RESUMO
High levels of mortality in Pacific oysters Crassostrea gigas in the Republic of Ireland were recorded during the summer of 2009. The new variant of Ostreid herpes 1 (OsHV-1 µVar) which first emerged in France in 2008 was identified from affected stocks. Retrospective data was collected from 70 oyster farmers through an interviewer-administered questionnaire to investigate the distribution and determinants of the mortality. Based on farmer recall, data were recorded at the batch level for cumulative mortality during 2009, start dates and duration of the mortality event and the age of animals affected. Observable mortalities were recorded in 109 out of 346 batches at 47 sites; 104 of the 109 batches were located in bays where OsHV-1 µVar had been detected. The records from bays where OsHV-1 µVar had been detected were analysed to characterize the pattern of mortality and potential risk factors. Batch mortality averaged 37% (18-65% quartiles) but showed a bimodal distribution (half the batches had mortality less than 45%). Mortalities started at the end of May and continued until early August, peaking in early July. On average oysters died over a period of 18 days. Mortality varied considerably both between and within bays. Mortality started in recently introduced batches and occurred later in the summer in established oysters, which is consistent with the introduction of an infectious agent. Mortality was significantly lower in adults compared with other age groups, which supports observations from France. Three variables were significantly (P<0.05) associated, in both bivariate screening and a logistic regression, with high batch-level mortality (>40%): oysters (i) introduced as juveniles, (ii) during or since the winter of 2008/9 and (iii) which spent less than 8h out of water (in a tidal cycle) (compared with oysters introduced as adults before the winter of 2008/9 and spending more than 8h out of water). Twenty-one percent of triploid batches experienced "high" (>40%) mortality compared with 10% for diploid batches which was significant (P<0.05) in the initial bivariate screening but not in the final logistic regression model. Future studies should develop improved methods to assess oyster mortality and follow stocks over time to better determine the influence of management and environmental factors on mortality.