EU adopts a new reference test for pasteurisation

Frank Harding OBE, formally Technical Director at the England & Wales Milk Marketing Board, examines recent changes.

It was recognised in the 1930s, that alkaline phosphatase (ALP), an enzyme naturally present in all milks, is inactivated at time temperature conditions slightly higher than those needed to kill Mycobacterium tuberculosis and most other pathogens in milk (71.7ºC/15seconds). These conditions are used commercially to pasteurise milk (HTST pasteurisation). Test methods based on ALP measurement were developed in order to demonstrate that milk has been correctly pasteurised.

Tests developed for ALP measurement in the 1930s and used for many years were based on colorimetric methods of analysis. Such tests are only semi-quantitative and relatively insensitive, with a limit of detection of about 0.1% raw milk.

In 1990, Rocco[1] published the results of a collaborative trial of a fluorimetric method, which was linear over a wide range of ALP values, quantitative, precise and sensitive to raw milk levels below 0.01%. The method, named Fluorophos ALP, was developed and marketed by Advanced Instruments, Inc., USA.

In a number of studies the Fluorophos has been compared with existing, less sensitive colorimetric reference methods and a relationship established around the 0.1% raw milk cut off point.

Table 1:
Determination of ALP by AOAC and Fluorophos collaborative study (Rocco [1]) using cow’s milk in the USA.

Milk Type	% of raw milk	Fluorophos Mean value mU/L	AOACMean value µg phenol/ml
Whole	0.05	256.2	0.83
	0.10	494.6	1.46
	0.20	960.2	2.70
Skim milk	0.05	262.3	0.60
	0.10	507.6	1.02
	0.20	994.6	2.14
Chocolate	0.05	262.8	0.38
	0.10	521.3	0.78
	0.20	1020.2	1.40

Lechner [2] studied the Fluorophos method using cow’s milk in Germany.

Table 2:
Study by Lechner

% raw milk in pasteurised milk	FluorophosmU/l
0.1	540
0.2	1026
0.3	1455

Langridge[3] summarised the results of a study involving 22 laboratories testing milk by Fluorophos, Aschaffenburg Mullen and the colourimetric EU reference method in use at that time. A comparison of the Fluorophos and EU colourimetric method based on results of blind duplicate samples is given below.

Table 3:
Comparison of EU (Rocco) method and Fluorophos (Langridge) using cow’s milk sourced in the UK

raw milk in pasteurised milk	Fluorophos methodmU/L EU	methodµg phenol /ml
Sample A 0.05%	257.4	1.26
Sample B	257.4	0.94
Sample A 0.10%	458.8	2.14
Sample B 0.20%	498.3	1.49
Sample A 0.15%	693.7	1.80
Sample B	713.5	2.08
Sample A 0.20%	911.6	2.29
Sample B	851.8	2.44

In the above studies, a Fluorophos value of 500mU/l was demonstrated to accurately assess the presence of un-pasteurised milk at the statutory cut off point (about 0.1% raw milk).

A study using French milk however gave different values for 0.1% raw milk in pasteurised milk:

Table 4:
October 2005 Study by EU reference laboratory (afssa).

% raw milk	IDF 63: 1970Colorimetric determinationµg phenol/ml	IDF 155-1: 2001(Fluorimetric method)mU/l
0.05	1.5	–
0.1	3.0	939

Where comparisons were made with colourimetric methods the variability of the colourimetric methods has been cited as the reason for poor agreement.

Changes to ALP testing within the EU

In May 2007 the EU approved the fluorimetric method (ISO 11816-1: IDF 155-1) as the “Official reference” method for alkaline phosphatase for applications related to EU sanitary regulations. The EU now requires that any other method being considered for use is validated against the fluorimetric method and lowered the statutory level for ALP in pasteurised milk from 500 to 350mU/l.

This change means that ALP in milk is now defined as the result obtained when ISO 11816-1: IDF 155-1, the fluorimetric method, is applied.

**Comparison of flourophos with other test methods

**Proficiency studies by an independent company in the UK have shown a significant bias between the Fluorophos and Charm methods of assessing ALP activity.

Comparison of the mean values obtained on the same samples tested under the Quality Management “Proficiency Testing Scheme”

Date	Test material	Fluorophos mean value	Number of samples tested	Charm mean value	Number of samples tested	Difference Charm – Fluorophos
Jan 05	31B105	173.91	189	306	40	187
July 05	31A111	242.61	157	261.3	52	19
July 05	31B111	80.62	155	95.04	52	15
Oct 05	31A144	472.89	258	768.29	52	296
Oct 05	31B144	447.35	256	997.65	52	550
Jan 06	31A177	80.41	161	123.36	35	43
Jan 06	31B117	134.82	161	234.47	34	100
	Mean of the means for all samples	222.94		398.0		175 (44%)higher than Fluorophos

The EU reference laboratory (afssa, Paris, France) compared Charm (NovaLum) and Fluorophos and also showed that Charm is positively biased when compared with Fluorophos. (see chart above right)

Scintu et al.[4] has compared the EU method with the Fluorophos for Ewes milk and showed that the R and r values were close to those quoted for cow’s milk.

Recommendation

It is recommended that the fluorimetric method is now adopted internationally as the reference method for measurement of Alkaline Phosphatase in pasteurised dairy products.

References:
[1]Rocco; J.ASSOC. OFF. ANAL. CHEM. Vol. 73 No 6. 1990
[2]Lechner & Ostertag Deutche Milchwirtschaft 23 (1146 – 1149) 1993.
[3]Langridge International Food Hygiene 10 (3) 31-33 1999)
[4]Scintu et al (J Food Protection, Vl63. No 9, 2000 Pages 1258-1261)