**Background:** Pesticide residues resulting from the use of plant protection products on crops or food products can potentially pose a risk to public health. The European Union has established a comprehensive legislative framework defining rules for the approval of active substances, their uses, and permissible residues in food. Under Regulation (EC) No 396/2005, Member States carry out official controls to ensure food placed on the market is compliant with legal limits, and EFSA is responsible for preparing an Annual Report analysing the data.
**Methods:** The 2021 report covers official control activities carried out in EU Member States, Iceland, and Norway. The EU-coordinated multiannual control programme (EU MACP) randomly sampled 12 food products (aubergines, bananas, broccoli, cultivated fungi, grapefruit, melons, sweet peppers, table grapes, virgin olive oil, wheat, bovine fat, and chicken eggs) — a total of 13,845 samples analysed for 190 pesticides. National control programmes (MANCP) added risk-based sampling, bringing the total to 87,863 samples. Dietary exposure was assessed using the deterministic Pesticide Residues Intake Model (PRIMo rev. 3.1) for acute and chronic risk to single substances. For the first time, a pilot probabilistic exposure assessment was introduced for 29 pesticides that exceeded the acute health-based guidance value (HBGV) in the deterministic assessment, using a two-dimensional method with 100,000 iterations per population group.
**Key Results:** Of the 87,863 total samples, 84,399 (96.1%) fell within legal limits. MRLs were exceeded in 3,464 samples (3.9%), decreasing from 5.1% in 2020. When accounting for measurement uncertainty, 2,207 samples (2.5%) triggered legal sanctions. Under the EU MACP, 13,550 samples (97.9%) were within legal limits, with MRLs exceeded in 295 samples (2.1%). The overall MRL exceedance rate rose from 1.4% in 2018 to 2.1% in 2021, but this was distorted by grapefruits from Türkiye — excluding these, the rate would have been 1.4%. Samples from third countries showed a 5-fold higher MRL exceedance rate (10.3%) compared to EU-grown samples (2%). In the acute deterministic assessment of 19,499 samples, 220 samples (1.1%) exceeded HBGVs across 29 different pesticides, most frequently lambda-cyhalothrin (46 samples), cypermethrin (44 samples), acetamiprid (43 samples), and indoxacarb (25 samples). In the chronic deterministic assessment, no consumer intake concern was identified under the lower-bound scenario; under the middle-bound scenario, dimethoate reached 101% of the ADI. The pilot probabilistic assessment revealed that for most of the 29 pesticides (38 active substances), the probability of exceeding the ARfD is close to 0%. For oxamyl, the percentage of toddlers exceeding the ARfD ranged from 6.7% to 17.3% across countries, and for dimethoate, from 18.3% to 27.8%, though these estimates were considered difficult to interpret due to high LOQ values relative to HBGVs.
**Clinical Implications:** For the vast majority of samples analysed, dietary exposure to pesticides for which HBGVs are available is very low and poses no risk to EU consumer health. In rare cases where exposure exceeded HBGVs, competent authorities took corrective measures including withdrawing samples from the market. The introduction of probabilistic methodologies will allow better quantification of real risk in future assessments. Key recommendations include: continued monitoring of specific pesticide/commodity combinations (e.g., chlorpyrifos in grapefruits from Türkiye, ethylene oxide in sesame seeds from India); improving analytical capabilities for single-residue methods; widening analytical scope for organic and imported samples; and developing processing/peeling factors to refine risk assessments.