Mineral nutrient composition of vegetables, fruits and grains: The context of reports of apparent historical declines
= 0.021 to <0 .001="" and="" em="" mg="" style="box-sizing: border-box; margin: 0px; padding: 0px;">P = 0.030 to =0.004), which coincided with the introduction of semi-dwarf, high-yielding cultivars. With regard to the hypothesis that soil nutrient levels are a causative factor, they found that the mineral concentrations in the archived soil samples either increased or remained stable. Reasons for this included inputs of Mg from inorganic fertilizer, Zn and Cu from farm yard manure, and Zn also from atmospheric deposition. The observed decreases in wheat grain mineral content were independent of whether the crop received no fertilizers, inorganic fertilizers or organic manure. Multiple regression analyses showed that the two highly significant factors associated with the downward trend in grain mineral concentration were increasing yield and harvest index (i.e., the weight of the harvested product, such as grain, as a percentage of the total plant weight of the crop, which for wheat was measured as the aboveground biomass due to the difficulty of obtaining the root biomass).
Fan et al. (2008b) noted that the Se concentration of the grain had a much larger range and was significantly higher (P < 0.001) in unfertilized plots compared to inorganic fertilizer or manure treated plots and higher in the unfertilized plots in periods before 1920 or after 1970 than during 1920–1970. These temporal and fertilizer-related patterns of Se decrease in the grain were influenced mainly by sulfur (S) inputs from fertilizers and atmospheric deposition of S, which increasedsulfate antagonism of selenium uptake, plus a small dilution effect. For these reasons, despite the observed long-term trend (not statistically significant) of an increase in soil Se concentration, primarily due to atmospheric Se deposition, the grain Se content did not increase.
Thus, the findings of Fan et al. from the Broadbalk Wheat Experiment are conclusive with regard to the lack of significant historical decreases in soil mineral levels in the fields they studied and that verified declines in mineral nutrient concentrations in wheat grain were associated with varieties having an increased grain yield. Nevertheless, it is still worthwhile exploring what role other potential causative factors could play in “apparent” historical mineral nutrient declines in vegetables and fruits.
5. Field trials to test hypotheses regarding historical mineral nutrient changes
McGrath (1985) used field experiments to look at the “dilution effect” of increased yields on the mineral nutrient concentrations in grain of winter wheat. He noted that concentrations of P, K, sulfur (S), Ca and Mg varied twofold (n = 238); Fe, Zn and Cu varied threefold (n = 236); and Mn varied by a factor of 5 (n = 236). While the potential for decreases was predictable, e.g., for Zn and Fe which move slowly from the soil into plant roots and thus might not meet the demand of a rapidly-growing crop, a decrease in these minerals with increased yield was not found in these analyses. There were small, statistically significant (P < 0.01) varietal differences but they were not large enough to be of agricultural importance; the overall changes in levels in crops with increased yields were positive, except for Mn which did not change.
Farnham et al. (2000) examined variations in Ca and Mg concentrations in a USDA collection of 19 inbred and 27 commercial F1 hybrids of broccoli grown side-by-side. Broccoli was chosen because it is a good vegetable source of Ca and Mg, and the bioavailability of Ca from broccoli is comparable to that from milk. Levels varied for Ca (1.99–4.35 mg/g dry weight) and Mg (1.94–3.74 mg/g dry weight) among the hybrids within the same growth year due to genetic differences—concentrations were significantly negatively correlated (P < 0.05) with broccoli head weight due to greater density, not size. With the inbred lines, the concentration of Mg, but not of Ca, was negatively correlated (P < 0.05) with head density. These mineral nutrient content differences correlated with head density are examples of adilution effect. However, there was also a significant (P < 0.05) environmental effect on both Ca and Mg concentrations when comparing two different growing seasons (1996 and 1997). Environmental and genotype-by-environment components of variance for Ca concentration were equal and both were ten times greater than the genotypic component of variance, while for Mg the environmental, genotypic and genotype-by-environment components of variance were of a similar magnitude.
Garvin et al. (2006) also used two replicated field trials rather than published data to examine historical shifts in mineral micronutrient concentration (Fe, Zn, Cu and Se) in 14 different varieties of U.S. hard red winter wheat from production eras ranging from 1873 to the late 1990s. They found significant effects on micronutrient content of cultivation location (Fe, Cu, P and Zn with P < 0.001, Se was not significantly different) and significant differences between the genotypes (P < 0.001), whose genetic profiles differ due to more than a century of crop development. When the data was organized by release date and yield, Zn content was seen to have decreased significantly with both increasing yield and more recent variety release date at both locations (P < 0.0001 and P < 0.05); Fe content decreased significantly with increasing yield and more recent variety release date at one location (P < 0.05); and Se content decreased significantly with more recent release date at one location (P < 0.01). With regard to Cu content, it was lower in grain from one site compared to the other (P < 0.001) but there was no correlation with variety release date.
Murphy et al. (2008) used a randomized complete block design nursery to grow 56 historical spring wheat cultivars widely grown in the Pacific Northwest region of the U.S.A. from 1842 to 1965, and 7 modern spring wheat cultivars widely grown in Washington State in 2003. Thirty-seven cultivars were in the soft white wheat market class, 20 were hard red, four were hard white and two were soft red. There were three replicates of each cultivar in one growing season and four replicates of each the following year. Yield and concentrations of Ca, Cu, Fe, Mg, Mn, P, Se and Zn were measured. They found that the modern cultivars had higher yields than the historical cultivars (P < 0.0001). The historical cultivars had significantly higher grain mineral concentrations than the modern cultivars: Cu (P < 0.001), Fe (P < 0.01), Mg (P < 0.001), Mn (P < 0.05), P (P < 0.001), Se (P < 0.05) and Zn (P < 0.001), the exception being Ca for which the decline in modern cultivars was not statistically significant (P = 0.07). There were highly significant variations in the concentrations of each mineral between cultivars (P < 0.0001) and a significant genotype-by-year interaction for each mineral as well, although statistical analyses showed that most of the variation was due to genotype rather than year. Overall yield was negatively correlated with mineral concentration for Ca (P < 0.001), Cu (P < 0.001), Mg (P < 0.001), Mn (P < 0.01), P (P < 0.001) and Se (P < 0.001), but not significantly for Fe and Zn. These results can be compared to those of Fan et al., 2008a, Fan et al., 2008b described above, who reported that levels of Cu, Mg, Fe and Zn were steady in cultivars with release years from 1845 to the mid-1960s and then declined significantly in more modern semi-dwarf cultivars with high yields. Murphy et al. postulated that the lack of a negative correlation between yield and concentrations of Fe and Zn in their analyses may be because the content of these two mineral nutrients is influenced by the high protein gene Gpc-B1 which may be subject to positive selection pressure for higher yield where protein content is a consideration. Regressions of mineral concentrations on year of cultivar release, separated into market classes, showed significant decreases among soft white cultivars for all minerals except Ca and Mg. However, among hard red cultivars, only Zn decreased with release date whereas Mg increased slightly over time. All other mineral nutrients remained stable among the hard red cultivars released over the past 120 years. Murphy et al. suggested that the decline in concentrations of most minerals in soft wheat might be due to selection pressure for lower ash content, since high ash content in flour gives a darker colour to finished products which is undesirable with regard to product quality. However, they noted that generally the correlations were weak and exceptions existed for high yielding cultivars with moderately high levels of certain minerals, such as P, Fe, Mg, Mn and Se, indicating that there is genetic potential for development of cultivars with high mineral nutrient levels, particularly for Cu, Zn and Mn.
Ficco et al. (2009) used side-by-side cultivation in two locations and two growing seasons to study mineral nutrient (Ca, K, Mg, Mn, Na, Cu, Fe, and Zn) and phytatelevels in Italian durum wheat cultivars. They studied 10 old genotypes released between 1900 and 1973, 58 cultivars released after 1974 that carried semi-dwarfing reduced height Rht genes, and 17 advanced breeding lines with high yield potential. They noticed a direct soil content effect on levels of Na and K in the grain, and at one site a higher soil level but lower grain level of Mn. Of the two genotypes with the highest grain Fe content, one was a modern genotype and one an old genotype; for Cu the modern genotypes had a higher content. For inorganic P, Cu, Fe, Na and Zn, the modern genotypes had the widest ranges. No clear trends for historical declines in mineral nutrient composition were observed comparing modern genotypes and advanced breeding lines with old genotypes. Their results suggested a significant dilution effect only for Mg and Zn (both P < 0.001) and not for Fe.
Rosanoff (2013) combined analytically determined Mg food content change results such as those of Fan et al. (2008a), Murphy et al. (2008) and Ficco et al. (2009), with a comparison of the Mg content listings in food composition tables of the U.K., USDA, and Health Canada from different publication dates, to conclude that a historical decrease in the Mg content of grains, fruits and vegetables has occurred. From these data sources she derived estimates that grain Mg concentrations have dropped by 7 to 25% and vegetable Mg concentrations have dropped by 15 to 35%. She associated her calculations of food Mg content and food supply data from the USDA with rates of cardiovascular disease (CVD) mortality data from the U.S. NIH. Rosanoff concluded that the results suggested a causal relationship between CVD mortality peaking in the U.S. in 1968 when Mg in the U.S. food supply reached its nadir and then gradually declining as food Mg supply rose in the years up to the present. She recognized that the decline in CVD mortality can be explained mostly by medical treatments and medications, increased exercise and decreased smoking. However, she drew a parallel between the U.S. trend and the rise of rates of CVD mortality, obesity, metabolic syndrome and non-communicable diseases in societies transitioning from traditional diets to modern processed food diets. Rosanoff’s main conclusion was that rising global mortality from CVD may be due to lower dietary intakes of Mg (and other nutrients) caused by declining crop content, which she attributed primarily to the change to high-yield varieties, and also to food processing losses. While the Institute of Medicine (1997) has recognized that Mg is a required cofactor for over 300 enzyme systems and that Mg depletion is linked to CVD, neuromuscular diseases, diabetes mellitus, and renal wasting syndromes, Rosanoff’s use of data from food composition tables published in different years as supporting evidence for a historical decline in food Mg content is not valid. The linking of CVD mortality rates to postulated historical declines in the Mg content of foods is perhaps oversimplified.
6. Analysis and discussion
Davis (2009) provided a summary and reanalysis of the scientific evidence available up to the time of writing regarding apparent historical decreases in fruit and vegetable nutrient composition and its potential causes. This reanalysis involved calculating the ratios (R) and distribution-independent 95% Confidence Intervals of the nutrient content between new/old varieties of the food using the nonparametric approach of testing the null hypothesis that ratios of group medians equaled 1. Davis preferred a nonparametric approach that provides more conservative results over the statistical approach used in previous studies. Calculating group geometric means and especially the use of a t-test (the statistical approach used in the articles by Mayer, 1997; and by White and Broadley, 2005) was determined to be insufficient to account for the skew of mineral nutrient analysis data, which were shown to deviate significantly from a normal distribution (Davis et al., 2004, Davis, 2006). Of the 33 median R values Davis (2009) recalculated, only 11 (33%) of them indicated a statistically significant (P < 0.05) apparent nutrient content decline (i.e., R < 1). No statistically significant increases in mineral nutrient content (i.e., R > 1) were observed. Among the statistically significant ratios (R), the most pronounced apparent declines in mineral nutrients were seen in vegetables. They ranged from approximately 80% for Cu (questionably large but strongly subject to the dilution effect) to approximately 17% for Ca. The decline in Na appeared to be about 40% and the decline in Mg appeared to be about 23%. A statistically significant decline in Fe in vegetables was seen only in U.S. data for a larger group of vegetables (Davis, 2009; Fig. 6). For the content of P in vegetables, the ratio of medians showed a small but statistically significant apparent decline in U.S. data for a large group of vegetables (Davis, 2009; Fig. 3) but no significant change in U.K. data for mixed crops including vegetables, fruits and nuts (Davis, 2009; Fig. 6). In fruits, apparent declines in the median mineral nutrient content were relatively small and not statistically significant (P > 0.05, Davis, 2009; Figs. 2 and 5).
Food composition tables and databases, such as Health Canada’s Canadian Nutrient File (2015a), the U.S. Department of Agriculture’s National Nutrient Database for Standard Reference, Release 28 (SR28) (2015), Public Health England’s Composition of Foods Integrated Dataset (2015), the FSANZ Nutrient Tables for Use in Australia (2015), and the Food and Agriculture Organization (FAO) of the United NationsInternational Network of Food Data Systems Food Composition Databases (2015), provide the foundations for the development of educational programmes on choosing healthy diets that help consumers to make informed choices with regard to thenutritional quality of foods. These databases also provide the basis for assessing population nutrient intake in combination with food intake surveys. However, comparison of historical food composition tables is not a reliable way to determine changes in nutrient composition of foods over time. Their data represent snapshots of nutrient content for foods available on the market at a particular time. There are changes in the genetic varieties of crops on the market over time, large ranges of variation in content of different nutrients from variety to variety of the same crop, and differences in geographic origin, season, degree of ripeness, sample sizes, sampling methods, analytical methods, statistical methods, etc.
With regard to analytical methods, the levels of minerals for most foods in SR28 (USDA, 2016a) were determined by AOAC Official Methods of Analysis, such asinductively coupled plasma − emission spectrophotometry (AOAC 984.27) for Ca, Fe, Mg, P, Na, K, Zn, Cu and Mn, a method that was the subject of Final Action consideration by AOAC in 1986. Preparation involves digesting the test samples in HNO3/HClO4 (a wet oxidation process). For some records in SR28, minerals except for P were determined by an atomic absorption method (AOAC 985.35, Revised First Action 1997) for which samples are prepared by dry ashing in a muffle furnace at 525 °C; determination of P in these cases was by a colorimetric method (AOAC 2.019, 2.095 and 7.098, published in AOAC (1980) but involving a method dating from 1957) with sample preparation by wet ashing. This colorimetric method was originally developed for determination of total P in fertilizers; it is listed as AOAC 957.02 in the current Official Methods database (AOAC International, 2016) and the latest revision (in 1998) for colorimetric determination of total P in foods is AOAC 995.11. Additional details on the mineral nutrient analytical methods used by the USDA are available from the SR28 Documentation and User Guide (USDA, 2016a).
Much of the food mineral nutrient values in the Canadian Nutrient File are derived from the SR database (Health Canada, 2015b) so the data was obtained mostly through AOAC Official Methods for atomic absorption (AOAC 985.35) or ICP-ES (AOAC 984.27). One exception is for data from the Canadian Sampling and Nutrient Analysis Program, which were obtained by an inductively coupled plasma/mass spectrometry (ICP-MS) method based on EPA 3051A (U.S. EPA, 2007). With regard to sample preparation, EPA 3051A uses microwave assisted (instead of conventionally heated) digestion with HNO3 or a combination of HNO3 and HCl.
As Mayer (1997) noted, the analytical methods used for the U.K. Composition of Foodsdatabase similarly have evolved and there is no clear conclusion on the extent to which that has influenced reported differences in mineral nutrient composition. The current seventh edition of the Composition of Foods database includes data for many foods analyzed in surveys carried out since the publication of the sixth edition in 2002 and updates where necessary from industry sources, other food composition datasets, and the scientific literature. In the case of carry-over of previous values, they were reviewed by members of the project team to ensure they are still representative of foods currently consumed or else calculations were done to update values. Editions prior to the sixth included data for foods both with and without inedible material or material that may be discarded as inedible by some consumers but in the seventh edition all nutrient values apply to the edible portion as specified in the food name, with edible conversion factors provided in an appendix. Thus, the data reflect changes in food preparation methods, advances in analytical methods, analytical variation, natural ranges of nutrient composition variation, and new varieties of plant foods (Roe et al., 2015).
Attributing different values of mineral nutrient levels in a given food to actual changes in the vegetable, fruit or grain composition over time, when the data are coming from different editions or sources of food composition databases, is inherently a flawed approach since official methods for the analysis of the mineral nutrient content of food have changed over time as the science advances. Some of the very early wet chemistry methods that would have been used in old publications include the 1928 titrimetric method for Al and Fe in plants (AOAC 928.03), 1935 colorimetric method for P in fruits (AOAC 935.45), and 1937 colorimetric method for Fe in plants (AOAC 937.03). Later improvements to these methods such as the 1970 spectrophotometric molybdovanadate method for P in fruits (AOAC 970.39), 1970 gravimetric quinolone molybdate method for P in fruits (AOAC 970.40), and dry ashing/sodium molybdate modifications to the colorimetric measurement of P in method AOAC 995.11 (Pulliainen and Wallin, 1994, Pulliainen and Wallin, 1996) allowed for the ongoing use of less expensive but still reliable methods with less toxic reagents. However, the application of modern chemical analytical instrumentation to determine the composition of mineral nutrients in plant-derived foods started with such methods as the 1975 atomic absorption spectrophotometric method AOAC 975.03 for Ca, Cu, Fe, Mg, Mn, K and Zn in plants, a 1980 direct reading spectrographic method AOAC 980.03 for metals in plants, the 1984 ICP-ES method mentioned above, a 1999 variant on the atomic absorption method (AOAC 999.11) for Pb, Cd, Cu, Fe and Zn in foods, and in 2015 an ICP-MS method AOAC 2015.06 for minerals and trace elements in infant formula and adult/pediatric nutritional formula was published (AOAC International, 2016).
The FDA has also published its own Elemental Analysis Manual for Food and Related Products with methods such as ICP-Atomic Emission Spectrometry with microwave assisted digestion (U.S. FDA, 2010) for the determination of 22 nutritive and toxic minerals in foods and an ICP-MS method with microwave assisted digestion (U.S. FDA, 2015) for Cr, Mn, Ni, Cu, Zn, As, Se, Mo, Cd, Hg and Pb levels in foods.
It is important to keep in mind that not all of the data in the food composition databases comes from laboratories using official methods of analysis; in part they are derived from articles in the peer-reviewed scientific literature for which researchers used alternate methods. Some of the more recent developments for plant-derived food mineral nutrient analysis include Neutron Activation Analysis, which does not involve chemical preparation techniques (e.g., Baidoo et al., 2014), visual-near-infrared spectroscopy, to a lesser extent mid-infrared spectroscopy andultraviolet spectroscopy, chlorophyll a fluorescence, X-ray fluorescence, and laser-induced breakdown spectroscopy (Mir-Marqués et al., 2016, van Maarschalkerweerd and Husted, 2015, Schmitt et al., 2014).
Castanheira et al. (2016) have published European Food Information Resource (EuroFIR) guidelines for the assessment of methods of analysis and proficiency testing with regard to the quality of data to be entered into food composition databases. They note that for some nutrients values from different food composition tables are not comparable mainly due to differences in analytical procedures. Prioritized nutrients for methodological guidance include the minerals and trace elements: Ca, Cu, I, Fe, Mg, Mn, P, K, Se, and Na. They recommended the use of only AOAC, European Committee for Standardization (CEN) and International Organization for Standardization (ISO) methods of analysis for minerals and trace elements. The methods considered appropriate were grouped into ICP-MS for trace elements (Se, I, Zn, Mn), ICP-Optical Emission Spectrometry (OES is also known as atomic emission spectrometry) for minerals present in higher quantities in foods (Fe, K, Na, P, Cu, Ca, Mn), atomic absorption spectrometry (AAS has similar performance to ICP-OES, which is more expensive but has largely replaced AAS), and graphite furnace atomic absorption spectroscopyfor Se, although that has largely been replaced by ICP-MS.
Regarding methods of sample preparation, Castanheira et al. (2016) note that extraction or destruction of organic matter before measurement of minerals and trace elements is generally required but this is a slow process and large sources of contamination can occur. Currently, food organic matrix destruction/removal or extraction is conducted through dry ash, wet digestion or pressure digestion procedures which are available at CEN and AOAC. However, each food matrix may require a different strategy and optimization for complex matrices to separate inorganic from organic components.
The laboratory mineral nutrient analyses of wheat grain varieties and soil samples archived over the last 160 years by the Broadbalk Wheat Experiment (Fan et al., 2008a, Fan et al., 2008b), completed with identical sample preparation and analytical methods, have helped to demonstrate that historical declines in mineral nutrient content of food crops can be real but that these changes are correlated with increased yield and harvest index, not soil mineral content. This was borne out by other comparisons of historical food composition data (grouped into vegetables, fruit or grains rather than single food comparisons, and adjusted for moisture content) with mineral nutrient analyses from side-by-side plantings of low- and high-yield cultivars, and fertilization studies. These studies have all demonstrated consistent negative correlations between yield and concentrations of mineral nutrients. In fruits, vegetables and grains, usually 80% to 90% of the dry weight yield is carbohydrate so when breeders select for high yield they may be selecting mostly for an increase in carbohydrate with no assurance that other nutrients will increase proportionately.
The Organisation for Economic Co-operation and Development (OECD), an organisation of 34 countries whose mission is to promote policies that will improve the economic and social well-being of people around the world, publishes international Consensus Documents on compositional considerations for new varieties of crops. At the time of writing, for evaluating the safety of novel foods and feeds there are OECD consensus documents on 18 plant and 2 mushroom crops. These documents are authoritative sources of information on the natural range ofmacro- and micro-nutrient content in staple food crops, providing necessary contextual information for assessing whether a new variety of a crop, grown in various locations and conditions, is likely to be as nutritious, more, or less nutritious than conventional varieties of the same crop.
For example, regarding bread wheat, the consensus (OECD, 2003) is that “the average mineral content of a given wheat grain varies significantly from one part of the world to another. This appears to be a function of a number of factors, including the wheat variety, the growing and soil conditions, and fertilizer application. The mineral composition of wheat has more to do with environmental conditions, rather than varietal characteristics.” Davis et al. (1984), whose work is cited in OECD (2003), determined the mean and range of Ca, Mg, P, K, Cu, Fe, Mn, and Zn content, and also chromium (Cr) and Se, in statistically valid numbers of samples of wheat of the following classes (each of which has multiple varieties; 231 varieties in total were analyzed, from 49 growing locations): hard red winter, hard red spring, soft red winter, durum, soft white winter, soft white spring, hard white winter and hard white spring.
Results from Davis et al. (1984) and other authors for selected grains, vegetables and fruits are presented in Table 1 to illustrate how widely mineral nutrient content varies among samples and varieties of a single crop species.
Table 1. Ranges of mineral nutrient content in examples of grains, vegetables and fruits.
a
- Range units are mg/100 g dry weight; percent as maximum/minimum × 100%; n/a: not available.
- b
- n values are not available from OECD consensus document tables since they are compilations of published data.
Data sources: 1 Davis et al. (1984); 2 Organisation for Economic Cooperation and Development OECD (2016);3 OECD (2002); 4 OECD (2004); 5 OECD (2015); 6 OECD (2012); 7 OECD (2010a); 8 Farnham et al. (2000); 9OECD (2008); 10 OECD (2010b).
Eng / Jamal h Jahlan
ليست هناك تعليقات:
إرسال تعليق