The clinical laboratory uses commercial products with limited shelf life or certain
expiry dates. As such, laboratories constantly need to change reagents, a process
which is usually transparent to health care providers. Reagent lot-to-lot variation
can cause a significant analytical error, with changes that may generate a significant
shift in patient data. The Clinical and Laboratory Standards Institute (CLSI) document
for reagent lot verification of performance is primarily targeted for clinical chemistry
tests, which have more robust and standardized performance characteristics than most
hemostasis-based assays.[1] Other guidance for laboratories includes documents provided by international regulatory
organizations such as the International Organization for Standardization (ISO) 15189
document[2]
[3] or regional regulatory agencies such as the Food and Drug Administration (FDA)[4] and the Centers for Medicare and Medicaid Services[5] in the United States. However, these regulatory agencies provide little details
or recommendations about how to perform new lot verification of reagents. ISO 15189
(section 5.3.2.3 Reagents and consumables) states that “Each new formulation of examination
kits with changes in reagents or procedure, or a new lot or shipment, shall be verified
for performance before use, or before release of results, as appropriate. Consumables
that can affect the quality of examinations shall be verified for performance before
placing into use..”[2] In 2018, the FDA updated their guidance document for industry, noting that “partial
validation” can be used to “evaluate modifications of already validated bioanalytical
methods” and can be limited to assessing a single intra-assay accuracy and precision
evaluation although recognizing “nearly full validation” may also be suitable.[4] The Clinical and Laboratory Improvement Amendments (CLIA, section §493.1255 Standard:
Calibration and calibration verification procedures) specifies calibration verification
procedures if a “complete change of reagents for a procedure is introduced, unless
the laboratory can demonstrate that changing reagent lot numbers does not affect the
range used to report patient test results, and control values are not adversely affected
by reagent lot number changes.” but also indicates (section §493.1256: Standard: Control
procedures) to “Perform control material testing as specified in this paragraph before
resuming patient testing when a complete change of reagents is introduced….”[5] No other laboratory guidance related to reagent changes made in a coagulation laboratory
is addressed, including reagents or test platforms that are not calibrated. The laboratory
accrediting agency College of American Pathologists (CAP) require “New reagent lots
and shipments are checked against old reagent lots or with suitable reference material
before or concurrently with being placed in service.”[6] providing additional guidance for acceptable samples or material used for quantitative
and qualitative tests. However, this guidance is limited to recommendations for a
number of samples to be tested, statistical acceptability, or other requirements.
CAP has additional requirements for the confirmation of specific methodology calibrations
for verifying the analytical measurement range (AMR), which will be described later.
In the clinical coagulation laboratory, the commonly used testing principles are either
chronometric (clot based), chromogenic, immunologic, or latex particle agglutination/aggregation.[7] Coagulation test reagents may be either (1) reagents for assays that do not require
calibration (e.g., activated partial thromboplastin time [aPTT] or lupus anticoagulant
[LAC] screen), with results reported in raw units (e.g., seconds); (2) reagents for
tests that require calibration (e.g., D-dimer or coagulation factor assays), with
results reported as quantitative values usually determined from a calibration curve
(e.g., IU/dL); (3) reagents for assays where a reported result is either qualitative
(i.e., fibrin monomer), semiquantitative (i.e., some heparin-induced thrombocytopenia
assays), or used as an ancillary reagent (normal pooled plasma, NPP); (4) internal
quality control (IQC) that may have assigned test values (and acceptable result ranges)
or require local determination of target mean and ranges, and (5) calibrator materials
with assigned values. Molecular diagnostics or flow cytometry are beyond the scope
of this document and will not be addressed.
A particular challenge for hemostasis laboratories is the verification of performance
for new lots of prothrombin time (PT) or aPTT reagents, as these tests are often the
first-line investigations performed on patients with query hemostatic challenges.
They have multiple indications for use, including screening for clotting factor deficiencies
and measuring the efficacy of treatment, which may be either anticoagulation or replacement
therapy. Presumably, tests that require calibration, such as factor assays, may mitigate
any biases associated with the variability of new reagent lots, especially if the
same calibrator material is used for multiple lots of reagents. However, even with
calibrated tests, the instrument limitations for thresholds (e.g., lower limit of
quantitation) may fluctuate between reagent lots. Other potential variables that may
introduce the risk of introducing testing bias, such as poor instrument precision,
are beyond the scope of this document.
Regional and international regulatory agencies identify that the verification of new
reagents should be performed, but little guidance is provided for laboratories.[2]
[3]
[5] The purpose of this document, from the International Council for Standardization
in Haematology (ICSH), is to provide laboratory guidance for the verification of the
performance of new reagent lots used for hemostasis testing. These ICSH Working Party
consensus recommendations are based on good laboratory practice, regulatory recommendations,
evidence emerged from scientific publications, and expert opinion. The resources (e.g.,
laboratory staffing, sample availability, and financial constraints) required for
new lot reagent verification of performance will vary, and thus, we will provide tiered
recommendations (minimal and optimal).
This document is not intended to supersede any national, regional, local, or institutional
requirements. Laboratories that modify regulatory-approved reagents or reagents that
are used outside their intended use may constitute a laboratory-developed test (LDT),
which may have more stringent requirements for performance verification. The verification
of performance for new reagents or instrument platforms has already been addressed
in previous ICSH documents and, therefore, will not be addressed here.[8]
[9] If a new reagent supplied by the same manufacturer is introduced, this should be
treated as a new method, needing more robust evaluation and performance verification.
New Lot Reagent Selection—General Guidance
The selection of new lot reagent material should be predicated on both laboratory
needs and manufacturer capacity. Manufacturers generally follow their own described
quality control procedures to ensure reagent performance. Nevertheless, the laboratory
should have the option of choosing which lot may give optimal performance for their
given patient population. With each new reagent batch, the instructions for use (IFU)
should be reviewed for any changes to the performance claims of reagents. Any such
changes require evaluation by the laboratory and may impact lot-to-lot variation.
Considerations for laboratories should include the expiry date and manufacturer inventory
for providing the longest dating possible for clinical use before another new lot
of reagents is required. For some automated analyzers, creating duplicate or shadow
test protocols would allow laboratory professionals to program both existing lot reagents
and new lot reagent tests concurrently on patient samples that would expedite method
comparison and avoid potential bias associated with delays in new lot reagent testing.
For laboratories with multiple analyzers, designating a single analyzer could be considered,
with results from each analyzer correlated to a predicate device. Such an approach
may ensure consistency across all analyzers and institutions. Given the variable ex-vivo
stability of coagulation factors, treatments, etc., the comparison of new lot reagents
to existing lot reagents should occur concurrently (within 1 hour).
Lastly, each laboratory should develop a written protocol, process, or plan for performance
verification of new reagent lots. Testing of new lot reagents should be performed
to mimic laboratory practice, preferentially encompassing different shifts and different
testing personnel and laboratory environment, although this process may be unwieldy
in networks with a vast number of testing sites and/or instruments. Creating a biorepository
of normal and abnormal samples may expedite the evaluation process, although surrogate
(contrived) samples and external quality assurance material can also be used. For
hemostasis laboratories that provide an aPTT-based heparin therapeutic range (HTR),
it may be prudent to coordinate with reagent manufacturer to evaluate two different
reagent lots to avoid potential HTR changes, which will likely have downstream effects
(e.g., heparin dose changing algorithms). This approach would provide a choice for
selecting which lot to be selected based on recommended performance criteria for HTR
determination.[10] The use of statistical methods (i.e., correlation, regression, and slope) is certainly
appropriate, but whether differences between reagent lots are “clinically significant”
may be a more important measure for patient care.[7] There is conflicting guidance as to the acceptability of simultaneous changing of
IQC material with new lot reagents.[1]
[2] Whether to change IQC material concurrently with new lot reagents or modifying existing
IQC thresholds with new lot reagents may be restricted by regional regulatory authorities.
However, IQC testing (precision assessment) should be part of the new lot verification
of performance, whether changed concurrently or not.
The local measurement of uncertainty is a valuable tool to use as a threshold for
comparing differences between results on new and existing lot numbers. Results falling
outside the acceptable limits may be significant as they exceed the normal day-to-day
biological and analytical variation. These should be assessed on an individual basis,
with regard to reference range cut-offs and clinical decision points.
General Recommendations
-
New reagent lots based on the longest expiry date, especially for screening tests
(PT, aPTT, LAC, and thrombin time [TT]), which may require more robust verification
of performance should be selected.
-
Appropriate PT and aPTT reagents with the consideration of PT ISI (International Sensitivity
Index), factor sensitivity, LAC sensitivity or insensitivity, as well as heparin sensitivity
should be determined.
-
For automated analyzers, duplicate test protocol(s) should be generated to allow for
concurrent analysis of new lots of reagents with the existing lot when testing patient
samples.
-
When more than one analyzer is regularly used in a single diagnostic laboratory for
patient testing, verification should be performed across all instruments.
-
For laboratories with multiple analyzers across multiple institutions, new reagent
lots should preferably be verified on each analyzer.
-
The testing of new and existing lots of reagents should be performed concurrently
(optimally, within 1 hour).
-
Reagent verification of performance should be performed by at least two different
testing personnel over at least 5 working days.
-
Precision assessment should be part of the new lot verification of performance, whether
IQC is changed concurrently or not.
-
Given the fluctuations in throughput and temperature over the course of a day, patient
comparison and internal QC testing should not be limited to specific times in the
day; rather, it should preferably be representative of the operating times of a laboratory,
standardizing any environmental influence.
-
Each laboratory should develop a protocol, process, or plan for the performance of
verification of new reagent lots that mimics laboratory practice and defines desired
statistical outcomes.
-
The plan should identify the intended use of each test to ensure identification of
the appropriate patient population and samples, reflective of a wide spectrum of conditions
(e.g., mild-severe disease, anticoagulant status, etc.) that shall be used for comparison
testing, also reflecting the performance characteristics of the reagents.
-
Samples should represent results across the reportable range, as well as samples from
patients under specific conditions (LAC, factor deficient, heparin, liver disease,
low fibrinogen, high fibrinogen, etc.).
-
Surrogate samples, including sample pools or contrived, can be used for new lot reagent
comparisons. If testing is to be delayed, or cannot be performed within optimal time
limits, frozen storage of plasma aliquots should be considered.
-
New lots of more stable reagents such as buffers, diluents or calcium chloride, saline
and water, and other consumables that can affect the quality of examinations should
be verified for performance before placing into use.2 Readership should consult regulatory
authorities for clarification and local regulatory requirements.
-
Testing of residual or banked external quality assurance (EQA) samples would be recommended
if available and can be used to demonstrate new lot accuracy.
-
Each laboratory must have a documentation system in place to address performance verification
of new reagent lots. This system must have director or designate approval prior to
use and shall be maintained in accordance with local accreditation or regulatory requirements.
This system must include a general performance policy for reagent lot verification,
including specific requirements for each measurand (or groups of assays by method)
and new lot acceptability requirements.
Noncalibrated Coagulation Tests
There are several tests in the coagulation laboratory that are not calibrated, including,
but not limited to, the PT with international normalized ratio (INR), aPTT, TT, reptilase
time, dilute Russell's viper venom time, other clotting tests that are reported in
seconds, platelet aggregation or function studies, or point-of-care devices such as
the activated clotting time (ACT) or thromboelastography or rotational thromboelastometry.
Tests that are not calibrated should be scrutinized for both the intended use and
population at each testing site, to assure that proper patient populations are considered
throughout the verification process. Tests such as the PT and aPTT have multiple clinical
uses (clotting factor deficiency assessment, anticoagulant monitoring, or replacement
therapy monitoring) and likely the complexity of the verification process will be
secondary to the local intended use. Laboratories should incorporate into their evaluation
a mix of abnormal samples (including surrogate samples when patient samples are unavailable)
that reflect the expected patient population to be tested. Tests with limited or specific
applications, such as LAC, may require specific patient populations for method comparison,
and each laboratory should consider obtaining formal approval by accredited bodies
(i.e., local ethics committee) for storing abnormal samples during the course of routine
testing aimed at creating a sufficient bank of abnormal plasmas to expedite performance
verification of new reagent lots. If multiple sites need to be correlated, plasma
pools may be created combining patient samples with similar results, conditions, or
therapies.
Prothrombin Time Reagents
The PT is used for evaluating factor abnormalities (deficiencies or dysfunctional
proteins) in the extrinsic and common pathway (factors II, V, VII, X, and fibrinogen),
monitor vitamin K antagonists (VKA) therapy, such as warfarin, or monitor efficaciousness
of reversal (e.g., prothrombin complex concentrates) or factor replacement therapy.
The INR was designed to standardize the monitoring of VKA anticoagulation[11] and may not be suitable for other indications, such as liver disease,[12] although sometimes the laboratory may report both the PT and INR values for any
given patient. With new reagent lots, the mean normal prothrombin time (MNPT) should
also be determined for the new lot, using 20 fresh samples from healthy individuals,[13] although the World Health Organization publications for INR guidance have indicated
the use of frozen citrated normal plasma for MNPT determination[14] and other methods have been described and may be considered after local validation.[15]
[16]
[17]
[18] After the calculation of MNPT, it may be prudent to have at least two experienced
laboratory staff review the INR calculations and instrument test protocol modifications
to assure manual entry accuracy. Afterward, INR performed over a range of INR targets
is necessary to ensure no significant change in INR between reagent lots. Prior to
initiating the verification of the reagent lot, verify the manufacturer's correct
ISI using certified plasma or other described methods as appropriate.[15]
[16]
[17]
[18] Commercially prepared frozen normal donor citrated plasma or lyophilized plasma
material may also be suitable for MNPT if the citrate concentration used for donor
collection is the same concentration used locally. Patient samples are preferred when
comparing different reagent lots to avoid issues with the commutability of IQC materials.
To evaluate the INR performance of the new thromboplastin lot, a minimum of 20 samples
of patients on stable VKA therapy should be tested in parallel and evaluated with
new and existing reagent lots, unless the laboratory uses an alternate method that
has been locally validated.[15]
[16]
[17]
[18] To evaluate the PT performance of a new thromboplastin lot, samples should be selected
to include normal subjects, patients on oral VKA therapy representing both the therapeutic
and nontherapeutic ranges, and if possible, at least samples of 20 patients with conditions
affecting the extrinsic and common pathway should be included (e.g., patients with
history of liver disease, consumptive coagulopathies [e.g., disseminated intravascular
coagulation; DIC], low fibrinogen (hypofibrinogenemia), or dysfunctional fibrinogen
(dysfibrinogenemia or hypodysfibrinogenemia) and tested with new and existing reagents.
If possible, samples from neonates or newborns may also provide sources for an abnormal
PT. If new lot PT reagent comparison statistical thresholds are not met, then assessing
factor sensitivity may be useful, where samples of known factor concentration, optimally
a commercial calibrator, are diluted in respective factor-deficient plasma and PT
testing is then performed. Alternatively, and if available, characterized patient
samples could be used. The factor sensitivity determination should be performed using
surrogate or patient samples having approximately 30 to 60% factor activity to help
determine at what factor level the PT becomes prolonged. Evaluation of new lot PT
reagents using samples from patients taking direct oral anticoagulants (DOACs) are
not recommended, given their variability for assessing this class of oral anticoagulants.[19] The use of INR calibrants or material that can verify the local INR/ISI used for
testing should be considered if available and applicable to the reagent and instrument
platform being used.
Recommendations for verification of performance of new lot PT testing:
-
Manufacturer ISI results should be verified using appropriate INR calibrators or equivalent
material or an alternative method that has been locally validated. If ISI verification
fails,
-
Ensure that correct ISI and MNPT have been properly recorded.
-
If verification still fails, consult with the reagent manufacturer for resolution
or seek a new reagent lot.
-
The MNPT for the new lot number using either 20 fresh or frozen samples from healthy
adult individuals should be determined. The use of alternative plasma sources or an
alternative method can be considered if locally validated.
-
The INR/PT reference interval (RI) should be verified using 20 samples collected from
nonanticoagulated patients using the CLSI-approved transference method[20] or an alternative method that has been locally validated. Failure to achieve statistical
limits may require a new RI determination.
-
At least 20, but ideally at least 40, normal and described abnormal samples should
be tested concurrently with existing and new lot reagents for both PT (in seconds)
and INR (for VKA therapy samples if reported) testing. Surrogate samples (i.e., pooling
or modified) are acceptable when patient samples are unavailable, but commutability
should be considered in this situation.
-
For patient comparisons, the testing should be performed within stability limits described
for PT[21] or longer if locally validated.
-
Verification should preferably be performed across all analyzers used for diagnostic
testing, unless there are alternate procedures in place to assess comparability of
performance.
-
Regression acceptability should be determined locally; however, linear correlation
(e.g., Spearman's) and regression (e.g., Passing and Bablok) are recommended, with
acceptable criteria of >0.95 (coefficient of correlation) and slope between 0.90 and
1.10, respectively.[22] Bias estimation may be a useful determination, with expected bias similar as observed
during the method verification of performance. For normal samples, at least 90% should
be within current or manufacturer-defined RI to verify the RI.
-
MNPT may also require changing for those sites reporting INRs.[17]
Note: It is good laboratory practice to always establish a new MNPT with a new lot
of PT reagent when reporting INR. For laboratories reporting the prothrombin index,
verification of the reporting method with each new reagent lot is required.
-
If statistical thresholds fail, then factor sensitivity testing may be useful to identify
factor sensitivity differences between reagent lots.
-
Direct INR IQC values (targets and limits) should be assessed for the new lot number
using appropriate quality control material.
Activated Partial Thromboplastin Time Reagents
Similar to the PT, the aPTT is used for evaluating factor abnormalities in the intrinsic
and common pathway (factors II, V, VIII, IX, X, XI, XII, other contact factors and
to a lesser degree, fibrinogen), monitoring unfractionated heparin (UFH) therapy,
monitoring direct thrombin inhibitor (DTI) therapy (e.g., argatroban), or monitoring
the efficaciousness of some therapies (e.g., factor VIII replacement therapy or reversal
agents such as protamine sulfate).
To evaluate abnormal samples for aPTT testing, these could be collected from patients
on heparin therapy or liver disease, consumptive coagulopathies, known intrinsic clotting
factors deficiencies, or neonates, which should then be tested with new and existing
reagent platforms. As with the PT, the aPTT has variable sensitivity for assessing
DOACs, and thus, the use of DOAC samples for new reagent lot comparisons is usually
not recommended.[19] If the aPTT is used for UFH monitoring, then at least 20 samples from patients on
active UFH treatment (not prophylaxis) are required for HTR determination.[10] For HTR determination, avoid samples with marked inflammatory conditions, as raised
fibrinogen and factor VIII may factitiously repress the aPTT clotting time while on
UFH therapy. Similarly, avoid the use of samples showing aPTT prolongation compounded
by secondary factors (e.g., factor deficiencies, on concomitant VKA therapy).
Parenteral DTI monitoring is likely to be predicated on baseline aPTT and target ratios
after the initiation of infusion and is not the same therapeutic range as determined
for UFH.[23] Samples collected from normal subjects or samples that are within normal limits
of existing reagent should also be incorporated in the method comparison. aPTT reagents
that are used for purposes of LA testing, which may or may not be the primary screening
aPTT reagent, would require additional population-defined samples to be evaluated
(see LA reagent section below).[24] If the aPTT comparison statistical thresholds are not met, then assessing factor
sensitivity may be considered in the same fashion as described for the PT, although
alternative strategies have also been described.[25] The aPTT factor sensitivity determination should be performed using multiple surrogate
samples ranging from approximately 30 to 60% factor activity. For laboratories that
perform testing for diagnosis and monitoring of hemophilia, verification of new lot
aPTT reagent sensitivity to factors VIII, IX, and possibly XI should also be considered,
regardless of whether statistical thresholds are met.
Recommendations for verification of performance of new lot aPTT testing:
-
The aPTT RI should be verified using 20 samples collected from nonanticoagulated patients
using the CLSI-approved transference method.[20]
[26] Failure to achieve statistical limits (at least 90% of the results using new lot
reagent on samples collected from nonanticoagulated normal subjects are within the
existing normal range) may require a new RI determination.
-
At least 20, but ideally at least 40, normal and abnormal samples (as described above)
should be tested concurrently (within 1 hour) with existing and new lot reagents for
aPTT (in seconds).
-
If the aPTT is used for monitoring UFH infusion, then a minimum of 20 samples from
patients on heparin therapy are required for HTR verification.[10]
-
Acceptability and generation of HTR is outside the scope of this document, but additional
guidance is available.[10]
[27]
-
HTR verification should preferentially be performed across all analyzers used for
diagnostic testing.
-
Samples should be used on patients with UFH treatment doses only, with no concomitant
additional anticoagulant therapy.
-
Normal to near-normal INR should be used, with no more than two samples from the same
patient.
-
Failure to meet acceptable thresholds would require new HTR determination comparing
aPTT times to anti-FXa (activated factor X) measurements.
-
HTR changes should be communicated as soon as possible to the appropriate health care
units, to allow modifications of UFH dosing protocols. The laboratory must provide:
-
For patient sample comparisons, the statistics used should be determined locally;
however, linear correlation (e.g., Spearman's) and regression (e.g., Passing and Bablok)
is recommended with acceptable criteria of >0.95 (coefficient of correlation) and
slope between 0.90 and 1.10, respectively.[22] A bias estimation may be a useful adjunct.
-
IQC values (targets and limits) should be assessed for the new lot reagents using
either assayed or unassayed control material.
Thrombin Time Reagents
The TT test is similar to fibrinogen determination, in that an exogenous thrombin
reagent is added to a patient plasma sample. For TT, the test sample is typically
neat (undiluted) or slightly diluted, and the thrombin reagent has a relatively low
concentration (approximately 2–10 NIH U/mL), whereas the thrombin reagent used for
fibrinogen testing will have higher concentrations of thrombin (approximately 35–100
NIH U/mL), and additionally use diluted patient plasma.[28] The TT measures the clotting time after thrombin addition, thus evaluating the conversion
of fibrinogen to fibrin or clot formation. Abnormal TTs are seen in patients with
low or dysfunctional fibrinogen (hypofibrinogenemia and dysfibrinogenemia, respectively,
or combination of both) or drugs that will inhibit thrombin such as heparins, and
parenteral or oral DTIs. Abnormal TTs can also be seen in patients with DIC, where
elevated fibrin(ogen) degradation products can inhibit fibrin polymerization thus
preventing clot formation.
Recommendations for verification of performance of new lot TT testing:
-
At least 20 normal and 20 abnormal samples (low fibrinogen or on UFH therapy) should
be tested concurrently with existing and new lot reagents for TT.
-
Correlation (e.g., Spearman's) and regression (e.g., Passing and Bablok) acceptability
should be >0.95; slope between 0.90 and 1.10, respectively, and bias estimation.[22]
-
For normal samples, at least 90% should be within current or manufacturer-defined
RI.
-
Failure to achieve statistical limits may require a new RI determination.
Normal Pooled Plasma
NPP is commonly used in the coagulation laboratory for the performance of PT or aPTT
mixing studies, as well as for normalizing LAC test results. The manufacturers of
NPP do not usually provide a certificate of analysis detailing factor activity levels
for each lot but rather assure that at least 80% (80 U/dL or 0.80 IU/mL) factor levels
are present. NPP used for more targeted applications (i.e., factor VIII Bethesda assay)
should be assessed prior to use to determine factor level(s). While an optimal level
(near 100%, 100 U/dL or 1.00 IU/mL) is recommended,[29] each laboratory should determine their optimal NPP performance for the given indication.
Recommendations for new lot NPP assessment:
-
For laboratories that use NPP for mixing studies, at least three known inhibitor (e.g.,
samples with LAC or factor inhibitors) and three known factor deficient (e.g., samples
from patients on VKA or with specific factor deficiency) samples should be tested.
Lupus Anticoagulant Reagents
LAC are part of a heterogenous group of autoantibodies directed against anionic phospholipids
(PL).[31] First described in a patient with systemic lupus erythematosus (SLE), these autoantibodies
are associated with thrombotic risk and pregnancy complications.[32] LAC screening is part of a panel of antiphospholipid antibody measurements used
to identify patients suspected of having the antiphospholipid antibody syndrome (APS)[32] and also used for SLE diagnosis.[33] LAC is a common cause for a prolonged aPTT (and sometimes PT) that often demonstrates
as an inhibitor (noncorrection of mixing studies). There are several types of LAC
assays that could be used, including those based on snake venoms (i.e., Russell viper,
ecarin, and taipan), hexagonal PL configurations, and neutralization of antibodies
using excess PL, platelets, or platelet derivatives.[31]
[32] Given the heterogeneity of these antibodies, combined with the lack of test standardization,
the methods used for screening the presence of LAC are PL type and concentration dependent,
which requires the prolongation of a screening method to proceed with establishing
LAC characteristics of inhibitor and PL dependence.[31]
[32] The use of integrated systems (screening and confirmation from the same method or
source) and calculating normalized ratios for results reporting may aid in mitigating
new reagent lot differences. Consult the manufacturer's IFU for guidance as some IFUs
indicate reference ranges, cutoffs, or both, but local verification of performance
for these assays has already been described.[9] Appropriate determination or confirmation (transference) of manufacturer cutoff
value (seconds or ratios) for any LAC method is critical for diagnostic accuracy.
Comparison studies should include samples near or at decision or cutoff values. Although
samples with LAC are recommended to be reported as positive (not semiquantitative
such mild or strong),[32] assessing a breadth of abnormal samples should be considered. Laboratories that
normalize LAC test results must incorporate the material used (i.e., NPP, normal mean,
and mixed normal mean) concurrently during new lot reagent verification of performance.
It may be beneficial to concurrently evaluate new lots of NPP during new lot verification
of LAC reagents.
Given the heterogeneity of these antibodies and the lack of international standards
for this assay, the statistical thresholds (as described below) may not be met. As
such, the use of percentage agreement (total, negative, and positive) could be considered,
with an optimal goal of 100% agreement. Failure to achieve 100% may suggest a required
change in the cutoff value. New lot numbers for LAC confirmatory reagents should be
tested to confirm that the new lot number corrects the abnormal control, thereby demonstrating
the PL-dependent nature of LAC. However, if the laboratory has the capacity to do
so, it might be preferred to apply similar recommendations to those of the screen
test reagent described below.
Recommendations for verification of performance of new lot LA assays:
-
At least 20, but ideally at least 40 patient, samples being investigated for possible
APS or SLE should be tested concurrently with existing and new lot reagents for the
LAC screening test (in seconds). For patient comparisons, the statistics used should
be determined locally; however, liner regression is recommended, with the acceptable
criterion of >0.95 (coefficient of correlation) and 0.90 to 1.10 (slope).[22] A bias estimation may be a useful adjunct.
-
At least a minimum of 20%, but ideally 40%, of the patient comparison samples should
be LAC positive and the full lupus LAC panel should be performed (e.g., screen, mix,
and confirm).
-
IQC values (targets and limits) for both screening and confirmatory reagents should
be assessed for the new lot number using either assayed or unassayed control material.
-
For laboratories that provide normalized ratio reporting using the mean of the normal
RI, at least 20 samples collected from normal donors should be used to help establish
appropriate normalized cutoffs.
-
Frozen normal donor samples may be suitable if citrate concentration is the same as
used for patient testing.
-
Alternative strategies such as the use of NPP for normalizing results can be used
once locally validated.
Qualitative (Screening) Assays
There are several coagulation assays that provide qualitative (positive or negative)
results (e.g., some heparin induced thrombocytopenia tests) or results that are associated
with a given diagnostic threshold (e.g., some activated protein C resistance, APCR
assays). These assays may or may not have kit-provided quality control material.
Recommendations for qualitative (screening) assays:
-
At a minimum, testing of a positive and negative sample that is independent of the
control material provided (if applicable)
Note: Regulatory agencies may not accept kit-provided control material as acceptable
performance verification of new reagent lots. For these laboratories, verification
of performance can be achieved by:
-
Refer to the manufacturer's IFU for additional testing recommendations.
Slide Agglutination Assays
Slide agglutination assays are immunologic methods that use either antibody-complexed
latex bead (latex immunoassays) or coated red blood cells (hemagglutination assays)
that are performed on reusable glass/ceramic (reusable) or plastic/treated paper (disposable)
plates. These tests may be semiquantitative or purely qualitative. Controls for these
tests are usually qualitative (positive and negative controls).
Recommendations for verification of performance of new lot slide agglutination assays:
-
At a minimum, testing of new reagents with existing control material. Acceptability
of the new lot would be achieving the expected results for both positive and negative
controls.
-
Stored plasma from previously tested using semiquantitative methods should be additionally
considered to assure longitudinal continuity of result reporting, with the acceptability
of new lot demonstrating equivalence (locally determined).
Platelet Aggregation or Platelet Testing Reagents
There are several platforms that assess platelet function, including rapid whole blood
methods, impedance methods, flow cytometry methods, and traditional platelet-rich
plasma methods. The veracity of these methods for assessing platelet function is outside
the scope of this document. Viscoelastic measurements used for global coagulation
assessment, including platelet function, will be discussed below.
Reagents for platelet function assessment may be single-use cartridges or separate
reagents (agonists) that are used as part of a profile for assessing platelet function
or determine the efficacy of antithrombotic therapy.[34]
[35]
[36]
[37]
[38] In general, the platelets present in patient whole blood or plasma are exposed to
an agonist(s) that, in normally functioning platelets, cause the platelets to undergo
activation, shape change, agglutination, degranulation, and aggregation. The testing
conditions may include shear rates (closed vacuum or cone/plate system) or more static
platforms such as magnetic stirring within a glass or plastic container. The difficulty
associated with quality assurance for platelet function testing is the ability to
assess abnormal samples, given the limited stability of platelets in citrated blood
or plasma (maximum 4 hours). Quality control for platelet function testing may be
limited to either (1) electronic assessment of instrumentation (primarily limited
to close system methods) or (2) normal donor samples tested concurrently for each
day of use. While normal donor testing may assure that there are no factitious abnormal
results, testing a normal donor does not assure the suitability of reagents for detecting
platelet function defects. This may be of particular concern when reagents are prepared
locally, in lieu of commercial sources. Contrived samples can be considered, using
normal donor blood enriched with aspirin or specific antithrombotic drugs that do
not require in-vivo metabolizing targeting adenosine diphosphate or platelet glycoprotein
IIb-IIIa receptors, to create abnormal aggregation responses or profiles.
Recommendations for performance verification of new lot of reagents or cartridges
for platelet testing.
-
For platelet testing using reagent cartridges, at a minimum one normal and abnormal
donor should be tested and compared with the existing lot of reagent(s).
-
For platelet aggregation reagents, at a minimum one, but optimally at least three
normal donors should be tested and compared with existing lot of reagent(s).
-
Optimally, in addition to normal samples, testing three (3) abnormal patient samples
or surrogate (contrived) samples specimens should be considered, especially when local
reagent preparation is used for testing to assure longitudinal continuity of result
reporting.
-
Laboratories should collate historical data from normal donors to aid in the reassessment
of new batches of reagents.
-
Due to limitations in QC and difficulty in evaluating batch-to-batch variation, consideration
should be made on how to limit batch change (use of a large batch of frozen reagents,
etc.)
-
As there are no commercial manufacturer control materials for platelet-rich plasma
aggregation, an individualized quality control plan (IQCP), or as regional regulations
required, should be considered for platelet function testing.
Viscoelastic Measurements
Viscoelastic measurements are tests that measure the physical properties of clotting
blood.[39]
[40] The measurements of clotting blood encompass either resistance (torque), sound,
or other physical properties that are measured continuously in real time, ultimately
resulting in graphical representation of clot initiation, fibrin polymerization, and
platelet aggregation, with thrombin generation, and potentially assessment of clot
lysis 30 to 60 minutes after maximal clot formation. Reagents are typically instrument-specific,
may be single-use cups or cartridges, with some systems allowing local modifications
or enhancements using different agonists than provided by the manufacturer. Most commercial
sources for viscoelastic reagents provide suitable control material (two levels, low
and high), although these control materials may not address each available reagent
or address each reportable parameter.
Recommendations for the verification of performance of new lot viscoelastic measurement
reagents:
-
At a minimum, testing of new reagents with commercially available control material,
prior to clinical use. Acceptability of new lot would be recovering expected results
(provided by the manufacturer) for each parameter.
-
Ideally, for reagents or kits that do not have quality control or do not provide any
abnormal results for reported parameters, concurrent testing (with existing lot reagents)
with at least one normal and one abnormal patient (or surrogate sample) should be
considered to assure longitudinal continuity of result reporting. Refer to regional
regulatory authorities for further guidance if an individualized quality control plan
is required.[41]
Other Point-of-Care Tests
Point-of-care tests (POCT) methods using citrate or native whole blood for hemostasis
testing have been available for decades, which are mostly targeted for use by nonlaboratory
professionals or patient self-assessment (after appropriate training). The most common
coagulation test menu using POCT platforms include the ACT, PT (and/or INR), aPTT,
and D-dimer.[42]
[43] POCT methods are single-use cartridges that are instrument-specific, and some cartridges
(primarily PT/INR or aPTT) have internal features or controlling mechanisms to aid
in determining the veracity of the sample collection and test result. POCT cartridges
are lot-specific and typically packaged to contain multiple test cartridges (e.g.,
25 per box).
Recommendations for the verification of performance of new lot POCT reagents (not
POCT for patient self-assessment):
-
At a minimum, testing of new lot POCT reagents with the specific quality control material
available from cartridge manufacturer if required prior to clinical use. Acceptability
of a new lot would be recovering expected results (provided by the manufacturer).
-
Verification of new lot performance using EQA material would be an additional consideration,
especially if the quality control material fails to meet expectations. Acceptability
of a new lot would be defined by the respective EQA summary.
-
Unexpected INR result(s) from a new lot POCT INR cartridge should be confirmed with
a laboratory INR determination prior to dose adjustment.
Calibrated Coagulation Tests
With the exception of some POCTs, the aforementioned coagulation tests are mostly
assays that report measurements of time (chronometric) or some coagulation function
over time. Tests that are reported in units per volume (e.g., mg/dL and IU/mL) are
calibrated and mitigate to some degree the need for new RI assessment but still require
accuracy assessment between reagent lots to assure transparency to clinicians. The
frequency of calibration may be regionally defined by regulatory agencies or by the
reagent manufacturer (e.g., every 6 months), but all new reagent lots require calibration
prior to clinical use. Acceptability of calibration is beyond the scope of this document
but is likely defined by the instrument manufacturer IFU. Consider performing precision
checks regularly (e.g., monthly) to reduce the risk of introducing bias by calibrating
a test on an imprecise instrument. Where appropriate, using calibrator(s) and controls
from different commercial sources may provide additional assurance of testing accuracy.
However, this may be considered an LDT by local regulatory authorities, which may
require additional validation.
Calibrated coagulation tests include chronometric, chromogenic, immunoturbidimetric,
immunochemiluminescent, immunofluorometric, enzyme-linked immunosorbent assay, and
others. Commercially available calibrated tests are commonly provided as kits containing
the required reagents, which may or may not also contain the required calibrator(s)
or controls. Factor assays are commonly performed using one-stage clot-based methods,
where each required reagent is obtained separately and changing all reagent lots concurrently
may be considered. Evaluation of new lots of more stable reagents such as buffers,
diluents or calcium chloride, saline, and water may not be required prior to clinical
use.[2] For laboratories with multiple analyzers (within institution or institution network),
it may be of value to check clotting times of factor curves between analyzers as well
as checking optical density readings of chromogenic and immunoturbidimetric curves
between analyzers (and between calibration runs), especially when there is a deviation
in final results between testing sites. Solid phase assays often utilizing chemiluminescent
technology have calibration curves pre-established by reagent manufacturers. For some
tests, international consensus RIs are available (e.g., anticardiolipin or anti-β2
glycoprotein I antibodies) so that batch-to-batch performance is partially relevant
where reassessment of RI is unfeasible.
Recommendations for verification of performance of new lot reagents that require calibration.
-
After calibration, at a minimum, each laboratory should assess new reagents with commercially
available (ideally assayed) control materials or in-house QC prior to clinical use.
Acceptability of a new lot would be recovering expected results (provided by the product
manufacturer) for each parameter.
Note: Regulatory agencies may not accept kit-provided control material as acceptable
new lot reagent verification of performance. For these laboratories, the verification
of calibration and new lot reagent performance can be achieved by:
-
Alternative (third party) sources of control or calibrator material that cover the
AMR can be used if the commutability of the controls or calibrators is demonstrated.
-
A more robust new lot evaluation would include an imprecision and accuracy assessment
by testing three replicates of three different levels covering the AMR.[4]
-
A minimum of 3 to 5 patient or commercial samples at medical decision levels, but
ideally 10 to 20 patient samples tested using previously verified reagents spanning
the reportable range.[45]
-
Surrogate and stored frozen plasma samples may be used. Frozen plasma samples should
be tested within 2 hours of thaw.
-
Calibrator or control material from another reagent or kit lot, if allowed by manufacturer
IFUs (no target value specificity to a given lot indicated).
-
Testing of certified reference materials.
Note: Excluding clot-based assays, some regulatory agencies may require the AMR range
to be verified prior to clinical use, especially when a single calibrator is used
to create AMR. AMR verification can be achieved using previously tested patient samples
or commercially available assayed material.
-
IQC values (targets and limits) should be assessed for the new lot reagents using
assayed (optimal) control material.
-
Factor-deficient plasma used for one-stage factor assay testing should be evaluated
to ensure the deficient material contains <0.01 IU/dL (<1%) of factor.
-
Laboratories servicing specialized patient populations (e.g., emergency departments
and hemophilia treatment centers) may consider assessing/verifying lower limit of
quantitation, especially if calibrated tests are used for screening purposes.[29]
[46]
[47]
-
For methods using preestablished calibration curves, a local calibration is performed
and the acceptable limits are set by the manufacture.
