Biology Reference
In-Depth Information
Table 22.1 Within- and between-run precisions and linearities of the Hydragel and biochemical
methods
Test/isoform
Linearity (
R
2
)
Within-run
precision (CV)
Between-run
precision (CV)
Hydragel method
BALP
7.2
8.0
0.99
LALP
9.6
10.8
0.97
CALP
3.7
3.7
0.96
Biochemical method
BALP
4.2
4.2
0.99
LALP
4.8
3.9
0.99
CALP
1.1
1.0
0.99
Table 22.2 Results of simple linear regression models fit to measurements of ALP isoforms
determined by Hydragel and biochemical methods
Isoform
R
2
Slope (95% CI)
Intercept (95% CI)
Error
BALP
0.98 (0.95-1.02)
4.91 (
10.59-0.76)
0.99
No
LALP
1.21 (0.54-1.87)
11.57 (
56.01-32.87)
0.36
No
CALP
0.39 (
0.04-0.82)
60.05 (8.89-111.22)
0.27
Constant and proportional
BALP (group C)
0.38 (
0.80-0.04)
17.93 (4.59-31.26)
0.22
Constant and proportional
significantly different (
P
¼
0.005) between the two test methods. The CALP
isoform was only found in group C.
Simple linear regression models fit to the data demonstrated good correlation
(
R
2
close to 1) only for BALP (Table
22.2
). For the CALP isoform, constant and
proportional errors were present. These findings were supported by a Bland-Altman
plot (mean bias
22.6). Deming regression analysis also demonstrated constant
and proportional errors for BALP in group C.
¼
22.4 Discussion
The Hydragel system was easy and fast to operate, providing standardized analyses
and diminished procedural errors. Overall, correlation between the Hydragel and
COBAS methods was low. When all 40 dogs were included, BALP measurements
for the two methods were highly correlated; but if only data from group C were
analyzed, the correlation was poor, and constant and proportional errors were
present. The Hydragel system was capable of obtaining good electrophoretic
patterns only for healthy dogs. Electrophoretic migrations of dogs with hyperadre-
nocorticism had incorrect subdivisions, causing constant and proportional errors for
CALP and BALP isoform measurements. The Hydragel system has good reproduc-
ibility, precision, and specificity when analyzing human serum (Van Hoof et al.
1993
). The main difference between human and canine serum ALP is the presence
