Transcript Evaluation of the Presage™ ST2 ELISA Jun Lu

Evaluation of the Presage™ ST2 ELISA
Jun Lu1, David G. Grenache1,2
1ARUP
Institute for Clinical and Experimental Pathology, Salt Lake City, UT
2Department of Pathology, University of Utah Health Sciences Center, Salt Lake City, UT
Analytical Validation
Background: Soluble ST2 (sST2) is a protein in the interleukin-1 receptor family
Introduction
The Presage ST2 sandwich immunoassay (Critical Diagnostics, New York, NY) was
developed to quantify the concentration of ST2 in human serum or plasma, a novel
prognostic marker of heart failure. We evaluated the assay’s performance characteristics
including precision, linearity, recovery, analytical sensitivity, limit of quantification,
stability, and sample type comparisons. Reference intervals were established from serum
obtained from 120 male and 120 female self-reported healthy individuals.
High
Medium
Low
Within Day
CV, %
3.8
2.4
7.6
Total
CV, %
6.3
14.0
11.5
Table 1. Within day and total precision were determined by
testing three specimens in triplicate each day for five days.
Research and Development -- ARUP
Percent of sST2 compared with base line after
RT
Refrigerated
Frozen
30 days
1 day
2 days
10 days
96.4%
94.3%
97.7%
96.0%
Referenceinterval,
Interval U/mL
Estimation: Co
Reference
Estimation:
CombinedReference
N Interval
Median
sST2, U/mL
Male
120
Female 120
Overall
240
Nonparametric
(CLSI C28-A)
Value
8.367
21.7
17.3
Lower
95% CI
18.9
<7.087 to 9.808
Alternatives:
Transformed Parametric
Parametric
Table 2. sST concentration of a patient sample was tested at
time 0, sample pools were stored at room temperature, 4 C
and -20 C respectively, sST2 was tested after storage.
10.4-52.1
8.4-33.6 Lower
Upper
Confidence
Value
95% CI
Ratio
Value
95% CI
8.9-41.3
33.603 25.539
to >39.843
>0.34
Nonparametric
(CLSI
C28-A)
10.363 <8.543 to 13.218
52.139
Alternatives:
31.243 Transformed
28.306 to 34.486
Parametric0.18
28.792 Parametric
27.024 to 30.560
0.16
47.621
43.842
Linearity
Confidence Limits for Nonparametric CLSI C-28A method computed by exact formula.
Histogram
15%
10%
180
800
y = 1.07x + 3.20
R2 = 0.93
700
100
80
60
-10
600
0
10
20
30
40
None
None
Sele
Boun
Filter
Statistics:
30%
Mean
SD
Median
Range
20%
N
Distinct values
Zeroes
Central 95% Index
10%
17.4926
5.7650
17.2745
7.087 to 39.843
120 of 120
119
0
3.0 to 118.0
Stati
Mean
SD
Medi
Rang
N
Distin
Zero
Cent
Analyst
Expt. Date
Jun
17 Jul 2009
Anal
Expt
0%
50
-20
0
20
0
0
ST2, U/mL
Probability Plot
(Original Data)
500
Probability Plot
(Transformed Data)
Probability Plot
(Original Data)
Frequency plots of 120 females
Nonparametric 95% CI
Parametric
Parametric
60
54.6
Figure 3. sST2 serum concentrations in US self-reported healthy 120 male
Normalizing Transformation
and 120 female individuals . Concentrations
of sST2 were not correlated by
age in either gender (r=-0.09; p=0.31 for all). Mean sST2 values show
significant difference between males and females (24.2 U/mL vs. 17.5 U/mL,
respectively, p<0.0001), and are significant higher than those reported for
an Austrian cohort (13.0 U/mL vs. 8.0 U/mL, respectively) (1).
50
33.1
20.1
40
20.1
30
Exponent
Constant20
10
0.00 (log)
0.00
10
100
-3
0
0
20
40
60
80 100 120 140 160 180
0
100 200 300 400 500 600 700 800
Expected concentration, U/mL
Figure 2. Different sample types for sST2 were compared by
testing serum and plasma specimens obtained from each of
13 patients.
Analytical sensitivity:
Specimen
sST2, U/mL
Volume, µL
Expected sST2,U/mL
Measured sST2, U/mL
Recovery, %
High
Low
107.4
29.8
8
12
61.6
57.8
93.8
High
Low
107.4
29.8
12
8
76.6
73.2
95.6
Table 3. Recovery was evaluated by combining two specimens
with sST2 concentrations of 107.4 (high) and 29.8 U/mL (low)
in ratios of 2:3 and 3:2 (high:low) in singlet.
Limit of Absence (LOA) was investigated by measuring 10
replicates of the zero calibrator and calculating as:
LOA (Mean + 3SD) = 0.94 + 3x0.22 = 1.6 U/mL
Limit of Detection (LOD) was investigated by measuring 10
replicates of a concentration near the expected LOD and
calculating as:
LOD (LOA + 3SD) = 1.6 + 3x0.14 = 2.0 U/mL
Limit of Quantification (LOQ) was investigated by measuring
10 replicates of a serial dilution of a low sample and
determining the concentration which produces a CV 20%. At
concentrations of 2.5, 3.3 and 4.0 U/mL, CVs are 23, 20 and
12%, respectively.
LOQ = 3.3 U/mL
-2
-1
0
1
True Gaussian SDI
Accepted by:
Serum, U/mL
Figure 1. Linearity of sST2 was determined by serially
diluting a specimen with an sST2 concentration of 214.8
U/mL and testing each sample in duplicate. Error bars
represent the SD.
Ex
Co
12.2
7.4
0
Nor
20
200
20
80
Probability Plot
(Transformed Data)
Nonparametric 95% CI
Parametric
30
300
60
Frequency plots of 120 males
Parametric
33.1
400
40
ST2, U/mL
12.2
40
Histogram
0
120
0%
0
Plasma, U/mL
140
5%
0
160
Confidence Limits for Nonparametric CLSI C-28A method computed by e
Selection Criteria:
40%
Bounds
Filter
25%
Paired samples
y = 0.95x + 2.25
R2 = 1.00
Value
8.841US
8.009
to 9.758
10.547 9.373
to 11.866
Table 4. Summary of
self-reported
healthy
cohort
reference
intervals
6.193
4.426 to 7.961
4.487
1.409 to 7.566
calculated using a non-parametric percentile method (95% double-sided).
20%
Measured concentration, U/mL
secreted by myocytes in response to mechanical strain. Like NT-proBNP, elevated serum
concentrations of sST2 are strongly prognostic in patients with heart failure. Although
analogous for prognosis the two markers have been shown to be independent and
complementary.
Objective: To perform an analytical validation of the Presage™ ST2 enzyme-linked
immunosorbent assay (ELISA) for the quantitative determination of sST2 in human
serum or plasma.
Methods: sST2 was measured using the Presage ST2 (Critical Diagnostics, New York,
NY) ELISA. Performance characteristics including imprecision, linearity, recovery,
analytical sensitivity, limit of quantification, stability, and sample type comparisons were
established using residual specimens sent to ARUP Laboratories. Gender-specific
reference intervals were established by non-parametric analysis using serum specimens
obtained from 120 healthy males and 120 healthy females. The project was approved by
the University of Utah’s Institutional Review Board.
Results: Within day and total precision were determined by testing three specimens in
triplicate each day for five days. At sST2 concentrations of 11.6, 26.9, and 88.0 U/mL,
the within-day CV was 7.6, 2.4, and 3.8%, respectively and the total CV was 11.5, 14.0,
and 6.3%, respectively. Linearity of sST2 was determined by serially diluting a
specimen with an sST2 concentration of 214.8 U/mL and testing each sample in
duplicate. The assay was linear over a concentration range of 2.8-161.1 U/mL
(y=0.95x+1.37; r=1.0). Recovery was evaluated by combining two specimens with sST2
concentrations of 107.4 (high) and 29.8 U/mL (low) in ratios of 2:3 and 3:2 (high:low).
Measured concentrations were 57.8 and 73.2 U/mL, which were 95.0 and 95.9%,
respectively of the expected concentrations. Analytical sensitivity was determined to be
1.6 U/mL by measuring 10 replicates of the zero calibrator, calculating the mean, and
adding 3 standard deviations. The limit of quantification, defined as the concentration
that produced a CV <20%, was determined to be 3.3 U/mL. sST2 stability was
determined using pooled patient specimens with an sST2 concentration of 12.3 U/mL.
After storing aliquots for 2 days at room temperature, 10 days at 4 C, and 30 days at -20
C, sST2 concentrations were 94.3, 97.7, and 96.0% of the original concentration,
respectively. The mean (SD) concentration of sST2 determined in serum and plasma
obtained from each of 13 patients was 141.9 (194.7) and 155.6 (217.2) U/mL, respectively
and were not significantly different (p=0.43) and showed excellent agreement
(y=1.07x+3.2; r=0.96). Concentrations of sST2 were significantly greater in males
compared to females (24.2 vs. 17.5 U/mL, respectively; p<0.0001) but were not correlated
by age in either gender (r=-0.09; p=0.31). Reference intervals for sST2 assay were
determined to be 10.4-52.1 U/mL for males and 8.4-33.6 U/mL for females.
Conclusions: The Presage ST2 ELISA demonstrates excellent performance
characteristics for quantifying sST2 in serum or plasma. The assay is linear over a wide
sST2 concentration range and can precisely measure low concentrations of sST2.
Concentrations of sST2 are unaffected by age but are greater in males compared to
females.
Mean Activity
U/mL
88.0
26.9
11.6
Reference Intervals
Research and Development -- ARUP
0
Abstract
Signature
2
3
-3
-2
-1
0
1
True Gaussian SDI
2
3
-3
-2
-1
0
1
True Gaussian SDI
Accepted by:
Date
2
3
-3
-2
-1
0
1
2
3
True Gaussian SDI
Signature
Conclusions
•
•
•
The Presage sST2 ELISA demonstrates excellent performance
characteristics for quantifying sST2 . The assay is linear over a wide
sST2 concentration range and can precisely measure low concentrations
of sST2.
Concentrations of sST2 are unaffected by age but are greater in males
compared to females.
US healthy population reference values are considerably higher than
those in Austria.
References
1. Dieplinger B, Januzzi Jr JL, Steinmair M, Gabriel C, Poelz W, Haltmayer
M, Mueller T. Analytical and clinical evaluation of a novel high-sensitivity
assay for measurement of soluble ST2 in human plasma—the Presage ST2
assay. Clin Chim Acta 2009;409:33-40.
Acknowledgement
Support for this study was provided by the ARUP Institute for Clinical and
Experimental Pathology and Critical Diagnostics.