Using an approach involving physiologically based toxicokinetic (PBTK) modelling and Monte Carlo simulation (MCS), we investigated the impact of the biological variability affecting the parameters (e.g. physiological, physicochemical, biochemical) which determine toluene (TOL) kinetics on two exposure indicators (EIs): urinary excretion of o-cresol (o-CR), measured at the end of an 8 h exposure at 50 ppm, and unchanged TOL in blood (B-TOL) sampled prior to the last shift of a 5 day workweek. Population variance was described by assuming normal, or lognormal, distribution of parameter values and assigning to each one+/-2 S.D. (or+/-2 G.S.D.). PBTK-MCS (n=1000) resulted in a geometric mean (G.M.) of 0.635 mmol/mol creatinine for urinary o-CR, upper and lower limits (95%) ranging from 0.23 to 1.75, whereas, the GM for B-TOL was 120.6 microg/l (95% limits: 64.5-225.7). Overall, the results showed that this approach facilitates the prediction of the range of BEI values that could be anticipated to occur in a group of workers exposed to a chemical.