* Corresponding author. Tel.: #358-17-201-347; fax: #358- 17-201-184. E-mail address: rufus.edwards@ktl." (R.D. Edwards). Atmospheric Environment 35 (2001) 1411}1420 Benzene exposure in Helsinki, Finland R.D. Edwards*, M.J. Jantunen Department of Environmental Hygiene, KTL Finnish National Institute of Public Health, P.O. Box 95, 70701 Kuopio, Finland Received 19 April 2000; received in revised form 27 June 2000; accepted 7 July 2000 Abstract Personal exposures and microenvironmental concentrations of benzene were measured in the residential indoor, residential outdoor and workplace environments for 201 participants in Helsinki, Finland, as a component of the EXPOLIS-Helsinki study. Median benzene personal exposures were 2.47 (arithmetic standard deviation (ASD)"1.62) gm for non-smokers, 2.89 (ASD"3.26) gm for those exposed to environmental tobacco smoke in any microenvironment and 3.08 (ASD"10.04) gm for active smokers. Median residential indoor benzene concentra- tions were 3.14 (ASD"1.51) gm and 1.87 (ASD"1.93) gm for environments with and without tobacco smoke, respectively. Median residential outdoor benzene concentrations were 1.51 (ASD"1.11) gm and median workplace benzene concentrations were 3.58 (ASD"1.96) gm and 2.13 (ASD"1.49) gm for environments with and without tobacco smoke, respectively. Multiple step-wise regression identi"ed indoor benzene concentrations as the strongest predictor for personal benzene exposures of those not exposed to tobacco smoke, followed sequentially by time spent in a car, time in the indoor environment, indoor workplace concentrations and time in the home workshop. Relationships between indoor and outdoor microenvironment concentrations and personal exposures showed consider- able variation between seasons, due to di!erences in ventilation patterns of homes in these northern latitudes. Automobile use-related activities were signi"cantly associated with elevated benzene levels in personal and indoor measurements when tobacco smoke was not present, which demonstrates the importance of personal measurements in the assessment of exposure to benzene.  2001 Elsevier Science Ltd. All rights reserved. Keywords: Air quality; Environmental monitoring; Seasonal variation; Exposure; Automobiles; Environmental tobacco smoke (ETS) 1. Introduction Studies have shown benzene indoor concentrations to be higher than outdoor levels, personal exposures to be higher than indoor air levels and large variation in per- sonal benzene exposures within the population (Hartwell et al., 1992). In addition the US EPA VOC-TEAM study reported that the main sources of human exposure to benzene for much of the US population were associated with personal activities and indoor sources and not the large outdoor point sources that had previously been suspected (Wallace, 1989; Ott and Roberts, 1998). It is increasingly apparent, therefore, that if signi"cant reduc- tions in benzene exposures in the general population are required, we should examine which sources are the largest contributors to personal exposures in repre- sentative-population-based samples, and where the most signi"cant and cost-e!ective strategies could be implemented. The "rst stage in such a process is the identi"cation of various source contributions in speci"c microenvironments that an individual spends his time, and their contribution to the total exposure. These may vary depending on the activities of the individual and the nature of the environment that he spends his time. Helsinki presents a unique opportunity for the study of urban air pollutants related to automobile use and their contribution to personal exposures as it is located on the Gulf of Finland and background sources are low. More interestingly, almost all of the building stock in the city has winter heating supplied by co-generating power 1352-2310/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 2 - 2 3 1 0 ( 0 0 ) 0 0 3 5 9 - 9