between air pollution and regular consumption (at least once per week) of each food item to measure departure from additivity of effects on mortality.

Results We observed consistently negative ERRI between all of the four pollutants and regular consumption of vegetables, fruits and soy. The effects of PM₁₀, NO₂ and O₃ were significant smaller in the subjects who regularly consumed fruits than those who never or seldom consumed such food. The effect modification of soy consumption on PM₁₀, NO₂ and SO₂ associated mortality was also found statistically significant. However, regular consumption of dairy products was associated with significant increased effects of PM₁₀ and NO₂.

Conclusions This study provides insight into dietary habit as one of the modifiers of health effects of air pollution. Our findings merit further studies to characterise the influence of diet on air pollution-related health and elucidate the underlying mechanisms.