Biomonitoring of possible human exposure to environmental genotoxic chemicals: lessons from a study following the wreck of the oil tanker Braer.

In January 1993 the oil tanker Braer ran aground in the Shetland Islands, Scotland. Approximately 80,000 tons of crude oil were released. Exceptionally high winds caused extensive pollution and exposure of the local population to crude oil. We describe the study which was immediately set in place to examine the exposed population for evidence of genotoxic exposure. Blood samples were taken and primary DNA damage was measured in the mononuclear cell fraction by the butanol modification of the 32P-postlabelling method. Mutation was measured at the hprt locus in T lymphocytes. No evidence of genotoxicity was obtained for either end point, but nevertheless, we believe that useful lessons were learnt, which should be incorporated into the design of future studies: (1) A rapid response is essential, and even if sufficient funds are not immediately available, it is still worth attempting to obtain samples quickly and use cryopreservation, also to attempt to estimate exposure. (2) Adequate numbers of volunteers must be sought, together with enough controls, not just to allow meaningful analysis but to overcome loss of samples and failure of things to go according to plan. (3) Points concerning laboratory practice include: (i) samples should be coded, (ii) clearly defined and proven protocols should be used, (iii) irreplaceable samples should not be used for method development, (iv) should a problem become apparent during the study, work on such samples should cease immediately until the problem is solved, (v) all critical experimental components should be pretested against a laboratory standard. (4) The study design should include replicate experiments to monitor experimental variability and reproducibility, as well as internal standards and cryopreserved "in house" samples. Care must be taken that samples from any one exposure group are spread between a number of independent experiments and that each experiment includes samples from a number of exposure groups. (5) A computerised data base should be maintained with full details of experimental variables, donor attributes, and raw data so that any contribution of experimental artefacts to "outlier" results can be monitored. (6) Because of the nature of the statistical variation for many environmental genotoxicity end points, only a large-scale study is likely to be capable of yielding useful information.