The use of anti-neoplastic drugs is increasing and nursing staff are evidently concerned about the risk of hazardous exposure. General methods are available for measuring exposure to carcinogens. These methods, however, lack specificity. One group of anti-neoplastic drugs, based on cisplatin derivatives, contain platinum, which can be analysed at low concentrations with the use of adsorptive voltammetry. The aim of this study was to employ this technique for assessment of occupational exposure by selective determination of platinum in workroom air and in blood and urine samples from medical staff nursing cancer patients. The voltammetric method for determination of platinum has been further developed for analysis of blood, urine and air samples. The effects of different materials in crucibles and different acids for sample pre-treatment, and of the ashing temperature programme were investigated for optimum conditions for analysis of blood, urine and filter samples. Occupational exposure to cisplatin derivatives was also investigated among medical staff. Air sampling in the workroom, as well as analysis of blood and urine samples from the exposed subjects, was carried out during the process of drug preparation and administration and when attending treated patients. After development of the experimental method by the introduction of synthetic quartz crucibles and the optimization of acid composition, and of the temperature programme for sample pre-treatment, the recovery has been improved and the method has proved to be adequate for determination of platinum (Pt) in blood, urine and air-filter samples. Increased Pt blood levels were found in both graduate (2.2 +/- 1.7 ng ml-1) and staff nurses (3.8 +/- 4.0 ng ml-1), but not in pharmacists (0.47 +/- 0.31 ng ml-1), in comparison with unexposed subjects (1.2 +/- 0.69 ng ml-1). The variation was, however, high with CV > 50% for all groups. From the median values it can be seen that a few subjects with high values had a large impact on the mean values. The mean Pt level in urine samples was 126 +/- 92 ng l-1, which was found to be close to the pooled reference urine (110 ng l-1). No increased Pt air levels were found in any of the measurements. Moreover, the results showed difference in mean Pt blood level between staff at the investigated hospitals. The staff at hospital A had 1.2 +/- 0.53 ng ml-1, at hospital B 3.8 +/- 4.3 ng ml-1 and at hospital C 2.0 +/- 2.1 ng ml-1. The method has proved to be adequate for determination of platinum in blood, urine and air-filter samples. No increased airborne Pt levels were found. However, increased Pt blood levels were found. Staff nurses had a higher mean level than graduate nurses, which indicates that possible exposure occurs while attending treated patients rather than during the preparation and administration of drugs. There was a noticeable variations in the mean blood level for the investigated groups as a whole. This variation might reflect an actual difference in the exposure situation, but more probably depends on differences in skill and experience between the subjects. Routines and facilities for handling anti-neoplastic drugs vary between the clinics and this also probably has an impact on exposure.