Determination of Total Dietary Fiber in Foods and Products with Little or No Starch, NonenzymaticGravimetric Method: Collaborative Study BETTY

W. LI and MARIA S. CARDOZO Nutrient Composition Laboratory, U.S. Department of Agriculture, Agricultural Research Service, Beltsville, MD 20705 Collaborators: K. E. Bach Knudsen; R. Brassard; R. Fein; D. Gamblin; F. Greenfield; K. Hodel; C. J. Huang; D. Jones; S. Lee; J. Marlett; R. McDonald; R. Mongeau; L. H. Rasmussen; J. B. Robertson; M. Sudler; R. Vincent; N. Vollendorf; C. Wo; D. Zemaitis

A collaborative study was conducted to validate a nonenzymaticgravimetric method for the determination of total dietary fiber (TDF) of samples containing little or no starch such as most fruits, and vegetables and many purified polysaccharides. This simple procedure involves suspension of freeze-dried, ground samples in deionized water and incubation at 37C for 90 min, followed by precipitation with 4 volumes of 95% ethanol. The weight of the dilute alcohol-insoluble residues after correcting for crude protein and ash corresponds to the TDF content of the sample. Six samples in blind duplicate (apples, apricots, cabbage, carrots, onions, and soy fiber) were sent with Celite to 10 laboratories. The reproducibility relative standard deviation (RSDR) of the TDF values for 9 laboratories ranged from 2.92 to 6.25%. The repeatability standard deviation (RSDr) for the 9 laboratories ranged from 1.50 to 2.70%. The method has been adopted first action by AOAC INTERNATIONAL. The official AOAC enzymaticgravimetric method, 985.29, for the determination of total dietary fiber (TDF) has been used on a variety of foods and food products, including fruits, vegetables, and cereals (1). In an earlier study (2), we showed that it was not neccessary to include a protease digestion step for a number of selected foods. More recently, we found that for fruits and vegetables containing little (<2% dry weight) or no starch, the steps for gelatinization with Termamyl, and incubation with amyloglucosidase could also be eliminated. Starting with an alcohol-insoluble residue in a crucible, we followed the procedures as described in the AOAC method, and obtained very similar TDF values (3). The present collaborative study was conducted to determine the reproducibility and repeatability of the much simplified procedure. Collaborative Study The 10 collaborating laboratories, including 1 Canadian, 1 Danish, and 8 American laboratories, are representatives from food manufacturers, government commercial testing laboratories, and universities. Six test samples in blind duplicates (4 g) and Celite (20 g) were sent to each laboratory. The 6 foods were (a) apples, (b) apricots, (c) cabbage, (d) carrots, (e) onions, and (f) FIBRIM 1450 (soy fiber), all of which had been used as test samples for other studies on dietary fiber analysis. Samples (b) through (e) were provided by Leon Prosky (4) of the U.S. Food and Drug Administration; sample (a) had been used in a study coordinated by Ruth Matthews of the U.S. Department of Agriculture, Human Nutrition Information Service, and sample (f) was

obtained from Protein Technologies International, courtesy of Grace Lo. The freeze-dried samples were further dried 4 h in a vacuum oven at 60C before bottling and shipping; samples were analyzed as received. 993.21 Total Dietary Fiber in Foods and Food Products with <2% StarchNonenzymaticGravimetric Method. First Action 1993 (Applicable to determination of 10% total dietary fiber in foods and food products with 2% starch, dry wt. basis.) Method Performance: See Table 993.21 for method performance data. (Caution: See Appendix: Laboratory Safety for safe handling of organic solvents.) A. Principle Dried fruit, vegetable, or isolated fiber sources are suspended in H2O and incubated 90 min at 37 to solubilize sugars and other water-soluble components. Water-soluble fiber components are then precipitated with ethanol. Residue is washed sequentially with 78% ethanol, 95% ethanol, and acetone and then dried at 105. One duplicate is analyzed for crude protein, the other for ash. Total dietary fiber (TDF) is calculated as weight of residue less weight of protein and ash. B. Apparatus (a) Analytical balance.Capable of weighing to 0.1 mg. (b) Air oven.Capable of maintaining 105 0.5. (c) Beakers.250 mL. (d) Desiccator.Containing mixture of colorless and indicating desiccant (Drierite is suitable). (e) Filtering flask.1 L capacity. (f) Fritted crucible.Porosity No. 2 (coarse ASTM 4060 m). Wet 0.5 g filter aid and evenly distribute by swirling with 78% ethanol; then apply vacuum to form even mat. Heat crucible containing filter aid in muffle furnace 1 h at 525, let cool in desiccator, and weigh before use. (g) Incubator or waterbath.Capable of maintaining 37 0.5. (h) Muffle furnace.Capable of achieving 525. C. Reagents (a) Ethanol.(1) 95% (without any organic additive). (2) 78%. Dilute 207 mL H2O with 95% ethanol to 1 L. (b) Acetone.

(c) Analytical filter aid.Acid-washed diatomaceous silica, ca 97.5% SiOH, ca 5% retained on 150 mesh screen [Celite , Fisher Scientific, 711 Forbes Ave, Pittsburgh, PA; or C.A.F.A. (Celite Analytical Filter Aid), Manville Products Corp., PO Box 5108, Denver, CO, is suitable]. D. Determination Accurately weigh to nearest 0.1 mg duplicate 500 mg freeze-dried, ground (30 mesh) samples or homogenized (by food processor) wet samples (containing ca 0.5 g dry matter) into separate 250 mL beakers. Add 25 mL (or volume necessary to bring wet sample to 25 mL) H2O to each beaker; sonicate or gently stir suspensions until samples are thoroughly wet, i.e., no clumps remain. Scrape down any particles on inside wall of beaker with rubber policeman, and rinse walls with 1 2 mL H2O. Cover beakers with Al foil and let stand 90 min without stirring in 37 incubator or water bath. Add 100 mL 95% ethanol to each beaker and let stand 1 h at room temperature (25 2). Collect residue under vacuum in preweighed crucible containing filter aid. If and when filtration becomes very slow, use closed-end Luer needle, or any small pointed object, to gently scratch matted sample without disturbing filter aid. Positive pressure may also be used if available. Wash residue 2 with 20 mL 78% ethanol, 2 with 10 mL 95% ethanol, and 1 with 10 mL acetone. Final rinsing with D. Determination Accurately weigh to nearest 0.1 mg duplicate 500 mg freeze-dried, ground (30 mesh) samples or homogenized (by food processor) wet samples (containing ca 0.5 g dry matter) into separate 250 mL beakers. Add 25 mL (or volume necessary to bring wet sample to 25 mL) H2O to each beaker; sonicate or gently stir suspensions until samples are thoroughly wet, i.e., no clumps remain. Scrape down any particles on inside wall of beaker with rubber policeman, and rinse walls with 1 2 mL H2O. Cover beakers with Al foil and let stand 90 min without stirring in 37 incubator or water bath. Add 100 mL 95% ethanol to each beaker and let stand 1 h at room temperature (25 2). Collect residue under vacuum in preweighed crucible containing filter aid. If and when filtration becomes very slow, use closed-end Luer needle, or any small pointed object, to gently scratch matted sample without disturbing filter aid. Positive pressure may also be used if available. Wash residue 2 with 20 mL 78% ethanol, 2 with 10 mL 95% ethanol, and 1 with 10 mL acetone. Final rinsing with acetone should be done in fume hood, collecting acetone wash in separate filtering flask for proper disposal. Dry crucible containing residue 2 h at 105. Cool crucibles 2 h in desiccators and weigh to nearest 0.1 mg. Ash residue from one duplicate 5 h at 525. Cool crucible 2 h in desiccator and weigh to nearest 0.1 mg. Analyze residue from remaining duplicate for crude protein by Kjeldahl nitrogen determination, 960.52 or 992.15, using %N 6.25.

Ecuacion where Wr = mg residue, P = % protein in residue, A = % ash in residue, and Ws = mg sample. Ref.: JAOAC 77, 687 (1994)

Results and Discussion Ten laboratories had agreed to participate in this collaborative study, but one withdrew due to lack of time and personnel. The results of duplicate analyses of 6 test samples from 9 laboratories are shown in Table 1. All values were used for statistical evaluation (5). The average TDF (%) values, the repeatability relative standard deviation (RSDr ), and the reproducibility relative standard deviation (RSDR) are given in Table 993.21. The respective values are as follows for apples, 12.89, 1.72, 5.58; apricots, 26.56, 2.70, 5.12; cabbage, 26.56, 1.70, 2.92; carrots, 29.60, 1.50, 5.10; onions, 17.31, 2.58, 6.25; FIBRIM 1450, 76.66, 1.75, 3.60. Laboratory 4 was a Cochran outlier for carrots only because the withinlaboratory variability was very small; therefore, data from all 9 laboratories were included in the final results. Comparisons of the average values from this study with those from our laboratory using the same procedures or an AOAC method (1) on the same test samples are presented in Table 2. This method is rapid and economical and has excellent precision both within- and between-laboratories.

Recommendation It is recommended that this nonenzymaticgravimetric method for the determination of total dietary fiber be adopted first action.