Currently, there is no single reference material commercially available with certified concentrations of Co, Cr, Ni and Mo in bodily fluids of unexposed subjects. This study provides validation data for the low level determination of Co, Cr, Mo and Ni in whole blood, serum and urine, highlighting the severe restrictions of currently available reference materials for this kind of analysis. The riverine water reference material SLRS-4 has been considered here as well because certified values for this matrix are available in contrast to the biological reference materials from Sero used throughout this study. In addition, spiking experiments with whole blood, serum and urine at concentrations of 0.25 µg l−1 and 0.50 µg l−1 revealed recoveries of 100 ± 5% underpinning the quality of the generated analytical data. All samples and standards have been diluted with an alkaline diluent containing 0.01 M NH3, 0.2 mM (NH4)2EDTA and 0.07% Triton X-100. Measurements were carried out with a sector field ICP-MS (ICP-SMS) operated under clean room conditions in the medium resolution mode (m/Δm ∼ 4000) to separate the analyte signals from severe polyatomic spectral interferences hampering the reliable determination of Co, Cr, Ni and Mo. Indium was used as internal standard element to correct for potential drifts in the ICP-SMS response. While the analysis of whole blood, serum and urine worked well for a few hours (serum and urine up to 24 h) using dilutions of only 1 + 9, dilutions of 1 + 19 gave the most robust instrumental performance for long-term measurements (>24 h). Considering a dilution factor of 20, conservative method detection limits of 0.005 µg l−1, 0.02 µg l−1, 0.10 µg l−1 and 0.01 µg l−1 have been calculated for Co, Cr, Mo and Ni. The inter-day reproducibility varied from 2% (Co, Mo) to 5% (Cr) and 10% (Ni) for low level blood samples. For serum and urine these numbers improved to 2–4%.