7. CONCLUSIONS

1. Both farmed and wild fish species in Finland were found to be contaminated by C. botulinum type E, with prevalence levels of 17% and 20%, respectively. Eight percent of fish roe samples and 4% of various types of fishery products were positive for type E. Vacuum-packaged hot-smoked whitefish was identified as a particularly high risk product with a contamination level of 10%. The results indicate that current fish processing practices are inadequate to eliminate type E spores from fishery products.

2. RAPD analysis with two arbitrary primers (OPJ 6 and OPJ 13) was found to be a discriminating and reproducible method for genotyping type E strains. It had a higher discriminatory power for certain sets of isolates than PFGE with two macrorestriction enzymes (SmaI/XmaI and XhoI), although in general, MRPs generated by PFGE were more discriminating and easier to interpret. The distinct advantages of RAPD as compared to PFGE were 100% typeability and rapid performance. Rep-PCR was not discriminatory enough to be a useful tool for type E subtyping.

3. High genetic biodiversity was observed among type E isolates regardless of the isolation source or geographical origins. Combined results of PFGE and RAPD analyses yielded 62 subtypes among 92 isolates. Extensive genetic variation was observed among strains isolated from different fish species as well as among isolates from one fish species or even from one individual fish. Processing did not seem to favor survival of any particular genotype. The wide genetic biodiversity observed among C. botulinum type E strains facilitates the use of DNA-based typing methods as a tool in contamination studies in the food industry and in investigations of botulism outbreaks.

4. At the mildly abusive storage temperature of 8^oC, C. botulinum type E was observed to produce toxin in vacuum-packaged unprocessed and cold-smoked rainbow trout in two and three weeks, respectively, with an inoculum corresponding to the natural contamination levels of fish. Even at 4^oC, vacuum-packaged cold-smoked rainbow trout became toxic after four weeks. In each case, toxin production occurred with very low or no detectable growth by type E. Vacuum-packaged raw pickled (gravad) rainbow trout stored at 6^oC was determined to be safe with respect to type E, provided that the declared shelf life is within reasonable limits. The results strongly indicated that unless the constant maintenance of a chill chain temperature below 3^oC can be guaranteed, the shelf life of vacuum-packaged cold-smoked rainbow trout in its current formulation should be limited to ten days.

5. The applicability of currently available predictive microbiological models in safety evaluation of fishery products with respect to C. botulinum type E was found to be highly questionable. The models tested were hampered by limitations of the controlling environmental factors set by the programs which had an adverse effect on the reliability of predictions.

6. The use of nitrite or nitrate in vacuum-packaged cold-smoked rainbow trout was found to reduce the hazard arising from C. botulinum type E at slightly abusive storage temperatures. Additionally, an extension in the microbiological and sensorial shelf life of the product was observed with both preservatives.