Phytochemical Characterization of Extracts From Aesculus hippocastanum L. Seeds and Quantification of Saponins Using HPLC‐DAD and a Novel Tensiometric Approach Authors: Khadijeh Nosrati Gazafroudi, Rieka Dreher, Morwen Liebscher, Peter Lorenz, Rolf Daniels, Florian C. Stintzing, Dietmar R. Kammerer Published: 2026 ABSTRACT ABSTRACT Aesculus hippocastanum L. is a well‐established medicinal plant used for the treatment of venous disorders and inflammation. Despite its wide therapeutic application, systematic comparisons of the phytochemical profile of horse chestnut seed extracts prepared according to pharmacopeial extraction protocols remain scarce. Consequently, this study aimed to characterize and compare the metabolite profiles of hydroalcoholic and aqueous fermented seed extracts prepared according to official pharmacopeial procedures. Extracts were analyzed using high‐performance liquid chromatography coupled to diode array and electrospray ion‐trap mass spectrometric detection (HPLC‐DAD‐ESI‐MSn) and gas chromatography‐mass spectrometry (GC‐MS). Additionally, a novel approach was developed for the semi‐quantification of saponins by tensiometry, which was compared with the quantification of UV‐active aescin derivatives by RP‐HPLC‐DAD. The HPLC‐DAD‐MSn analyses revealed different phytochemical compositions that include triterpene saponins, flavonoid glycosides, and indoleacetic acid‐N‐glycosides. GC‐MS analyses mainly revealed the presence of fatty acids, monosaccharides, and lactic acid. The hydroalcoholic extract showed higher aescin contents by RP‐HPLC‐DAD compared to the aqueous fermented extract, indicating better recovery rates with this extraction procedure. These findings underline the importance of process parameters in pharmaceutical manufacturing and support the use of complementary analytical approaches for comprehensive quality evaluation of complex herbal extracts. Introduction Medicinal plants represent an ancient foundation of healthcare [1]. Among these, the seeds of Aesculus hippocastanum L. commonly known as horse chestnut have a long history in European traditional medicine [2]. It belongs to a group of approximately 13 deciduous species in the genus Aesculus L. (Family: Sapindaceae) [3]. The horse chestnut tree is native to the Balkan Peninsula. The Ottomans introduced it to Europe in the 16th century as horse feed and for medicinal purposes. Upon spreading throughout Central and Western Europe, A. hippocastanum became a popular ornamental tree in parks, on boulevards, and in public gardens [2]. Horse chestnut trees are characterized by their impressive stature and palmately compound leaves. Their seeds are inedible and covered in spiny shells [4]. Historically, the seeds, bark, and leaves of horse chestnut trees have been employed for medicinal purposes [2]. In European folk medicine, preparations of the seeds and bark have been used to treat venous disorders, swelling, hemorrhoids, and inflammatory diseases [5]. These ancient applications form the basis for its use in modern phytotherapy. Contemporary assessments, including those by the European Union's Committee on Herbal Medicinal Products (HMPC), support this use. The HMPC has recognized the application of A. hippocastanum seed extracts for treating chronic venous insufficiency as “well‐established use” [6]. Today, horse chestnut extracts, particularly those obtained from seeds are among the most widely used herbal remedies for treating chronic venous insufficiency (CVI), varicose veins, and related conditions [7, 8]. The therapeutic effects of horse chestnut are primarily attributed to aescin [9], which is a mixture of more than 30 triterpenoid saponins [10]. Its individual constituents have an amphiphilic structure consisting of a triterpenoid aglycone, the apolar moiety of the molecule, which is linked to one or more glycan substituents, representing the polar part of the molecule [11]. β‐Aescin has been described as the main bioactive component of aescin [12]. The amphiphilic nature of triterpenoid saponins is responsible for their diverse biofunctional properties. For example, aescin exhibits anti‐inflammatory, antiexudative, and vasoprotective effects. It reduces capillary permeability, improves venous tone, and promotes lymphatic drainage [13, 14]. In addition to saponins, A. hippocastanum also contains other primary and secondary metabolites, including flavonoids such as quercetin and kaempferol glycosides, coumarins, such as esculin and fraxin, tannins, sterols, fatty acids, and polysaccharides [15, 16, 17]. The field of natural product developments has attracted considerable interest from scientists worldwide. One of the primary advantages of natural products over synthetic medicines is their reduced incidence of side effects [18]. Additionally, medicinal plants are the primary source of natural drugs for the majority of the human population [19]. In the 2019 Global Report on Traditional and Complementary Medicine, the World Health Organization (WHO) highlighted the increasing global relevance of traditional and complementary medicine as therapeutic options. According to WHO data, 170 of the 194 member states officially acknowledge traditional and complementary medical practices [20]. This emphasizes the growing acceptance of traditional healing methods within modern healthcare systems, with medicinal plants playing a key role in therapy. Thus, improving knowledge of plant constituents and their medical potential remains an important scientific issue. As an example, in Germany, the regulation of herbal medicinal products is outlined by the Medicinal Products Act, which determines requirements concerning efficacy, safety, and quality [21]. The European Pharmacopoeia, as well as national pharmacopoeias such as the German Homoeopathic Pharmacopoeia (GHP) provide comprehensive monographs for the manufacture of pharmaceutical extracts. However, specific extraction conditions result in multifaceted plant extracts comprising complex mixtures of substances, which play a crucial role in the treatment of various diseases. The aim of this work was to study different pharmaceutical extraction methods described in authentic national pharmacopeia. Despite the widespread use of herbal medicines, there is comparatively little knowledge about the complex phytochemical composition of such extracts. Commercial preparations obtained according to the European Pharmacopeia vary in solvent polarity and processing, ranging from oil‐based extracts to hydroalcoholic macerations to aqueous fermented extracts. In our study we focused on hydroalcoholic and aqueous fermented extracts. Advanced chromatographic and mass spectrometric methods such as HPLC‐DAD‐ESI‐MSn and GC‐MS were carried out to achieve a comprehensive phytochemical characterization of polar and low‐molecular weight compounds. Moreover, we developed a novel approach for the semi‐quantification of aescin by surface tension measurements via bubble pressure tensiometry. For comparing different methods, we further quantified aescin concentrations by RP‐HPLC‐DAD, which is used today as a gold‐standard technique. While the chromatographic approach represents a well‐established method, its high specificity may limit the detection of constituents without UV‐activity. The tensiometric method provides a rapid and cost‐effective alternative, which can be used for for example in‐process controls. Accordingly, the present study aimed to (i) characterize and compare the metabolite profiles of hydroalcoholic and aqueous fermented A. hippocastanum seed extracts, (ii) determine their major UV‐active aescin constituents by RP‐HPLC‐DAD, and (iii) evaluate bubble‐pressure tensiometry as a complementary semiquantitative screening approach for surface‐active constituents expressed as aescin equivalents. To the best of our knowledge, comprehensive phytochemical analyses and a comparison of hydroalcoholic and aqueous fermented pharmaceutical extracts recovered from A. hippocastanum L. seeds have not been conducted so far. Results and Discussion Analysis of Polar Secondary Metabolites by RP‐HPLC‐DAD‐ESI‐MSn To expand our knowledge of polar and semi‐polar secondary constituents of different horse chestnut pharmaceutical extracts, a qualitative analysis was performed using RP‐HPLC‐DAD‐ESI‐MSn in negative ionization mode. For this purpose, we investigated hydroalcoholic (HE) and aqueous fermented (AE) extracts from horse chestnut seeds prepared according to GHP methods 34c and 12m. Representative base peak and UV chromatograms (BPC and UVC) of the extracts are illustrated in Figure 1. In total 42 compounds were assigned based on their UV/vis characteristics, retention times, and fragmentation patterns obtained in collision induced dissociation (CID) experiments in comparison with data reported in bibliographic references. This data are summarized in Table 1 (peak numbers refer to Figure 1). However, it is important to acknowledge the limitations of UV spectroscopy. Signal intensity depends on both the analyte concentration and the individual molar extinction coefficient. These factors can lead to irregularities, thus hampering compound assignments. In this context, the complementary use of HPLC‐DAD‐MS helps to overcome these limitations and improves the reliability of compound characterization. Moreover, formate adducts ([M‐H+46]−) and lactic acid adducts ([M‐H+90]−) were formed due to the presence of formic acid in the mobile phase and the synthesis of lactic acid in the fermented extract. At low retention times, several indole‐3‐acetic acid (IAA)‐N‐glycosides were assigned in both extracts (compounds 1–4). These substances were primarily observed at retention times between 20 and 26 min. Their UV absorption maxima occurred at approximately 200–220 nm and 270 nm, which is consistent with IAA absorption behavior described in the literature [22]. The specific molecular ions corresponded to di‐ and triglycosylated derivatives. Neutral losses of 162 and 132 Da upon CID indicate the presence of hexosyl and pentosyl moieties. The detection of a base peak signal at m/z 174 in the MS2 or MS3 spectra could be attributed to the loss of the glycoside residue. Moreover, the base peak at m/z 130 in the MS3 spectra is an indication of a 3‐methylindole fragment, which is formed from the indole acetyl moiety. While indoleacetic acid derivatives have occasionally been reported in various plants, among others in some Aesculus species [23], they have not yet been documented or described for extracts of A. hippocastanum L. To the best of our knowledge, this is the first time that IAA‐N‐glycosides have been detected in extracts of A. hippocastanum L. IAA is a phytohormone that belongs to the auxin class. Although auxins are associated with several biofunctional properties, including the modulation of oxidative stress and inflammation [24], this class of compounds has thus far received considerably less attention than other compound classes, such as triterpene saponins and flavonoids. The structural diversity observed here suggests that IAA‐N‐glycosides represent a previously underestimated class of metabolites in this species. From a phytochemical perspective, this broadens the chemotaxonomic profile of A. hippocastanum and raises questions regarding their potential biological relevance in pharmaceutical preparations derived from such plant species. Furthermore, at low retention times also coumarin and phenylpropanoid derivatives were detected, including 6‐methoxy‐3,4‐dihydrocoumarin, fraxetin, and p‐coumaric acid glycosides. 6‐Methoxy‐3,4‐dihydrocoumarin glycoside (compounds 8 and 9) was present in both extracts. Slight differences in retention times may have been caused by matrix effects, with co‐eluting matrix components interacting with the stationary phase, resulting in changes of retention characteristics [25]. The coumarin glycoside derivatives (m/z 357) produced fragment ions at m/z 177 (MS2) and m/z 133 (MS3), thus indicating glycosidic cleavage and loss of CO2, respectively, which is typical of coumarins upon fragmentation [26]. Benzoyl‐p‐coumaroyl glycoside derivatives (compounds 13 and 14) exhibited fragment ions at m/z 163 and m/z 145 corresponding to the p‐coumaroyl moiety and thus, confirming ester‐linked phenylpropanoid structures. However, several compounds formed adducts with lactic acid (90 Da) in the AE. The presence of lactic acid derivatives indicates interactions in the aqueous preparation, which may only occur upon fermentation, thus providing analytical evidence that manufacturing processes directly affect the chemical composition of the resulting preparations. While such process‐dependent chemical modifications have rarely been considered in phytochemical studies, they may be relevant for the quality assessment of fermented herbal preparations. The retention‐time range from 37 to 62 min was dominated by flavonoid glycosides. Quercetin and kaempferol glycosides dominated both extracts. Neutral losses observed upon CID can be used to determine the type of glycosidic substitution. Thus, differences of 132 Da are associated with an O‐pentosyl moiety, 146 Da with a deoxyhexosyl moiety, 162 Da with the loss of an O‐hexosyl substituent, 294 Da with a pentosyl‐hexoside moiety, 308 Da with a deoxyhexosyl‐hexoside moiety, and finally 324 Da with a di‐hexoside. Upon neutral loss of the glycosyl moieties, quercetin glycosides produced the characteristic aglycone fragment at m/z 301, whereas kaempferol derivatives resulted in the corresponding aglycone fragment at m/z 285. Comparison of the UV chromatograms revealed components 22 and 23 to be more abundant in HEs as compared to fermented extracts. Conversely, component 25 was found in higher concentrations in fermented extracts. RP‐HPLC‐DAD‐ESI‐MSn base peak chromatograms (BPC) (negative ionization mode) and UV chromatograms (UVC) of aqueous fermented and hydroalcoholic extracts from Aesculus hippocastanum L. TABLE 1Physicochemical characteristics, occurrence, and peak assignment of metabolites detected in aqueous fermented (AE) and hydroalcoholic (HE) extracts from Aesculus hippocastanum L. using RP‐HPLC‐DAD‐ESI‐MSn (negative ionization mode).Peak no. t R Peak assignmentλmax Mass spectrometric data [m/z]ReferenceOccurrence(min)(nm)MS1 MS2 MS3 HEAE120.7IAA‐N‐(Hex)2 222, 270498454, 336, 292200, 161, 130, 101[37]—x221.6IAA‐N‐(Hex)2 222, 272498336, 292, 174200, 161, 130, 101[37]x—324.0IAA‐N‐Pent‐Hex204, 270468, 426336, 292200, 161, 130, 101[38], t.a.—x424.6IAA‐N‐Pent‐Hex204468424, 306262, 130[38], t.a.—x524.8Catechin‐hexoside220, 274451289, 245245, 205, 179[39, 40]x—625.6IAA‐N‐Pent‐(Hex)2 270630468, 424, 306262, 200, 174, 130[37], t.a.—x726.3IAA‐N‐Pent‐(Hex)2 222, 270630586, 468, 424, 306262, 174, 130(37), t.a.x—826.96‐Methoxy‐3,4‐dihydrocoumarin‐8‐O‐glycoside204 e 357177133[26, 41], t.a.—x927.86‐Methoxy‐3,4‐dihydrocoumarin‐8‐O‐glycoside206, 238, 312357177133[26, 41], t.a.x—1027.8Fraxetin‐Pent‐Hex a , e 501250, 206164t.a.—x1128.9Di‐O‐methyl‐ellagic acid‐deoxyhexosyl‐pyranoside d 204 e 565475, 419329, 161[42], t.a.—x1232.3IAA‐N‐Pent‐Hex220, 270468306, 262262, 200, 174, 130[37]xx1334.3 p‐Coumaric acid benzoyl‐glycoside d a , e 519, 475429265, 163, 145[43]—x1434.8 p‐Coumaric acid benzoyl‐glycoside c 232, 280sh, 362475429, 325, 265, 235163, 145[43]x—1535.5Dihydroferulic acid‐HMG‐glycoside222, 280501305, 195180[44, 45, t.a.x—1636.5IAA‐N‐(Pent)2 204, 270438306262, 200, 174, 131[38], t.a.—x1736.9Quercetin‐O‐hexoside‐pentosyl‐hexoside202, 266, 350757595301, 271, 179t.a.—x1837.0Gallocatechin‐pentoside226, 268, 358438306262, 174, 130[44], t.a.x—1937.3Di‐coumaroyl‐pentoside d a , e 549459, 265235, 163t.a.—x2037.7Di‐coumaroyl‐pentoside c 228, 274, 360505459, 265235, 205, 163, 145t.a.x—2139.1Gallocatechin 3‐(4‐methyl‐gallate)230, 270, 362471427, 305, 165150[b]x—2241.1Quercetin‐O‐dihexosyl‐pentoside204, 256, 354757625, 301301[46]xx2342.0Quercetin‐O‐hexosyl‐pentoside256, 352595301271, 255, 151[2]xx2446.1Quercetin‐O‐hexoside204, 256 354463301271, 229, 179, 151[45]—x2546.5Kaempferol‐O‐dihexosyl‐pentoside266, 350741609285, 254, 151[44, 47]xx2647.2Flavonoid derivative204, 266, 350367331, 303286, 269, 187t.a.—x2747.4Kaempferol‐O‐hexosyl‐pentoside266, 346579429, 284255, 227, 151[2]x—2850.8Quercetin‐3‐O‐deoxyhexosyl‐hexoside (Rutin)204, 256, 350609301271, 229, 179, 151[2]x—2951.4Quercetin‐7‐O‐deoxyhexosyl230, 358447301271, 229, 179, 152[48]x—3055.8Quercetin‐O‐hexoside204, 350463301193, 179, 151[45]—x3156.0Kaempferol‐3‐O‐ deoxyhexosyl‐hexoside264, 342593285257, 241, 151[2]x—3257.3Kaempferol‐3‐O‐deoxyhexosyl‐pyranoside (Afzelin)234, 266sh, 358431285255, 229, 163[2]x—3357.6Procyanidin B1204575525, 485, 395359, 381t.a.—x3458.3unknown204567387, 351—[a]—x3560.3Aesculuside B204 e 1048687, 615555, 547, 507[49]—x3662.4Quercetin256, 370301257, 179, 151107[45, 50]—x3765.8Aesculiside A23610901047, 866, 729, 657641, 609, 589, 523[51, 52], t.a.x—3869.5Escin Ia/Ib21011301086, 950, 906, 769, 697629, 563, 525Standard, [30]xx3969.9Escin Ia/Ib20811301086, 950, 906, 769, 698654, 629, 564Standard, [30]xx4070.4Isoescin Ia/Ib22011301086, 950, 906, 698631, 629, 599Standard, [30, 53]xx4170.6Isoescin Ia/Ib22411301086, 906, 769, 697650, 629, 597Standard, [30, 53]xx4273.2Aesculioside C/D2281170, 1140946, 809, 738669, 650, 573[51], t.a.xxAbbreviation: t. a.: tentatively assigned; x: detected and—not detected. a ND—not detected. b Characterized based on MSn data and public database of FooDB and MassBank. c Formate adduct ([M‐H+46]−). d Lactic acid adduct ([M‐H+90]−). e Due to very low amounts, the typical absorption maxima are lacking.John Wiley & Sons, Ltd. Physicochemical characteristics, occurrence, and peak assignment of metabolites detected in aqueous fermented (AE) and hydroalcoholic (HE) extracts from Aesculus hippocastanum L. using RP‐HPLC‐DAD‐ESI‐MSn (negative ionization mode). Abbreviation: t. a.: tentatively assigned; x: detected and—not detected. ND—not detected. Characterized based on MSn data and public database of FooDB and MassBank. Formate adduct ([M‐H+46]−). Lactic acid adduct ([M‐H+90]−). Due to very low amounts, the typical absorption maxima are lacking. This finding is indicative of a higher concentration of quercetin glycosides in the HE and of higher concentrations of kaempferol glycosides in the fermented extract. The presence of quercetin and kaempferol glycosides in the extracts examined is consistent with the findings of previous investigations of horse chestnut seeds. Kapusta et al. identified 13 flavonoids in A. hippocastanum seeds, with di‐ and tri‐glycosides of quercetin and kaempferol as the principal forms [27]. Furthermore, Dias et al. demonstrated in their study that flavonol glycosides constituted a percentage of over 90% of the phenolic fraction of seed‐kernel extracts, and that the ethanol concentration was the most substantial extraction parameter [28]. Nevertheless, it is important to acknowledge that quercetin glycosides present in the fermented extract were partially hydrolyzed by microorganisms upon fermentation, thus releasing their aglycone [29]. This finding underscores the importance of solvent polarity and manufacturing process in modulating the qualitative and quantitative polyphenol profile of pharmaceutical extracts. The retention‐time range from 60 to 73 min was dominated by triterpene saponins, including aesculuside B and aescin derivatives. In particular, the four major saponins (aescin Ia, aescin Ib, isoaescin Ia, and isoaescin Ib) could be assigned in both extracts. To ensure clear compound characterization and classification, a comparison of MSn spectra with those of authentic standards and literature data was conducted. All aescin derivates exhibited deprotonated pseudomolecular ions at m/z 1130 ([M–H]−). Subsequently, fragmentation in the MS2 experiments indicated characteristic losses of acyl substituents yielding fragments at m/z 1086 ([M–H–C2OH3]−) and sugar residues at m/z 906 ([M–H–C6H11O6]‐−). The fragment at m/z 629 in the MS3 experiment indicates the loss of a sugar residue (C6H11O6 −) and an acetylacetonate substituent (C5H7O2 −). Furthermore, the signal intensities of aescin components were higher in the UVC of the HE, suggesting that this extraction method yields higher concentrations of amphiphilic triterpene saponins. This observation is of particular importance for pharmaceutical quality, as, according to the European Pharmacopoeia, aescin is the main compound class of Aesculus hippocastanum seeds displaying pharmacological activity. These results are consistent with those reported in previous studies of A. hippocastanum seeds. Abudayeh et al. identified four major saponins and determined that aescin Ia was the most abundant constituent in both the seed endosperm and seed skin samples [30]. Furthermore, Dias et al. demonstrated in their study that flavonol glycosides constituted more than 90% of the phenolic fraction of seed‐kernel extracts, and that the proportion of ethanol was the most significant extraction parameter. Consequently, the present LC‐MS profiles are consistent with the previously established composition of the aescin fraction. In addition, the higher UV absorption exhibited by the aescin constituents in the hydroalcoholic extract is consistent with the significant influence of ethanol on saponin extraction, as reported by Dias et al. [28]. In sum, the comparison of different manufacturing methods revealed that the preparation method has a notable impact on the secondary metabolite profile of A. hippocastanum seed preparations. Thus, these findings underscore the importance of analytical characterization of pharmaceutical extracts to guarantee their reproducible and well‐defined chemical composition. In addition to the confirmation of the presence of previously characterized major constituents, for example, aescin and flavonoid derivatives, this study expands our knowledge on the chemical understanding of A. hippocastanum seeds by demonstrating a novel compound class, that is indoleacetic acid‐N‐glycosides, which were detected for the first time in A. hippocastanum. Altogether, these findings highlight the complexity of horse chestnut seed phytochemical composition and provide a significant basis for subsequent bioactivity investigations. Analysis of Low‐molecular Weight Constituents by GC‐MS GC‐MS investigations of hydroalcoholic and aqueous fermented horse chestnut seed extracts were carried out to analyze the composition of low‐molecular weight metabolites. To facilitate analysis, all samples were silylated with BSTFA (N,O‐bis(trimethylsilyl)trifluoro‐acetamide) to obtain trimethylsilyl (TMS) derivatives. Derivatization was performed to enhance the volatility of low‐molecular weight compounds and to make them GC compatible. All detected constituents exhibited a characteristic fragment ion at m/z 73, corresponding to the [(CH3)3Si]+ cation. Figures 2 and 3 illustrate the corresponding chromatograms. Compound assignments, retention times, molecular and fragment ions detected upon EI are detailed in Table 2. Furthermore, identification of the compounds was facilitated by automated comparison with the NIST (National Institute of Standards and Technology) database. The chromatogram of the AE was dominated by lactic acid (t R 7.79 min), which exhibited a molecular ion at m/z 234 and characteristic fragments at m/z 147 (base peak) and m/z 117. Previous studies have shown that the transformation of aqueous extracts performed according to GHP relies on spontaneous fermentation by plant‐specific microbial communities. The procedure involves a specific temperature regime with periodic cooling during the first seven days and continued fermentation for at least six months [31]. Furthermore, the study by Sauer et al. comprised variants in which whey was incorporated as a fermentation additive prior to the start of fermentation, in accordance with GHP Method 34c. In these preparations, the used whey contains members of the genera Lactococcus, Brachyspira, Lactobacillus, and Clostridium. Interestingly, on day three of fermentation, most of these taxa were not detected any more, indicating that they did not further affect communities and bacterial fermentation [31]. Our finding of lactic acid strongly indicates fermentation‐driven metabolic conversion, as lactic acid is not a primary storage metabolite of horse chestnut seeds. Specifically, lactic acid is produced by Lactobacillus species upon fermentative metabolism of saccharides (glucose/lactose), which are transformed to generate energy [32]. This is consistent with the results displayed in the chromatograms, which revealed significantly more saccharide derivatives in HE compared to AE. Furthermore, erythritol (t R 19.72 min) and sorbitol (t R 31.26 min) were identified in the chromatogram of the fermented extract. These two sugar alcohols (polyols) are formed as metabolic by‐products during microbial fermentation due to reduction reactions, thus partially converting carbohydrates into polyols [33]. Additionally, succinic acid (t R 14.17 min, m/z 262 as TMS derivative) was detected in both extract types. These findings indicate that succinic acid is an intrinsic seed metabolite. Semi‐quantitative evaluation based on AUC values revealed significantly higher amounts of succinic acid in the fermented extract. Previous research on lacto‐fermentation has demonstrated succinic acid to be a metabolite produced by lactic acid bacteria upon fermentative conversion of saccharides [34, 35]. This observation is consistent with the higher levels of succinic acid observed in the fermented extract. GC‐MS total ion current chromatogram of A. hippocastanum L. aqueous fermented extract after derivatization with BSTFA. GC‐MS total ion current chromatogram of A. hippocastanum L. hydroalcoholic extract (HE) after derivatization with BSTFA. TABLE 2Compound assignment of substances detected by GC‐MS in (a) aqueous fermented, and (b) hydroalcoholic extracts. Corresponding GC chromatograms are illustrated in Figures 2, 3.tR (min)Peak assignmentMW (Da) b Mass spectrometric dataFragment m/z (Intensity %) (a) Aqueous fermented extract 7.79Lactic acid23459 (10), 66 (16), 73 (93), 74 (12), 75 (19), 88 (10), 117 (82), 118 (14), 132 (17), 147 (100), 148 (22),149 (12), 190 (45), 191 (38), 219 (15)11.82Pentanoic acid27659 (11), 69 (12), 73 (100), 75 (23), 103 (71), 133 (11), 147 (79), 148 (10), 159 (90), 160 (15), 177 (10)12.85Glycerol30873 (100), 103 (24), 116 (30), 133 (19), 147 (76), 148 (10), 205 (60), 218 (19)14.17Succinic acid26273 (46), 75 (21), 147 (100), 148 (15), 247 (14)16.18Lactic acid, dimer30673 (100), 75 (23), 117 (87), 147 (34), 163 (13), 191 (14), 218 (14), 219 (16)16.29Lactic acid, dimer30673 (100), 75 (18), 117 (98), 118 (10), 146 (36), 191 (13), 218 (13), 219 (15)19.72Erythriol41073 (100), 103 (18), 116 (18), 147 (40), 205 (31), 217 (52)22.18Hydrocinnamic acid31073 (89), 75 (12), 91 (10), 147 (63), 193 (100), 194 (14), 220 (18), 267 (12)29.45Saccharide derivative54073 (74), 147 (23), 191 (42), 204 (100), 205 (16), 217 (29)29.63Pinitol55473 (100),133 (24), 147 (42), 159 (14), 191 (21), 207 (12), 217 (57), 247 (19), 260 (39), 305 (22), 318 (22)30.35Mannose45273 (70), 147 (20), 191 (41), 204 (100), 205 (16), 217 (16)30.59Xylose43873 (100), 147 (41), 179 (76), 204 (88), 217 (30), 308 (25)31.26Sorbitol61473 (100), 103 (25),147 (56), 205 (69), 217 (50), 307 (19), 319 (82)32.41 D‐Galacturonic acid55473 (100), 75 (10), 147 (48), 191 (26), 204 (84), 217 (13), 292 (61), 293 (10), 306 (19) (b) Hydroalcoholic extract 10.41 N, N‐Dimethylamino‐methylidene‐trifluoro‐acetamide a 168168 (2), 99 (100), 69 (21), 56 (10)12.87Glycerol30873 (100), 75 (11), 103 (22), 117 (25), 133 (20), 147 (73), 205 (56), 218 (16), 299 (11)14.17Succinic acid26273 (81), 75 (22), 147 (100), 148 (11)24.18Lauric acid27255 (15), 73 (100), 75 (77), 117 (80), 129 (34), 131 (13), 132 (26), 145 (18), 257 (92)27.33Saccharide derivative54073 (72), 75 (12), 103 (13), 147 (26), 217 (100), 218 (15), 257 (22), 393 (14)28.25Saccharide derivative54073 (100), 75 (10), 147 (27), 217 (97), 218 (10), 437 (14)28.42Saccharide derivative54073 (100), 75 (10), 147 (33), 204 (88), 205 (17), 217 (34), 437 (18), 437 (10)28.61Protocatechuic acid37073 (79), 192 (100), 194 (16), 223 (12), 281 (10), 311 (21), 355 (32), 370 (56), 371 (12)30.36Mannose45273 (62), 147 (20), 191 (35), 204 (100), 205 (15), 217 (13)30.62Syringic acid34273 (100), 141 (12), 253 (37), 283 (11), 297 (71), 312 (65), 326 (85),327 (14), 342 (59)31.39Saccharide derivative49673 (60), 147 (27), 204 (100), 217 (8)32.44 D‐Galacturonic acid55473 (100), 75 (15), 147 (52), 191 (23), 204 (62), 217 (11), 292 (85), 293(12), 306 (25), 364 (13)34.31Palmitic acid32855 (10), 73 (100), 75 (49), 116 (77), 128 (27), 132 (21),145 (13), 313 (59)37.83Fatty acid derivative41857 (12), 73 (41), 75 (10), 147 (31), 183 (30), 315 (100), 316 (12)38.73Sinapic acid36873 (45), 75 (10), 279 (10), 323 (14), 338 (100), 353 (30), 368 (77)45.063,5‐Bis(trifluoromethyl)‐4‐hydroxyacetophenone a 27273 (100), 103 (11), 147 (30), 204 (19), 217 (26), 259 (97), 272 (74)45.28Glutaconic acid27473 (100), 103 (18), 147 (38), 207 (20), 217 (68), 259 (75), 348 (11)50.43Catechin65073 (48), 207 (16), 355 (42), 368 (100), 369 (21) a Silylation artifact. b Molecular weight of the TMS derivative.John Wiley & Sons, Ltd. Compound assignment of substances detected by GC‐MS in (a) aqueous fermented, and (b) hydroalcoholic extracts. Corresponding GC chromatograms are illustrated in Figures 2, 3. Silylation artifact. Molecular weight of the TMS derivative. However, the occurrence of monosaccharides eluting between 27 and 31 min was a notable characteristic of the HE. The fragmentation patterns of these compounds exhibited characteristic daughter ions at, e. g., m/z 217 (C3H3(TMSiO)2 +) and m/z 204 (C2H2(TMSiO)2 +), indicative of trimethylsilyl derivatives of carbohydrates and which are formed through cleavage of the sugar ring [36]. Moreover, the presence of various sugar derivatives suggests that either partial hydrolysis of polysaccharides or the extraction of free monosaccharides has occurred during extraction. Compared with the AEs (two monosaccharide derivatives), HE preparations exhibited five monosaccharides derivatives, indicating that in the AEs, microorganisms metabolized monosaccharides through fermentative processes, leading to their depletion. Among phenolic constituents, syringic acid (m/z 342) and protocatechuic acid (m/z 370) were identified as benzoic acid derivatives. The detection of sinapic acid (m/z 368) is consistent with TMS‐derivatized phenylpropanoids and supports previous reports describing the occurrence of phenolic acids in horse chestnut seeds [2]. However, their relatively low abundance compared to carbohydrate derivatives indicates that they represent minor constituents in the low‐molecular weight fraction. Moreover, GC‐MS analyses revealed the presence of fatty acids (tR 24.18 min and tR 34.31 min) in the HE. This occurrence and their absence in the fermented extract is consistent with their physicochemical characteristics, thus going along with a better solubilization of lipophilic compounds in ethanol‐containing systems. Finally, catechin (m/z 650 as penta‐TMS derivative) was detected at tR 50.43 min in the HE. Its mass spectrum exhibited a base peak signal at m/z 368. The detection of catechin by GC‐MS is consistent with the findings described for LC‐MS analyses, thereby demonstrating that derivatization facilitates the analysis of even nonvolatile compounds by GC‐MS. A comprehensive analysis of GC‐MS data indicates that the HE is distinguished by the presence of saccharide derivatives, fatty acids, and minor phenolic compounds, while the AE is predominantly characterized by lactic acid and sugar alcohols. While the HE is characterized by a metabolite profile more closely aligned with the intrinsic horse chestnut seed composition, the AE is a chemically modified preparation shaped by bacterial modifications. The presence of palmitic acid in the hydroalcoholic extract is consistent with the seed‐composition study of Čukanović et al., who identified 16 fatty acids in A. hippocastanum seeds and found palmitic acid to be the predominant saturated fatty acid [54]. The findings of the present study extend the results observed in the raw seed matrix to pharmacopeial hydroalcoholic preparations. The absence of detectable palmitic acid in the aqueous fermented extract may be related to their limited recovery in an aqueous matrix and/or to processing‐related changes. Nevertheless, a direct comparison with seed‐lipid studies remains limited by variations in sample matrix and analytical preparation. Furthermore, combining the results of LC‐MS analyses with those of GC‐MS analyses underscore that different extraction and manufacturing procedures result in preparations that may significantly differ in their phytochemical composition and, accordingly, their pharmacological profile, as demonstrated in our study. Quantification of Aescin Aescin, the principal bioactive constituent of Aesculus hippocastanum L. seeds, is recognized by the European Pharmacopoeia for its pharmacological relevance and is therefore widely accepted as a quality marker. In the present study, aescin concentrations of the aforementioned extracts were determined by two different analytical methods: (i) RP‐HPLC‐DAD quantification based on UV detection, and (ii) a novel semi‐quantitative approach based on dynamic surface tension measurements (tensiometry), exploiting the amphiphilic character of aescin. For both approaches, calibration was carried out with an external aescin standard. Table 3 shows the results of aescin quantitation in horse chestnut seed extracts. TABLE 3Aescin concentrations determined by HPLC‐DAD and bubble pressure tensiometry in Aesculus hippocastanum extracts (Ex), listed in descending order. Data are presented as mean ± standard deviation (n (HPLC‐DAD) = 2; n (Tensiometry) = 3).Dry matter per 1 mL ExHPLCTensiometrySample(mg)Aescin (µg E/ mg Ex)RSD (%)Aescin (µg EE/ mg Ex)RSD (%) n‐BuOHn.a375.1 ± 1.910.51558.0 ± 33.365.98HE55.0312.1 ± 0.040.33183.5 ± 11.266.14AE49.596.6 ± 0.060.8517.1 ± 1.096.36E: Aescin; EE: Aescin equivalents; RSD: Relative standard deviation.John Wiley & Sons, Ltd. Aescin concentrations determined by HPLC‐DAD and bubble pressure tensiometry in Aesculus hippocastanum extracts (Ex), listed in descending order. Data are presented as mean ± standard deviation (n (HPLC‐DAD) = 2; n (Tensiometry) = 3). E: Aescin; EE: Aescin equivalents; RSD: Relative standard deviation. Quantitative RP‐HPLC‐DAD investigations revealed significant differences in aescin contents among the various extract types. Results were expressed as µg aescin (E) per mg dry weight. The n‐butanol (BuOH) fraction, which was utilized as a reference extract, exhibited highest aescin concentration with 375.1 ± 1.91 µg E/ mg dry weight and a relative standard deviation (RSD) of 0.51%, followed by the HE with 12.1 ± 0.04 µg E/ mg dry weight (RSD 0.33%) and the AE with 6.6 ± 0.06 µg E/ mg dry weight (RSD 0.85%). The high aescin content of the BuOH fraction indicates a significant enrichment, and thus indicates n‐butanol to be a suitable solvent for the reference extract. Aescin is characterized by an amphiphilic structure, comprising a triterpenoid aglycone that is linked to single or multiple sugar moieties [11]. Due to its moderate polarity, aescin is efficiently recovered by medium‐polar organic solvents, such as n‐butanol. This physicochemical behavior aligns with findings of prior phytochemical studies on triterpene saponin‐containing plants [55]. The AE exhibited a notably lower aescin concentration, indicative of the reduced solubility of amphiphilic saponins in aqueous systems. In contrast, the HE exhibited approximately two‐fold higher aescin levels compared to the fermented extract, underscoring the importance of ethanol in extracting triterpene saponins. A novel tensiometric approach was developed for the semi‐quantification of surface‐active compounds, such as aescin. These components are characterized by their amphiphilic nature. Thus, they accumulate in aqueous systems at the interface between air and water, leading to a reduction in surface tension. In static surface tension measurements, the aforementioned effect is proportional to the logarithm of the concentration of the surfactants, below the critical micelle concentration (CMC) [56, 57, 58]. In our study, we used dynamic surface tension measurements and performed a calibration with aescin standard solutions, which enables the semi‐quantification of amphiphilic compounds in the extracts as aescin equivalents. The dynamic measurements provide a continuous relationship between surface tension and concentration over a broader concentration range, which is particularly advantageous for semi‐quantitative analysis. Results were expressed as µg aescin equivalents (EE) per mg dry weight. The BuOH fraction showed a high surface tension reduction, which was equivalent to a content of surfactants of 558.0 ± 33.36 µg EE/ mg dry weight (RSD 5.98%). The HE exhibited 183.5 ± 11.26 µg EE/ mg dry weight (RDS 6.14 %) followed by the AE, which showed a concentration of surfactants of 17.1 ± 1.09 µg EE/ mg dry weight, with a relative standard deviation of 6.36 %. Upon comparison of the two quantitation methods, tensiometry generally revealed higher values compared to RP‐HPLC‐DAD for all samples. This discrepancy can be explained by the analytical principles employed. RP‐HPLC‐DAD quantification is selective for UV‐absorbing aescin isomers, which are individually quantitated after chromatographic separation using an external aescin standard. In contrast, tensiometry provides a non‐specific quantification of all surface‐active constituents present in the extracts. Thus, higher aescin contents determined by tensiometry likely reflects contributions of other saponins or non‐saponin surface‐active compounds. These latter compounds affect surface tension but are not covered by RP‐HPLC‐DAD if they are not UV‐active and have not been characterized as surface‐active substances and calibrated using appropriate standards. Comparison of the results obtained in this study with previously published aescin contents of Aesculus seed extracts demonstrates comparable amounts for both AE and HE [2, 30, 53]. However, direct comparison of these values is hampered by the fact that compound concentrations are affected by a variety of factors, including solvent type, extraction method, and temperature. Overall, tensiometric analysis was found to be a novel, straightforward approach for aescin semi‐quantitative, whereas chromatographic methods are recognized as the gold standard for the selective quantification of individual components [59]. Thus, surface tension measurement provides a rapid and cost‐effective alternative that directly reflects the physiochemical property, which could be used for for example in‐process controls. Further, it may function as a complementary approach for conducting rapid screening and comparative assessment of different extracts. Conclusions In the present study, pharmaceutical extracts from Aesculus hippocastanum L. seeds were examined to provide a comprehensive phytochemical characterization and quantification. For qualitative purposes, complementary analytical approaches, including GC‐MS and RP‐HPLC‐DAD‐MSn were applied. RP‐HPLC‐DAD was used for quantitative saponin analysis and the results were compared with those obtained using a novel approach based on surface‐tension measurements. The complementary analytical approaches used in the present study enabled a comprehensive understanding of the phytochemical composition and extraction‐dependent variations of pharmaceutical preparations from horse chestnut seeds. The chromatographic analyses illustrated that plant‐derived extracts are multi‐constituent mixtures with a complex metabolite profile. Furthermore, LC‐MS analyses revealed the presence of flavonoid glycosides, coumarin derivatives, and triterpene saponins in both extract types. Beyond the aforementioned compound classes, a number of indole‐3‐acetic acid‐N‐glycosides (auxin derivatives) carrying one to three sugar moieties were detected in the HE and AE. To the best of our knowledge, this compound class has been detected for the first time in A. hippocastanum seed extracts, thus expanding the known secondary metabolite profile of this species. Furthermore, GC‐MS analyses of the extracts provided a detailed insight into the low‐molecular weight compound fraction. The HE was characterized by monosaccharide derivatives and fatty acids. Conversely, the AE was dominated by lactic acid, highlighting the impact of microbial fermentation. Minor short‐chain organic acids and sugar alcohols were also found in the AE. For quantitative investigations of saponins, two methods were carried out, an RP‐HPLC‐DAD method and a novel tensiometric approach. The results of both methods revealed differences between the extracts, with the n‐butanol extract (reference extract) exhibiting highest values. Based on the RP‐HPLC‐DAD results, the HE showed twofold higher aescin concentrations as compared to the fermented extract. Tensiometry exploits the intrinsic surface‐active properties of aescin. Although this non‐specific approach cannot replace chromatographic methods, it offers a novel, rapid, and simple alternative for assessing the contents of amphiphilic constituents. In conclusion, the findings of this study demonstrate that the extraction solvent and manufacturing process have a significant impact on the qualitative and quantitative phytochemical composition of Aesculus hippocastanum L. seed preparations. Hence, this study provides a foundation for future research on their pharmacological relevance, stability, but also for innovative formulations based on natural constituents. Future research should further investigate the relationship between extract composition and pharmacological activity, assess the stability of relevant constituents during processing and storage, and explore formulation strategies that preserve or enhance the characteristic phytochemical profiles. Moreover, the complementary analytical approaches applied in this study may support the development of improved quality assessment and process monitoring strategies for complex herbal preparations. Experimental Section Chemicals and Reagents The solvents n‐butanol (n‐BuOH), ethanol, acetonitrile, ethyl acetate (EtOAc), and dichloromethane (DCM) were from Chemsolute (Th. Geyer GmbH & Co., KG, Renningen, Germany). Anhydrous sodium sulfate was also obtained from Chemsolute (Th. Geyer GmbH & Co., KG, Renningen, Germany). Dimethylformamide (DMF) and N,O‐bis(trimethylsilyl)‐trifluoroacetamide (BSTFA) were from Sigma–Aldrich (St. Louis, Missouri, USA), and formic acid was from Fluka (Sigma Aldrich, St. Louis, MO, USA). Aescin was provided by abcr GmbH (Karlsruhe, Germany). Ultrapure water (ELGA Purelab Classic, High Wycombe, UK) was used for all experiments. Plant Material Aesculus hippocastanum L. seeds were collected in the Kurpark of Bad Boll (48°38'27.7''N 9°35'59.7''E, Baden‐Württemberg, Germany) in October 2023. The seeds were thoroughly cleaned to remove dirt and surface impurities, and damaged or contaminated seeds were discarded. Subsequently, the plant material was sealed in polyethylene zip bags and stored at −70°C for further investigations. The identity of the plant material was confirmed by Dr. R. Duque Thüs, and voucher specimens (voucher no.: HOH‐022978) were deposited at the herbarium of the Institute of Botany, Hohenheim University (Stuttgart, Germany). Authentic dried plant material was supplied by rosarum.leben gUG (Heiligenberg, Germany) and used for the preparation of the hydroalcoholic extract. The identity of the plant material was confirmed by the quality control department of WALA Heilmittel GmbH. Plant Material Extraction Hydroalcoholic extract Dried plant material was used for the preparation. The manufacturing process was conducted in accordance with GHP method 12 m. The seeds were crushed with a hammer and minced with a blender (MB 800 Microtron, Kinematica, Malters, Switzerland), until their size was small enough to get through a sieve with 1 mm pore size. For obtaining hydroalcoholic extracts, one part of the plant material was mixed with 3.1 parts (m/m) of purified water and 0.9 parts (m/m) of ethanol (96 per cent, v/v). The mixture was stored at a temperature of 20°C ± 2°C for 7 days. During maceration the mixture was stirred for approx. five minutes every morning and evening. After seven days the solution was filtered through filter paper (Whatman, Buckinghamshire, UK) yielding a clear solution, which was stored at a temperature of −70°C until further analyses. Two independent hydroalcoholic extraction replicates were prepared. Aqueous Fermented Extract Two commercially produced aqueous fermented extracts from WALA Heilmittel GmbH (Bad Boll/Eckwälden, Germany) were used for comparison. These were obtained according to the German Homeopathic Pharmacopoeia (Aesculus hippocastanum e semine ferm, monograph 34c). The two aqueous fermented preparations represented independently manufactured extract batches. n‐Butanol Extract The extraction process involved 50 g of frozen plant material, which was crushed using a hammer prior to extraction. A two‐step extraction process was carried out using a mixture of acetone and water (70/30, v/v). A volume of 250 mL of solvent was used for each step. In order to facilitate extraction, the suspension was subjected to a three‐minute mincing process using an Ultra‐Turrax device at a speed of 15000 rpm (IKAWerke GmbH and Co., KG, Staufen, Germany). Furthermore, in order to prevent oxidative degradation of plant compounds, the mixture was aerated with nitrogen for 10 min both prior to and after mincing. Subsequently, the mixture was placed at a temperature of 5°C overnight and then filtered using Celite (Carl Roth GmbH + Co. KG, Karlsruhe, Germany). The resulting residue underwent a second extraction as described above. The brownish filtrates resulting from both steps were combined, and acetone was removed through vacuum rotary evaporation. Subsequently, the aqueous extract underwent a liquid–liquid extraction with three portions of 100 mL each of dichloromethane, ethyl acetate, and n‐butanol using a separating funnel. After phase separation, the n‐BuOH fraction was further used for tensiometry and HPLC‐DAD. Therefore, the solvent was removed by vacuum rotary evaporation to obtain dry extracts for further analyses. Two independent n‐BuOH extraction replicates were prepared. Silylation of Extracts for GC‐MS Analyses For GC‐MS analyses, 10 mL aliquots of fermented aqueous and hydroalcoholic extracts were extracted with three portions of 20 mL EtOAc each. The combined organic layers were dried over anhydrous sodium sulfate, filtered, and evaporated using a rotary evaporator. Subsequently, 5‐10 mg of the residue were dissolved in 500 µL DMF and derivatized with 300 µL BSTFA. The solution was subsequently heated at 105°C for 15 min in order to obtain trimethylsilyl derivatives. GC‐MS Analysis For GC‐MS analyses, a PerkinElmer Clarus 500 gas chromatograph (PerkinElmer, Inc., Shelton, CT, USA) with split injection (split ratio 30:1, injection volume 1.0 µL) coupled to a single quadrupole mass spectrometer operating in electron ionization (EI) mode at 70 eV was used as described previously [45, 60]. The stationary phase for chromatographic separation was a Zebron ZB‐5MS capillary column (60 m × 0.25 mm i.d., 0.25 µm film thickness, 5 % phenylpolysiloxane and 95% dimethylpolysiloxane coating; Phenomenex, Torrance, USA) and helium was used as carrier gas at a flow rate of 1 mL/min. The injector temperature was 250°C, the temperature program of the column oven was 100°C–320°C, applying a linear gradient of 4°C /min and a final holding time of 30 min. Data were acquired and processed using the software TurboMass (v.5.4.2, PerkinElmer, Inc., Waltham, MA, USA). Each prepared sample solution was analyzed in duplicate under identical instrumental conditions. Compound assignments were based on comparison of the EI mass spectra with entries in the NIST mass spectral library (match factor > 800). RP‐HPLC‐DAD‐ESI‐MSn Analysis For HPLC analyses, the extracts were directly injected after filtration through a syringe filter (perfect‐flow RC, 0.45 µm, WICOM, Heppenheim, Germany). Reversed‐phase high‐performance liquid chromatography was carried out using an Agilent 1200 HPLC system (Agilent Technologies, Inc., Palo Alto, CA, USA) equipped with a binary pump, micro vacuum degasser, autosampler, thermostatic column compartment, and UV/vis diode array detector (DAD). The stationary phase was a Kinetex C18 reversed‐phase column (2.6 µm particle size, 150 mm × 2.1 mm i.d., Phenomenex Ltd., Aschaffenburg, Germany) and a pre‐column of the same material at 25°C with a flow rate of 0.21 mL/min. The mobile phase consisted of 0.1 % formic acid in water (eluent A) and of 100% acetonitrile (eluent B). An injection volume of 10 µL was used for each sample. The gradient protocol was as follows: 0–10 min, 0%–10% B; 10–22 min, 10% B; 22–53 min, 10%–23% B; 53–72 min, 23%–60% B; 72–80 min, 60%–100% B; 80–85 min, 100% B; 85–90 min, 100%–0% B; and 90–100 min, 0% B as established previously [45, 60]. The LC system was coupled to an HCTultra ion trap mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) equipped with an ESI source. All extracts underwent analysis in negative ionization mode with a capillary voltage of 4000 V, a dry gas (N2) flow of 9.00 L/min, capillary temperature of 365°C, and nebulizer pressure of 35 psi. Full‐scan mass spectra (mass range m/z 50–1000) of HPLC eluates were recorded during chromatographic separation. MSn data were acquired in the auto MS/MS mode through collision‐induced dissociation (CID). Instrument control was managed using ChemStation for LC 3D systems (Rev. B01.03 SR1 (204)) by Agilent and EsquireControl software (v7.1) by Bruker Daltonics. All analyses were run in duplicate. Quantification of Aescin by RP‐HPLC‐DAD Analysis For quantification of aescin contents in the extracts, HPLC‐DAD analyses were performed on a Dionex Ultimate 3000 RSLC system (Thermo Fisher Scientific GmbH, Dreieich, Germany), which was equipped with a binary pump HPG‐3400 A, an autosampler WPS‐3000 TSL, a thermostated column compartment TCC‐3000 SD and a diode array detector DAD‐3000. The aforementioned stationary phase and eluent system were also used for quantitation purposes applying an injection volume of 10 µL of each sample. The gradient protocol was as follows: 0–5 min, 0%–10% B; 5–10 min, 10%–35% B; 10–15 min, 35%–50% B; 15‐20 min, 50% B; 20‐30 min, 50%–55% B; 30‐40 min, 55‐60 % B; 40‐50 min, 60%–70% B; 50‐55 min, 70%–100% B; 55‐58 min, 100%; 58‐60 min, 100%–0% B, 60‐65 min, 0% B. Each prepared sample solution was analyzed in duplicate under identical instrumental conditions. All analyses were run in duplicate. Samples were centrifuged for 10 min at 12000 rcf (10,309 rpm) prior to injection into the HPLC system. A five‐point calibration curve was obtained by injection of aescin standard calibration solutions (2.00 mg/mL; 1.50 mg/mL; 1.00 mg/mL; 0.5 mg/mL; 0.25 mg/mL; 0.125 mg/mL). Detection of aescin isomers were performed at 210 nm. The AUC of total aescin, calculated as the sum of aescin Ia, aescin Ib, isoaescin Ia, and isoaescin Ib, were plotted against their concentrations. Aescin contents were calculated for all extracts using a linear regression equation (y = 304.01x – 5.0679; R 2 = 0.9997). In the regression equation, y represents the summed peak area of aescin Ia, aescin Ib, isoaescin Ia, and isoaescin Ib, and x represents the aescin concentration in mg/mL. The n‐BuOH extracts were dissolved in deionized water to achieve a concentration of 3 mg/mL before injection. The aqueous fermented extracts were diluted with deionized water at a ratio of 2:1 (v/v), whereas the hydroalcoholic extracts were diluted with deionized water at a ratio of 1:2 (v/v). Data acquisition and processing were performed with Chromeleon V 7.3.2 software (Dionex, Idstein, Germany). Semi‐Quantification of Aescin by Tensiometric Measurements For tensiometric measurements, the n‐BuOH, hydroalcoholic, and aqueous fermented extracts were used. All measurements were performed at a temperature of 22°C ± 1°C using 20 mL per measurement. The standard calibration solutions (2.00 mg/mL; 1.00 mg/mL; 0.5 mg/mL; 0.25 mg/mL; 0.125 mg/mL; 0.0625 mg/mL) were prepared by dissolving aescin standard in deionized water, homogenizing for one minute in an ultrasonic bath and vortexing until a clear solution was obtained. The calibration solutions were measured in triplicate using the dynamic measurement mode of the tensiometer (pro line t15, SITA Messtechnik GmbH, Dresden, Germany) with a capillary made of PEEK (polyether ether ketone). The dynamic surface tension demonstrates the time‐dependent adsorption of surfactants at newly formed interfaces, resulting in a concentration‐dependent signal over a broad concentration range. The surface tension values recorded at the specific bubble lifetimes (0.015 s – 20.00 s) were plotted against the logarithm of the concentrations of the calibration solutions to obtain calibration curves. For further evaluation the linear regression with the highest correlation coefficient was used (y = −2.126x + 69.407 R 2 = 0.9722). The n‐BuOH extracts were diluted with deionized water (0.25 mg/mL—1 mg/mL) and the aqueous fermented extracts were diluted with deionized water (1:4, v/v) until the obtained surface tension was in the linear range of the calibration curve. Prior to measurement, the hydroalcoholic extracts were evaporated to dryness using a rotary evaporator in order to eliminate any interference from ethanol. The dry extracts were then redissolved in deionized water and diluted (1:16, v/v) until the surface tension was within the linear range of the calibration curve. Each prepared calibration and extract solution was measured three times under identical conditions. For semiquantitative evaluation, the bubble lifetime yielding the highest correlation coefficient over the investigated calibration range was selected. In the regression equation, y represents the dynamic surface tension in mN/m and x represents the logarithm of the aescin concentration in mg/mL. The concentrations of aescin equivalents were calculated from the calibration equation, corrected for the respective dilution factors, and normalized to the extract dry mass. Determination of Dry Matter Contents A total of 25 mL porcelain crucibles were heated at 105°C for 24 h, cooled in a desiccator and weighed. Plant extracts were centrifuged, 5 mL aliquots of the clear supernatants were pipetted into the porcelain crucible in duplicate, dried at 80°C for 1 h and subsequently at 105°C for another 2 h. After cooling in a desiccator, the dry matter was quantitated gravimetrically. Duplicate determinations were performed using separately pipetted aliquots. Dry matter contents were calculated from the increase in crucible mass divided by the 5‐mL aliquot volume and expressed in mg/mL. Author Contributions K.N.G. and D.R.K. designed the study. K.N.G. and M.L. prepared the extracts and performed the analyses. K.N.G. wrote the draft. K.N.G., R.D., P.L., M.L., R.D., D.R.K., and F.C.S. evaluated the results and proofread the manuscript. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest K.N.G., R.D., P.L., M.L., D.R.K., and F.C.S. are employed at WALA Heilmittel GmbH. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.