Comparison of organic and conventional Italian cheeses chemical composition from parallel production

Although there are several studies comparing organic and conventional milk characteristics, very few focused on dairy processed products such as cheese. Thus, this study aimed for a detailed controlled examination of gross composition, minerals, and the fatty acid profile of organic ( ORG ) and conventional ( CON ) Italian cheeses from parallel production. Four Italian cheese types were analyzed: ‘Latteria’ (ORG, n = 9; CON, n = 10); ‘Asiago Protected Denomination of Origin ( PDO ) fresco’ (ORG, n = 9; CON, n = 9); ‘Caciotta’ (ORG, n = 8; CON, n = 8); and ‘Mozzarella Traditional Special-ties Guaranteed ( TSG )’ (ORG, n = 14; CON, n = 14). Cheese samples were collected from September 2020 to August 2021. Gross composition, minerals, and fatty acids were determined using infrared spectroscopy. Within each cheese type, paired ORG and CON samples were compared using a non-parametric Wilcoxon signed-rank test. ‘Latteria’ showed lower PUFA, n-3, and n-6 content, and greater Fe, K, C10:0, C12:0, and C16:0 content in ORG than CON ( P < 0.05). ‘Asiago PDO fresco’ showed lower protein and Zn content, and greater salt, ash, and Na content in ORG than CON ( P < 0.05). ‘Caciotta’ showed lower ash, n-3, and n-6 content and greater K, C4:0, C8:0, C10:0, C14:0, and C16:0 content in ORG than CON ( P < 0.05). ‘Mozza-rella TSG’ showed lower fat and, therefore, fatty acids content, and greater moisture, ash, and Mg content in ORG than CON ( P < 0.05). In conclusion, few significant differences in chemical composition were observed between ORG and CON cheeses, regardless of the type


INTRODUCTION
Cheese production from whole cow milk has been steadily increasing in the last 5 yr in the European Union (EU) up to 7.8% (FAOSTAT, 2022).Italy is one of the Top-3 cheese producing countries in the EU, with a 14.6% share, after France and Germany (FAOSTAT, 2022).Asiago Protected Denomination of Origin (PDO) 'fresco' (>20 d of ripening) or 'stagionato' (>90 d of ripening) is produced from cow milk in Northeastern Italy, being the fourth most produced cheese type in that country (CLAL, 2022).Mozzarella Traditional Specialties Guaranteed (TSG) is a soft cheese from cow milk.Caciotta is an aged or semi-aged soft cheese in a cylinder shape from cow's, sheep's, goat's, and buffalo's milk or a mix of these milks.Latteria is a general term to describe a semi-hard cheese from cow milk made in the Veneto region in Northern Italy.On the other hand, all organic (ORG) retails sales in the EU continued its growth in 2020, where Italy is among the Top-3 countries after Germany and France with 3,872 million €, and is the country with the largest number of ORG producers (>70,000) and ORG processors (>22,000) considering all goods (Willer et al., 2022).These data support the relevance of Italy as a cheese producer as well as an ORG producer in the EU considering all goods.The difference between ORG and conventional (CON) cheese making procedure is only related to the different type of milk at beginning of cheese making.The cheese making conditions (e.g., temperature, time, operation), the type of rennet, and bacteria are the same between ORG and CON cheeses.Thus, ORG cheese should be made from ORG milk and manufactured avoiding cross-contamination with CON cheese (i.e., produce first on the morning after the machinery was clean the day before).The ORG milk come from ORG farms which mainly differ from CON ones on the feed (ORG feed), stocking density, and veterinary treatments while all other inputs and management remain unchanged.
Research found that people with university education and greater concern about environmental sustainability are the consumer group with the highest willingness to pay more for ORG than CON food (Gracia and de-Magistris, 2016).For ORG cheese it was found a greater expected liking score that CON cheese, probably due to a satisfactory consumer trust in ORG legislation (Napolitano et al., 2010).Consumers perceive ORG being better (i.e., healthier and better quality) than CON products (Rodríguez-Bermúdez et al., 2020), but there is a lack of studies on the intrinsic quality traits of cheese comparing both production systems.Studies are usually conducted in fluid milk rather than in processed dairy products (Średnicka-Tober et al., 2016).The very few studies on ORG cheese used samples from different dairy producers for the CON cheese (da Silva Cândido et al., 2020;Abreu et al., 2021) or focus only on ORG cheese (Popovic-Vranjes et al., 2016;Łepecka et al., 2022).In particular, da Silva Cândido et al. ( 2020) and Abreu et al. (2021) focused on the isolation of Staphylococcuss sp. and the bacterial diversity from ORG and CON fresh minas cheese, respectivily, and did not evaluate cheese chemical composition.Those authors concluded that the microbiological differences observed cannot be attributed to the production system itself -i.e., being ORG or CON-because the cheesemaking process was different (da Silva Cândido et al., 2020); Abreu et al., 2021).To the best of our knowledge, there are not further studies comparing ORG and CON cheese.Therefore, the aim of this study was to evaluate compositional (gross composition, and mineral and fatty acid content) differences between ORG and CON Italian cow cheeses 'Latteria', 'Asiago PDO fresco', 'Caciotta', and 'Mozzarella TSG' from parallel production.Variation do to different processing and milk sourcing regions were removed by our research design using the same processing plant, cheese types, and milk source in the region, however differently produced on ORG or CON farms.

Sample Collection
Respecting the mandatory rules reported by organic standards (e.g., separation of milk, traceability of batch and wheels), samples (100 g) of ORG and CON cheese of each type paired by factory, production day, batch, cheesemaking process, and ripening time were collected.Because of ORG cheese processing regulation for what is called 'parallel production' ORG and CON processing on the same factory line the ORG cheese was always made first after cleaning the line the day before.Cheeses were sampled almost monthly across 1 yr (September 2020 to August 2021).Cheese types included in the present study were: 'Latteria' (ORG, n = 9; CON, n = 10); 'Asiago PDO fresco' (ORG, n = 9; CON, n = 9); 'Caciotta' (ORG, n = 8; CON, n = 8); and 'Mozzarella TSG' (ORG, n = 14; CON, n = 14).
Samples arrived refrigerated (4°C) at the laboratory of the Department of Agronomy, Food, Natural resources, Animals, and Environment of the University of Padova (Legnaro, Italy) and 1.5 cm of the rind were removed in 'Latteria', 'Asiago PDO fresco', and 'Caciotta' samples.Each cheese sample was homogenized with a knife mill Retsch Grindomix GM200 (Retsch GmbH & Co, Haan, Germany) and the ground cheese was kept in a sealed plastic bag at 4°C and analyzed within 2 h to avoid variations in cheese composition depending on moisture loss.

Chemical Analyses
Samples were analyzed using a near-infrared spectrophotometer in transmittance mode FoodScan TM Dairy Analyzer (Foss Electric A/S) which worked at room temperature (20°C).A 50-g ground sample of each cheese was placed in the cup glass of FoodScan TM (diameter 140 mm, depth 20 mm) assuring that the sample covered all the bottom surface of the plate without leaving empty spots, and was scanned from 850 to 1,050 nm every 2 nm.Each spectrum was the average of 16 subspectra collected during the automatic rotation of the cup and recorded as log(1/transmittance) using ISIscan Nova and Mosaic software (Foss Electric A/S).

Statistical Analysis
Compositional data for all cheese varieties were not normally distributed (Shapiro-Wilk test) and it was not possible to normalize the data using Box-Cox transformations (Box-Cox and Cox, 1964;Osborne, 2010).Therefore, differences in composition between ORG and CON within each cheese type were tested pairwise using the non-parametric Wilcoxon Singed-Rank test through the UNIVARIATE procedure of SAS ver.9.4.(SAS Institute Inc., Cary, NC).Results are reported as median and 95% confidence interval of the estimate (95% CI).Significance was established at P < 0.05 unless otherwise stated.

Gross Composition
'Latteria' gross composition of 5 traits (moisture, protein, fat, ash, and salt content in percentage as sampled) was similar between ORG and CON cheese.However, 'Latteria' tended to a lower ash content and a greater protein content in ORG than CON (P < 0.10; Table 1).'Asiago PDO fresco' showed lower protein content, and greater salt and ashes content in ORG than CON cheese (P < 0.05; Table 1).'Caciotta' presented greater protein and lower ashes content in ORG than CON cheese (P < 0.05; Table 1).However, 'Caciotta' tended to a lower moisture content in ORG than CON (P < 0.10; Table 1).'Mozzarella TSG' had lower fat content, and greater moisture and ash content in ORG than CON cheese (P < 0.05; Table 1).

Mineral Profile
Among all the mineral evaluated, few differences were observed (Table 2).'Latteria' had (P < 0.05) greater K and Fe, and tended (P < 0.10) to lower Mg content in ORG than CON cheese (Table 2).'Asiago PDO fresco' showed greater Na and lower Zn content in ORG than CON (P < 0.05).'Caciotta' presented greater K content in ORG than CON cheese (P < 0.05), and tended to lower Na and greater S content in ORG than CON cheese (P < 0.10).'Mozzarella TSG' revealed greater Mg content in ORG than CON cheese (P < 0.05), and tended to greater Ca and lower K content in ORG than CON cheese (P < 0.10).

DISCUSSION
Milk composition is often affected by factors over which the cheesemaker has no control at all.Such are animal feeding, breed, and season (Law and Tamime, 2011).Moreover, due to the requirements established by the ORG regulation (European Union, 2018), ORG farms usually have different management and feeding practices than CON farms.To standardize the cheese manufacture and ripening process, the cheesemaker standardizes milk composition by adding milk solids or removing cream to give continuity in composition and protect consumers and the manufacturing process (Law and Tamime, 2011).Thus, the cheesemaker has complete control on the cheesemaking process and ripening (Law and Tamime, 2011).However, milk standardization is not allowed in the production of PDO cheese.To be sure that the cheesemaking process and the ripening time were the same for both ORG and CON samples from each cheese type, they were collected from the same factory through a complete year.Therefore, in our study results in the end product mirror milk origin (animals' management practices, breed, etc.) and not the cheesemaking process and ripening time.We are not aware of previous studies comparing ORG and CON cheese composition, thus results will be discussed in relation to milk differences between both systems.
Before discussing the results of ORG vs. CON cheese, we would like to point-out that values of gross composition, minerals, and fatty acids for both ORG and CON cheese (Tables 1 to 4) were in line with those reported by Manuelian et al. (2017).Those authors characterized the mineral profile through ICP-OES after mineralization with nitric acid, and the fatty acid profile using GC of 18 varieties of commercial cheeses including soft, semi-hard, and hard products.The comparable results of the present study with those obtained by Manuelian  2017) are supportive of the adequacy of infrared prediction models used in our study for the quantification of minerals and fatty acids content.Differences reported in the literature between ORG and CON milk are related to several factors which are known to impact milk quality such as feeding, breed, and stage of lactation (Schwendel et al., 2015).However, studies related to the evaluation of ORG and CON products often neglect to consider other factors beyond the farming system (Schwendel et al., 2015).A clear example is when milk fatty acid composition is evaluated, because the fatty acid profile responds to the farming practices (e.g., high input vs. low input) rather than being the milk ORG or CON (Schwendel et al., 2015).

Gross Composition
None or up to 3 traits significantly differed between the ORG and CON cheese samples and, despite being significant, those differences represented < 11%.The literature reports contradictory results regarding milk fat (Schwendel et al., 2015), protein (Schwendel et al., 2015;Średnicka-Tober et al., 2016), and lactose (Schwendel et al., 2015) content, which could explain the different trends observed across the cheese types we analyzed.Some studies reported greater fat content (% or g/kg of milk) in ORG than CON milk (Butler et al., 2011;Schwendel et al., 2015), while other observed lower content in ORG, or no differences at all (Schwendel et al., 2015;Średnicka-Tober et al., 2016).The diversity of results related to fat content can be explained by the different breeds used by ORG (often non-Holstein-Friesian breeds) and CON (often Hol- because as being a PDO cheese it should follow the PDO guidelines.The other cheese varieties evaluated significantly differed between the ORG and CON cheese samples up to 46%.On the other hand, the differences for the determined individual fatty acids were in general < 11%.Milk fatty acids are the main compounds evaluated in ORG milk, as they are influenced by the diet (including pasture), stage of lactation, and season (Schwendel et al., 2015).It has been reported that minor dietary differences can modify the fatty acid profile of both ORG and CON milk (Schwendel et al., 2015) without affecting the total SFA percentage (Średnicka-Tober et al., 2016).In the current study, few differences in the fatty acid composition between ORG and CON cheeses were observed, and the majority of those differences were detected in 'Caciotta' and 'Mozzarella TSG'.In fact, 'Caciotta' and 'Mozzarella TSG' were the only cheese types that tended to present lower moisture content in ORG than CON cheese.Moreover, 'Mozzarella TSG' had lower fat content in ORG than CON cheese which explains the lower amount of several individual and groups of fatty acids in the former than the latter.Thus, our results partially agreed with previous results in bulk milk collected in the same area where the fatty acid profile was almost identical in ORG than CON milk (Manuelian et al., 2022).The similarity of the fatty acid content is mainly related to the animals' feeding regimens (Manuelian et al., 2022).The cheese factories involved in the present study collect milk from the wider Po Valley, where pasture is scarce and feeding of the lactating cows -both in ORG and CON-is based on a total mixed ration that includes corn meal and/or corn silage.The inclusion of corn, which is a C4 plant, in ORG dairy farms does not allow to discriminate ORG from CON milk, because it is identified using the carbon stable isotope ratio milk method which relays on the fact that most CON farms use corn, whereas ORG farms does not (Schwendel et al., 2015;Inácio and Chalk, 2017).Therefore, the few differences we detected are in line with previous studies on ORG milk (Średnicka-Tober et al., 2016;Manuelian et al., 2022).Nevertheless, Średnicka-Tober et al. ( 2016) reported greater n-3 and PUFA, and lower n-6 percentage in ORG milk, whereas we observed a lower content (g/100g cheese) of n-3 in 'Latteria' and 'Cacciota', and of PUFA in 'Latteria'.

CONCLUSIONS
Our study of 4 different cheeses sampled over many month, showed that ORG and CON cheese differed slightly in terms of gross composition, and mineral and fatty acid content, depending on the cheese type being analyzed.'Asiago PDO fresco' and showed fewer differences between ORG and CON samples than 'Latteria', 'Caciotta', and 'Mozzarella TSG'.Nevertheless, fatty acid content differences between ORG and CON samples of 'Mozzarella TSG' are influenced by the lower fat content in the former than the latter.The huge similarity between ORG and CON of 'Asiago PDO fresco' could be as a results of dealing with a PDO cheese that should follow specific guidelines.This study contributes to fill the gap of information regarding the differences between ORG and CON cheese as there are very few studies on the subject.

Table 1 .
Manuelian et al.: ORGANIC vs.CONVENTIONAL CHEESE Gross composition (% as sampled) of organic and conventional cheese.Significant differences between gross composition of organic and conventional cheese are reported in bold

Table 2 .
Manuelian et al.: ORGANIC vs.CONVENTIONAL CHEESE Mineral composition (mg/kg) of organic and conventional cheese.Significant differences between mineral content of organic and conventional cheese are reported in bold 295% CI = 95% confidence interval of the estimate.

Table 4 .
Manuelian et al.: ORGANIC vs.CONVENTIONAL CHEESE Individual fatty acids (g/100 g of cheese) of organic and conventional cheese.Significant differences between individual fatty acids of organic and conventional cheese are reported in bold