Zsuzsanna Horváth

Characterization of distribution of pesticide residues in crop units.

The characteristic features of distribution of pesticide residues in crop units and single sample increments were studied based on more than 19,000 residue concentrations measured in root vegetables, leafy vegetables, small-, medium- and large-size fruits representing 20 different crops and 46 pesticides. Log-normal, gamma and Weibull distributions were found to provide the best fit for the relative frequency distributions of individual residue data sets. The overall best fit was provided by lognormal distribution. The relative standard deviation of residues (CV) in various crops ranged from 15–170%. The 100–120 residue values being in one data set was too small to identify potential effects of various factors such as the chemical and physical properties of pesticides and the nature of crops. Therefore, the average of CV values, obtained from individual data sets, were calculated and considered to be the best estimate for the likely variability of unit crop residues for treated field (CV = 0.8) and market samples (CV = 1.1), respectively. The larger variation of residues in market samples was attributed to the potential mixing of lots and varying proportion of non-detects. The expectable average variability of residues in composited samples can be calculated from the typical values taking into account the sample size.

Estimation of sampling uncertainty for pesticide residues in root vegetable crops.

The sampling uncertainty for pesticide residues in carrots, parsley leaves and selected medium size crops was estimated with simple random sampling by applying range statistics. The primary samples taken from treated fields consisted of individual carrots or a handful of parsley leaves. The samples were analysed with QUEChERs extraction method and LCMS/MS detection with practical LOQ of 0.001 mg/kg. The results indicate that the average sampling uncertainties estimated with simple random sampling and range statistics were practically the same. The confidence interval for the estimated sampling uncertainty decreased with the number of replicate samples taken from one lot and the number of lots sampled. The estimated relative ranges of sampling uncertainty are independent from the relative standard deviation of the primary samples. Consequently the conclusions drawn from these experiments are generally applicable. There is no optimum for sample size and number of lots to be tested for estimation of sampling uncertainty. Taking a minimum of 6 replicate samples from at least 8–12 lots is recommended to obtain a relative 95% range of sampling uncertainty within 50%. The cost of sampling/analyses, the consequences of wrong decision should also be taken into account when a sampling plan is prepared.

Limitations in the determination of maximum residue limits and highest residues of pesticides: Part I

The pesticide usages are controlled by comparing residue concentrations in treated commodities to legally permitted maximum levels (MRLs) determined based on supervised trials designed to reflect likely maximum residues occurring in practice following authorised use. The number of trials available may significantly affect the accuracy of estimated maximum residues. We conducted a study with synthetic lognormal distributions with mean of 1 and standard deviations of 0.8 and 1.0, which reflect the residue distributions observed in practice. The likely residues in samples were modelled by drawing random samples of size 3, 5, 10 and 25 from the synthetic populations. The results indicate that the estimations of highest residues (HR), used for calculation of short-term intake, and the MRLs, serving as legal limits, are very uncertain based on 3–5 trials indicated by the calculated HR0.975/HR0.025 and MRL0.975/MRL0.025 ratios of 12 and 9, and 13 and 10, respectively, which question the suitability of such trials for the intended purpose. As the 95% range of HR and MRL rapidly decreases with number of trials, ideally ≥15 but minimum 6–8 trials should be used for estimation of HR and MRL according to the current typical practice of Codex Alimentarius.

Estimation of Sampling Uncertainty of Pesticide Residues Based on Supervised Residue Trial Data

Typical sampling uncertainties were calculated as the average of relative standard deviations (CV) of residues measured in individual crops tested in supervised residue trials and from their pooled variance for crop groups. The relative confidence intervals of the sampling uncertainty for different crops were estimated from the random duplicate composite samples generated with computer modeling from residues in 182 independent primary sample sets, each consisting of 100-320 residue data. The relative 95% confidence intervals were found to be independent from the CV of primary residue data populations; therefore, the calculated values are generally applicable. In view of the potentially serious consequences of underestimated sampling uncertainties, their upper confidence limits are recommended for practical use to verify the compliance of products and for planning statistically based sampling programs. Sampling uncertainties are reported for 24 crop groups and 106 individual crops.

Nature of the field-to-field distribution of pesticide residues

The supervised trial datasets (1950), consisting of a minimum of five residue values and selected by the experts of FAO/WHO Joint Meeting on Pesticide Residues for recommending maximum residue levels between 1997 and 2011, were evaluated to obtain information on the typical spread of residue values in individual datasets. The typical relative standard deviation, CV, of field-to-field variation of pesticide residues was about 80%. The spread of residues in datasets is independent from the chemical structure of pesticides, residue level, pre-harvest interval and number of values in the datasets. The CV ranges within the Codex commodity groups and between groups overlapped and their difference were not statistically significant. The number of residues below the limit of quantification (LOQ) affects the CV at various extents depending on the ratio of LOQ/R mean. The combined uncertainty of the highest residue in a dataset significantly affects the CV of the dataset. The lowest and intermediate ones have less influence. The residues in different fields receiving the same treatment vary within large range: 55%, 72%, 78%, 86% and 89% of the 25,766 residues values were, respectively, within 3, 4, 5, 6 and 7 times the median value of the corresponding dataset.

Analysis of industry-generated data. Part 2: Risk-based sampling plan for efficient self-control of aflatoxin M1 contamination in raw milk

Aflatoxin M1 (AFM1) contamination in 21 969 milk samples taken in Italy during 2005–08 and 2010 provided the basis for designing an early warning self-control plan. Additionally, 4148 AFM1 data points from the mycotoxin crisis (2003–04) represented the worst case. No parametric function provided a good fit for the skewed and scattered AFM1 concentrations. The acceptable reference values, reflecting the combined uncertainty of AFM1 measured in consignments consisting of milk from one to six farms, ranged from 40 to 16.7 ng kg−1, respectively. Asymmetric control charts with these reference values, 40 and 50 ng kg−1 warning and action limits are recommended to assess immediately the distribution of AFM1 concentration in incoming consignments. The moving window method, presented as a worked example including 5 days with five samples/day, enabled verification of compliance of production with the legal limit in 98% of the consignments at a 94% probability level. The sampling plan developed assumes consecutive analyses of samples taken from individual farms, which makes early detection of contamination possible and also immediate corrective actions if the AFM1 concentration in a consignment exceeds the reference value. In the latter case different control plans with increased sampling frequency should be applied depending on the level and frequency of contamination. As aflatoxin B1 increases in feed at about the same time, therefore a coordinated sampling programme performed by the milk processing plants operating in a confined geographic area is more effective and economical then the individual ones. The applicability of the sample size calculation based on binomial theorem and the fast response rate resulting from the recommended sampling plan were verified by taking 1000–10 000 random samples with replacement from the experimental databases representing the normal, moderately and highly contaminated periods. The efficiency of the control plan could be substantially enhanced if the dairy farms used feed with a tolerable level of AFB1.

Factors affecting the quantitative uncertainty of the estimated short-term intake. Part I—Calculation methods

ABSTRACT The calculation of the combined uncertainty of the international estimated short-term intake (IESTI) of ethephon residues in apples is shown as an example. The ethephon residues in apples were reported by the Joint FAO (Food and Agriculture Organization of the United Nations)/WHO (World Health Organization) Meeting on Pesticide Residues (JMPR). The apple consumption data were taken from the IESTI (international short-term intake) calculation template used by the JMPR. The IESTI was calculated with the currently used method (case 2a) and a proposed one recommended by the EFSA (European Food Safety Authority)/RIVM (Dutch National Institute for Public Health) Scientific Workshop co-sponsored by FAO and WHO. In this example, the ratio of IESTIproposed/IESTIcurrent and their combined relative uncertainty are about 2.8, and 1.7, respectively. The larger IESTI and uncertainty obtained with the proposed equation are the consequence of calculation only with the large portion (LP) instead of its combination with unit mass, and the MRL instead of the highest residue (HR). The LP is the major contributor to the combined uncertainty. Both the calculated IESTI and its combined uncertainty depend on the actual food – pesticide residue combination, and should be calculated for each case.