Labelling control of pasture milk

Authentication by multivariate analysis?

Unlike organic milk and hay milk, which are clearly regulated in producer guidelines and EU regulations, there are no legal requirements for designations such as Alpine milk, farm milk or pasture milk. However, producers of pasture milk guarantee “natural” grazing for at least 120 or even 150 days a year for at least six hours a day. In order to provide options for protecting consumers from false claims, scientists at the Max Rubner-Institute have now examined analytical approaches for identifying pasture milk in detail.

Contrary to common consumer expectations, “pasture milk” does not come from animals that feed exclusively on grass directly from the pasture. To achieve a higher milk yield, significant proportions of silage and concentrates are usually fed in the barn in addition. While analytical criteria for authenticating organic milk have already been developed at MRI, analyses of pasture milk did not reveal a composition similar to that of organic milk despite the potentially increased content of pasture feed. Levels of omega-3 fatty acids were not significantly higher than in conventional milk without a pasture label (conventional milk).

To determine suitable parameters for identifying pasture milk, around 300 commercial samples of pasture milk and conventional milk as well as a further 26 organic milk samples were collected in a project over a period of 2.5 years. The pasture milk came from dairies in the German states of Schleswig-Holstein, Mecklenburg-Western Pomerania and Saxony. The analysis of the stable isotope ratios of carbon (δ13C) and nitrogen (δ15N) as well as the fatty acid and triglyceride composition showed that none of the 90 individual parameters or their bivariate correlations were suitable for differentiating between pasture milk and conventional milk. Even considering only the typical grazing period from May to October did not allow pasture milk to be identified. However, combining δ13C and δ15N confirmed the unambiguous authentication of organic milk throughout the year.

A multivariate analysis with the aim of authenticating pasture milk was carried out both with the individual data sets of fatty acids and triglycerides as well as with combinations of the data sets of all three analytical methods used. The most accurate prediction model was obtained using PLS discriminant analysis of the fatty acid data. Excluding organic milk, the prediction accuracy for pasture/non-pasture was 92 percent. More conventional samples were assigned to pasture milk (10 %) than pasture milk was assigned to a conventional origin (6 %). Although the term “pasture milk” for milk produced in the pasture-free period from November to April is questionable in principle, the model demonstrated the ability to differentiate independently of the season. Overall, the findings indicated a regional rather than a husbandry-related distinction.