It was one of five labs that incorrectly said a spiked sample was compliant.
Findings come from results of a proficiency test (PT) organised by the European Commission's Joint Research Centre (JRC) with DG SANTE following a request by the Belgian Government in August.
The aim was to assess the competence of Official Control Laboratories (OCLs) and National Reference Laboratories (NRLs) in EU Member States to determine content of the insecticide fipronil in eggs around the regulated Maximum Residue Level (MRL).
Target analytes were fipronil, fipronil sulfone and the sum (fipronil and fipronil sulfone) expressed as fipronil.
Analysis of 100-500 egg samples for lab with unsatisfactory result
Individual laboratory performance was expressed in terms of z scores according to ISO 13528:2015.
Sample A was spiked with Fipronil and Fipronil Sulfone and Sample B was the unaltered starting material with the SUM of Fipronil and Fipronil Sulfone (expressed as Fipronil) far below the MRL.
A total of 94.2% of the participants obtained satisfactory z scores. Only 5% (13 individual results reported by six labs) fell into the unsatisfactory performance range.
It found 80 datasets (out of 87) are rated with satisfactory z scores while the results of four participants for the SUM parameter (F+FS, expressed as fipronil) were non-satisfactory.
Five poor performing labs were advised to conduct a thorough root-cause analysis to identify the reasons of such under performance to avoid future inaccurate results.
JRC said this is of special relevance to labs performing routinely such analyses.
“One participant with a very unsatisfactory result for Fipronil Sulfone analysed 100-500 egg samples since July 2017, which might have resulted in false positive results.”
It is not known which lab provided this result due to the confidential nature of the test.
Chicken eggs were found to contain from 0.0031 to 1.2 mg/kg fipronil. Several hundred farmers, food producers and supermarkets pulled millions of eggs off the shelves.
The European Commission has a MRL for fipronil in eggs and poultry meat of 0.005 mg/kg (or 5μg/kg).
A third of labs ‘heavily involved’ in fipronil analysis
Two well-characterised, homogeneous and stable sets of frozen liquid eggs samples were purchased from a local supermarket and distributed for analysis to evaluate capability to identify non-compliant food commodities.
Eighty-five NRLs and OCLs from 22 EU Member States, Norway, Serbia and Albania participated.
One third of participants were heavily involved during the past couple of months in analysis of fipronil in eggs (from 100 to 1,000 samples).
Fifty-one participants were accredited for determination of fipronil in eggs while 68 were accredited for the analysis of fipronil in food in general.
No influence from the analytical techniques used (GC-MS/MS or LC-MS/MS) could be identified.
“54 of the participants applied a multi-residue method for analysis, while 28 used a method targeted only to Fipronil and metabolites,” said JRC.
“No significant difference in the performance is observed for both classes of methods, neither for both major instrumental measurement techniques (LC-MS/MS with 48 datasets; GC-MS/MS with 14 datasets; and applying both methods with 13 datasets.”
Measurement uncertainty reporting
Laboratories had also to report measurement uncertainties in μg/kg. However, 20 labs reported seemingly erroneous uncertainties (expressed in %) and 14 did not report uncertainties at all.
A total of 43% of participants set their measurement uncertainty to the maximum tolerable uncertainty (of 50%) recommended for regulatory compliance assessment.
Sample A was designed to be non-compliant even at the expanded uncertainty of 50% recommended to be used by regulatory authorities in enforcement decisions.
Fipronil content in Sample A (0.017 mg/kg), resulting from the sum of fipronil and fipronil sulfone (expressed as fipronil), was higher than the maximum residue level (0.005 mg/kg), even at 50% uncertainty level.
Seventy nine participants (93%) classified Sample A as non-compliant; most provided the correct justification for their compliance assessment, while nine provided no justification.
Five participants wrongly declared Sample A to be compliant: Two labs based on the low measurement results reported and three gave no justification for the assessment.
The low content of fipronil sulfone in Sample B (of 1.1mg/kg) triggered the reporting of truncated values ('less than') by many participants.