Analysis of Ergot Alkaloids in Cereal-based Food Products From the US Market using LC-MS/MS

Abstract

Cereal grains such as rye, wheat, oat, and barley are susceptible to ergot fungus that produce toxins known as ergot alkaloids (EAs), posing potential health risks to consumers. Despite growing global concern, occurrence data on EA contamination in cereal-based food products within the US market remains scarce. To address this data gap, a targeted LC-MS/MS method was developed and validated to quantify six predominant EAs and their corresponding epimers across a variety of cereal-based food matrices. Samples were prepared by extraction using a mixture of acetonitrile and water under alkaline conditions and cleanup using a MycoSep 150 Ergot SPE column. This method was single laboratory validated, demonstrating reliable performance with limits of quantification (LOQs) of 1.0 μg/kg for all analytes. The validated method was then applied to the analysis of sixty commercially available cereal-based products collected as a convenience sampling from the US market, revealing that 42% of food products (25 out of 60) analyzed contained at least one detectable EA and seven products contained all 12 targeted EAs. Concentrations of individual EAs ranged between <1.0 and 231 μg/kg, while total EA concentrations ranged from <1.0 to 755 μg/kg in tested products. Among the product categories analyzed, flour and rye-based products showed the highest incidence and concentrations of total EAs.

Introduction

Ergot alkaloids (EAs) are mycotoxins produced by several fungi species in the genus Claviceps, which infect various grasses and small grain cereals such as wheat, barley, oats, and especially rye (Mavungu et al. 2011, Agriopoulou 2021). During fungal infection, the healthy kernels on grain ears are replaced with dark colored crescent-shaped structures, known as ergots or ergot sclerotia, containing toxic alkaloids (EFSA 2012, Silva et al. 2023). Harvesting grains with sclerotia can lead to the contamination of cereal grains and consequently, cereal-based food products with EAs. Consumption of these contaminated products can cause harmful health effects in humans, resulting from both acute and chronic exposure (EFSA 2012, Silva et al. 2023).

While improvements in grain sorting and cleaning techniques have largely reduced sclerotia content in cereals, EAs can still be found in cereal grains, even in the absence of visible sclerotia (EFSA et al. 2017). Monitoring of ergot contamination solely based on the presence of sclerotia can be less reliable as the size, weight, and composition of the sclerotia may vary greatly (Mulder et al. 2012). Additionally, this approach is not applicable for analyzing processed food products. Therefore, the development of analytical methods specifically targeting EAs is essential to obtain accurate occurrence data, thereby ensuring reliable dietary exposure and risk assessments.

More than 50 different EAs have been identified in cereals infected with Claviceps spp., with varying amounts and profiles of EAs depending on fungal strains, host plants, geographical locations, and the climatic conditions (EFSA 2012, Chung 2021). Common EAs share a tetracyclic ergoline ring structure and those containing a double bond between C9 and C10 can undergo epimerization, resulting in two epimers, R-epimer (suffix: -ine) and S-epimer (suffix: -inine) (Crews 2015). The most predominant EAs and their corresponding epimers are ergometrine (Em) / ergometrinine (Emn), ergosine (Es) / ergosinine (Esn), ergotamine (Et) / ergotaminine (Etn), ergocornine (Eco) / ergocorninine (Econ), ergocryptine (Ekr) / ergocryptinine (Ekrn), and ergocristine (Ecr) / ergocristinine (Ecrn) (EFSA 2012, Mulder et al. 2012). While Em and Emn are simple lysergic acid derivatives, other aforementioned EAs belong to the ergopeptines/ergopeptinines class, which contain tripeptide moieties attached to the ergoline ring (Figure S1) (Chung 2021). Differences in bioactivity between R- and S- epimers have been reported. The R-epimers are considered biologically active, while the S-epimers have almost no pharmacological effects and are thus less toxic; however, their rapid interconversion occurs under various conditions including protic solvents, pH, temperature, and UV light (Schummer et al. 2020). It is therefore recommended to conduct analysis on both forms, to avoid the potential underestimation of total EA content.

Over the years, different analytical methods have been developed for the separation and quantification of main EAs and their epimers. Among these, liquid chromatography tandem mass spectrometry (LC-MS/MS) and liquid chromatography high resolution mass spectrometry (LC-HRMS) have emerged as the most widely used techniques for analyzing EAs in food, due to their high selectivity and sensitivity (Di Mavungu et al. 2012, Liao et al. 2015, Guo et al. 2016, Babic et al. 2020, EURL Mycotoxins and Plant Toxins 2020, Huybrechts et al. 2021, Poapolathep et al. 2021, Uhlig et al. 2021, Carbonell-Rozas et al. 2021a, European Standard 2021, García-Juan et al. 2023, Nam et al. 2024). Extensive studies have been conducted to investigate the occurrence of EAs in a range of cereals and cereal-based food products. In a 2017 report, the European Food Safety Authority (EFSA) presented human dietary exposure to EAs based on analytical results of 12 EAs in food samples collected between 2011 and 2016 across 15 different European countries. The highest levels of EAs were reported in rye and rye-containing commodities, with EAs also detected in other cereal grains, including wheat, spelt, oats, corn, and their processed derivatives.

Other studies have also reported the occurrence of EAs from countries including Italy, Albania, Belgium, Spain, Netherlands, Algeria, Canada, and China (Lattanzio et al. 2021, Carbonell-Rozas et al. 2023, Peloso et al. 2024, Topi et al. 2017, Huybrechts et al. 2021, Carbonell-Rozas et al. 2021a, Mulder et al. 2015, Versilovskis et al. 2020, Carbonell-Rozas et al. 2021b, Walkowiak et al. 2022, Guo et al. 2016). These investigations have consistently reported the presence of EAs in rye, with an increasingly higher incidence observed in other cereals such as wheat, barley, and oats over time. This may reflect changes in agricultural practices, climatic conditions favoring fungal growth, or improved detection techniques. Since cereals may be a key component in the diet of infants and young children, several studies have focused on cereal-based foods for this age group (Mulder et al. 2015, Huybrechts et al. 2021). While positive identification of EAs has been reported, the concentrations found in foods commonly consumed by infants and young children are significantly lower compared to those detected in other cereal-based foods.

Given potential health risks associated with EA consumption, various authorities, including Codex, the US, Canada, and the EU, have established maximum levels for ergot sclerotia in cereal grains (Walkowiak et al. 2022). Moreover, Commission Regulation (EU) 2023/915 set maximum levels for the sum of 12 EAs in a range of cereals and cereal-derived products, with the highest limit for rye milling products at 500 μg/kg and the lowest limit for cereal-based food for infants and young children at 20 μg/kg (European Commission 2023).

Currently, regulatory limits for EAs have not been established in the US and information on the prevalence and occurrence of EAs in cereals available in the US market remains limited. Therefore, the aim of this study was to 1) develop and validate a selective and sensitive LC-MS/MS method for the quantitative analysis of predominant EAs including their epimers in a variety of cereal-based food matrices, and 2) evaluate the levels of EAs in a convenience sampling of sixty cereal-based products commercially available in the US marketplace. To ensure the method is suitable for routine use and applicable across different matrices, we incorporated a semi-synthetic ergot alkaloid as a surrogate standard and used wheat flour as a representative matrix for calibration, effectively compensating for matrix effects and variability in extraction and ionization. This strategy enables accurate and precise quantification of EAs across a range of cereal-based matrices, including wheat, rye, barley, oat, and multigrain products, as well as baby food, without requiring matrix-matched calibration curves for each product type. The established workflow supports the collection of analytical data on the occurrence of EAs in commercial food products, contributing to future assessments of dietary exposure and associated potential health risks.

Materials and Methods

Reagents and Materials

Extraction solvents acetonitrile and water were LC/MS Optima grade and purchased from Fisher Scientific (Waltham, MA). Ammonium carbonate and ammonium hydroxide were both purchased from Sigma-Aldrich (St. Louis, MO). The mobile phase additive ammonium bicarbonate was purchased from Acros Organics (Morris Plains, NJ). The Mycosep 150 Ergot column for extraction clean-up was purchased from Romer Labs (Getzersdorf, Austria). Z-Sep+ and C18 sorbents for clean-up were from Supelco (Bellefonte, PA) and Waters (Milford, MA), respectively. 0.2 μm polytetrafluoroethylene (PTFE) filter vials were purchased from Thomson Instrument Company (Oceanside, CA).

Standards

Analytical reference standards of ergometrine (Em), ergosine (Es), ergotamine (Et), ergocornine (Eco), α-ergocryptine (α-Ekr), ergocristine (Ecr), their corresponding epimers ergometrinine (Emn), ergosinine (Esn), ergotaminine (Etn), ergocorninine (Econ), α-ergocryptinine (α-Ekrn), ergocristinine (Ecrn), and the surrogate standard dihydroergocristine (DHEC) were purchased from Romer Labs (Getzersdorf, Austria). Following the manufacturer’s instructions, stock standard solutions were prepared by reconstituting each compound in 5 mL acetonitrile to achieve concentrations of 100 and 25 μg/mL for the R- and S- epimers, respectively. These individual stock standard solutions were then combined and diluted in acetonitrile to generate working solutions for the preparation of the calibration curve. Separately, the surrogate stock solution was reconstituted in acetonitrile at 100 μg/mL and subsequently diluted to prepare working solutions used both for the calibration curve and for sample fortification prior to extraction. All standard solutions were stored in amber glass vials (15 mL, Supelco, Bellefonte, PA) at −20°C in a dark environment.

Samples

Product selection was guided by available occurrence data in the literature. A non-representative sampling of sixty (60) unique products, including single grain [barley (8), oat (11), rye (14), and wheat (21)] and multi-grain (6) foods, were purchased from online retailers in June 2023. Products were divided into 4 categories, including breakfast cereal (14), cracker (15), flour (17) and grain/flake (14). Among these products, 24 products were labeled with an organic statement and 4 products were marketed for infants and young children. Two proficiency test materials were obtained from FAPAS, including rye flour and baby food (a mixture of wholegrain oat flour, wholegrain wheat flour, and organic dark rye flour), both naturally contaminated with EAs.

Sample Preparation

Samples were homogenized prior to sample extraction. For breakfast cereal and cracker products, the entire package was ground using a blender mixer (Robot Coupe, BLIXER 4, Ridgeland, MS) to achieve a fine powder consistency. For flour products, the entire package was initially divided into 10 fractions, and this process was repeated three times using a sample divider (Retsch PT-100, Verder Scientific, Newtown, PA). From these, three fractions were randomly selected, combined, and then re-divided into 10 subsamples. For grain/flake products, the entire package was cryo-ground using a cryogenic grinder (Cole-Parmer, CG-900 Cryo-Blade^™, Vernon Hills, IL) with liquid nitrogen until a fine powder consistency was achieved.

A portion of 4 g homogenized sample was spiked with 40 μL of 1 μg/mL DHEC (surrogate standard) and extracted using 20 mL acetonitrile/200 mg/L ammonium carbonate (85:15, v/v). The mixture was shaken for 5 minutes using a Geno/Grinder 2010 (SPEX CertiPrep, Metuchen, NJ) at 1000 strokes/min, followed by centrifugation (Sorvall legend XTR centrifuge, Thermo Fisher Scientific, Waltham, MA) at 4000 × g for 5 minutes at 10 °C. 3 mL of the supernatant was transferred to a glass tube for clean-up of the extract using a MycoSep 150 Ergot column. The purified extract was diluted (1:1) with 200 mg/L ammonium carbonate in water and transferred into a PTFE filter vial.

Preparation of Calibration Standards and Quality Control Samples

Matrix calibration standards were prepared at concentrations of 0.05, 0.1, 0.5, 1, 5, and 10 ng/mL (equivalent to 0.5, 1, 5, 10, 50, and 100 μg/kg) using blank oat flour extract for oat-based samples and blank wheat flour extract for all other samples. DHEC was added to each calibration standard at 1 ng/mL as a surrogate standard. A calibration check standard at the midpoint of the calibration curve was prepared and analyzed every 10^th injection, to evaluate for carryover or instrument deviations.

A solvent blank sample (acetonitrile/water (50/50, v/v)) was analyzed prior to the start of each batch and following high-concentration samples or calibration standard(s). A method blank sample (acetonitrile/water (50/50, v/v)) was prepared for each sample batch and processed according to the method protocol. For quality control, spike recovery samples were prepared in duplicate by fortifying a representative matrix with target EAs at a final concentration above the method LOQ (e.g. 2–10x LOQ) or significantly different from the native concentration (e.g. 50–100%).

Peak area response in any type of blank samples (e.g., solvent, method, or matrix blank) should be less than 50% that of the lowest calibration standard. The calculated concentration of the calibration check standard should be within 80–120% of expected concentration. Percent recovery for spike recovery samples should be between 60–120% of the known value for acceptable performance. Precision expressed as relative percent difference (RPD), should be < 20%.

LC-MS/MS

An aliquot (1 μL) of each prepared sample extract was injected into a Shimadzu Nexera LC-40 system coupled with a Sciex 6500+ QTRAP mass spectrometer. Chromatographic separation was performed on an ACQUITY UPLC BEH C18 Column (130Å, 1.7 μm, 2.1 mm × 100 mm) with an ACQUITY UPLC BEH C18 VanGuard Pre-column (130Å, 1.7 μm, 2.1 mm × 5 mm) as the guard column (Waters, Milford, MA). Prior to sample analysis, the column was equilibrated according to manufacturer’s instructions and further conditioned with matrix samples to ensure the optimal performance and reproducibility. The mobile phases were comprised of 10 mM ammonium bicarbonate in water, pH 9.0 (A) and acetonitrile (B). A linear gradient was applied starting with 5% B. At 1.0 min, % B was increased to 40%, followed by a further increase to 70% at 11.0 min. At 11.5 min, the composition was returned to 5% B for column re-equilibration, which was held for 6.5 minutes. The total run time was 18 minutes. A diverter valve was used to divert the flow to the mass spectrometer between 2 and 10 minutes of the gradient run. The flowrate was 300 μL/min and the column temperature was 40 °C.

Mass spectrometric analysis was performed in a scheduled multiple-reaction monitoring (MRM) mode with a 45 s detection window and a target cycle time of 1 s. Source parameters were set as follows: curtain gas (CUR) at 40 psi, ion spray voltage at 5500 V, collision gas (CAD) at medium, temperature at 500 °C, and ion source gases 1 and 2 (GS1 and GS2) at 30 and 50 psi, respectively. Two transitions were monitored for each analyte with optimal values for declustering potential (DP), entrance potential (EP), collision energy (CE), and collision cell exit potential (CXP) summarized in Table 1. All experimental MS/MS data were analyzed using Analyst^™, version 1.7 and Sciex OS version 3.0.

Table 1.

Scheduled MRM Transitions and Parameters for a Sciex 6500+ QTRAP Mass Spectrometer^a.

Compound	Retention Time (min)	MRM Transitions	DP (eV)	CE (eV)	CXP (eV)
Em	2.5	326.2 -> 223.1/208.1	86	31/35	16/16
Emn	3.0	326.2 -> 208.1/223.1	96	37/35	14/18
Es	4.3	548.3 -> 223.1/268.1	126	43/33	16/18
Esn	6.8	548.3 -> 223.1/277.1	126	41/35	16/18
Et	4.6	582.3 -> 223.1/208.1	111	45/49	14/14
Etn	7.3	582.3 -> 223.1/277.1	96	41/35	14/20
Eco	5.4	562.3 -> 268.1/223.1	126	33/47	22/14
Econ	8.0	562.3 -> 305.1/277.1	101	37/37	22/18
α-Ekr	6.0	576.3 -> 268.1/223.1	106	35/45	18/16
α-Ekrn	8.8	576.3 -> 305.1/223.1	101	37/43	24/14
Ecr	6.2	610.3 -> 268.1/208.1	111	37/51	18/14
Ecrn	9.2	610.3 -> 305.1/325.1	111	37/37	24/22
DHEC	5.7	612.3 -> 270.1/350.1	116	39/35	18/24

^aEntrance potential (EP) is set as 10 eV for all analytes.

DP: declustering potential; CE: collision energy; CXP: collision cell exit potential.

Quantitation ion.

Results and Discussion

Method Optimization

Optimization of LC-MS Method

The identification of EAs remains challenging due to the presence of epimers which share the same accurate mass and produce indistinguishable fragment ions. Although alternative techniques have been explored (Carbonell-Rozas et al. 2022, Liang et al. 2023, Narváez et al. 2024), the separation of EA epimers is best achieved by reversed-phase LC using C18 columns (Guo et al. 2016, Versilovskis et al. 2020, Carbonell-Rozas et al. 2021b). In this study, the separation of EA epimers was investigated on a C18 column under two gradient mobile phase conditions: (1) methanol with an acidic aqueous solution (see Supplementary Information for method details) and (2) acetonitrile with an alkaline aqueous solution. The use of alkaline pH prevents the protonation of the basic nitrogen in these epimers, leading to increased chromatographic retention and improved separation. This resulted in complete separation of all EA epimers, along with improved peak shape and increased signal intensity (Figure 1). Further optimization of the alkaline mobile phase was performed by varying additive concentrations and pH levels. A mobile phase consisting of 10 mM ammonium bicarbonate, adjusted to pH 9.0, yielded the best overall performance and was selected for the final method (Figure S2). This trend aligns with a recent study, which similarly reported more sensitive signals and better peak shapes under alkaline conditions (Rollo et al. 2025).

Figure 1.

Extracted ion chromatograms of EA standards eluted using different mobile phase conditions (a) methanol with acidic aqueous solution – 0.3% formic acid and 5 mM ammonium formate (b) acetonitrile with alkaline aqueous solution – 10 mM ammonium bicarbonate pH 9.

As reported in previous studies, the separation of α- and β-Ekr/Ekrn presents a significant challenge when using conventional reverse-phase chromatographic techniques. Currently, there are no commercially available analytical standards for the β-forms, further complicating their individual quantification. Here, the separation of the a- and β-epimers was evaluated using a rye reference sample naturally contaminated with both epimers (Figure S3). While successful resolution of α- and β- Ekr was achieved, α- and β- Ekrn could not be separated under the same conditions. Consequently, the epimers were quantified and reported as a sum for α- and β-Ekr and α- and β-Ekrn, which is consistent with other studies in the absence of β-form standards.

Optimization of Sample Preparation

Several extraction methods for EAs have been reported, such as Quick, Easy, Cheap, Effective, Rugged, and Safe (QuEChERS) and extraction using organic solvent mixtures under acidic or alkaline conditions (Krska et al. 2008, Kokkonen and Jestoi 2010, Guo et al. 2016, Arroyo-Manzanares et al. 2018, EURL Mycotoxins and Plant Toxins 2020). In this study, a method was developed based on a modified QuEChERS procedure using an acetonitrile mixture under alkaline conditions. In the absence of available labeled internal standards, an effective clean-up process is essential to reduce matrix effects and ensure accurate quantification. Both solid phase extraction (SPE) columns/cartridges and dispersive SPE sorbents have been used for cleaning EA extracts in various studies (Krska et al. 2008, Di Mavungu et al. 2012, Bryla et al. 2015, Guo et al. 2016, Carbonell-Rozas et al. 2021a). Moreover, a commercial push-through SPE column (MycoSep 150) has also been applied (Kokkonen and Jestoi 2010, Guo et al. 2016). Here, we evaluated two different sorbents (C18 and Z-Sep+) and the MycoSep 150 SPE column for clean-up of wheat flour extracts (Figure 2). Quantification was performed using calibration standards prepared in solvent. Among the clean-up procedures evaluated, the MycoSep 150 SPE column provided the best overall recovery results across all target analytes and was therefore selected for further method development. While C18 and Z-Sep+ sorbents also demonstrated acceptable recoveries for most analytes, C18 showed poor recovery for Ecrn and Z-Sep+ resulted in reduced recovery for Emn.

Figure 2.

Recovery results for wheat flour fortified at 2.5 μg/kg EAs and extracted using different clean-up procedures. Error bars represent deviations between triplicate sample preparations. Dashed lines indicate EA percent recoveries between 80–120%.

Method Validation

Following method development, a single laboratory validation was performed per the Guidelines for the Validation of Chemical Methods for the FDA Foods Program (U.S. FDA 2019). The performance of the method was evaluated, including matrix effects, linearity, determination of method detection limits (MDL), accuracy, and precision, for all 12 monitored EAs in six fortified matrices and two proficiency test (PT) samples. Six cereal-based foods, including grains (barley, oats, rye flour, wheat flour), dry infant cereal (oatmeal), and processed grain (multi-grain breakfast cereal), were evaluated at three different concentrations (2.5, 10, and 40 μg/kg), with each concentration analyzed in triplicate. Target EAs were identified by retention time alignment (within ±5%) and product ion transition confirmation (two unique, structurally specific ions within ±20% relative unit ion ratio tolerance) with calibration standards. The concentration was determined using the peak area ratio of response of each EA’s quantitation transition to that of the surrogate standard (DHEC), adjusted for dilution.

The matrix effects (ME) were evaluated for each of the six food matrices by comparing the signal of each EA fortified in solvent to an equivalent concentration fortified post-extraction in matrix at three different concentrations (Table 2). For most EAs, matrix suppression was lower than 25%. However, significant matrix suppression was observed for Es in rye flour, pearl barley, and steel-cut oats, Et in rye flour and pearl barley, α-Ekrn in baby oatmeal, and Ecrn in all matrices except wheat flour. Therefore, to compensate for matrix suppression impacting the quality of analytical results, calibration standards were prepared using a blank wheat flour extract, which demonstrated the lowest and most consistent ME for all EAs among the matrices evaluated. In contrast, the other matrices showed stronger and analyte-specific suppression. This approach allowed the use of a representative matrix-based calibration that effectively mitigated ME while maintaining calibration applicability across all matrices analyzed. All EA calibration standards prepared in wheat matrix achieved a good linearity over the working range of 0.1 – 100 μg/kg (1/x weighting, R² > 0.99, Table S1). Matrix-specific and compound-specific MDLs were determined following Environmental Protection Agency (EPA) procedure [40 CFR 136, Appendix B (U.S. EPA 2017)] and the mean MDL across all analytes and matrices was 0.22 μg/kg (range: 0.09–0.43 μg/kg) (Table S1). For reporting, the MDL was established as 0.3 μg/kg and the limit of quantification (LOQ) as 1.0 μg/kg for each analyte.

Table 2.

Average %Matrix effects (ME)^a observed for individual EAs in six food matrices (n = 3^b).

Analyte	Wheat Flour	Rye Flour	Pearl Barley	Steel-Cut Oats	Baby Oatmeal	Multi-Grain Flakes
Em	−2	−21	−17	−13	−15	−24
Emn	−7	−7	−10	−5	−3	−3
Es	−12	−30	−30	−27	−21	−23
Esn	−6	−10	−9	−12	−6	−4
Et	−12	−30	−28	−25	−21	−19
Etn	−13	−16	−12	−16	−10	−9
Eco	3	−11	−14	−11	−10	−8
Econ	−13	−19	−21	−20	−11	−11
α-Ekr	1	−11	−13	−17	−10	−5
α-Ekrn	−13	−21	−21	−23	−27	−10
Ecr	1	−14	−11	−15	−11	−9
Ecrn	−22	−45	−43	−53	−54	−28

^a$\mathit{\text{ME}}=\frac{\mathit{\text{signal}}\mathit{\text{of}}\mathit{\text{analyte}}\mathit{\text{in}}\mathit{\text{matrix}}-\mathit{\text{signal}}\mathit{\text{of}}\mathit{\text{analyte}}\mathit{\text{in}}\mathit{\text{solvent}}}{\mathit{\text{signal}}\mathit{\text{of}}\mathit{\text{analyte}}\mathit{\text{in}}\mathit{\text{solvent}}}\times 100\mathit{\text{\%}};$ an ME below 0% indicates signal suppression while an ME above 0% indicates signal enhancement.

^bCalibration standards were prepared at three concentrations (1, 10, 100 μg/kg).

Across all six matrices, the majority of target EAs achieved recoveries between 80–120% with a relative standard deviation (RSD) less than 20% (Table 3). However, lower recoveries of Em and Ecrn were observed in oat-based matrices (steel-cut oats and baby oatmeal), likely due to ME and possibly to analyte loss during sample preparation. Significant ME for Em/Emn and Ecr/Ecrn in oat-based matrices has been previously reported (Carbonell-Rozas et al. 2021a). To enhance quantitative accuracy, recoveries were re-evaluated using a calibration curve prepared in an oat matrix extract, which demonstrated acceptable linearity over the range of 0.5 to 100 μg/kg (1/x weighting, R² > 0.99). When this matrix-matched calibration was applied, recoveries for EAs in both oat-based matrices fell within the range of 87–121% (Table 4). Based on these findings, the validated method recommends oat-based samples be quantified using a calibration curve prepared in an oat flour matrix extract. For all other sample types, calibration using a wheat flour matrix extract is sufficient to ensure reliable quantification.

Table 3.

%Recovery (%RSDr) of EAs fortified in six food matrices at concentrations of 2.5, 10, and 40 μg/kg (n=3).

Analyte	Fortified Conc. (μg/kg)	Wheat Flour	Rye Flour	Pearl Barley	Steel-Cut Oats	Baby Oatmeal	Multi-Grain Flakes
Em	2.5	100 (5)	81 (7)	93 (4)	93 (4)	67 (5)	68 (4)
	10	101 (8)	90 (7)	96 (2)	96 (3)	67 (4)	74 (3)
	40	96 (1)	82 (5)	95 (6)	94 (5)	67 (4)	70 (3)
Emn	2.5	106 (5)	94 (5)	94 (7)	101 (5)	81 (3)	84 (2)
	10	113 (10)	100 (7)	98 (4)	106 (5)	83 (1)	89 (3)
	40	104 (1)	93 (3)	101 (2)	102 (6)	80 (3)	85 (1)
Es	2.5	102 (2)	92 (3)	103 (3)	104 (7)	97 (6)	102 (2)
	10	112 (7)	101 (7)	105 (2)	109 (5)	103 (2)	106 (0)
	40	99 (0)	93 (4)	108 (5)	104 (9)	102 (7)	101 (8)
Esn	2.5	102 (5)	101 (3)	99 (2)	100 (4)	106 (4)	103 (6)
	10	106 (3)	114 (6)	104 (0)	108 (2)	111 (3)	107 (4)
	40	100 (3)	106 (3)	106 (3)	106 (6)	108 (3)	98 (4)
Et	2.5	101 (6)	104 (4)	109 (4)	111 (5)	110 (5)	111 (1)
	10	111 (10)	110 (3)	112 (5)	119 (4)	110 (2)	110 (4)
	40	100 (6)	101 (1)	114 (5)	113 (9)	114 (1)	109 (6)
Etn	2.5	102 (8)	104 (6)	100 (2)	98 (8)	100 (1)	104 (5)
	10	103 (5)	110 (3)	108 (4)	108 (3)	101 (1)	110 (1)
	40	99 (6)	99 (4)	110 (4)	110 (8)	100 (5)	106 (5)
Eco	2.5	97 (5)	100 (6)	94 (3)	97 (2)	101 (3)	94 (5)
	10	104 (4)	109 (3)	101 (2)	102 (3)	103 (2)	100 (4)
	40	96 (2)	100 (4)	102 (3)	101 (3)	102 (1)	92 (2)
Econ	2.5	100 (9)	105 (1)	105 (2)	102 (6)	111 (3)	113 (9)
	10	104 (10)	113 (7)	107 (1)	110 (3)	110 (2)	112 (3)
	40	95 (4)	105 (3)	114 (1)	101 (5)	109 (7)	109 (4)
α-Ekr	2.5	99 (8)	102 (3)	101 (2)	100 (3)	104 (4)	100 (4)
	10	99 (7)	112 (6)	105 (2)	102 (1)	111 (3)	104 (3)
	40	99 (1)	102 (2)	107 (4)	103 (1)	103 (2)	101 (2)
α-Ekrn	2.5	97 (12)	98 (5)	97 (1)	79 (8)	91 (9)	111 (3)
	10	100 (6)	100 (4)	101 (4)	90 (3)	93 (4)	109 (2)
	40	92 (1)	95 (2)	102 (3)	82 (8)	83 (4)	104 (5)
Ecr	2.5	107 (3)	101 (5)	98 (2)	97 (4)	102 (2)	95 (4)
	10	102 (6)	111 (6)	102 (1)	104 (2)	107 (2)	98 (4)
	40	99 (1)	100 (4)	102 (3)	102 (3)	104 (2)	96 (3)
Ecrn	2.5	107 (10)	85 (9)	87 (3)	62 (8)	78 (2)	100 (9)
	10	103 (5)	93 (14)	88 (1)	67 (6)	77 (5)	104 (8)
	40	93 (0)	88 (6)	91 (5)	67 (5)	72 (3)	102 (7)

Table 4.

%Recovery (%RSDr) of EAs fortified in oat-based food matrices at concentrations of 2.5, 10, and 40 μg/kg (n=3)^a.

Analyte	Fortified Concentration (μg/kg)	Steel-Cut Oats	Baby Oatmeal
Em	2.5	106 (4)	87 (7)
	10	110 (3)	94 (6)
	40	112 (1)	92 (1)
Emn	2.5	97 (2)	89 (5)
	10	102 (1)	97 (4)
	40	103 (5)	93 (6)
Es	2.5	108 (5)	107 (5)
	10	116 (6)	115 (6)
	40	116 (5)	112 (9)
Esn	2.5	109 (6)	108 (6)
	10	110 (3)	124 (3)
	40	110 (1)	115 (4)
Et	2.5	106 (3)	110 (5)
	10	117 (6)	115 (7)
	40	112 (4)	107 (3)
Etn	2.5	111 (5)	110 (5)
	10	114 (6)	124 (6)
	40	118 (3)	114 (4)
Eco	2.5	106 (2)	103 (3)
	10	108 (2)	109 (2)
	40	108 (2)	105 (3)
Econ	2.5	107 (1)	115 (1)
	10	116 (4)	126 (2)
	40	121 (3)	119 (1)
α-Ekr	2.5	104 (0)	106 (3)
	10	110 (2)	113 (1)
	40	109 (4)	110 (5)
α-Ekrn	2.5	97 (3)	102 (6)
	10	104 (5)	113 (7)
	40	103 (6)	104 (0)
Ecr	2.5	107 (1)	109 (3)
	10	109 (4)	114 (1)
	40	109 (3)	114 (4)
Ecrn	2.5	106 (10)	105 (2)
	10	106 (4)	116 (4)
	40	104 (10)	108 (1)

^aQuantification was performed using calibration curve prepared in oat flour extract.

Furthermore, the method was tested in two PT samples obtained from FAPAS, including EA contaminated rye flour and baby food. Acceptable z-score values between −2 and 2 were obtained for both PT samples, further confirming the suitability of the method (Table S2).

Method Application for Analysis of Cereal-based Food

Following validation, the method was applied to determine EA levels in a convenience sampling of sixty (60) commercially available cereal-based food products from the US market. In this study, the performance of the method was assessed according to quality assurance and quality control processes executed throughout the study. These processes included defined criteria to evaluate method accuracy (spike recovery), precision (triplicate analysis), and interferences (matrix blank) for individual products within each batch and product category.

Occurrence of Individual Ergot Alkaloids

Table 5 presents the incidence of individual EAs in the analyzed samples. At least one EA was detected (≥ LOD 0.3 μg/kg) in 25 of the 60 (42%) products with a range of <1.0 – 231 μg/kg for an individual EA. The most frequently detected EAs in all analyzed products were Ecrn (23 of 60 products) and α-Ekrn (22 of 60 products). The least frequent EAs were Em and Emn, both present in 10 of 60 products. When considering only concentrations above the LOQ, Ecr, α-Ekr, Et, Eco, and Es were the most predominant EAs, with mean values for samples with quantifiable results ranging from 25.4 – 63.5 μg/kg. These findings align with previous studies. The EFSA report (2017) identified the highest average contributors of EAs in food samples as Et (18%), Ecr (15%), and Es (12%) (EFSA et al. 2017). Ecr and Es were the predominant EAs in Italian food commodities (Debegnach et al. 2019), while Es, Ekr and Ecr were the most common EAs in cereals from European countries (Malysheva et al. 2014) and in wheat from Algeria (Carbonell-Rozas et al. 2021b). In contrast to our findings, Em and Emn were the most commonly detected EAs in wheat samples from Italy (Debegnach et al. 2019). These differences in the prevalence of specific EAs may be related to the influence of geographic and commodity-specific factors on EA profiles.

Table 5.

Detection and concentration of individual EAs in cereal-based food.

Analyte	Number (Percentage) of Positive Samplesa	Range of Individual EAs (μg/kg)b	Mean of Individual EAs (μg/kg)c	Median of Individual EAs (μg/kg)c
Em	10 (17%)	<1.0 – 51.4	18.0	9.24
Emn	10 (17%)	<1.0 – 15.5	7.30	5.84
Es	13 (22%)	<1.0 – 58.6	25.4	16.3
Esn	15 (25%)	<1.0 – 16.8	7.95	4.78
Et	17 (28%)	<1.0 – 81.7	28.5	6.92
Etn	17 (28%)	<1.0 – 21.3	8.91	9.01
Eco	11 (18%)	<1.0 – 91.4	28.2	10.0
Econ	20 (33%)	<1.0 – 37.5	11.2	4.27
α-Ekr	13 (22%)	<1.0 – 122	37.0	14.0
α-Ekrn	22 (37%)	<1.0 – 72.7	12.5	2.92
Ecr	19 (32%)	<1.0 – 231	63.5	25.7
Ecrn	23 (38%)	<1.0 – 53.3	14.7	3.74

^aAt least one EA detected (≥ LOD); number of total samples = 60.

^b<1.0: Individual detectable EA concentration between LOD (0.3 μg/kg) and LOQ (1.0 μg/kg).

^cMean and median concentrations calculated using quantifiable EA concentrations above the LOQ.

Additionally, among the 25 positive samples (at least one EA detected) in this study, 12 samples (48%; 8 rye-based, 2 oat-based, 1 wheat-based, and 1 barley-based) contained ten or more different EAs, while only 3 samples (12%) were identified with only one EA (Figure 3a). Current literature lacks a consensus regarding the distribution of EA in cereal-grain commodities. One study reports the presence of one EA (27%) predominantly in wheat and rye derived products, with only 13% of positive samples reported to contain more than nine EAs (Debegnach et al. 2019). In contrast, the EFSA report (2012) concluded that co-occurrence of all six major ergot alkaloids was most frequent in rye food samples collected in mills and in wheat- and rye- based food products from Belgian shops, whereas the co-occurrence of all six was less common in wheat flour products (EFSA 2012). Although the current study provides preliminary information on the prevalence of multiple EAs in tested cereal-based foods, findings may not be representative of the US market and could be influenced by geographic location and environmental factors at the time of sampling.

Figure 3.

(a) Distribution of number of individual EAs detected per sample (b) Average percentage of -ine (R)-epimers (blue) and -inine (S)-epimers in each food category.

In addition to the number and diversity of EAs, the distribution of their epimeric forms also revealed interesting trends. The average contribution of all -ine (R)-epimers to the total concentration was higher than that of -inine (S)-epimers, accounting for 76% and 24%, respectively. These values are consistent with those reported in the EFSA report (2017), which documented proportions of 73% and 27%, respectively (EFSA et al. 2017). Notably, a shift toward -inine (S)-epimers was observed in breakfast cereal and cracker categories, increasing from 26% in flour and 24% in grain/flake to 59% in breakfast cereal and 50% in cracker (Figure 3b). While this shift may suggest epimerization induced by food processing, such as baking, further investigation is required to validate these findings, due to the limited number of samples included in this study. The EFSA report (2017) observed a difference in the average contribution of the epimeric forms between unprocessed (inine (S)-epimers: 23%) and processed (inine (S)-epimers: 41%) foods (EFSA et al. 2017). Another study on wheat and rye products sourced from Italy reported a distinct ratio between the -ine/-inine epimers in flour and bread samples (Debegnach et al. 2019). Altogether, these differences in EA patterns may be attributed to variations in cereal types, geographical regions, fungal strains, and processing methods.

Distribution of Total Ergot Alkaloids

Total EA concentrations, grouped by product category and grain type, are summarized in Table 6. Total EA concentrations were determined as the sum of the 12 monitored EAs and ranged from <1.0 – 755 μg/kg in tested products. The mean concentration was 153 μg/kg and the median was 12.0 μg/kg, calculated only from positive samples with quantifiable values above the LOQ (1 μg/kg). To further illustrate the associations between total EA concentrations and product characteristics, a rank plot is presented in Figure 4, organized by product category (a), grain type (b), and organic status (c).

Table 6.

Detection of EAs and total EA concentrations in cereal-based food.

Product Category	Grain Type	Number of Samples	Number (Percentage) of Positive Samplesa	Range of Total EAs (μg/kg)b	Mean of Total EAs (μg/kg)c	Median of Total EAs (μg/kg)c
Breakfast Cereal	Multi-Grain	4	1 (25%)	4.15	NA	NA
	Oat	3	2 (67%)	4.51 – 45.2	24.8	24.8
	Wheat	7	0 (0%)	ND	-	-
Cracker	Multi-Grain	2	1 (50%)	2.62	NA	NA
	Oat	2	0 (0%)	ND	-	-
	Rye	4	4 (100%)	<1.0 – 13.3	8.87	10.7
	Wheat	7	4 (57%)	<1.0 – 5.20	3.95	3.95
Flour	Oat	2	2 (100%)	3.62 – 6.80	5.21	5.21
	Rye	8	7 (88%)	<1.0 – 755	428	413
	Wheat	7	2 (29%)	7.28 – 41.4	24.3	24.3
Grain/Flake	Barley	8	1 (13%)	347	NA	NA
	Oat	4	1 (25%)	<1.0	NA	NA
	Rye	2	0 (0%)	ND	-	-
Total	Total	60	25 (42%)	<1.0 – 755	153	12.0

^aAt least one EA detected(≥ LOD); number of total samples = 60.

^b<1.0: Products with detectable EA concentrations between LOD (0.3 μg/kg) and LOQ (1.0 μg/kg).

^cTotal EA mean and median concentrations calculated using quantifiable EA concentrations above the LOQ.

ND: Not-detected (< LOD); NA: Not-applicable.

Figure 4.

Rank plot of total EA concentrations in cereal-based food. Total EA concentrations are organized from lowest to highest, with respect to product category (a), grain type (b), and organic status (c). Only positive products with at least one EA detected are plotted.

Among the 25 positive products, 11 products (44%) were flour products, with the highest concentration of total EA (755 μg/kg) in a rye flour sample (Table 6). Notably, among the ten samples with the highest EA concentrations, seven were flour products (Figure 4a), highlighting a potential link between flour products and elevated EA levels. Previous studies have reported EA degradation in EA contaminated flours during baking (EFSA et al. 2017), which may account for the higher EA concentrations observed in flour products compared to other more extensively processed foods.

Overall, rye-based products displayed the highest EA occurrence (11 of 14 rye-based products were positive) and concentration (mean and median total EA concentration of 289 and 219 μg/kg, respectively) in tested products (Figure 4b, Table S3). This observation is consistent with literature findings on EA occurrence in cereal-grain products, reporting the highest levels of EAs in rye and rye-containing commodities (EFSA et al. 2017).

Among the 60 analyzed products, 24 products were labeled with an organic statement. Of these, 9 organic products contained at least one detectable EA and 6 ranked among the top 10 highest total EA concentrations (Figure 4c). Both the mean and median total EA concentrations were higher in the organic group, with a mean of 305 μg/kg and a median of 219 μg/kg, versus a mean of 71.9 μg/kg and a median of 5.2 μg/kg in products without an organic label (Table S4). As rye flour samples had the highest EA concentrations overall, this difference is likely related to the higher number of organic rye flour samples (6) compared to non-organic rye flour samples (2) as well as the study not being statistically representative of the U.S. market. Previous studies on rye-based products in the UK did not find significant differences in ergot levels between organic and non-organic products (Crews et al. 2009). However, a separate study examining cereals and cereal products from European countries reported significantly lower frequencies and concentrations in organic samples compared to products that were not labeled as organic (Malysheva et al. 2014).

Four products marketed towards infants and young children were collected in this study (Table S5). At least one EA was detected in 3 of 4 products (75%) with a range of <1.0 – 45.2 μg/kg. The highest total EA concentration in this class (45.2 μg/kg) was determined in an oat-based breakfast cereal, in which all 12 EA residues were detected. However, given the small sample size of products in this convenience sampling marketed for infants and young children, it is difficult to draw any further conclusions from these data.

Unlike ergot sclerotia in grains, EAs are not regulated in the US; however, the EU has implemented regulatory limits for them. While these are not used in the US, when comparing the convenience sampling data with EU maximum levels (ML) for various cereals and cereal products, four products were found to exceed permitted limits in the EU for total EA concentrations. These include two rye flour products with concentrations of 647 and 755 μg/kg, respectively (EU ML: 500 μg/kg), one barley grain product with a concentration of 347 μg/kg (EU ML: 150 μg/kg), and one oat-based breakfast cereal product marketed for infants and young children with a concentration of 45.2 μg/kg (EU ML: 20 μg/kg). Nevertheless, further data collection is needed to better understand the incidence and distribution of EAs in cereal-based food products from the US market.

Conclusions

Herein, we presented the development and validation of a targeted LC-MS/MS method for the determination of six predominant EAs and their epimers in cereal-based food matrices. The performance of the method, in terms of precision, accuracy, and quantification limits, was shown to be suitable for the quantification of individual EAs in cereal-based food. The validated method was applied to a convenience sampling of sixty cereal-based products purchased in the US, in which EAs were detected in 25 products (42%). Ecrn and α-Ekrn were the most detected EAs, while Em and Emn were the least detected in all analyzed samples. Notably, a significant percentage (48%) of positive samples contained more than 10 EAs per sample, indicating the high co-occurrence of multiple EAs in this small sampling. Similar to findings from other studies, an increasing ratio of epimers toward the -inine (S)-form was observed in breakfast cereals and crackers, suggesting a potential effect of baking on epimerization. When concentrations of all 12 monitored EAs were summed, the highest incidence and concentrations were found in flour (by product category) and rye (by grain type) products, consistent with findings reported in the literature. Additionally, EAs were detected in 3 out of 4 samples intended for infants and young children. Overall, these findings highlight the need for more comprehensive analytical data on the occurrence of EAs in cereal-based food products to better estimate dietary exposures and support regulatory decisions regarding food safety in the US marketplace.