The proximate composition was determined according to AOAC methods: moisture (925.10) (13), ash (923.03) (14), protein (920.87, CF=6.25) (15), total fat determined by solvent extraction (Soxhlet method) (920.85) (18), TDF, IDF and SDF (991.43) (16). Carbohydrates were calculated by difference, according to the following equation:

Each analysis was conducted in triplicate. Energy density (kcal/g) was estimated from proximate composition using the Atwater factors (4 kcal/g for proteins and carbohydrates, and 9 kcal/g for lipids).

The physical and technological properties of the cookies (diameter, thickness, spread ratio, texture and color) were evaluated as described by Santuccione et al. (19). Cookie dimensions were measured with a caliper (Ruhlmann 200 mm, Buenos Aires, Argentina), and the spread ratio was calculated as the diameter-to-thickness ratio. Texture was determined by a three-point bending test using a motorized testing machine (Multitest 2.5-d; Mecmesin Ltd., Slinfold, UK) at a crosshead speed of 180 mm/min, with hardness defined as the maximum force at fracture. Color parameters (L*, a* and b*) were obtained by digital image analysis under standardized illumination, and the total color difference (ΔE*) was calculated to assess perceptible differences between samples:

where L₀*, a₀* and b₀* are the values for the 25/30 cookie formulation, which had the lowest oil and sugar reductions, while L*, a* and b* are the values for the formulations with higher reductions (25/60, 50/30 and 50/60, respectively).

All measurements were performed in quintuplicate (N=5).

The in vitro gastrointestinal digestion of cookie samples (CP1, CP2 and CP3 and their reduced formulations) was performed according to the standardized INFOGEST 2.0 protocol (20), simulating the oral, gastric and intestinal phases. Simulated salivary, gastric and intestinal fluids (1.25×) were prepared from stock saline solutions and adjusted to pH=7.0, 3.0 and 7.0, respectively. Fresh enzyme solutions (amylase, lipase, pepsin and pancreatin) and bile salts were prepared immediately before digestion. Aliquots of 5 g of cookies were subjected to sequential digestion at 37 °C under constant agitation (2 min oral phase; 120 min gastric and intestinal phases), with pH monitored and adjusted as required. After digestion, samples were centrifuged (VT3216; Cavour, Córdoba, Argentina) at 1500×g, and the soluble fractions were collected and stored at −20 °C until analysis. Enzyme blanks were included, and all experiments were performed in duplicate.

Antioxidant capacity was determined using the ABTṠ⁺ decolorization assay according to Llopart et al. (21), with minor modifications. The ABTṠ⁺ radical was generated by reacting 7 mM ABTS with 2.45 mM dipotassium peroxydisulfate in the dark at room temperature for 16 h and then diluted with a φ(ethanol,water)=50 % to an absorbance of 0.70±0.02 at 730 nm. For the assay, 60 µL of the soluble fraction obtained after in vitro digestion were mixed with 840 µL of the ABTṠ⁺ working solution and incubated in the dark for 25 min. Absorbance was measured at 730 nm using a UV/Vis spectrophotometer (model 330; Metrolab S.A., Buenos Aires, Argentina). Antioxidant capacity was quantified using a Trolox calibration curve (0.01–0.1 mg/mL) and expressed as mg of Trolox equivalents (TE) per 100 g of sample. All measurements were performed in triplicate.

Total phenolic content (TPC) was determined using the Folin-Ciocalteu method according to Llopart et al. (21), with slight modifications. An aliquot of 40 µL of the soluble fraction obtained after in vitro digestion was mixed with 2 mL of Folin-Ciocalteu reagent (1:10) and 1 mL of 7.5 % Na₂CO₃, and incubated at 40 °C for 10 min. Absorbance was measured at 760 nm using a UV/Vis spectrophotometer (model 330; Metrolab S.A.). TPC was quantified using a gallic acid calibration curve (0.025–0.700 mg/mL) and expressed as mg gallic acid equivalents (GAE) per 100 g of sample. All determinations were performed in triplicate.

A quantitative descriptive analysis (QDA) was conducted with eight panelists (3 male and 5 female), aged between 30 and 50, trained in tasting baked goods, recruited from the Faculty of Biochemical and Pharmaceutical Sciences at the National University of Rosario (Rosario, Argentina). All participants were informed about the study procedures and signed a written informed consent. Training included aroma, texture and taste recognition and discrimination, as well as descriptive assessments of baked products and pomace obtained from fruit and vegetable blends. A consensus list of attributes was subsequently established (Table S1). The attributes assessed were color, hardness, sweet aroma, oil aroma, foreign aroma, sweet taste, oil taste, foreign taste, chewiness, graininess, and overall acceptability, with “foreign aroma” and “foreign taste” referring to the pomace used in the formulations. A 10-cm unstructured line scale, anchored at 1 cm from each end, was used for each descriptor, with anchors indicating minimum and maximum intensities. Panelists marked the perceived intensity on the scale, and values were recorded as the distance (cm) from the left end (22). Evaluations were carried out in individual booths, in accordance with the International Organization for Standardization (ISO 11132:2021; ISO 8589:2007) (23, 24).

The color parameters and visual appearance of the different formulations are shown in Fig. 1. The incorporation of fruit and vegetable pomaces significantly influenced the color attributes of the cookies, with the effect varying depending on the type of pomace used. For CP1 formulations, a progressive reduction in lightness was observed as pomace amount increased, with values ranging from the highest in the 25/30 sample to the lowest in the 50/60 sample, confirming a darkening effect. This trend is consistent with the findings of Andrejko et al. (43), who reported that the inclusion of carrot and apple pomace in oat cookies significantly decreased L* values while increasing a* and b* components due to the presence of carotenoids, which impart reddish and yellowish tones. In our case, the increases in both a* and b* with the addition of P1 may also be related to the influence of carotenoid-rich residues, in line with earlier reports showing that β-carotene from carrot remains sufficiently stable during baking to affect final product color (44). Moreover, color differences (ΔE*>5) in all CP1 formulations (Table 4) indicate perceptible variations to consumers, similar to the perceptibility thresholds reported by Andrejko et al. (43), where ΔE values above 5 were associated with visible differences when carrot pomace was added to cookies.

In CP2, L* values decreased with higher pomace addition, while both a* and b* increased, resulting in an intense reddish-pink hue. The most pronounced differences were observed in the 50/60 formulation, which exhibited the highest contrast compared with 25/30. These results agree with Mitrevski et al. (33), who demonstrated that increasing beetroot powder (15–25 %) in biscuits significantly reduced L* and elevated a*, producing cookies with a darker and more reddish appearance due to the strong pigmentation effect of betalains. Similarly, Kruczek et al. (9) reported that betalains not only withstand thermal processing but also intensify red coloration in bakery matrices. Our results, showing ΔE* values above 5 for all CP2 formulations (Table 4), reinforce the strong visual impact of beet residues as dominant drivers of bakery product color.

In contrast, CP3 displayed a different pattern. L* values increased up to the 50/30 formulation, while a* and b* decreased, suggesting a dilution of the typical brownish shades generated during baking. However, the 50/60 formulation deviated from this trend, showing minimal color difference relative to 25/30. This subtle effect is likely due to the relatively neutral pigmentation of apple and pear residues. Similar observations were reported by Naseem et al. (40), who showed that substituting wheat flour with apple pomace powder (5–15 %) decreased L* only moderately and induced slight changes in a* and b*, depending on the substitution amount. Andrejko et al. (43) also highlighted that apple pomace has less effect on cookie coloration than carrot, corroborating our findings for CP3.

In addition to pigment composition, the sugar content of the pomaces likely contributed to the observed color changes. Although P2 contained less total and reducing sugars, its high betalain content may have dominated the color profile, leading to pronounced decreases in L* and increases in a* and b*. In contrast, P1 and P3 supplied not only pigments but also higher mass fractions of total and reducing sugars, which could promote Maillard reaction and caramelization during baking (45). In CP1, this combination intensified browning and color intensity, while in CP3, the relatively neutral pigments from apple and pear diluted the natural color development of the matrix, partially offsetting the darkening effect of Maillard reactions. This interplay may explain the non-linear trends observed for CP3, particularly in formulations with higher pomace amounts. Furthermore, the reduction of added sucrose (30 and 60 %) and its partial replacement by pomace-derived sugars may have further modulated Maillard reactions, highlighting the complex interaction between endogenous sugars and formulation adjustments in determining the final cookie color beyond the direct contribution of pigments.

Texture analysis revealed a clear trend among the different formulations, as shown in Table 4. Cookies with lower sugar reductions (25/30 and 50/30) showed significantly higher hardness values than formulations with greater sugar reductions (25/60 and 50/60), regardless of the type of pomace. These findings are consistent with previous studies reporting that sugar content contributes to cookie firmness and crispness by dissolving in the available water in the dough and, upon baking, recrystallizing and forming an amorphous glassy phase that strengthens the structure. This glassy sugar phase limits molecular mobility in the baked matrix, resulting in increased rigidity and hardness (46). However, no significant differences were observed between formulations with the same sugar content but different mass fractions of oil. Reducing oil mass fraction did not significantly affect cookie hardness in this study. This observation aligns with Sudha et al. (47), who found that fat reduction in soft dough cookies led to increased dough hardness but did not necessarily translate into harder baked products. The minimal impact of oil reduction on hardness in this study may be attributed to the compensatory effects of pomace addition, which could influence moisture content and overall texture more strongly than oil content alone. In particular, fiber–water interactions associated with pomace incorporation may have contributed, partially offsetting the expected effects of lipid reduction on texture.

Formulation and pomace type significantly influenced the technological parameters of the cookies (Table 4). Regarding diameter, the 25/30 formulation generally produced the largest cookies across pomace types. This behavior can be attributed to the higher sugar and oil contents, which may favor dough spreading during baking by enhancing system mobility and delaying structure setting (46). In CP1 and CP2, a consistent reduction in diameter was observed with greater oil and sugar reduction, with significantly lower values for 25/60, 50/30 and 50/60 formulations than for 25/30. In contrast, CP3 showed no significant differences among formulations, suggesting a compensatory effect of this pomace on dough spread, possibly linked to its fiber composition. Differences among pomace types were also evident: CP1 and CP2 yielded larger diameters than CP3, particularly in 25/30 and 50/60 formulations. Certain fibers can contribute to the liquid phase and soften the dough, thereby promoting expansion, while others can retain absorbed water and increase dough cohesiveness, limiting spreading (48).

Thickness was primarily influenced by formulation, tending towards slightly higher values at greater oil and sugar reduction, particularly in the 25/60 and 50/30 formulations of CP1 and CP2. However, these differences were not entirely consistent across pomace types. Thickness values of CP3 were generally higher than of CP2 in the 25/30 formulation, but the overall pattern remained variable. Spread ratio reflected the combined effects of diameter and thickness. The highest values were observed for CP2 with 25/30 formulation, consistent with its larger diameter and lower thickness. In most other cases, spread ratios decreased with greater reduction of oil and sugar, especially for CP1 and CP3, indicating reduced dough expansion during baking. CP3 consistently showed the lowest spread ratios, in agreement with its smaller diameters and relatively higher thickness.

Overall, the results confirm that cookie dimensions and spread behavior are strongly affected by the balance between fat, sugar and fiber in the formulation. However, the patterns observed were not always linear or consistent across pomace types, reflecting the complex interactions between ingredients during dough processing and baking. Similar variability has been reported in other studies on sugar and fat reduction, where differences in fiber source, water distribution and matrix composition can lead to divergent technological responses (49, 50). These findings highlight that the technological behavior of reformulated cookies cannot be predicted solely from individual ingredient functions, but depends on the combined and sometimes compensatory effects arising from the specific formulation and fiber source used.

Following in vitro gastrointestinal digestion, all formulations enriched with pomace (up to 19 % substitution) showed similarly high antioxidant activities, expressed as TE, ranging from approx. 3600 to 3900 mg/100 g of cookies (Fig. 2). Only the formulation with the greatest oil and sugar reduction (50/60 with 23 % pomace substitution) showed a slightly lower value, suggesting that excessive substitution may not further enhance and could even slightly limit antioxidant potential. This pattern suggests that, beyond an optimal substitution amount, excessive fiber addition can alter cookie microstructure, restricting antioxidant compound diffusion or extraction during digestion. Such non-linear behavior has been reported by Yagci et al. (51), who studied the effect of adding tomato pomace powder to extruded snacks. They found that increasing pomace from 0 to 20 % enhanced antioxidant capacity by up to 15 %, but no further improvement was observed at 20 %, indicating that higher pomace contents do not necessarily lead to proportional gains in antioxidant potential.

When comparing pomace types, CP1 and CP2 consistently delivered the highest ABTS values, which likely reflect differences in pomace composition, including the nature and proportion of antioxidant compounds, as well as their interactions with the food matrix. Thus, pomaces containing more stable, readily releasable bioactives tended to show higher values, while those with phenolics more closely associated with fiber may exhibit slightly weaker post-digestive activity (47).

Consistent with these findings, total phenolic content (TPC) after digestion, expressed as GAE, ranged from approx. 200 to 500 mg/100 g (Fig. 2). The highest TPC values were found in formulations containing 12–17 % pomace (25/30 and 50/30), suggesting that moderate substitutions favor the retention or release of phenolic compounds during digestion. In contrast, higher pomace incorporation (19–23 % in 25/60 and 50/60) led to a decline in TPC, possibly because the higher fiber content limited phenolic accessibility. The observed differences among pomace types further emphasize the influence of their inherent compositional characteristics, including phenolic profile and compound stability, on their recovery and activity after digestion (52). Moreover, Rocchetti et al. (53) investigated the effect of the addition of Moringa oleifera L. leaf powder on the phenolic bioaccessibility of durum wheat fresh pasta. and found a non-linear trend between polyphenol bioaccessibility and increasing amounts of powder added to the formulation.

Taken together, the ABTS and TPC results could be related not only to the release of phenolic compounds bound to the fiber-rich matrix, as mentioned above, but also to their potential interactions with other matrix components, which may be influenced by the reductions in oil and sugar content. Lower lipid content may reduce the solubilization of lipophilic antioxidant compounds during digestion, while changes in sugar content can modify water availability, indirectly affecting phenolic release. Additionally, the formation of new antioxidant compounds during processing could further contribute to the observed activity. It should also be noted that the ABTS assay may detect other redox-active molecules beyond phenolics, which could partly explain the measured values. However, excessive fiber/pomace seems to limit extractability (54). Accordingly, these methods, widely accepted for estimating the antioxidant potential of digested food matrices, provide an overall measure and do not capture the full antioxidant profile or the contribution of individual compounds. Therefore, the results should be interpreted as indicative of general antioxidant capacity and phenolic availability rather than a complete characterization of antioxidant activity.

The results of the quantitative descriptive analysis (QDA) of cookies incorporating pomaces P1 (CP1), P2 (CP2) and P3 (CP3) are shown in Fig. 3.

Color was one of the attributes showing marked differences among formulations. For CP1 and CP2, the 50/60 formulation exhibited a significantly darker color than the others, with a progressive reduction in color observed as the pomace content decreased, as indicated by the ∆E* values previously reported (Table 4), which are perceptible to the human eye. In the case of CP3, 25/30 formulation was the darkest, following the trend observed and previously explained (Fig. 1).

Regarding hardness, cookies with lower sugar reductions (25/30 and 50/30) were the hardest, consistent with instrumental hardness results (Table 4). Sweetness was significantly higher in cookies with lower sugar reductions, regardless of the type of pomace added. Differences in oil taste were detected in some formulations with lower oil reductions, particularly between CP1 and CP2. Foreign taste showed significant differences in cookies, being greater in those formulations with the highest pomace content (50/60), reflecting the influence of both the type and quantity of pomace incorporated.

Other attributes, such as sweet aroma, oil aroma, foreign aroma, chewiness and graininess, did not show significant differences, although certain trends were observed. Formulations with higher pomace content (CP1 50/30, CP2 50/60 and CP3 50/60) tended to exhibit greater chewiness, graininess, and perception of foreign aroma or taste. Overall acceptability showed significant differences in CP1 and CP2, with formulations containing the highest sugar and oil reductions (25/60 and 50/60) receiving lower scores. For CP3, no significant differences were observed, although the 50/60 formulation was perceived as the least acceptable.

These results demonstrate that the type of pomace and the degree of sugar and oil reduction selectively influence sensory attributes and overall acceptability, allowing identification of formulations that best combine sensory attributes and functional value. In CP1 and CP2, formulations with moderate sugar and oil reductions (25/30 and 50/30) generally showed higher overall acceptability, preserving adequate sweetness and texture while limiting the perception of foreign taste associated with higher pomace amounts. In contrast, formulations with the highest pomace incorporation and greatest reduction (formulation 50/60) tended to have lower overall acceptability, mainly due to increased hardness, reduced sweetness, and greater foreign taste perception. For CP3, no significant differences in overall acceptability were observed, although the 50/60 formulation was consistently perceived as the least acceptable. This suggests that pomaces P1 and P2 may better tolerate higher amounts of incorporation. Taken together, formulations with intermediate pomace amounts and moderate sugar and oil reductions emerged as the most promising candidates for further optimization, combining improved nutritional quality with sensory attributes positively evaluated by trained panelists.

Notably, while sweetness was significantly lower in formulations with greater sugar reduction, non-nutritive sweeteners were not used to match amounts typically preferred by consumers. Instead, a gradual lowering of sweetness perception thresholds was applied, following Lingiardi et al. (55), who suggested that maintaining high sweetness intensity in reformulated products may reinforce consumer preference for highly sweet foods, thereby limiting the effectiveness of policies aimed at reducing daily sugar exposure. Thus, the reformulation strategy in this study was designed not only to improve the nutritional profile of the product but also to encourage sustainable, long-term changes in dietary patterns.

Finally, it should be acknowledged that the sensory evaluation was conducted with a limited number of trained panelists and did not include a consumer hedonic test. Therefore, conclusions regarding sensory evaluation should be interpreted as a description of the formulations according to the attributes evaluated, rather than as indicators of general consumer acceptance or preference, which should be confirmed in future studies involving a larger group of untrained consumers.