Following a request from the European Commission, EFSA assessed the scientific information on some neonicotinoids (i.e. thiamethoxam, clothianidin and imidacloprid) and fipronil which the Italian authorities gathered with a project named “APENET”. APENET is a multidisciplinary monitoring and research project, mainly aimed at evaluating the bee health status, the dust dispersal during the sowing of treated maize seeds and the lethal and sub-lethal effects on bees exposed to this dust; potential synergism between some neonicotinoids and bee pathology was also considered. This project was funded by the Italian authorities following the losses of honeybee colonies reported in Italy since spring 2008. The project was launched after the temporary suspension in Italy, as precautionary measure, of the placing on the market of maize seeds treated with plant protection products containing thiamethoxam, clothianidin, imidacloprid or fipronil.
The research undertaken within APENET was described and reported in 3 different reports: APENET 2009, 2010 and 2011. However, the most comprehensive report was the APENET 2011, which was brought to the attention of the European Commission. Therefore, as requested by the European Commission, EFSA performed an in-depth evaluation of the latter and considered the other reports as background documents.
To evaluate the bee health status, within APENET a monitoring network was set up. Hives situated in different geographic areas were monitored by means of periodic sampling and laboratory analysis of different matrices like dead bees, live bees, brood, honey, wax and pollen. Both pathogens and chemicals were analysed. In relation to the monitoring of pathogens EFSA identified gaps in terms of the data provided (e.g. the total number of stations, apiaries and hives). The reasons behind the sampling plan chosen and the conclusions drawn were presented without describing the levels of representativeness and uncertainty of the estimates obtained. Some important pathogens have not been included in the sampling plan. In relation to the chemicals analyses, the concentrations found are reported as range and not per active substance. An appropriate analysis of the results was also difficult due to the lack of environmental characteristics such as the agricultural landscape around the sampling points and the weather conditions.
Several trials were performed to measure the dust dispersal of thiamethoxam, clothianidin, imidacloprid and fipronil. These trials were conducted with a precision pneumatic seeder machine equipped with deflectors further modified (i.e. Prototype 1 and Prototype 2) to investigate the dust drift reduction. EFSA noted some deficiencies on the reporting of the results. EFSA concluded that a detailed analysis of these results could not be performed, but some general trends could be observed. In particular, the dust and therefore the deposition of residues in the off-crop area decreased with the distance; however, no decrease with the distance was apparent in the air concentration. This was attributed by the authors to the very fine fractions of the dust. The reduction in dust deposition at soil level for imidacloprid was 89% for Prototype 1 and 95.4% for Prototype 2. The overall reduction reached for the other active substances for the Prototype 2 deflector system was 74.4% for clothianidin, 88.6% for thiamethoxam and 94.8% for fipronil. The reductions in air concentration of imidacloprid were 53.1% and 72.4% for Prototype 1 and Prototype 2, respectively. The reductions in air concentration were 86% for clothianidin, 90% for thiamethoxam and 96% for fipronil with the Prototype 2.
Field tests to evaluate the effects on bees directly exposed to dust produced during the sowing of coated maize seed were performed. Different trials were carried out by using different protocols: 1) bees inside cages placed at different heights were exposed to dust produced during the maize clothianidin-coated seeds sowing with a “modified” machine (Prototype 2); 2) free flying bees were exposed to dust produced during the sowing of maize coated seeds with thiamethoxam, clothianidin, imidacloprid and fipronil; 3) bees inside mobile cages were exposed to dust during the sowing of maize coated seeds with thiamethoxam, clothianidin and imidacloprid. For trials under points 2) and 3) an unmodified seeder machine was used for sowing.
As regards the tests with bees inside cages and the Prototype 2, EFSA noted that in each trial only a small number of bees was used and the number of repetitions was very small, therefore the data available on mortality were considered insufficient for a robust statistical assessment, although some data analysis was performed by the authors. Furthermore, some mortality occurred also in the untreated control group, which may have added uncertainty to the data analysis. Due to such deficiencies, it was difficult to make any firm conclusion. Although the authors concluded that the mortality rates were higher in all treated groups than in the untreated control, they also argued that the results could not be generalised and represented worst case exposure conditions for flying bees with respect to sowing line and wind direction.
As regards the tests on free flying bees with unmodified machine, it was concluded that forager bees are at risk if they fly through dust clouds emitted by seeders sowing maize seeds coated with either fipronil, thiamethoxam, clothianidin or imidacloprid (in these trials no deflector system was used). The elevated air humidity increased the mortality rate of the bees that had been exposed to dust containing fipronil, thiamethoxam, clothianidin or imidacloprid. EFSA considered these conclusions supported by the data provided.
As regards the tests with bees inside mobile cages with unmodified machine it might be concluded that the level of dust emission is in the range where detrimental effects on bees cannot be excluded. However, it was noted that the exposure in these trials was unrealistically high.
Effects of contaminated dust were reported on the short and long-term learning and olfactory memory abilities of bees for all the four molecules tested at concentration levels found with both unmodified and modified machines located at 5 m from bees. With the use of modified machines emitting between 80 and 90% less dust than unmodified machines, it was unclear why the authors considered a worst-case exposure for clothianidin alone (i.e. 20% of contaminants in dust versus 10% of contaminants in dust with thiamethoxam, imidacloprid and fipronil) which might have under-estimated the effect of dust exposure with the three other tested molecules. However, the observed effects on bees could not be validated because, from the information provided, it was not possible to guarantee that the protocol was developed in fully controlled conditions and with appropriate statistical testing.
The sub-lethal effects observed and measured on bees in studies dealing with orientation and homing behaviour were regarded more as proposals for implementing the current protocols rather than definitive testing studies because of the small number of bees tested (e.g. study with the simple labyrinth), the exploratory nature of the study design (e.g. study with different exposure scenarios for testing homing and foraging in the field and study on orientation in a complex labyrinth) or simply because the authors mentioned that the study was still ongoing. The incompleteness of the description of these studies and their results did not allow a proper assessment of the methodology and data presented. Nonetheless, the proposed protocols were found innovative and interesting because they attempted to take into account the variability found in the environment of bees (e.g. different exposure scenarios, different field conditions). Such studies warrant further development and fine-tuning for the testing of the effects of pesticides on bee behaviour in field conditions.
The study of the interaction between DWV prevalence in bees and clothianidin exposure was in line with some recent findings showing the potential interaction and/or synergy between various factors involved in bee health (e.g. Varroa and DWV or Nosema and pesticides). The authors claimed that drs-GFP expression was significantly reduced in Drosophila larvae when exposed to clothianidin at LD50 concentrations. Nonetheless the insufficient quality of the reporting did not allow an appraisal of the methodology used and the conclusions drawn by the authors could not be supported. In addition the mechanisms investigated in Drosophila still need to be demonstrated in honeybees. The underlying mechanisms involved in the interaction between pesticides and infection level in bees would merit further investigation.
Overall, it was not possible to draw a firm conclusion on all the scientific information in the APENET report, due to some deficiencies in the study designs and weakness in the statistical analysis and conclusions drawn as reported, or due to the incompleteness in the reporting of the results. However, within this project some potential concerns such as lethal effects on bees exposed to dust, sub-lethal effects and interactions between clothianidin and pathogens were identified suggesting that a change in the assessment of the substances thiamethoxam, clothianidin, imidacloprid and fipronil as regards their effects on bees might be required.
EFSA recently received a mandate from the European Commission for scientific and technical assistance and was requested to provide an EFSA conclusion with an updated risk assessment to bees for the neonicotinoids thiamethoxam, clothianidin, imidacloprid, acetamiprid and thiacloprid. The results for the neonicotinoids investigated in the APENET project might be re-considered, within this mandate, provided the identified deficiencies of the reports will be addressed. For this purpose the papers mentioned in the report and in the process of being published, might be useful. However, since fipronil does not belong to the neonicotinoids, it will not be considered in the new mandate.