BACKGROUND: Adverse cardiovascular effects associated with air pollution exposure have been widely demonstrated. However, inconsistent cardiovascular responses were observed from reducing indoor air pollution exposure. We aimed to assess whether short-term air filtration intervention could benefit cardiovascular health in elderly living in high pollution area. METHODS: A randomized crossover intervention study of short-term indoor air filtration intervention on cardiovascular health was conducted among 35 non-smoking elderly participants living in Beijing in the winter of 2013, as part of Beijing Indoor Air Purifier StudY (BIAPSY). Portable air filtration units were randomly allocated to active filtration for 2 weeks and sham filtration for 2 weeks in the households. Twelve-hour daytime ambulatory heart rate variability (HRV) and blood pressure (ABP) were measured during active and sham filtration. Concurrently, real-time indoor and outdoor particulate matter with diameter less than 2.5鈥痬icrom (PM(2.5)) and indoor black carbon (BC) concentrations were measured. We applied generalized additive mixed models to evaluate the associations of 1- to 10-h moving average (MA) exposures of indoor PM(2.5) and BC with HRV and ABP indices, and to explore whether these associations could be modified by air filtration. RESULTS: We observed decreases of 34.8% in indoor PM(2.5) and 35.3% in indoor BC concentrations during active filtration. Indoor PM(2.5) and BC exposures were significantly associated with reduced HRV and increased ABP indices, and greater changes were observed during sham filtration. In specific, each 10鈥痬ug/m(3) increase in indoor PM(2.5) at MA8-h was associated with a significant reduction of 1.34% (95% CI: -2.42, -0.26) in SDNN during sham filtration, compared with a non-significant reduction of 0.81% (95% CI: -6.00, 4.68) during active filtration (P(inter)<鈥?.001). Each 1鈥痬ug/m(3) increase in indoor BC at MA8-h was associated with a significant increase of 2.41% (95% CI: 0.38, 4.47) in SBP during sham filtration, compared with a non-significant increase of -1.09% (95% CI: -4.06, 1.96) during active filtration (P(inter) =鈥?.135). Nonlinear inverse exposure-response relationships of indoor air pollution exposures with predicted HRV and ABP indices also confirmed some cardiovascular benefits of short-term air filtration intervention. CONCLUSIONS: Our results suggested that short-term indoor air filtration intervention can be of some cardiovascular benefits in elderly living with high pollution episodes.