Milk sampling

It was a stratified sampling with proportional allocation, taking into account the number of production units (PU) in each municipality, and as reference variable the number of cows by PU. Samples (n= 198 ) of raw milk from different PU locations were taken: Tierra Blanca (114), Tres Valles (6), Carlos A. Carrillo (18), Laguna de lagarto (30), Cosamaloapan (6), Acula (12) and Ixmatlahuacan (12), belonging to the DDR 008 Cd. Alemán, Veracruz, México, which together provide 90.5 % of the milk production in the district¹⁰. Samples were taken from the storage tanks in 60 ± 10 min, after milking and moved in a cooler box at 4 ± 1 °C for analysis in a time span of 3 ± 1 h¹¹.

Physicochemical characteristics

Fat content (FT), protein (PR), lactose (LA) and nonfat solids (NFS), were evaluated. Additionally, density (DE) and cryoscopy (CR) by ultrasound using a scanner Lactoscan S (Milkotronic Ltd, 4 Narodni Buditeli str. 8900 Nova Zagora, Bulgaria) previously calibrated and validated, presenting an average error of 0.025% with determinations made in accordance with the AOAC¹².

Contamination indicators groups

A mixture of milk (10 mL) and 90 mL of sterile peptone solution was placed in a homogenizer (Stomacher®, model 400 circulator Seward Limited, UK) for a minute at 265 rpm. Dilutions of 10^-1, 10^-2 and 10^-3 were prepared to assess the total bacterial count (TBC) and total coliform counts (TCC)¹². Values obtained in colony forming units per milliliter (cfu/mL) were transformed into log₁₀ scale¹³ to be used for further analysis. Titratable acidity (TA) was reported as the average value in grams of lactic acid per liter of milk (g/L)¹⁴.

Management practices

In each PU a questionnaire was applied to obtain information on the production system (PS), predominantly breed, type of milking, supplementation practices, use of oxytocin and milk yield (L·cow^-1).

Statistical analysis

A univariate analysis was performed to describe management practices and milk characteristics. Two-way analysis of variance (ANOVA) were applied to determine effects of PS, breed, type of milking, supplementation and use of oxytocin on milk characteristics and milk yield. Cluster analysis served to group milk based on their quality from FT, NFS, CR and TA variables, establishing as reference the Euclidian distance, using the full linkage algorithm. The test of least significant difference of the Fisher procedure established the difference among the identified groups to a significance level of 5%. Statistics statistical package version 7.0 was used.

Physicochemical characteristics

The contrast of results with standard references it shows in Table 1. The FT content showed a high variability with an average of 34.8 g/L, which accomplished the national standard NOM-155-SCFI-2012^(14-) and the international standard signed by FAO¹⁵. The average values of PR (29.5 g/L), LA (41.9 g/L), NFS (79.2 g/L), CR (- 0.490 °C) and DE (1027.45 g/L) did not meet the minimum intervals of the standard NOM-155-SCFI-2012¹⁴ and those established by FAO¹⁵, except the PR that met the minimum international standard average.

Table 1 Physicochemical quality, spoilage indicators groups of milk and comparison with national and international standards 

¹NOM-155-SCFI-2012, ²Calderón et al. (2006), ³NMX-F-700-COFOCALEC-2004, ⁴PMO, 1995, ⁵Draaiyer et al. (2009). The results represent the average of three determinations ± standard deviation

Contamination indicators groups

Table 1 presents the results of bacterial counts, TA and comparison with reference standards. The TBC (1.08 x 10⁴ cfu/mL) and TCC (5.80 x 10² cfu/mL) met with regulatory values. However, there was a wide variability in TCC, reaching 28 % of the samples above the standards set out in the standard NOM-155-SCFI-2012¹⁴. The average value of TA complied with the values set by the standard NOM-155-SCFI-2012 (14) and FAO¹⁵, but in the interval (0.67 to 1.75 g/L) it was observed that 2 % of the samples had values above those established by Mexican standards.

Management practices and relationship with the characteristics of the milk

Breed crosses present in the PU were Swiss x Zebu (61 %), Holstein x Zebu (9.4 %), Zebu (7.8 %) and crosses not defined (22 %). The herd was handled under a grazing system (87. 5 %) and semi-stabled (12.5 %). Milking was carried out manually (94 %), and less frequently in a mechanical manner (6 %). In 20.5 % of cases, cows receive balanced feed supplementation. The milk ejection was induced: without using oxytocin (47 %), with partial use, only in hard milking cows (30 %) and application to the entire dairy herd (23 %). The milking herd averaged 34 cows, with an average yield of 4.95 L·cow^-1.

The results of the ANOVA (Tables 2 and 3) showed that in the UP where the Holstein x Zebu crossing dominated, yield increased with a lower content of FT and PR in milk and CR elevated values (P<0.05). Type of milking affected (P<0.05) bacterial counts, since the milk obtained by mechanical milking presented counts and TA lower than those by manual milking did. Furthermore, mechanical milking apparently increases the yield; however, the two-way ANOVA revealed that it relates more to a mix of factors, as supplementation and oxytocin application. Application of oxytocin in the entire herd increased (P<0.05) the yield when compared to the partial or no use of, in relation with a low content of solids.

Table 2 Effect of management practices on the physicochemical milk quality and productive yield 

ZE= Zebu. SZ= Swiss x Zebu, UD= Undefined breed, HZ= Holstein x Zebu. HAN= Hand, MEC= Mechanic. NO= Not apply, PAR= Partial application, TOT= Total application. NO= Not supplement, YES= Supplement. GR= Grazing, SES= Semi-stabled.

^ab Values with different superscripts within the same column differ (P<0.05).

Table 3 Effect of management practices on the spoilage indicators groups of milk 

TBC= Total bacterial count; TCC= Total coliform count; HA= Hand; MEC= Mechanic. GR= Grazing. SES= Semi-stabled.

^ab Values with different superscript within the same column differ (P<0.05).

The PS based on grazing was related to elevated values of NFS, PR, LA, DE and lower values of CR, when compared with the semi-stabled system. There were no differences (P>0.05) in milk yield between both systems. Grazing-based system related to a higher bacterial load and TA (P<0.05) to that seen in milk from semi-stabled systems

Classification of milk quality

Cluster analysis allowed classifying milk samples in five groups of quality based on the variables FT, NFS, CR and TA. Dendrogram in Figure 1 shows association of PU with each group. The distance between the groups fluctuated between 3 and 39. The distance between the groups of milk from very poor quality with groups of excellent quality was 39; good quality was 33; poor quality 24 and very poor quality was 3. According to the ANOVA, differences were observed (P<0.05) among quality groups for the different variables (Figure 2). The so-called group of excellent quality, represented 14 % of the sample and presented the highest values of FT (43.60 ± 2.49 g/L), NFS (83.34 ± 1.11 g/L) and TA (1.36 g/L) and the lowest in CR (-0.520 ± 0.005 °C); this group met with the values set by the standard NOM-155-SCFI-2012 and FAO^14,15. The good quality group accounted for 39 % of the sample and presented values of FT (32.57 ± 2.80 g/L) and NFS (82.13 ± 1.24 g/L) less than the excellent quality group, values of CR (-0.512 ± 0.010 °C) superiors and TA (1.39 ± 0.14 g/L) similar to that group. This group met the minimum values set out in international and national standards, with the exception of the of NFS content which was below that required by the standard NOM-155-SCFI-2012¹⁴. Groups of poor (21 %), bad (3 %) and very poor quality (23 %) showed similar values of NFS (between 77.20 and 77.80 g/L) below the established standards of reference and similar high values of CR (between -0.460 and -0.480 °C). They presented values below established standards for FT (between 15 and 10 g/L), except for the group of poor quality which presented high values of FT (42.80 ± 5.30 g/L) similar to the values of the excellent quality group.

Figure 1 Dendrogram of milk samples by productive units using the complete linkage method 

^abcd Values with different superscript differ (P<0.05).

Figure 2 Characteristics of groups formed based on milk quality 

Physicochemical characteristics

The FT content was similar to the interval reported in tropical zones (32.6 to 34.5 g/L)¹⁶ and can be linked to the breed pattern Bos taurus x Bos indicus⁶. The content of PR, LA and NFS was lower than the range reported in other studies^16,17. This low solids content, often is related to a dilution effect ⁴ by the high-performance caused by the B. Taurus component⁷. However, the content of solids found in conjunction with the CR values superior to -0.530 °C and DE below 1,028 g/L, reveal practices of adulteration by addition of water⁶, which has serious implications for the processing industry and generates risks to health¹⁸.

Contamination indicators groups

TBC values were similar to those observed in another study, which suggest good hygiene milking practices¹⁹. Twenty eight (28) percent of samples out of the norm in TCC reflect failures in post milking management and lack of the elimination of milk or residual water from the tanks. The values found in the analyzed milk contamination indicators groups guaranteed an efficient pasteurization process¹³, but it does not endorse their safety, so the consumption of milk and unpasteurized products represents a health risk for the possible presence of pathogenic bacteria²⁰.

Management practices and relationship with the characteristics of the milk

Yield increase related to the Holstein x Zebu crossing has been reported previously²¹ and in this study was also related with a low content of solids (Table 2), however, the CR and DE values are more related to adulteration practices than to physiological factors²². Milk yield increased with the use of oxytocin. It has been observed previously, and is attributed to a greater alveolar contraction of the myoepithelial cells, moving the milk to the cavities available for ejection²³. The use of oxytocin increases milk production, but must take into account the low solids content which compromises its quality for being processed²⁴; a better option would be the supplementation of milking cows, which increases milk yield, from 4.80 to 5.50 L·cow^-1, without affecting (P>0.05) the solids content.

The effect of mechanical milking on the bacterial count has been reported²⁵, but it is necessary to emphasize that cleaning of the milking equipment with detergents is essential to avoid the milk residues⁴ that relate to contamination by coliforms²⁶. PS of grazing with high bacterial counts was related to the lack of milking parlous, since cows must be milked in open lots, which increases the chances of environmental contamination²⁵.

Classification of milk quality

Excellent and good quality groups (53 % of the sample) met the values required by reference standards, so actions should focus on increasing the level of technology implemented in the PU, which is related to efficiency and competitiveness²⁷. Good quality group failed to meet the minimum value of NFS in accordance with NOM-155-SCFI-2012. This problem could be related to small amounts of residual water²⁸ and can be solved through the training of operators in the milking and storage procedures.

The problematic groups named as poor, bad, and very poor quality (47 % of the sample) presented low NFS and high CR values. The adequate actions for these groups should be to satisfy the demands of nutrients from the dairy herd, taking into account the variability in the availability of forage in tropical conditions²¹ and that the nutritional aspect is a key component to improve production of dual purpose systems²⁹. Another important aspect is the awareness by the producers, that the intentional addition of water to the milk represents a fraud¹⁸.