QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Online Supporting Material Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Conway, R. E. Articles by Powell, J. J. Search for Related Content PubMed PubMed Citation Articles by Conway, R. E. Articles by Powell, J. J.

---

Methodology and Mathematical Modeling

Serum Iron Curves Can Be Used to Estimate Dietary Iron Bioavailability in Humans^1,2

Rana E. Conway^*,3, Catherine A. Geissler^*,3,4, Robert C. Hider^*, Richard P. H. Thompson^* and Jonathan J. Powell

^* The Iron Metabolism Interdisciplinary Research Group, King's College, London, UK and MRC Human Nutrition Research, Cambridge, UK

⁴ To whom correspondence should be addressed. Email: catherine.geissler{at}kcl.ac.uk.

	ABSTRACT

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Erythrocyte incorporation of isotopic iron (Fe) is the standard method for assessing iron bioavailability, but the process is expensive, technically difficult, and gives no information on the kinetics of absorption. The main objective of this study was to validate serum Fe curves as measures of dietary iron absorption because previous work demonstrated that serum iron curves can be generated with iron doses as low as 5–20 mg and that up to 20 mg iron can be added to meals without affecting relative absorption. In 3 studies, groups (n = 10, 10, 21) of Fe-deficient, mildly anemic women consumed meals of varying calculated Fe bioavailability, with and without added ferric chloride (10 mg Fe). Blood samples were collected at baseline and every 30 min for 4 h after the meal. Serum Fe concentrations were measured. Areas under the serum Fe curves and peak concentrations were used in different models to estimate Fe absorption and uptake. In 21 subjects, ⁵⁸Fe-enriched ferric chloride was added to the meals, and blood was taken 2 wk later to calculate red cell isotope incorporation. The addition of 10 mg Fe to test meals produced measurable serum iron curves even when the meal Fe bioavailability was low. Serum Fe curves were highly reproducible and were affected as expected by food composition. Even the single measurement at the estimated time of peak iron concentration was correlated significantly with erythrocyte incorporation of ⁵⁸Fe (r = 0.72, P < 0.0001). Hence the extent and rate of absorption of nonheme iron from meals, rather than in individuals, can be investigated with such subjects without the need for isotopes.

KEY WORDS: • serum iron • iron bioavailability • nonheme iron absorption • ⁵⁸Fe

Bioavailability can be defined as the degree to which a substance is absorbed or becomes available at the site of physiological activity. It is a function of both the source and the subject. In pharmacology, the standard method for measuring the bioavailability of iron compounds involves whole-body counting after dosing with radiolabeled compounds. In nutrition, the standard method for measuring the bioavailability of iron⁵ from food is the incorporation of the radioactive isotopes ⁵⁵Fe or ⁵⁹Fe, or the stable isotope ⁵⁸Fe, into hemoglobin 2 wk after consumption of a test meal (1–3). This method is technically demanding, does not measure the rate of absorption, and assumes equal utilization of iron by hemoglobin among different subjects.

In contrast to the standard method, the sequential changes in serum concentrations of Fe after an oral dose encompass only the kinetics of intestinal absorption and clearance from the blood and thus may reflect differences in the characteristics of the ingested source more than the subject. Serum iron is simple and inexpensive to measure, and if serum iron concentrations peak at a constant time after consumption, then the number of blood samples can be reduced.

In several studies, serum iron curves after ingestion of pharmaceutical preparations of Fe were compared with whole-body counting, with fair agreement (4–7). Similar validation studies were not carried out for dietary iron because this method is commonly considered too insensitive for detecting changes in serum concentrations (8). Recently, however, after the ingestion of a roll fortified with 100 mg ⁵⁵Fe-enriched iron, Hoppe et al. (9) found a high correlation between serum Fe responses and whole-body retention. Moreover, the serum iron method was used to detect iron deficiency using low dose (5–20 mg Fe) Fe supplements (10); in a small study, changes in serum iron concentrations mirrored expected bioavailability in iron-deficient subjects who consumed a fortified breakfast cereal with or without milk and/or orange juice (11).

Here, we conducted a series of studies using meals with differing predicted iron bioavailability to evaluate the use of the serum iron method for estimating dietary iron absorption. Iron-deficient, mildly anemic women were chosen because they have high iron requirements and therefore increased iron absorption and reduced subject variability. Iron bioavailability is also very important in this group. In some meals, Fe was added as ⁵⁸Fe, and the serum iron curves were compared with red cell incorporation. New information provided by this study includes the use of serum iron curves after food consumption as opposed to pharmaceutical preparations, at concentrations of iron that could be found in the diet, if fortified, and the use of normal iron as opposed to expensive isotopes.

	SUBJECTS AND METHODS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
    Subjects. Subjects were recruited from the staff and students of King's College London using posters, leaflets, and circular e-mails. Venous blood samples were taken and questionnaires were completed to select volunteers with the following inclusion criteria: female aged 18–45 y; hemoglobin 95–125 g/L; serum ferritin <20 µg/L; healthy with no history of gastrointestinal illness; taking no drugs that might affect iron absorption; ability to eat all of the proposed foods. The study was approved by King's College Research Ethics Committee (7/797). Subjects gave their informed consent to all procedures.

    Study protocol. There were 3 studies: Study 1 tested reproducibility of the serum iron curves after the addition of Fe to the test meals; Study 2 compared serum iron curves after test meals with different estimated iron bioavailability; Study 3 compared iron absorption estimated from serum iron curves with that from ⁵⁸Fe red cell incorporation.

In each study, subjects consumed 1–4 meals of different calculated iron bioavailability on separate days within a 2-wk period to reduce variation in iron status. Test meal iron bioavailability was calculated according to the method of Hallberg and Hulthen (2). Subjects fasted for 12 h before each study and abstained from alcohol for 24 h. They were also asked not to exercise on the morning of the study and to sleep well before the study because exercise and sleep deprivation may affect serum iron (12).

From 0900, the subjects rested in a seated position for 15 min, and 2 baseline venous blood samples were taken 5 min apart for serum iron determination. After the meal was consumed, blood samples were drawn at 30-min intervals for 4 h while subjects remained seated and consumed only water. Participants in Study 3 also had a further blood sample taken 14 d later for the determination of red cell ⁵⁸Fe incorporation.

The composition of the meals in the 3 studies is shown in Supplemental Table 1. In Study 1, 10 subjects each received the same high iron bioavailability meal containing 3.1 mg intrinsic iron on 3 separate occasions. On the first occasion, no Fe was added, but 10 mg Fe was added to the later meals (i.e., 13.1 mg iron in total) to ensure higher serum iron curves. Reproducibility was assessed by comparing the results of the 2nd and 3rd meals.

In Study 2, 10 other subjects each received 4 different test meals (A–D)⁶ designed to have high to low bioavailability, respectively, with 10 mg Fe added (Table 1). The least bioavailable meal (D) was given first to ensure that a serum iron curve could be detected. The other 3 meals were then given in random order.

    Test-meals (Supplemental Table 1). In Study 1, the high bioavailability test meal contained vegetable shepherd's pie made from 300 g mixed vegetables (vegetable oil, onions, mushrooms, broccoli, cauliflower, tomatoes, vegetable stock, sugar, tomato puree) and 150 g mashed potatoes (potatoes, margarine), 180 g fruit salad (orange, banana, pineapple, raspberries), and 200 mL orange juice.

In Study 2, all 4 meals contained the same vegetable shepherd's pie but with modifications as follows: meal A contained 26 g added fat (total fat 41 vs. 15 g) to determine its effect on the timing of the serum iron peak. To reduce the bioavailability of the meal in a progressive manner, the fruit salad and orange juice were replaced with wholemeal bread in meal B; meal C comprised meal B + lentils; meal D comprised meal C + tea.

In Study 3, the 11 meals contained either rice or potatoes with vegetables and one or more of white bread, wholemeal bread, cheese, yogurt, strawberries, tea, or white fish were added to provide a range of predicted relative bioavailabilities.

The potato or rice dishes were prepared in advance, frozen, and then reheated in a microwave oven on the day of consumption. Other foods (e.g., fruit salad) were freshly prepared. For the addition of 10 mg Fe, anhydrous FeCl₃ was dissolved in 2 mL deionized water; this was added drop-wise to the mixed vegetables, and the vial was rinsed with a further 2 mL water, which was also added, and the food mixed. There was no detectable change in the flavor of the meal. For the addition of ⁵⁸Fe in Study 3, ⁵⁸Fe-enriched ferric chloride was used; it contained 2–8 mg ⁵⁸Fe in the total 10 mg added Fe. Initially, 8 subjects received 2, 3, 5, or 8 mg ⁵⁸Fe (2 subjects at each concentration) in a dose-response study. Because 2 mg gave adequate red cell incorporation allowing detection of absorption of 3% of total iron, this amount was given to the remaining 13 subjects.

⁵⁸Fe was from Chemgas, isotopic composition was 93.15% ⁵⁸Fe, 0.35% ⁵⁶Fe, and 6.50% ⁵⁷Fe. It was converted to ferric chloride by being dissolved in concentrated hydrochloric acid, dried by rotary evaporation, and reconstituted with water. It was added to food as for unlabeled Fe.

    Food analysis. The vitamin C content of meals was estimated using the values from McCance and Widdowson (13) (Table 1). Iron and calcium were determined by inductively coupled plasma (ICP) optical emission spectrometry (OES) on a Leeman's DRE-ICP-OES using wavelengths of 317.33 and 259.940 nm, respectively. Food samples were digested with nitric acid and hydrogen peroxide at 200°C using an Ethos Plus Microwave (Milestone). Food samples for phytate analysis were freeze-dried and inositol phosphates (IP) extracted using concentrated hydrochloric acid; the IP were estimated by the Lehrfeld method but using phytic acid hydrolysate to permit peak identification (14). Phytate was then estimated as the sum of IP3–IP6 by HPLC because IP1 and IP 2 have little iron-binding capacity, and the inhibition of iron absorption is closely related to the sum of IP3–6 (15). Polyphenols were extracted from foods with acidified methanol (16,17) and estimated colorimetrically from food samples employing Folin-Ciocalteu reagent (18) at 765 nm. Figures for total polyphenols were adjusted to correct for the small interference by the ascorbic acid content of meals and were expressed as gallic acid equivalents (GAE).

    Iron analysis. Serum was transferred to cryovials using disposable plastic pipettes, and samples (1.5–2 mL) were frozen at –20°C on the day of collection. Serum iron was later determined using the Ferrozine colorimetric method (19) on a Cobras Mira autoanalyzer at 562 nm.

The parameters of the serum iron curves used as indicators of iron absorption were as follows: the maximum increase in serum iron concentration ( iron max), the area under the curve (AUC), and the percentage iron recovery at peak (max) and at fixed time points. These were compared with erythrocyte ⁵⁸Fe incorporation. The iron max was calculated from the difference between the mean of the 2 basal measurements and the peak serum iron value. The AUC was calculated from the 9 serum iron time points over 4 h using Simpson's rule (20). The percentage iron recovery, the percentage of intrinsic and extrinsic iron in the meal present in serum at the peak of the serum iron curve (max), or at other time points, was calculated by a method based on Henley et al. (21) Details of this method are available as Online Supporting Material.

    Red cell ⁵⁸Fe incorporation. Isotope ratios (⁵⁸Fe:⁵⁷Fe) in blood were determined by dynamic reaction cell ICP-MS (Model Elan DRC⁺, Perkin-Elmer Sciex). Blood samples were centrifuged and the red cells washed 3 times with saline, centrifuged, and stored at –70°C. Aliquots of 1 mL defrosted red cells were digested by adding 2 mL concentrated nitric acid and 2 mL hydrogen peroxide (30% wt:v) and incubating overnight at 40°C. The digest was then diluted with water to produce a solution of 4% HNO₃. The ICP-MS was calibrated using a standard natural iron solution of known ⁵⁸Fe:⁵⁷Fe ratio, checked by thermal ionization MS, to allow automatic correction of sample readings for mass discrimination. The analysis was run with a reaction gas of 5% hydrogen in argon to remove the argon oxide and hydroxide interferences. For each run, a blank of ultra-high-purity water was used, followed by the standard solution, and then the samples containing 0.2 mL digested red cells and 2.5 mL water. The standard was rerun after every 4 samples and the instrument recalibrated if necessary; nitric acid was used for rinsing between readings. The details of the calculations used to estimate red cell iron absorption are shown in the Online Supporting Material.

    Effects of extrinsic Fe on molar ratios. To confirm that, as previously shown (22,23), adding 10 mg Fe to each meal would not alter the fact that the main promoters or inhibitors of iron absorption would remain in molar excess to iron, we calculated the molar ratios of iron: phytate, polyphenols, and vitamin C for meals A–D, with and without the addition of 10 mg extrinsic Fe, assuming a value of 5 mg intrinsic iron in the diets. For phytate, the calculation was based on the assumption that all phytate was triphosphoinositol.

    Statistical analysis. Friedman's and Mann-Whitney nonparametric tests were used to test differences among meals for AUC, iron max, and % iron recovery. In addition, a repeated-measures ANOVA was used to test differences between the serum iron profiles of subjects after consumption of different meals within a single study. Bonferroni's correction was used in post hoc tests for specific comparisons. Intercorrelation between the parameters in Study 3 was analyzed using Pearson's product-moment correlation. Statistical analysis was performed using SPSS for Windows V10.0 (Statistical Package for the Social Sciences) and STATA (Stata release 7.0, 2001). Values in the text are means ± SD. Differences were considered significant at P < 0.05.

	RESULTS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
At screening, all subjects had a hemoglobin concentration of 114 ± 7.6 g/L, a serum ferritin concentration of 9 ± 3.8 µg/L, and a body weight of 66 ± 16.8 kg.

In Study 1, the peak serum iron concentration in women who consumed the high bioavailability test meal without added Fe was greater than the one at baseline (P < 0.0001), but in the women who consumed the meal to which 10 mg Fe was added, the increase from baseline was significantly greater (P < 0.0001; Fig. 1). Iron was therefore added in all subsequent experiments because some would include meals of lower bioavailability in which intrinsic iron absorption may not be easily observed. The serum iron curves in women who consumed meals with and without extrinsic iron all peaked at 180 min, suggesting that the added iron formed a common pool with intrinsic iron as proposed in previous radio-isotope studies (24,25).

View larger version (13K): [in this window] [in a new window]   FIGURE 1  Serum iron curves in women during the 4 h after consuming the same test meal with (A) and without (B) 10 mg added Fe (Study 1). Values are means ± SE, n = 10. AUC, % recovery at max, or iron max for curves without a common letter differ, P < 0.05.

The molar ratios of iron:inhibitors or promoters of iron absorption are shown in Table 2 for meals A–D. Without the addition of extrinsic Fe, there was an excess of each factor in all meals, except phytate in meal A; because this meal contained virtually no phytate, the ratio approached zero. With the addition of 10 mg extrinsic Fe, a reduced excess of all factors remained except for phytate:iron which became approximately equimolar for meals B–D, which would still be expected to affect absorption (26–28)

View larger version (12K): [in this window] [in a new window]   FIGURE 2  Serum iron curves in women during the 4 h after consuming 4 meals of decreasing predicted bioavailability (A > B > C > D) (Study 2). Values are means ± SE, n = 10. Calculated relative absorption (2) of A:B:C:D = 1:0.42:0.28:0.14, respectively. AUC, % meal iron in serum at the maximum point on the serum iron curve (% recovery at max), and increase in serum iron from baseline at the maximum point on the serum iron curve ( iron max) for curves without a common lowercase letter differ, P < 0.05.

In Study 3, iron absorption estimated from erythrocyte incorporation was significantly correlated with data from the serum iron curves (Fig. 3). In all 21 subjects, the most closely correlated was % iron recovery at max, followed by percentage iron recovery at the fixed time of 180 and 210 min. When the higher value of either 180 or 210 min was used, the correlation improved to 0.74 (P < 0.0001; data not shown). The AUC showed the lowest correlation. Regression equations for selected serum iron parameters against erythrocyte incorporation were used to calculate an equivalent erythrocyte absorption value (Supplemental Table 2).

View larger version (15K): [in this window] [in a new window]   FIGURE 3  Correlation of absorption methods: erythrocyte incorporation against parameters of serum iron (Study 3). Iron absorption estimated by 14-d incorporation of 58Fe into erythrocyte hemoglobin (Hb) (% from 58Fe incorporation into Hb). Parameters from the serum iron method are: % iron recovery at max (% meal iron in serum at the maximum point in serum iron curve) (A), % iron recovery at 180 min post iron dose (B), % iron recovery at 210 min after the iron dose (C), area under serum iron curve (D). Negative values indicate that absorption was undetectable (noise).

	DISCUSSION

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The serum iron curve produced after eating a meal of estimated high iron bioavailability showed a significant increase. However, the increase was small and likely to be unmeasurable after meals of low bioavailability (Fig. 1). It was therefore considered necessary to add extrinsic iron. The 10-mg amount was chosen because a previous study with meals using fortified breakfast cereal of similar total Fe content had given distinguishable curves (11).

It can be justifiably argued that the extrinsic iron will alter the molar ratios of inhibitors and promoters to Fe and thus alter their effect. To examine this issue, we calculated the molar ratios of phytate, polyphenols, or vitamin C to Fe in 4 meals (A–D) that cover a range from very high to very low theoretical iron bioavailability. This was done both with and without added Fe, and only the effect of phytate was at risk of being masked by the addition of this amount of Fe (Table 2). However, the calculations were based on the worst-case scenario of all iron being associated with only one of these factors that inhibited or promoted at one time, which is unlikely. Moreover, our practical results showed that the inhibitory effect of phytate could still be observed with the added iron (Fig. 2). Therefore, even with this wide range of iron bioavailability from meals, 10 mg added iron does not appear to overwhelm the action of promoters or inhibitors of iron absorption. This is in agreement with other studies (9,22).

The curves produced after meals of differing estimated relative bioavailability were clearly and significantly distinguishable, with ratios calculated from the percentage Fe recoveries at max closely comparable to the values calculated using the equations from Hallberg and Hulthen (2) from studies of radio-iron isotopes. Moreover, almost identical mean serum Fe curves were produced on 2 occasions in the same 10 subjects when we investigated reproducibility (Fig. 1).

The serum curve method did appear to show good agreement with the gold standard method of ⁵⁸Fe erythrocyte incorporation, although at least 9% iron absorption appears to be required to elicit detectable serum iron curves (i.e., +8.75 is the minimum constant in the regression equations in Supplemental Table 2) explaining some of the negative values (noise) where absorption was undetectable (Fig. 3). Indeed, each of the selected serum Fe parameters was significantly correlated with red cell incorporation.

The best parameter from the serum iron curves was the % iron recovery at max, which requires the collection of a series of serum iron samples. However, taking the greater of the values at 2 fixed time points, namely, 180 or 210 min, gave a similar level of correlation and would allow bioavailability to be fairly accurately measured from only 3 blood samples: baseline, 180, and 210 min. In certain studies, bloodletting is restricted; thus, for only 2 blood samples, baseline and 200 min may be the better choice.

In conclusion, these studies show that a simple serum iron test can be used to estimate iron bioavailability from mixed meals without the use of isotopes, at least in iron-deficient, mildly anemic subjects (for whom intersubject variability is reduced). Serum iron data provided information on both relative bioavailability and a good estimate of absolute iron absorption, depending on the number and time of blood samples taken and the calculations used. This methodology may be especially useful to compare foods or meals as opposed to absorption in individuals, where facilities are not available for isotope measurements, where measures of direct iron uptake into serum (i.e., kinetics) are required, or where there are concerns that red cell incorporation methods may not adequately reflect absorption.

	ACKNOWLEDGMENTS

 
We thank Mark Sykes for the nutrient analyses of the meals and for the ICP-MS measurements of ⁵⁸Fe, Wendy Clarke for cannulation and blood sampling of the subjects, Peter Milligan and Tom Marshall for statistical advice, and Maryanne Thompsen who provided expertise and support in the use of the DRC-ICP-MS.

	FOOTNOTES

 
¹ Supported in part by the Food Standards Agency research grant number NO5017.

² Supplementary Tables 1 and 2 and details of the calculations used to estimate RBC iron absorption are available with the online posting of this paper at www.nutrition.org.

³ These authors contributed equally to this work.

⁵ Throughout the text, Fe is the symbol used to denote extrinsic unlabelled iron, ⁵⁸Fe denotes labelled extrinsic iron and "iron" alone denotes total iron (extrinsic + intrinsic).

⁶ Abbreviations used: AUC, area under the curve; ICP, inductively coupled plasma; IP, inositol phosphate; meals A–D, meals with decreasing predicted iron bioavailability; iron max, maximum increase in serum iron concentration; OES, optical emission spectrometry; % recovery at max, 180, 210, percentage of meal iron present in serum at the maximum serum iron concentration, at 180 and 210 min.

Manuscript received 7 September 2005. Initial review completed 24 October 2005. Revision accepted 5 April 2006.

	LITERATURE CITED

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 

1. Sandstrom B, Fairweather-Tait S, Hurrell R, Van Dokkum W. Methods for studying mineral and trace element absorption in humans using stable isotopes. Nutr Res Rev. 1993;6:71–95.

2. Hallberg L, Hulthen L. Prediction of dietary iron absorption: an algorithm for calculating absorption and bioavailability of dietary iron. Am J Clin Nutr. 2000;71:1147–60.[Abstract/Free Full Text]

3. Reddy MB, Hurrell RF, Cook JD. Estimation of nonheme-iron bioavailability from meal composition. Am J Clin Nutr. 2000;71:937–43.[Abstract/Free Full Text]

4. Ekenved G, Norry A, Solvell L. Serum iron as a measure of iron absorption—studies on the correlation with total absorption. Scand J Haematol Suppl. 1976;28:31–49.[Medline]

5. Dietzfelbinger H. Bioavailability of bi- and trivalent oral iron preparations/investigations of iron absorption postabsorption serum iron concentration curves. Arzneim-Forsch. 1987;37:107–12.[Medline]

6. Kaltwasser JP, Werner E, Neichzial M. Bioavailability and therapeutic efficacy of bivalent and trivalent iron preparation. Arzneim-Forsch. 1987;37:122–9.[Medline]

7. Gonzalez H, Mendoza C, Viteri FE. Absorption of unlabeled reduced iron of small particle size from a commercial source. A method to predict absorption of unlabeled iron compounds in humans. Arch Latinoam Nutr. 2001;51:217–24.[Medline]

8. Hurrell RF. Bioavailability of iron. Eur J Clin Nutr. 1997;51: Suppl:S4–8.

9. Hoppe M, Hulthen L, Hallberg L. The validation of using serum iron increase to measure iron absorption in human subjects. Br J Nutr. 2004;92:485–8.[Medline]

10. Crosby WH, O'Neil-Cutting MA. A small-dose iron tolerance test as an indicator of mild iron deficiency. JAMA. 1984;251:1986–7.[Abstract/Free Full Text]

11. Geissler C, Agbley G. Iron absorption from fortified breakfast cereals [abstract]. Proc Nutr Soc. 1997;2:284A.

12. Eastham RD, Slade R. Clinical haematology. 7th edition. Oxford: Butterworth-Heinemann; 1992.

13. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson's The composition of foods. 5th edition. London: The Royal Society of Chemistry and Ministry of Agriculture, Fisheries and Foods;1991.

14. Burbano C, Muzquiz M, Osagie A, Ayet G, Cuadrado C. Determination of phytate and lower inositol phosphates in Spanish legumes by HPLC methodology. Food Chem. 1995;51:321–5.

15. Brune M, Rossander-Hulthen L, Hallberg L, Gleerup A, Sandberg AS. Iron absorption from bread in humans: inhibiting effect of cereal fiber, phytate and inositol phosphate with different numbers of phosphate groups. J Nutr. 1992;122:442–9.[Abstract/Free Full Text]

16. Deshpande SS, Cheryan M. Determination of phenolic compounds of dry beans using vanillin, redox and precipitation assays. J Food Sci. 1987;52:332–4.

17. Carmona A, Seidl DS, Jaffé WG. Comparison of extraction methods and assay procedures for the determination of the apparent tannin content of common beans. J Sci Food Agric. 1991;56:291–301.

18. Singleton VL, Rossi JA. Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. Am J Enol Vitic. 1965;16:144–58.[Abstract/Free Full Text]

19. Ruutu R. Determination of iron and unsaturated iron binding capacity in serum with ferrozine. Clin Chim Acta. 1975;61:229–32.[Medline]

20. Abramowitz M, Stegun IA. Handbook of mathematical functions with formulas, graphs, and mathematical tables. 9th printing. New York: Dover; 1972. p. 886

21. Henley ED, Christenson WN, Grace WJ, Wolff HG. Absorption of iron from the gastrointestinal tract: a comparative study of the oral iron tolerance test in human beings using stable and radioactive iron. Am. Clin Nutr. 1956;4:609–18.[Abstract]

22. Björn-Rasmussen E, Hallberg L, Isaksson B, Arvidsson B. Food iron absorption in man. Application of the two-pool extrinsic tag method to measure heme and nonheme iron absorption from the whole diet. J Clin Invest. 1974;53:247–55.[Medline]

23. Björn-Rasmussen E, Hallberg L, Rossander L. Absorption of ‘fortification’ iron: bioavailability in man of different samples of reduced Fe, and prediction of the effects of iron fortification. Br J Nutr. 1977;37:375–88.[Medline]

24. Björn-Rasmussen E, Hallberg L, Magnusson B, Rossander L, Svanberg B, Arvidsson B. Measurement of iron absorption from composite meals. Am J Clin Nutr. 1976;29:772–8.[Abstract/Free Full Text]

25. Cook JD, Layrisse M, Martinez-Torres C, Walker RB, Monsen E, Finch CA. Food iron absorption measured by an extrinsic tag. J Clin Invest. 1972;51:805–15.[Medline]

26. Cook JD, Reddy MB. Effect of ascorbic acid intake on nonheme-iron absorption from a complete diet. Am J Clin Nutr. 2001;73:93–8.[Abstract/Free Full Text]

27. Tuntawiroon M, Sritongkul N, Brune M, Rossander-Hulten L, Pleehachinda R, Suwanik R, Hallberg L. Dose-dependent inhibitory effect of phenolic compounds in foods on nonheme-iron absorption in men. Am J Clin Nutr. 1991;53:554–7.[Abstract/Free Full Text]

28. Gillooly M, Bothwell TH, Torrance JD, MacPhail AP, Derman DP, Bezwoda WR, Mills W, Charlton RW, Mayet F. The effects of organic acids, phytates and polyphenols on the absorption of iron from vegetables. Br J Nutr. 1991;49:331–42.

This article has been cited by other articles:

				C. Hutchinson, C. A Geissler, J. J Powell, and A. Bomford Proton pump inhibitors suppress absorption of dietary non-haem iron in hereditary haemochromatosis Gut, September 1, 2007; 56(9): 1291 - 1295. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Online Supporting Material Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Conway, R. E. Articles by Powell, J. J. Search for Related Content PubMed PubMed Citation Articles by Conway, R. E. Articles by Powell, J. J.