Modeling the time-lag effect of sea surface temperatures on ciguatera poisoning in the South Pacific: Implications for surveillance and response. (30th July 2020)

• Record Type: Journal Article
• Authors: Zheng, Lingfeng; Gatti, Clémence Mahana iti; Garrido Gamarro, Esther; Suzuki, Aya; Teah, Heng Yi
• Is Part Of: Toxicon. Volume 182(2020)
• Page Start: 21
• Page End: 29
• Publication Date: 2020-07-30
• Subjects: Toxins -- Periodicals; Venom -- Periodicals; 615.9
• DOI: 10.1016/j.toxicon.2020.05.001
• Languages: English
• ISSNs: 0041-0101

Abstract

Abstract: Ciguatera poisoning (CP), arising from ciguatoxins produced by toxic dinoflagellate Gambierdiscus, is one of the most common food-borne diseases in the South Pacific. Climate change as well as its related events have been hypothesized to a higher abundance and wider presence of toxic dinoflagellates, hence a higher risk of the disease. Yet existing studies assessing the relationship between climate factors and CP are limited or based on old data. In this study, we used prewhitened cross-correlation analysis and auto-regressive integrated moving-average (ARIMA) modeling to develop predictive models of monthly CP incidence in Cook Islands and French Polynesia, two ciguatera-endemic regions in the South Pacific, utilizing the latest epidemiological data. Results reveal the significant time-lagged associations between the monthly CP incidence rate and several indicators relating to sea surface temperature (SST). In particular, SST anomaly is proven to be a strong positive predictor of an increased ciguatera incidence for both countries. If these time-lags can be supported by more investigations, it will allow health authorities to take appropriate actions, to limit or avoid an epidemic risk, especially on high-risk climate scenarios. Highlights: Lags between sea surface temperatures and ciguatera incidence rate were estimated. Predictions made by best-fit models were used to compare with the actual values. Sea surface temperature anomaly is the most significant variable for prediction. More investigations are recommended to confirm and explain the lagged correlations.