A wide range of Croatian monovarietal red wines was analyzed for OTA and BA contamination. The sample represented geographically diverse sources, as it included both wines from continental (Eastern Continental Croatia) and maritime regions of Croatia (Istria and Dalmatia), characterized by distinguished climatic conditions. Eastern Continental Croatia has a continental climate with hot summers and cold winters and enough rainfall to make it a major agricultural area. The coastal line has Mediterranean climate, however, Istria is influenced also by the Alpes and has hot and humid summers and mild winters, while Dalmatia has hot and dry summers and mild winters. The investigated wines can further be traced by their protected designation of origin (PDO) to the wine-making sub-regions: Slavonia and Croatian Podunavlje within Eastern Continental Croatia; Croatian Istria; Northern Dalmatia, Dalmatian Inland, Central and Southern Dalmatia and Dingač within Dalmatia. Due to greater prevalence of red varieties in maritime regions [
33] more wines in this research were sourced from Istria and, especially, Dalmatia.
3.1. Ochratoxin A
OTA concentrations in analyzed wines ranged from below LOD to 0.163 µg/L, while the average and median values were 0.040 and 0.026 µg/L, respectively, with only 9% of wines having OTA concentration higher than 0.1 µg/L (
Table 3 and
Table 4). Since this is the first comprehensive report of OTA in Croatian wines, the results are extensively compared with those from different European and Mediterranean countries (
Table 5). However, comparisons with results of other authors have to be done with caution, since methodologies of analytical procedures as well as of reporting of results differ notably. In an attempt to present more uniform results from previous studies,
Table 5 contains the results obtained by methods using mostly IAC clean-up of samples followed by HPLC with fluorescence detection, although results obtained by different sample clean-up procedures or those with no clean-up are also presented (direct injection, C18-SPE) resulting mostly in higher LOD or limit of quantitation (LOQ) values. In addition, Remiro et al. [
34] argued that the obtained results were much better represented by median than by average values, so both values were included, where possible.
The range of ochratoxin A concentrations in wines in our research was somewhat larger than that previously reported for Croatian wines [
25,
26] (
Table 5), however, the average and median values showed only a small increase in spite of the substantial span of investigated years (2002 in previous studies-2015 in present study), contrary to the observations of De Jesus et al. [
28] who found nearly two orders of magnitude increase on a decadal scale, although any conclusion based on these findings must be drawn with caution due to much smaller number of red wine samples in previous Croatian studies (seven and ten samples, respectively). The most significant OTA producing species are
Penicillium verrucosum,
Aspergillus ochraceus,
Aspergillus niger and
Aspergillus carbonarius [
35], the last one being the main producer [
36]. Southern Europe is more favorable for growth of ochratoxigenic
Aspergillus than
Penicillium species, since black aspergilli are very resistant to sunshine exposure and to a hot and dry climate [
37,
38] which gives them competitive advantage in comparison with other fungi. However, climate change, especially the rise of average temperatures, will probably lead to substantial changes in OTA production on a decadal scale. Different OTA producing species differ in optimal growth and optimal OTA producing temperatures. Higher OTA concentrations were measured in grapes at 30 °C (upper limit of optimal growth temperature span for
Aspergillus carbonarius and lower limit of optimal growth temperature span for
Aspergillus niger) than in grapes at 20 °C [
38,
39]. With climate change, it is likely that
Aspergillus niger will become prevalent over
Aspergillus carbonarius since it is better adapted to extreme high temperatures and dry conditions [
40] but produces OTA less frequently [
41,
42]. Therefore, the rise of temperatures in moderate climates will most probably lead to a rise of OTA concentrations, while a rise of temperatures to extreme values such as in the Southern regions of Europe will probably lead to decline in OTA concentrations. However, other mycotoxins might become more present in wine, such as aflatoxin, since it is produced by more thermoresistant fungi [
38]. Therefore, small increase of OTA concentration in Croatian wines in comparison with previous reports supports these assumptions, since most of the samples originate from regions with an already hot climate, where further increase of temperatures will probably not lead to further substantial increase in OTA concentrations. Future studies are needed to investigate this proposed trend.
Compared to numerous reports on internationally sourced wines (
Table 5), Croatian wines were seemingly more comparable to wines from Central European countries [
43,
44,
45] although Croatian South (Dalmatia), from which the majority of our samples originated, has typical Csa climate (warm temperature-summer dry-hot summer) of Koppen-Geiger climate classification [
46] which corresponds to climate characteristics of Southern European and Mediterranean countries [
47,
48,
49,
50,
51,
52,
53,
54,
55] (
Table 5) with higher reported concentrations of OTA. Owing to the low limit of detection, the incidence of contamination occurrence was high (92.8%) but comparable to those previously reported for Croatian wines [
25,
26,
34]. According to geographic origin of investigated wines, there was a gradual southwards increase in OTA concentrations with the highest individual concentrations measured in samples produced in southernmost areas of investigated regions (0.163 µg/L in wine from island Vis and 0.141 µg/L in wine from island Hvar). In addition, the frequency of occurrence of OTA contamination increased southwards, as all Dalmatian wines were contaminated (
Figure 1), as well as the prevalence of high level of contamination vs. low level of contamination (
Table 4).
These results strongly support previous findings, as reviewed by Blesa et al. [
57], where wines from southern European regions contained higher OTA concentrations, although a more recent study [
34] has questioned this trend. Higher concentrations of OTA in southern regions, not only on international scale, but also on a within-country scale, were also observed by Labrinea et al. [
49] and Brera et al. [
51]. When comparing the results from different PDOs, the highest values of OTA were measured along the Dalmatian coast, while values measured in wines originating from geographically close and the climatologically very similar Dalmatian Inland were comparable to those measured in continental wines. Although the average daily temperatures and the total rainfall amount are similar in those regions, the Dalmatian Inland does not enjoy the temperature regulation provided by the sea, and is characterized by sharp differences between day-night temperatures. It can be hypothesized that lower night temperatures were responsible for lower measured concentrations of OTA in wines originating from Dalmatian Inland. In fact, Clouvel et al. [
58] have suggested that minimum air temperature of 15 °C during the berry contamination period might correspond to a lower limit below which the fungal growth is restricted. In addition, higher concentration of OTA in coastal wines can be explained by higher air humidity provided by the sea, as previously observed [
53]. It is interesting to note that highest OTA concentrations (highest average, highest maximum) (data not shown) were measured in wines from the climatologically unfavorable year of 2014. In addition, all the wines from 2014 were contaminated with OTA. Year 2014 was the most humid among studied years (+32% on annual level) with temperatures slightly lower than average or without deviation from the average. The greatest increase of rainfall amount occurred in July–September period which probably favored the growth of the ochratoxigenic species [
58].
It is difficult to make any conclusion about the variety-dependent level of contamination since each variety (with exception of Merlot and Cabernet Sauvignon) originated from one region and, therefore, different varieties were grown under different climatic conditions. However, the frequency of occurrence of OTA contamination pointed towards varietal dependence. Blaufränkisch and Merlot were the only wines with non-contaminated samples (22.2%). Considering the dominant effect of climate on OTA contamination, Merlot wines were compared with Cabernet Sauvignon wines (all samples contaminated) since both of these were the only varieties with samples originating from all three studied regions. The result points to variety-dependent occurrence of OTA contamination and suggests the influence of employed viticultural practices. This could be particularly true when considering the time of harvest [
59] since Merlot grapes are known to ripen earlier than Cabernet Sauvignon, which reduces the risk of optimal growth of OTA producing strains at higher temperatures. In addition, frequency of occurrence of OTA contamination in Blaufränkisch wines was compared to that of Pinot noir wines (all wines contaminated) since both grape varieties were grown in the same region (Eastern Continental Croatia). This difference could be explained with the fact that Pinot noir has much thinner skin in comparison to Blaufränkisch and is highly susceptible to pathogenic fungi, like
Botrytis cinerea, which damage the berry and create conditions favorable for colonization of ochratoxigenic fungi [
59].
3.2. Biogenic Amines
All studied wines were contaminated with putrescine and cadaverine, 88.2% with histamine and 82.7% with tyramine. Cabernet Sauvignon wines seemed to be the most and Merlot wines the least susceptible to BA contamination (
Table 6).
In all investigated wines, putrescine was the most abundant, followed by histamine, tyramine and cadaverine (
Table 6) which is in agreement with previous findings [
23]. The ranges of reported results for individual BA were mostly consistent with recent results reported for Italian wines [
60]. However, Mitar et al. [
27] found much lower BA values in Croatian wines. Whether these differences are real features of investigated wines (the study of Mitar et al. included also wines from Western Continental Croatia which were not part of our study), or are consequence of methodologic discrepancies, must remain open for the time being. However, in previous reports, as reviewed by Anzin-Azpilicueta et al. [
11], BA concentrations varied greatly, with ranges from below detection limit to the following amounts: 55 mg/L for putrescine, 25 mg/L for histamine, 28 mg/L for tyramine and 14 mg/L for cadaverine. Konakovsky et al. [
23] reported even higher values for putrescine (122 mg/L) and histamine (27 mg/L) measured in Austrian high-quality red wines. This variability is expected, since BA concentrations depend on various factors, such as the quality of raw material (depending on climate and soil characteristics of wine-making region), aminogenic capabilities of present microbial strains, the vine-growing conditions (including irrigation and use of nitrogen fertilizers), wine-making techniques (including type of maceration and fermentation conditions) and storage conditions [
17]. Due to great number of influencing factors, we observed high within-group (variety) variability which is in agreement with the findings of Konakovsky et al. [
23] who, for the same reason, did not find statistically significant differences between wines produced from seven red varieties.
When observing geographic origin of the samples, wines with highest average histamine concentrations had higher pH values: PDO Dingač (4.2 mg/L, pH = 3.74); PDO Central and Southern Dalmatia (2.4 mg/L, pH = 3.53); PDO Northern Dalmatia (4.3 mg/L, pH = 3.50), while wines with lower pH values had lowest average histamine concentrations: PDO Dalmatian Inland (0.9 mg/L, pH = 3.22); PDO Croatian Podunavlje (1.1 mg/L, pH = 3.38). Average pH values increased southwards which is consistent with the fact that grapes which grow in warmer climate have less acids in comparison with grapes from colder climate. Higher pH values (above 3.5) are connected with higher microbial surviving abilities, while lower pH values (below 3.3) may cause difficult malolactic fermentation [
61]. Data of malic and lactic acid concentrations show that all the wines had gone through partial or complete malolactic fermentation: The malic acid concentrations of investigated wines ranged from not detected to 1927 mg/L, while lactic acid concentrations ranged from 519 to 3982 mg/L. Higher pH values (demonstrated with positive correlation of lactic acid and negative correlation of malic acid with pH values, as shown in
Table 7) in wine that had undergone malolactic fermentation were expected, since microbial malic acid degradation and formation of lactic acid is accompanied by slight increase in pH values. Concomitant synthesis of BA was suggested as microbial strategy to survive acidic environment or to provide alternative metabolic energy [
62]. Positive correlation between pH and BA concentrations was previously reported [
12,
60,
61]. Consistent with our findings, Landete et al. [
12] measured higher histamine concentrations in wines with pH values above 3.5 while the influence of pH was less pronounced for tyramine and not observed for other BA. In our study, statistically significant correlation coefficients between pH values and histamine and tyramine concentrations support these findings (
r = 0.304,
p < 0.05;
r = 0.300,
p < 0.05, respectively). In addition, malic acid was negatively correlated with histamine (
r = −0.328,
p < 0.05), lactic acid was positively correlated with tyramine (
r = 0.302,
p < 0.05), while both histamine and tyramine were highly correlated (
r = 0.605,
p < 0.05) indicating their common origin in malolactic fermentation (
Table 7). Although some authors found positive correlation between BA concentration and ethanol, considering it as enhanced risk for some consumers [
61], we observed no connection between BA concentrations and neither ethanol content nor SO
2 concentrations. Considering changes of biogenic amines concentrations during conservation, no correlations were observed between age of wine and biogenic amines in investigated wines (
Table 7), contrary to the several previous reports where histamine was found either to increase in the first six months and then decrease afterwards [
12], increase in the first 105 days [
63] or to decrease [
64] during storage. However, these studies were conducted over a relatively short period of time (maximum of 12 months) so studies evaluating longer periods of conservation, especially important for high quality wines, should be performed.