1047

Journal of Food Protection, Vol. 68, No. 5, 2005, Pages 1047–1053

Copyright Q, International Association for Food Protection

Collaborative Evaluation of a Fluorometric Method for

Measuring Alkaline Phosphatase Activity in Cow’s, Sheep’s, and
Goat’s Milk
FRANK HARDING1* AND EILEEN GARRY2

18 Sharon Close, Bookham, Surrey KT233LB, UK; and 2Advanced Instruments, Inc., 2 Technology Way, Norwood, Massachusetts 02062, USA

MS 04-374: Received 26 August 2004/Accepted 10 December 2004

ABSTRACT
Pasteurization of raw milk was introduced to extend product shelf life and destroy pathogens. The measurement of alkaline
phosphatase (ALP) activity has been used as an indicator of proper pasteurization in dairy products for more than 65 years.
This study was undertaken to evaluate six different fluid dairy products at lower phosphatase levels than previously verified
using the Fluorophos Test System, a sensitive and precise method for ALP activity detection. Thirteen laboratories participated
in this collaborative, international study to evaluate the fluorometric test at 20, 40, 100, 350, and 500 mU/liter and extend the
scope of the method to include milk from not only cows but also goats and sheep. Initially, the statutory level of ALP
measured fluorometrically was set to equivalent levels of colorimetric test standards (500 mU/liter). The European Union
recently announced its intention of lowering the legal limit from 500 to 350 mU/liter and, in addition, setting a target value
of 100 mU/liter, which if exceeded would trigger an investigation into the pasteurizer plant performance. At 500 mU/liter of
ALP, this trial generated relative standard deviation of repeatability values of 6.48, 5.69, and 1.74% and relative standard
deviation of reproducibility values of 14.66, 13.30, and 5.33% for all cow’s, sheep’s, and goat’s milk samples, respectively.
Data from this study are comparable to data from previous studies and indicate the suitability of the Fluorophos Test System
method for measuring ALP activity in milk from cows, sheep, and goats not only at the current European statutory level of
500 mU/liter but also at much lower levels.

Alkaline phosphatase (ALP), an enzyme widely dis-
tributed in nature and naturally present in raw milk, is used
as an indicator of proper milk pasteurization. The measure-
ment of ALP activity has been used to assess correct milk
pasteurization for more than 65 years (4). Initially, colori-
metric assays were developed to allow the dairy industry
to monitor milk quality (8). In 1990, a more sensitive and
precise fluorometric method for the quantitative measure-
ment of ALP in milk (Fluorophos Test System) was intro-
duced (5, 6). During the past 14 years, this method has been
widely investigated and extensively used and has received
international approval by governing bodies, including the
Association of Official Analytical Chemists (7), the Inter-
* Author for correspondence. Tel: 1101372 454257; Fax: 1101372
454257; E-mail: fharding@globalnet.co.uk.
national Dairy Federation (IDF) (1), the International Or-
ganization for Standardization (ISO) (3), and the National
Conference on Interstate Milk Shipments.
The European Union has stated its intention to lower
the statutory ALP level for pasteurized milk from 500 to
350 mU/liter and to set a maximum target value of 100
mU/liter. This has stimulated the need to subject the Fluo-
rophos Test System to international, interlaboratory evalu-
ation at lower ALP values and to widen the scope to include
milk from cows, goats, and sheep. This study was under-
taken to validate the test at lower ALP levels (20, 40, 100,
350, and 500 mU/liter) than previously reported. The eval-
uation covered six different types of milk tested in 13 lab-
oratories in Europe and America. The data will be of im-
portance for practical applications in routine use, to legis-

TABLE 1. Comparison of data from this trial with previous international trial data at 500 mU/liter of alkaline phosphatasea
Trial RSDr (%) RSDR (%) Repeatability (%) Reproducibility (%)

IDF (1992) 62 98
AOAC (1990) 4.4 7.00
FSA (2002) 4.4 9.20 57.6 146
Current trial cow’s milk 6.48 14.66 93.3 211.1
Current trial sheep’s milk 5.69 13.30 99.76 233.10
Current trial goat’s milk 1.74 5.33 28.56 87.51
a RDSr, relative percentage standard deviation of repeatability; RSDR, relative percentage standard deviation of reproducibility; IDF,
International Dairy Federation; AOAC, Association of Official Analytical Chemists; FSA, United Kingdom Food Standards Agency.
1048 HARDING AND GARRY J. Food Prot., Vol. 68, No. 5

TABLE 2. Comparison of repeatability and reproducibility values for interlaboratory trials evaluating Fluorophos across seven levels
of alkaline phosphatase activitya
Alkaline phosphatase (mU/liter)

Trial 20 40 100 200 350 500 900

FSA trial 2002 (%)

r 19 12 61 57 134
R 19 30 40 146 322
Pretrial 2003 (%)
r 4.79 4.77 10.84 29.89 37.73
R 6.77 16.64 28.08 80.03 95.70
Current trial 2004, cow (%)
r 21.5 22.1 89.6 93.3
R 31.8 51.0 136.4 211.1
Current trial 2004, sheep (%)
r 10.43 16.26 33.67 96.82 99.76
R 16.63 20.34 46.63 170.24 233.10
Current trial 2004, goat (%)
r 8.63 7.98 26.20 42.83 28.56
R 10.69 20.55 28.71 127.89 87.51
a FSA, United Kingdom Food Standards Agency; r, repeatability; R, reproducibility.

TABLE 3. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of whole cow’s milk samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 2 6 3 10 5 7 1 9 4 8

1 12.0 12.0 46.9 45.1 103.9 105.7 352.1 352.1 545.7 557.6
2 ,10a 13.3 41.4 46.0 107.1 109.4 363.2 361.3 565.0 560.4
3 17.0 12.4 44.1 50.6 108.9 107.1 354.9 367.3 589.8 583.4
4 12.0 14.3 54.7b 79.1b 101b 127.8b 425.2c 893.2b 553.0 684.5
5 15.6 12.0 49.2 39.1 96.1 98.4 367.3 342.0 532.3 512.6
6 12.0 12.0 57.5b 49.6b 148.5b 146.6b 376.5b 527.3b 747.0b 745.2b
7 32.6b 23b 18.4 44.1 125.0 106.7 405.5 359.0 556.2 576.9
8 16.1 12.4 58.8 30.3 118.1 108.0 382.5 351.7 868.4b 683.6b
9 ,10 13.3 44.1 44.1 101.6 103.9 360.4 333.7 552.6 554.9
10 11.0 11.0 44.6 39.1 98.4 96.5 354.0 352.1 544.7 544.7
11 12.4 ,10 34.9 43.2 90.1 84.1 299.7b 312.1b 469.8 498.8
12 ,10 ,10 34.5 32.2 81.8 78.1 337.0 321.8b 403.6 456.9
13 ,10 18.8 41.8 43.2 105.3 109.9 350.8 336.5 570.5 565.9
a Pair results were not used.
b Identified as stragglers by Cochran’s test for cell standard deviations or Grubbs’ test for one and two outlying cell means.
c Identified as outlying cell means by Cochran’s test for cell standard deviations for one and two outlying cell means.

TABLE 4. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using whole cow’s milk
samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20 13.0 2.10 16.20 6.06 2.104 16.20 6.06

40 44.5 10.01 22.51 28.83 11.317 25.44 32.59
100 106.5 7.23 6.79 20.83 16.990 15.96 48.93
350 350.8 16.15 4.60 46.50 23.144 6.60 66.65
500 577.8 48.38 8.37 139.33 99.911 17.29 287.75
a sr, repeatability standard deviation; RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.
J. Food Prot., Vol. 68, No. 5 FLUOROMETRIC METHOD FOR MEASURING ALKALINE PHOSPHATASE IN MILK 1049

TABLE 5. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of semiskimmed cow’s milk samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 12 16 13 20 15 17 11 19 14 18

1 12.4 ,10a 46.9 40.9 92.4 104.4 403.6 389.4 547.0 522.2
2 ,10 ,10 38.6 40.5 103.9 99.3 334.7 399.9 562.7 547.0
3 13.3 ,10 49.2 53.3 104.8 113.5 411.9 407.8 561.3 588.0
4 12.4 17.0 63.0 48.3 109.4 113.5 564.1 422.9 585.2 582.9
5 12.4 ,10 49.2 55.6 104.8 102.5 428.9 412.4 542.0 553.9
6 ,10 ,10 61.6 51.5 137.5 151.7 428.9b 613.7b 812.7c 556.2c
7 22.1 24.4 51.0 60.7 114.5b 145.7b 523.1 462.0 617.8 656.9
8 20.7 20.7 63.9 45.1 117.2 113.1 422.0 376.5 547.5 531.0
9 ,10 13.8 47.8 50.1 106.7 104.4 397.6 395.3 524.5 540.6
10 ,10 11.5 40.9 45.1 103.4 103.4 380.4 408.2 538.8 533.3
11 ,10 ,10 39.1 41.4 86.9 84.6 349.4 365.9 450.5 465.2
12 ,10 ,10 32.2 40.5 74.9 85.0 301.1 319.5 424.3 461.1
13 ,10 15.2 47.3 45.5 108.0 106.2 404.5 398.1 548.9 566.8
a Pair results were not used.
b Identified as stragglers by Cochran’s test for cell standard deviations or Grubb’s test for one and two outlying cell means.
c Identified as outlying cell means by Cochran’s test for cell standard deviations for one and two outlying cell means.

TABLE 6. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using semiskimmed cow’s
milk samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20 19.6 2.57 13.15 7.41 4.751 24.30 13.68

40 48.0 6.34 13.19 18.25 8.260 17.19 23.79
100 107.4 8.09 7.53 23.30 17.658 16.45 50.86
350 412.4 52.47 12.72 151.10 69.545 16.86 200.29
500 541.6 16.10 2.97 46.37 53.207 9.79 152.72
a sr, repeatability standard deviation; RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.

TABLE 7. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of skimmed cow’s milk samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 22 26 23 30 25 27 21 29 24 28

1 ,10a ,10 34.0 26.7 98.4 86.0 355.8 339.7 531.4 525.9
2 ,10 ,10 34.5 21.6 65.3c 141.1c 399.0 327.3 585.7 553.0
3 ,10 ,10 45.1 51.5 117.2 123.7 399.0 394.9 629.8 619.2
4 ,10 ,10 50.6 46.9 115.4 126.9 380.6 416.0 638.5 589.8
5 ,10 ,10 61.6 51.9 125.0 125.0 421.1 421.1 621.1 633.0
6 ,10 ,10 58.4 41.8 102.1 109.4 414.2 410.5 718.5 736.9
7 ,10 ,10 45.1 No result 160.9b 184.8b 817.3c 461.5c 723.1 612.8
8 ,10 ,10 42.3 46.0 103.0 104.8 308.9 308.9 557.2 465.2
9 ,10 ,10 53.8 47.8 97.5 119.5 372.4 362.2 541.1 579.2
10 ,10 ,10 46.0 44.1 114.0 114.0 397.2 388.0 553.5 564.5
11 ,10 ,10 36.3 30.3 117.2 129.2 347.5 367.8 549.8 519.5
12 ,10 ,10 40.0 34.5 97.5 107.1 315.4 359.5 476.2 483.6
13 12.9 ,10 69.0 67.6 150.8b 134.7b 436.7 412.4 600.4 630.7
a Pair results were not used.
b Identified as stragglers by Cochran’s test for cell standard deviations or Grubbs’ test for one and two outlying cell means.
c Identified as outlying cell means by Cochran’s test for cell standard deviations for one and two outlying cell means.
1050 HARDING AND GARRY J. Food Prot., Vol. 68, No. 5

TABLE 8. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using skimmed cow’s milk
samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20
40 45.1 5.90 13.08 16.99 12.271 27.21 35.34
100 118.1 9.38 7.94 27.01 22.447 19.01 64.65
350 397.5 21.13 5.32 60.85 38.989 9.81 112.29
500 586.1 34.25 5.84 98.65 72.525 12.37 208.87
a sr, repeatability standard deviation; RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.

TABLE 9. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of strawberry flavored cow’s milk
samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 32 36 33 40 35 37 31 39 34 38

1 ,10a ,10 51.0 45.1 102.1 104.4 433.0 433.0 641.7 600.8
2 ,10 ,10 48.3b 25.3b 101.6 103.9 438.1 389.8 620.1 617.8
3 ,10 ,10 64.8 71.3 115.8 104.8 438.6 450.0 654.6 630.2
4 12.9 ,10 58.4 60.2 116.3 125.0 454.6 420.2 644.0 614.2
5 ,10 ,10 58.4 58.4 116.8 116.8 443.6 474.0 653.2 663.3
6 ,10 ,10 67.1 63.0 111.7 126.0 461.5 502.5 630.7 630.7
7 ,10 ,10 59.8 49.2 115.4 113.5 459.7 446.8 707.4 650.0
8 ,10 ,10 60.2 68.5 116.3 122.3 521.3 481.8 697.8 735.1
9 ,10 ,10 54.2 49.6 111.2 113.5 432.6 430.3 605.4 595.8
10 ,10 ,10 60.2 51.9 116.3 105.7 483.6 442.2 610.5 606.3
11 ,10 ,10 44.6 31.7 97.5b 86.9b 370.1b 378.3b 556.2 522.2
12 ,10 ,10 34.5b 38.6b 79.5b 83.7b 338.3b 332.4b 493.7 491.4
13 ,10 ,10 56.1 63.9 98.8 104.8 428.9 442.2 609.6 579.2
a Pair results were not used.
b Identified as stragglers by Cochran’s test for cell standard deviations or Grubbs’ test for one and two outlying cell means.

TABLE 10. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using strawberry flavored
cow’s milk samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20
40 53.6 6.90 12.86 19.86 11.865 22.13 34.17
100 108.1 5.53 5.12 15.94 12.167 11.26 35.04
350 435.7 20.39 4.68 58.72 45.628 10.47 131.41
500 617.8 20.56 3.33 59.21 58.682 9.50 169.00
a sr, repeatability standard deviation, RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.

lators and regulatory bodies, and for approval by interna-
tional standardization agencies.
MATERIALS AND METHODS
Sample preparation. The following samples were heated to
958C for 5 min to produce ALP-free milk: pasteurized cow’s milk,
including whole, semiskimmed, skimmed, and strawberry fla-
vored; raw sheep’s milk; and raw goat’s milk. The fat content of
various cow’s milks are set by legislation in Europe. Whole cow’s
milk contains approximately 3.5% fat, semiskimmed cow’s milk
contains 1.5 to 1.8% fat, and skimmed cow’s milk contains
,0.3% fat. Samples were then cooled and spiked with varying
amounts of the relevant species raw milk to give samples with
ALP values close to 20, 40, 100, 350, and 500 mU/liter of ALP.
These values are of greatest commercial interest, 500 being the
current legal level, 350 and 100 being the values to be used by
the European Union, and 20 to 40 being the values most likely
to be observed in routine testing of properly pasteurized milk.
Bulk samples were carefully mixed and tested for ALP using the
Fluorophos Test System before being subsampled into tubes for
refrigeration. Random testing of subsamples confirmed homoge-
neity of ALP before sample dispatch.
J. Food Prot., Vol. 68, No. 5 FLUOROMETRIC METHOD FOR MEASURING ALKALINE PHOSPHATASE IN MILK 1051

TABLE 11. Summary of mean results for whole, semiskimmed,
skimmed, and strawberry flavored cow’s milk samples from 13
participantsa
Alkaline phosphatase (mU/liter)
,20 40 100 350
RSDr (%) 18.66 7.69 8.89
RSDR (%) 27.6 17.69 13.53
a RSDr, relative percentage standard deviation of repeatability;
RSDR, relative percentage standard deviation of reproducibility.
Sample transportation and laboratory participation. Sam-
ples were prepared on a Monday and distributed, using interna-
tional parcel distribution companies, to 13 laboratories, four in the
United States, four in the United Kingdom, and one each in
France, The Netherlands, Norway, Italy, and Switzerland. The
sample boxes used were designed to maintain temperatures below
48C during a 3-day period. Participating laboratories were asked
to record the temperature and time of sample receipt. All were
asked to test samples on Wednesday morning in an attempt to
balance any sample deterioration effect. The ISO/IDF draft fluo-
rometric method (ISO/CD 11816-1/IDF 155-1) was followed by
each laboratory. A pretrial run was completed to familiarize an-
alysts with the test method.
Data analysis. Samples were supplied as blind duplicates,
yielding 60 samples per laboratory. Single determinations were
made on the 60 samples (30 samples in blind duplicate). Results
were returned on report sheets provided. Data collected from the
collaborative trial were collated and analyzed following the sta-
500 tistical procedures defined in ISO 5725-2 (2). This comprised the
following steps. First, data from each laboratory were collated into
6.48 the appropriate pairs. Second, cell means and spread for each pair
14.66 of samples were calculated. Third, Cochran’s test was used to
identify outlying values for cell differences for each test material
type. Sample pairs identified as outliers were excluded from fur-
ther statistical analysis, and the test was applied iteratively to each
set of data until all outliers were identified. Fourth, for each test
material type, Grubbs’ test was used to identify outlying individ-
ual values for cell means. Sample pairs identified as outliers were
excluded from further statistical analysis, and the test was applied
iteratively to each set of data until all outliers were identified.
Where no further outliers were identified, Grubbs’ test for two
outlying values was applied, and pairs of results identified as out-
liers were excluded from further statistical analysis. Fifth, preci-
sion values for each type of test material were calculated.
RESULTS AND DISCUSSION
Sample matrix selection. Although it is more difficult
to transport and maintain the integrity of liquid milk sam-
ples than freeze-dried milk samples, liquid milk samples

TABLE 12. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of goat’s milk samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 42 46 43 50 45 47 41 49 44 48

1 25.3 25.3 37.7 40.0 113.5 137.0 410.5 382.9 559.9 542.9
2 19.8 19.8 46.0 43.7 120.0 120.0 408.7 423.8 557.2 543.8
3 22.5 22.5 46.9 44.6 114.0 131.9 402.7 422.9 566.4 566.4
4 27.6 25.7 51.5 51.5 124.1 137.0 432.1 412.8 570.9 562.2
5 24.8 24.8 53.8 49.6 133.8 137.9 449.6 422.5 597.6 597.6
6 23.4 17.0 60.2 57.9 137.5 131.0 446.8 429.8 614.2a 625.2a
7 23.4 18.8 58.4 49.2 135.6a 145.3a 479.9 491.4 704.7b 819.2b
8 14.7 24.8 39.5 41.4 132.9 122.3 396.7 386.1 535.6 565.0
9 21.6 23.9 47.8 45.1 121.4 123.7 405.0 414.6 566.8 557.2
10 22.5 22.5 49.2 43.2 135.6 121.4 422.0 390.7 590.7 580.1
11 19.3 19.3 45.5 43.2 112.6b 112.6b 381.1a 363.6a 519.5b 486.8b
12 17.9 12.0 31.7 34.0 89.6b 93.3b 296.5a 294.7a 428.0b 422.0b
13 21.1 23.9 49.2 51.0 137.5 125.5 421.5 391.7 572.8 566.8
a Pair results were not used.
b Identified as stragglers by Cochran’s test for cell standard deviations or Grubbs’ test for one and two outlying cell means.

TABLE 13. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using goat’s milk samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20 21.7 3.00 13.81 8.63 3.712 17.11 10.69

40 46.6 2.77 5.94 7.98 7.136 15.31 20.55
100 124.9 9.10 7.29 26.20 9.968 7.98 28.71
350 406.9 14.87 3.65 42.83 44.406 10.91 127.89
500 570.0 9.92 1.74 28.56 30.386 5.33 87.51
a sr, repeatability standard deviation; RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.
1052 HARDING AND GARRY J. Food Prot., Vol. 68, No. 5

TABLE 14. Five levels of alkaline phosphatase results from 13 participants using blind duplicates of sheep’s milk samples
Alkaline phosphatase (mU/liter) by duplicate no.

,20 40 100 350 500

Participant no. 52 56 53 60 55 57 51 59 54 58

1 35.4 25.7 41.4 43.2 112.2 104.4 407.8 405.9 551.6 596.7
2 26.2 28.5 41.8 52.9 109.9 101.1 388.4 425.7 594.8 599.4
3 27.1 31.7 45.5 51.9 113.1 126.9 479.9 441.8 593.5 631.6
4 33.1 37.2 49.6 49.6 118.1 114.0 438.1 461.1 658.1 599.9
5 34.5 34.5 52.9 48.7 115.8 126.0 499.7 483.1 722.6 700.6
6 37.2 30.3 50.1 65.3 111.7 152.6 518.5 412.8 712.1 740.6
7 41.4 39.1 54.2 52.4 117.7 141.6 572.8 494.6 660.1 747.5
8 34.0 30.3 42.3 38.2 106.2 92.4 383.4 447.7 600.4 638.5
9 33.1 30.8 47.3 45.1 104.4 111.2 426.6 419.2 580.6 597.2
10 31.7 29.9 41.4 47.3 104.8 103.0 421.5 419.2 587.5 642.7
11 24.8 33.6 35.4 51.9 96.1 91.9 331.4 375.6 498.3 456.9
12 15.6a 19.8a 35.4 35.4 92.4 74.5 331.9 325.9 516.7 436.3
13 31.7 29.0 47.8 51.9 103.9 101.1 406.4 405.0 575.5 579.7
a Pair results were not used.

were used because they better reflect the samples that will
be tested in routine practice by the method under investi-
gation. Consideration was given to the use of freeze-dried
samples, because this would have minimized sample dete-
rioration during transit.
Comparative data from current study to previous
fluorometric ALP studies in fluid dairy products. Table
1 compares the repeatability and reproducibility of the
Fluorophos Test System at 500 mU/liter for cow’s, sheep’s,
and goat’s milks with previous reported studies by the IDF,
the Association of Official Analytical Chemists, and the
United Kingdom Food Standards Agency (9) on cow’s
milk. Repeatability data generally agreed with earlier stud-
ies, whereas reproducibility values, with the exception of
goat’s milk, were somewhat higher, possibly due to changes
in ALP during extended sample transportation times. Table
2 compares the repeatability and reproducibility findings of
the current study over a wide range of ALP values for
cow’s, sheep’s, and goat’s milk with those reported on
cow’s milk from a pretrial international study and a study
undertaken by the United Kingdom Food Standards Agen-
cy. Repeatability and reproducibility values, although
somewhat higher than previous studies, were considered to
be suitable for the control of ALP in pasteurized milk.
ALP results from 13 participants using milk from
cows, goats, and sheep that cover a range of 20 to 500
mU/liter. The performance of the Fluorophos Test System
method was studied over a range of cow’s milk of com-
mercial importance (Tables 3 through 11). Tables 3, 5, 7,
and 9 detail individual participant test results on blind du-
plicate samples. Tables 4, 6, 8, and 10 present calculated
repeatability and reproducibility values for each category
of cow’s milk under study, and Table 11 summarizes the
repeatability and reproducibility values for all cow’s milk
samples tested. Tables 12 and 14 report individual partici-
pant test results on blind duplicate samples for goat’s and
sheep’s milk, and Tables 13 and 15 summarize the repeat-
ability and reproducibility values for goat’s and sheep’s
milk.
Participating laboratories 4 and 6 had a number of re-
sults that were high in ALP. There was no evidence that
they had incorrectly calibrated their test instrument; hence,
it was considered possible that sample deterioration may
have contributed to the elevated ALP values. Also, a poor
pipetting technique might have contributed to these higher
than expected results. Where appropriate, statistical outliers
were eliminated from calculated values.
Little difference existed between the relative standard

TABLE 15. Repeatability and reproducibility values for each of the five levels of alkaline phosphatase results using sheep’s milk
samplesa
Enzyme Mean sr RSDr Repeatability sR RSDR Reproducibility
(mU/liter) (%) (%) (%) (%) (%) (%) (%)

,20 31.0 3.62 11.68 10.43 5.775 18.62 16.63

40 46.9 5.65 12.04 16.26 7.063 15.07 20.34
100 109.5 11.69 10.68 33.67 16.191 14.79 46.63
350 427.8 33.62 7.86 96.82 59.112 13.82 170.24
500 608.4 34.64 5.69 99.76 80.936 13.30 233.10
a sr, repeatability standard deviation; RSDr, relative percentage standard deviation of repeatability; sR, reproducibility standard deviation;
RSDR, relative percentage standard deviation of reproducibility.
J. Food Prot., Vol. 68, No. 5 FLUOROMETRIC METHOD FOR MEASURING ALKALINE PHOSPHATASE IN MILK
deviation of repeatability (RSDr) and the relative standard
deviation of reproducibility (RSDR) values for cow’s milk
types, except that the value for whole cow’s milk at 500
mU/liter gave a high reproducibility value. This was due to
the fact that some results used fell just inside the outlier
values. The RSDR for whole milk at 500 mU/liter was
17.29% (Table 4) compared with 7.00% (Table 1) reported
by the Association of Official Analytical Chemists collab-
orative study from 1990. Statistical stragglers from labo-
ratories 6 and 11 contributed to this high value, because
they were included in calculations and significantly affected
the reproducibility value due to the small sample popula-
tion. Although the Fluorophos Test System validly recorded
one or both results as ,10 mU/liter, no statistical procedure
exists for dealing with such results.
Based on this collaborative trial, the fluorometric meth-
od has been submitted to the ISO and IDF for approval as
a final international standard method. This trial demon-
strates the applicability of the method for routine control
and legislative surveillance. Data generated in this study
were generally comparable to data from previous studies,
with the exception of reproducibility data at 500 mU/liter
for whole milk. This international trial provides results in-
dicating that the Fluorophos Test System method is a sat-
isfactory control test for detecting ALP activity in pasteur-
ized milk from cows, sheep, and goats down to 20 mU/
liter.
ACKNOWLEDGMENTS
We gratefully acknowledge Quadra Chem Laboratories (Forrest
Row, East Sussex, UK) for sample preparation and distribution, L. Cov-
eney (Savant Technologies, Powys, UK) for expertly undertaking all of
the statistical analyses, and Linda Snyder for her technical assistance in
completing the manuscript for submission. We thank Advanced Instru-
ments, Inc., for supplying all reagents and consumables for this study. We
also thank the following for participating in this collaborative international
1053
interlaboratory study: Marina Nicolas (Agence Francaise de Sanitaire Des
Aliments, Paris, France), Tony Woods and Mark Whatton (Quadra Chem
Laboratories), Frea Woltjes (Centrale Organisatie voor de Kwaliteitsaan-
gelegenheden in de Zuivel, Leusden, The Netherlands), Jaques Meyer
(Eidg. Forschungsanstalt fur Milchwirtschaft, Liebefeld, Switzerland), Lu-
isa Pellegrino (Department of Food Science and Technology, Milan, Italy),
Kevin Williamson (Lancashire NHS Trust, Lancashire, UK), Melody
Greenwood (Health Protection Agency, South East Laboratory, South-
ampton, UK), Mary Bulthaus (Dairy Quality Control Institute, Mounds
View, Minn.), Robert Bradley (Wisconsin Centre for Dairy Research,
Madison), Doug Laker (Express Dairy Company, East Sussex, UK), Karl
Eckner (The Norwegian Institute for Food & Environmental Analysis,
Oslo), and Kurt Mangione (New York State Department of Agriculture,
Albany).
REFERENCES
1. International Dairy Federation (IDF). July 1992. Standard 155:1992:
milk and milk-based drinks—determination of alkaline phosphatase
activity fluorometric method. IDF, Brussels.
2. International Organization for Standardization (ISO). 1994. Inter-
national standard ISO 5725-2: accuracy (trueness and precision) of
measurement methods and results—part 2: basic method for the de-
termination of repeatability and reproducibility of a standard mea-
surement. ISO, Geneva.
3. International Organization for Standardization (ISO). 1997. Inter-
national standard ISO 11816-1: milk and milk products—determi-
nation of alkaline phosphatase activity using a fluorometric meth-
od—part 1: milk and milk-based drinks. ISO, Geneva.
4. Kay, H. D., and W. R. Graham. 1935. The phosphatase test for pas-
teurised milk. J. Dairy Res. 6:191–203.
5. Marshall, R. T. (ed.). 1992. Standard methods for the examination
of dairy products. 16th ed. American Public Health Association,
Washington, D.C.
6. Rocco, R. M. 1990. Fluorometric analysis of alkaline phosphatase
in fluid dairy products. J. Food Prot. 53:588–591.
7. Rocco, R. M. 1990. Fluorometric determination of alkaline phos-
phatase in fluid dairy products, collaborative study. J. Assoc. Off.
Anal. Chem. 73:842–849.
8. Scharer, H. 1938. A rapid phosphomonoesterase test for control of
dairy pasteurization. J. Dairy Sci. 21:21–34.
9. Wood, R., A. Damant, and F. Harding. 2003. Method to estimate
alkaline phosphatase in pasteurised milk. Submitted for publication.