Predicting the Safe Storage of Fresh Fish Under Modified Atmospheres with Respect to Clostridium botulinum Toxigenesis by Modeling Length of the Lag Phase of Growth

Abstract

The toxigenesis of non-proteolytic C. botulinum, types B, E, and F, in a variety of fresh fish stored between 4 and 30°C for up to 60 days was evaluated in a five year study. Muscle homogenates and fillets were inoculated with 104-10−2 spores from multiple strain pools, then modified atmosphere (MA) packaged under vacuum, 100% CO2, or 70% CO2 and 30% N2. Our conclusions are based on all data generated in 927 experiments comprising 18,700 samples. The earliest lag times (LT), i.e. sampling period prior to observation of toxicity, among all experiments conducted at 30, 20, 16, 12, 8, and 4°C, were 0.5, 1, 2, 3, 3, and 18 d, respectively. The LT across experiments was significantly affected by temperature (p<0.00001), inoculum level (p<0.00001), fish type (p<0.0001), spore pool (0.0001), and MA (p<0.02). The LT was not affected by homogenization of the fish muscle. We established a general formula that provides the most conservative model for the prediction of LT. This formula yielded calculated LT's that were very close to observed data and models based on individual experiments. Temperature explained 74.6% of experimental variation in the final multiple linear regression model (r2=0.883). The spore inoculum level accounted for 7.4%, while other factors explained <3% together. The utility of the model is demonstrated by its ability to predict the time before toxigenesis in qualitative studies of C. botulinum toxicity in inoculated fish stored under different MA's, reported in the international literature.