Scientific opinion on animal exposure to zearalenone in feed materials and compound feed in the Republic of Croatia

The Croatian Agency for Agriculture and Food (HAPIH), through the work of the Centre for Food Safety in preparing scientific opinions on risks in food and feed, provides scientific and technical support to the authorities responsible for food and feed safety in making informed risk‐management decisions. This activity is based on national legislation, among which the Ordinance on Feed Safety and the Official Controls Act are particularly important in the context of preparing this scientific opinion, as well as on alignment with relevant European Union regulations. In this context, the assessment of risks posed by mycotoxins is especially important, with zearalenone (ZEN) standing out as one of the significant contaminants in animal feed. ZEN is an estrogenic mycotoxin produced as a secondary metabolite of moulds of the genus Fusarium, most commonly contaminating cereals and plant‐derived products. Its presence poses a risk to animal health, particularly to pigs, which are the most sensitive species, but also to fattening lambs, broiler chickens, and fattening turkeys. Indirectly, ZEN may also affect the safety of food of animal origin. Given its frequent occurrence and potential consequences, its monitoring and risk assessment are important components of the animal feed safety system. The scientific opinion on ZEN was prepared based on an analysis of national data on feed contamination collected during the period 2020–2024. The data included the results of laboratory analyses of feed samples taken during official controls and analysed at the Croatian Veterinary Institute (HVI), as well as the results of laboratory analyses carried out at HAPIH within the implementation of Measure 14 – Animal Welfare and Intervention 70.06. A total of 815 samples from official controls were analysed, of which 677 were included in the assessment, and 674 samples from HAPIH, of which 521 were included. The reduction in the number of processed samples resulted from the need to harmonise feed categories, exclude insufficiently defined samples, and ensure a minimum number of samples per category. Data analysis revealed significant inconsistencies in feed naming and categorisation, requiring additional processing and manual harmonisation. A relatively high proportion of left‐censored (LC) data, i.e. results below the limit of quantification, was also observed, which further increased the uncertainty of the assessment. The proportion of such data varied – in official controls it reached up to approximately 70%, while in HAPIH data it reached up to 80% in certain categories. The assessment of animal exposure to ZEN was carried out according to the methodology of the European Food Safety Authority (EFSA), combining data on ZEN concentrations in feed, daily feed intake, and body weight of animals. Three approaches (LB, MB, and UB) were applied to account for the uncertainty associated with left‐censored data. The assessment covered fattening piglets, fattening pigs/gilts, broiler chickens, fattening turkeys, and fattening lambs. The results showed that ZEN levels in all analysed samples, both from official controls and HAPIH analyses, were below the guidance values established by legislation. Furthermore, the estimated exposure of animals was significantly lower than the toxicological reference values (NOAEL) established by EFSA. In most cases, exposure ranged from 0% to a maximum of about 20% of the NOAEL, indicating a low risk to animal health during the observed period. A comparison of the two datasets revealed certain differences in estimated exposure. In general, exposure levels were slightly higher in the results from official controls than in HAPIH data, except in the case of fattening lambs, where the opposite trend was observed. However, these differences were not significant in terms of risk assessment. They were likely influenced by factors such as the geographical origin of samples, climatic conditions, and differences in sampling patterns and data collection systems. An important limitation of the assessment is the unequal representation of different feed categories. While basic feed components (e.g. maize, wheat, barley) were well represented, data for compound feeds and certain specific categories were limited. Given that compound feeds often constitute a key component of animal diets, this imbalance may affect the representativeness of the exposure assessment. Despite the estimated low level of risk, the results confirm the continuous presence of ZEN in animal feed, which is consistent with the known influence of climatic and storage conditions on mould growth and mycotoxin production. Therefore, continuous monitoring is necessary to ensure timely detection of potential changes in contamination levels. Given the low exposure to ZEN and its limited transfer to muscle and other tissues, very low or non‐detectable levels are expected in meat; therefore, the potential risk to consumer health is considered negligible. Based on the obtained results, the need to improve the monitoring system was highlighted, including better sampling design, standardisation of feed categorisation, reduction of left‐censored data through the use of more sensitive analytical methods, and inclusion of information on the geographical origin of samples. Strengthening coordination between different data collection systems and further incorporation of actual animal diets into exposure assessment were also recommended. Although current levels of ZEN in animal feed in the Republic of Croatia do not pose a significant risk to animal health, maintaining and further improving monitoring systems remains essential for ensuring a high level of feed safety and protecting both animal and human health in the future.