Effect of Temperature of Two-Year Storage of Varietal Honeys on 5-Hydroxymethylfurfural Content, Diastase Number, and CIE Color Coordinates

Abstract

This study aimed to evaluate the effect of two-year storage of varietal honeys (buckwheat, linden, rapeseed, honeydew, and multifloral) at various temperatures (4 °C, −18 °C, −40 °C, and −80 °C) on the content of 5-hydroxymethylfurfural (5-HMF), diastase number (DN), and color assessed in the CIE L*a*b* system. The control samples were stored at room temperature (RT, ca. 20 °C). The results indicate that storing honey at low temperatures effectively mitigates undesirable quality changes, particularly enzymatic degradation and color alterations, while preventing excessive 5-HMF accumulation. After storage, a significant (p ˂ 0.01) decrease was noted in the diastase number (DN) of the honeys, regardless of the temperature (by ca. 66.7% at RT and by 53.1% to 58.3% at low temperatures, p > 0.05). Low storage temperatures led to higher enzymatic activity in buckwheat, linden, and honeydew honeys compared to rapeseed honeys. RT significantly (p ˂ 0.01) increased 5-HMF concentration by 79.3%, whereas the cold and frozen storage conditions increased 5-HMF concentration only by 25.1% at −18 °C and 33.2% at 4 °C. The greatest color changes manifested by significant (p ˂ 0.01) darkening, with a decrease in the h° value (p ˂ 0.01), and a lower contribution of the yellow color and a greater contribution of red color (p > 0.05) in the color profile were noted in the honeys stored at RT. Storage at this temperature resulted in a significantly (p ˂ 0.01) higher total color difference of the honeys (ΔE = 9.53) compared to the other temperatures tested (3.71 < ΔE < 5.58). The low storage temperatures may elicit a positive and comparable effect on preserving the satisfactory quality of the analyzed varietal honeys. It is noteworthy that this positive effect could already be achieved at a storage temperature of +4 °C without the need to apply frozen storage temperatures, which is essential given the economic and environmental concerns.

1. Introduction

Storage is the last stage of the honey distribution chain before its consumption. Its conditions affect the course of natural chemical and enzymatic reactions in honeys, which determine their ultimate quality, including their physical parameters and composition. The activity of α-amylase (expressed as the diastase number, DN) and the content of 5-hydroxymethylfurfural (5-HMF) are monitored at each stage of honey production and distribution to ensure its quality and freshness [1]. Both of these indicators are also stipulated in EU legal acts [2] as major criteria indicating changes that occur in honeys over the storage period, particularly in the long term. A strong negative correlation has been observed between a higher 5-HMF content and a lower diastase number, which intensifies with the storage time [3,4,5].

Most enzymes found in honeys are derived from worker bees (from salivary and pharyngeal glands), whereas the others are plant-derived enzymes that occur in nectar and pollen. The honey enzymes include α- and β-glucosidase (invertase), α- and β-amylase, and glucose oxidase. They catalyze reactions leading to the conversion of nectar and honeydew (honey raw materials) into honey (the final product). Their activities lead to the synthesis of multiple biologically active compounds that affect the properties of individual honeys. The enzyme’s inactivation or suppression of activity is largely dependent on its specific nature as well as processing methods and storage conditions for honey. Among the multiple honey enzymes, the European Union legal acts stipulate the minimal threshold of 8 Schade units for the diastase activity [2]. However, for honeys with a low natural enzyme activity, DN should be not less than 3 Schade units [6]. Diastase activity leads to starch and dextrin conversion to oligosaccharides. A significant decrease in its value may indicate the overheating of honey, its excessively long storage in inappropriate conditions, or even its adulteration [5,7,8].

5-Hydroxymethylfurfural present in honeys is formed from monosaccharides (glucose or fructose) in the course of non-enzymatic reactions as a derivative of furfuran or as an intermediate product of Maillard reactions [9,10]. Negligible amounts of this compound are found naturally in fresh honey [5]. During long-term storage of honey, the content of 5-HMF increases. This is the combined result of acidic pH (between 3.2 and 4.5) and water activity, which are beneficial for Maillard reactions (aw of 0.5–0.8) and facilitate 5-HMF formation [11,12]. In addition, 5-HMF concentration is an indicator of not only overly long and inappropriate storage of honeys but also excessive heat treatment or potential adulteration [10,13]. The acceptable 5-HMF content of honey should be less than 40 mg kg⁻¹, except for honeys (or blends) originating from tropical areas, for which the permissible 5-HMF level is equal to or lower than 80 mg kg⁻¹ [6]. For longer storage periods, the initial 5-HMF level is recommended to be low so that its final (probable) level after storage does not exceed the permissible limit [14]. So far, no standard has been developed for 5-HMF level determination within a specified period of honey storage [15].

The sensory properties of honey, particularly color, are important criteria for consumers [16]. Honey has a characteristic color, specific to a particular variety resulting from various contents of carotenoids, anthocyanins, phyto-chlorophylls, and phenolic compounds, and the presence of pollens and minerals [1,17]. However, the color may turn darker (brown) during storage, often at various rates. This browning phenomenon is mainly ascribed to the Maillard reactions (accumulation of brown-colored melanoidin compounds) or the reaction of polyphenols with iron salts. The rate of these reactions is chiefly due to the physicochemical composition, pH, water activity, heat treatment, temperature and duration of honey storage [9,11,18,19]. The process of crystallization as well as the shape and size of formed crystals contribute to changes in color attributes and, consequently, to color perception [17]. The phenomenon of crystallization may be completely inhibited by cooling honey to temperatures below its glass transition temperature (from −40 to −50°C), which however, poses certain technical challenges [20].

As a food product, honey needs to meet high standards of quality and health safety. Changes likely to occur in honeys during storage include, among other things, darkening, 5-HMF accumulation, and suppressed diastase activity [21]. Appropriately selected storage conditions, particularly temperature, are key factors that extend the shelf life of honeys and enable the preservation of all biological properties [9,19]. Honeys available in retail have best-before dates of no longer than 3 years [22]. Chou et al. [10] recommended storing honeys at lower temperatures for no longer than one year. Undoubtedly, choosing the optimal storage temperature of honeys is also essential from an economic standpoint as it may ensure lower energy consumption (at lower storage temperatures) and financial savings. Likewise, heat treatment and cooling are viable methods for food preservation [23]. However, frozen temperatures are hardly ever applied during honey storage both by beekeepers and in trade [19].

We hypothesized that storing honey at low temperatures, including freezing conditions, would limit quality changes and help preserve the characteristics of fresh honey. The scientific literature provides an extensive overview of research on the influence of positive temperatures on the quality of honeys. To the best of the authors’ knowledge, however, there is a paucity of data on the effect of storage at negative temperatures on honey quality. Given the above, the present study aimed to evaluate the effect of a broad range of storage temperatures, i.e., from 20 °C, through 4 °C, −18 °C, and −40 °C, to as low as −80 °C, on 5-HMF content, diastase number, and color of varietal honeys coupled with a long storage period of up to 24 months.

2. Materials and Methods

2.1. Sampling

The experimental materials included 15 honeys that were purchased in 2019 directly from beekeepers operating in southeastern Poland (Lublin region) and were produced in home apiaries. The apiaries were located in 6 districts from different parts of the Lubelskie Voivodeship, namely, Bialski, Zamojski, Biłgorajski, Włodawski, Chełmski, and Lubelski districts. Lublin Voivodeship is the easternmost region of the country. The natural potential for honey production in the region is significantly more favorable compared to the average values in Poland. The quality of the soils, water conditions, agroclimate, and landscape contribute to the fact that Lublin Voivodeship is a leader in many agricultural and horticultural crops. It also leads in terms of the number of bee families. Freshly collected honeys were poured into glass jars. Based on the dominant pollen content [24], 3 honeys of each of the 5 following varieties were distinguished: buckwheat (Fagopyrum esculentum Moench), linden (Tilia spp. L.), rapeseed (Brassica napus L.), honeydew, and multifloral (without the dominant pollen).

2.2. Stage I of Study

The first stage of this study aimed to assess the maturity and quality of fresh honeys based on the following determinations: water and extract contents (%), free acid content (mval kg⁻¹), pH, water activity (a_w), and specific electrical conductivity (mS cm⁻¹). The percentage of water and the total content of substances dissolved in honeys (in % Brix) were determined using a RePo-4 refracto-polarimeter (ATAGO Co. Ltd., Fukaya, Japan) [25]. In turn, a Pioneer 65 Meter (Radiometer Analytical, Villeurbanne, CEDEX-France) was used to measure the specific electrical conductivity (mS cm⁻¹) with a conductometric chamber (CDC 30T), and pH and free acid content were measured with a combined pH electrode (E16M340) [24,25]. The water activity (a_W) of honeys was determined as described earlier [26] using a HygroLab C1 water activity meter (Rotronic, Bassersdorf, Switzerland).

Subsequently, the honeys’ 5-HMF content and diastase number were obtained, as well as the instrumental measurement of color coordinates in the CIE L*a*b* system. The content of 5-HMF (5-(hydroxymethyl-)furan-2-carboxaldehyde) (mg kg⁻¹) [25,27] was determined using a Carry 300 Bio spectrophotometer (Varian Australia PTY, Ltd., Mulgrave, Australia) and computed using the following formula: HMF = (A₂₈₄ − A₃₃₆) × 149.7, assuming the A₂₈₄ and A₃₃₆ values as the readout absorbance values and a constant value of 149.7. Diastase number (DN, in Schade units per 1 g honey) was determined using Phadebas tablets (Honey Diastase Test, Magle AB, Lund, Sweden) containing conjugated starch with a blue dye. The absorbance of the color solutions was measured at a wavelength of 620 nm using a Carry 300 Bio spectrophotometer [24]. The results were computed according to the following formula: DN = 28.2 × ∆A₆₂₀ + 2.64, assuming ∆A₆₂₀ as a difference between the absorbance of the analyzed honey solution and a blind sample, with constant values reaching 28.2 and 2.64. A detailed methodology for the aforementioned chemical analyses was provided by Czernel et al. [28].

Determinations of honey color coordinates in the CIE L*a*b* system (L*, metric lightness; a*, contribution of redness-greenness; b*, contribution of yellowness-blueness, and h°, hue angle) were performed using a CM-600d spectrophotometer (Konica Minolta Sensing, Inc., Osaka, Japan) according to Kędzierska-Matysek et al. [26]. Samples of honey (in 3 replicates) were transferred to CR-A504 Tube Cells (35 mm × 34 mm) and analyzed 3 times to compute the mean value.

The results of analyses conducted for fresh honeys, including the 5-HMF content, diastase number (DN), and values of color components on the CIE L*a*b scale, were assumed as control values.

2.3. Stage II of Study

To determine the effect of various storage temperatures on selected quality attributes, the fresh varietal honeys were divided into 5 batches (ca. 250 g each). Analyses were conducted twice, i.e., immediately before and after storage for 24 months. The first batch of honeys (denoted as RT) was stored under typical consumer household conditions, i.e., at room temperature (20 °C ± 2 °C) and in a dark place. The second batch of honeys (denoted as FR) was stored at refrigerated temperature conditions (ca. 4 °C), whereas the last three batches were frozen and stored at three temperatures (−18 °C, −40 °C, and −80 °C). The honey samples were stored in 240 mL PFA (perfluoroalkoxy alkane) containers approved for contact with food and constant freezing (Savillex, Eden Prairie, MN, USA) for 24 months at the following temperatures and in the following appliances: at 4 °C in a Whirlpool WBC3525 A + NFX refrigerator, at −18 °C in a LIEBHERR GG 5210 ProfiLine upright freezer (Liebherr-Hausgeräte GmbH, Ochsenhausen, Germany), at −40 °C in an LT U250 upright freezer (Nordic Lab, Vaerloese, Denmark), and at −80 °C in a Telstar Boreas U445 upright ultra-low-temperature freezer (AZBIL TELSTAR Technologies S.L.U, Terrassa, Spain).

After 24 months of storage, the honey samples were analyzed again to determine the content of 5-hydroxymethylfurfural (5-HMF) and diastase number (DN), and they were subjected to measurements of color coordinates in the CIE L*a*b* space accordingly to the aforementioned methodology. Furthermore, the total color difference (ΔE) was calculated using the CIE L*, a*, and b* values from Equation (1):

$$\mathit{\Delta }\mathit{E}=\sqrt{\left[\right({{\mathit{L}}^{\mathit{*}}}_{\mathit{C}}-{{\mathit{L}}^{\mathit{*}}}_{\mathit{S}}{)}^2+({{\mathit{a}}^{\mathit{*}}}_{\mathit{C}}-{{\mathit{a}}^{\mathit{*}}}_{\mathit{S}}{)}^2+({{\mathit{b}}^{\mathit{*}}}_{\mathit{C}}-{{\mathit{b}}^{\mathit{*}}}_{\mathit{S}}{)}^2]}$$

(1)

where L*_C, a*_C, and b*_C are values of control honeys and L*_S, a*_S, and b*_S are values of stored samples.

The above analyses were conducted on honeys brought to room temperature.

2.4. Statistical Analysis

The results achieved from the determination of both fresh honeys and those stored at various temperatures were subjected to statistical analysis using Statistica ver. 13 software (TIBCO Software Inc., Palo Alto, CA, USA). The initial characterization of fresh honeys involved computations of descriptive statistics (mean value and standard deviation) for individual honey varieties. The one-way analysis of variance (ANOVA) was applied to determine the effect of variety, and differences between mean values in groups were verified with the Tukey HSD test. The results of these calculations are presented in

Table 1. Physicochemical characteristics of fresh varietal honeys ($\overline{\mathrm{x}}$ ± SD).

Properties	Buckwheat	Linden	Rapeseed	Honeydew	Multifloral
Water, %	18.5 ± 1.0	17.4 ± 1.4	17.6 ± 0.8	17.3 ± 1.2	16.9 ± 1.7
Brix, %	79.8 ± 1.0	80.8 ± 1.5	80.7 ± 0.8	81.0 ± 1.2	81.3 ± 1.6
Free acids, mval kg−1	45.5 ± 2.8 b	42.7 ± 3.5 b	21.8 ± 4.5 a	32.7 ± 14.2 ab	30.0 ± 7.7 ab
Electrical conductivity, mS cm−1	0.51 ± 0.06 A	0.51 ± 0.23 A	0.19 ± 0.02 A	1.10 ± 0.18 B	0.40 ± 0.11 A
pH	3.65 ± 0.19	3.65 ± 0.21	3.75 ± 0.06	4.06 ± 0.39	3.87 ± 0.12
Water activity	0.581 ± 0.010	0.538 ± 0.036	0.564 ± 0.009	0.571 ± 0.020	0.532 ± 0.015

Means with different letters in the same column are significantly different: ^a,^b—p ˂ 0.05; ^A,^B—p ˂ 0.01.

. The preliminary statistical analysis of the results achieved for the honey sample stored at refrigeration and freezing temperatures in a two-way ANOVA with interactions (honey variety × storage temperature) demonstrated a significant effect of both factors (except on ΔE), and their interaction was significant for all assessed variables. Due to a high number of sub-groups (5 honey varieties × 6 storage temperatures) and obvious substantial differences in the chemical composition of honeys, a one-way analysis was conducted to verify only the main effects in various systems. The results of 5-HMF, DN, and color (CIE L*a*b*) measurements obtained for the stored honeys were tentatively analyzed for all samples together (regardless of variety) to verify the effect of storage temperature (main effect) using a one-way analysis of variance and Tukey HSD test. The results of this analysis are presented in figures. The subsequent stage of the statistical analysis aimed to determine the effect of honey variety (as the main effect) on the honey samples stored at the analyzed temperatures. The results of these computations are presented in figures, which additionally demonstrate the effect of storage temperature separately on individual honey varieties.

Correlations between physicochemical properties and variability indicators were evaluated using multi-dimensional principal component analysis (PCA) with the covariance matrix. The percentage of the explained variance exceeding 80% was adopted as the criterion for the number of principal components.

3. Results

3.1. Fresh Honeys

The results of the selected physicochemical properties of fresh honeys are presented in Table 1. They demonstrate no effect of the variety on the water and extract contents of the honeys. Fresh honeys were characterized by collective maturity as the water content between 16.9% (multifloral honeys) and 18.5% (buckwheat honeys) was lower than the 20% threshold stipulated by the European Union [2]. The average soluble solid content in honeys ranged from 79.8% to 81.3%. In turn, the contents of free acids differed significantly (p ˂ 0.01) between the honey varieties, with the lowest acidity determined in the rapeseed honeys (21.8 mval kg⁻¹) and the highest one in the linden and buckwheat honeys (42.7 and 45.5 mval kg⁻¹, respectively). The permissible acidity level of 50 mval kg⁻¹ or less [2] was not exceeded in any of the analyzed samples. The variety of honeys had significantly different pH values, which ranged from 3.65 (buckwheat honeys) to 4.06 (honeydew honeys). The measurements of the specific electrical conductivity performed in the present study enabled confirming the origin of nectar honeys (from 0.19 mS cm⁻¹ in rapeseed honeys to 0.51 mS cm⁻¹ in buckwheat and linden honeys) and honeydew honeys (1.1 mS cm⁻¹). The recommended values of the specific electrical conductivity of the nectar honeys are lower than 0.8 mS cm⁻¹ [2], whereas those of the honeydew honeys are higher. The water activity in the analyzed honeys did not exceed the critical level of 0.6 and ranged, on average, from 0.532 (multifloral honeys) to 0.581 (buckwheat honeys). The physicochemical properties of the analyzed fresh honeys (Table 1) confirmed their high quality and usability for the long-term 24-month storage period assumed in this study.

3.2. Stored Honeys

3.2.1. Diastase Number

Figure 1 presents mean DN values determined for the analyzed honeys (regardless of variety) before (control group) and after 24-month storage at various temperatures (20 °C, 4 °C, −18 °C, −40 °C, and −80 °C). The mean initial DN of all honeys was high and reached 33.16. After 24 months, it decreased significantly (p ˂ 0.01) in all honey samples, regardless of the storage temperature, with the greatest drop (to 11.05 DN) accounting for 66.7% of the initial DN value determined in the honeys stored at room temperature (RT). At the low storage temperatures (refrigeration and freezing), diastase activity was also observed to decrease in the honeys (from 53.1% to 58.3%); however, the determined DN values were similar (p > 0.05) and ranged from DN = 15.54 at 4 °C (FR) to DN = 13.84 at −18 °C. It is noteworthy that regardless of the storage temperature, the mean DN value did not exceed the threshold adopted for this honey quality criterion (DN = 8).

Figure 2 depicts changes in the diastase number of honeys of particular varieties depending on the storage temperature. All varietal honeys exhibited high initial enzymatic activity, which was many times greater than the value (DN = 8) recommended by the European Union [2]. Based on the mean initial DN value, they were ranked according to the enzymatic activity in descending order, as follows: buckwheat (62.8 ± 8.9, p ˂ 0.01) > linden (35.1 ± 7.4) > honeydew (29.1 ± 9.3) > multifloral (21.0 ± 2.2) > rapeseed (16.8 ± 4.7).

Generally, this order of the DN in the evaluated honey varieties was observed at all storage temperatures tested (Figure 2). Samples of all honey varieties stored at RT showed lower DN values. Diastase activity decreased significantly compared to the control samples, i.e., by 75.7% in buckwheat honeys, 66.7% in linden honeys, 61.3% in honeydew honeys, 56.0% in multiflora honeys, and 55.6% in rapeseed honeys. In the case of the latter honey variety, the mean DN value was lower (by 0.6 units) than the recommended level. Compared to RT, the refrigerated and frozen storage had a positive effect on the preservation of higher enzymatic activity in the honeys of all tested varieties. Nonetheless, better outcomes were observed in the case of buckwheat honeys, followed by linden and honeydew honeys. In addition, the low storage temperatures allowed for maintaining diastase activity in rapeseed honeys above the critical value (DN = 8). Summing up, among the evaluated honeys, the buckwheat honey samples had the highest DN (p ˂ 0.01), and the rapeseed ones had the lowest DN at all storage temperatures.

3.2.2. 5-HMF

The initial mean 5-HMF content reached 5.91 mg kg⁻¹ in the honeys of all analyzed varieties (Figure 3). After 24 months, a significant (p ˂ 0.01) increase of 79.3% (28.61 mg kg⁻¹) was noted only in the honeys stored at RT. In the cold and frozen stored honey samples, the 5-HMF content increased from 25.1% (7.89 mg kg⁻¹ at −18 °C) to 33.2% (8.85 mg kg⁻¹ at FT); however, the observed differences were statistically non-significant compared to the control sample.

The effect of storage temperature on the 5-HMF content in varietal honeys is presented in Figure 4. Assuming 5-HMF < 40 mg kg⁻¹ as the criterion for high-quality honeys [2], it may be concluded that its concentration in fresh honeys of all examined varieties was low (buckwheat honeys) or very low (other varieties). Initially, the highest content of 5-HMF was determined in buckwheat honey (13.97 mg kg⁻¹), whereas significantly lower contents (p ˂ 0.01) were determined in the other varieties (6.02 mg kg⁻¹ in multifloral honeys, 4.01 mg kg⁻¹ in linden honeys, 3.43 mg kg⁻¹ in honeydew honeys, and 2.1 mg kg⁻¹ in rapeseed honeys). As expected, room temperature turned out to be the least beneficial for the storage of honeys, compared to the low temperatures tested. This finding was confirmed by the highest 5-HMF content in the honeys of all varieties stored at RT. It is noteworthy that the mean 5-HMF content of buckwheat honeys stored at RT reached 50.3 mg kg⁻¹, i.e., exceeding the permissible value. In the honeys of the other varieties, the 5-HMF content increased due to the influence of room temperature, e.g., to 32.3 mg kg⁻¹ (more than a 5-fold increase in multifloral honeys) or 16.26 mg kg⁻¹ (more than an 8-fold increase), but still did not exceed the permissible value. The results presented in Figure 4 indicate a positive and similar effect of low storage temperatures (both refrigeration and freezing temperatures) on the inhibition of 5-HMF formation in varietal honeys. This positive effect was especially evident in the case of buckwheat honey, where low storage temperatures allowed for maintaining the 5-HMF content at the initial level. A similar observation was made for linden and honeydew honeys. Due to the varying 5-HMF concentrations in individual samples of varietal honeys, statistical analysis did not confirm any significant differences between the mean values in the analyzed range of low temperatures of storage (4 °C, −18 °C, −40 °C, and −80 °C).

3.2.3. Color

Figure 5, Figure 6, Figure 7 and Figure 8 show changes in the color coordinates within the CIE color space (L*, a*, b*, and h°) for all analyzed honey varieties depending on the storage temperature. The lightness (achromatic component) of honeys stored at RT (L* = 31.98) was significantly lower (p ˂ 0.01) compared to the control sample (L* = 38.91) (Figure 5). A similar, although non-significant trend, was observed after storage at FT (L* = 36.73). The lightness value of honey stored at −18 °C (L* = 38.36) was numerically comparable to that of the control samples. As the frozen storage temperature progressively decreased (below −18 °C), the honeys became brighter, exhibiting higher L* values (p > 0.05), as evidenced by L* = 39.74 for honeys stored at −40 °C and L* = 41.61 for those stored at −80 °C.

Regardless of storage conditions, the a* coordinate values were generally higher, reflecting a greater contribution of redness than in the control samples (a* = 1.96 in fresh honeys). However, significant differences were found for honeys stored at −18 °C (a* = 3.02) and −40 °C (a* = 2.97) (Figure 6). Storage at room temperature (RT) and refrigeration temperature (4 °C) resulted in comparable a* values (2.83 < a* < 2.87). In turn, in the honey samples stored at −80 °C, redness (a* = 2.41) appeared to be more stable compared to the fresh honeys.

The analysis of the results for yellowness (b* coordinate) demonstrated a lower value (p > 0.05) in honeys stored at RT (b* = 7.57) compared to the control samples (b* = 10.60), and its initial value was maintained in the honey samples stored at FT (b* = 10.69) (Figure 7). As the frozen storage temperature decreased, the yellowness values increased (p > 0.05) to b* = 11.31 (−18 °C) and b* = 11.90 (−40 °C and −80 °C).

In the case of the hue angle (h°), a significantly lower value compared to the control group (h° = 77.5) was determined only for honey samples stored at RT (h° = 67.5, p ˂ 0.01) (Figure 8). A similar but statistically non-significant trend was observed at the low storage temperatures (refrigeration and freezing). However, lower temperatures elicited a stabilizing effect on the values of this color attribute (from h° = 70.5 at 4 °C to h° = 76.0 at −80 °C).

To sum up, the greatest changes, including significant darkening (lower L* value) and simultaneous color shift toward red (significantly lower h°, a lower contribution of the yellowness, and a greater contribution of the redness), were determined in the honeys stored at RT. The smaller extent of changes mentioned above was noted in the honeys stored at low temperatures.

According to Equation (1), the ΔE value was influenced by the changes in the three CIE coordinates (L*, a*, and b*) in fresh honeys (control group) and the honey samples stored at different temperatures (Figure 9). The ΔE value differed significantly (9.53, p ˂ 0.01) between honey samples stored at RT and those stored at the other temperatures, showing the highest value in the first ones. It is noteworthy that the decrease in L* value during storage at RT had the greatest contribution to the color shift compared to the fresh honeys. The total color difference determined for the honeys stored at low temperatures ranged from 3.71 (−18 °C) to 5.58 (−80 °C).

The instrumental color analysis showed that variety caused significant (p ˂ 0.01) differences in the values of the L* component for both the control honeys and the ones stored at various temperatures (Figure 10). Among the fresh honeys, the highest lightness (p ˂ 0.01) was noted for the rapeseed honey (L* = 53.61), whereas the multifloral honey was darker (L* = 44.12), and the honeydew, linden, and buckwheat honeys were the darkest (L* = 33.48, 33.26, and 30.06, respectively). Generally, this order of honey varieties, from the lightest rapeseed honey to the darkest buckwheat honey, was observed at all storage temperatures. The same order of significant differences (as in the control samples) was repeated for the samples stored at −18 °C and −40 °C and those stored at RT (except for the linden honey). In the case of the samples stored at FT, there were no significant differences between the multifloral and linden honeys, whereas, at −80 °C, there was not a significant difference between linden and honeydew honeys. An additional analysis of variance also confirmed the significant effect of storage temperature on changes in the lightness of individual honey varieties, indicating their darkening after storage at RT.

The effect of variety on changes in the redness contribution to the CIE color profile of honeys stored at the analyzed temperatures is presented in Figure 11. In fresh honey samples, the highest a* value was determined in the buckwheat honey (3.50), followed by multifloral and honeydew honeys (2.27 and 2.20, respectively) and linden honey (1.11), while the lowest one was observed in rapeseed honey (0.70). A similar order of varieties was noted in the honeys stored at all three sub-zero temperatures. However, a different arrangement of significant differences was found. The greatest changes in the contribution of redness occurred in the honeys stored at RT. They resulted in a reversed order of varieties, i.e., the highest a* value was determined in rapeseed honey (4.43), and the lowest one was found in buckwheat honey (0.39). In turn, storage at FT resulted in a similar contribution of the redness in four honey varieties (i.e., from 2.79 to 3.31 in honeydew, linden, buckwheat, and multifloral honeys), and the lowest a* value was noted in rapeseed honey (a* = 1.87). Temperature (considered as the main effect) caused significant differences in the redness of honeys of individual varieties (except for honeydew honey, Figure 11).

Variety had a significant influence on yellowness contribution in the color profile of honeys stored for 24 months at all temperatures (Figure 12). Generally, the highest b* coordinate values were determined in multifloral and rapeseed honeys. An exceptional situation was observed during the storage of honeys at a temperature of −80 °C, where the yellowness value was similar (12.2 < b* < 14.8) in the honeys of four varieties (multifloral, linden, rapeseed, and honeydew). In contrast, the lowest value of b* was typical of the buckwheat honey (except for the sample stored at −18 °C). Storage temperature also appeared to be a significant factor in differentiating the yellowness between individual varieties of honey.

The results presented in Figure 13 demonstrate a significant effect of honey variety on the hue angle (h°) at all storage temperatures. The lowest value of h° was determined for buckwheat honey (46.7 < h° < 58.8) in fresh, cold, and frozen stored honeys, which differed significantly from the values determined for all other honey varieties. In turn, at the mentioned temperatures, rapeseed honey showed the highest h° values (74.5 < h°< 86.9), i.e., hue angle values shifted towards the range typical of the yellow hue (90°). The storage of honeys at RT resulted in a significant increase in the h° value compared to the control samples only in the buckwheat honey (from 58.8 to 73.3). This change followed a different trend than that observed for the remaining varieties, which showed a significant decrease in the values of this parameter, which may be interpreted as a shift in vividness, with a greater contribution of redness (Figure 11). Statistical analysis confirmed a significant (p ˂ 0.01) effect of storage temperature, which also had the main effect on the honey samples of individual varieties.

Variety significantly affected only the total color difference (ΔE) of honeys stored at RT (Figure 14). The smallest color change, compared to the fresh honeys, was noted for linden honeys (ΔE = 4.9), whereas the greatest one was observed for rapeseed honeys (ΔE = 15.5), with the difference found to be significant (p < 0.01). In the case of the refrigeration and freezing storage temperatures, statistical analysis showed no effect of honey variety on the total color difference, which was ascribed to the high variability of this parameter. The ΔE values of the honeys stored at FT ranged from 3.1 to 6.7, with those stored at −18 °C ranging from 1.5 to 4.9, those stored at −40 °C ranging from 1.4 to 6.5, and those stored at −80 °C ranging from 1.9 to 10.3. The influence of storage temperature (as the main effect) on ΔE turned out to be significant only in the case of rapeseed honey. In addition, the tendencies of changes in ΔE were noted for linden (p ˃ 0.05) and multifloral (p ˃ 0.05) honeys.

3.2.4. PCA

PCA allowed for the reduction of six active variables into three principal components that helped explain 89.90% of the total variance, with PC1 accounting for 62.73%, PC2 for 15.33, and PC3 for 11.84%. The first principal component was strongly negatively correlated with the h° value (−0.955), L* value (−0.907), and b* value (−0.783). The second PC was negatively correlated with DN (−0.787) but positively correlated with 5-HMF (0.650). The third PC showed a positive correlation with 5-HMF (0.692) and DN (0.514).

Three groups of variables could be distinguished according to the data shown in Figure 15a. The 5-HMF was located in the positive areas of both components, the DN was distributed in the positive area of PC1 and negative area of PC2, while hue°, L*, and b* coordinates were in the negative area of PC1 and PC2. The positioning of the chemical criteria of honey quality and color coordinates according to the PC1 x PC2 design represented their importance in relation to storage temperatures (Figure 15b). Based on the visualization of the cases, three groups can be distinguished. The control honey samples are clearly separated at the bottom of the plot and, thus, are negatively correlated with PC1 (linked to the h°, L*, and b* values) and PC2 (and thus with DN). In turn, the RT samples are placed mostly in the upper right area of the plot; hence, they are positively correlated with PC2 and HMF content. The cases represented by the samples stored both at chilled temperatures and at all temperatures below 0 °C are distributed centrally along the PC2 axis and on both (positive and negative) sides of the PC1 axis. Correlation analysis further showed that variety (an additional variable included in the PCA) was most strongly and negatively associated with PC1 (−0.809), while it correlated positively with L* (0.815), h° (0.713), and b* (0.697).

4. Discussion

4.1. DN and 5-HMF

Comparing the results achieved in the present study to findings reported by other authors may be challenging due to variety-specific differences between honeys, their geographical origin, ranges of storage periods and temperatures, and analytical methods and equipment applied. Most often, honeys have been investigated after a defined period of storage or after high-temperature treatment [29,30,31]. Generally, prolonged storage time and/or heat treatment leads to a decrease in diastase activity and an increase in 5-HMF formation in honeys. These undesirable changes may be mitigated by storing honeys at low temperatures or by applying heat treatment at lower temperatures. In the present study, particularly unbeneficial changes in the quality criteria were noted in the honeys stored at RT for 2 years, where a 3-fold decrease in DN (Figure 1) and a 5-fold increase in 5-HMF concentration were noted (Figure 3) compared to the fresh samples. Similar observations were stated in our previous study [3] conducted with rapeseed honey stored for 18 months at RT (20–26 °C) and at a temperature of −20 °C. In the RT group, the 5-HMF content increased more than sixfold, whereas DN decreased by 25%. In contrast, in the second group (stored in a freezer), a downward trend (p > 0.05) was noted in both parameters, i.e., by −50.5% and −7.3%, respectively, compared to the fresh samples (5-HMF = 3.07 mg kg⁻¹ and DN = 28.37). In the case of the honey of this variety analyzed in the present study after 2 years of storage at RT, the 5-HMF level increased by eight times, whereas the diastase number accounted for only 44% of its initial value determined in fresh honeys (Figure 4). Czipa et al. [32], who stored acacia honeys for 24 months at a temperature of 20 °C, observed a 7-fold increase in HMF concentration (from 1.9 to 13.6 mg kg⁻¹) and a 15% decrease in DN (from 17.8 to 15.1). In turn, Wesołowska and Dżugan [33], who stored domestic honeys (Subcarpathian region, Poland) at RT (20 ± 2 °C) for 24 months, demonstrated the effect of variety on the decrease in diastase number, which fluctuated between 17% (rapeseed) and 42% (linden), with the final DN values ranging from 12.75 (rapeseed) to 20.92 (goldenrod). In turn, Panseri et al. [8] showed the optimal temperature of acacia honey storage for 550 days (18 months) to be 15 °C (compared to temperatures of 25 °C and 35 °C), which ensured that the HMF concentration and diastase activity were maintained at acceptable levels. Honey storage at temperatures higher than RT (25–27 °C) for 24 months caused a significant increase (10-fold) in HMF concentration (3.5 vs. 34.8 mg kg⁻¹) and a concomitant decrease of 33% in diastase activity (DN ranging from 19.13 to 12.85) [34]. Similar findings were reported by Seraglio et al. [35] during the storage of ‘Bracatinga’ honeydew at a temperature of 23.0 °C (±2.3 °C); after 20 months, the HMF concentration increased from the initial ‘not detected’ level (<LOD) to the value exceeding the permissible level for this compound (<40 mg kg⁻¹) [35]. In turn, Bhalchandra et al. [36] observed a decrease in DN from 22.41 to 3.88, i.e., below the recommended level (DN = 8.0) in honeys stored at RT for 24 months. In the present study, after the two-year storage of honeys at RT, the DN decreased below the recommended value but only by 0.6 units in the rapeseed honey (Figure 2). In contrast, the 5-HMF content exceeded the permissible level by 10 mg kg⁻¹ only in the buckwheat honey (Figure 4). The higher HMF concentration in honeys is due to a decreased content of fructose as a result of its intensive breakdown observed during storage [29]. In the present study, the highest 5-HMF concentration was demonstrated for buckwheat honey, the major saccharide of which is fructose [37].

The application of low temperatures offers a viable means for food stability preservation. Chou et al. [10] demonstrated that storing honey (regardless of access to light) for up to two years at refrigeration temperature (4 °C), compared to RT (25 °C and 35 °C), reduced diastase activity suppression. In contrast, the final 5-HMF concentration determined in the honeys stored at 4 °C and 25 °C remained at a low level (<10 mg kg⁻¹). Storage at low temperatures (−20 °C and +4 °C) was reported to result in significant changes in HMF concentration due to the low effectiveness of its formation [14]. Apart from the present study, such observations were confirmed earlier by Ghoniemy et al. for HMF [38] after 12-month storage of honeys in a freezer (−20 °C) and a refrigerator (+4 °C) compared to the samples stored at RT (+25 °C). Regardless of the package type (black glass vs. white glass), the HMF concentration increased from 3.72 to 3.77 mg kg⁻¹ at −20°C, from 4.30 to 4.43 mg kg⁻¹ at +4 °C, and from 8.78 to 9.13 mg kg⁻¹ at RT. In contrast, Raweh et al. [39] demonstrated no significant differences in HMF level and diastase number between Talh honey samples stored for 8 months at a temperature of 0 °C and fresh honey samples, i.e., HMF = 2 ± 0.4 mg kg⁻¹ and DN = 15 ± 3.0 vs. HMF = 1 ± 0.2 mg kg⁻¹ and DN = 16 ± 0.3, respectively. On the other hand, storage at a temperature of 25 °C caused a significant increase only in HMF concentration (15 ± 4.9 mg kg⁻¹).

It is noteworthy, however, that HMF content is affected by multiple factors, and the cause of its changes, as reported in the literature, seems highly complex [15]. In the case of diastase activity, Sancho et al. [40] demonstrated three types of temporal behavior in 115 varietal honeys (Basque Country, Spain). After their storage at RT (mean: 20 °C, range: 15–25 °C) for 28 months, its decrease followed a logarithmic trend (68.7% of the samples), a linear trend (21.7% of the samples), or an asymptomatic trend (9.6% of the samples).

4.2. CIE Color

The variability of color coordinates (CIE L*, a*, and b*) determined in the present study for fresh Polish varietal honeys was within the ranges reported by others. Halagarda et al. [41], using a CM-5 bench-top spectrophotometer (Konica Minolta), revealed a lightness range of 23.18 (for honeydew honey) to 59.91 (for multifloral honey), redness of −1.61 (for multifloral honey) to 11.6 (for multifloral honey), and yellowness of 7.59 (for rapeseed honey) to 31.33 (for multifloral honey). Kuś et al. [42] reported the following values for six native floral honeys: L* value ranging from 41.5 (buckwheat honey) to 86.2 (black locus honey), a* value of −1.6 (for black locust honey) to 31.9 (for buckwheat honey), and b* value of 19.6 (for black locust honey) to 69.2 (for buckwheat honey). In turn, Piotraszewska-Pająk and Gliszczyńska-Świgło [17] achieved the following mean values of the color coordinates for six fresh nectar honeys using a CR-310 Chroma Meter (Konica Minolta): L* of 49.62 (heather honey) to 70.56 (rapeseed honey), a* of −3.56 (linden honey) to 6.32 (buckwheat honey), b* of 9.32 (acacia honey) to 26.62 (buckwheat honey), and h° of 76.24 (buckwheat honey) to 105.13 (acacia honey). Although the statistical analysis did not confirm any significant differences between the honey varieties in their L* and b* values, the light-colored varieties (rapeseed, acacia, linden, and multifloral honeys) differed significantly from the dark-colored honeys (heather and buckwheat honeys) in the redness value (−a* for light honeys vs. +a* for dark honeys) and hue angle (h° >90 for light honeys vs. h° < 85 for dark honeys) [17]. Similar observations were determined for these color coordinates in the present study between the dark buckwheat honey and other light-colored honey varieties (Figure 11 and Figure 13). It needs to be emphasized that the values of color coordinates in the CIE L*a*b* space for varietal honeys in the present study differ from those reported by the abovementioned authors, particularly with respect to the negative a* values, which were not confirmed in our study. The positive a* values are indicative of the prevailing contribution of redness, typical of flavonoids (anthocyanins in particular) [43] found in plant nectar. In contrast, the negative a* values are ascribed to the presence of green pigments (e.g., chlorophylls) [44]. The values of the b* coordinate were positive in all honey varieties (both in those analyzed in the present study and by other authors), which points to the prevalence of yellow pigments, including mainly carotenoids [45]. Although carotenoids are generally claimed to occur in high quantities in dark honeys [1], in the present study, significantly higher values of the b* coordinate were determined in the light honeys (multifloral and rapeseed) compared to the dark buckwheat honey (Figure 12). The differences observed in the L*, a*, b*, and h° values could result, on the one hand, from the obvious differences in the chemical composition of the evaluated honeys and, on the other hand, from differences in the methodology of sample preparation, sample form, and measuring instruments used (colorimeters vs. spectrophotometers).

After collection, honey that is properly stored in sealed containers is stable and safe for consumption for weeks, months, and years or may even ‘have eternal shelf life’ [15]. Despite this, physical changes and chemical processes take place (at varying rates) in honey, resulting in a darker color and loss of flavor. Since this is a temperature-dependent phenomenon, it is not easy to define the correct period for which honey can be stored. In practice, a two-year shelf life is most often assumed [46]. After processing, honey should be kept at room temperature (18–24 °C), while unprocessed honey should be stored at less than 10 °C. In contrast, the recommended storage temperature for both types of honey mentioned earlier is 0 °C (or lower), which best preserves the color and flavor [47]. It should be noted that water does not freeze in honey but is only strongly bound [48]. This phenomenon is referred to as glass transition, which occurs below 233 K (−40.15 °C) [49].

Typically, honey is stored at ambient temperature, which is unlikely to be controlled [19]. In Europe, (increasingly) during the summer, ambient temperatures exceed 35 °C during the day. Potentially, honey can be exposed to such conditions for shorter or longer periods. Sancho et al. [40] stated that honey should be consumed within the year following its harvesting. For instance, the maximum durability of honey in the Hungarian market must not exceed 2 years. However, honey stored over this period became darker in color [32]. Jones et al. [19] reported that the color of honey after 5 years of storage at RT (ambient temperature, ±22 °C) was significantly darker compared to the fresh samples (average of 1064 vs. 475 mAU, respectively). In an earlier study, Kędzierska-Matysek et al. [50] evaluated changes in color indices of multifloral honeys stored at RT, 2 years after their minimal stability date. They noted significant decreases in the values of all color coordinates, except for a*; lightness (L*) decreased by 31%, yellowness (b*) by 52%, and hue angle (h°) by 28%, whereas redness (a*) increased by 43%.

The values of L*, a*, and b* obtained in the present study for multifloral honey, as well as the trends in their changes, are similar to those obtained by Amarei et al. [51], who assessed the color of multifloral honey (using the Konica Minolta Chroma Meter CR-410, illuminant D65) stored in the dark for 12 months at a temperature of 14–16 °C. In turn, Piotraszewska-Pająk and Gliszczyńska-Świgło [17], who stored domestic varietal honeys for 9 months in refrigeration conditions (4–6 °C) and at RT (22–24 °C, in the dark), obtained higher L* values compared to the fresh samples. The only exception was buckwheat honey, which turned darker after storage at RT, and the same observation was made in the present study (Figure 10); however, the numerical values for lightness were significantly lower. After storage at RT, buckwheat honey had a lower b* value and rapeseed honey had a higher b* value than the control samples [17], which is in line with the results of the present study (Figure 12). It has been shown that storage time, regardless of temperature, resulted in a decrease in the hue angle, i.e., the value in the color wheel shifted from the range for yellow (90°) to orange-red (45°) and then finally to a red hue (0 °C) [44,52]. The same relationships (except for buckwheat honey stored at RT) were also demonstrated in the present study (Figure 13).

One of the physical factors influencing the color of honeys is the process of natural crystallization, which alters the structure of the product. Moreover, the consistency of honey (liquid vs. solid), as well as the size and shape of crystals formed, affect the behavior of light in the object. Piotraszewska-Pająk and Gliszczyńska-Świgło [17] reported that liquid honeys exhibited lower L* values (from 1.5 to 10 units) and b* values (from 2 to 6 units) compared to solid honeys. Rapeseed honey with a high glucose-to-fructose ratio (G/F) tends to crystallize quickly [53], within 1–2 months [51]. Glucose, being less soluble than fructose, separates faster in honey by interacting with water molecules. The precipitated white crystals determine the light color of rapeseed honey due to their greater ability to reflect light [54] compared to the multifloral honey containing large and uneven crystals [51]. On the contrary, because fructose is the major sugar in buckwheat honey, the rate of crystallization in this honey variety is much slower [37], and crystallization may not occur at all, as in acacia honey [51]. In addition to modifying the color of honey, crystallization also changes its appearance and consistency, making it waxy and opaque [55].

Practically, if the samples are not next to each other, a total color difference of less than 1 is imperceptible. However, Frances and Clydesdale [56] pointed out that the value ΔE = 2 can already be a perceptible sensation in the visual assessment of many products. From a technical point of view, the color differences (in CIE L*a*b* units) between 1.1 and 2.8 align with rigorous tolerances and those between 2.8 and 5.6 align with normal tolerances, while values over 5.6 ought to be easily distinguished by the human eye [57]. Piotraszewska-Pająk and Gliszczyńska-Świgło [17] reported that during 9-month storage at refrigeration temperatures (4–6 °C) and at RT (22–24 °C), rapeseed and buckwheat honeys exhibited greater stability of color components (ΔE from 8 to 10) than the other analyzed honey varieties (acacia, linden, multifloral, and heather honeys). In contrast, our results indicated the most severe color loss for rapeseed honey (ΔE = 15.5) at RT compared to buckwheat honey (ΔE = 7.2) (Figure 14). The degradation rate of color quality in terms of total color difference at RT was significantly (p < 0.01) faster in all honey varieties (ΔE between 4.9 and 15.5) than those after cold storage (ΔE ranged from 3.1 to 6.7) and frozen storage at −18 °C, −40 °C, and −80 °C (ΔE from 1.5 to 4.9, from 1.4 to 6.5, and from 1.9 to 10.3, respectively) (Figure 14). Kędzierska-Matysek et al. [3] reported that the color of rapeseed honey changed from 417 to 476 mAU after 18 months of storage at −20 °C, whereas at RT, it increased to 662 mAU. Using the color space in the CIE L*a*b* system, these authors demonstrated that the change in the color of buckwheat honey was significantly (p < 0.01) more strongly affected by storage at RT (ΔE = 20.37) than by freezing storage (ΔE = 8.07). Jones et al. [19] did not report any significant color change (mAU) during the 16-week storage of honeys at three temperatures (4 °C, −20 °C, and −80 °C), i.e., at temperatures lower than RT/ambient temperature. These results indicate that chemical reactions still take place in honey at low (and very low) temperatures, only much slower. Despite the proven benefits of low temperatures, their use may be limited from a technical point of view due to the undesirable consistency changes associated with the crystallization of honey under household conditions [58].

In the present study, principal component analysis showed that the variables that best identify fresh honey of different varieties were the diastase number and color components, such as lightness and hue. In contrast, the best determinant of honey stored at RT was the 5-HMF content.

5. Conclusions

The results obtained in the present study clearly indicate a negative effect of the 2-year (long-term) storage at RT on the quality of honeys and, in the case of buckwheat honeys, even on their health safety when the permissible 5-HMF level is exceeded. On the other hand, a beneficial and comparable effect of low temperatures on maintaining the satisfactory quality of honeys of the assessed varieties was confirmed. It should be emphasized that such a beneficial effect can be obtained at a temperature of +4 °C without the need to store honeys at freezing temperatures, regardless of their adopted level (−18 °C, −40 °C, or −80 °C). This finding has a measurable value because it can significantly contribute to lowering the costs of storing this food product. However, due to the scope of the analyses conducted, which were limited to determining several important physicochemical parameters of a few Polish honey varieties from one region, it is necessary to continue making observations over a broader spectrum of research. In the authors’ opinion, further studies should focus beyond the nutritional aspect and include the bioactive properties (antioxidant, anti-inflammatory, immunomodulatory, and antibacterial activities) of honey as a raw material used in the food industry as well as the pharmaceutical and cosmetic sectors.