Particulate-Matter Carcinogenic Risk in Thirteen European Indoor Micro-Environments: The Limits of Pre–Post Evaluation of Portable Air Cleaning (TwinAIR)

Indoor particulate matter is a Group 1 carcinogen with no identified exposure threshold, and the portable air cleaners now widely deployed in European schools, hospitals and offices are almost always evaluated by a change in mean concentration measured before and after installation in the same room. We applied a probabilistic incremental lifetime cancer risk (ILCR) framework to size-resolved particulate matter in thirteen occupied micro-environments in five countries, monitored continuously at five-minute resolution within the TwinAIR pilot network, and then tested that evaluation design against the measurements it rests on. Risk was uniformly low: every site-by-fraction estimate fell below the 10−6 de minimis threshold by between one and three orders of magnitude, and the conclusion survived recomputation from the measured concentration distributions (11.2 to 353 times below the threshold across forty-two estimates) and the substitution of paediatric exposure parameters. The evaluation design did not survive. In the two rooms that received air cleaners, measured PM2.5 rose by 7.8% and 21.0%, while a thermo-hygrometric regression explaining 1–6% of the variance—the concentration model underlying the probabilistic estimates—predicted reductions of 6.7% and 2.3%. Applying the same phase boundary to four sensors in three micro-environments that received no device produced changes from −74.3% to +37.2%, exceeding in magnitude and spanning in direction the contrast attributed to the intervention. The novelty of this work is therefore twofold: a harmonised probabilistic risk assessment of thirteen real occupied interiors on a single scale, and a demonstration, from the same network’s data, that within-room pre–post designs without concurrent controls cannot identify an air-cleaning effect.