Occurrence of Aflatoxin M₁ in Milk and Dairy Products Traded in São Paulo, Brazil: An Update

Abstract

The aim of this study was to conduct an up-to-date investigation on the occurrence levels of aflatoxin M₁ (AFM₁) in samples of raw milk (n = 40), pasteurized milk (n = 44), ultra-high temperature (UHT) milk (n = 27), Minas cheese (n = 57), and yogurt (n = 44) traded in São Paulo state, Brazil. AFM₁ was extracted from fluid milks and dairy products using immunoaffinity columns and determined by high performance liquid chromatography. AFM₁ was detected at the mean level of 0.080 ± 0.071 µg/L or kg in 72 samples (34.0%) evaluated in the study (n = 212). Detectable levels of AFM₁ were observed in five samples of raw milk (12.5%), 16 samples of pasteurized milk (36.4%), 13 samples of UHT milk (48.1%), 27 samples of cheese (47.4%), and 11 samples of yogurt (25.0%), although none of them had concentrations above the maximum permitted levels (MPL) for AFM₁ adopted in Brazil. However, 11.7% (n = 13) of samples of raw, pasteurized, and UHT milks would have AFM₁ concentrations above the MPL of 0.05 μg/L adopted in the EU. The maximum level was detected in one cheese sample containing 0.695 µg/kg. Although none of the samples exceeded the Brazilian MPL, the high frequencies of AFM₁ in Brazilian milk products warrant concern about their contribution to the human exposure to aflatoxins. Because aflatoxins are among the most potent carcinogens known, the results of this trial stress the need for stringent measures in the milk production system to avoid AFM₁ in milk and derived products.

1. Introduction

Mycotoxins are important contaminants produced as secondary metabolites by certain toxigenic fungi during their growth on food and feedstuffs, which can develop several toxic effects such as hepatotoxicity, nephrotoxicity, neurotoxicity, teratogenicity, immunosuppression, mutagenicity, and carcinogenicity, among others [1,2,3]. Aflatoxins are produced by fungal species from Aspergillus genus, especially A. flavus, A. parasiticus, and A. nomius [1]. As broadly dispersed, these fungi species and their harmful metabolites are a major problem to the food and feed industries [1]. Up to now, more than 18 compounds have been classified as “aflatoxin—AF” [2,3]. In food and feed materials, the highly toxic compound that normally occurs under any natural conditions is the aflatoxin B₁ (AFB₁), followed by G₁ (AFG₁), B₂ (AFB₂), and G₂ (AFG₂). After the ingestion of AFB₁ by animals, the hepatic biochemical transformation of the molecule produces the hydroxylated derivative known as aflatoxin M₁ (AFM₁), which is excreted into milk and other body fluids [4,5]. Both AFB₁ and AFM₁ are classified as group 1 human carcinogens by the International Agency for Research on Cancer (IARC) [6,7,8,9]. Many countries still lack a proper system to avoid food contamination with aflatoxins, especially milk and derived products that have been shown to be risky for consumption [10]. Milk is largely consumed by humans, therefore, the AFM₁ level in milk is constantly regulated worldwide [11,12]. The amount of AFM₁ excreted in milk of lactating animals has been expressed as a percentage of the AFB₁ intake from contaminated feed. The European Food Safety Authority has estimated that the transfer rate from AFB₁ in feed to AFM₁ in cow’s milk is 1–2% on average, although in high-productivity cattle, it can increase to 6% [13].

As extensively indicated in the literature, milk products such as pasteurized milk and ultra-high temperature (UHT), if not controlled properly, may contain AFM₁, thus posing a significant risk to consumers [3,4,9,10]. In addition, AFM₁ binds to casein and remains bound to it, even in the production processes of dairy products including powdered milk, cheese, and yogurt [14,15,16]. In line with this tight binding capacity and the high heat resistance of the toxin, AFM₁ in dairy products cannot be completely removed by standard heat-based pasteurization or sterilization processes [14,17]. Thus, studies have demonstrated that human exposure to AFM₁ can be associated with the consumption of the given products [14,18,19,20]. For instance, human exposure to an exceeded limit of AFM₁ in any of these products can contribute as causative agents for human diseases such as hepatocellular carcinoma (HCC) [21]. Thus, determination of AFM₁ in milk and dairy products are essential for control purposes, which requires appropriate analytical techniques such as enzyme-linked immunosorbent assay (ELISA) [22,23], high performance liquid chromatography (HPLC) [24], and HPLC coupled tandem mass spectrometry (MS/MS) [25].

AFM₁ levels in yogurt, pasteurized, and UHT milk, along with related health factors have been recently described in important milk-producing countries including China [26] and Brazil [5]. High frequencies (59–83%) of AFM₁ in yogurts and pasteurized and UHT milks exceeding the maximum permitted level (MPL) of 0.05 μg/L of the European Union (EU) were reported in China [26]. Moreover, about 79% of pasteurized and UHT milk were found to be contaminated with AFM₁, with high seasonal variation, whereas a high AFM₁ prevalence was observed in the winter [27]. A recent study confirmed that higher risk concerns regarding AFM₁ contamination of milk in Croatia were associated with autumn and winter [23]. However, a survey conducted in Bangladesh indicated no or small significant differences correlated to the seasonality [22]. In Brazil, previous studies regarding the occurrence of AFM₁ showed higher incidence rates from 63% to 100%, and levels ranging from 0.0002 to 0.106 µg/L among diverse milk products [28,29,30]. According to a Brazilian study [18], UHT milk, powdered milk, and infant formula on evaluation have exhibited 0.150–1.02 μg/kg of AFM₁, whose maximum level surpassed the MPL of 0.5 μg/L adopted in Brazil [11,12]. Recently, raw milk used in the manufacture of Minas cheese were evaluated in cheese processing plants from the Brazilian state of São Paulo, and the results indicated that 39 and 29% of samples were contaminated with AFM₁ at mean levels of 0.028 ± 0.009 µg/L and 0.113 ± 0.092 µg/kg, respectively [31]. However, there is little information available on the recent occurrence data regarding AFM₁ in pasteurized milk, UHT milk, Minas cheese, and yogurts commercially available in retail markets for human consumption in Brazil. Therefore, the objective of this study was to conduct an up-to-date investigation on the occurrence levels of AFM₁ in raw milk collected in dairy processing plants and in dairy products traded in São Paulo state, Brazil.

2. Materials and Methods

2.1. Sampling Design

A total of 212 samples of raw milk (n = 40), pasteurized milk (n = 44), UHT milk (n = 27), Minas cheese (n = 57), and yogurt (n = 44) were collected between January and August of 2022. Samples of raw milk were collected in 10 dairy plants located in the northeast region of São Paulo state, Brazil, while samples of pasteurized milk, UHT milk, cheese, and yogurt were collected in supermarkets from the same region. In each location, duplicate samples of raw milk (500 mL), processed milks (pasteurized and UHT: 1-L original bottle), Minas cheeses (500-g original package), and yogurts (900-g original bottle) were collected and transported to the laboratory in a cooler with dry ice and stored at 4 °C until the analysis of AFM₁.

2.2. Reagents and Instruments

Analytical grade reagents were used in all determinations of AFM1 in the milk and dairy samples. Ultra-pure water was provided by a Milli-Q^® system (Millipore, Bedford, MA, USA). The AFM₁ standard was purchased from Sigma Chemical Company (St Louis, MO, USA). Methanol and acetonitrile of HPLC grade were obtained from EM Science (Gibbstown, NJ, USA).

All chromatographic analyses were conducted on a Shimadzu^® (Kyoto, Japan) 10VP high performance liquid chromatographic (HPLC) system equipped with a Shimadzu^® (Kyoto, Japan) 10 AXL fluorescence detector (excitation at 360 nm and emission above 440 nm). The chromatographic column used was a Kinetex C₁₈ column (Phenomenex, Torrance, CA, USA) 4.6 × 150 mm with 2.6 μm particle size, coupled to a pre-column with a 0.5 μm filter. An isocratic mobile phase composed by methanol:water:acetonitrile (61.4:28.1:10.5, v/v/v) was used in all chromatographic runs, with a flow rate of 0.50 mL/min.

2.3. Sample Preparation

AFM₁ was extracted from samples of milk and dairy products using immunoaffinity columns (Aflatest^®, Vicam, Watertown, MA, USA) with minor modifications as described and validated in the laboratory by Jager et al. [32]. For raw, pasteurized, and UHT milks, 40 mL aliquots of each sample were transferred to Falcon tubes and pre-heated at 37 °C for 10 min. Then, 1 g of NaCl was added to the tube and centrifuged (SOLAB SL-700, Piracicaba, Brazil) at 3500 rpm for 5 min. After separating the upper fat layer, the sample was submitted to a second centrifugation step to remove the remaining fat. Samples were filtered through Whatman no. 4 filter paper, and 30 mL-aliquots were passed into immunoaffinity columns attached to a manifold (Supelco, Bellefonte, PA, USA) and vacuum pump (Primatec, Itu, Brazil) at a flow rate of 2–3 mL/min. Twenty mL of ultra-pure water (Milli-Q^®, Millipore, Bedford, MA, USA) was used to wash the column before eluting AFM₁ with 1 mL of methanol. The eluate was recovered in glass vials and submitted to evaporation under nitrogen to near dryness. The final extract was resuspended in 1 mL of methanol/water (50:50, v/v), and 25 μL was injected into the HPLC system, as described in Section 2.2. The retention time for AFM₁ was 5.9 min.

AFM₁ in yogurt and Minas cheese was determined by weighing 8 g of ground cheese in a Falcon tube and then adding 2 g of sodium chloride, 22 mL of methanol, and 13 mL of water. The extraction was conducted by mixing the mixture in a mixer for 1 min, followed by centrifugation at 3500 rpm for 10 min. The upper liquid phase was filtered through Whatman no. 4 filter paper, and 20 mL of the filtrate was diluted with 40 mL of ultra-pure water (Milli-Q^®, Millipore, Bedford, MA, USA). The mixture was passed through the immunoaffinity column (2–3 mL/min) attached to a manifold (Supelco, Bellefonte, PA, USA) and vacuum pump (Primatec, Itu, Brazil), and the subsequent extraction steps were identical as described for the milk samples.

Calibration curves to quantify AFM₁ in the milk and milk products (yogurt and cheese) were prepared using AFM₁ standard solutions (Sigma^®) prepared in acetonitrile at levels ranging from 0.1 to 1 μg/L and 0.1 to 20 μg/L, respectively. The AFM₁ concentrations in the samples were linearly correlated with the integrated peak areas of the standards. The limits of detection (LOD) and quantification (LOQ) were determined for each analytical method considering a signal-to-noise ratio of 3:1 and 10:1, respectively. Linearity was verified in all chromatographic runs by checking the coefficient of determination (r²) through visual observation of the residual plots of the calibration curves.

The performance of the method for AFM₁ in the milk and dairy products samples has been presented elsewhere [32], describing LOD and LOQ values of 0.0025 and 0.0080 μg/L for AFM₁ in fluid milk, respectively. LOD and LOQ of AFM₁ in the cheese and yogurt samples were 0.017 and 0.055 μg/L, respectively.

3. Results and Discussion

The occurrence levels of AFM₁ in samples of raw milk collected at dairy plants and the processed milks and dairy products commercially available in São Paulo state are summarized in

Table 1. Aflatoxin M₁ (AFM₁) levels in raw milk collected at dairy plants, and pasteurized milk, UHT milk, cheese, and yogurt commercially available in São Paulo, Brazil.

Product	Number of Samples According to the Level of AFM1										Mean 3	Range 4
	<LOD 1		LOD–0.050 2		0.051–0.250 2		0.251–0.500 2		0.501–1.00 2
	n	%	n	%	n	%	n	%	n	%
Raw milk (n = 40)	35	87.5	1	2.5	4	10.0	0	0	0	0	0.114 ± 0.070	0.014–0.182
Pasteurized milk (n = 44)	28	63.6	10	22.7	6	13.7	0	0	0	0	0.032 ± 0.033	0.003–0.117
UHT milk (n = 27)	14	51.9	10	37.0	3	11.1	0	0	0	0	0.080 ± 0.038	0.020–0.148
Minas cheese (n = 57)	30	52.6	16	28.1	7	12.3	1	1.7	3	5.3	0.122 ± 0.190	0.017–0.695
Yogurt (n = 44)	33	75.0	5	11.4	6	13.6	0	0	0	0	0.050 ± 0.025	0.017–0.091
Total (N = 212)	140	66.0	42	19.8	26	12.3	1	0.5	3	1.4	0.080 ± 0.071	0.003–0.695

¹ LOD: Limit of detection of the analytical methods: 0.0025 μg/L (raw, pasteurized and UHT milks) and 0.017 μg/kg (cheese and yogurt). ² µg/L or µg/kg. ³ Values (µg/L or µg/kg) are reported as mean ± standard deviation, for samples analyzed in duplicate containing detectable levels of AFM₁. ⁴ Minimum and maximum levels of AFM₁ (µg/L or µg/kg). UHT: Ultra high temperature.

. Throughout the analyses, AFM₁ was detected in 72 samples (34.0%) of milk and dairy products evaluated in the study (N = 212). Detectable levels of AFM₁ were observed in five samples of raw milk (12.5%; n = 40), 16 samples of pasteurized milk (36.4%; n = 44), 13 samples of UHT milk (48.1%; n = 27), 27 samples of Minas cheese (47.4%; n = 57), and 11 samples of yogurt (25.0%; n = 44). The maximum level was detected in one cheese sample containing 0.695 µg/kg. These results indicate a high incidence of low levels of AFM₁ in milk and dairy products, especially in fluid milks (pasteurized and UHT) and in cheese.

Toxigenic fungal molds are widely disseminated, and inherently affect the food and feed chain. In particular, milk and further dairy products are globally consumed by people of all ages (e.g., toddlers to the elderly), thus unacceptably higher frequencies and levels of AFM₁ in milk products can pose risks to human health. In this study, the raw, pasteurized, and UHT milks showed the occurrence of AFM₁ at the mean levels of 0.114 ± 0.070, 0.032 ± 0.033, and 0.080 ± 0.038 μg/L, respectively. All samples had AFM₁ levels below the Brazilian MPL of 0.5 μg/L [12], which is in contrast to the higher occurrence levels of AFM₁ in milk from several countries [26,33,34,35,36]. Compared with the data reported here, much higher AFM₁ concentrations in the pasteurized milks at maximum levels of 4.81, 2.33, and 1.14 μg/L were determined in Pakistan [36], India [34], and Ethiopia [35], respectively. Additionally, in contrast with our results, UHT milks with higher AFM₁ concentrations of 1.54 [36], 2.58 [34], and 0.56 μg/L [18] were previously reported in Brazil. In these studies, AFM₁ concentrations in the pasteurized or UHT milks exceeded the established EU (0.05 μg/L) or the Brazilian limits (0.5 μg/L) [8]. In the same way, higher incidence rates (63–100%) of AFM₁ at levels ranging from 0.0002 to 0.106 µg/L were observed in milk products traded in Brazil [28,29,30]. Thus, compared with previous data on the frequencies and levels of AFM₁ in milk and dairy products, our results indicate that control measures regarding the occurrence of AFB₁ in feed for dairy cows have been improved in Brazil. However, 13 samples (34%) of fluid milk analyzed (raw, pasteurized and UHT) in our study had AFM₁ levels above the MPL of 0.05 μg/L adopted in the EU [11]. Moreover, the mean concentration of AFM₁ in yogurt samples was 0.050 ± 0.025 μg/L, and none of the samples analyzed exceed any established limits for raw milk used for the manufacture of dairy products (0.5 μg/L of Brazil or 0.05 μg/L of the EU), which corroborates the low occurrence levels of AFM₁ in this type of milk product. Much higher mean concentrations of AFM₁ at 1.63, 0.40, and 0.09 μg/L have been described in yogurt samples from Ethiopia [35], Yemen [33], and Lebanon [37], respectively. However, AFM₁ concentrations or its variabilities in milk products from different countries can indicate geographical or climatic effects, poor-quality control, and unsupervised feed and dairy products, respectively.

Minas cheese is another largely consumed, traditional Brazilian dairy product that has been studied regarding AFM₁ contamination [20,38]. The mean level of AFM₁ in cheese samples collected from retail markets in our study was 0.122 ± 0.190 μg/L, although only three samples (5%) showed a moderate concentration range between 0.50 and 1.00 μg/kg, which was below the normal limit of 2.5 μg/kg settled in Brazil [12]. This is consistent with the frequency of positive samples (29%) and the mean concentration of AFM₁ (0.113 ± 0.092 µg/kg) in the Minas cheese collected in cheese processing plants from São Paulo [31]. However, studies over the past decades have demonstrated higher AFM₁ levels in the stated samples. For example, 74.7% cheese samples demonstrated AFM₁ at the levels of 0.02 to 6.92 μg/g [39], hence indicating some samples with non-compliance levels when compared to the MPL of 2.5 μg/L adopted for cheeses by the Brazilian regulations [12,40]. In agreement with the data reported in this study, lower levels ranging from 0.050 to 0.31 μg/g were observed in Iranian cheeses, which were related to variabilities in the environmental or seasonal conditions [38].

The lack of awareness regarding the AFB₁ contamination of feed for dairy cattle and limitations in analytical services remains the most important causes of the high incidence of AFM₁ in dairy products in Brazil. In addition, there are no regulations regarding MPL for aflatoxins in feed for dairy cattle in Brazil. Franco et al. [41] found AFB₁ at levels ranging from 8.7 to 390 μg/g in 13% of the 45 feed samples collected in Brazilian small-scale farms. However, Brazilian milk production has undergone several modernization processes over the past decades, leading to significant improvements in the overall quality of milk produced in the country [35], which may have contributed to the low levels of AFM₁ within Brazilian MPL values for milk and dairy products, as determined in the present study.

4. Conclusions

High incidences (12.5–47.4%) of low levels of AFM₁ (mean: 0.080 ± 0.071 µg/L or kg) were observed in raw milk collected from dairy plants as well as in pasteurized milk, UHT milk, yogurt, and cheeses commercially available in São Paulo state, Brazil. Although these levels were below the limits established by Brazilian regulations (0.5 µg/L), 34% of the samples of raw, pasteurized, and UHT milks had AFM₁ concentrations above the MPL of 0.05 μg/L adopted in the EU. The frequent occurrence of AFM₁ in Brazilian milk products warrants concern about their contribution for the human exposure to dietary aflatoxins, taking into account their hepatocarcinogenity. Hence, the results of this trial indicate that stringent measures such as good agricultural practices and the adequate storage of dairy cow feed are needed to avoid AFM₁ in milk and derived products. Further studies should be carried out to evaluate the occurrence levels of AFM1 in other dairy products such as milk powder and other types of cheese available for human consumption in Brazil.