Predictive score and probability of CTX-like toxicity in fish samples from the official control of ciguatera in the Canary Islands

This research identifies factors associated with the contamination by ciguatoxins (CTXs) in a population of fish and proposes a predictive score of the presence of CTX-like toxicity in amberjack samples from the official control program of ciguatera in the Canary Islands of the Directorate-General (DG) Fisheries (Canary Government). Out of the 970 samples of fish studied, 177 (18.2%) samples showed CTX-like toxicity. The fish were classified according to the species, amberjack (Seriola dumerili and S. rivoliana) (n = 793), dusky grouper (Epinephelus marginatus) (n = 145) and wahoo (Acanthocybium solandri) (n = 32). The data were separated by species category and statistically examined, resulting in 137 (17.3%) amberjack and 39 (26.9%) grouper samples showing CTX-like toxicity; regarding wahoo species, only 1 toxic sample (3.1%) was found. According to fishing location the contamination rates suggested grouping the islands in four clusters; namely: {El Hierro: HI; La Gomera: LG; La Palma: LP}, {Gran Canaria: GC; Tenerife: TF}, {Fuerteventura: FU} and {Lanzarote: LZ}. For the amberjack species, the multivariate logistic regression showed the factors that maintained independent association with the outcome, which were the warm season (OR = 3.617; 95% CI = 1.249-10.474), the weight (per kg, 1.102; 95% CI = 1.069-1.136) and the island of fish catching. A prediction score was obtained for the probability of contamination by CTX in amberjack fish samples. The area under de curve (AUC) obtained using the validation data was 0.747 (95% CI = 0.662-0.833). Regarding grouper species, the island of fishing was the only factor that showed significant differences associated with the presence of CTX-like toxicity. We provide herein data for a better management and prediction of ciguatera in the Canary Islands, suggesting a review of the minimum limits of fish weight established by the Canary Government for the control program.