Ann. din. Biochem.

13 (1976) 327-335

Quality Control of Serum Alkaline Phosphatase Assays:

Project Report and Discussion of Some Factors Affecting the Assay

G. ELLIS
Department of Clinical Chemistry, General Hospital, Steelhouse Lane, Birmingham, B4 6NH

The factors which affect the standardisation and quality control of serum alkaline phosphatase
assays are discussed. A quality control project was designed to test the performance of seven
Birmingham laboratories in the assay of sera with alkaline phosphatase activities outside the
range normally tested by the National Quality Control and Wellcome schemes. The results showed
that the precision of individual laboratories was satisfactory. Differences between the results of the
laboratories were considerable and could be accounted for by differences in methodology. Auto-
Analyzer methods employing phenyl phosphate as substrate would best be standardised by adopt-
ing the optimised reaction conditions of Buch and Buch (1939); but the borate buffer of these
authors should be replaced by the carbonate-bicarbonate buffer of Moss et al. (1971). To avoid the
confusion which may arise in future if alkaline phosphatases are reported in U/1 irrespective of the
substrate, it is suggested that some substrate-indicative nomenclature may be advisable.

Purified alkaline phosphatases (orthophosphoric universally agreed method of standardisation of

monoester phosphohydrolase E.C. 3.1.3.1) from phosphatase activity even when essentially the same
many sources exhibit three types of activity: methodological conditions are used. This is par-
ticularly true of assays adapted for AutoAnalyzer
use and will be illustrated by reference to methods
phosphatase employing phenyl phosphate. Standardisation may
ROP + H.OH ---.. ROH + PI (1) utilise either (a) phenol standards in conjunction
phosphotransferase with the measured time of incubation in the Auto-
ROP + R'.OH ---.. ROH + R'OP(2)
pyrophosphatase Analyzer, or (b) "Serum" standards. These may have
PPI + HIO - - - - - - - - - . . 2PI (3) values assigned by a manual technique as recommen-
ded by Moss et al. (1971), or they may be secondary
standards the values of which have been assigned by
Clearly, pyrophosphatase activity will proceed only method (a) above. "Serum" standards are difficult
in the presence of inorganic pyrophosphate, and this to prepare with suitable high activity. Commercial
condition will not obtain in a phosphatase assay sys- sera may be enriched with animal alkaline phospha-
tem. However, the inclusion of an organic phospho- tases or human isoenzymes (e.g., placental), which
monoester in the reaction mixture provides substrate have different properties from the human bone or
for both phosphatase and phosphotransferase liver isoenzymes most commonly encountered in the
activities, and the relative importance of each general hospital laboratory. Additional difficulties
activity is determined largely by the availability of may also arise when lyophilised quality control
suitable alcohol groups (R'.OH, e.g., Tris) to par- serum is employed, since this may considerably
ticipate in phosphotransferase activity. This means increase in activity over the period following recon-
that a particular isoenzyme of alkaline phosphatase stitution (Brojer and Moss, 1971; Smith and Fogg,
may exhibit widely different activity when acting on 1972).
the same substrate at the same concentration in dif- The present study will illustrate the difficulties
ferent buffers if those buffers differ in their ability to encountered in alkaline phosphatase standardisation
act as phosphotransferase acceptors (McComb and in seven laboratories in the Birmingham area. All
Bowers, 1972). hospitals regularly analysed serum from the National
One reason why alkaline phosphatase is difficult to Quality Control scheme. Their mean variance index
standardise is that over the past 45 years so many for alkaline phosphatase in 1974 was 33.0 compared
methods have been published for its measurement, with a mean variance index for all laboratories par-
yet none has been internationally accepted as the ticipating in that scheme of 78. Three laboratories
reference method. The methods differ not only in the participated in the Wellcorne Quality Control scheme.
diversity of their substrates but also in the variability Their mean position in the "league table" was 189 out
of the conditions of assay-e.g., in the use of buffers of 485laboratories. This project sought to measure the
which permit transphosphorylation, or in the inelu- performance of the laboratories with sera outside the
sion of metal activators (usually MgH). There is no range of activity normally tested by these schemes.
327
328 G. Ellis

MATERIALS obtained from a fresh cadaver. Alkaline phosphatase

Fresh human sera
At fortnightly intervals over a period of several
months sera with raised alkaline phosphatase
activities were pooled and aliquots were distributed
to the seven laboratories. The laboratories were asked
to store the sera at 4°C and analyse them the follow-
ing day,
Commercial control sera
The following sera were employed in this study:
Horse Serum Number 3 and Wellcomtrol QCP from
Wellcome Reagents Ltd., Beckenham, Kent;
Versatol-A and Validate-A Abnormal Human
Serum Control from the General Diagnostics Divi-
sion, Warner Lambert Company, Eastleigh, Hamp-
shire; and Q-Pak Chemistry Control Serum II from
Hyland-Travenol Laboratories Inc., Costa Mesa,
California, U.S.A. These sera were reconstituted
with water in accordance with the manufacturers'
instructions. Samples were pooled then aliquoted
and distributed. They were analysed after overnight
storage at 4°C.
Sera with butanol extract of human liver
A specimen of histologically normal liver was
was prepared by aqueous homogenisation and
butanol extraction as described by Morton (1950).
The alkaline phosphatase was concentrated by
addition of Lyphogel (Gelman Instrument Co.,
from Hawksley & Sons, Lancing, Sussex) to the
aqueous material. This preparation was then added
to Wellcome Horse Serum No.3 which had been
heat-inactivated at 56DC from 1 hour to give a series
of 8 pools of activity 20 to 50 King-Armstrong units/
100m1 (determined by the manual reference assay;
Moss et al., 1971). Serum was tested for Australia
antigen and was found to be negative by complement
fixation. The serum was divided into aliquots distri-
buted to the laboratories and stored at - 20°C. It
was thawed on the day required. Laboratories were
asked to analyse one or two specimens per week
according to a proforma. Dilution, where required,
was 1 in 2 vl» with 0.15 mmol/l saline.
No laboratory knew how many pools were being
analysed, and the protocol randomised the pools so
that cheating or deductive or intuitive modification of
results before reporting to the author were not
possible.
METHODS
The methods employed in the seven laboratories
are shown in Table I. All laboratories assayed at

Table 1. Alkaline phosphatase methods used in the quadrant"

Substrate Activator Technique Standardisation Hospital

(Final concentration mmolfl)

{Children'S
Phenyl Phosphate (4.75) - Manual Phenol standards + time Accident
Eye
Phenyl Phosphate (6.75) - SMA 12/60 Versatol abnormal high Queen
(assayedmanually) Elizabeth
(SMA
12(60)
Phenyl Phosphate (6.75) MgO+ (l mmol/l) Auto- Hyland QPII (assayedon General
Analyzer AutoAnalyzer against
phenol + time)
Phenolphthalein Phosphate (2.3) - Manual Warner Validate Maternity
(manufacturer's stated
value)
p-Nitrophenyl Phosphate Mg2+ (8 mmol/l) Manual Rate of optical density Queen
mannitol change standardarised Elizabeth
against human serum (Emergency
method
used for
urgent
specimens
only)
.For the purposes of quality control and education the Midland Region of the Association of Clinical Biochemists is
divided into four quadrants.
 Quality control of serum alkaline phosphatase assays 329

Table 2. Results obtained from the assay offresh human serum and quality control preparations

Serum Children's Accident Eye MaternityQueen Queen General

Elizabeth Elizabeth
(SMA 12/6)) (p-nitrophenyI
phosphate)
------ --------------
Human 1 - 17 18 23 20 21 22
2 - 27 27 26 32 33 39
3 - 23 22 22 31 33 41
4 - 24 24 37 29 32 30
5 - 21 18 37 27 31 31
6 - 21 21 41 30 34 34
7 - 23 26 39 29 30 31
8 - 20 21 34 29 29 32
9 - 55 57 102 'i2 86 102
10 24 26 26 29 30 3\ 37
11 27 31 28 32 30 32 39
12 56 54 51 68 65 68 90
I3 25 25 25 27 29 29 34
Wellcorne Horse 3 13 15 17 - 15 18 16
Wellcome ACP 15 12 12 - 12 13 14
Warner Validate A 22 20 23 - 22 21 21
Warner Validate A 19 18 19 20 19 19 23
Hyland Q-Pak 30 29 30 - 26 20 27
Wyland Q-Pak 30 29 32 29 27 23 32
Versatol-E 38 40 40 67 45 38 46

37°C and reported their results in "King-Armstrong
units/IOOml".
The manual method of Belfield and Goldberg
(1971) was performed as published. Although not
described by the original authors, the presence of
serum constituents decreased the stability of the
final coloured solutions. For this reason all solutions
were read within 15 minutes. Colour instability was
particularly notable with certain batches of 4-
aminophenazone, and such batches were avoided.
Phenyl phosphate concentrations in this paper
were calculated on a molecular weight of 254
(CuH5N a 2P04.2H20).
RESULTS
Phenol standards
A phenol standard was analysed by those labora-
tories employing phenyl phosphate as substrate. This
was found to agree with the standard employed in
each laboratory.
Assay of fresh human serum and commercial quality
control preparations
In view of the large number of methodological
variations between the various participating labora-
tories an initial unsystematic exchange of fresh
human sera and reconstituted quality control pre-
parations was undertaken as described above and
the results are shown in Table 2.
The laboratories employing phenyl phosphate
with AutoAnalyzer methodology produced higher
results than manual laboratories using phenyl phos-
phate. The difference is clearly seen Fig. 1. This
trend was not evident with commercial quality con-
trol serum (Fig. 2). Clearly, there is some danger in
employing such material as calibration "standards"
for AutoAnalyzer methods if the "standard value" is
derived under some alternative methodological
conditions, e.g., a manual assay at some other sub-
strate concentration; in the presence or absence of
activator, e.g., M g 2 + and at some other serum
dilution (see later).
Assay of horse serum enriched with human liver
alkaline phosphatase
All laboratories reported their results in King-
Armstrong units/IOO ml, but since three of these
were using the manual method-the "reference"
method of Moss et al. (l971)-the mean value of the
three laboratories was regarded as the reference
value-the "manual mean" result. The mean value
of two of these laboratories was employed if the
third was considered to be in error. The results
determined by the various methods were plotted in
relation to the actual activity of the sample deter-
mined by the reference technique (Figs. 3 to 6).
330 G. Ellis

HUMAN S E RU M
ACTIVITY COMPARISON
(

AA
Result
Mat ern it y

.,,
5

.. ,

"
,., ,r:.
. .
. •
5

... .
,•
,.' I
••
•

••
• • ,~
,
..
,'

"~' •a E Manual Mean

50
"

-.
'

.......---....-_-~~-
Manual Mean
Fig. 3.-The relation of the activities obtained at the
Maternity Hospital to the absolute manual mean values•

A CT I V I T y e O M PAR IS 0 N

OE (PN~~)
(Mg )
AA COMMERCIAL. SERUM,.
Result

50 ...
50 , ,, ..
:'t

.,
•
,vers
, • GHB
./

,Ao •a E
...
• I' q.2
I,~al
I
, h Manua I Mean
, w
Manual Mean
50
50

Fig. 2.-As Fig. 1 but using the following commercial sera:

Wellcontrol QCP (w), Horse Serum No.3 (h), Validate A Fig. 4.-As for Fig. 3 but sbowing the Queen Elizabeth
(val), Q-Pak II (q2), and Versatol-A (vers), Hospital (emergency laboratory) results.
 Quality control of serum alkaline phosphatase assays 331

ACTIVITY COMPARISON
ACT I V IT Y COM PA R I SON

GHB
aE( SMA 12/60)

..
... -- -
I., --
50
..
...- -'" ~
5
.- .
-
0(:' I.

....-. .. ,
...... -I .. ,
..- -
.
.... ,
•'
.,.'
,,

.: Manual Mean Manual Mean

50 50

Fig. 5.-As for Fig. 3 but showing the Queen Elizabeth Fig. 6.-As for Fig. 3 but showing the General Hospital
Hospital (SMA 12/60) results. results.

Precision of the individual laboratories Effect of various factors on alkaline phosphatase

Eight serum pools have been analysed in this study
over 23 weeks.
Pools were analysed between 3 and 5 times during
the study, and precision was derived by taking 8
pairs of results. Each pair represented the value
obtained for a given pool. The two results analysed
over the shortest time interval were selected to mini-
mise any disparity between pairs due to deterioration
of the specimens. Precision data are shown in Table 3.
activity with phenyl phosphate as substrate
A large number of methods have been proposed
for measuring alkaline phosphatase with phenyl
phosphate as substrate. The rationale behind the
modifications to the original King-Armstrong con-
ditions are often not stated. For this reason the effect
of M g 2 + and phenyl phosphate concentrations and
also the effect of buffer pH were briefly investigated
to determine optimum assay conditions at 37°C.

Table 3. Precision 0/ alkaline phosphatase determination based on 16 results

Hospital Mean value % of Standard deviation Coefficientof

(King-Armstrong manual mean (paired duplicates) variation
units/100 ml) co
Accident 33.6 100.9 2.7 7.9
Children's 33.1 99.3 4.6 13.8
Eye 33.2 99.6 1.5 4.5
Maternity 31.6 94.9 2.4 7.4
Queen Elizabeth
(SMA 12/60) 35.8 107.5 2.1 S.8
Queen Elizabeth
(Emergency) 40.9 122.8 2.5 6.1
General 43.5 130.6 2.3 S.2
332 G. Ellis

Table 4. The effect of variation of the serum dilution on the magnesium concentration in four alkaline phosphatase
methods

Method ml Serum ml Substrate Total volume Serum Final MgH concen-

dilution tration (if serum is
0.83 mmol/l)

Belfield and Goldberg

(1971) (manual) 0.05 2.0 2.05 1:41 0.020

Mossetal, (1971)
(manual) 0.1 2.0 2.1 1 :21 0.040

Axelsson et al, (1966) 0.32 2.5 2.82 1: 8.8 0.095

(AutoAnalyzer) rnl/min (no added Mg)

Price and Woodman (1971) 0.32 1.6 1.92 I: 6.0 0.139

(AutoAnalyzer) ml/min

(a) Significance of serum di/ution.-Since serum bicarbonate buffers over the range of pH 8.9 to 10.2
contains Mg2+ (reference values 0.70--0.95 mmol/l), (measured at 37"C) at 10 mmol/I phenyl phosphate
the final M g 2+ concentration of the reaction mixture concentration in the presence of 1 rnmol/l magnesium.
of an alkaline phosphatase assay system is dependent The pH optimum was 9.9. This buffer may be pre-
not only on Mg2+ incorporated in the substrate mix- pared as described by Moss et al. (197]).
ture but also on the dilution of the serum in the
2
reaction mixture. Table 4 shows that two Auto- ACTIVATION BY Mg +
Analyzer methods and two manual methods differ
in respect of the dilution of serum employed and that
this affects the Mgz+ concentration in the assay. o
(b) Effect of MgH concentration.-The effect of
7a ofAclivity in I mmol/l Mg 2 +

Mg2+ concentration on four sera was assessed using

the method of Belfield and Goldberg (1971). Mg2+
was added to the substrate mixture to give final con-
centration of 0.01-1.0 mmoljl. (Activation is not
o ••
increased by exceeding 1.0 mmoljl.) The results are
shown in Fig. 7.
(c) Effect of phenyl phosphate concentration in the
presence of MgH.- The effects of substrate concen-
tration and magnesium were investigated. Ten pooled
sera (mean activity 30.8, range 15.2 to 54.5 King-
Armstrong units/100 ml) were assayed using the
method of Belfield and Goldberg (1971) at 4 sub-
strate concentrations in the presence and absence of
magnesium (l mmol/I), The activities were expressed
as the % age of their activity at 18 mmoljl phenyl
phosphate concentration (the highest concentration Mg 2 +
tested) in the presence of magnesium (Fig. 8). There ... - ....,... ....__
m_mOI/ I
was only slightly greater activity at 18 rnmol/l, rec-
ommended by Fishman and Ghosh (1967) and Henry 0·5 1·0
et al. (1974), than at 10 mmol/l, recommended by Fig. 7.-The activation of serum alkaline pbosphataK by
BuchandBuch(l939) and Price and Woodman (1971). magnesium: 0, e, _, and'" represent four sera. The
These results would support an optimum substrate activities are expressed as the percentage of the activity
concentration of 10 mmo!jl in the presence of 1 obtained at 1 mmoljl MgH concentration. Arrows shows
rnmol/l magnesium. the MgH concentration calculated from the serum dilution
(Table 4) for three methods: M = Moss et af. (1971), A =
(d) Effect of buffer pH.-The pH activity curves Axelsson et al. (1966), P & W = Price and Woodman
were.drawn for 3 serum pools in 50 mmol/I carbonate- (1971).
 Quality control of serum alkaline phosphatase assays
DISCUSSION
Some practical considerations on the protocol of this
study
(a) The use of fresh human serum.-This material
would seem to have two main advantages over
alternative quality control materials. It has similar
isoenzyme composition to the material being ana-
lysed, and its use eliminates time-dependent errors
whichmight be introduced associated with the recon-
stitution of lyophilsed material. The disadvantages
in its use are that the selective pooling of sera known
to have high values is time-consuming. Requesting a
laboratory to analyse fresh specimens within 24 hours
imposes constraints on that laboratory. Preparation
and delivery of suitable serum pools tends to be
sporadic and consequently quality control is
sporadic. The delivery of 4--8 additional specimens
may represent a negligible increase in work for an
automated laboratory but the increase may be size-
able for a manual laboratory. Unless Australia anti-
gen testing is performed in the laboratory from which
the sera are distributed it is unlikely that routine
testing of all fresh serum pools for this antigen may
be practicable, and consequently pooling of sera
from patients with liver disease is associated with an
increased hepatitis risk,
(b) The use of liver preparation added to horse
serum.-This material has been employed in a pre-
vious study by Teasdale (1973), and is recommended
as a potential "serum standard" for AutoAnalyzer
calibration by Moss et al. (1971). Teasdale (1973)
observed a deterioration of 1 % per week, but in the
present study no clear evidence of deterioration of
this material was evident from examination of the
results. However, towards the end of the 23-week
period the author was receiving a number of reports
from the participating laboratories of specimens
whichwere turbid and required centrifugation before
analysis, and it is suggested that material of this type
should not be stored much longer than the period
employed in this study.
Performance of the various laboratories
The precision of individual laboratories was satis-
factory (Table 3). These data may be compared with
those of Teasdale (1973). In that study (see that
author's Table 6) the nine manual laboratories
showed a mean coefficient of variation of 16.5 %
(range 6.2 % to 29.9 %. level = 15.3 King-Armstrong
unitsflOO ml), while the eight automated laboratories
had a mean coefficient of variation of 9.9 % (range
2.2% to 18.9%).
While the above data indicate that laboratories
333
are consistent in the values which they report, even
cursory examination of Figs. 3 to 6 shows that there
are gross differences in the activities reported by the
individual laboratories. Manual laboratories employ-
ing the reference manual technique agreed well
(Table 3). The Maternity Hospital laboratory agreed
remarkably well with the reference manual labora-
tories considering (a) that phenolphthalein phosphate
was substrate and (b) that the method was calibrated
on the stated value of commercial control serum.
Differences between the Queen Elizabeth Hospital
emergency technique employing p-nitrophenyl phos-
phate, M g2+, and mannitol as transphosphorylation
acceptor were hardly unexpected. Differences
between the reference manual technique and the two
AutnAnalyzer methods were more surprising, since
phenyl phosphate is the substrate in all these methods
and the reasons for these discrepancies will be dis-
cussed more fully.
Possible reasons for discrepancies between alkaline
phosphatase activities measured on the AutoAnalyzer
and by manual techniques
The reaction mixtures employed by the manual
laboratories differed from those of the two auto-
mated laboratories. Of particular significance is the
inclusion of Mg2+ in that used at the General Hospi-
tal, since this ion is a known activator of alkaline
phosphatase. The present study has shown that diffe-
rences exist between the results obtained on a SMA
12/60 Analyzer at the Queen Elizabeth Hospital and
100
,
!f0
I
,/ ,
,
,
I :
I
r ,
I'
I
,.,
10
Fig. 8.-The effect of substrate concentration on serum
alkaline phosphatase activity in the presence • and absence
• of 1 mmol/I MgH. Symbols indicate mean activity of 10
serum pools. Bars give the range of values obtained.
334 G. Ellis
manual techniques-s-even in the absence of added
Mg 2 + . The substrate concentration (6.75 mrnol/l)
employed on the SMA 12/60 is higher than that of
the reference manual technique (4.75 mmol/l), but
experiments reported above and also work performed
independently by Culank and Whitehead (974) sug-
gest that the increased substrate concentration and
reduced incubation time alone may not be sufficient
to account for the differences between manual and
AutoAnalyzer results, Previous authors have also
found differences between manual and automated
procedures-though the differences between the pro-
cedures may, on occasion, have been the reverse of
those just described (manual method giving higher
values). Fishman and Ghosh (1967) considered that
"Disparate findings between AutoAnalyzer and
manual data on the same specimen may be explained
by the presence of heat sensitive enzymes". In con-
trast Horn (1972) found that, after elimination of
phenol absorption on to the walls of the manifold,
the results obtained on the AutoAnalyzer (which
previously had required correction by factors of 0.67
to 0.90 to compare with manual results) then became
identical with those obtained by the manual method.
However, an additional factor must be the serum
dilution. Table 4 shows that two AutoAnalyzer
methods and two manual methods differ in respect of
the dilution of serum employed in their reaction
mixtures and that this affects the Mg 2 + concentration
of their reaction mixtures. It would appear that
increasing the amount of serum effectively increases
the Mg 2 + concentration, and from Fig. 7 this would
be expected to produce a slightly increased alkaline
phosphatase activity. Similarly, dilution of serum
with 0.15 M NaCl on an AutoAnalyzer system would
be expected (after correction for dilution) to a yield a
slightly low result relative to the "true" activity in
that system.
Applicability of the standard reference assay of Moss
et al. (1971) to the automated laboratories
Moss et al. (1971) recommended that automated
systems measuring human serum alkaline phospha-
tase should be calibrated against reference serum
containing human alkaline phosphatase. These
authors indicated that such serum should be assayed
by means of the reference manual assay.
Alkaline phosphatase is measured at the Queen
Elizabeth Hospital on a Technicon SMA 12/60.
Commercial calibration serum is employed, and the
machine automatically calibrates on set values every
10 samples. Clearly the method of Moss et al. would
not be applicable in this scheme unless the recom-
mended human alkaline phosphatase control serum
were run every ten samples additionally to the cali-
bration standard and the alkaline phosphatase
results subsequently recalculated.
At the General Hospital alkaline phosphatase is
measured on an AutoAnalyzer AAI type channel as
part of a scheme of grouped analysis (Gaddie et al.,
1967). One technician performs five assays (including
alkaline phosphatase) simultaneously from a com-
mon sample plate containing commercial calibration
standards and unknown sera. The standard curve (7
points) is repeated every 40 samples and drifts run
every 10 samples. The recommended procedures of
Moss et al. would require an additional "human
alkaline phosphatase" drift sample every tenth
specimen and a second set of "human alkaline
phosphatase" calibration standards every 40 speci-
mens, seriously reducing throughput. Alternatively,
alkaline phosphatase would require running as a
separate. channel with inherent loss of the efficiency
gained in grouping analyses.
Optimum conditions for the assay of alkaline phos-
phatase in serum using phenyl phosphate as substrate
The manual assay of Moss et al. (1971) gives
satisfactory precision in routine laboratory use
(Table 3) and may be universally employed. How-
ever, those authors' proposals for automated assays
would seem applicable only to laboratories which
have the necessary expertise to prepare human tissue
alkaline phosphatases and which employ single chan-
nel AutoAnalyzers dedicated to alkaline phosphatase.
Many automated laboratories consider commercial
quality control serum as the most satisfactory form
of standardisation. Figs. 1 and 2 show that such sera
do not behave as human serum does when reaction
conditions are changed-s-e.g., commercial serum q2
shows slightly lower activity under the assay con-
ditions of the General Hospital than under manual
reference conditions (Fig. 2), whereas human serum
has activity 20 to 40 % higher under General
Hospital conditions (Fig. 1). For this reason it is
suggested that commercial sera are calibrated using
phenol standards (observing the precautions of Horn,
1972) and timing the period of analysis from the
entry of the sample into the reagent stream to a point
at the centre of the path through the dialyser,
Since AutoAnalyzer systems can never precisely
simulate manual King-Armstrong conditions (e.g.,
serum dilution, incubation time), there seems little
point in operating them under suboptimal conditions.
Optimum conditions may be defined on the follow-
ing considerations.
Since it is desired to measure alkaline phosphatase
activity and not alkaline phosphotransferase activity, a
buffer-e.g., 50 mmolJl carbonate-bicarbonate, pH 9.9
at 37°C-which does not act as phosphotransferase
acceptor should be employed.
 Quality control of serum alkaline phosphatase assays 335
In the absence of added Mgl+-, variations in the
dilution of the serum in the reaction mixture produce
changes in the MgH concentration of the reaction
mixture sufficient to slightly alter the measured activity.
This can be counteracted if the recommendations of
Buch and Buch (1939) and Baron (1965) are followed
and MgH (1 mrnolfl) is included as activator.
Three groups of authors (Buch and Buch (1939),
Price and Woodman (1972), and Henry et al. (1974»
considered that the phenyl phosphate concentration
should be increased to 10 rnmol/l. This is not as high
as the 18 mmol/l recommended by Fishman and Ghosh
(1967), but would seem optimal in the presence of mag-
nesium (Fig. 8).
Since the assay conditions are no longer King-
Armstrong conditions, it is recommended that the
results be reported in V/1. In order to avoid the con-
fusion which may occur if results quoted in U/l are
obtained with alternative substrates, it is suggested
that a suitable substrate indicative nomenclature be
adopted, e.g., alkaline phthymolphthalein phos-
phatase, alkaline phenyl phosphatase, alkaline p-
nitrophenyl phosphatase.
The author wishes to thank the following heads of
departments for helpful discussion and for permission to
publish results obtained in their laboratories: Dr. A. M.
Bold (Queen Elizabeth Hospital), Mrs. J. M. Crawley
(Maternity Hospital), Miss S. Littlejohn (Accident Hos-
pital), Dr. J. B. Marsters (Eye Hospital), Dr. B. E.
Northam (General Hospital), Dr. D. N. Raine (Children's
Hospital).
Dr. L. S. Culank kindly loaned a copy of his M.Sc.
thesis and gave permission to refer to some of his data.
The author thanks the following members of the Quad-
rant Quality Control committee for many valuable sug-
gestions: Dr. E. C. Albutt, Mr. R. Beetharn, Mr. G. O.
Evans, Mrs. A. Green, Mr. P. W. Lewis, Miss L. O.
Morris, Mr. G. R. Shuttleworth, and Dr. H. G. Worth.
REFERENCES
Axelsson, H., Ekman, B., Knutsson, D., Determination
of SGOT and SGPT and alkaline phosphatase with a
simplified AutoAnalyzer technique not requiring blank
runs. In Automation in Analytical Chemistry, p. 603.
New York, Mediad Inc. (1966).
Baron, D. M., Magnesium and serum alkaline phos-
phatase. Proc. Ass. din. Biochemists 3 (1965) 215.
Belfield, A., Goldberg, D. M. Revised assay for serum
phenyl phosphatase activity using 4 amino-antipyrine.
Enzyme 12 (1971) 561.
Brojer, B., Moss, D. W. Changes in the alkaline phos-
phatase activity of serum samples after thawing and
after reconstitution from the Iyophilised state. Clin.
chim. Acta 3S (1971) 511.
Buch, 1., Buch, H. An improved King and Armstrong
method for the determination of phosphatase activity
in blood serum. Acta. med. Scand. 101 (1939) 213.
Culank, L. S., Whitehead, T. P. Serum alkaline phos-
phatase activities in adolescent boys. Clin, Sci. mol.
Med. 47 (1974) 22P.
Fishman, W. H., Ghosh, N. K. Isoenzyrnes of human
alkaline phosphatase. Advan. din. Chern. 10 (1967)
255.
Gaddie, R., Northam, B. E. and Roberts, L. B., The
development of a system for multiple clinical bio-
chemical analysis. In Automation in Analytical Chemi-
stry, vol. 2, p. 63. New York, Mediad Inc. (1967).
Henry, R. J., Cannon, D. C., Winkelman, J. W. Clinical
Chemistry Principles and Technics, 2nd ed. p. 924.
Maryland, New York, San Francisco, and London,
Harper and Row (1974).
Horn, D. B. Standardisation of mechanised serum
alkaline phosphatase determinations. Clin. chirn. Acta
37 (1972) 43.
McComb, R. B., Bowers, G. N., Jr. Study of optimum
buffer conditions for measuring alkaline phosphatase
activity in human serum. Clin. Chern. 18 (1972) 97.
Morton, R. K. Separation and purification of enzymes
associated with insoluble particles. Nature (Lond.) 166
(1950) 1092.
Moss, D. W., Baron, D. N., Walker, P. G., Wilkinson, J.
H. Standardisation of clinical enzyme assays. J. din.
Path. 24 (1971) 740.
Price, C. P., Woodman, D. D. An improved AutoAnaly-
zer technique for the determination of serum alkaline
phosphatase. cu« chim. Acta 3S (1971) 265.
Smith, A. F., Fogg, B. A. Possible mechanisms for the
increase in alkaline phosphatase activity of lyophilised
control material. Clin. Chem. 18 (1972) 1518.
Teasdale, P. R. A report on the regional quality control
of serum alkaline phosphatase. Ann. din. Biochem. 10
(1973) 57.