Biology Reference
In-Depth Information
was divided, with half pretreated with wheat-germ agglutinin (WGA) to precipitate
BALP, aiding identification of the LALP fraction. Both unadulterated and
pretreated samples were electrophoresed concurrently in adjacent wells. The elec-
trophoretic curves were generated using computer software (Phoresis, Sebia, Issy-
les-Molineaux, France), with relative ALP isoform concentrations calculated as the
percentage of optical adsorbance of the absolute ALP concentration (IU/L).
A chemical autoanalyzer previously determined appropriate for veterinary use
(COBAS MIRAS-Roche, Basel, Switzerland) was used as a biochemical reference
method to measure TALP (Syakalima et al.
1997b
). CALP and BALP were
measured by heat inactivation and WGL precipitation (Sigma Chemical Co.,
St. Louis, MO, USA), respectively, as described in the literature (Teske et al.
1986
; Syakalima et al.
1997b
). LALP concentration was calculated by subtraction
of BALP and CALP from TALP (Syakalima et al.
1998
). Within-run and between-
run precisions were calculated for all ALP isoform measurements obtained by both
biochemical and electrophoretic methods [coefficient of variation (CV)
¼
SD/
mean
100].
To establish within-run precision, two samples containing high (449 U/L) and
low (98 U/L) ALP concentrations were analyzed with both biochemical and
electrophoretic methods 10 times on the same day. Duplicate ALP isoform
measurements for the same high and low ALP samples were obtained by both
biochemical and electrophoretic methods, using the same batch of gels, on 5 differ-
ent days during a 1-week period to estimate between-run precision. To assess
linearity, one sample containing a high ALP concentration (2,400 U/L) was serially
diluted 1:1, 1:2, 1:8, 1:16, and 1:32, and ALP isoforms were determined with both
methods on every dilution. Linearity for each isoform was determined by linear
regression of the two test methods. ALP isoform measurements obtained by the two
methods were compared using Kruskal-Wallis tests (
P
0.05). Agreement
between the two methods was analyzed using linear and Deming regression
methods to assess correlation (
R
2
values), and the presence of constant (intercept
6¼
<
1) errors using a Bland-Altman plot with bias was
defined as the mean difference between methods. Statistical analyses were
performed in MedCalc for Windows, version 9.2.1 (MedCalc Software,
Mariakerke, Belgium).
0) or proportional (slope
6¼
22.3 Results
Within- and between-run precisions and linearities of the two test methods are
presented in Table
22.1
. The Hydragel method demonstrated separate migration
patterns for unadulterated and pretreated samples only for groups A and B: unclear
migration pattern divisions, with more fractions than expected occurred in group C.
BALP and LALP concentrations were not significantly different between the two
test methods in groups A and B (group A:
P
BALP
¼
0.701,
P
LALP
¼
0.107; group B:
P
BALP
¼
0.915,
P
LALP
¼
0.428).
In group C, BALP concentrations were
