Dose-response modelling of paralytic shellfish poisoning (PSP) in humans

A systematic review has been conducted to identify and collect raw data of reported cases of human poisoning associated with the ingestion of shellfish contaminated with PSP toxins. Two scales were developed to classify the... [ view full abstract ]

A systematic review has been conducted to identify and collect raw data of reported cases of human poisoning associated with the ingestion of shellfish contaminated with PSP toxins. Two scales were developed to classify the symptoms, one based on the most severe symptom observed and another based on the calculation of a disability weight for all the symptoms observed. In many cases, key information was missing (such as body weight) and when possible, assumptions were made to include the case. Data from 198 exposed individuals (158 with symptoms, 40 without symptoms) from 15 studies were analyzed but showed no clear dose-response. We decided to focus on 3 key studies for which data were available for exposed individuals, both with and without symptoms. Using a classical Benchmark dose modelling of the frequency of exposed individuals with symptoms according to the exposure, the BMDL_10% was estimated at 1.83 µg STX eq./kg b.w. This value is closed to the lowest-observed-adverse-effect-level (LOAEL) established by the European Food Safety Authority (EFSA, 2009) in the region of 1.5 μg STX eq./kg b.w. But our BMDL should be considered as a crude estimate because it was not possible to model individual exposure data (only mean) and to take into account uncertainty of exposure due to the use of PROAST software. We are currently working on a new modelling based on Bayesian inference to take into account variability both in exposure and in the severity of the symptoms. The results will be presented at the conference.