OBJECTIVE This study assessed personal exposure of pregnant women to fine particles (PM(25)) and benzo[a]pyrene (B[a]P) and the relationship between pollutant concentrations in ambient and indoor air. METHODS In a group of 78 pregnant women, simultaneous 48 h measurements of personal, indoor, and outdoor exposure to PM(25) and B[a]P were carried out in the second trimester of pregnancy. The results show that participants were exposed to varying concentrations of PM(25) and B[a]P, with higher exposure in the winter season. Overall, the mean personal PM(25) level was 30.4 microg/m(3) and B[a]P 2.1 ng/m(3). The winter/summer ratios for mean personal exposures were 1.4 (35.6 microg/m(3) vs. 25.8 microg/m(3)) and 5.4 (4.9 ng/m(3) vs. 0.9 ng/m(3)), respectively. As for indoor levels, the winter/summer ratios were 1.4 (33.2 microg/m(3) vs. 24.4 microg/m(3)) for PM(25) and 5.4 (4.3 ng/m(3) vs. 0.8 ng/m(3)) for B[a]P, and for outdoor concentrations, the respective values were 1.5 (40.3 microg/m(3) vs. 26.4 microg/m(3), and 6.8 (6.1 ng/m(3) vs 0.9 ng/m(3)). A stronger correlation was found between personal PM(25) exposure and the pollutant concentration indoors (r = 0.89; 95% CI: 0.83-0.93) than outdoors (r = 0.75; 95% CI: 0.64-0.83). The correlations between personal B[a]P exposure and its indoor or outdoor levels were similar (0.95-0.96) and significant. The markedly higher exposure to B[a]P in Kraków in winter than in summer can be explained by the massive use of coal for heating in the cold season. CONCLUSIONS We conclude that although ambient PM(25) measurements provide an adequate indicator of outdoor air quality for use in epidemiologic studies, they may not be adequate for studies on relationship between non-ambient pollution and health effects. Since only about 20% of variability in personal B[a]P exposure could be explained by personal PM(25) level, the extrapolation of personal exposure to B[a]P from personal PM(25) data may be greatly underestimated.