Introduction

Rice is the seed of a monocot plant, Oryza sativa L., from the Poaceae family. It is a staple food for half of the earth’s population. Among the world’s most important grains, rice is second after maize. Global rice production is estimated to be 508.7 million metric tons [1]. Most of the world’s rice (75%) is transplanted under flooded conditions, while the remainder is direct-seeded [2]. In the 2021/22 fiscal year, Americans and US residents consumed 14.5 kg of rice per capita [3]. In the USA, rice is cultivated in five states, but the USA must also import a large amount of rice from Asian countries, including Thailand, India, Pakistan, and Bangladesh every year to meet consumption demands [4]. Rice is an important source of both essential and potentially toxic elements [5]. Selected elements, including Cu, Fe, Mn, Se, and Zn, are responsible for the synthesis of several enzymes and gene expression regulation [6,7,8]. Other elements, such as As and Cd, are considered toxic, posing a health risk to humans [9, 10].

Some rice-growing regions command a premium price in the world market and are vulnerable to economic adulteration and mislabelling. For this reason, cultivation area and rice cultivar are crucial factors in market prices [11]. In this sense, consumers’ interest in foods with traceable origins and geographic indication status is growing, and robust analytical tools are necessary to authenticate them [11,12,13]. Food safety monitoring programs require the authenticity of the origin of food to be tracked. Market regulations have been introduced in several countries to guarantee food safety [14, 15]. In 2009, the US House of Representatives passed the Food Safety Enhancement Act (HR2749 RS) requiring the disclosure of full information regarding the geographical origin of foods grown in the USA or foods intended for import into the USA [16].

However, there is still a lack of scientific information to ensure the implementation of safety regulations. Previous studies widely evidenced and recognized that the multi-element composition of rice reflects the soil composition and agro-climates in which it is grown and, thus, could be used as an indicator of its geographical origin [11, 12, 17]. However, the traceability of food depends not only on reliably measured data but also on appropriate chemometric data-processing methods such as principal component analysis (PCA) and linear discriminant analysis (LDA) [17]. Because rice is a newly emerging crop in terms of cultivation in the USA, it is reasonable to verify the quality and geographic origins of US-grown rice using analytical methods in combination with statistical analysis. However, to the best of our knowledge, no studies have been previously reported on the geographic discrimination of rice on sale in Las Vegas. Thus, the aims of this study were to determine the metal(loid) profile of rice on sale in Las Vegas and to discriminate US-grown rice from imported rice grains based on the metal(loid) profile coupled with the LDA model.

Materials and Methods

Chemicals and Reagents

Nitric acid (69% HNO3) was obtained from J.T. Baker (Center Valley, PA, USA). The wash-nitric acid blank and tune solution for ICP-MS were obtained from Agilent Technologies (Santa Clara, CA, USA). An ICP-MS grade solution containing 1000 μg mL−1 single-element from CPI International (Santa Rosa, CA, USA) was used to prepare the calibration curve. An internal standard solution (Li, Sc, Y, In, Tb, and Bi) from CPI International (Santa Rosa, CA, USA) was also used during the ICP-MS analysis. Rice flour (NIST-1568b) certified reference material (CRM) used for method validation was obtained from the National Institute of Standards and Technology (Gaithersburg, MD, USA).

Collection of Rice Samples

A total of 54 rice samples were purchased from stores in Las Vegas, USA. Based on the package label information, these samples represented 8 countries (USA (20), Thailand (12), India (7), Bangladesh (6), Pakistan (6), Argentina (1), Spain (1), and Sri Lanka (1)). For milling, 100 g of individual samples was taken and powdered using ADL multi-purpose grinder. The samples were packed in polyethylene bags and stored at room temperature prior to digestion.

Sample Digestion and Analyses

Rice samples were digested using the Hot Block digestion method (EPA Method 3050b). Briefly, 0.25 g of rice flour sample was placed in a 50 ml polyethylene vial (Environmental Express, Charleston, SC) followed by 5 ml of concentrated HNO3. The vials were placed in a heating block (Environmental Express, Charleston, SC) and loosely covered with caps then heated at 95 °C for 2 h. After cooling to room temperature, 5 ml of HNO3 was added and heated until the disappearance of brown fume appeared. Then, after cooling to room temperature, 2 ml of distilled water and 3 ml of 30% H2O2 were added and heated and 30% H2O2 was added and heated again until effervescence subsided. After cooling, the digestate was made into 20 ml with distilled water and filtered using a 0.22-μm pore size syringe filter. Elemental analyses (Ag, Al, As, Cd, Co, Cr, Cu, Fe, Mn, Ni, Se, Sr, V, and Zn) were carried out by using an Agilent 7850 ICP-MS (Agilent Technologies, CA, USA) equipped with a Peltier-cooled quartz spray chamber (Scott type), a quartz torch (2.5 mm i.d.), and a low flow concentric nebulizer (0.2 mL min−1). Each sample was analyzed in triplicate. The ICP-MS operating conditions are presented in Tables S1 and S2. The detection limit (DL) and quantification limit (QL) were calculated as the concentration that provided 3 and 10 times the signal of the blank for each element, respectively. Method DLs are given in Table S3.

Quality Control and Assurance

For quality control and quality assurance purposes, new polyethylene vials were used throughout the sample digestion and preparation steps. The reagents used for the preparation of the calibration standard solutions and treatments were analytical grades. Internal standards were automatically added using the ICP-MS instrument. The mean values and RSD for the elements in the CCV-100 solution are 98.9–102 and 0.5–1.4 μg L−1, respectively. The CRM was prepared and analyzed in the same way as the samples, observing good agreement between certified and analyzed values with recovery values ranging from 89 to 117% for Al, As, Cd, Cu, Fe, Mn, Se, and Zn. Rice flour samples were also spiked with Ag, Co, Cr, Ni, Sr, and V for recovery tests. The recoveries of the six elements were ranged from 80 to 121%.

Statistical Analysis

Data were analyzed using Statistical Package for Social Sciences (SPSS) version 23 and XLSTAT (Addinsoft, NY, USA). The normal distribution and homogeneity of variance were tested using Kolmogorov-Smirnov’s and Levene’s tests, respectively. In cases of non-normal distribution, the data were log-transformed. Whenever data/transformed data satisfied these assumptions, the mean concentrations of analyzed elements were statistically evaluated by analysis of variance (ANOVA) or Student’s t test. The Kruskal-Wallis test was used in combination with Dunn’s multiple comparisons otherwise. Linear discriminant analysis (LDA) was used to discriminate the geographic origins of rice grains. LDA is a supervised classification method that maximizes the variation between groups and minimizes variation within a group by creating new variables (discriminant functions), which are linear combinations of the original variables. Statistical significance was established at p < 0.05.

Results and Discussion

Metal(loid) Concentration in Rice

Aluminum

Aluminum is ubiquitous in the environment. Aluminum has no recognized biological function in humans [18]. Prolonged exposure to Al is related to neurodegenerative diseases such as Alzheimer’s [19]. On average, an adult in the USA eats roughly 7–9 mg of aluminum per day [20]. Cereals are the main Al contributors [21]. In this study, US rice exhibited a significantly lower mean Al content (3.2 mg kg−1) compared with rice from the other geographic origins. The mean Al concentrations were 5.4, 4.8, and 4.2 mg kg−1 in black, brown, and white rice, respectively (Fig. S1). Antoine et al. [21] found higher levels of Al in rice samples from Jamaica, with concentrations ranging from 6.1 to 23.4 mg kg−1 in white rice and 7.4 to 43 mg kg−1 in brown rice. The mean values of Al for basmati and jasmine rice were 9.30 and 14.7 mg kg−1, respectively (Fig. S1).

Arsenic

Arsenic is a metalloid and its presence in the environment is the result of both natural geologic processes and environmental pollution. Arsenic has been linked to a wide range of deleterious health impacts, including cancers [22, 23], diabetes [24], and cardiovascular diseases [25]. Furthermore, As exposure is also associated with elevated genotoxic effects [26]. Rice, a staple crop that is grown in paddy fields, may pose a particular health concern as it accumulates tenfold higher levels of inorganic arsenic (i-As), an established human carcinogen, than any other grain [27, 28]. The mean As concentrations in USA (103 μg kg−1) and Thai (115 μg kg−1) rice samples were significantly higher than concentrations found for Bangladesh, India, and Pakistan rice samples. Similarly, rice grown in some regions of the USA has been reported to have higher mean total As concentrations than rice grown in other geographic regions [27, 29], which may be related to the increase in As contamination in paddy soils. Likewise, a two-and-a-half-fold higher mean As concentration was obtained in US rice samples compared to Pakistani rice samples sold in Almadinah Almunawarah markets, KSA [30]. On the other hand, three individual rice samples from Argentina, Spain, and Sri Lanka exhibited As concentrations of 176 μg kg−1, 314 μg kg−1, and 50 μg kg−1, respectively. For minimizing As exposure due to rice consumption, knowing where As is elevated in rice grains can help choose lower concentration grains based on growing regions [31]. A maximum residue limit (MRL) of 200 μg kg−1 is set for inorganic-arsenic (i-As) [32]. Although only total As concentrations were determined in this study, one white rice sample may only exceed the MRL if the i-As of this sample exceeds 63.6% of the total As. Moreover, all the analyzed rice samples contained total As concentrations far lower than the MRLs value stipulated by the Food Safety and Standards Authority of India (FSSAI) [33]. In fact, rice grown in the USA contains lower concentrations of i-As [27, 29].

The As content in rice also varies with color. In the present study, black and brown rice contained significantly higher concentrations than white rice (Fig. 1). Consistent with this, Chen et al. [34] reported that brown rice has two times higher level of As than white rice. Only recently, TatahMentan et al. [4] studied the toxic element contents of rice samples from various geographic origins, including the USA and China, and found that brown and black rice contained higher levels of As than white rice. In the case of white rice, the bran and germ (or most of it) are removed during polishing, whereas black and brown rice with the bran still intact may have a higher fiber content than white rice but also have higher As levels [4, 35]. Arsenic accumulation also varies by brand. Basmati rice contained less As than rice samples under the “other” category (Fig. 1). On the other hand, basmati rice exhibited significantly similar As content to jasmine rice. This is not consistent with the results of Zavala et al. [36] who reported that basmati and jasmine rice exhibited higher As levels. Conversely, jasmine rice from Thailand had a higher mean As (130 μg kg−1) compared with basmati rice from India and Pakistan (60 μg kg−1) [37].

Fig. 1
figure 1

Box plot of As concentration in rice samples based on (A) color and (B) brand. + indicates mean value; dash line indicates median value; bars represent the 1st and 3rd quartile

Full size image

Cadmium

Cadmium is a toxic heavy metal and a carcinogenic environmental pollutant [38]. Food represents the major source of cadmium exposure pathway [39]. Long-term Cd exposure may be associated with bone defects, high blood pressure, and myocardial dysfunction [40]. Irrespective of rice varieties, the mean concentrations of Cd in rice on sale in Las Vegas ranged from 3.64 μg kg−1 (Pakistan) to 41 μg kg−1 (Bangladesh). The mean Cd concentrations in rice samples from the USA (5.33 μg kg−1), Thailand (7.18 μg kg−1), and Pakistan were significantly lower than those rice samples originating from Bangladesh and India (16.5 μg kg−1). Similarly, several previous studies reported that the mean Cd concentrations in Bangladeshi rice were the highest (88 μg kg−1, [39]; 73 μg kg−1, [41]). A far higher mean concentration of Cd, 750 μg kg−1, in Bangladeshi rice samples was also observed by Kormoker et al. [42]. A particular concern from Cd exposure is infants; stricter standards are in place in the EU for Cd in processed baby foods (40 μg kg−1) [43], and 100% of Bangladeshi rice samples failed this standard, making it unsuitable for use in making baby foods. The mean Cd concentration in rice from the USA was 5.33 μg kg−1, which is considerably comparable with the value (6.2 μg kg−1) reported by Rahman et al. [39]. While countries such as Argentina, Spain, and Sri Lanka have been less considered, the results presented were for single samples, and Cd concentrations were 4.8, 6.7, and 28.5 μg kg−1, respectively. Shi et al. [31] established a global context for Cd in rice grain. They found that the Americas were typically low in cadmium, but the Indian sub-continent and certain regions of Bangladesh were universally elevated. The mean Cd concentrations in rice samples from five different geographic origins were well below the European ML of 200 μg kg−1 [43].

The mean Cd concentrations were 8.6, 11.5, and 11.6 μg kg−1 in black, brown, and white rice, respectively (Fig. 2). In a 2001 market survey on rice on the Swedish market, the mean Cd concentration was similar between white (20 μg kg−1) and brown (21 μg kg−1) but the variation in Cd among geographic origins of rice was considerable, indicating the influence of soil, climatic, and management variables [37]. The mean concentrations of Cd in basmati and jasmine rice were 10.3 μg kg−1 and 4.64 μg kg−1, respectively. Previously, Jorhem et al. [37] reported non-considerable variation in Cd between basmati (23 μg kg−1) and jasmine (18 μg kg−1) rice. Basmati rice from Pakistan contained remarkably higher Cd concentration than the concentration presented in the present study [44].

Fig. 2
figure 2

Cd content in rice based on (A) color and (B) brand. + indicates mean value; dash line indicates median value; bars represent the 1st and 3rd quartile

Full size image

Cobalt

Cobalt is an essential oligo element, particularly as a constituent of vitamin B12 (cobalamine) and some metalloproteins; however, excessive exposure to this element causes goiter and reduced thyroid activity [45, 46]. For example, cobalt intoxication has been a well-documented cause of cardiomyopathy for over 50 years; however, it has mostly been known in the context of so-called Quebec beer drinkers’ cardiomyopathy [47, 48]. Historically, Quebec beer drinkers’ cardiomyopathy was characterized by congestive heart failure caused by the addition of small amounts of cobalt chloride as a foam stabilizer by certain breweries in Canada and the USA [47, 48]. Moreover, cobalt intoxication is an increasingly recognized and life-threatening problem associated with hip prostheses [49]. Yet Co is needed in trace amounts in cobalamine to avoid megaloblastic anemia [48]. In the present study, the mean Co concentration varied between 5.53 μg kg−1 (Bangladesh) and 11.9 μg kg−1 (Thailand), with a significant difference between Thai (12.0 μg kg−1) and USA (6.52 μg kg−1) rice. Similar lower concentrations were previously reported by Rahman et al. [39] and Shraim et al. [30]. Conversely, Naseri et al. [50] reported a range of mean Co concentrations (110–440 μg kg−1) in domestic and imported rice from Iran. Black and brown rice contained significantly higher concentrations than white rice (Fig. S2). The mean content of white rice is somewhat lower than reported in other studies [5, 21]. The mean concentration of Co in jasmine rice (11.5 μg kg−1) was about two times higher than in basmati rice (6.2 μg kg−1) (Fig. S2).

Chromium

Chromium has high redox potential, and it can exist in various oxidation states, including the most important +3 (trivalent), and +6 (hexavalent) states [51]. Several studies have shown that exposure to hexavalent chromium [Cr (VI)] is associated with several types of cancer and is increasingly being recognized as a neurotoxicant [52]. On the other hand, Cr(III) is generally regarded as less toxic and is argued to be potentially therapeutic for type II diabetes [53]. In the USA, Cr(VI) is released to the environment primarily from stationary point sources resulting from human activities, and these sources contribute an estimated 900–970 tons of Cr(VI) emissions per year [54]. In this study, mean Cr concentrations in rice on sale in Las Vegas varied from 0.117 (Thailand) to 0.140 mg kg–1 (India). Likewise, the mean Cr concentration in rice from the USA was 0.120 mg kg–1. Naseri et al. [50] found an average of 0.47 mg kg–1 in Thai rice samples collected from retail stores in Shiraz, Iran. On the other hand, some studies have indicated excessive levels of Cr in rice from the wastewater-irrigated sites in India [55, 56].

Despite the non-statistical significance among black, brown, and white rice for Cr (Fig. S3), brown rice presented the highest mean concentration (0.131 mg kg–1) followed by white rice (0.122 mg kg–1). Likewise, a relatively lower mean Cr concentration (0.056 mg kg–1) was obtained in the USA white rice samples sold in Brisbane, Australia [30]. Similarly, Jorhem et al. [37] reported a remarkably lower mean Cr concentration (0.008 mg kg–1) for brown, white, and parboiled rice from the Swedish market study. Basmati rice presented the highest mean Cr content (0.141 mg kg–1) against 0.118 mg kg–1 for jasmine and “other” rice brands.

Copper

Copper is an essential trace element that is required by humans in several metabolic processes, including hemoglobin synthesis, iron oxidation, neurotransmitter biosynthesis, cellular respiration, and pigment formation [57, 58]. For adults, the recommended dietary allowance for Cu is 0.9 mg per day. Consequently, Cu deficiency is rare; however, the hematologic manifestations associated with Cu deficiency, such as anemia, osteoporosis, leukopenia, and neutropenia, are well documented [59]. The mean Cu contents in Indian (1.9 mg kg−1) and Pakistani (1.8 mg kg−1) rice on sale in Las Vegas were significantly higher than those found in Thai (1.5 mg kg−1) and USA (1.5 mg kg−1) rice. A relatively comparable mean Cu (1.01 mg kg−1) concentration was found in US rice collected from Almadinah Almunawarah markets, KSA [30]. Conversely, the mean Cu concentration in US rice of the present study was about twofold lower than that reported by TatahMentan et al. [60]. Similarly, Shraim [30] found a mean Cu concentration of 4.4 mg kg−1 in US rice collected from Brisbane, Australia. Previous studies reported a lower mean Cu concentration in Bangladeshi rice [39, 41], which is comparable with the present study. The mean concentration of Cu in brown rice was significantly higher than that in white rice (Fig. 3). A recent study by TatahMentan et al. [60] showed that the mean Cu concentration in brown rice was nearly 1.4-fold higher than found in white rice. The Cu content in basmati rice differed significantly from jasmine (Fig. 3).

Fig. 3
figure 3

Box plot of Cu concentration in rice samples based on (A) color and (B) brand. + indicates mean value; dash line indicates median value; bars represent the 1st and 3rd quartile

Full size image

Iron

Iron deficiency is the most common nutritional deficiency. Globally, more than two billion people suffer from Fe deficiency annually [61, 62]. Iron deficiency and anemia are associated with impaired neurocognitive development and immune function in different life stages [63]. The mean concentrations of Fe in rice on sale in Las Vegas ranged from 3.0 mg kg−1 (Bangladesh) to 5 mg kg−1 (India). Iron was observed in lower concentrations ranging from 1.4 to 9.1 mg kg−1, with a mean value of 4 mg kg−1 in US rice. These results agree with those obtained by Shraim [30] who reported a mean level of 3.3 mg kg−1 for Fe, in US rice samples collected from Brisbane, Australia. The literature is quite extensive for Fe in rice and reports concentrations between 1.75 and 47 mg kg−1 [30, 60, 64]. It is worth noting that regular rice consumption in different forms would cause Fe malnutrition among vegetarians. On the other hand, significant differences in the contents of Fe were observed among rice colors (Fig. 4). Likewise, the mean Fe concentration in brown rice was about fourfold higher than that in white rice [60]. Although genetic variations exist for Fe content in grains of rice germplasms, commonly grown rice varieties are poor sources of Fe [62]. However, when comparing the mean Fe concentration between basmati and jasmine rice, the values are very similar and showed no significant difference (Fig. 4). Dumitrascu et al. [65] also observed comparable Fe concentrations between basmati and jasmine rice.

Fig. 4
figure 4

Box plot of Fe concentration in rice samples based on A color and B brand. + indicates mean value; dash line indicates median value; bars represent the 1st and 3rd quartile

Full size image

Manganese

Manganese is an essential micronutrient that plays a primary role in the activation of several enzymes [7]. Because of its ubiquity in the environment, Mn deficiency is unlikely. The mean Mn contents in US (10.2 mg kg−1) and Thai (10.4 mg kg−1) rice on sale in Las Vegas were significantly higher than those found in Pakistani rice (4.2 mg kg−1). On the other hand, relatively low Mn concentrations have been reported for US rice samples collected from Almadinah Almunawarah markets in KSA with a mean Mn concentration of 5.2 mg kg−1 [30]. Conversely, TatahMentan et al. [60] found an average of 15 mg kg–1 in rice samples cultivated in four states (California, Louisiana, Arkansas, and Texas) in the USA. The mean contents of black and brown rice samples were significantly higher than white rice. TatahMentan et al. [60] reported mean Mn concentrations of 11 mg kg−1 and 29.2 mg kg−1 for white and brown rice, respectively. Basmati rice had a significantly lower mean Mn content than the “other” category, whereas there was no significant difference between basmati and jasmine (Fig. S4).

Nickel

Nickel is generally not considered an essential nutrient for higher animals or humans because it lacks a clearly defined specific biochemical function [66]. Excessive exposure to Ni may cause oxidative stress in various organs in animals and humans, including liver and kidney damage [67, 68]. The mean Ni content is significantly lower in US rice (0.214 mg kg−1) compared to Bangladeshi rice (0.362 mg kg−1) and Indian rice (0.412 mg kg−1), respectively. The mean Ni concentration in samples from the five geographic origins was below the ML set forth by the United Nations Food and Agriculture Organization and World Health Organization (FAO/WHO) [69]. Nickel concentrations in rice in the present study were comparable to the levels reported for Chinese and Spanish rice [70, 71]. On the other hand, relatively high Ni concentrations have been reported for US rice samples collected from Brisbane, Australia, with a mean Ni concentration of 1.93 mg kg−1 [30]. Conversely, nearly all rice samples collected from Almadinah Almunawarah markets in KSA contained no detectable Ni [30].

Figure S5 also compares the mean content of Ni among black, brown, and white rice. Rahman et al. [39] reported the highest mean Ni concentration in brown rice (0.268 mg kg−1) compared to white rice (0.07 mg kg−1). Kruskal-Wallis analysis showed that the Ni content was significantly similar among black, brown, and white rice. For Ni, basmati rice had a significantly higher Ni content than jasmine, whereas there was no significant difference between jasmine and “other” brands.

Selenium

Selenium is an essential constituent of many antioxidant enzymes such as glutathione peroxidase and selenoprotein P in humans and animals [72, 73]. The deficiency of this element has been associated with various clinical consequences including cardiovascular diseases, type 2 diabetes, and an increased risk of certain cancers [73, 74]. In fact, a trace quantity of Se is required to maintain good health in both animals and humans. The Institute of Medicine has set the Recommended Dietary Allowance for Se at 55 μg per day for adults [75]. Despite the importance of Se in promoting growth and maintaining healthy heart muscle, rice is scarcely investigated in terms of total Se content and Se speciation. In this study, the mean Se content in rice ranged from 0.02 (Bangladesh) to 0.07 mg kg−1 (India). The mean Se content in rice grown in the USA was 0.035 mg kg−1, regarded as seriously low-selenium rice. There are factors that limit rice Se levels, including geochemistry, and soil composition [76].

According to Kelly et al. [77], rice grown in one of the recognized rice-growing regions in the USA (Arkansas) is characterized by relatively high concentrations of Se (0.125–0.4 mg kg−1). Likewise, considerable regional diversity in Se concentrations in US rice was observed by Williams et al. [76]. On the other hand, Zhang et al. [78] analyzed Se content in rice grown in a Se-rich environment in China and reported that Se concentrations in rice were between 0.05 and 0.3 mg kg−1.

For Se in rice grain, the difference between white and brown rice was non-significant (Fig. S6). The mean Se concentration in white rice observed in this study for US rice is nearly five times lower than the result of Williams et al. [76] who found a mean value of 0.180 mg kg−1. Mean Se content in basmati rice was significantly higher than in jasmine rice (Fig. S6). The mean Se concentration found in this study for basmati rice is supported by Naeem et al. [79] who found mean levels from 0.04 to 0.05 mg kg−1.

Strontium

Strontium is an alkali earth metal with no reported biological role in the human body [80]. The mean Sr contents detected differed significantly between Bangladeshi (0.68 mg kg−1) and US rice (0.88 mg kg−1). The mean Sr in brown rice was significantly higher than that in white rice (Fig. S7). The mean content of Sr in brown rice was 0.93 mg kg−1, which was lower than in the previous study conducted by Antoine et al. [21]. On the other hand, the mean contents of basmati and jasmine differed non-significantly.

Zinc

Zinc is an essential component of the catalytic site or sites of at least one enzyme in every enzyme classification, which has a diverse range of functions, including the synthesis of nucleic acids and specific proteins [6]. The mean Zn concentrations in rice samples from India and Pakistan were significantly lower than Zn in Thai rice. The US rice had a mean Zn concentration of 18 mg kg−1. This result is consistent with a recent study by TatahMentan et al. [60]. However, Zn concentrations in US rice were generally higher in this study than those by Shraim [30]. In this study, black and brown rice contained significantly higher Zn concentrations than white rice (Fig. 5). Similar results had been reported in US rice grown in California, Louisiana, Arkansas, and Texas [30]. They found that brown rice contained a 1.5-fold higher Zn concentration than white rice. Basmati rice contained significantly lower mean Zn than jasmine rice (Fig. 5).

Fig. 5
figure 5

Box plot of Zn concentration in rice samples based on (A) color and (B) brand. + indicates mean value; dash line indicates median value; bars represent the 1st and 3rd quartile

Full size image

Linear Discriminant Analysis

Linear discriminant analysis was performed to check whether metal(loid) profile could be used to discriminate samples according to the country of origin of the rice samples. Metal(loid)s that demonstrated significant variation among countries of origins were used as chemical descriptors in the statistical analysis to classify the different rice samples based on their geographic origin: USA, Thailand, India, Bangladesh, and Pakistan. The LDA model showed a Wilks’ λ value proximal to zero (0.00004), suggesting that group means were different [81, 82]. The first discriminant function (F1) accounted for 98.3% of the total variance, while the second discriminant function (F2) accounted for 0.93% of the total variance. Prior and posterior classifications and membership probabilities summarize the observations’ reclassification and allow one to quickly see the percentage of well-classified observations [83]. The influential standardized canonical discriminant function coefficients for F1 were Ag (1.332), Cd (0.739), and Mn (0.698) and for F2, Zn (1.046), Sr (−0.961), Cd (−0.776), Se (−0.724), and As (0.690). Previous studies adopting discriminant models found that Cu, Ag, Zn, and Cr were the critical elemental markers for determining rice geographic authentication [84, 85]. The results of the LDA showed 100% correctly classified samples. LDA indicated that all US rice samples formed a dense cohort that was well separated from the Asian rice samples (Fig. 6). On the other hand, LDA showed that Indian rice formed an overlapped cohort with Pakistani rice. Nonetheless, Thailand was properly separated from the studied Asian region.

Fig. 6
figure 6

Linear discriminant analysis scatter plot of the first two discriminant functions (F1 and F2) for rice grown in the USA, Thailand, India, Pakistan, and Bangladesh

Full size image

To check the reliability of the developed classification model, a cross-validation method was used to compute the classification and probability of rice samples [11]. According to the confusion cross-validation results, 100% of US rice samples were correctly classified. In addition, Thai rice samples were cross-validated with a higher accuracy (100%) than Pakistani (83.3%) and Bangladeshi (66.7%) (Table S4). This confirms that US-grown and Thai rice samples show unique elemental fingerprint characteristics that can be differentiated from other geographic origins. LDA struggles to group and separate the samples from India, where the accuracy is low (57.1%). As shown in Table S4, three Indian rice samples were misclassified as Pakistani rice samples. Similarly, two Bangladeshi rice samples were misclassified as Indian samples. On the other hand, one Pakistani rice sample was misclassified as an Indian sample. Therefore, for Indian/Bangladeshi rice samples, it is worth noting that the narrow variance of the metal(loid) profile and its origin can be misclassified. This can be attributed to the closeness of India, Bangladesh, and Pakistan in terms of location. Yet the total correct results of cross-validation reached 82.24% These results show that soil has a great influence on rice element composition and, consequently, geographic origin discrimination [86]. Furthermore, local climate and agricultural practices also influence the elemental composition of rice. Using LDA, it is impossible to misclassify Bangladeshi/Indian/Pakistani/Thai rice for US rice, which resolves fraud issues for rice cultivated in the USA.

Conclusion

This research reports on the metal(loid) composition of 51 rice samples from five countries on sale in Las Vegas. It was found that the concentrations of metal(loid)s in black and brown rice were statistically similar, except for Mn and Se. The concentrations of potentially toxic elements in the analyzed rice samples were below the MRL values, indicating no potential health risks associated with the consumption of rice on sale in Las Vegas. In fact, potential health risks associated with dietary intakes of toxic elements from the consumption of rice are related to both concentrations and intake patterns. For Las Vegas, intake patterns by different community groups should be studied to estimate dietary intakes of potentially toxic elements. Distinct separations between samples cultivated in the USA and those in the four Asian countries were observed using LDA models and selected metal(loid)s were identified as descriptors that could determine rice geographic origin discrimination. The confusion matrix showed that 100% of US and Thai rice samples were correctly classified. On the other hand, the distribution of metal(loid) composition of Indian and Pakistani rice showed some similarity, making it difficult to do 100% discrimination for these samples. The metal(loid) marker found in this study may provide critical insight regarding the geographic authentication of rice on sale in Las Vegas to improve food safety with potential application for other cereals following further validation.