Comparison of ELISA vs GC results
In this study results were compared both qualitatively and quantitatively. Quantitative
comparison means that respective results for both methods were mathematically and
graphically compared by calculating and analyzing slope, intercept, and correlation
coefficients. For the qualitative comparisons the respective ELISA and GC results were
stratified by concentration range and compared as to the respective public health
interpretations that could be made based on the results. For example, the paired results
could be stratified and compared as to whether or not they are above or below the MCL.
Table 3 is a side-by-side compilation of all the results generated in this study. The sample
number is in column 1, the diaminoatrazine ELISA results in column 2, and the GC
diaminoatrazine results are in column 5. As a first step in understanding the data, the
ELISA diaminoatrazine results were graphed against the GC diaminoatrazine results and
the slope, intercept, and correlation coefficients were calculated. For this analysis the
"below detection" results are assumed to be zero. The results are presented in figure 1.
This data analysis suggests that the concentration results for the two methods match quite
well. The correlation coefficient is 0.9134 and the slope is 0.5037. A correlation of
greater than .90 is indicative that a values from column 2 (ELISA diaminoatrazine
results) will accurately track with the values in column 5 (GC diaminoatrazine results).
The slope, however tells you the GC value will track at approximately 50% of the ELISA
value, or the ELISA results are biased towards yielding results that are double the GC
results. The higher ELISA concentrations are probably due to low detection limits of the
ELISA method, better recovery at lower concentrations, and some cross-reactivity to
other triazine compounds.
In order to better understand the public health implications of the bias identified in figure
1, the data was stratified into ranges of values based on the levels that are currently used
in advising well owners about actions they might wish to take based on levels of atrazine
found in their wells. Currently, owners are advised that the water is potable if no atrazine
is detected. If detectable levels are found that are below the MCL (3.0 ppb), owners are
told to continue periodic testing to see if levels increase and to consider not using the
water for drinking. If the level is above the 3 ppb MCL, they are advised to not consume
the water. A comparison of the two methods with the data stratified into groups is
presented in table 4. In addition to the strata based on health advisories an additional
strata was include based on the limit of detection of the GC method. The data indicates
that using the diaminoatrazine ELISA test as a surrogate for the GC test would result in
11 of the 74 wells testing above the MCL. Similarly, the ELISA test would have found
only 13 wells were below detection, while the GC test would have produced 53 wells
where no diaminoatrazine was detected.
While it is interesting to understand how the new diaminoatrazine ELISA test system
performed when compared to the diaminoatrazine GC test, it is more important from a
drinking water safety standpoint to make comparisons with total atrazine. "Total atrazine"
is the sum of the atrazine, desethylatrazine, diaminoatrazine, and deisopropylatrazine
concentrations. Colunm 4 of table 3 contains the sum concentrations of atrazine and
diaminoatrazine determined using ELISA. Colunm 9 contains the sum of atrazine plus all