vvEPA
United States
Environmental Protection
Agency
Method 546: Determination of Total Microcystins and
Nodularins in Drinking Water and Ambient Water by
Adda Enzyme-Linked Immunosorbent Assay

-------
Questions concerning this document should be addressed to:
Steven C. Wendelken. Ph.D.
U.S. EPA, Office of Ground Water and Drinking Water, Standards and Risk Management Division,
Technical Support Center, 26 W. Martin Luther King Dr., Cincinnati, OH 45268
Phone:(513)569-7491
wendelken.steve@epa.gov
Office of Water (MS-140)
EPA 815-B-16-011
EPA contract EP-C-12-013
August 2016
Authors
Alan Zaffiro, CB&I Federal Services, LLC (Cincinnati, OH)
Contractor's role did not include establishing Agency policy
Laura Rosenblum, Ph.D., CB&I Federal Services, LLC (Cincinnati, OH)
Contractor's role did not include establishing Agency policy
Steven C. Wendelken, Ph.D., U.S. EPA (Cincinnati, OH)
Acknowledgements
Ohio EPA Division of Environmental Services (Reynoldsburg, OH): Ohio EPA demonstrated the utility of
the Adda ELISA technique for the monitoring of microcystins in Ohio lakes and in drinking water when
concern over harmful algal blooms surfaced in the State. The U.S. EPA used the standard operating
procedure (SOP) published by OH EPA (Ohio EPA. 2015) as a foundation for developing the guidance in
this method. Ohio EPA provided analytical support as a participant in the Second Laboratory Validation
of Method 546 and served as a peer reviewer for the method.
Abraxis, Inc. (Warminster, PA): Abraxis served as a technical resource and as a peer reviewer for the
method.
North East Ohio Regional Sewer District (NEORSD, Cuyahoga Heights, OH): NEORSD provided analytical
support as a participant in the Second Laboratory Validation of Method 546 and served as a peer
reviewer for the method.
Eurofins Eaton Analytical, Inc. (Monrovia, CA): Eurofins Eaton Analytical provided analytical support as a
participant in the Second Laboratory Validation of Method 546 and served as a peer reviewer for the
method.
Mark Domino, CB&I Federal Services, LLC (Cincinnati, OH): Mr. Domino served as a technical resource
and laboratory analyst during the development of Method 546.
i

-------
Contents
1	Scope and Application	1
2	Summary of the Method	1
3	Definitions	2
4	Interferences	4
5	Safety	5
6	Equipment and Supplies	5
7	Reagents and Standards	7
8	Sample Collection, Preservation, and Storage	7
9	Quality Control	8
10	Calibration	12
11	Procedure	12
12	Data Analysis and Calculations	13
13	Method Performance	13
14	Pollution Prevention	15
15	Waste Management	15
16	References	16
17	Tables, Figures, and Method Performance Data	17
ii

-------
1 Scope and Application
1.1	Method Description
Method 546 is a procedure for the determination of "total" microcystins (MC) and nodularins (NOD) in
finished drinking water and in ambient water using enzyme-linked immunosorbent assay (ELISA). The
term "Total microcystins and nodularins" is defined as the sum of the congener-independent,
intracellular and extracellular microcystin and nodularin that is measurable in a sample. Method 546
measures the total concentration based on detection of a characteristic feature common to microcystin
and nodularin congeners (structural variants), specifically, the Adda amino acid side chain: (4£,6£)-3-
amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid (Fischer et al.. 2001). To ensure
comparability between laboratories, the ELISA is calibrated against one congener, MC-LR (CASRN
101043-37-2).
1.2	Required Operator Expertise
This method is intended for use by analysts skilled in the performance of ELISA and in the interpretation
of the associated data. The analyst should be familiar with the calibration model used for indirect
competitive ELISA, the four-parameter logistic fit, which is described in Section 3.3.
1.3	Supporting Data
1.3.1	Single-Laboratory Lowest Concentration Minimum Reporting Levels (LCMRLs)
LCMRLs for MC-LR are presented in Section 17, Table 1. The LCMRL is the lowest spiking concentration
such that the probability of spike recovery in the 50 to 150% range is at least 99%. Laboratories using
this method are not required to determine LCMRLs, but they must demonstrate that the Minimum
Reporting Level (MRL) (Sect. 3.11) for MC-LR meets the requirements described in Section 9.1.3.
1.3.2	Single-Laboratory Precision and Accuracy Data
Precision and accuracy data for MC-LR fortified in reagent water, drinking water, and ambient water are
presented in Section 17. Table 2, Table 3, and Table 4.
1.4	Method Flexibility
The Adda-ELISA technique must be used for the congener-independent detection of microcystins and
nodularins. An indirect competitive assay must be used. Changes may not be made to sample collection,
preservation, and storage procedures (Sect. 8). or to the quality control (QC) requirements (Sect. 9).
2 Summary of the Method
This method is based on the 96-well microtiter plate format. In these wells, microcystins and nodularins
in the samples and a microcystin-protein analogue immobilized in the wells compete for the binding
sites of a primary detection antibody in solution. After a wash step, an enzyme-conjugate is added to the
wells and binds to the primary antibody in an inverse relationship to the original concentration of
microcystins and nodularins in the sample. After a second wash step, tetramethylbenzidine substrate is
added to develop color via an enzyme-mediated reaction. After a set period, an acidic solution is added
to each well to stop color generation. Finally, the absorbance of each well is measured using a plate
reader. The concentration of microcystins and nodularins is calculated using a four-parameter logistic
calibration curve.
1

-------
3 Definitions
3.1	Analysis Batch
Standards, samples, and quality control elements are assayed on a single 96-well plate using identical
lots of reagents and wells. Each plate by definition is an Analysis Batch, regardless of the number of
wells included. Quality control samples must be analyzed in each Analysis Batch at the frequencies
prescribed in Section 9.2. Each Analysis Batch includes the following elements:
•	Calibration standards
•	Laboratory Reagent Blanks
•	Low-Range Calibration Verification standard
•	Laboratory Fortified Blanks
•	Field samples (drinking water and ambient water)
•	Laboratory Fortified Sample Matrix and Laboratory Fortified Sample Matrix Duplicates
3.1.1	Well Replicates
Within the Analysis Batch, this method requires each calibration standard, field sample, and QC sample
to be assayed in two wells. These two wells are called well replicates. Two values are associated with
each well replicate: an absorbance measured by the plate reader, and a concentration calculated from
this absorbance.
3.1.2	Use of Well Replicate Absorbance Values
For each set of well replicates, the percent coefficient of variation (%CV) is calculated from the two
absorbance values. The %CV of the absorbance values for calibration standards must meet QC criteria
stated in Section 10.2. The %CV of the absorbance values for all field and QC samples must meet the
limits stated in Section 9.2.1.
3.1.3	Use of Well Replicate Concentrations
For each set of well replicates, the mean is calculated from the two concentration values. The mean
concentration must be used for reporting field sample results. The mean must be used in all method
calculations and for evaluating results against QC limits.
3.2	Calibration Standards
Solutions of MC-LR provided in the ELISA kit or prepared in the laboratory that are appropriate for the
measurement range of the ELISA kit.
3.3	Calibration Curve
The calibration points are modelled using a four-parameter logistic function (Sect. 3.3.1), relating
concentration (x-axis) to the measured absorbance in the wells (y-axis). A calibration curve generated
during method development is presented in Section 17, Figure 1. Note the inverse relationship between
concentration and response. The zero calibration standard gives the highest absorbance and the highest
calibration standard gives the lowest absorbance. Note also that the slope, or sensitivity, of the ELISA
response is greatest in the middle of the curve and tends toward zero slope at extreme low and high
concentrations. For a more detailed explanation of the four-parameter calibration model, see Maciel
(1985) and Sasaki (web resource).
2

-------
3.3.1 Four-Parameter Logistic Equation
(a-d)
y=	TZZh +d
^(f)
y = absorbance
x = concentration
a = absorbance at the bottom plateau
b = slope related term at the inflection point
c = concentration at the inflection point = EC50
d = absorbance at the top plateau
The coefficients, a, b, c, and d, are calculated by the data reduction software using regression analysis.
3.4	ECso
The EC5o is the concentration of microcystin that yields an absorbance halfway between the bottom
plateau of the calibration curve (coefficient a) and the top plateau (coefficient d). The EC50 is the
concentration at the inflection point (Section 17, Figure 1). and is in the center of the most reliable
measurement range (i.e., the greatest slope) of the ELISA. The EC50 is determined and used as described
in Sections 3.4.1 and 3.4.2.
3.4.1	Determining the EC50
For each calibration curve, the EC50 is equal to the coefficient, c, of the four-parameter logistic fit. The
EC50 is found on the calibration report generated by the plate reader.
3.4.2	UseoftheEC5o
Because the EC50 is in the center of the most reliable measurement range, guidance for fortifying QC
samples is based on this value. For this method, QC samples requiring fortification with MC-LR should
have concentrations near the EC50. These QC samples include Laboratory Fortified Blanks and Laboratory
Fortified Sample Matrix.
3.5	Laboratory Fortified Blank (LFB)
An aliquot of reagent water to which a known quantity of MC-LR is added. The LFB is lysed and filtered
to match the analytical procedure for field samples. The LFB is used during the IDC to verify method
performance for precision and accuracy. The LFB is also a required QC element with each Analysis Batch.
The results of the LFB verify method performance in the absence of sample matrix.
3.6	Laboratory Fortified Sample Matrix (LFSM)
An aliquot of a field sample to which a known quantity of MC-LR is added. The purpose of the LFSM is to
determine whether the sample matrix contributes bias to the analytical results.
3.7	Laboratory Fortified Sample Matrix Duplicate (LFSMD)
A second aliquot of the field sample used to prepare the LFSM that is fortified and assayed in the same
Analysis Batch as the LFSM. The LFSMD is used to verify method precision in both ambient water and
drinking water matrixes.
3

-------
3.8	Laboratory Reagent Blank (LRB)
An aliquot of reagent water that is lysed and filtered to match the sample processing procedure. The
LRB is used to determine if microcystins or other interferences are introduced from the sample
containers, sample processing equipment, or the reagents used in the assay.
3.9	Lowest Concentration Minimum Reporting Level (LCMRL)
The single-laboratory LCMRL is the lowest spiking concentration such that the probability of spike
recovery in the 50 to 150% range is at least 99% (U.S. EPA. 2004 and U.S. EPA. 2010).
3.10	Low-Range Calibration Verification (Low-CV)
The Low-CV is a calibration standard with a concentration equal to, or less than, the MRL. The purpose
of the Low-CV is to confirm the accuracy of the calibration at concentrations near the MRL.
3.11	Minimum Reporting Level (MRL)
The minimum concentration that can be reported by a laboratory as a quantified value for total
microcystins and nodularins in a sample following analysis. This concentration must meet the criteria
defined in Section 9.1.3 and must be no lower than the concentration of the lowest calibration standard.
3.12	Primary Dilution Standard (PDS)
A solution of MC-LR in methanol prepared from the MC-LR Stock Standard Solution. The PDS solutions
are used to fortify QC samples (LFB, LFSM, and LFSMD).
3.13	Quality Control Sample (QCS)
A solution containing MC-LR at a known concentration that is obtained from a source different from the
source of calibration standards. The purpose of the QCS is to verify the accuracy of the primary
calibration standards.
3.14	Reagent Water
Purified water that does not contain any measurable quantity of microcystins, nodularins, or interfering
compounds at or above one-half of the MRL.
3.15	Stock Standard Solution
A concentrated standard in methanol that is prepared in the laboratory from purified MC-LR or that is
purchased from a commercial source with a certificate of analysis.
4 Interferences
4.1	General Considerations
The accuracy of the ELISA procedure is dependent upon analyst technique, the accuracy of pipetted
volumes, and consistent incubation periods across the wells of each plate.
4.2	Assay Drift
Assay drift refers to systematic rather than random imprecision in measured analyte concentrations, the
magnitude of which depends on the sample position within the plate. A possible cause for assay drift is
slight differences in incubation times as reagents are added sequentially across the plate (Davies, 2005).
Identical control samples distributed throughout the plate are necessary to detect assay drift. This
method includes QC measures to evaluate assay drift. During the IDC, laboratories must assay five LRBs
distributed across the plate and each Analysis Batch must include two LRBs placed on opposite sides of
4

-------
the plate. Because the LRBs give absorbance values near the upper plateau of the calibration curve, the
calculated LRB concentrations are sensitive to slight changes in the measured absorbance. If the
distributed LRBs pass the QC limit of one-half the MRL, then plate drift is minimal and under control.
4.3	Sample Matrix Effects in Ambient Water
During method development, a positive bias of approximately 30% in ambient water from a single
source was observed. Two ambient waters from other sources did not cause matrix bias.
4.4	Sample Matrix Effects in Drinking Water
During method development, six drinking water sources were evaluated for matrix effects. Of these,
three exhibited positive bias within a range of 12 to 15%. Three sources did not cause matrix bias.
4.5	Cross Contamination
This method covers the analysis of samples collected from both ambient water and drinking water. To
avoid cross contamination, segregate glass syringes used to filter ambient water, which may contain
high levels of microcystins, from those used to filter drinking water. Alternately, use plastic, disposable
syringes. Thoroughly clean glass sample containers if they are reused. Do not reuse septa from bottles
containing ambient water samples.
4.6	Interpretation of Results
The results reported by this method represent total microcystins and nodularins based on Adda ELISA
calibrated with MC-LR.
5	Safety
Each reagent used in these procedures should be treated as a potential health hazard and exposure to
these materials should be minimized. Each laboratory is responsible for maintaining an awareness of
OSHA regulations regarding safe handling of any chemicals used in this method. A reference file of
Safety Data Sheets should be made available to all personnel involved in the analysis. The primary
hazard when conducting the procedures in this method is exposure to cyanotoxins in samples and
concentrated standards. Handle samples and standards using appropriate personal protective
equipment.
6	Equipment and Supplies
References to specific brands and catalog numbers are included as examples only and do not imply
endorsement of the products. Such reference does not preclude the use of other vendors or
manufacturers. Specific references are intended to represent suitable specifications for the items.
6.1	Adda ELISA Test Kits
Indirect competitive assay based on detection of the Adda epitope (Abraxis Part Number 5200110H, or
equivalent). Automated systems for processing the 96-well plate may be used.
6.2	Sample Containers
Amber glass bottles fitted with polytetrafluoroethylene (PTFE)-lined screw caps. The single use of
sample containers is recommended. Good laboratory practices for cleaning are expected to be followed
if sample bottles are reused. Do not use bottles treated at a high temperature in a muffle furnace
(400 °C and above) as a cleaning procedure. Studies conducted during method development indicate a
5

-------
tendency for microcystins and nodularins to adsorb to the surface of glassware repeatedly cleaned by
heating in a muffle furnace.
6.3	Plate Reader
Microtiter plate reader and associated software capable of reading absorbance at 450 nm and
constructing a calibration curve using a four-parameter logistic function (Sect. 3.3.1).
6.4	Adhesive Plate Covers
Clear adhesive film for sealing wells during the incubation steps of the assay, VWR Catalog Number
60941-120.
6.5	Pipette for Loading Wells
Adjustable or fixed-volume manual pipette with 50 microliter capacity and manufacturer-recommended
pipette tips. A manual pipette is recommended for addition of standards, field samples, and QC samples
to the wells.
6.6	Repeating Pipette
Repeating pipette with 5 mL capacity disposable tips, HandyStep® electronic (Wertheim, Germany). An
electronic repeating pipette is recommended for addition of antibody, enzyme conjugate, substrate, and
stop solution to the wells.
6.7	Multi-Channel Pipette
Eight-channel pipette with 250 piL capacity per channel and polypropylene pipette tips. An eight-channel
pipette is recommended for addition of wash solution to the wells during the washing steps.
6.8	Basin for Wash Solution
Plastic basin designed for multi-channel pipettes, VWR Catalog Number 21007-970. A basin with 55 mL
capacity is recommended.
6.9	Syringes for Filtering Samples after Lysing
Glass: gas-tight Luer-lock, 5 mL capacity, Hamilton Co. Model Number 1005TLL. Plastic: Luer-lock, 3 mL
capacity, polypropylene barrels with polyethylene plungers, Thermo Fisher Scientific, Inc., Catalog
Number S7515-3.
6.10	Syringe Filters for Filtering Samples after Lysing
25 mm glass fiber, 1.2 pim pore size, polypropylene housing, Environmental Express (Charleston, South
Carolina) Catalog Number SF012G. 25 mm glass fiber, 0.45 pim pore size, polypropylene housing, GE
Healthcare Life Sciences/Whatman (Marlborough, MA) Catalog Number 6894-2504.
6.11	Fifteen Milliliter to Forty Milliliter Vials
Borosilicate glass, clear or amber, with PTFE-lined closures. Vials with capacity in this range are
recommended for the lysing procedure. Do not use bottles treated in a muffle furnace at a high
temperature as a cleaning procedure. See caution in Section 6.2.
6.12	Four Milliliter Vials
Borosilicate glass, with PTFE-lined closures. Four milliliter vials are recommended for receiving and
storing sample filtrate after lysing.
6

-------
7 Reagents and Standards
7.1	Reducing Agent for Residual Chlorine
Sodium thiosulfate (CASRN 7772-98-7) is used to reduce residual chlorine in drinking water samples at
the time of collection.
7.2	Stock and PDS Solvent
Methanol (CASRN 67-56-1) is used to reconstitute MC-LR if this material is purchased as a solid, and to
dilute the MC-LR stock to prepare PDS solutions.
7.3	MC-LR Stock and PDS Solutions
Obtain MC-LR (CASRN 101043-37-2) as a solution with a concentration of at least 10 ng/mL or as the
neat material. Reconstitute the neat material in methanol to obtain a stock concentration of at least
10 ng/mL. Dilute the MC-LR stock with methanol to prepare PDS solutions. Select MC-LR concentrations
for PDS solutions such that at least 5 piL are used to fortify QC samples, or to prepare calibration
standards. More than one PDS concentration may be necessary to meet this requirement.
7.4	Calibration Standards
Calibration standards supplied in the ELISA kits are recommended. Calibration standards prepared in the
laboratory are permitted. Laboratories should use appropriate QC practices to determine when
standards need to be replaced.
8 Sample Collection, Preservation, and Storage
8.1	Sample Collection for Drinking Water
Prior to shipment to the field, add sodium thiosulfate to each sample bottle. The final concentration of
sodium thiosulfate in the sample should be 100 mg/L. Do not dilute sodium thiosulfate in water when
preparing sample bottles. The reducing agent must be added to the empty bottle in solid form. In the
field, open the tap and allow the system to flush for approximately 5 minutes. Fill each bottle, taking
care not to flush out the sodium thiosulfate, and invert several times to mix the sample with the
reducing agent.
8.2	Sample Volume
Collect enough sample to meet the requirements in this method. Considerations for sample size are
sufficient volume for preparing QC samples and appropriate volume for frozen storage.
8.3	Prohibition Regarding Ascorbic Acid
Do not use ascorbic acid to reduce chlorine in drinking water samples. During studies to evaluate analyte
stability during transport and storage, the authors discovered that microcystins degrade in the presence
of ascorbic acid.
8.4	Sample Collection for Ambient Water
The addition of sodium thiosulfate is not required for ambient water samples, but may be added if the
laboratory chooses to prepare only one type of sample container.
8.5	Sample Shipment and Storage
Samples must be chilled during shipment and must not exceed 10 °C during the first 48 hours after
collection. Freeze samples upon arrival at the laboratory.
7

-------
8.5.1	Verification upon Receipt
8.5.1.1	Temperature
Samples must be confirmed to be at, or below, 10 °C when they are received in the laboratory. A
temperature of greater than 10 °C is acceptable if transit time is short and the samples do not have
sufficient time to chill. In this case, examine the ice packs in the cooler. If they remain frozen, the
samples are valid.
8.5.1.2	Residual Chlorine
For drinking water, analyze one sample from each cooler using common assays for total residual
chlorine, for example, the /V,/\/-diethyl-p-phenylenediamine (DPD)-colorimetric technique. The total
chlorine concentration should be less than the detection limit of the assay. A duplicate sample may be
collected for conducting the residual chlorine assay.
8.5.2	Considerations for Frozen Storage
Use clear, or amber, borosilicate glass bottles with PTFE-lined septa. Select bottle capacity and sample
volume to prevent breakage of bottles during freezing. Plan ahead to retain enough volume for
preparing QC samples as required in this method.
8.6	Sample Holding Time
Analyze samples as soon as possible. Samples that are collected and stored as described in Section 8
must be analyzed within 14 days of collection. Fourteen days is set as a safe holding time based on the
empirical evidence: during method development, degradation of microcystins in two drinking water
matrixes was observed after three weeks of frozen storage (Sect. 13.2.3 and Table 9).
8.7	Storage of Samples after Lysing
Samples may be filtered and assayed any time after lysing if within 14 days of collection. If not assayed
immediately, store lysed samples by freezing in glass vials with PTFE-faced septa, for example, 1 mL of
lysed and filtered sample held in a 4 mL vial.
9 Quality Control
QC requirements include the IDC, and QC elements associated with each Analysis Batch. This section
describes each QC parameter, its required frequency, and the performance criteria that must be met in
order to satisfy EPA data quality objectives. These QC requirements are considered the minimum
acceptable QC protocol. Laboratories are encouraged to institute additional QC practices to meet their
specific needs.
9.1 Initial Demonstration of Capability (IDC)
The IDC must be successfully performed prior to analyzing field samples. The IDC includes four
determinations: demonstration of precision and accuracy, demonstration of acceptable system
background, MRL confirmation, and a second-source verification of the calibration standards (Quality
Control Sample). The IDC requirements are outlined in Section 17, Table 10.
When conducting the IDC, the analyst must meet the calibration requirements specified in Section 10.
The lowest calibration standard used to establish the initial calibration must be at, or below, the
concentration representing the MRL. The four determinations necessary to complete the IDC may be
included in a single Analysis Batch, i.e., processed on a single ELISA plate.
8

-------
9.1.1	Demonstration of Precision and Accuracy
Prepare seven replicate LFBs (Section 3.6) by fortifying each with MC-LR at 0.50 ng/L. Sodium thiosulfate
must be added as described in Section 8.1. Lyse, filter, and assay (Section 11) the LFBs in a single
Analysis Batch (Sect. 3.1). For this Analysis Batch, include the LRBs prepared in Section 9.1.2 to fulfill the
LRB QC requirement. A Low-CV (Sect. 9.2.3) is also required. The percent relative standard deviation
(%RSD) for MC-LR in the seven LFB replicates must be less than, or equal to, 15%. The mean recovery for
the seven replicates must be greater than, or equal to, 70% and less than, or equal to, 130%.
Average Measured Concentration
Average % Recovery =		—-	x 100
Fortified Concentration
Standard Deviation of Measured Concentrations
%RSD =			x 100
Mean Concentration
9.1.2	Demonstration of Acceptable System Background
In the same Analysis Batch constructed for the precision and accuracy IDC, lyse, filter, and assay five
LRBs (Section 3.8). The LRBs must contain sodium thiosulfate. The LRBs must be distributed throughout
the plate. The result obtained for each LRB must be less than one-half of the MRL.
9.1.3	Minimum Reporting Level (MRL) Confirmation
Establish a target concentration for the MRL based on the intended use of the method. Confirm the MRL
following the procedure outlined below.
9.1.3.1	Prepare and Assay MRL Samples
Fortify seven replicate LFBs with MC-LR at, or below, the proposed MRL concentration. The LFBs must
contain sodium thiosulfate as specified in Section 8.1. Lyse, filter, and assay the samples in a single
Analysis Batch. The Analysis Batch must include two LRBs (Sect. 9.2.2) and a Low-CV.
9.1.3.2	Calculate MRL Statistics
Calculate the mean and standard deviation of the seven replicates. Determine the Half Range for the
Prediction Interval of Results (HRpir) using the following equation:
HRpir = 3.963s
Where,
s is the standard deviation and 3.963 is a constant value for seven replicates (U.S. EPA. 2004).
Calculate the Upper and Lower Limits for the Prediction Interval of Results (PIR = Mean ± HRpir) as shown
below.
Mean + HRP,P
Upper PIR Limit =		—-	 x 100%
Fortified Concentration
Mean — HRPir}
Lower PIR Limit =		—-	 x 100%
Fortified Concentration
These equations are only defined for seven replicate samples.
9

-------
9.1.3.3 MRL Acceptance Criteria
The MRL is confirmed if the Upper PIR Limit is less than, or equal to, 150% and the Lower PIR Limit is
greater than, or equal to, 50%. If these criteria are not met, the MRL may be set too low and the
confirmation should either be repeated, or the MRL established and confirmed at a higher
concentration.
9.1.4 Quality Control Sample (QCS)
Assay a mid-level QCS, prepared as described in Section 9.2.6. to confirm the accuracy of the primary
calibration standards.
9.2 Analysis Batch QC Requirements
This section describes the QC elements that must be included with each Analysis Batch (Sect. 3.1). The
Analysis Batch QC requirements are summarized in Section 17, Table 11.
9.2.1	Criterion for Replicate Wells
All field and QC samples are added to at least two wells (Sect. 11.3). See Section 3.1.1 for a definition of
well replicates. The %CV of the absorbance values measured for the well replicates must be less than, or
equal to, 15%. Calculate the %CV as follows:
Standard Deviation of Absorbances
%CV = 					j	 x 100%
Mean Absorbance
If the %CV exceeds 15% for a field sample or QC sample (Low-CV, LRB, LFB, LFSM, and LFSMD), then that
sample is invalid. Note that the well replicates of calibration standards must meet a different set of
criteria for %CV. These criteria are found in Section 10.2.
9.2.2	Laboratory Reagent Blank (LRB)
For each Analysis Batch, prepare, lyse, and filter one LRB. The LRB must contain sodium thiosulfate if
drinking water samples are included in the Analysis Batch. Assay the LRB in duplicate by placing one pair
of well replicates on opposite sides of the plate (four total wells). The total microcystin and nodularin
concentration in each LRB must be less than one-half the MRL. If the concentration is equal to, or
greater than, this level, then any samples that yielded a positive result in the Analysis Batch are invalid.
Subtracting blank values from sample results is not permitted.
9.2.3	Low Calibration Verification (Low-CV)
With each Analysis Batch, assay one Low-CV. The Low-CV is a calibration standard prepared at a
concentration equal to, or less than, the MRL. A calibration standard from the kit may be used. Do not
add sodium thiosulfate, do not lyse, and do not filter the Low-CV. The assayed concentration in the
Low-CV must be greater than, or equal to, 50% and less than, or equal to, 150% of the true value. If the
result fails this criterion, then the entire Analysis Batch is invalid.
9.2.4	Laboratory Fortified Blank
At least two LFBs, fortified at an identical concentration, are required with each Analysis Batch. Add
sodium thiosulfate if drinking water samples are included in the Analysis Batch. Fortify the LFB near the
EC5o of the calibration curve. Lyse and filter each LFB in a separate vial. See Section 3.4.1 for help
determining the EC5o.
10

-------
9.2.4.1 Acceptance Criterion for LFBs
The percent recovery for each LFB must be greater than, or equal to, 60% and less than, or equal to,
140% of the true value. If either LFB fails this criterion, then the entire Analysis Batch is invalid.
9.2.5 Laboratory Fortified Sample Matrix (LFSM) and Laboratory Fortified Sample Matrix Duplicate
(LFSMD)
One LFSM and LFSMD set is required with each Analysis Batch in a drinking water sample. If more than
20 drinking water samples are present in the Analysis Batch, two sets are required. One LFSM and
LFSMD set is required with each Analysis Batch in an ambient water sample. If more than 20 ambient
water samples are present in the Analysis Batch, two sets are required. The native background
concentration in the sample matrix must be determined in a separate field sample. Over time, distribute
LFSMs among the various drinking water and ambient water sources that the laboratory receives.
9.2.5.1	Prepare the LFSM and LFSMD
Three separate aliquots of a field sample are required for determining the native background
concentration, and for preparing the LFSM and the LFSMD. Prepare the LFSM and LFSMD by fortifying
two aliquots of the same field sample with an appropriate amount of MC-LR. If frozen, thaw the
samples. Mix thoroughly to homogenize the sample before distributing into the three vials. Choose a
spiking concentration such that the total microcystin and nodularin result will fall near the EC5o of the
calibration curve, and fortify at least twice the native concentration, if known. Lyse and filter the
samples, or if the samples were frozen, complete two more cycles of lysing.
9.2.5.2	Special Considerations for Ambient Water
If the concentration in ambient water is unknown, randomly select samples and fortify with
approximately 1.0 ng/L MC-LR. If the initial concentration in the randomly selected LFSMs is high, the
fortified sample result may fall outside of the range of the ELISA calibration, or fail to meet the
requirement of fortifying at a concentration at least twice the native value. In these cases, the QC results
are deemed unusable and may be discarded. However, the laboratory should attempt to collect valid
LFSM data over time for representative ambient waters.
9.2.5.3	Calculate the Mean Percent Recovery
Calculate the mean percent recovery (%R) for each LFSM and LFSMD set using the equation:
04-5)
o/0 R =	x 100%
Where,
A = mean measured concentration of the LFSM and LFSMD,
B = measured concentration in the unfortified sample, and
C = fortification concentration.
In order to obtain meaningful percent recovery results, correct the mean value of the LFSM and LFSMD
for the native concentration in the unfortified sample, even if the native value is less than the MRL.
9.2.5.4 Evaluate Recovery for Fortified Matrix
The mean percent recovery for each LFSM and LFSMD set should be greater than, or equal to, 60% and
less than, or equal to, 140% of the true value. If the percent recovery falls outside this range, and the
laboratory performance is in control for the LFBs within the same Analysis Batch, the recovery may be
matrix biased. Qualify the result for the sample from which the LFSM was prepared as "suspect-matrix".
11

-------
9.2.5.5	Calculate the RPDfor the LFSM and LFSMD
Calculate the relative percent difference (RPD) using the equation:
| LFSMD - LFSM \
RPD = 	-V X 100%
(LFSMD + LFSM)/2
9.2.5.6	Evaluate the RPDfor Fortified Matrix
The RPD for each LFSM and LFSMD set should be less than, or equal to, 40%. If the RPD falls outside this
range, and the laboratory performance is in control for the LFBs within the same Analysis Batch, the
precision may be matrix biased. Qualify the result for the sample from which the LFSMD was prepared
as "suspect-matrix".
9.2.6 Quality Control Sample (QCS)
A QCS (Sect. 3.13) must be assayed during the IDC, and then with each new lot of calibration standards.
The MC-LR used for the QCS must be obtained from a source independent from the source of the
calibration standards. Prepare the QCS in reagent water near the EC5o. A QCS supplied with the ELISA kit
may be used if these criteria are met. The percent recovery of MC-LR in the QCS must be greater than,
or equal to, 70% and less than, or equal to, 130%.
10	Calibration
10.1	Calibration Procedure
A calibration is required with each Analysis Batch. Use the concentrations stated in the kit instructions.
Do not add additional calibration levels or eliminate any levels. Laboratories may prepare calibration
standards in house; however, the number of levels and concentrations must match those in the original
kit. Each calibration standard must be added to at least two wells. The lowest calibration standard must
be at, or below, the concentration of the MRL.
10.2	Calibration Acceptance Criteria
The calibration curve is validated by evaluating the %CV of the absorbance values for the well replicates
representing each calibration level, and the correlation coefficient of the four-parameter logistic curve.
Calculate the %CV (Equation 9.2.1) for each of the paired absorbance values, including the "zero"
standard. The %CV for each pair must be less than, or equal to, 10%. However, one pair is allowed to
exceed 10% providing the %CV is less than, or equal to, 15%. The square of the correlation coefficient
(r2) of the four-parameter curve must be greater than, or equal to, 0.98.
If the calibration fails the %CV limits or r2 is less than 0.98, then the entire Analysis Batch is invalid. Assay
the samples in a subsequent Analysis Batch. Freeze the filtered samples if this Analysis Batch cannot be
completed on the same day as the original attempt. Each sample must be within the 14-day holding
time for the repeat assay.
11	Procedure
This section describes the procedure for preparing samples and processing the microtiter plate to
conduct the ELISA. Fortify QC samples prior to the lysing step (LFBs, and LFSM and LFSMD sets). If
samples were frozen, it is acceptable to fortify after this first lysing cycle. See Section 9.2 for the
frequency and procedure for preparing fortified samples.
12

-------
11.1	Lysing by Freeze-Thaw
Mix thoroughly and immediately transfer 5 to 10 mL of each field sample into a 40 mL vial to begin three
freeze-thaw cycles. If the sample was previously frozen, only two freeze-thaw cycles are needed. Smaller
vials may be used, but reduce the sample volume to less than 25% of vial capacity. Ensure samples are
completely frozen and completely thawed during each cycle. Thaw samples at approximately 35 °C in a
water bath and mix after each cycle.
11.2	Filtering the Sample
Filter 1 to 2 mL of each lysed sample into a 4 mL vial using a glass-fiber syringe filter.
11.3	Conducting the ELISA
Follow the manufacturer's instructions for adding samples and reagents to the plate. Fill two wells with
each calibration standard, field sample, and QC sample. Use Table 11 as a summary guide for
constructing the Analysis Batch to ensure that all required QC elements are included. Drinking water and
ambient water samples may be assayed in the same Analysis Batch.
11.3.1	Color Development Step
If processing plates manually, use a technique that ensures each well is incubated with substrate for
exactly the same period. To accomplish this, add stop solution in the same sequence and at the same
rate as the addition of substrate.
11.3.2	Reading the Plate
Measure the color by reading the absorbance at 450 nanometers using a microtiter plate reader.
12	Data Analysis and Calculations
12.1	Quantitation
A four-parameter logistic curve fit must be used (Sect. 3.3). Other curve-fitting models are not
permitted. Calculate the sample concentration for each well using the multipoint calibration. For each
field and QC sample, average the two concentration values from each well. Use this mean to report
sample results and to evaluate QC results against acceptance limits. Report only those values that fall
between the MRL and the highest calibration standard. Final results should be rounded to two
significant figures.
12.2	Exceeding the Calibration Range
If a result exceeds the range of the calibration curve, dilute the sample with reagent water. Based on the
estimated concentration, select a dilution factor that results in a diluted sample concentration near the
EC5o of the calibration curve. The concentration in the diluted sample must fall between the MRL and
the highest calibration standard. Analyze the diluted sample in a subsequent Analysis Batch. Incorporate
the dilution factor into the final concentration calculations. Report the dilution factor with the sample
result.
13	Method Performance
13.1 Precision, Accuracy, and LCMRL Results
Tables for these data are presented in Section 17. The single-laboratory LCMRL, based on the response
of MC-LR, is presented in Table 1. Single-laboratory precision and accuracy data are presented at low
and high concentrations within the measurement range: near the MRL (Table 2). and at a mid-range
13

-------
concentration (Table 3). Three matrixes were studied: reagent water, lake water, and drinking water
processed from lake water. Precision and accuracy results in reagent water at a concentration near the
EC5o are presented in Table 4.
13.2 Analyte Stability Studies
The stability of microcystins and nodularins was studied in both ambient water and drinking water.
Because MC-LR is only one of many possible congeners, separate experiments were designed to
evaluate the stability of MC-LR and a more representative mixture of microcystins and nodularins. These
experiments are described in the sections that follow.
13.2.1	Experimental Conditions
Each matrix was fortified with MC-LR and the representative mixture of seven congeners for a total of
four experimental conditions:
•	Drinking water fortified with 0.50 ng/L MC-LR
•	Ambient water fortified with 0.50 ng/L MC-LR
•	Drinking water fortified with 0.35 ng/L of a representative microcystin and nodularin mixture
•	Ambient water fortified with 0.35 ng/L of a representative microcystin and nodularin mixture
Ambient water was obtained from a lake known to experience algal blooms and used to supply a
drinking water treatment plant. Drinking water was obtained from a treatment plant fed by a lake
known to experience algal blooms. Drinking water samples were inoculated with microbial rich water
from an impacted surface source to incorporate an additional challenge to the efficacy of the storage
procedure. A bulk sample was prepared for each experimental condition, then distributed into separate
40 mL vials for storage. The ELISA was calibrated with MC-LR.
13.2.2	Composition of the Mixture of Congeners used in the Experiments
Seven microcystin and nodularin congeners were used: MC-LR, RR, LW, YR, LA, 3-desmethylated-MC-RR
(3-dm-MC-RR) and NOD-R. Individual congener concentrations were 0.050 ng/L. The six microcystins
were selected to represent a range of chemistries for microcystin congeners known to occur in ambient
water in terms of polarity, ionic charge and backbone methylation. In addition, MC-LW contains
tryptophan (W) which is highly susceptible to oxidation, and MC-YR contains tyrosine (Y) which is
moderately susceptible to oxidation. NOD-R was included to represent nodularin congeners.
13.2.3	Storage Conditions, Duration, and Results
All samples were preserved and stored as required in this method. Seven samples from each
experimental set were analyzed after 0, 2, 7, 15, and 22 days of storage. For drinking water, analyte
recovery at each time point was calculated by comparing the mean concentration measured in the
stored samples to the mean concentration measured in five LFBs fortified at the same concentration and
prepared on the day of assay. This approach cancels other factors contributing to the variability in the
sample response and isolates the experimental variable of storage time. For ambient water, the percent
change in analyte concentration at each time point was calculated by comparing the mean of the stored
samples to the mean of seven identical ambient samples analyzed at T-zero (zero storage time). This
alternate means of assessing stability was necessary because the ELISA response was biased by the
matrix in this particular lake water compared to calibration standards.
14

-------
Table 5 presents the results for drinking water fortified with MC-LR, and Table 6 presents the results for
ambient water fortified with MC-LR. Table 7 presents the results for drinking water fortified with the
mixture of congeners, and Table 8 presents the results for ambient water fortified with the mixture of
congeners.
13.2.4 Degradation of Microcystins and Nodularins during Storage
Microcystins and nodularins degraded during frozen storage in two drinking waters studied during
method development. These data (Table 9) are included in the method to support the 14-day holding
time recommendation.
13.2.4.1	Experimental Design
Each matrix was fortified with a representative mixture of congeners (Sect. 13.2.2) for a total of two
experimental conditions:
•	Drinking water from central Ohio fortified at 0.35 ng/L
•	Drinking water from eastern Pennsylvania fortified at 0.35 ng/L
A bulk sample was prepared for each experimental condition, then distributed into separate 40 mL vials
for storage. All samples were preserved and stored as required in this method. Seven samples from each
experimental set were analyzed after 0, 6, 9, 16, 23, 43, and 91 days of storage. Analyte recovery at each
time point was calculated by comparing the mean concentration measured in the stored samples to the
mean concentration measured in seven LFBs fortified at the same concentration and prepared on the
day of assay. The ELISA was calibrated with the same MC and NOD mixture used to fortify the drinking
water samples and the LFBs.
As an additional control, LFBs fortified at the same concentration as the study samples were stored
under identical conditions. Three vials from the frozen LFB pool were sacrificed at each sampling event
to evaluate stability in the absence of matrix.
13.2.4.2	Experimental Results
Table 9 presents the results of the analyte stability study during which degradation was observed.
Microcystin recovery in the two drinking water matrixes is near 70 percent after three weeks of frozen
storage. These losses appear to occur during frozen storage and not as a result of the process of
freezing. Note that the T-zero samples were lysed (frozen and thawed three times) and no losses were
detected. In addition, the LFB controls do not exhibit analyte loss, ruling out the sample container as a
contributing factor. These facts point to the matrix as the cause for analyte degradation during storage.
14	Pollution Prevention
For information about pollution prevention applicable to laboratory operations described in this
method, consult: Less is Better, Guide to Minimizing Waste in Laboratories, a web-based resource
available from the American Chemical Society at www.acs.org.
15	Waste Management
The Agency requires that laboratory waste management practices be consistent with all applicable rules
and regulations, and that laboratories protect the air, water, and land by minimizing and controlling all
15

-------
releases from fume hoods and bench operations. In addition, compliance is required with any sewage
discharge permits and regulations, particularly the hazardous waste identification rules and land
disposal restrictions.
16 References
Davies, Chris. Concepts. In The Immunoassay Handbook, 3rd ed.; Wild, David, Ed.; Elsevier, Ltd.: Oxford,
UK, 2005; p 119 (section titled "Assay Drift")
Fischer, Werner J. et al. Congener-Independent Immunoassay for Microcystins and Nodularins. Environ.
Sci. Techno!. 2001, 35, 4849-4856
Ohio EPA Division of Environmental Services. Ohio EPA Total (Extracellular and Intracellular)
Microcystins - ADDA by ELISA Analytical Methodology; Method 701.0 Version 2.2 (and previous
versions); Ohio EPA: Reynoldsburg, OH, November 2015.
Maciel, Robert J. Standard Curve Fitting in Immunodiagnostics: a Primer. Journal of Clinical
Immunoassay. 1985, Vol. 8, 98-106.
Sasaki, Diane and Mitchell, Robert A. How to Obtain Reproducible Quantitative ELISA Results. Oxford
Biomedical Research website. http://www.oxfordbiomed.com/sites/default/files/ELISAArticle.pdf
(accessed August 2016).
U.S. EPA. Statistical Protocol for the Determination of the Single-Laboratory Lowest Concentration
Minimum Reporting Level (LCMRL) and Validation of Laboratory Performance at or Below the
Minimum Reporting Level (MRL); EPA 815-R-05-006; Office of Water: Cincinnati, OH, November 2004.
U.S. EPA. Technical Basis for the Lowest Concentration Minimum Reporting Level (LCMRL) Calculator;
EPA 815-R-11-001; Office of Water: Cincinnati, OH, December 2010.
16

-------
17 Tables, Figures, and Method Performance Data
Data presented in this section were obtained using Abraxis, Inc., "Microcystins/Nodularins (Adda) ELISA"
kits, Part Number 5200110H. Samples were added to the 96-well plate using manual pipettes. Antibody,
enzyme conjugate, substrate, and stop solution were added using an electronic stepping pipette.
Absorbance was measured using an 8-channel microplate reader (Abraxis, Inc., Model 4300).
Table 1. Lowest Concentration Minimum Reporting Limit for MC-LR
Analyte
Fortification Levels, |ig/L
LCMRL
MC-LR
0.0, 0.10, 0.15, 0.20, 0.25, 0.30a, 0.35, 0.40
0.26
a' The authors confirmed the MRL at 0.30 ng/L using the statistical calculations in Section 9.1.3.
Table 2. Accuracy and Precision near the MRL for MC-LR, Seven Replicates
Matrix
Fortification,
Hg/L
Native
Concentration, ng/L
% Recovery
%RSD
Reagent water
0.30
n/a
105
8.8
Drinking water3
0.30
0.037
117
12
Ambient waterb
0.30
0.063
110
6.3
a' Central KY, total organic carbon = 3.0 mg/L, pH = 7.55, free chlorine = 0.89 mg/L, total chlorine =
1.17 mg/L, hardness = 126 mg/L as CaC03.
b' Southern Ohio lake, pH = 7.97
Table 3. Accuracy and Precision for MC-LR at a Mid-Range Concentration
Matrix
Fortification,
Hg/L
Native
Concentration, ng/L
% Recovery
%RSD
Reagent water, n=10
2.2
n/a
93.4
8.6
Drinking water3, n=7
2.2
0.037
107
5.1
Ambient waterb, n=7
1.5
0.063
96.8
9.1
a' Central KY, total organic carbon = 3.0 mg/L, pH = 7.55, free chlorine = 0.89 mg/L, total chlorine =
1.17 mg/L, hardness = 126 mg/L as CaC03.
b' Southern Ohio lake, pH = 7.97
Table 4. Accuracy and Precision near the EC50 for MC-LR in Reagent Water, Seven
Replicates
Matrix
Fortification, ng/L
% Recovery
%RSD
Reagent water
0.50
115
5.4
17

-------
Table 5. Stability of MC-LR Fortified at 0.50 |ig/L in Drinking Water3
Storage Time, days:
0
2
7
15
22
% Recovery13:
91.8
107
98.9
109
110
a- Central KY, total organic carbon = 3.0 mg/L, pH = 7.41, free chlorine = 1.25 mg/L, total chlorine =
1.60 mg/L, hardness = 134 mg/L as CaC03.
b- % Recovery = (mean concentration of seven fortified tap water samples stored for X days) / (mean
concentration of five Laboratory Fortified Blanks at 0.50 ng/L prepared on the day of assay) * 100%;
where X is the storage time.
Table 6. Stability of MC-LR Fortified at 0.50 |ig/L in Ambient Water3
Storage Time, days:
0
2
7
15
22
Units:
Hg/L
% Changeb
% Change
% Change
% Change
Results:
0.689
-18.6
14
1.5
-12
a- Southern Ohio lake, pH = 7.97
b- % Change from initial, analyzed concentration = (mean concentration of seven samples stored for X
days - mean concentration of seven samples analyzed on Day-zero) / (mean concentration of seven
samples analyzed on Day-zero) * 100%; where X is the storage time.
Table 7. Stability of a MC and NOD Mixture Fortified at 0.35 |ag/L in Drinking Water3
Storage Time, days:
0
2
7
15
22
% Recovery13:
108
109
104
123
109
a- Central KY, total organic carbon = 3.0 mg/L, pH = 7.41, free chlorine = 1.25 mg/L, total chlorine =
1.60 mg/L, hardness = 134 mg/L as CaC03.
b' % Recovery = (mean concentration of seven fortified tap water samples stored for X days) / (mean
concentration of five Laboratory Fortified Blanks at 0.35 ng/L prepared on the day of assay) * 100%;
where X is the storage time.
Table 8. Stability of a MC and NOD Mixture Fortified at 0.35 |ag/L in Ambient Water3
Storage Time, days:
0
2
7
15
22
Units:
Hg/L
% Changeb
% Change
% Change
% Change
Results:
0.673
-19.5
-0.42
13
-11
a' Southern Ohio lake, pH = 7.97
b- % Change from initial, analyzed concentration = (mean concentration of seven samples stored for X
days - mean concentration of seven samples analyzed on Day-zero) / (mean concentration of seven
samples analyzed on Day-zero) * 100%; where X is the storage time.
18

-------
Table 9. Stability of a MC and NOD Mixture Fortified at 0.35 pig/L in Drinking Water3
Storage Time, days:
0
6
9
16
23
43
91
Drinking water processed from
lake waterb, % Recovery:
107
89.6
86.9
85.8
72.6
75.3
63.7
Drinking water processed from
ground waterc, % Recovery:
99.6
78.0
88.6
77.6
70.1
71.4
72.4
Laboratory Fortified Blanks
(LFBs) stored identically to
matrix samplesd:

119
94.5
118
105
115
110
a' % Recovery = (mean concentration of seven fortified tap water samples stored for X days) / (mean
concentration of seven Laboratory Fortified Blanks at 0.35 ng/L prepared on the day of
assay) * 100%; where X is the storage time.
b' Central OH, total organic carbon = 1.2 mg/L, pH = 9.77, total chlorine = 1.68 mg/L, hardness = 127
mg/L as CaC03.
c" Eastern PA, total organic carbon = 0.69 mg/L, pH = 7.37, Hardness = 193 mg/L as CaC03.
d- % Recovery = (mean concentration of three LFBs stored for X days) / (mean concentration of seven
Laboratory Fortified Blanks at 0.35 ng/L prepared on the day of assay) * 100%; where X is the
storage time.
Table 10. Initial Demonstration of Capability (IDC) QC Requirements
Method
Reference
Requirement
Specification
Acceptance Criteria
9.1.1
Demonstration of
precision and
accuracy
Lyse and assay 7 replicate Laboratory
Fortified Blanks (LFBs) at 0.50 ng/L.
Percent relative standard
deviation <15%. Mean
percent recovery >70%
and <130%.
9.1.2
Demonstration of
acceptable system
background
Lyse and assay 5 Laboratory Reagent
Blanks (LRBs) distributed throughout
a plate.
MC concentration must
be less than one-half the
Minimum Reporting
Level (MRL) in each LRB.
9.1.3
MRL confirmation
Fortify and analyze 7 replicate LFBs
at the proposed MRL concentration.
Confirm that the Upper Prediction
Interval of Results (PIR) and Lower
PIR meet the recovery criteria.
Upper PIR <150%
Lower PIR >50%
9.1.4
Quality Control
Sample (QCS)
Prepare a QCS near the EC5o with
MC-LRfrom a source independent
from the calibration standards.
Percent recovery >70%
and <130% of the true
value
19

-------
Table 11. Analysis Batch QC Requirements
Method
Reference
Requirement
Specification and Frequency
Acceptance Criteria
10
ELISA Calibration
Use kit-recommended levels and
concentrations. Two well replicates
(Sect. 3.1.1) per standard.
%CV of absorbance <10%;
<15% allowed for 1 pair.
r2> 0.98.
9.2.1
Well replicates
Assay field and QC samples in two
wells.
Sample invalid if %CV of
absorbance values >15%
9.2.2
Laboratory
Reagent Blank
(LRB)
Lyse one LRB per Analysis Batch.
Assay in duplicate on opposite sides
of the plate.
MC/NOD concentration must
be less than one-half the
Minimum Reporting Level
(MRL) in each LRB.
9.2.3
Low Calibration
Verification
(Low-CV)
Calibration standard at, or below,
the MRL concentration. One per
Analysis Batch.
Percent recovery >50% and
<150% of the true value
9.2.4
Laboratory
Fortified Blank
(LFB)
Reagent water fortified near the
EC5o. Lyse and assay 2 per Analysis
Batch.
Percent recovery for each
LFB >60% and <140% of the
true value
9.2.5
Laboratory
Fortified Sample
Matrix (LFSM)
and LFSM
Duplicate
Fortify near the EC5o and twice
native concentration. One set in
Analysis Batches containing
drinking water; two if 20 or more
field samples. One set in Analysis
Batches containing ambient water;
two if 20 or more field samples.
Mean percent recovery of
LFSM and LFSMD pair >60%
and <140%. Relative percent
difference (RPD) <40%.
Qualify results for samples
failing these limits as
"suspect-matrix".
9.2.6
Quality Control
Sample (QCS)
Assay 1 QCS for each new lot of
calibration standards. Prepare the
QCS near the EC5o with MC-LR from
a source independent of the
calibration standards.
Percent recovery >70% and
<130% of the true value
20

-------
1.5-
1.4-
1.3-
1.2-
50% response
0.9-
0.8-
0.7-
0.6-
o
LO
0.5-
0.4
.15
.4
1
2
5
cone MC (ng/mL)
Figure 1. ELISA calibration curve with calculated EC50. Note: x-axis uses a logarithmic
scale.
21

-------