Associations of air cleaning interventions in primary school classrooms with indoor environmental quality and perceived symptoms

• We conducted an air cleaner intervention in occupied primary school classrooms. • Air cleaners significantly reduced TVOC levels and PM2.5 I/O ratio. • Microbial levels were generally unaffected, while fungal diversity was reduced. • Pupils’ perceived symptoms were reduced both during air cleaner and SHAM operation. • Future studies should explore possible placebo effect and high-exposure settings. Air cleaners are used to remove various pollutants from air in different types of indoor environments. Studies on the air cleaners’ efficiency to remove pollutants are largely performed in chamber and test room settings under controlled conditions, rarely in occupied, real-life environments. The objective of this study was to determine the effect of portable air cleaners on several IEQ parameters in Finnish classrooms as well as on the perceived IEQ and symptoms reported by the pupils. We carried out an intervention study in 27 occupied classrooms (6 schools) where air cleaners equipped with either EPA or HEPA and activated carbon filters were operated, in both regular and SHAM, i.e. without filters, mode. We performed a comprehensive exposure assessment, including measurement of particles, microbes and VOCs, in addition to pupils’ reports on perceived IEQ and symptoms. We found that air cleaner operation (ON versus OFF) reduced average concentration of TVOC by 52.9 % and PM 2.5 I/O ratio by 24.8 %. In addition, air cleaner operation resulted in a 25.8 % and 8.2 % reduction in fungal richness and diversity, respectively, and a 16.8 % and 45.3 % increase in the relative abundance of two dominant bacterial genera. We observed small reductions in the pupils’ reports of upper and lower respiratory symptoms as well as other symptoms during air cleaner operation (4.7–7.5 %). However, we found similar reductions during SHAM operation (6.1–8.7 %). Considering the overall low pollutant concentrations observed in the examined classrooms, a placebo effect in symptom reporting is likely to explain these findings.