1. INTRODUCTION

Development of new-generation foods, which are mildly processed, contain few or no preservatives, are packaged in vacuum or modified atmospheres to ensure long shelf life and rely primarily on refrigeration for preservation, has raised concerns of potential increases in botulism risk caused by psychrotrophic nonproteolytic group II Clostridium botulinum (Peck, 1997). In the fish industry, this hazard was recognized as early as in the 1960s, soon after the introduction of vacuum-packaging technology to fishery product manufacturing. Three successive serious C. botulinum type E outbreaks caused by vacuum-packaged hot-smoked fish from the Great Lakes area in the Unites States (Anonymous, 1960; Anonymous, 1963; Anonymous, 1964; Conner et al., 1989) lead to surveys on type E prevalence in aquatic environments both in North-America and Europe. These studies demonstrated that type E was most prevalent at cold and temperate regions of the northern hemisphere, with particularly high contamination levels in the Baltic Sea area (Johannsen, 1963; Cann et al., 1965; Huss et al., 1980). A high prevalence of type E in fish farms was also observed (Huss et al., 1974a; Cann et al., 1975; Tjaber and Håstein, 1975). In the United States attempts were made in 1964 to improve the safety of fishery products when guidelines for the production of smoked fish were established and vacuum-packaging was severely restricted (City of Milwaukee, 1964). Despite wide recognition of the emerging risk, few other actions were taken to improve the safety of fishery products or to further elucidate the epidemiology of type E. In Finland, only one small-scale survey including two fish farms was performed (Ala-Huikku et al., 1977). Interest in the issue gradually decreased, since additional significant outbreaks had not occurred. However, a cluster of four outbreaks (Anonymous, 1991; Öberg, 1994; Anonymous, 1998; Korkeala et al., 1998) associated with vacuum-packaged hot-smoked fish in northern Europe in the 1990s confirmed the emerging botulism risk which appeared to be increasing with wider introduction of various types of minimally processed foods.

In northern Europe, raw or mildly processed fishery products are commonly consumed and widely exported which increases the possibility that botulism outbreaks, like most other foodborne outbreaks, will affect more than one small geographic area or country (Majkowski, 1997). The negative publicity caused by a botulism outbreak can severely damage the economics of the fish industry in a small country such as Finland. Therefore, appropriate measures to control and decrease the risk caused by type E needs to be undertaken. Contamination studies in fish farms and fish manufacturing plants should to be performed to facilitate the introduction and maintenance of the hazard analysis critical control point (HACCP) system. Before this can be accomplished, the basic epidemiology of type E must be investigated to obtain updated data on the prevalence of type E in different fish species and fishery products. Furthermore, the genetic biodiversity of type E strains must be studied in order to evaluate the applicability of DNA-based typing methods in epidemiological studies concerning type E. The diagnostics of type E has traditionally concentrated on the detection of botulinal neurotoxin (BoNT) from food and clinical samples by a mouse lethality assay (Hatheway, 1995). During this past decade, PCR-based techniques have facilitated the detection of C. botulinum without laborious toxin detection (Szabo et al., 1992; Campbell et al., 1993; Fach et al., 1993; Hielm et al., 1996). However, very few efforts have been made to develop modern research tools, which could be used to type botulinal strains to the subspecies level (Lin and Johnson, 1995; Hielm et al., 1998a).

Development of new types of products and even minor changes in formulation of existing products warrants a fresh look into their microbiological safety. Traditionally, the risk of the growth of pathogenic microorganisms and possible toxin production in foods has been determined through the use of inoculated pack studies (Roberts, 1997). Now, however, there are too many products, alternative ingredients, and process variations to conduct a complete laboratory evaluation of each possible contingency and potential foodborne pathogen for each product. Therefore, predictive food microbiology, the modeling of microbial populations, particularly those of foodborne pathogens, has become an active field of research. However, before predictive models can be considered reliable, they must be thoroughly validated by conducting challenge tests using various types of products (Whiting and Buchanan, 1994). With respect to C. botulinum type E in fishery products, validation data are limited. Additionally, the toxin production mechanism of C. botulinum in foods remains unclear due to the laborious nature of the conventional enumeration procedure, which employs the combination of either most probable number (MPN) method or conventional plating and the mouse lethality assay (Doyle, 1991; Austin et al., 1998). Moreover, there is a need to evaluate the use of additional hurdles to preserve fishery products and improve their safety.