Introduction

Bangladeshi’s have amongst the highest consumption rates of rice globally, circa. 450 g/d for an adult (Meharg and Zhao 2012). Although rice is high in calories it is generally low in macro- and micro- trace elements, and in vitamins, leading to the population of Bangladesh being widely impacted by malnourishment (ICN2, 2014). Globally, Bangladesh has amongst the highest micro-nutrient deficiencies of any country (Bromage et al. 2016; ICN2, 2014).

Dietary dependence on rice can be considered as an exemplar of hidden hunger, i.e., where a population receives sufficient calories but the diet is deficient in essential elements and vitamins (FAO 2014; Kimmons et al. 2005; Murgia et al. 2011). With respect to essential elements, rice contributes to phosphorus, calcium, copper, iron, potassium, and zinc, amongst others (Narukawa et al. 2014). For vitamins it is B1, B3, B5, B6 (Lui et al. 2017; Kam et al.; 2012; Nur Haqim et al. 2013; Rezaei et al. 2022) and E (Aguilar-Garcia et al 2007; Goufo and Trindade 2017) where rice is thought to be an important dietary contributor. The vitamin E analogues, the oryzanols, as their name suggests, are elevated in rice (Aguilar-Garcia et al 2007; Liu et al. 2019; Xu et al. 2001). Rice consumption in Bangladesh, besides hidden hunger, is also of concern as rice is elevated in the toxicants and carcinogens arsenic and cadmium compared to other dietary staples at concentrations where ill-health effects should be observed in high consuming populations (Meharg and Zhao 2012; Narukawa et al. 2014; Shi et al. 3) were calculated for elements and vitamins that had dietary reference intakes set by the US National Academy of Science (2006), assuming a per capita rice consumption rate of 0.45 kg for Bangladesh, (Meharg and Zhao 2012) for adult males as they tend to have higher RDAs than females, with the exception of iron, where females DRI is 18 mg/d as compared to males 8 mg/d. For and adult male Bangladeshi, rice contributes significantly, in decreasing order for medians, for copper (97%), phosphorus (77%), B1 (59%), zinc (37%), iron (27%), B3 (21%), B6 (20%), potassium (10%), and poorly for E (1.2) and calcium (0.7%). The ranges in contribution to DRI are important with copper being the widest 89–177%, the only nutrient to oversupply RDA from rice consumption alone. B1 also has a very wide range, 56%, ranging from 27 to 83%. Similarly, medians for arsenic (61 μg/d) and cadmium (13 μg/d) (Table 4) also had wide ranges, 46 and 19 μg/d, respectively.

Table 3 Percentage adult male DRIs (%) for consumption of 0.45 kg/d rice calculated for elements and vitamins that had dietary reference intakes set by the US National Academy of Science (2006)
Full size table
Table 4 Daily intake or total arsenic and cadmium (µg) per day based on a per capita consumption 0.45 kg of rice per day
Full size table

Discussion

The only vitamins detected above LoD for Bangladesh market rice were B1, B3, B5, B6 and α-tocopherol (vitamin E). B1 was reported at ~ 0.4–0.7 mg/kg in two polished Chinese rice cultivars that underwent different cooking approaches (Liu et al. 2019), while a median of 1.54 mg/kg was reported here for market purchased and uncooked rice. That same Chinese study found B2 at concentrations of ~ 0.2 mg/kg, higher than our reported LoD for B2, 0.06 mg/kg, with Bangladeshi samples below this value. These higher B1 and lower B2 may be due to cultivar differences in these vitamins between studies, or due to rice processing, as the Chinese investigation was on cooked non-parboiled rice, were the Bangladesh study reported here was for uncooked parboiled rice. Liu et al. (2017) reported that there was a linear decrease in thiamine and riboflavin on milling, with up to 90 and 60% losses, respectively, when milled to 15% biomass loss. Rezaei et al. (2022) report B1, B2 and B6 in market purchased (polished) Iranian rice. All 3 vitamins varied ~ fivefold between rice types. They did not report B5, which was found to be above LoD in the current study. Here B2 was below LoD (0.06 mg/kg) using state-of-the-art triple quadrupole detection, while with their analysis the LoD was ~ 10 mg/kg, suggesting methodological differences, as they used outmoded diode array detection. Their B1 and B6 were considerably higher than the findings reported here, by at least threefold for B1 and an order of magnitude for B6. For Philippines rice, using older (non-MS) technologies, Mangel et al. (2022) reported very similar levels of B1 and B6 in polished to those reported here, finding significant differences in concentrations between cultivars. For unpolished Thai rice (Lichanporn et al. 2020), B3 was ~ 60% of that reported here, while they report B9, which we did not find above the LoD (0.36 mg/kg), again using older UV-detection technologies. B9 in Australian parboiled rice was below the limits of quantification (0.5 mg/L) (Kam et al. 2012), confirming the low levels, < LoD, reported here. B1 was reported in Malaysian uncooked polished grain at concentrations of ~ 2 mg/kg and niacin at 0.2 mg/kg or less (Nur Haqim et al. 2013). Here we report B1 at 1.5 mg/kg B2 at 8.8 mg kg 0.31 mg/kg. Thus, B1 measurements reported here are similar to Nur Haqim (2013), while B2 concentrations are an order of magnitude higher here. This may be cultivar or regent dependent, or it may be methodological. Our method used LC–MS, while theirs had spectroscopic detection. From this literature survey of B-vitamins reported in rice, there is agreement and disagreement with our current study, and this is probably due to those literature studies being based on non-mass spectroscopy detector techniques, though more work needs to be conducted to build a more comprehensive knowledge of B-vitamins in rice. Cultivar differences appear apparent, but there is little geographic information beyond our Bangladesh study reported here.

Similarly, characterization of vitamin E analogues is poor in rice, though more attention has been paid to the oryzanol analogues as they are found at high concentrations in rice bran. The polyphenol complex that represents vitamin E analogues is ~ fourfold higher in rice bran compared to brown rice flour (Aguilar-Garcia et al. 2007). The tocopherols, tocotrienols and γ–oryzanols are all thought to have antioxidant activity, and because the oryzanols dominant in rice, up to tenfold higher in concentration, they have been thought to be more important than other vitamin E analogues for rice (Aguilar-Garcia et al. 2007; Xu et al. 2001). Shobana et al. (2011) investigated the tocopherols and tocotrienols present in parboiled Indian rice that underwent various degrees of milling. The reported sum of tocopherols and tocotrienols in non-milled, non-parboiled rice was 220 mg/kg, and slightly higher, 22 mg/kg, for non-milled parboiled rice, declining to 5 mg/kg in 8% milled non-parboiled, and 4.4% milled parboiled, rice. Here we report 2.5 mg/kg for the sum of tocopherols and tocotrienols. However, it is only the 2R-stereoisomesr of α-tocopherol that can be considered as vitamin E as other analogues are not converted to α-tocopherol in the body, and have poor affinity for the liver α-tocopherol transfer protein (National Academies 2006). Furthermore, the non-vitamin E tocotrienols, tocopherols and oryzanols of rice, according to a recent overview, appear to have limited antioxidant function (Goufo and Trindade 2017). Vitamin E was reported in Thai rice ~ tenfold than the median reported here, using older analytical technologies, UV-detection (Lichanporn et al. 2020).

Peng et al. (2023) looked at nutrient element availability in paddy fields across Eastern China, and while there were no strong trends for iron, copper and manganese, zinc did appear to have higher availability in the north, but grain concentrations were not measured. Joy et al. (2017) reported a survey of a transect of wholegrain and polished rice along a N-S transect for Malawi for calcium, iron, selenium, zinc and arsenic species. They did not report any geospatial trends even though their sampling strategy had a good N-S distribution, but the lack of spatial tends may be due to the fact that their samples numbers were low; 33 and 21 for wholegrain and polished rice, respectively.

Abdu et al. (2022) and Gashu et al. (2021) surveyed a N-S transect of Ethiopian districts for nutritional content (calcium, iron, selenium and zinc) of barley, maize, sorghum, teff and wheat. They found that while all four of these elements varied significantly with respect to daily consumption, there was no clear geographical pattern, though a composite consumption for the grains considered was used. For zinc in Malawian maize (Botoman et al 2022), and other grains including maize (Gashu et al. 2021), concentrations increased by ~ 100% in the south as compared to the north. Soil factors did not explain well the spatial variation in grain zinc, though it was thought pH did contribute a small proportion of spatial effects. There was no correlation with soil organic carbon, but there did seem to be a positive association with temperature. There is a positive relationship between pH and grain maize zinc from a range of studies (Botoman et al. 2022). For rice, however across a transect from Vietnam to Spain, soil pH was strongly negatively correlated with grain zinc, and cadmium (Perera et al. 2022). The Pleistocene soils of Bangladesh are more leached and therefore of lower pH than the Holocene (Lu et al. 2009), and thus this negative correlation.

Chen et al. (2018) found a geographical trend in arsenic speciation in Chinese rice with higher percentages of DMA and lower inorganic arsenic compared to more southern provinces, but no change in sum of arsenic species. Here for Bangladesh, both DMA and inorganic arsenic decreased towards the north significantly, and for sum of species and total arsenic concentrations (graphs not shown), but there was no significant change in speciation.

Chen et al. (2018) also report that cadmium decreased significantly at northern Chinese latitudes, where the reverse was true here. Unlike Chen et al.’s (2018) study, there was a clear edaphic gradation in the current study with Holocene sediments dominating in the south and Pleistocene in the north (Lu et al. 2009). The more oxidized and weathered Pleistocene soils release more cadmium and less arsenic in paddy systems (Meharg and Zhao 2012), while the reverses is true for the less oxidized and less weathered Holocene soils, and this explains the different opposite trends in cadmium and arsenic observed for Bangladesh. Cadmium did not show any strong geographical trends, rather having country-specific elevation (Meharg et al. 2022; Shi et al. 2020). Otero et al. (2020) found that inorganic arsenic concentrations did not differ between South America and Spain. Al-Rmalli et al. (2012) found a mean of 0.037 mg/kg for raw rice in an eastern Bangladesh survey (Dhaka and Sylhet regions), not dissimilar to the mean of 0.032 for parboiled rice reported here.

No geographical trends were observed for arsenic species from paddy rice surveyed from southern Spain through to southern France (Signes-Pastor et al. 2016). Mu et al. (2019) surveyed total arsenic and cadmium in rice grain across the eastern growing band of China. They did not show any trend in total arsenic but found that cadmium was higher in grain in the south, which they attributed to higher soil cadmium availability rather than to total soil cadmium concentration. Characterization of a wide range of grain nutrient elements and arsenic and cadmium, from East Africa, South America, Asia and Europe found strong regional differences in grain elemental profile, but no strong north–south relationships with the exception of grain inorganic arsenic (Meharg et al. 2022), as already shown by Carey et al. (2020).

PCA analysis found that there may be potential synergistic and antagonistic interactions were observed between vitamins and elements, but there is no literature on this that we could find, unlike previous identified interactions within trace elements for Bangladeshi grown rice (Williams et al. 2009). Also, climatic and soil gradients would need to be considered in this context (Williams et al. 2009). How grain vitamin and elemental chemistry interrelate with soils and climatic gradients was beyond the scope of the present study and is identified as a key area for future investigation.

To conclude, this current study found strong geographical gradients in essential elements, vitamins and toxic elements across a north–south and east–west gradient for the western half of Bangladesh. Thus, any strategies to counteract malnutrition or toxic element exposure of those on a rice subsistence diet within this populace (Bromage et al. 2016; FAO 2014; ICN2, 2014; Kimmons et al. 2005; Murgia et al. 2011) must consider the regional differences identified here for Bangladesh market rice.