*2.1. Method Set Up and In-House Validation*

The aim of this work was to set up and validate a fit-for-purpose method for the routine monitoring of EA in official control. Since the CEN standard [16] was not yet available at the time of the study, a new method was set up, starting from the procedure developed by Kokkonen et al. 2010 (https://doi.org/10.1002/jssc.201000114, accessed on November 2021). The primary improvements adopted to make the method suitable for routine analysis were a shorter extraction time (shaker time of 30 min vs. 60 min) and the use of a calibration curve, in the mobile phase, instead of a matrix-assisted calibration curve for quantification. This last point was very important for official controls, considering that different types of food products are generally analysed in the same batch.

Prior to the validation study, the chromatographic separation of target EAs, was optimized. Special attention was paid to EKR and EKRI, which have been shown to be particularly challenging under conventional reverse-phase chromatographic conditions, leading to chromatographically unresolved double peaks for both compounds, corresponding to the α- and β-forms [14].

Within this study, two different reverse-phase columns were selected and tested to improve EKR and EKRI separation: a Kinetex EVO C18 (100 × 2.1 mm, i.d. 2.6 μm) and an Acquity UPLC BEH C-18 (150 × 2.1 mm, i.d. 1.7 μm). Complete separation of 12 EAs was achieved using column Acquity UPLC BEH C-18 as reported in Figure 2.

Both columns were able to separate α and β isomers of EKR, while for EKRI, the Kinetex column did not provide any separation. For this reason, the Acquity UPLC BEH C-18 column was chosen. The EKR and EKRI results are shown in Figure 3.

Although the separation of the a- and β-isomers, co-occurring in real samples, would be desirable, a joint quantification (estimating the sum expressed as α isomer) might still be acceptable, in routine monitoring, considering the lack of available reference standards for the β forms.

**Figure 2.** LC-MS/MS chromatogram of (**a**) wheat sample spiked with 2.5 μg/kg of each EA (EKR and EKRI alpha isomer only) and (**b**) naturally contaminated barley sample with EM (127 μg/kg), EMI (50 μg/kg), ESI (197 μg/kg), ET (858 μg/kg), ETI (209 μg/kg), EC (266 μg/kg), ECI (141 μg/kg), sum of α + β EKR (262 μg/kg), sum of α + β EKRI (119 μg/kg), ECR (459 μg/kg) and ECRI (161 μg/kg).

**Figure 3.** Extracted ion chromatograms (quantifier and qualifier transitions) for EKR and EKRI in standard solution (0.04 μg/mL) (**a**) and wheat sample naturally contaminated with EKR (mixture of α-EKR and β-EKR) (351 μg/kg) and EKRI (141 μg/kg) analysed for proficiency test using an Aquity BEH C-18 column (**b**).

Given that EAs are more likely occurring in cereals and relevant derived products, the in-house method's performance was evaluated with wheat at concentrations as low as possible (e.g., the estimated LOQ) and at higher levels, taking into account available occurrence data.

Data obtained from in-house validation with wheat are summarized in Table 2.

**Table 2.** In-house analytical performances of the LC-MS/MS method for EAs, including spiking levels, limits of detection (LOD) and quantitation (LOQ), average recovery %, repeatability (RSDr) and within-laboratory reproducibility (RSDWLR). <sup>a</sup> Spiking levels were set at LOQ (2.5 μg/kg), 2xLOQ (5 μg/kg) and 4xLOQ (10 μg/kg).


Abbreviation: ergometrine (EM), ergometrinine (EMI), ergosine (ES), ergosinine (ESI), ergotamine (ET), ergotaminine (ETI), ergocornine (EC), ergocorninine (ECI), α isomers of ergocryptine (α EKR), α isomers of ergocryptinine (α EKRI), ergocristine (ECR) and ergocristinine (ECRI).

Taking into account that no acceptability criteria for linearity were set in the EU legislation regarding performance criteria for mycotoxins analysis, the authors used residuals to evaluate linearity, and the criteria was met for all 12 compounds [17].

The estimated LOQs (see Section 4.6.1) ranged from 0.6 to 2.3 μg/kg for each compound and were compliant with CEN TR 16059 criteria. According to this guideline, when the legal maximum limit (ML) is set for a sum, the LOQs suitable for enforcement of the legal limit shall be equal to or less than ML divided by 2n (where n is the number of

compounds involved). Therefore, the desired values for the monitoring of wheat-milling products and bran-milling products/grain for human consumption (other than rye) shall be set at 4.0 and 6.0 μg/kg for each ergot compound respectively. The values calculated according to the Guide are above the LOQs obtained for the present method.

Then, taking into account the experimentally determined LOQs values, the lowest validation level was set at 2.5 μg/kg for each individual toxin, whereas the others were set at 2 and 5 × LOQ. Mean recoveries ranged from 87 to 119%, whereas repeatability (RSDr) and within-laboratory reproducibility (RSDWLR) were lower than 13% and 15%, respectively (Table 2). Overall, very satisfactory performances were obtained for the proposed method.

A further confirmation of the reliability of the results obtained with the validated method should be sought in the positive outcome of the participation in four Proficiency Tests (FAPAS 22158, Rye Flour 2019, individual EA range 3–65 μg/kg; Bipea 99-1 Barley 2020, individual EA range 50–858 μg/kg; Bipea 99-2 Rye 2020, individual EA range 18–329 μg/kg; Bipea 99-3 Wheat 2020, individual EA range 76–1030 μg/kg).

Acceptable z-score values (−2 ≤ z ≤ 2) were obtained in all PTs (for a total of *n* = 49 provided results), even in cases where the values of the individual molecules were close to or even slightly lower than the estimated method LOQs.

#### *2.2. Applicability of the New Method for Official Control Purposes*

To provide evidence of the applicability and fitness for purpose of the presented method for official controls, data generated within the Italian national monitoring program on the period 2017–2020 are reported herein. Occurrence data for EAs are summarized in Table 3, whereas individual data for each toxin in all analysed samples are provided as Supplementary Material (Table S1).

**Table 3.** Concentration of EAs in cereal grains and cereal products (67 samples analysed). <sup>a</sup> Values calculated on positive samples. LC (left-censored data).


Abbreviations: ergometrine (EM), ergometrinine (EMI), ergosine (ES), ergosinine (ESI), ergotamine (ET), ergotaminine (ETI), ergocornine (EC), ergocorninine (ECI), ergocryptine (EKR), ergocryptinine (EKRI), ergocristine (ECR) and ergocristinine (ECRI).

The analysis of the EAs returned a high percentage of left-censored data (>86%). EM was the most abundant compound, followed by ECR and ES. The individual highest concentration was detected for ECR at 94.2 μg/kg in wheat bran. One sample only (wheat bran) contained all 12 EAs, with a sum of EAs of 271 μg/kg, which could be labelled as non-compliant under the new EU ML of 150 μg/kg [10]. All the other 16 positive samples were compliant, according to the relevant EU ML.

The data in Table 3 were then compared with previously generated ones. The most recent occurrence data for food samples, available in EFSA reports, cover the period 2011 and 2016 and show the highest average contributors to the total concentration to be ET (18%), ECR (15%) ES (12%) and EM (11%) [4]. Considering the large amount of left-censored data, present in the EFSA data set (86%), to minimize the impact of presence of relatively high LODs/LOQs on the UB (upper bound) scenario, a value of 20 μg/kg was selected as a LOQ cut off for each individual EAs, permitting the exclusion of those samples analysed by methods with poor sensitivity but without excessively compromising the number of available samples. In this respect, the LOQs of the method validated and applied in this manuscript were around 10 times lower than the above cut off level (Table 1) and, for this reason, can be considered fit for the purpose of an accurate occurrence evaluation.

The method was also applied to detect the presence of 12 EAs in cereal products for infants. The method did not report any particular issue; therefore, a future validation in cereal products for infants could be demonstrate its suitability for these product categories.

Available literature data on EAs occurrence in food samples, collected in the period 2015–2021, are summarized in Table 4. Results presented in this work are globally in line with previous studies. EM was also reported as the most common EA in wheat sample from Italy by Debegnac et al. [18], moreover, ECR was predominant in cereal samples from Luxembourg [19] and in French cereals [20].

The literature data provide a limited picture of EAs distribution worldwide. This could be partly attributed to the analytical challenges to be undertaken in analysing EAs. Therefore, the availability of isotopically internal standards could improve the accuracy of quantification. Moreover, the difficult chromatographic separation of alpha and beta EKR and EKRI isomers [14], the carefulness needed in samples and standard management (to avoid the epimerization of EAs during sample treatment) [13,21], make the analysis of EAs very tricky. The highest EA incidence is reported for rye and rye-based products, whereas an incidence lower than 10% was observed for other cereals and derived products, and, therefore, comparable to the data presented herein.


**Table 4.** Overview of representative studies on the occurrence of EAs in food samples collected worldwide over the period 2015–2021. The selected studies are relevant to data obtained from sets of more than 15 samples.

N: number of analysed samples, <sup>a</sup> Values calculated on positive samples. NA: Not available in the publication.

Overall, this comparison demonstrated that general applicability of the proposed method and, specifically, that (i) the ranges selected for method's validation encompassed the natural contamination of EAs, not only in Italy, but also in other countries; and (ii) method quantification limits are also suitable to assess EAs contamination in samples for other countries.

The proposed method was suitable to monitor the natural occurrence of EAs in grain and cereal and derived products. Although the method was not validated on cereal products intended for infant consumption, it was applied to the analysis of nine cerealbased food for infants. From the results obtained, the method seems compliant, however further efforts are needed to lower the LOQ.

#### **3. Conclusions**

A fit-for-purpose LC-MS/MS method has been developed and validated for the determination of EAs in official control. The method's performances were proven to be suitable in assessing the compliance of cereals and cereal-based foods with the recently issued EU maximum permitted levels (Commission Regulation 2021/1399/EU). Furthermore, the method's applicability was evaluated by implementing it for EAs analysis in the national monitoring program, which included a total of 67 cereal-based samples collected in Italy over the period 2017–2020. Both the generated data and a comparison with previously reported occurrence data indicate that the method's performances, in terms of precision, accuracy, applicability range and quantification limits, are suitable for assessing EAs natural contamination of cereals and derived products.

#### **4. Materials and Methods**

#### *4.1. Chemicals and Reagents*

EAs were obtained from Romer Labs (Tulln, Austria). Acetic acid and ammonium carbonate were purchased from Honeywell (Wunstorferstrasse, Germany). Ethyl acetate (EtOAc), Methanol (MeOH) and Acetonitrile (ACN) were obtained from Carlo Erba reagent Srl (Milan, Italy). All solvents used were of LC–MS or analytical grade. Water was purified by a Milli-Q system (Millipore, Merck KgaA, Darmstadt, Germany). The MycoSep 150 Ergot columns were purchased from Romer Labs.

#### *4.2. Samples*

Sixty-seven official samples were collected in the period between 2017 and 2020 from three six Italian Regions (Umbria, Marche and Puglia) and analyzed by Istituto Zooprofilattico Sperimentale of Umbria and Marche "Togo Rosati". The samples consisted of 18 cereal grains, 16 flours (14 of wheat and 2 of spelt) and 33 other types of cereals derivatives (including 9 for infants) respectively.

Samples were ground by a knife mill (GRINDOMIX GM 300, Restek, Haan, Germany) with dry ice and split in aliquots of 25 g for the analysis. The samples were stored at −20 ◦C until analysis.
