ABSTRACT:Since 1999, large-scale mortalities of fish-eating birds have been observed on the Great Lakes, and more specifically on Lake Erie. Type E botulism has been established as the primary cause of death. The mechanism of type E botulism exposure in fish-eating birds is unclear. Given that these birds are thought to eat live fish exclusively, it seems likely that their prey play a key role in the process, but the role of fish as potential transport vectors of botulinum neurotoxin type E (BoNT/E) to birds has not been adequately investigated. Between June 2003 and April 2004 a methodological model for exposing fish to Clostridium botulinum was developed and used to compare the sensitivity of rainbow trout (Oncorhynchus mykiss), round goby (Neogobius melanostomas), walleye (Stizostedion vitreum), and yellow perch (Perca flavescens) to four doses (0, 800, 1,500, and 4,000 Mouse Lethal Doses) of Clostridium botulinum type E neurotoxin. Each fish species expressed unique changes in both behavior and skin pigmentation prior to death. Yellow perch survived significantly longer (P,0.05) than the three other species at all toxin treatments. Results of this study suggest that live fish can represent a significant vector for transfer of BoNT/E to birds.

In a 4-mo study (June 2004-September 2004), round gobies (Neogobius melanostomas) were dosed orally every 72 hr for up to 21 days with Clostridium botulinum neurotoxin type E (BoNT/E) at one of four doses: 0, 50, 250, and 500 mouse lethal doses (MLD). Fish were observed for changes in pigmentation and behavior for the duration of the experiment. Mortality was observed with all treatments, with the exception of the 0 MLD control. Clinical signs observed were consistent with prior research and appeared to occur in a threshold manner. The mean times to death and percent mortalities were dose dependent. Hazard ratios were determined to have a significant positive (parameter estimate = 0.03) linear relationship with dose. The hazard ratio showed that per one unit dose increase, the instantaneous probability of a fish dying increased 1.02%. Postmortem analysis of experimental fish demonstrated that 11% (3/27) of fish contained detectable BoNT/E in their visceral fraction. The other 89% tested negative for BoNT/E, despite the fact that all fish died as a result of BoNT/E exposure. Therefore, botulism should not necessarily be ruled out as the cause of a fish kill, even if the fish test negative for BoNT/E.

Rainbow trout (Oncorhynchus mykiss), round gobies (Neogobius melanostomas), yellow walleye (Stizostedion vitreum), and yellow perch (Perca flavescens) were given Clostridium botulinum neurotoxin type E (BoNT/E) at four doses (0, 800, 1,500, and 4,000 mouse lethal doses). BoNT/E was sought in the fish tissues at death or at the conclusion of the experiment (10 days after treatment). Fish were divided into a ''fillet'' (axial musculature) and a ''nonfillet'' sample before testing for BoNT/E toxicity with a mouse bioassay. BoNT/E was detected in all species. The percentage of positive BoNT samples ranged across the species and doses from 0 (trout, perch, and walleye) to 17% (round goby) in fillet tissues and from 0 (perch) to 92% (round goby) in nonfillet tissues. The lack of positive fillet samples in three key commercial fish species suggests that the public health implications of eating these fish are minimal. However, the presence of toxin in the nonfillet compartment of a high proportion of fish supports the hypothesis that live intoxicated fish are a vehicle for the transfer of BoNT/E to fisheating birds, which are then in turn, intoxicated.