Bioavailability of Iron in Cottonseed Meal, Ferric Sulfate, and Two Ferrous Sulfate By-Products of the Galvanizing Industry1 STEPHANIE D. BOLING, HARDY M. EDWARDS, III, JASON L. EMMERT, ROBERT R. BIEHL, and DAVID H. BAKER2 Department of Animal Sciences, University of Illinois, Urbana, Illinois 61801 ABSTRACT Iron depletion-repletion assays were carried out with young chicks to establish Fe bioavailability values for Fe2(SO4)3·7H2O (22.7% Fe), FeZnSO4·H2O (20.2% Fe, 13.0% Zn), Zn-FeSO4·H2O (20.2% Zn, 14.2% Fe), and cottonseed meal (200 mg Fe/kg). Standard hemoglobin response curves were established using feed-grade FeSO4·H2O (28.8% Fe) or reagent-grade FeSO4·7H2O (20.1% Fe) as standards such that relative bioavailability (RBV) could be assessed for the experimental sources of Fe. Weight gain, hemoglobin, and hematocrit responded linearly (P < 0.05) to Fe supplementation in all assays. Using hemoglobin as the response criterion, slope-ratio calculations established Fe RBV values of 126% for Fe-ZnSO4·H2O and 93% for ZnFeSO4·H2O. The 126% value for Fe-ZnSO4·H2O was greater (P < 0.05) than the FeSO4·H2O standard (100%), but the 93% value for Zn-FeSO4·H2O was not different
(P > 0.10) from the standard. However, evaluation of all criteria of response (hemoglobin, hematocrit, weight gain) suggested that neither Fe-ZnSO4·H2O nor ZnFeSO 4 ·H 2 O had different Fe RBV values than FeSO4·H2O. Standard-curve calculations were used for assessment of Fe RBV in Fe2(SO4)3·7H2O and cottonseed meal, as only a single level of Fe addition was studied for each of these products. Iron RBV in Fe2(SO4)3·7H2O was estimated to be 37%, whereas Fe RBV in cottonseed meal was found to be 56%. Both of these values were lower (P < 0.05) than the FeSO4 standard. The data suggest that the two new products, representing combinations of FeSO4·H2O and ZnSO4·H2O by-products of the galvanizing industry, are excellent sources of bioavailable Fe, whereas ferric sulfate and cottonseed meal are relatively poor sources of usable Fe.
(Key words: iron, bioavailability, ferric sulfate, ferrous sulfate, cottonseed meal) 1998 Poultry Science 77:1388–1392
INTRODUCTION Two new sources of supplemental Fe and Zn have emerged, and questions have been raised concerning Fe and Zn bioavailability in these products. These products are popularly referred to as iron-zinc sulfate and zinciron sulfate.3 Both products are by-products of the galvanizing industry; FeSO4 and ZnSO4 are crystallized from the pickling fluid that is used for cleaning the scale from galvanized steel. After drying, two Fe-Zn products result, both of which contain varying proportions of FeSO4·H2O and ZnSO4·H2O. All of the Fe in these products is ferrous (+2) Fe, but oxidation of ferrous sulfate to ferric sulfate can occur under certain conditions. Thus, we evaluated Fe bioavailability in the two Fe-Zn products as well as in reagent-grade ferric sulfate,
Received for publication December 23, 1997. Accepted for publication April 14, 1998. 1Supported in part by Southeastern Minerals, Bainbridge, GA 31717, and Nutri-Flo Co., Sioux City, IA 51107. 2To whom correspondence should be addressed: 290 ASL, 1207 West Gregory Drive, Urbana, IL 61801;
[email protected] 3Nutri-Flo Co., Sioux City, IA 51107.
i.e., Fe2(SO4)3·7H2O. Also, because Fe bioavailability in cottonseed meal is not known for poultry (Henry and Miller, 1995), this product was evaluated for Fe bioavailability. Hemoglobin depletion-repletion assays were carried out with young chicks fed graded doses of the Fe sources under investigation. Linear hemoglobin responses were evaluated for each experimental source of Fe, and these were compared to reagent-grade FeSO 4 ·7H 2 O or feed-grade FeSO 4 ·H 2 O standard response curves.
MATERIALS AND METHODS Male chicks resulting from the cross of New Hampshire males and Columbian females were used for all experiments. Chicks were fed an Fe-deficient caseindextrose basal diet (Biehl et al., 1997) analyzed to contain 20 mg Fe/kg (Table 1) for the first 7 d (Assays 1 and 2) or 8 d (Assay 3) posthatching to deplete Fe stores. Following an overnight period without feed, birds were
1388
Abbreviation Key: RBV = relative bioavailability.
IRON BIOAVAILABILITY TABLE 1. Composition (as-fed basis) of iron-deficient basal diet1 Ingredient
Percentage
Dextrose Casein (85% CP) Mineral mix (Fe-free)2 Soybean oil NaHCO3 Glycine L-Arginine DL-Methionine Vitamin mix3 Choline chloride DL-a-tocopheryl acetate (20 mg/kg) Ethoxyquin (125 mg/kg)
to 100.00 20.00 5.38 3.00 1.00 2.00 1.00 0.50 0.20 0.20 + +
1Contained
(by analysis) 20 mg Fe/kg. mix provided the following (per kilogram of diet): CaCO3, 22.00 g; KH2PO4, 19.00 g; KHCO3, 10.24 g; NaCl, 8.8 g; CuSO4·5H2O, .02 g; ZnCO3, 0.10 g; MgSO4·7H2O, 3.5 g; MnSO4·H2O, 0.65 g; H3BO3, 9 mg; Na2MoO4·2H2O, 9 mg; KI, 40 mg; CoSO4·7H2O, 1 mg; Na2SeO3, 0.2 mg. 3Vitamin mix provided the following (per kilogram of diet): thiamin·HCl, 20 mg; niacin, 50 mg; riboflavin, 10 mg; D-calcium pantothenate, 30 mg; vitamin B12, 40 mg; pyridoxine·HCl, 6 mg; Dbiotin, 0.6 mg; folic acid, 4 mg; ascorbic acid, 250 mg; retinyl acetate, 1,789 mg; cholecalciferol, 15 mg; menadione dimethylpyrimidinol bisulfite, 2 mg. 2Mineral
weighed, wing-banded, and randomly assigned to experimental diets. Each group had a similar mean initial weight. Chicks were housed in thermostatically controlled stainless-steel starter batteries with raised wire floors. In order to minimize Fe contamination, stainless-steel waterers, feeders, and floor grids were used. Feed and deionized, distilled water were provided for ad libitum consumption. A constant 24-h lighting schedule was maintained throughout the studies, which were conducted until Day 21 or 22 posthatching. At the termination of each experiment, birds were gassed lightly with CO2 and then bled by cardiac puncture with heparinized syringes equipped with stainless-steel needles. Blood samples were pooled by replicate and analyzed for blood hemoglobin and blood hematocrit. Previous research had established that young chicks fed semi-purified diets based on casein will show linear hemoglobin responses to supplemental Fe doses up to at least 20 mg Fe/kg (Aoyagi and Baker, 1995; Biehl et al., 1997). The basal diet (Table 1) was formulated to meet or exceed all NRC (1994) nutrient recommendations, with the exception of Fe. Dietary additions were made at the expense of dextrose. The basal diet, cottonseed meal, and all experimental and standard inorganic Fe sources were analyzed for Fe by flame atomic absorption spectrophotometry 4 following HNO3 wet ashing (Aoyagi and Baker, 1995). Four replications of two birds were fed the experimental diets in Assay 1. Reagent-grade FeSO4·7H2O additions were made to the basal diet to provide 0, 10 or 20 mg Fe/kg. As it is possible for ferrous sulfate to be 4Model
306, Perkin-Elmer Corp., Norwalk, CT 06850.
1389
oxidized to ferric sulfate, we were interested in the bioavailability of ferric sulfate in chicks. Therefore, reagent-grade Fe2(SO4)3·7H2O was added to the basal diet to provide 20 mg Fe/kg in order to determine the bioavailability of ferric sulfate. In Assay 2, five replications of five chicks were utilized. Dietary additions of feed-grade FeSO4·H2O were made to provide 0, 9.4 or 18.8 mg Fe/kg to establish a standard curve. Previous research had established that feed-grade ferrous sulfate and analytical-grade ferrous sulfate were equivalent in bioavailability (Henry and Miller, 1995). Two dietary inclusions of feed-grade Zn-FeSO4·H2O (14.2% Fe, 20.2% Zn) were made to supply 10.3 or 20.7 mg Fe/kg. We also included a treatment that provided 20 mg Fe/kg from cottonseed meal so that the bioavailability of Fe in cottonseed meal could be determined. Assay 3 was conducted with five replications of five chicks. Dietary additions were made to the Fe-deficient casein-dextrose diet to provide 0, 10 or 20 mg Fe/kg. Supplemental Fe sources were either feed-grade FeSO4·H2O (the standard) or the Fe-ZnSO4·H2O product (20.2% Fe, 13.0% Zn). All data were analyzed using the General Linear Models procedure of SAS (SAS Institute, 1985). Iron bioavailability was determined using FeSO4·H2O as a standard by means of common-intercept multiple linear regression (slope-ratio), and, in some cases, by standardcurve methodology (Wedekind and Baker, 1990; Biehl et al., 1997). The multiple linear regression model consisted of two straight lines with a common intercept. Blood hemoglobin, blood hematocrit, and weight gain were the dependent variables and were regressed on supplemental Fe intake.
RESULTS AND DISCUSSION Weight gain, blood hemoglobin, and blood hematocrit responded linearly (P < 0.05) to supplemental Fe from the standard FeSO4·7H2O in Assay 1 (Table 2). Blood hemoglobin (Y in grams per deciliter) regressed on supplemental Fe intake (X in milligrams) yielded a standard curve equation Y = 6.21 + 0.426 ± 0.04X (r2 = 0.89). Using this equation, the relative bioavailability in Fe2(SO4)3·7H2O was estimated to be 37.0%. This value is considerably lower than the 65% value determined earlier by Fritz et al. (1970), who also used hemoglobin repletion as a criterion of Fe utilization. We have no explanation for this difference. In Assay 2, weight gain, hemoglobin and hematocrit responded linearly (P < 0.05) to increasing doses of supplemental Fe from feed-grade FeSO4·H2O or ZnFeSO4·H2O (Table 3). Also, addition of 20 mg Fe/kg from cottonseed meal increased (P < 0.05) weight gain, hemoglobin, and hematocrit over values observed for chicks fed the unsupplemental negative-control diet. Hemoglobin (Y in grams per deciliter) was regressed on supplemental Fe intake (X in milligrams) from FeSO4·H2O (X1) and Zn-FeSO4·H2O (X2). This regression yielded the equation Y = 4.10 + 0.312 ± 0.04 X1 + 0.290 ±
1390
BOLING ET AL. TABLE 2. Bioavailability of iron in ferric sulfate, Assay 11
Diet2
Supplemental Fe source3
Weight gain4
1. 2. 3. 4.
None FeSO4·7H2O FeSO4·7H2O Fe2(SO4)3·7H2O
(g) 195b 258a 272a 253a
Basal (B) B + 10 mg Fe/kg B + 20 mg Fe/kg B + 20 mg Fe/kg Pooled SEM
Blood analyses Supplemental Hemoglobin4,5 Hematocrit4 Fe intake4 (mg) 0c 4.14b 8.54a 7.91a 0.30
(g/dL) 6.21c 7.95b 9.89a 7.46b 0.25
(%) 20.9c 25.4b 29.3a 23.9b 0.7
a–cMeans
in a column with no common superscript differ significantly (P < 0.05). are means of four pens of two male chicks during a 14-d feeding period (8 to 22 d posthatching); average initial weight was 98 g. 2The basal Fe-deficient diet (Table 1) was analyzed to contain 20 mg Fe/kg, and this diet was fed during the 8-d pretest period to deplete Fe stores. 3Supplemental Fe sources contained, by analysis: reagent-grade FeSO ·7H O, 20.1% Fe; reagent-grade 4 2 Fe2(SO4)3·7H2O, 22.7% Fe. 4Linear (P < 0.05) response to increasing Fe doses from FeSO ·7H O. 4 2 5Hemoglobin (Y in grams per deciliter) regressed on supplemental Fe intake (X in milligrams per 14 d) from the standard FeSO4·7H2O product yielded the equation Y = 6.21 + 0.426x (r2 = 0.89). Using this as a standard curve, relative Fe bioavailability in Fe2(SO4)3·7H2O was estimated to be 37.0%. 1Data
0.04 X2 (R2 = 0.71). The ratio of slopes indicated that the RBV of Fe in Zn-FeSO4·H2O was 93%, a value not significantly different (P > 0.10) from the standard, which suggests that this product is an excellent source of bioavailable Fe. Large ratios of dietary Zn:Fe have been known to decrease Fe utilization (Cox and Harris, 1960; Bafundo et al., 1984), although this antagonism does not seem to manifest in chicks fed purified diets based on casein (Parsons et al., 1989). Our assay system (Table 3) involving graded doses of Fe from FeSO4·H2O has been
tested in both the presence and absence of 30 mg Zn/kg from ZnSO4·H2O (unpublished data). Slope of the hemoglobin response curve was virtually the same as a function of supplemental Fe intake, regardless of whether ZnSO4·H2O was present or absent. Thus, as shown in Assay 2, the (excess) Zn provided in ZnFeSO4·H2O does not reduce the utilization of Fe in this product. Assay 2 also was used to estimate the bioavailability of Fe in cottonseed meal. Because only one level of the ingredient was added, standard-curve calculations for
TABLE 3. Bioavailability of iron in zinc-iron sulfate and cottonseed meal, Assay 21
Diet2
Supplemental Fe source3
Weight gain4
Blood analyses Supplemental 4 Hemoglobin4,5 Hematocrit4 Fe intake
1. 2. 3. 4. 5. 6.
None FeSO4·H2O FeSO4·H2O Zn-FeSO4·H2O Zn-FeSO4·H2O Cottonseed meal
(g) 172c 245a 264a 243a 259a 205b 8
(mg) 0 3.58 7.46 3.85 8.20 6.93 0.34
Basal (B) B + 9.4 mg Fe/kg B + 18.8 mg Fe/kg B + 10.3 mg Fe/kg B + 20.7 mg Fe/kg B + 20.0 mg Fe/kg Pooled SEM a–cMeans
(g/dL) 3.96d 5.47b 6.32a 5.09c 6.54a 5.30b,c 0.26
(%) 19.1c 21.8b 23.8a 21.8b 25.0a 23.4a,b 0.6
in a column with no common superscript differ significantly (P < 0.05). are means of five pens of five male chicks during a 13-d feeding period (8 to 21 d posthatching); average initial weight was 89 g. 2The basal Fe-deficient diet was analyzed to contain 20 mg Fe/kg, and this diet was fed during the 8-d pretest period to deplete Fe stores. 3Supplemental Fe sources contained, by analysis: feed-grade FeSO ·H O, 28.8% Fe; feed-grade Zn4 2 FeSO4·H2O, 14.2% Fe and 20.2% Zn; cottonseed meal, 200 mg Fe/kg. 4Linear (P < 0.05) response to increasing Fe doses from both FeSO ·H O and Zn-FeSO ·H O. 4 2 4 2 5Hemoglobin (Y in grams per deciliter) regressed on supplemental Fe intake (milligrams per 13 d) from FeSO4·H2O (X1) and Zn-FeSO4·H2O (X2) yielded the equation Y = 4.10 + 0.312 ± 0.04 X1 + 0.290 ± 0.04 X2 (R2 = 0.71). Standard curve methodology indicated that the Fe bioavailability in cottonseed meal was 55.5%. 1Data
1391
IRON BIOAVAILABILITY TABLE 4. Bioavailability of iron in iron-zinc sulfate, Assay 31 Blood analyses
Diet2
Supplemental Fe source
Weight gain4
Supplemental Fe intake4
Hemoglobin4,5 Hematocrit4
1. 2. 3. 4. 5.
none FeSO4·H2O FeSO4·H2O Fe-ZnSO4·H2O Fe-ZnSO4·H2O
(g) 153 237 256 251 267 6
(mg) 0 3.73 7.63 3.80 8.14 0.11
(g/dL) 3.84 5.60 6.87 6.61 7.47 0.19
Basal (B) B + 10 mg Fe/kg B + 20 mg Fe/kg B + 10 mg Fe/kg B + 20 mg Fe/kg Pooled SEM
(%) 18.4 22.6 24.2 22.8 24.2 0.7
1Data are means of five pens of five male chicks during a 13-d feeding period (9 to 22 d posthatching); average initial weight was 118 g. 2The basal Fe-deficient diet was analyzed to contain 20 mg Fe/kg, and this diet was fed during the 9-d pretest period to deplete Fe stores. 3Supplemental Fe sources contained, by analysis: feed-grade FeSO ·H O, 28.8% Fe, feed-grade Fe4 2 ZnSO4·H2O, 20.2% Fe and 13.0% Zn. 4Linear (P < 0.05) response to increasing Fe doses from both FeSO ·H O and Fe-ZnSO ·H O. 4 2 4 2 5Hemoglobin (Y in grams per deciliter) regressed on supplemental Fe intake (milligrams per 13 d) from FeSO4·H2O (X1) and Fe-ZnSO4·H2O (X2) yielded the equation Y = 4.26 + 0.343 ± 0.047 X1 + 0.433 ± 0.044 X2 (R2 = 0.82).
the hemoglobin data were used to derive an Fe RBV value of 55.5% for cottonseed meal. We are unaware of any previous estimate of Fe RBV in cottonseed meal for chicks. However, Pellett et al. (1990) conducted an Fe depletion-repletion assay in rats and found that the Fe in cottonseed protein isolate had an RBV of 43% relative to their FeSO4·7H2O standard. The low values in both chicks and rats confirm that cottonseed contains inhibitors that reduce Fe bioavailability to the animal. In the Pellett et al. (1990) work, the RBV of Fe in a protein isolate from cottonseed was lower than that in a protein isolate from soybeans. In contrast, the Fe RBV found herein for cottonseed meal (55.5%) was higher than the Fe RBV value of 38.5% we found earlier for soybean meal in chickens (Biehl et al., 1997). Obviously, soybean meal, like cottonseed meal, contains factors that reduce the absorption efficiency of Fe. Cottonseed meal is higher in phytate than soybean meal (de Boland et al., 1975; Erdman, 1979), and it also contains gossypol, which can bind Fe. Hence, it is somewhat surprising that Fe bioavailability in cottonseed meal (55.5%) is higher than that in soybean meal, i.e., 38 to 45% (Chausow and Czarnecki-Maulden, 1988a,b; Biehl et al., 1997). The effects of phytate on Fe utilization are not clear (Welch and Van Campen, 1975; Hallberg and Rossander, 1982). Indeed, Chausow and Czarnecki-Maulden (1988b) reported that Fe RBV in sesame seed meal and rice bran (both very high in phytate) were 96 and 77%, respectively. They also reported that Fe RBV in corn was only 20%, but it was 85% in corn gluten meal and 90% in kibbled corn that had undergone both heat treatment and expansion. These data, together with the soy heat treatment data of Morck et al. (1982), suggest that heat treatment per se increases the bioavailability of Fe in both corn and soy products. In Assay 3, linear (P < 0.05) responses in weight gain, hemoglobin, and hematocrit occurred when graded increments of Fe from either FeSO4·H2O or Fe-
ZnSO4·H2O were added to the Fe-deficient basal diet (Table 4). Multiple linear regression plots of hemoglobin and hematocrit are shown in Figure 1. The ratio of slopes for hemoglobin indicated an Fe RBV for FeZnSO4·H2O of 126%, which was greater (P < 0.05) than the FeSO4·H2O standard. However, the hematocrit data indicated that Fe-ZnSO4·H2O slope was only 95% (P > 0.10) of the FeSO4·H2O slope. Moreover, regressing weight gain data in the same manner (data not shown) produced an Fe RBV estimate of 101% for FeZnSO4·H2O. Thus, we believe the proper conclusion is that the bioavailability of Fe in Fe-ZnSO4·H2O is equivalent to, and not greater than, the Fe present in feed-grade ferrous sulfate. Previous hemoglobin depletion-repletion work in chicks using an Fe-deficient casein-dextrose diet (Chausow and Czarnecki-Maulden, 1988a, b; Biehl et al., 1997) provided clear evidence that blood hemoglobin is a better response parameter than either blood hematocrit or weight gain. Weight gain is particularly questionable as an Fe response criterion when standardcurve methodology is used for estimating Fe RBV in intact ingredients such as the cottonseed meal we evaluated in Assay 2 (Table 3). Thus, dietary addition of 10% cottonseed meal to an Fe-deficient semipurified diet could either increase or decrease weight gain for reasons other than its contribution of bioavailable Fe. As shown by Chausow and Czarnecki-Maulden (1988a,b), however, blood hemoglobin responds to feed ingredient supplementation based on its contribution of usable Fe. Hence, we believe our estimate of Fe RBV in cottonseed meal, although perhaps less defendable than those for our mixtures of FeSO4·H2O and ZnSO4·H2O, is valid. In fact, our Fe RBV estimate for cottonseed meal was done in essentially the same manner as was done previously for corn, corn gluten meal, meat by-products, and soybean meal (Chausow and Czarnecki-Maulden 1988a, b; Biehl et al., 1997) and for various human foods by Fritz et al. (1970).
1392
BOLING ET AL.
REFERENCES
FIGURE 1. Multiple linear regression plot of hemoglobin (top) and hematocrit (bottom) regressed on supplemental Fe intake from FeSO4·H2O (π = X1) and Fe-ZnSO4·H2O (◊ = X2). Data from Assay 3 were fitted to a multiple linear regression model involving two straight lines with a common intercept.
Aoyagi, S., and D. H. Baker, 1995. Iron requirement of chicks fed a semipurified diet based on casein and soy protein concentrate. Poultry Sci. 74:412–415. Bafundo, K. W., D. H. Baker, and P. R. Fitzgerald, 1984. The iron-zinc interrelationship in the chick as influenced by Eimeria acervulina infection. J. Nutr. 114:1306–1312. Biehl, R. R., J. L. Emmert, and D. H. Baker, 1997. Iron bioavailability in soybean meal as affected by supplemental phytase and 1a-hydroxycholecalciferol. Poultry Sci. 76: 1424–1427. Chausow, D. G., and G. L. Czarnecki-Maulden, 1988a. The relative bioavailability of plant and animal sources of iron to the cat and chick. Nutr. Res. 8:1041–1050. Chausow, D. G., and G. L. Czarnecki-Maulden, 1988b. The relative bioavailability of iron from feedstuffs of plant and animal origin to the chick. Nutr. Res. 8:175–185. Cox, D. H., and D. L. Harris, 1960. Effect of excess dietary zinc on iron and copper in the rat. J. Nutr. 70:514–520. de Boland, A. R., G. B. Garner, and B. L. O’Dell, 1975. Identification and properties of “phytate” in cereal grains and oilseed products. J. Agric. Food Chem. 23:1186–1189. Erdman, J. W., 1979. Oilseed phytates: nutritional implications. J. Am. Oil Chem. Soc. 56:736–741. Fritz, J. C., G. W. Pal, T. Roberts, J. W. Boehne, and E. L. Hove, 1970. Biological availability in animals of iron from common dietary sources. J. Agric. Food Chem. 18:647–651. Hallberg, L., and L. Rossander, 1982. Effect of soy protein on nonheme iron absorption in man. Am. J. Clin. Nutr. 36: 514–520. Henry, P. R., and E. R. Miller, 1995. Iron bioavailability. Pages 169–200 in: Bioavailability of Nutrients in Animals: Amino Acids, Minerals, and Vitamins. C. B. Ammerman, D. H. Baker, and A. J. Lewis, ed. Academic Press, San Diego, CA. Morck, J. A., S. R. Lynch, and J. D. Cook, 1982. Reduction of the soy-induced inhibition of nonheme iron absorption. Am. J. Clin. Nutr. 36:219–228. National Research Council, 1994. Nutrient Requirements of Poultry. 9th rev. ed. National Academy Press, Washington, DC. Parsons, C. M., D. H. Baker, and C. C. Welch, 1989. Effect of excess zinc on iron utilization by chicks fed a diet devoid of phytate and fiber. Nutr. Res. 9:227–231. Pellett, L. J., M. I. Schnepf, P. E. Johnson, and N. M. DiMarco, 1990. A comparative study of iron bioavailability in rats from soybean and cottonseed protein isolate diets. J. Food Qual. 13:419–433. SAS Institute, 1985. SAS Users Guide: Statistics. Version 5 Edition. SAS Institute Inc., Cary, NC. Wedekind, K. J., and D. H. Baker, 1990. Zinc bioavailability in feed-grade sources of zinc. J. Anim. Sci. 68:684–689. Welch, R. M., and Van Campen, 1975. Iron availability to rats from soybeans. J. Nutr. 105:253–256.
