Human biomonitoring provides an integrated measure for individual exposure to environmental pollutants. Better insight in inter-individual variability of biomarkers of exposure may help in the interpretation of biomonitoring studies. The aim was to study the impact of outliers, determine the optimal unit for fat-soluble biomarkers in serum and quantify the major determinants for biomarkers of exposure to polychlorinated aromatic hydrocarbons (PCAHs) in three age groups. Data were obtained from the Flemish Environment and Health Study (2002-2006). Marker PCBs (sum of 138, 153, 180), hexachlorobenzene (HCB) and p,p'-DDE were measured in cord blood samples of 1196 newborns, in serum samples of 1679 adolescents (14-15 years) and 1583 adults (50-65 years). Exclusion of influential outliers in multiple linear regression models lead to models that are better applicable to the general population. In terms of adjusted R2, the regression model with the pollutant expressed in volume-based units and blood fat as a separate independent variable was superior compared to models with other units. We found highly consistent relationships between the serum concentration of PCAHs and blood fat, age, changes in body weight, animal fat in the diet, local vegetable consumption (HCB and p,p'-DDE only) and being breastfed as a baby (in adolescents only). The impact of sex and BMI differed by age. For biomarkers of persistent pollutants that reflect long-term exposure, the relation between the covariates and the biomarkers can be well quantified.