Chromatographic Considerations and Maltitol Interference in Dietary Fiber Analysis Authors: Tadas Kargelis, Zhen K Lam, David Mangan Published: 2026 ABSTRACT AbstractBackgroundAOAC 2017.16 and AOAC 2022.01 are CODEX Type I methods for total dietary fiber (TDF) that integrate enzymatic–gravimetric steps with high-performance liquid chromatography (HPLC) quantification of low molecular weight soluble dietary fiber (SDFS). As implementation expands, chromatography-related limitations and matrix-specific interferences can affect SDFS determination.ObjectiveTo evaluate performance of AOAC-recommended and alternative HPLC columns for DP2/DP3 boundary assignment across a broad carbohydrate set, and to investigate maltitol-associated interference leading to apparent SDFS overestimation in starch-containing matrixes.MethodsSeventy-four carbohydrate standards (including oligosaccharides, polyols, and sugar alcohols) were analyzed in duplicate across five HPLC column configurations under AOAC 2017.16/2022.01 and AOAC 2009.01/2011.25 conditions. Column performance was scored by correct peak-area assignment relative to the maltose demarcation. Maltitol interference was assessed using food products formulated with/without maltitol and mixtures of maltitol (and other polyols) with wheat starch or wheat flour using scaled AOAC 2017.16 and AOAC 2001.03 procedures.ResultsAcross 74 analytes, SugarPak (Waters) provided the highest overall assignment accuracy, with a two-column series configuration scoring highest. Maltitol alone could be excluded by adjusting the DP2/DP3 demarcation; however, maltitol combined with glucose-based polymers produced an additional peak area in the SDFS region, yielding up to ∼8% apparent SDFS overestimation in wheat starch mixtures.ConclusionsChromatographic selection materially affects DP2/DP3 assignment for SDFS quantification. Maltitol can cause matrix-dependent artifactual SDFS in starch-containing test materials, suggesting formation of low-molecular-weight species during enzymatic processing. Guidance on column equivalency would strengthen AOAC dietary fiber methodology, and further work is warranted to further investigate maltitol interaction with glucose-based polymers during TDF analysis.HighlightsSugarPak showed best DP2/DP3 peak-area assignment across 74 standards. Several analytes elute at/near the demarcation depending on column configuration. Maltitol + starch matrixes can generate artifactual SDFS and inflate TDF. Dietary fiber (DF) is a key nutritional component with well-established roles in human and animal health, influencing digestion, metabolism, and gut microbiota composition (1). Regulatory definitions of total dietary fiber (TDF) continue to evolve, requiring analytical methods that accurately quantify both soluble and insoluble fractions. AOAC Method 2017.16 (2) was first approved for the measurement of total dietary fiber, which is inclusive of dietary fiber soluble in 78% (w/w) ethanol/water (SDFS) and high-molecular-weight dietary fiber composed of insoluble dietary fiber (IDF) together with dietary fiber soluble in water, but insoluble in 78% (w/w) ethanol/water (SDFP), while AOAC Method 2022.01 (3) expanded the approach to separate the measurements of IDF and SDFP within a unified protocol (Figure 1). Under the CODEX Alimentarius framework, AOAC Method 2017.16 replaced AOAC Method 2009.01 (4) as the Type I reference method in 2021, and AOAC Method 2022.01 replaced AOAC Method 2011.25 (5) in a similar capacity in 2025. The implementation of these modern AOAC methods that use enzymatic digestion steps more closely aligned with physiological conditions and food transit times in the human small intestine has improved the accuracy of TDF measurement; however, analytical challenges remain—specifically, issues regarding chromatographic performance for SDFS analysis. In this area, while universal consensus has not been reached on the minimum degree of polymerization (DP) a nondigestible oligosaccharide (NDO) must have to be considered dietary fiber, the vast majority of national regulatory authorities have recognized a threshold of DP ≥3 (i.e., at least three monosaccharides in length). Rapid Integrated Total Dietary Fiber Method procedure overview with example HPLC chromatogram depicting demarcation line and DP3 (SDFS)/DP2 split. A key feature of the recent AOAC methods for dietary fiber is the integration of high-performance liquid chromatography (HPLC) for the quantification of low-molecular-weight soluble dietary fiber (LMWSDF/SDFS) fractions. This chromatographic step enables the separation and quantification of oligosaccharides and other nondigestible carbohydrates that cannot be accurately measured by gravimetric means. Columns such as the Waters® SugarPak I and Tosoh® TSKgel G2500PWxl are commonly recommended for these determinations and are critical to the accuracy of SDFS analysis. However, chromatographic performance can significantly influence resolution, particularly for low-molecular-weight soluble carbohydrates. A number of limitations have been identified for these columns. On TSKgel column, disaccharides such as lactose, lactitol, and iso-maltose coelute with the DP3 demarcation line, and on SugarPak, fructotriose does not elute in the DP ≥3 area of the chromatogram, despite being classified as SDFS (6). In an effort to resolve the known limitations with these two columns and to find a chromatography system that would be more universally applicable, five different HPLC columns were assessed. Seventy-four unique oligosaccharides with diverse monosaccharide compositions, polyols, and sugar alcohols were chosen to perform the investigation, particularly focusing on DP2-3 carbohydrates representing most common glycosidic linkages. Although chromatographic performance is crucial for accurate SDFS quantification, additional challenges can arise when analyzing complex food matrixes. One such challenge that was identified and investigated in this study occurs when maltitol is used in food formulations. Maltitol, a hydrogenated disaccharide polyol derived from maltose, is employed as a low-calorie sweetener and bulking agent in sugar-free and reduced-sugar products such as confectionery, bakery items, dairy products, and nutritional bars (7,8). Maltitol provides approximately 75–90% of the sweetness of sucrose but roughly half its caloric value, making it an increasingly common ingredient in modern formulated foods. An observation was made that the same product formulated with and without maltitol yielded significantly different results when analyzed for SDFS. Further investigation revealed the presence of artifactual peaks in the chromatographic region applied for SDFS quantification, suggesting that maltitol can interfere with HPLC-based measurements and lead to an overestimation of soluble dietary fiber content. Understanding the chromatographic behavior of maltitol and its interaction with dietary fiber quantification procedures is therefore essential for ensuring accurate analytical interpretation and maintaining the integrity of fiber measurements. Experimental Reagents Test kits for the measurement of total dietary fiber (K-TDFR-200A, 700004347; AOAC Method 2001.03 [9]) and rapid integrated enzymatic–gravimetric–liquid chromatography (K-RINTDF, 700004335; AOAC Methods 2022.01 and 2017.16) were obtained from Neogen (Bray, Ireland). Distilled water was laboratory grade unless otherwise specified. Ethanol (99%), sodium hydroxide pellets, and glacial acetic acid were obtained from Sigma-Aldrich (Darmstadt, Germany). Sodium maleate buffer (50 mM, pH 6.0, 2 mM CaCl2) was prepared in-house by dissolving 11.6 g of maleic acid (Sigma-Aldrich) in distilled water, adjusting the pH to 6.0 using 4 M (160 g/L) NaOH solution, adding 0.6 g of calcium chloride dihydrate (Sigma-Aldrich), and adjusting the final volume to 2 L. TRIS buffer solution (0.75 M) was prepared in-house by dissolving 90.8 g of TRIS (tris(hydroxymethyl)aminomethane) (Neogen) in 800 mL of distilled water, adjusting the pH to 11 with 4 M NaOH, and bringing the final volume to 1 L. Acetic acid solution (2 M) was prepared in-house by diluting 115 mL of glacial acetic acid to 1 L with distilled water. Cation-exchange resin (G-AMBOH, 700004248) and anion-exchange resin (G-AMBH, 700004247) were obtained from Neogen. MES/TRIS buffer (0.05 M each, pH 8.2) was prepared in-house by dissolving 19.52 g of MES [2-(N-morpholino)ethanesulfonic acid] (Neogen) and 12.2 g of TRIS buffer in 1.7 L of deionized water, adjusting the pH to 8.2 with 6 N (240 g/L) NaOH solution. Hydrochloric acid (0.561 N) was prepared in-house by diluting 93.5 mL of 6 N HCl (Sigma-Aldrich) to 1 L with distilled water. Ethylenediaminetetraacetic acid calcium disodium salt hydrate was obtained from Sigma-Aldrich. Hexane was obtained from Univar Solutions (Bradford, United Kingdom). Chemicals High-amylose maize starch (K-TDFC, 700004345), wheat starch, and wheat flour (K-TSCK, 700004349) were obtained from Neogen. The following compounds were also obtained from Neogen: 1,3-β-D-galactosyl-D-glucose (O-LAC3), arabinobiose (O-ABI, 700004930), arabinofuranosyl-xylobiose (O-A3X, 700004929), arabinotriose (O-ATR, 700004934), cellobiosyl-glucose (O-BGTRIB, 700004936), cellobiosyl-mannose (O-GGM), cellotriose (O-CTR, 700004949), diacetyl-chitobiose (O-CHI2, 700004938), digalacturonic acid (O-GALA2, 700004950), glucosyl-cellobiose (O-BGTRIA, 700004935), glucosyl-mannobiose (O-GMM), glucosyl-mannose (O-GM), inulotriose (O-INU3, 700004955), isomaltose (O-IMO2, 700004953), isomaltotriose (O-IMO3, 700004954), kestose (O-KTR, 700004958), laminaribiose (O-LAM2, 700004959), laminaritriose (O-LAM3, 700004960), levanbiose (O-LEV2, 700004964), levantriose (O-LEV3, 700004965), maltotriose (O-MAL3, 700004966), mannobiose (O-MBI, 700004972), mannosyl-glucosyl-mannose (O-MGM), mannotriose (O-MTR, 700004976), nigerose (O-NGR, 700004977), nigerotriose (O-NGR3, 700004978), panose (O-PAN, 700004982), sophorose (O-SOPH, 700004984), sophorotriose (O-SOPH3, 700004985), xylobiose (O-XBI, 700004991), xylotriose (O-XTR, 700004995). The following compounds were obtained from Sigma-Aldrich (Darmstadt, Germany): α-cyclodextrin (C4642), β-cyclodextrin (C4767), γ-cyclodextrin (861413), cellobiose (22150), diethylene glycol (93171), galactitol (G0049000), glycerol (G7893), isomalt (I0465000), lactose (L254), maltose (M9171), stevioside (CDS020802). Additional compounds were obtained from Biosynth (Compton, United Kingdom): allolactose (OG09259), 1,3-β-galactobiose (OG10186), 1,4-β-galactobiose (OG15071); α-galactobiose (OG06491), chitosanbiose (OC08301); chitosantriose (YC181132), D-iditol (MI07130), DL-arabitol (MA07031), DL-threitol (FT76625), D-mannitol (MM63310), D-sorbitol (MS00451), 1,6-β-galactobiose (OG05008), gentiobiose (OG05175), isopanose (OI09775), lactitol monohydrate (OL01305), lactulose (OL03995), L-fucitol (MF44825), maltitol (OM06867), maltotriitol (OM06979), melibiose monohydrate (OM06702), meso-erythritol (ME05658), neoagarobiose (ON29863), ribitol (MR01271), sucralose (OS04165), trehalose (α,α) (OT05313), trehalose (α,β) (OT00157), trehalose (β,β) (MT04989), triacetyl-chitotriose (OT06497), turanose (OT06692), volemitol (MG10117), xylitol (MX01771). Food test materials, with and without maltitol, were provided by Mérieux NutriSciences (Barcelona, Spain). Apparatus The following apparatus were used throughout the study: a Waters Alliance e2695 high-performance liquid chromatography (HPLC) system equipped with a refractive index detector (RID), external column heating module capable of reaching 90 °C (Waters Column Heater Module), and a 50 µL injection loop. Additional equipment included an analytical balance; 25 mL glass bottles (Fisherbrand® or equivalent); magnetic stirring bars; a positive displacement dispenser (e.g., Eppendorf® Multipette® with Combitip®); a temperature-controlled water bath; a thermometer; 0.2 µm pore size syringe filters (nylon or equivalent); 5 mL polypropylene tubes; HPLC vials compatible with the LC system (Waters or equivalent); a vacuum pump or aspirator with adjustable regulator (e.g., Edwards XDS 10); a 1 L filtering flask with side arm (e.g., Büchner flask or equivalent); and a Gooch fritted disk filter (Pyrex® 50 mL, pore size coarse, ASTM 40–60 µm; Corning® No. 32940-50C). HPLC Columns TSKgel® G2500PWXL, 7.8 mm i.d. × 300 mm (Tosoh). Two analytical columns connected in series with cation and anion exchange guard columns in H+ and CO2 3- forms, respectively (Bio-Rad Laboratories); SugarPak®, 6.5 mm × 300 mm (Waters), with Guard Pak® LC pre-column inserts (Waters). Two analytical columns connected in series were also evaluated; Ultrahydrogel DP 120A, 7.8 mm × 300 mm (Waters). Two analytical columns connected in series with cation and anion exchange guard columns in H+ and CO2 3- forms, respectively (Bio-Rad Laboratories); SuperOligo PW, 6 mm × 150 mm (Tosoh). Two analytical columns connected in series with cation and anion exchange guard columns in H+ and CO2 3- forms, respectively (Bio-Rad Laboratories); PolySep-GFC-P 2000 (7.8 mm × 300 mm, Phenomenex®) and PolySep-GFC-P 3000 (7.8 mm × 300 mm, Phenomenex) analytical columns connected in series with cation and anion exchange guard columns in H+ and CO2 3- forms, respectively (Bio-Rad Laboratories). See Table 1 for column summary. Table 1.HPLC column list and parametersColumnSeparation techniqueParticle size, µmID, mmLength, cmConnected in seriesTSKgel G2500PWXL (Tosoh)Size exclusion77.830YesSugarPak (Waters)Cation exchange106.5302 configurationsUltrahydrogel DP 120A (Waters)Size exclusion67.830YesSuperOligo PW (Tosoh)aSize exclusion36.015YesPolySep-GFC-P 2000 and 3000 (Phenomenex)Size exclusionn/a7.830 eachYesaFlow rate adjusted to 0.3 mL/min. HPLC column list and parameters Flow rate adjusted to 0.3 mL/min. Elution Conditions TSKgel, Ultrahydrogel, and PolySep column setups were evaluated according to AOAC Methods 2017.16/2022.01 (HPLC grade distilled water, 0.5 mL/min at 85 °C, and run time of 60 min), except for the SuperOligo PW column, where the flow rate was adjusted to 0.3 mL/min to account for the shorter column length. SugarPak column setups were evaluated according to AOAC Methods 2009.01 and 2011.25 (HPLC grade distilled water plus ethylene diamine tetraacetic acid; disodium calcium salt (50 mg/L), 0.5 mL/min at 90 °C, and run time of 30 min). Carbohydrate Retention Time Analysis All compounds were dissolved in distilled water at concentrations between 1 and 10 mg/mL and injected in duplicate across five chromatography setups using HPLC conditions outlined in AOAC Methods 2017.16/2022.01 and 2009.01/2011.25. A modification was applied to all methods: HPLC analysis was performed without enzyme incubation, ethanol precipitation, or subsequent filtration, as all compounds were soluble at the tested concentrations. An additional adjustment was made for the SuperOligo column by reducing the flow rate to 0.3 mL/min to accommodate its shorter length, in line with the manufacturer’s guidance. Further testing was conducted using a modified SugarPak configuration, comprising two columns connected in series and evaluated under the same conditions specified for AOAC Methods 2009.01/2011.25. The demarcation maltose standard provided in the Rapid Integrated Total Dietary Fiber (K-RINTDF) assay kit was initially injected to establish the demarcation standard retention time. This standard was subsequently injected after every 15 test solution injections to monitor potential retention time drift. Chromatographic data were acquired and processed using Waters Empower™ software. Peaks were manually integrated at the chromatographic baseline beneath each peak. When applicable, peaks were divided at the retention time determined by the demarcation standard. Determination of Concentration Impact During Chromatographic Analysis The TSKgel and SugarPak columns were evaluated to determine the effect of analyte concentration on peak width relative to the demarcation line. For TSKgel, the test compounds included maltotriose, arabinotriose, gentiobiose, and maltose, whereas SugarPak was assessed using mannotriose, arabinotriose, sophorose, and melibiose. Each compound was prepared at two concentrations—1 mg/mL and 43.3 mg/mL—in distilled water, transferred to HPLC vials, sealed, and injected in duplicate. All analyses and data processing were performed under the same chromatographic conditions as previously described. The peak area relative to the demarcation line was compared between the two concentrations, expressed as the percentage of peak area eluting before or after the demarcation line according to the degree of polymerization of the carbohydrate analyzed. Determination of Maltitol Interference Food products containing maltitol and corresponding control products without maltitol were analyzed using a modified AOAC Method 2017.16, incorporating a tenfold reduction in assay scale. Using the same procedure, maltitol was also tested in a 50:50 (w/w) mixture with wheat starch. Additional tests included individual analyses of wheat starch and maltitol, as well as 50:50 (w/w) mixtures of maltitol with other polyols—xylitol, sorbitol, lactitol, isomalt, and erythritol. Mixtures of wheat starch and maltitol were further prepared at ratios of 5:95, 25:75, 50:50, 75:25, and 95:5 (w/w). Wheat flour and maltitol mixtures (50:50 w/w) were also analyzed under the tenfold-reduced assay conditions following AOAC Methods 2017.16 and 2001.03 to assess whether enzyme incubation conditions influenced the level of interference. Test Material Preparation Food products with and without maltitol were supplied milled, and the test materials were further subjected to defatting by following standard defatting conditions; test materials were washed 3 times using 25 mL of hexane, decanted each time, and left to air-dry overnight, followed by additional overnight drying in a desiccator. The test portions were then weighed and used in the assays. All other compounds were analyzed as is. Results and Discussion The performance of the analytical columns was assessed based on their ability to fully resolve compound peaks relative to the demarcation line. For compounds with DP2 or below, ideal resolution corresponds to complete elution after the demarcation line, whereas for DP3 or above, it should occur before the line. Peak areas were expressed as percentages before or after the demarcation line: compounds with a DP ≥3 were quantified as peak area percentage before the demarcation line, whereas compounds with DP ≤2 were quantified as peak area percentage after the demarcation line. The compounds were divided into five categories: Category 1 = Correct peak area assignment Category 1 = Correct peak area assignment Category 2 = >95% correct peak area assignment Category 2 = >95% correct peak area assignment Category 3 = 75–95% correct peak area assignment Category 3 = 75–95% correct peak area assignment Category 4 = <75% correct peak area assignment Category 4 = <75% correct peak area assignment Category 5 = 0% correct peak area assignment Category 5 = 0% correct peak area assignment Each compound was analyzed and assigned to a category according to its elution profile. For example, using the TSKgel column, xylobiose eluted entirely after the demarcation line and was therefore placed in Category 1. Weighted scoring was applied, with Category 1 assigned 10 points and Category 5 assigned 0 points. This scoring system enabled an objective evaluation of each column’s performance. Carbohydrate Retention Time Analysis Analysis of all 74 carbohydrates indicated that the best performance was achieved with the SugarPak (Waters) configuration connected in series, which scored 592. This configuration provided the greatest number of compounds (48) eluting correctly within the chromatogram (Category 1) and also the highest number of compounds (58) in Category 1 and Category 2 combined (>95% peak area accuracy). The same column used individually, as described in the AOAC Methods 2009.01 and 2011.25, scored 560, marginally underperforming the proposed “in-series” column configuration. The difference between these scores reflects changes in compound category assignment. Category 5 remained unchanged, whereas SugarPak in series showed one fewer compound in Categories 3 and 4, with the main improvement driven by a reduction in Category 2 assignments (from 20 compounds to 10). Although this shift is notable, its practical impact is likely minimal, because >95% correct peak assignment generally falls within normal analytical variability. It is also worth noting that, although these results highlight specific chromatographic challenges associated with TDF analysis, the materials examined in this study were isolated compounds. This situation is uncommon in routine TDF determinations, where analyses are typically performed on complex ingredient or food matrixes. As a result, the practical performance of SDFS quantification in real-world test materials is expected to be more robust than what is indicated under the conditions reported here. Although connecting two SugarPak columns in series can marginally improve chromatographic performance, this represents a modification to AOAC Methods 2009.01 and 2011.25, which may complicate implementation. When SugarPak was first introduced in official methodology, it exhibited limitations—most notably, misassignment of inulotriose, a component of inulin-type fructo-oligosaccharides (FOS), which elutes in the non-SDFS region. Unfortunately, this issue persists even when two SugarPak columns are connected in series. Later AOAC Methods (2017.16 and 2022.01) addressed this by reverting to the TSKgel column originally introduced in AOAC Method 2001.03; however, this column configuration in our study exhibited a lower number of compounds tested displaying correct peak area assignment with only 45 compounds falling within Categories 1 and 2 combined. Enzymatic pretreatments can also be applied to selectively remove targeted carbohydrates from the test material matrix before SDFS analysis using a SugarPak column. Notably, this approach is incorporated in AOAC Method 2009.01, in which a combination of sucrase, α-glucosidase, and β-galactosidase is used to hydrolyze sucrose, maltose, and lactose, respectively, thereby enabling accurate quantification of the fructotriose peak. Complete removal of fructans through the use of a fructanase preparation would likewise eliminate interference from these oligomers; however, such treatment would necessitate separate, dedicated analysis for fructan quantification. Table 3 summarizes all analyzed compounds along with the best-performing column(s) for each in terms of peak assignment. For most compounds, both TSKgel and SugarPak columns deliver comparable results; however, certain compounds are quantified more accurately using one column over the other. The optimal column was selected based on carbohydrate category classification as shown in Figures 2–4. If a compound belonged to the same category, either column was assigned. Categories 1 and 2, were treated as a single group for this figure. For example, levantriose is classified as Category 2 in TSKgel (Figure 2) and Category 1 in SugarPak (Figure 3). Due to the relatively small difference in elution patterns, both columns were designated as optimal. This figure serves as a practical guide for selecting the most appropriate column when the food formulation is known. Individual categorisation of each analyzed compound using TSKgel column. Individual categorisation of each analyzed compound using SugarPak column. Individual categorisation of each analyzed compound using SugarPak column connected in series. A total of five compounds either did not elute or eluted entirely in the wrong region under the three highlighted configurations (chitosanbiose, chitosantriose, digalacturonic acid, stevioside, and β-cyclodextrin; Table 3). Carbohydrates containing acidic or basic moieties would not be expected to elute due to their capture in the deionization steps described in the AOAC Official Method procedures. However, β-cyclodextrin, a cyclic oligosaccharide composed of seven glucose units, eluted in the non-SDFS region using all column configurations. Coincidentally, this compound is a food additive (10,11) and not considered a dietary fiber by the United States Food and Drug Administration (FDA) (12) or the European Food Safety Authority (EFSA)/European Commission (EC) (13) and is not widely used in food formulations, making the observed elution profile a positive outcome overall, because it will not artificially inflate SDFS values. Nevertheless, it does highlight methodological limitations should the current scenario be reversed. Figures 2–4 illustrate the categorization of all compounds analyzed using the three best-performing column configurations. Assessing the industrial and analytical relevance of each compound is challenging, because this can vary significantly depending on target analytes and specific food formulations. Nonetheless, the superior performance of SugarPak is clearly demonstrated. None of the alternative columns sourced outside official methodologies showed improved performance (Table 2), although this does not preclude the possibility that improved alternatives may emerge in the future. Table 2.Column performance weighted scoringColumnNo. of columns in seriesWeightScore108620Cat 1Cat 2Cat 3Cat 4Cat 5SugarPak24810448 592 SugarPak13620558 560 TSKgel PWxl2500222237157 476 PolySep-GFC-P2000/30002191219168 432 Ultrahydrogel DP 120A217923196 418 SuperOligo PW2171121169 416 Column performance weighted scoring Table 3.Individual compound optimal resolution Individual compound optimal resolution These authors would suggest that AOAC dietary fiber methodology could be strengthened by allowing the use of equivalent columns or a defined selection of columns, which have already been subjected to interlaboratory studies for dietary fiber methods. AOAC Methods 2009.01 and 2011.25 already addressed this by permitting either TSKGel or SugarPak chromatography options, and allowing for the use of “equivalent” columns. Although column equivalency can be contentious without clear guidelines, this approach enables alternative column use without modifying the official method. Chromatographic determination in AOAC Methods 2017.16 and 2022.01 would benefit from adopting this strategy, as seen in earlier methods. Determination of Concentration Impact During Chromatographic Analysis Chromatographic peaks typically broaden as analyte concentration increases, which can lead to inaccuracies in measuring soluble dietary fiber (SDFS). To evaluate this effect, two concentrations were tested: 1 mg/mL, representing approximately 2.5% of the total weight (1 gram) in the soluble fiber fraction supernatant, and 43.3 mg/mL, equivalent to ∼100% of the total weight based on evaporating roughly one-third of the filtrate to allow for minor “over-concentration” when compared to the AOAC Official Method, which specifies evaporation of one-quarter of the filtrate. These concentrations were selected to simulate low and high SDFS content in food products. Carbohydrates were chosen based on their retention times relative to the demarcation line, focusing on compounds that elute close to or across this boundary. Results outlined in Table 4 indicate that varying the concentration of certain carbohydrates can introduce measurable differences in quantitation. This variation is influenced primarily by the carbohydrate type and the column used. Among the analyzed compounds, mannotriose exhibited the greatest variation, with ∼10% peak area shift at the lower concentration using the SugarPak column. This suggests that mannotriose is quantified more accurately at lower concentrations. In contrast, sophorose, analyzed on the same column, displayed a peak area shift of ∼6%, but, in this case, the lower concentration reduced quantitation accuracy. Comparable, though less pronounced, trends were observed with TSKgel for the selected carbohydrates. Table 4.Carbohydrate concentration impact during chromatographic analysisCarbohydrateConc., mg/mL% PA correctly assigned% CV SugarPak Mannotriose43.353.8711.55163.45Arabinotriose43.396.810.72197.8Sophorose43.365.867.35159.35Melibiose43.31000.001100 TSKgel Maltotriose43.392.250.18192.02Xylotriose43.358.322.86160.73Gentiobiose43.314.517.33113.08Maltose43.399.790.05199.72 Carbohydrate concentration impact during chromatographic analysis The most obvious explanation for this variation is peak tailing. Data in Table 4 support this interpretation: carbohydrates expected to elute before the demarcation line (DP3 and above) show less peak area in the correct region, as tailing pushes more of the peak beyond the demarcation line. As expected, DP2 compounds exhibit the opposite effect. Clearly, dilution and reinjection of any SDFS test solutions that contain compounds at such a level to induce peak tailing should be sufficient to resolve this issue. Determination of Maltitol Interference A collaboration with Mérieux NutriSciences identified a previously unreported issue with maltitol interference in the measurement of total dietary fiber determination. Mérieux NutriSciences discovered that the same food product, formulated with and without maltitol, generated significantly different dietary fiber content results when analyzed using AOAC Method 2017.16 (Table 5). The presence or absence of maltitol, a DP 2 sugar alcohol, should have no impact on dietary fiber content. Table 5.AOAC Method 2017.16 analytical results of two food productsTest materialHMWDF (IDF + SDFP) (gravimetric)SDFS (HPLC)TDFProduct with maltitol3.227.1610.39Product without maltitol3.141.404.54Amylose maize control53.972.0656.03 AOAC Method 2017.16 analytical results of two food products Our SDFS chromatography evaluation had already shown that maltitol elutes slightly early on TSKGel, with ∼10% of its peak area erroneously appearing in the SDFS region (Figure 2, Category 3 compound). To account for this discrepancy and eliminate the associated error, the DP2/3 demarcation line was adjusted to exclude the entire maltitol peak during all subsequent experiments investigating this specific issue (Figure 5). Further examination of the HPLC chromatograms for the two food products revealed a distinct additional peak area within the SDFS region (Figure 6). This observation prompted further investigation to determine the extent of interference caused by maltitol when applying AOAC Method 2017.16. Maltitol and wheat starch analyzed separately contribute less than 1% dietary fiber. However, a mixture containing 50% wheat starch and 50% maltitol resulted in approximately 7% dietary fiber, representing an overestimation of about 6% due to the combined presence of these compounds. Further experiments were performed using a range of sugar alcohols mixed in a 1:1 ratio with wheat starch, as well as varying ratios of maltitol and wheat starch mixtures (Table 6 and Figure 7). Results presented in Table 6 indicate that this interference is unique to maltitol when combined with wheat starch. Figure 7 demonstrates that the highest level of overestimation, approximately 8%, occurs in mixtures containing 70% wheat starch and 30% maltitol. Additional testing using a 50:50 wheat flour/maltitol mixture resulted in an overestimation of SDFS by roughly 3.5%. To further evaluate the effect of analytical conditions, the same test materials were analyzed using AOAC Method 2001.03, which involves α-amylase incubation at approximately 100 °C for 30 min. Surprisingly, the results were nearly identical to those obtained with AOAC Method 2017.16, again showing an overestimation of about 3.5% (Table 7). These results indicate that maltitol alone does not produce any artifactual SDFS provided that the DP2/3 demarcation line is adjusted as per Figure 5. However, when combined with glucose-based polymers, overestimation of SDFS is likely to occur. This suggests that maltitol and the polymer may undergo some form of enzyme-mediated transglycosylation reaction, generating material that is measured as SDFS. Further work is needed to fully determine the extent of this interference, including maltitol’s behavior in the presence of other carbohydrates. Maltitol elution pattern in a food sample. HPLC traces showing product without maltitol (top) and product with maltitol (bottom). Maltitol/wheat starch mixture overestimation graph. Table 6.SDFS as per AOAC Method 2017.16 results for mixtures of sugar alcohols and wheat starch#Test MaterialSDFS, g/100 g1Wheat starch0.3110Maltitol0.2911Wheat Starch/Xylitol (50% w/w)0.1413Wheat Starch/Sorbitol (50% w/w)0.2915Wheat Starch/Lactitol (50% w/w)0.1217Wheat Starch/Isomalt (50% w/w)0.4719Wheat Starch/Erythritol (50% w/w)0.1921Wheat Starch/Maltitol (95%/5% w/w)2.9522Wheat Starch/Maltitol (75%/25% w/w)7.746Wheat Starch/Maltitol (50%/50% w/w)6.9023Wheat Starch/Maltitol (25%/75% w/w)4.2424Wheat Starch/Maltitol (5%/95% w/w)1.20 SDFS as per AOAC Method 2017.16 results for mixtures of sugar alcohols and wheat starch Table 7.Results of maltitol, wheat flour, and wheat flour/maltitol mixture under AOAC Methods 2017.16 and 2001.03 conditionsTest materialAOAC Method 2017.16 SDFS, g/100 gAOAC Method 2001.03 SDFS, g/100 gWheat flour2.563.08Maltitol0.310.14Wheat flour w/maltitol 50% (w/w)6.216.62 Results of maltitol, wheat flour, and wheat flour/maltitol mixture under AOAC Methods 2017.16 and 2001.03 conditions Conclusion The evolution of AOAC methods for dietary fiber analysis, particularly AOAC Methods 2017.16 and 2022.01, has significantly improved the accuracy of total dietary fiber (TDF) measurements in alignment with the CODEX definition. These methods integrate enzymatic–gravimetric steps with high-performance liquid chromatography to quantify SDFS, a critical fraction that cannot be reliably measured by gravimetric techniques alone. However, this study demonstrates that chromatographic performance remains a key determinant of analytical reliability, and challenges persist despite methodological advancements. Among the five chromatography configurations evaluated, Waters SugarPak exhibited the highest overall performance for resolving carbohydrates relative to the demarcation standard, particularly when two analytical columns were connected in series. This configuration achieved the most accurate peak assignments for DP2 and DP3 carbohydrates, which are central to SDFS quantification. Nevertheless, limitations such as misassignment of inulotriose, a component of inulin-type fructo-oligosaccharides, remain unresolved even with improved configurations. The Tosoh® TSKgel G2500PWxl column, although widely adopted in AOAC methods, demonstrated lower resolution for a number of oligosaccharides. However, a small number of compounds exhibited more accurate resolution, such as isomaltotriose, inulotriose, and mannotriose. Alternative columns sourced globally did not outperform these established options, underscoring the need for clear equivalency guidelines to accommodate future innovations without compromising method integrity. The study also examined the impact of analyte concentration on chromatographic accuracy. Although peak tailing and minor shifts in peak area assignment were observed at higher concentrations, these effects are unlikely to introduce significant bias in routine food analysis, given the complexity of real-world matrixes and the prominence of mixed carbohydrate profiles, with no single compound typically present in extreme abundance. Nonetheless, these findings highlight the importance of maintaining consistent test material preparation and injection conditions to minimize variability. A more intractable challenge identified is the interference caused by maltitol, a sugar alcohol used in reduced-sugar and sugar-free product formulations. Experimental data revealed that maltitol, when present alongside glucose-based polymers such as wheat starch, can lead to substantial overestimation of SDFS—up to 8% in certain mixtures. Surprisingly, this effect was consistent across AOAC Methods 2017.16 and 2001.03. The mechanism behind this phenomenon remains unclear but may involve transglycosylation during enzymatic hydrolysis, producing low molecular weight fragments that elute in the SDFS region. Importantly, this interference appears unique to maltitol, as none of the other polyols tested exhibited comparable effects. These findings underscore a specific limitation of liquid chromatography–based dietary fiber methods in accurately labeling dietary fiber in maltitol-containing products. Further investigation is warranted to elucidate the underlying chemistry and to develop corrective measures. In conclusion, although current AOAC methods represent a robust framework for dietary fiber determination, refinements in chromatographic protocols and deeper understanding of polyol interactions are essential to maintain analytical integrity. Future work should focus on establishing column equivalency criteria, exploring alternative chromatographic solutions, and characterizing the mechanisms of maltitol-induced interference. Such efforts will ensure that dietary fiber analysis remains accurate and adaptable to evolving food formulations, safeguarding both regulatory compliance and consumer trust. CRediT Author Statement Tadas Kargelis (Conceptualization [Supporting], Data curation [Lead], Formal analysis [Lead], Investigation [Lead], Methodology [Equal], Project administration [Lead], Resources [Supporting], Supervision [Equal], Writing—original draft [Lead], Writing—review & editing [Lead]), Zhen Kian Lam (Data curation [Supporting], Formal analysis [Lead], Investigation [Supporting]), and David Mangan (Conceptualization [Lead], Methodology [Lead], Project administration [Supporting], Resources [Lead], Supervision [Supporting], Writing—original draft [Supporting], Writing—review & editing [Supporting])