ELISA Lab Report Discussion
Chapter 4: Discussion
Considering that this laboratory experiment aimed at developing and validating a competitive
ELISA assay for the measurement of cortisol in serum as well as making comparison between
the ELISA results with the results of the LC-MS/MS obtained from the RCPA Quality Assurance
Programs (QAP), it is clearly evident that it is possible to conduct a critical evaluation of the
most reliable, highly reproducible, highly sensitive, and accurate method for the quantification of
cortisol. (1) This is attributable to the fact that measurements of cortisol extremely require very
high specificity even at very low concentrations to monitor the hormone status in clinical
laboratories to ensure the diagnostic accuracy, appropriate treatment and follow-up. (2)
Therefore, since the experiment required extraction of cortisol it is very justifiable that the first
part of the project involved the development of the ELISA technique for cortisol extraction
which began by selection of the extraction solvent whereby three solvents such as hexane, ethyl
acetate and MTBE were utilized for the extraction of the cortisol subsequently followed by
running them on LCMS/MS. (3) However, out of the three solvents ethyl acetate was selected as
the extraction solvent for cortisol because it was in a non-liquid matrix and also because ethyl
acetate in organic and cortisol is organic soluble steroid meaning that an organic phase extraction
of cortisol could only have been achieved using ethyl acetate. However, using ethyl acetate as a
solvent also had another advantage because it was easy to completely remove the solvent from
cortisol using centrifugal vacuum devices. (4)
In order to succinctly determine the concentration of cortisol at any measurement ranging
between the minimum and maximum concentration, a calibration curve was plotted utilizing the
five parametric logistic calibration plots and the calibration curve showed a working range of
ELISA Lab Report Discussion 3
between 250nmol/L and about 750nmol/L. The results of the calibration curve implies that the
ELISA method is capable of measuring cortisol within a range of the least concentration to
highest concentration, that is, from the lowest concentration of 250 nmol/L to the highest
concentration of 750 nmol/L. (5) Therefore, the plotted calibration curve can be very essential in
the determination of the unknown concentrations of the sample solution in an experiment since
the sample solution is plotted against the observable variable on basis of their concentration and
absorbance or the calibration standards of several prepared solutions. After plotting the
calibration curve, unknown solution concentrations (which in this case are assumed to mean
cortisol concentrations) can be determined upon placing the sample solution on the calibration
curve mainly on the basis of its absorbance or on the basis of other observable variable. (6) In
this experiment, a calibration curve was vital for the determination of concentrations of cortisol
at varied ranges of absorbance on the basis of other observable variable. (7-10)
The experiment also considered the validation of the method through evaluation of the
imprecision, sensitivity, accuracy, and linearity for the ELISA. (8) This was done because it is
highly essential when it comes to making decisions or conclusions about the reliability and/or
validity of an experimental procedure. (3) For example, according to Lequin (2) in order to make
a clinically significant decision concerning an analyte, the detectable degree of change is
equivalent to the total allowable error (TEa). This means that accuracy or bias is an indication of
the lack of agreement between two methods under comparison. However, for a particular method
the detected systemic error is recorded as the negative or positive bias. The variation between
methods under comparison is described through a coefficient of variation which gives a good
idea concerning the assay or the method performance. According to Lumsden (7) the
performance of a good method is generally denoted by a CV of less than 5% and whenever the
ELISA Lab Report Discussion 4
CV is 10% or higher, this is an indication that the performance of the method is not satisfactory.
According to Yeh, Glock and Ryu (5) the upper limit of CV for cortisol is + 25%, while the
ELISA results showed a CV of 32% and a bias of 19 % for the high sample and CV of 24% and
a bias of 20.7% for the low sample. This shows that the performance of ELISA was not good for
the analyte. Therefore, for further analysis of this analyte, medical decision chart (sigma chart)
was used. The gold standard for defining the world class measure of quality is through
achievement of Six Sigma. This implies that when an assay is performed at the 3-sigma level, it
is regarded as the minimum acceptable level of quality; whereas when an assay or a method is
below the 3-sigma level, a new better method should be adopted mainly because the quality of
the test cannot be ensured and/or assured even if the QC runs are repeated. (7) Therefore, for this
experiment the ELISA was unacceptable for both the levels on basis of the sigma chart.
Moreover, the experiment results were also used to plot the linearity which according to the
obtained results showed that the ELISA was linear upto750 nmol/L. This meant that the upper
limit of detection was 750nmol/L. The results of the linearity plot showed that they are consistent
to the calibration curve results which also showed the upper limit as 750nmol/L. Therefore,
despite the fact that the performance of ELISA was not good for the analyte, there is a significant
correlation between the two methods such as the calibration curve and the linearity curve. (2)
Furthermore, the lowest concentration of the analyte is described by terms such as sensitivity,
analytical/functional sensitivity, LoB, LoD and LoQ and they can be measured by a method or
an assay. For instance, LoB shows that 95% of the observed values when a Gaussian distribution
is considered, while LoD gives an estimate of imprecision and bias at very low concentrations of
the analyte meaning that it is the smallest concentration of the analyte that is distinguishable
from the LoB implying that it is greater than LoB. (11) However, a consideration of the Gaussian
ELISA Lab Report Discussion 5
distribution, 95% of the values of the analyte exceeded the LoB whereas 5% values of the
analyte were below LoB. In addition, LoQ is considered to be the smallest concentration at
which the analyte can be measured and cannot be detected but where it meets the goals for
imprecision and bias. Furthermore, the functional sensitivity measures the precision of an assay
at low analyte concentrations that resulted in a CV of 20%. The LoQ may be equal to the LoD
but it cannot be lower. LoQ will be equal to LoD if the bias and imprecision at the LoD is within
the total allowable error. (1) This implies that a consideration of the results obtained for LoB,
LoD and LoQ can give some insights about that imprecision and biasness of the method and/or
analyte on basis of the total allowable error. Thus, in this experiment, the LoD is not equal to
LoQ, meaning that the bias and imprecision do not meet the requirements of the total allowable
error. (5-8, 19-21)
Finally, the experiment compared the results of the ELISA to those of the RCPA QAP through
an analysis using the Passing Bablok as well as the difference plot. The results showed a value of
0.829 and a slope of 0.604 as well as an intercept of 86 nmol/L, which implies that there is a
good agreement between the RCPA and the developed ELISA because the developed ELISA and
RCPA QAP showed a mean difference of -152 nmol/L. (8, 9, 12, 15-18) Furthermore, the Levey
Jennings chart showed that all the QCs didn’t work because the medium and high QC values
were within the acceptable limits even though the low QC was out. This is an indication of
improvement in the development of the assay, and the bad results may be attributable to the fact
that the errors in the experiment procedures such as extraction or pipetting. (1, 15-17) In
addition, this variation may also be attributed to the errors that may occur during the process of
labeling or transferring of the analyte or calibration deviations within the measurement
equipment and/or devices used during the experiment. Therefore, this means that during the
ELISA Lab Report Discussion 6
experiment the standard procedures should be strictly adhered to in order to ensure
reproducibility of the obtained results. (21)
ELISA Lab Report Discussion 7
References
- Hoppert M. Microscopic techniques in biotechnology. Weinheim: Wiley-VCH; 2003.
- Lequin R. Enzyme immunoassay (EIA)/enzyme-linked immunosorbent assay (ELISA). Clin.
Chem. 2005; 51(12): 2415–8. - Wide L, Porath J. Radioimmunoassay of proteins with the use of Sephadex-coupled
antibodies. Biochem Biophys Acta. 1966; 30:257–260. - Leng, S, McElhaney J, Walston J, Xie D, Fedarko N, Kuchel G. Elisa and Multiplex
Technologies for Cytokine Measurement in Inflammation and Aging Research. J Gerontol a
Biol Sci Med Sci. 2008; 63(8): 879–84. - Yeh CM, Glock M, Ryu S. An optimized whole-body cortisol quantification method for
assessing stress levels in larval zebrafish. PloS one. 2013; 8(11):e79406. - Schoenmakers CH, Naus AJ, Vermeer HJ, van Loon D, Steen G. Practical application of
Sigma Metrics QC procedures in clinical chemistry. Clinical chemistry and laboratory
medicine: CCLM / FESCC. 2011; 49(11):1837-43.
