Donkey Milk Quality and Safety: Challenges of Using the ISO 11816-1 ALP Method for Pasteurization Verification

TL;DR

This study examines whether the ISO 11816-1 method for verifying pasteurization is suitable for donkey milk, finding it ineffective and proposing a new threshold.

Contribution

Proposes a new ALP threshold for donkey milk pasteurization verification to preserve nutraceutical properties.

Findings

Residual ALP values in donkey milk were below bovine thresholds, making the ISO method unsuitable.

A provisional ALP threshold of 61 mU/L is suggested for donkey milk pasteurization verification.

Native ALP levels varied widely and were influenced by lactose and physiological parameters.

Abstract

Donkey milk is valued for its similarity to human milk, low allergenicity, and high nutraceutical content, particularly lysozyme. As for milk from other species, donkey milk requires pasteurization for commercial sale. The ISO 11816-1 alkaline phosphatase (ALP) method, developed for cow’s milk, is currently used to verify pasteurization in donkey milk. This study evaluated whether the method distinguishes compliant from non-compliant treatments and investigated associations of native and residual ALP with chemical and physiological parameters. Milk from 14 Ragusano donkeys was sampled three times at three-month intervals (42 samples) and analyzed for chemical and physiological composition, lysozyme, and total antioxidant capacity. Samples underwent compliant and non-compliant heat treatments, and ALP was determined. All residual ALP values, even for non-compliant treatments, were well below the bovine ISO threshold, and non-compliant treatments were not fully distinguished, precluding this method as a suitable approach. Pasteurization at 76 °C × 15 s is suggested, with an explorative and provisional threshold of 61 mU/L to achieve 100% specificity while preserving nutraceutical components. Considering a 9% analysis uncertainty at 95% confidence, ALP values below 50 mU/L are fully negative, with 97% of compliant samples below this value. Native ALP was highly variable (1758–6086 mU/L) and depended on physiological parameters and lactose, but not fat. Validation on larger datasets and across other breeds is required.

Figures