Comparing Extraction Techniques and Varieties in Grape Stems: A Chemical Assessment of Antioxidant Phenolics

Abstract

Grape stems are undervalued winemaking by-products that constitute a promising source of bioactive phenolics with notable antioxidant potential and diverse industrial applications, including food preservation, cosmetics, and pharmaceuticals. Effective valorisation of this resource requires not only efficient extraction strategies, but also the strategic selection of grape stem varieties to tailor phenolic profiles for specific high-value uses. In this study, a comparative assessment of three extraction techniques, pressurized liquid extraction (PLE), ultrasound-assisted extraction (UAE), and conventional solid–liquid extraction (SLE), across six grape stem varieties was conducted. By integrating spectrophotometric analyses of total phenolics and antioxidant capacity with HPLC-DAD profiling of individual phenolic compounds, the combined influence of extraction method and varietal composition on phenolic recovery was demonstrated. PLE and UAE significantly enhanced both yield and antioxidant capacity relative to SLE, with PLE providing the broadest spectrum of phenolic compounds. Varietal differences were also pronounced; e.g., Cabernet Sauvignon stems yielded higher antioxidant phenolic compound content, particularly under UAE, reinforcing the importance of aligning extraction technique and stem variety with the intended functional application.

1. Introduction

The global wine industry is one of the largest agro-industrial sectors worldwide, producing over 29 million tons of by-products annually [1]. Among these residues, grape stems represent nearly 8% of the solid waste generated during the winemaking process, making them a significant yet underutilized biomass fraction [2]. Traditionally, grape stems have been discarded through composting, incineration, or landfilling practices that contribute little to value creation and may even pose environmental challenges such as greenhouse gas emissions and leachate formation [1]. In recent years, however, these lignocellulosic materials have attracted growing scientific and industrial interest as a renewable source of bioactive compounds, aligning with the principles of green chemistry and circular economy.

Chemically, grape stems are composed of cellulose, hemicellulose and lignin, but their most valuable constituents are phenolic compounds, including flavan-3-ols (catechin, epicatechin, procyanidins, tannins), flavonols (quercetin, rutin, kaempferol), stilbenes (resveratrol, piceid), and phenolic acids (gallic, caffeic, ferulic, syringic acids) [3,4]. These compounds are widely recognized for their antioxidant, antimicrobial, anti-inflammatory, cardioprotective, and neuroprotective properties, making grape stems promising candidates for applications in functional foods, nutraceuticals, pharmaceuticals, and cosmetics. Moreover, phenolic compounds from grape stems have been linked to anti-aging effects, modulation of gut microbiota, and protection against oxidative stress-related diseases, further broadening their potential impact [5,6].

Recent studies highlight that the variety of grape stems plays a decisive role in determining phenolic composition and recovery efficiency. Dias-Costa et al. [7] compiled evidence showing that cultivars such as Touriga Nacional, Tinta Roriz, Castelão, Syrah, and Arinto differ markedly in their phenolic profiles. This variability directly impacts on the bioactivity of extracts, influencing antioxidant potential, antimicrobial efficacy, and even anti-aging properties. Similarly, Mitic et al. [8] reported significant differences in total phenolic content and antioxidant activity among nine Vitis vinifera varieties, with flavan-3-ols and phenolic acids predominating. These findings confirm that phenolic recovery is not uniform across grape varieties, and varietal composition directly influences both quantitative yield and qualitative bioactivity.

Equally critical is the choice of extraction technique, which governs not only yield but also the selectivity, composition, and stability of recovered compounds. Conventional solid–liquid extraction (SLE) remains widely used due to its simplicity and accessibility; however, it is constrained by long extraction times, high solvent consumption, and limited selectivity [5]. Pressurized liquid extraction (PLE), by contrast, employs elevated temperature and pressure to enhance solubility and mass transfer, thereby reducing solvent usage and improving efficiency. In addition, ultrasound-assisted extraction (UAE) represents a cost-effective and scalable alternative, where acoustic cavitation disrupts plant cell walls and accelerates phenolic release, with lower energy requirements compared to thermal methods [9,10]. Comparative studies confirm that extraction technology not only determines quantitative yield, but also modulates qualitative composition, antioxidant potential, and compound stability. Although emerging techniques such as supercritical fluid extraction (SFE) and microwave-assisted extraction (MAE) further expand the toolbox, offering environmentally friendly and highly efficient alternatives, their application to grape stems remains relatively underexplored. This limited adoption is largely due to higher operational costs, specialized equipment requirements, and challenges in scalability [9,11]. For instance, SFE often requires supercritical CO₂ combined with co-solvents to efficiently recover polar phenolics, which increases complexity and cost, while MAE may cause thermal degradation of sensitive compounds and is less easily adapted to industrial-scale continuous processes. In contrast, PLE and UAE strike a balance between efficiency, selectivity, and scalability, making them more suitable for both laboratory optimization and industrial implementation [5,9]. Their ability to deliver high yields of phenolics with reduced solvent consumption and lower energy input explains why they are currently the preferred techniques for the valorisation of food residues in both research and applied biorefinery contexts [11].

Despite these advances, most existing studies have either focused on varietal differences using a single extraction method or on methodological optimization without considering varietal diversity [12,13]. This fragmented approach leaves a critical gap in understanding how grape stem variety interacts with extraction technology to determine phenolic yield, composition, and bioactivity. Addressing this gap is essential for exploring tailored valorisation strategies that maximize recovery efficiency while ensuring sustainability.

The present study aims to bridge this gap by systematically comparing phenolic recovery from diverse grape stem varieties using three extraction techniques: SLE, PLE, and UAE. By evaluating yield, phenolic composition, and antioxidant capacity across varieties and extraction techniques, this research provides novel insights into the interplay between grape stem diversity and extraction technology. This perspective contributes to the broader goal of sustainable innovation in applied sciences, offering practical solutions for waste valorisation while deepening scientific understanding of the complex relationship between grape stem variety, extraction technology, and phenolic recovery. By situating grape stems within the framework of bioeconomy and resource efficiency, this work underscores their potential to move the wine industry toward a more sustainable, circular, and value-driven future.

2. Materials and Methods

2.1. Chemical and Reagents

HPLC-grade acetonitrile and formic acid were sourced from Labscan (Dublin, Ireland) and Acros Organics (Geel, Belgium), respectively. Standards including protocatechuic, syringic and vanillic acids; piceid (3,4′,5-trihydroxystilbene-3-O-β-D-glucopyranoside); (+)-catechin, (–)-epicatechin, epicatechin gallate; procyanidins B1 and B2; quercetin derivatives (3-O-galactoside, 3-O-glucuronide, 3-O-glucoside and quercetin dihydrate); as well as delphinidin-, cyanidin- and malvidin-3-O-glucosides were supplied by Extrasynthèse (Genay, France). Gallic acid, trans-caftaric acid, ferulic acid, ellagic acid, protocatechuic aldehyde, trans-resveratrol, trolox (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid), potassium persulfate, ABTS [2,2′-azinobis(3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt] and DPPH• were acquired from Sigma-Aldrich (Madrid, Spain). Sodium carbonate, Folin–Ciocalteu reagent, methanol and ethanol were obtained from Panreac (Barcelona, Spain).

2.2. Plant Material

Three white grape varieties, Riesling (RI), Gewürztraminer (GW), and Roussanne (RO), and three red grape varieties, Alicante Bouschet (AB), Cabernet Sauvignon (CS), and Petit Verdot (PV), of Vitis vinifera L. were selected for this study. All samples were sourced from the same geographic region, 2018 vintage, from a large number of grape bunches for each variety. They were provided by IMIDRA (Instituto Madrileño de Investigación y Desarrollo Rural, Agrario y Alimentario, Madrid, Spain) and originated from their vineyards located in the Madrid region (Alcalá de Henares, Spain).

Grape bunches were de-stemmed, and the stems were washed three times with cold water to remove residual must. The cleaned stems were subsequently lyophilized using a Telstar Lyobeta 15 freeze dryer (Telstar, Barcelona, Spain). The dried material was ground in a commercial blender, sieved to a particle size of ≤1 mm, homogenized, and stored at −20 °C until further analysis.

2.3. Conventional Extraction of Phenolic Compounds

Conventional extraction of phenolic compounds was performed by solid–liquid extraction (SLE), following the conditions described by Nieto et al. [14]. Briefly, 1 g of sample was extracted with 16 mL of ethanol/water (30:70, v/v) at room temperature for 24 h under magnetic stirring and protected from direct light. The resulting extracts were vacuum-filtered, and ethanol was removed by rotary evaporation at 37 °C using an IKA RV 10-control (IKA, Barcelona, Spain). Subsequently, the extracts were lyophilized using a Telstar Lyobeta 15 freeze dryer (Telstar, Barcelona, Spain) and stored at −20 °C in the dark until analysis.

2.4. Advanced Extraction Techniques

2.4.1. Pressurized Liquid Extraction

Pressurized liquid extraction (PLE) was performed using an ASE 350 system (Dionex Corporation, Sunnyvale, CA, USA) equipped with a solvent controller. Freeze-dried stem samples were extracted in 11 mL cells, each containing 1 g of dry grape stem powder and 1 g of diatomaceous earth. The extraction solvent, ethanol/water (30:70, v/v), was sonicated for 30 min to remove dissolved oxygen. Extractions were carried out at 1500 psi and 120 °C for 10 min, following an optimized Nieto et al. [14] protocol for phenolic compound recovery from grape stems. The liquid-to-solid ratio was 16 mL/g DW. Ethanol was removed by rotary evaporation at 37 °C using an IKA RV 10-control (IKA, Spain), and the resulting extracts were lyophilized in a Telstar Lyobeta 15 freeze dryer (Telstar, Barcelona, Spain). The lyophilized extracts were stored at −20 °C in the dark until analysis.

2.4.2. Ultrasound-Assisted Extraction

Ultrasound-assisted extraction (UAE) was performed following López-Padilla et al. [15] protocol with slight modifications. Briefly, 1 g of dry stem was mixed with 16 mL of ethanol/water (30:70, v/v) and subjected to extraction using a Branson Digital Sonifier model 250 (Branson Ultrasound, Brookfield, CT, USA), equipped with an ultrasonic probe and magnetic stirrer. The extraction consisted of three cycles of 15 min each, using a ½″ diameter disruptor horn probe at 70% amplitude (maximum power output 200 W, 60 Hz). The resulting extracts were vacuum-filtered, and ethanol was removed by rotary evaporation at 37 °C using an IKA RV 10-control (IKA, Barcelona, Spain). Extracts were then lyophilized in a Telstar Lyobeta 15 freeze dryer (Telstar, Barcelona, Spain) and stored at −20 °C in the dark until analysis.

All extractions were performed in triplicate from the homogenized grape stem samples to ensure reproducibility of the extraction procedure.

2.5. Total Phenolic Content (TPC)

The determination of total phenolic content (TPC) in the extracts was carried out by means of the Folin–Ciocalteu spectrophotometric method, following the protocol originally described by Singleton et al. [16] with slight adjustments. Briefly, an aliquot of 10 μL of each sample was combined with 0.6 mL of distilled water and 50 μL of Folin–Ciocalteu reagent in a test tube. After allowing the mixture to react for 1 min, 150 μL of a 20% (w/v) sodium carbonate solution and an additional 190 μL of distilled water were added. The reaction mixture was then kept in the dark at ambient temperature for 2 h. Absorbance readings were subsequently recorded at a wavelength of 760 nm. Quantification was performed using a calibration curve constructed with gallic acid under identical experimental conditions, and the TPC values were reported as mg of gallic acid equivalents (GAE)/g dry extract.

2.6. Antioxidant Capacity

The antioxidant potential of the extracts was evaluated through the DPPH and ABTS radical scavenging assays. The DPPH method was performed based on the approach proposed by Brand-Williams et al. [17], incorporating minor methodological adjustments. A DPPH stock solution was prepared by dissolving 23.5 mg of DPPH in 100 mL of methanol. To ensure a linear response, four different concentrations of each extract were prepared. For the reaction, 25 μL of the extract solution was mixed with 975 μL of the DPPH working solution to obtain a final volume of 1 mL. The reaction mixture was incubated at room temperature in the absence of light until completion (2 h), after which the absorbance was recorded at 516 nm.

For the ABTS assay, the ABTS radical cation (ABTS⁺) was generated by mixing a 7 mM ABTS solution with 2.45 mM potassium persulfate, and the mixture was left to react at room temperature for 12–16 h, following the procedure described by Nieto et al. [14]. The resulting ABTS⁺ solution was subsequently diluted with 5 mM phosphate buffer (pH 7.4) to obtain the working solution, which was adjusted to an absorbance value of 0.70 ± 0.02 at 734 nm. Four concentrations of each extract were prepared to ensure linearity. The assay was initiated by adding 10 μL of the extract solution to 990 μL of the ABTS⁺ working solution. After incubation at room temperature in dark conditions (120 min), the absorbance was measured at 714 nm.

Trolox was used as the reference antioxidant standard for both assays, and the antioxidant activity was expressed as Trolox equivalent antioxidant capacity (TEAC), reported in mmol Trolox/g dry extract. All measurements were conducted in triplicate.

2.7. Identification and Quantification of Phenolic Compounds by High-Performance Liquid Chromatography with Diode Array Detection (HPLC-DAD)

Phenolic constituents obtained through solid–liquid extraction (SLE), pressurized liquid extraction (PLE), and ultrasound-assisted extraction (UAE) were systematically identified and quantified by high-performance liquid chromatography using an Agilent 1260 Infinity system equipped with a diode array detector (DAD). Instrument control and data acquisition were performed with ChemStation software (version 6.8; Agilent Technologies Inc., Santa Clara, CA, USA), following the analytical approach described by Nieto et al. [14].

Chromatographic separation was achieved on a reversed-phase ACE C18-AR column (150 mm × 4.4 mm, 3 μm particle size; ACE, London, UK) fitted with an ACE 3 C18-AR guard column (7 mm × 13 mm), using a previously optimized elution method [14]. The injection volume was set at 40 μL, with a mobile phase flow rate of 1.0 mL/min and a column temperature maintained at 30 °C. Prior to analysis, extract samples were diluted with an ultrapure water/methanol mixture (1:1, v/v) and passed through a 0.45 μm PVDF membrane filter.

Compound identification was based on comparison of retention times and UV–Vis spectral characteristics with those of corresponding reference standards. Detection was carried out at multiple wavelengths depending on the phenolic subclass: 280 nm for hydroxybenzoic acids and flavan-3-ols, 320 nm for hydroxycinnamic acids and stilbenes, 360 nm for flavonols, and 520 nm for anthocyanins. Quantification was performed using external calibration curves generated from authentic standards at five concentration levels, each analyzed in triplicate. Phenolic contents were reported as mg of compound/g dry extract.

Each extract was subsequently analyzed in triplicate for total phenolics, antioxidant capacity, and targeted HPLC-DAD phenolic profiling to ensure analytical reliability.

2.8. Statistical Analysis

Statistical evaluation of the data was performed using Statgraphics Centurion XVII software (version 19.6.05) (Statistical Graphics Corp., The Plains, VA, USA). The effects of grape stem variety and extraction technique were assessed by analysis of variance (ANOVA). When significant differences were detected, means were compared using Fisher’s least significant difference (LSD) post hoc test, considering a significance threshold of p ≤ 0.05.

Multivariate statistical analysis was carried out with SIMCA 14.0 software (MSK Data Analytics Solutions, Umeå, Sweden). Principal component analysis (PCA) was employed and represented through score plots in order to differentiate the extract samples and to examine the impact of the extraction methods on both phenolic profiles and antioxidant capacity.

3. Results

This study aims to elucidate the influence of extraction technique on the recovery of phenolic compounds from grape stems by comparing extraction yields, total phenolic content, and antioxidant capacity of extracts obtained through SLE, PLE, and UAE from six grape stem varieties. In addition, the phenolic compounds of these six grape stem varieties were identified and quantified by HPLC-DAD.

3.1. Influence of the Extraction Technique and Grape Variety on the Recovery of Antioxidant Phenolic Compounds from Stems

The extraction technique and grape stem variety significantly influenced the extraction yield, TPC, and antioxidant capacity of the extracts. Overall, PLE consistently produced the highest extraction yields across all varieties, ranging from 47% in RO to 60% in AB. UAE achieved intermediate yields (38–53%), while conventional SLE consistently yielded the lowest values (25–39%) (see

Table 1. Extraction yields (g extract/100 g dry stem matrix).

Variety	SLE	PLE	UAE
AB	26 ± 6 b*C	60.1 ± 0.5 aA	53 ± 2 aA
CS	31 ± 1 bA,B,C	53.9 ± 0.6 aB,C	51 ± 3 aA
PV	39 ± 2 bA	58.0 ± 0.2 aA,B	43 ± 1 bA
RO	24.6 ± 0.3 cC	47 ± 2 aD	38 ± 2 bB
GW	27.4 ± 0.7 cB,C	56.2 ± 0.5 aA,B,C	48 ± 2 bA
RI	35 ± 5 bA,B	52 ± 4 aC	48.2 ± 0.6 aA

* Mean ± standard deviation; ^a,b,c Letters in superscript indicate significant differences among extraction methods (p < 0.05); _A,B,C,D Letters in subscript indicate significant differences among grape varieties (p < 0.05).

). Among the studied grape varieties, red stems from AB exhibited the highest extraction yields for all extraction techniques, except for SLE, where red stems from PV yielded the highest. These results allowed us to observe that the selection of the best extraction technique to obtain the highest extraction yield depends on the grape variety.

As can be observed in

Table 2. Total phenolic content (mg GAE/g dry extract) and antioxidant activity of the extracts obtained by DPPH and ABTS methods (mmol Trolox/g dry extract).

	TPC			DPPH			ABTS
Variety	SLE	PLE	UAE	SLE	PLE	UAE	SLE	PLE	UAE
AB	106 ± 9 bC	135 ± 8 aD	130 ± 11 aC	1.1 ± 0.1 cD	1.44 ± 0.04 bD	1.57 ± 0.04 aB,C	1.71 ± 0.09 bB	1.93 ± 0.03 aD	1.75 ± 0.08 bB
CS	178 ± 4 bA	179 ± 4 bA	222 ± 12 aA	1.89 ± 0.06 cA	2.03 ± 0.03 bA	2.4 ± 0.1 aA	2.36 ± 0.05 bA	3.1 ± 0.3 aA	2.5 ± 0.2 bA
PV	128 ± 7 cB	146 ± 14 bC	199 ± 1 aB	1.29 ± 0.02 bC	1.52 ± 0.03 bC	1.89 ± 0.01 aB,C	1.57 ± 0.08 cC	2.4 ± 0.1 bB	2.58 ± 0.07 aA
RO	129 ± 10 cB	158 ± 11 aB	144 ± 8 bC	1.41 ± 0.02 bB	1.83 ± 0.04 aB	1.43 ± 0.02 bC	1.47 ± 0.04 cD	2.27 ± 0.06 aB,C	1.8 ± 0.1 bB
GW	133 ± 5 bB	161 ± 6 aB	135 ± 9 bC	1.43 ± 0.05 bB	1.48 ± 0.06 bC,D	1.70 ± 0.04 aB	2.29 ± 0.04 aA	2.14 ± 0.07 bC	1.72 ± 0.05 cB
RI	61 ± 3 cD	109 ± 3 aE	101 ± 3 bD	0.63 ± 0.02 bE	1.10 ± 0.03 aE	0.59 ± 0.01 cD	0.95 ± 0.08 cE	1.39 ± 0.04 aE	1.30 ± 0.07 bC

^a,b,c Letters in superscript indicate significant differences among extraction methods (p < 0.05). _A,B,C,D,E Letters in subscript indicate significant differences among grape varieties (p < 0.05).

, PLE and UAE extraction techniques enhanced the recovery of TPC compared to SLE for all varieties. In several cases, UAE produced slightly higher TPC values than PLE (CS and PV), suggesting that ultrasonic cavitation enhances the extraction efficiency of phenolic compounds more effectively than the high-pressure conditions used in PLE. However, PLE extracts from AB, PV, RO, and GW presented the highest TPC values, consistent with their extraction yields. These results indicate a clear correlation between the amount of material extracted and its phenolic content for these last grape stems, emphasizing the importance of using advanced extraction methods for maximizing recovery.

On the other hand, the antioxidant capacity measured using the ABTS assay followed a trend consistent with that observed for TPC. In particular, CS and PV red grape stem extracts obtained by both PLE and UAE exhibited the highest radical-scavenging activity in the ABTS assay. However, when the antioxidant capacity was evaluated using the DPPH assay, the PLE-RO extract showed higher activity than the PV extract.

Taken together, these observations indicate that the recovery of antioxidant compounds is influenced not only by the extraction technique and grape variety, but also by the type of antioxidant assay employed. Because the DPPH and ABTS radicals differ in their chemical properties and reaction mechanisms, they may interact differently with the antioxidant constituents of the extracts.

In general, UAE was more effective for extracting antioxidant compounds as measured by the DPPH assay, except for the RO and RI varieties, where PLE showed a higher extraction capacity. Conversely, PLE was generally more effective for extracting antioxidant compounds when assessed using the ABTS assay. Finally, SLE proved to be the least efficient extraction technique overall, as it yielded lower TPC values and antioxidant capacities compared with both PLE and UAE.

3.2. Exhaustive Identification and Quantification of Phenolic Compounds by HPLC-DAD

The HPLC–DAD analysis revealed substantial qualitative and quantitative variation in the phenolic composition of grape stem extracts as a function of the extraction technique and grape variety. A total of more than twenty phenolic constituents were quantified across the three methods, including phenolic acids, stilbenes, flavan-3-ols, proanthocyanidins, flavonols, and anthocyanins. Overall, PLE produced the broadest phenolic profile, UAE maximized the extraction of flavan-3-ols and oligomeric procyanidins, and SLE generated the lowest phenolic yields.

3.2.1. Phenolic Profile of SLE Extracts

As can be seen in

Table 3. Phenolic composition of the SLE extracts (mg compound/g dry extract) (mean ± S.D.).

	SLE Extracts
	AB	CS	PV	RO	GW	RI
Gallic acid	0.280 ± 0.009 c	0.57 ± 0.02 a	0.60 ± 0.04 a	0.29 ± 0.01 c	0.50 ± 0.03 b	0.30 ± 0.02 c
Protocatechuic acid	Nd	Nd	Nd	0.001 ± 0.000 a	Nd	Nd
Vanillic acid	0.5 ± 0.2 a	Nd	Nd	Nd	Nd	Nd
trans-Caftaric acid	0.091 ± 0.006 c	0.24 ± 0.02 a	0.19 ± 0.01 b	Nd	0.040 ± 0.002 d	0.02 ± 0.01 d,e
Protocatechuic aldehyde	Nd	0.053 ± 0.008 a	Nd	Nd	Nd	Nd
Catechin	3.9 ± 0.4 c	5.6 ± 0.2 a	3.488 ± 0.002 b	1.6 ± 0.2 e	1.34 ± 0.01 d	1.93 ± 0.06 d,e
Epicatechin	1.54 ± 0.07 a	1.5 ± 0.1 a	0.902 ± 0.005 b	1.4 ± 0.4 a	0.63 ± 0.09 b	0.79 ± 0.03 b
Epicatechin gallate	Nd	0.08 ± 0.01 a	Nd	Nd	Nd	Nd
Procyanidin B1	1.4 ± 0.3 c	1.71 ± 0.01 b	0.75 ± 0.05 d	2.10 ± 0.05 a	Nd	0.80 ± 0.02 d
Procyanidin B2	0.7 ± 0.1 b	0.67 ± 0.02 b,c	1.97 ± 0.06 a	0.6 ± 0.1 b,c	0.69 ± 0.06 b	0.493 ± 0.006 c
Quercetin-3-O-glucuronide	1.01 ± 0.04 b	0.95 ± 0.01 b	1.096 ± 0.003 a	0.67 ± 0.01 d	0.87 ± 0.03 c	0.40 ± 0.02 e
Quercetin-3-O-glucoside	0.30 ± 0.01 b	0.207 ± 0.002 d	0.41 ± 0.01 a	0.218 ± 0.008 d	0.28 ± 0.01 c	0.199 ± 0.005 e
Malvidin glucoside	2.5 ± 0.3 a	0.109 ± 0.004 b,c	0.37 ± 0.02 b
Cyanidin glucoside	0.11 ± 0.02 a	Nd	Nd
Cyanidin derivative	2.0 ± 0.3 a	Nd	Nd
Unidentified anthocyanins	1.6 ± 0.2 a	0.196 ± 0.006 b,c	0.35 ± 0.02 b

Nd = not detected; ^a,b,c,d,e Letters in superscript indicate significant differences among grape stem varieties.

, a total of 16 phenolic compounds were identified across the SLE extracts obtained from the six grape stem varieties. These compounds belonged to five major phenolic classes: phenolic acids, flavan-3-ols, procyanidins, flavonols, and anthocyanins. Phenolic acids represented the least abundant group, with only four detected and generally at low concentrations; among them, gallic acid was the only compound consistently quantified across all varieties (0.28–0.60 mg/g dry extract). In contrast, flavan-3-ols were among the most abundant classes, with catechin showing the highest concentrations (1.34–5.60 mg/g dry extract), followed by epicatechin (0.63–1.54 mg/g dry extract). Procyanidins were also widely distributed, particularly procyanidin B1 and procyanidin B2, which reached maximum levels of 2.10 mg/g dry extract and 1.97 mg/g dry extract, respectively. Flavonols were moderately represented, with quercetin-3-O-glucuronide and quercetin-3-O-glucoside detected in all extracts (0.20–1.10 mg/g dry extract). Anthocyanins were restricted to the red grape varieties (AB, CS, and PV), with AB exhibiting the highest anthocyanin content. Overall, the phenolic profile of the six grape varieties was dominated by flavan-3-ols and procyanidins, whereas certain phenolic acids (like protocatechuic acid in RO) and anthocyanins such as cyanidin glucoside (in CS) occurred only at trace levels.

3.2.2. Phenolic Profile of PLE Extracts

A total of 25 phenolic compounds were identified in the PLE extracts obtained from the six grape stem varieties (see

Table 4. Phenolic composition of the PLE extracts (mg compound/g dry extract) (mean ± S.D.).

	PLE Extracts
	AB	CS	PV	RO	GW	RI
Gallic acid	0.13 ± 0.02 b	0.18 ± 0.08 b	0.316 ± 0.004 a	0.16 ± 0.01 b	0.28 ± 0.03 a	0.16 ± 0.04 b
Protocatechuic acid	0.004 ± 0.000 c	0.006 ± 0.001 a,b	0.006 ± 0.001 a,b,c	0.008 ± 0.001 a	0.005 ± 0.000 b,c	0.005 ± 0.001 b,c
Monogalloyl glucoside	0.015 ± 0.003 a	Nd	Nd	Nd	0.015 ± 0.002 a	Nd
Vanillic acid	0.32 ± 0.01 a	Nd	Nd	Nd	Nd	Nd
Syringic acid	0.19 ± 0.02 b	Nd	Nd	Nd	10.3 ± 0.8 a	Nd
trans-Caftaric acid	0.37 ± 0.02 b	0.40 ± 0.04 b	0.497 ± 0.008 a	0.130 ± 0.005 c	0.125 ± 0.008 c	0.14 ± 0.02 c
Ferulic acid	0.004 ± 0.000 b	0.004 ± 0.000 b	0.004 ± 0.000 b	0.007 ± 0.001 a	Nd	0.007 ± 0.000 a
Ellagic acid derivative	0.034 ± 0.001 b	0.034 ± 0.001 b	0.034 ± 0.003 b	0.11 ± 0.02 a	0.05 ± 0.01 b	0.047 ± 0.001 b
Protocatechuic aldehyde	Nd	0.043 ± 0.009 a	Nd	Nd	Nd	Nd
trans-Piceid	Nd	0.019 ± 0.001 c	Nd	Nd	0.89 ± 0.01 a	0.29 ± 0.02 b
trans-Resveratrol	0.039 ± 0.001 c	0.061 ± 0.007 b,c	0.050 ± 0.005 c	0.08 ± 0.02 b	0.083 ± 0.007 b	0.26 ± 0.01 a
Catechin	2.57 ± 0.05 c	3.9 ± 0.3 a	3.49 ± 0.04 b	1.667 ± 0.008 d	1.20 ± 0.02 e	1.9 ± 0.1 d
Epicatechin	2.2 ± 0.4 a,b	2.4 ± 0.1 a	1.67 ± 0.04 c,d	1.88 ± 0.09 b,c	1.2 ± 0.2 d,e	0.8 ± 0.1 e
Epicatechin gallate	0.22 ± 0.02 b,c	0.290 ± 0.001 b	0.204 ± 0.005 c	0.27 ± 0.05 b,c	0.265 ± 0.007 b,c	0.37 ± 0.06 a
Procyanidin B1	2.3 ± 0.1 b	3.2 ± 0.2 a	3.22 ± 0.08 a	1.72 ± 0.02 c	0.94 ± 0.05 d	1.1 ± 0.3 d
Procyanidin B2	0.62 ± 0.07 ab	0.50 ± 0.02 b,c	0.658 ± 0.006 a	0.57 ± 0.03 a,b,c	0.5 ± 0.1 c,d	0.339 ± 0.007 d
Quercetin-3-O-glucuronide	1.22 ± 0.06 b,c	1.1 ± 0.1 b,c	1.56 ± 0.02 a	1.24 ± 0.05 b	1.07 ± 0.03 c	0.54 ± 0.03 d
Quercetin-3-O-glucoside	0.159 ± 0.009 d	0.11 ± 0.01 e	0.277 ± 0.005 b	0.33 ± 0.02 a	0.20 ± 0.01 c	0.076 ± 0.004 f
Quercetin	0.001 ± 0.000 b	Nd	Nd	0.005 ± 0.000 a	0.005 ± 0.000 a	0.001 ± 0.000 b
Quercetin-3-O-galactoside	Nd	0.037 ± 0.009 a	Nd	Nd	Nd	Nd
Malvidin glucoside	2.1 ± 0.2 a	0.07 ± 0.02 c	0.35 ± 0.02 b
Cyanidin glucoside	0.11 ± 0.01 a	Nd	0.025 ± 0.001 b
Delphinidin derivative	0.104 ± 0.009 a	Nd	Nd
Cyanidin derivative	1.7 ± 0.1 a	Nd	0.07 ± 0.01 b
Unidentified anthocyanins	1.5 ± 0.1 a	0.180 ± 0.004 c	0.33 ± 0.03 b

Nd = not detected; ^a,b,c,d,e,f Letters in superscript indicate significant differences among grape stem varieties (p < 0.05).

), encompassing phenolic acids, flavan-3-ols, procyanidins, flavonols, stilbenes, and anthocyanins. Compared with SLE, PLE enabled the detection of nine additional compounds not previously observed, including syringic acid, ferulic acid, ellagic acid derivatives, protocatechuic aldehyde, trans-piceid, trans-resveratrol, monogalloyl glucoside, quercetin, and quercetin-3-O-galactoside (hyperoside). Phenolic acids were more diverse under PLE conditions, with trans-caftaric acid being the predominant acid (0.125–0.497 mg/g dry extract), and syringic acid showing exceptionally high levels in GW (10.3 mg/g dry extract) but also in lower amounts in the AB variety. Syringic acid was only detected in PLE extracts, SLE and UAE were totally inefficient for the recovery of this compound from grape stems (see Table 3, Table 4 and

Table 5. Phenolic composition of the UAE extracts (mg compound/g dry extract) (mean ± S.D.).

	UAE Extracts
	AB	CS	PV	RO	GW	RI
Gallic acid	0.164 ± 0.003 c	0.45 ± 0.05 a	0.58 ± 0.03 a	0.117 ± 0.003 c	0.25 ± 0.01 b	0.24 ± 0.01 b
Monogalloyl glucoside	0.03 ± 0.00 a	Nd	Nd	Nd	Nd	Nd
trans-Caftaric acid	0.602 ± 0.008 b	0.81 ± 0.08 a	0.4 ± 0.1 a,b	0.17 ± 0.03 c	0.20 ± 0.05 c	0.233 ± 0.003 c
trans-Piceid	Nd	0.019 ± 0.001 b	Nd	Nd	Nd	Nd
trans-Resveratrol	Nd	0.061 ± 0.007 b,c	Nd	0.04 ± 0.03 b	0.05 ± 0.01 b	0.110 ± 0.001 a
Catechin	3.88 ± 0.07 b,c	7.1 ± 0.7 a	3.5 ± 0.5 b	1.9 ± 0.3 d	1.8 ± 0.1 d	3.8 ± 0.1 c
Epicatechin	1.441 ± 0.001 a	1.4 ± 0.3 a	1.0 ± 0.6 b	1.1 ± 0.2 b	0.69 ± 0.04 b	0.665 ± 0.003 b
Epicatechin gallate	Nd	0.202 ± 0.002 b	Nd	Nd	Nd	0.32 ± 0.01 a
Procyanidin B1	4.02 ± 0.04 c	6.7 ± 0.5 a	0.64 ± 0.05 b	1.9 ± 0.3 d,e	1.3 ± 0.2 e	2.5 ± 0.2 d
Procyanidin B2	0.71 ± 0.03 c	1.1 ± 0.2 b	2.8 ± 0.4 a	0.69 ± 0.08 c	0.7 ± 0.2 c	0.78 ± 0.02 c
Quercetin-3-O-glucuronide	1.16 ± 0.08 a,b	1.08 ± 0.03 b	1.18 ± 0.08 a	1.1 ± 0.2 b	1.0 ± 0.1 b	0.493 ± 0.004 c
Quercetin-3-O-glucoside	0.41± 0.02 c	0.344 ± 0.007 c	0.46 ± 0.05 a	0.6 ± 0.1 a,b	0.43 ± 0.04 b,c	0.185 ± 0.002 d
Malvidin glucoside	3.2 ± 0.1 a	0.137 ± 0.007 c	0.36 ± 0.04 b
Cyanidin glucoside	0.16 ± 0.01 a	Nd	Nd
Cyanidin derivative	2.6 ± 0.1 a	Nd	0.06 ± 0.02 b
Unidentified anthocyanins	2.7 ± 0.4 a	0.30 ± 0.02 b,c	0.42 ± 0.03 b

Nd = not detected; ^a,b,c,d,e Letters in superscript indicate significant differences among grape stem varieties.

). In addition, it was observed that PLE enabled the extraction of trans-caftaric acid from the RO variety, whereas SLE was unable to extract this compound from that grape stem variety. Moreover, PLE allowed for the extraction of protocatechuic acid from all grape stem varieties, while SLE was only effective in extracting it from the RO variety and yielded lower amounts compared to PLE.

On the other hand, flavan-3-ols remained one of the major identified phenolic groups, with catechin (1.20–3.92 mg/g dry extract) and epicatechin (0.81–2.42 mg/g dry extract) exhibiting the highest abundances across all varieties and compared with the rest of the phenolic compounds identified in the extracts. However, catechin was more efficiently extracted using SLE, whereas the remaining flavan-3-ols exhibited higher extraction yields under PLE conditions.

As can be observed in Figure S1 and Table 4, procyanidins were also abundant, particularly procyanidin B1 (0.94–3.25 mg/g dry extract), but SLE extracts presented higher contents of procyanidin B2 than PLE extracts. PLE was the most efficient extraction technique for the recovery of flavonoids, particularly of quercetin-3-O-glucuronide and quercetin; however, higher extraction contents of quercetin-3-O-glucoside were obtained using SLE, except for RO variety, where PLE was more efficient. As in SLE, anthocyanins were restricted to the red varieties (AB, CS, and PV), with AB exhibiting the richest profile, dominated by malvidin glucoside, cyanidin derivatives, and unidentified anthocyanins. Overall, PLE enhanced the recovery of a wide diversity of phenolic compounds, providing a more comprehensive profile than SLE and revealing multiple compounds absent in the conventional extraction method.

3.2.3. Phenolic Profile of UAE Extracts

A total of 19 phenolic compounds were identified in the UAE extracts obtained from the six grape varieties (see Table 5). Compared with SLE and PLE, UAE extracts resulted in a distinct phenolic profile than those obtained by PLE. Notably, the UAE extracts lacked several phenolic compounds that were recovered under PLE conditions, such as protocatechuic acid, ferulic acid, ellagic acid derivatives, protocatechuic aldehyde, quercetin, and hyperoside. Particularly, phenolic acids were less diverse than in PLE, with only gallic acid and trans-caftaric acid consistently quantified (0.117–0.81 mg/g dry extract). In general, the extraction of trans-caftaric acid was more efficient using UAE than PLE. In addition, the UAE promoted the extraction of catechin, but other flavan-3-ols were more efficiently recovered using PLE, such as epicatechin and epicatechin gallate. UAE provided the most abundant content in procyanidins compared to SLE and PLE, particularly procyanidin B1, which reached its maximum concentration in CS (6.7 mg/g dry extract), and procyanidin B2, especially prominent in PV (2.8 mg/g dry extract). UAE also allowed the extraction of stilbenes, though at lower levels than PLE; trans-resveratrol ranged from 0.04 to 0.110 mg/g dry extract, while trans-piceid was detected only in CS at trace levels.

Regarding flavonols, although PLE extracts exhibited higher contents of quercetin-3-O-glucuronide, UAE yielded higher concentrations of quercetin-3-O-glucoside across all stem varieties. As in the other extraction techniques, anthocyanins were restricted to the red varieties (AB, CS, PV), with AB exhibiting the richest profile, dominated by malvidin glucoside, cyanidin, and cyanidin derivative, and unidentified anthocyanins, which were present at higher contents under UAE than under either SLE or PLE conditions.

Overall, UAE proved to be the most effective technique for maximizing the extraction of phenolic content, such as catechin, procyanidin B1 and B2, and anthocyanins, whereas PLE was the superior method for recovering the greatest structural diversity of phenolic compounds, enabling the extraction of nine additional phenolics absent in SLE and yielding the broadest compositional profile among the three extraction techniques.

3.3. Role of the Extraction Technique and Grape Variety on the Recovery of Antioxidant Phenolic Compounds from Stems

Principal Component Analysis (PCA) was applied to explore the multivariate structure of phenolic compound profiles extracted from grape stems of six varieties using three extraction techniques (SLE, UAE, PLE). The model captured a substantial portion of the dataset’s variability. The first two principal components (PC1 and PC2) accounted for 57.8% of the total variance (R²X [1] = 0.405; R²X [2] = 0.173), providing a robust basis for sample discrimination and visualization (see Figure 1A). The cumulative explained variance reached approximately 80% across four components, as shown in the R²X (cum) bar chart. However, the predictive ability (Q²) declined with each additional component and became negative for PC4, indicating that components beyond PC2 contribute little to model generalizability and may introduce noise. Therefore, interpretation was focused on the first two components, which effectively captured the dominant trends in phenolic composition and antioxidant capacity.

As can be observed, integrating Figure 1B (score plot) and Figure 1C (loading plot), clear relationships emerge among sample clusters and specific phenolic compounds. In this sense, CS and PV (all extraction techniques) clustered together in the positive PC1 region, strongly associated with TPC, antioxidant assays (DPPH, ABTS), and key phenolic acids and flavonoids (gallic acid, quercetin, protocatechuic acid). This positioning confirms that CS and PV stems exhibit the most robust phenolic and antioxidant profiles, consistently outperforming the other varieties regardless of extraction technique.

PLE extracts from AB, GW, and RO, formed a second cluster, also aligned with PC1 variables but slightly separated from CS and PV. This indicates that PLE significantly enhances phenolic recovery (TPC values and phenolic diversity), shifting these varieties toward higher antioxidant potential compared to their SLE and UAE. The relocation of AB-PLE from the intermediate group to the high-antioxidant cluster underscores the effectiveness of pressurized extraction in the recovery of antioxidant compounds.

RI (all techniques), RO-SLE, and AB-SLE/UAE grouped in an intermediate region, more closely associated with PC2 loadings, including flavonoids and anthocyanins (malvidin-3-glucoside, cyanidin-3-glucoside, epicatechin, resveratrol). These samples displayed diverse phenolic profiles but lower overall antioxidant capacity compared to CS and PV, reflecting varietal specificity rather than extraction efficiency.

Overall, the PCA clearly distinguishes grape stem varieties based on their phenolic content and extraction performance, identifying CS and PV as the richest sources of phenolics and highlighting PLE as the most effective technique for enhancing phenolic recovery and positioning extracts in regions associated with higher antioxidant potential.

4. Discussion

4.1. Role of Extraction Technique and Grape Variety in Phenolic Recovery and Antioxidant Capacity

The results confirm that the extraction technique significantly affects the extraction yield, with PLE > UAE > SLE for all grape stem varieties. In general, advanced extraction techniques have been associated with higher recovery of total phenolic compounds compared with SLE [14,18,19,20]. The enhanced performance of PLE can be attributed to its elevated temperature and pressure, which improve solvent diffusivity and promote the disruption of cell structures [21,22]. UAE could also increase extraction efficiency compared with SLE due to acoustic cavitation, which generates intense localized shear forces and microjets that disrupt cell walls and enhance mass transfer, although to a lesser extent than PLE. In general, these findings are consistent with previous reports showing that advanced extraction techniques allow for higher extraction yields than conventional methods, such as SLE [23,24].

When comparing grape stem color groups, red varieties consistently presented higher extraction yields, phenolic content, and antioxidant capacity than white varieties [25,26]. In this research, analytical differences among varieties cannot be due to soil and climate conditions of the cultivar or by vintage because all grape varieties belonged to the same geographical area and vintage. For this reason, differences among grape stems could be explained by well-established metabolic differences. This pattern may be explained by the higher activity of the phenylpropanoid and flavonoid biosynthetic pathways in red cultivars, which typically show increased expression of key structural genes such as PAL, CHS, F3′H and UFGT [27,28]. By contrast, white cultivars often carry loss-of-function mutations in the regulatory gene MYBA1, which prevent the activation of UFGT and other late flavonoid pathway genes, thereby reducing the metabolic flux toward anthocyanins and related flavonoids [29]. These regulatory and metabolic differences extend beyond the berry to supporting tissues such as stems, where red cultivars generally accumulate higher levels of flavan-3-ols and proanthocyanidins than white varieties [25,30]. In fact, Nieto et al. [14] obtained PLE extracts under the same extraction conditions as this study from the Merlot red grape stem variety, with similar TPC and antioxidant values to red grape stem varieties in this investigation. Therefore, the higher phenolic recoveries observed for red grape stems in the present study are consistent with their intrinsically greater phenolic biosynthetic capacity. These intrinsic differences between red and white grape stems highlight the importance of considering grape variety when selecting extraction strategies, as the naturally higher phenolic content in red stems allows for greater recovery of bioactive compounds.

The present study further contextualizes these findings within the broader literature. For instance, the PLE extract from CS stems exhibited higher yields than CS extracts obtained by SLE in the study by Spatafora et al. [23], illustrating the efficiency advantage of pressurized extraction. Simultaneously, the TPC and antioxidant activities of the PLE extracts obtained in this study were comparable to those reported for Callet, Manto Negro, Syrah, Tempranillo, Macabeu, Parellada, and Premsal Blanc stems [24], suggesting that advanced extraction methods provide a consistently high efficiency in recovering antioxidant phenolic compounds across diverse grape cultivars. Nevertheless, the limited data available in the literature for grape stem extracts show considerable variability in TEAC values, making direct comparisons among studies challenging. Among the few comparable reports, the extracts obtained in this study exhibited higher antioxidant capacity, as measured by the ABTS assay, than both the PLE and SLE extracts from grape stems previously studied by González-Centeno et al. [24] and Jiménez-Moreno et al. [18], respectively.

These results reinforce the conclusion that both the choice of extraction technique and the intrinsic characteristics of the grape stem variety are critical determinants of phenolic recovery.

4.2. Influence of the Extraction Technique and Grape Stem Variety in the Phenolic Profile of the Extracts

The comparative analysis of the three extraction techniques reveals that the differences observed are not merely quantitative but reflect the intrinsic mechanisms by which each extraction technique interacts with the grape stem matrix. PLE produced the most chemically diverse extracts, and this outcome can be explained by the combined effect of elevated temperature and pressure, which not only increases the solubility of compounds with medium-to-high polarity but also promotes the release of phenolics bound to the cell wall. This explains why protocatechuic acid, ellagic acid derivative, ferulic acid, syringic acid, trans-caftaric acid, and epicatechin reached their highest concentrations under PLE, since these compounds are often associated with structural components of the plant tissue and require harsher conditions to be liberated [31,32]. Similar observations have been reported in grape pomace valorisation, where PLE enhanced the recovery of bound phenolic acids and stilbenes compared to conventional methods [14,33]. Stilbenes are known to be present in glycosylated or bound forms, and the efficiency of PLE in disrupting these associations accounts for their superior recovery compared to UAE or SLE, particularly in RI and GW varieties [10]. This reinforces the notion that PLE is the most suitable technique when the goal is to maximize chemical diversity.

In contrast, UAE showed a clear affinity for flavan-3-ols, procyanidins, and anthocyanins. The mechanical effects of cavitation, which generate microjets and shockwaves, facilitate the rupture of cell walls and vacuoles, allowing the release of medium-sized phenolics that are otherwise less accessible [34,35,36]. As a result, compounds like catechin, procyanidin B1, and anthocyanins (malvidin-3-O-glucoside, cyanidin derivatives) are recovered at levels far exceeding those obtained by PLE, particularly in the AB variety. These findings are robustly supported by extensive peer-reviewed literature, microscopy evidence, and quantitative extraction data [37]. Anthocyanins are vacuolar pigments, and ultrasound is particularly effective in liberating them without requiring high temperatures, which could otherwise degrade their structure [38,39]. Malvidin-3-O-glucoside was the predominant anthocyanin, particularly enriched under UAE, which aligns with previous reports identifying it as the main pigment in grape stems [40,41]. Therefore, UAE emerges as the most suitable technique when the goal is to maximize pigment recovery or obtain extracts rich in oligomeric tannins. Anthocyanins were restricted to red varieties such as AB, CS, and PV, consistent with their pigmentation profiles. These differences may be attributed to the genetic regulation of phenolic biosynthesis, which determines the presence or absence of certain compounds in specific cultivars [42]. Such variability highlights the importance of tailoring extraction strategies not only to the target compounds but also to the grape variety under study. In fact, these differences were also observed in white varieties, due to syringic acid was abundant in GW but absent in red varieties.

SLE performed under mild conditions, consistently produced the lowest yields and the narrowest phenolic profile. This limitation is explained by the reduced solubilization capacity of conventional solvents at ambient conditions and the absence of mechanical or thermal forces to disrupt the plant matrix [14]. As a result, many compounds detected under PLE and UAE were absent in SLE extracts, and those present, such as gallic acid or catechin, appeared at moderate levels. In fact, stilbenes, particularly resveratrol and piceid, were absent in SLE extracts but efficiently recovered under PLE and UAE. Their presence is significant given their well-documented antioxidant and anti-inflammatory properties [43,44]. The superior performance of PLE in recovering stilbenes, especially in RI stems, can be attributed to its capacity to disrupt glycosidic bonds and liberate bound forms [10]. While SLE may be advantageous for preserving thermolabile compounds, its overall inefficiency makes it less suitable for the valorisation of grape stems when compared to advanced techniques [14].

Particularly, catechin emerged as the most abundant compound across most extracts, in agreement with previous studies that identified catechin as the predominant flavan-3-ol in grape stems [18,45,46]. Its prevalence is explained by its localization in solid grape tissues such as stems and seeds, from which it can be dissolved into wine during maceration. Quercetin derivatives were also consistently identified, with quercetin-3-O-glucuronide being the most abundant across extracts. This agrees with previous reports that glucuronidated and glycosylated quercetin forms are characteristic of grape stems [45,46]. The predominance of these derivatives is relevant since they exhibit strong antioxidant and anti-inflammatory activities [42].

Additionally, the PCA results demonstrate that both grape variety and extraction technique are decisive factors in shaping the phenolic landscape of grape stem extracts. CS and PV emerged as the most promising sources of antioxidant compounds, while PLE proved to be the most effective method for enhancing a wide diversity of phenolic recovery with high antioxidant capacity from grape stems. These findings highlight the potential of grape stems as valuable by-products for phenolic valorisation, with varietal selection and extraction optimization as key determinants of bioactive yield.

5. Conclusions

This study demonstrates that both grape variety and extraction technique critically determine the yield, composition, and antioxidant potential of grape stem extracts. CS and PV consistently exhibited the richest phenolic profiles and the highest radical scavenging activity, positioning them as the most promising candidates for valorisation. Among the extraction techniques, PLE produced the most comprehensive extracts, efficiently recovering phenolic acids, stilbenes, flavonols, flavan-3-ols, procyanidins, and anthocyanins. UAE excelled in the recovery of flavan-3-ols, procyanidins, and pigments, particularly in red varieties, while SLE remained limited in scope but provided moderate yields of certain phenolic acids such as gallic acid. The antioxidant capacity of the extracts was mainly driven by the presence of gallic acid, protocatechuic acid, catechin, epicatechin, procyanidins, quercetin derivatives, and anthocyanins, compounds especially abundant in CS and PV stems. The varietal specificity observed, such as syringic acid in GW, trans-resveratrol in RI, and malvidin-3-O-glucoside in AB, could be attributed to the genetic and metabolic diversity of grape cultivars. It was observed that the selection of the extraction technique was dependent on the grape stem variety studied. To our knowledge, this is the first time that phenolic compounds from AB, RO, GW, and RI stems have been identified and quantified by HPLC-DAD, expanding the chemical characterization of grape stem by-products.

In conclusion, grape stems represent sustainable and valuable agro-industrial by-products, rich in bioactive molecules with strong antioxidant potential, reinforcing their applicability in nutraceuticals, functional foods, cosmetics, and other bioactive products within a circular economy framework.