Keywords
antiphospholipid syndrome - anticardiolipin antibodies - anti-β2-glycoprotein I antibodies
- prothrombin - β2-glycoprotein I - interference - classification
The presence of antiphospholipid antibodies (aPL) defines the antiphospholipid syndrome
(APS).[1] The classification criteria act as surrogate markers for the diagnosis of APS, and
are based on the combination of clinical symptoms and the presence of circulating
aPL.[1] APS is an autoimmune disease with clinical symptoms characterized by thromboembolic
complications and pregnancy morbidity, requiring long-term anticoagulant therapy.[1]
[2] The primary form of the syndrome is defined in the absence of any underlying autoimmune
disease, whereas the secondary form is associated with another autoimmune disorder,
especially systemic lupus erythematosus (SLE).[2] Both forms of APS usually do not show diagnostic or management differences. Besides
the main clinical presentation of thrombosis and pregnancy morbidity, the syndrome
may be associated with a variety of other clinical symptoms: thrombocytopenia, cardiac
valve disease, renal, neurological, pulmonary and dermatological manifestations.[2] The aPL are not specific for APS and can be present in different clinical settings
as well as being sometimes detected in healthy individuals or in patients with other
autoimmune disease.[2]
The autoantibodies associated with APS, the aPL, include anticardiolipin antibodies
(aCL), anti-β2-glycoprotein I antibodies (aβ2GPI), and lupus anticoagulant (LA).[1]
[3] The first two groups, aCL and aβ2GPI antibodies, are measured by solid phase assays.[4] aCL are directed against cardiolipin, which is a phospholipid contained in cell
membranes. aβ2GPI are directed against β2-glycoprotein 1 (β2GPI), a cardiolipin-binding
cofactor of aPL. LA comprises a mixture of various aPL, which is detected by the prolongation
of phospholipid-dependent coagulation tests.[3]
[5] Measurement of aCL and aβ2GPI results in an antibody titer, while LA testing is
reported with a final conclusion as positive or negative, interpreted toward local
cut-off values of the clotting times, depending on antibody characteristics and reagents
used.[4]
[5]
The Sydney criteria were intended to be classification criteria for clinical trials
and studies to identify patients with high probability of APS. Meanwhile, the laboratory
criteria are now used as diagnostic criteria as well.[1] These laboratory criteria were further defined over the past years, and are still
subject of discussion and regular updates.[3]
[6] The diagnosis of APS remains a challenge from the clinical side, since clinical
criteria of APS, vascular thrombosis, and pregnancy morbidity, are common in the general
population, and often not triggered by the presence of aPL. Thus, the requirement
for reliable aPL assays for the diagnosis of APS is high. Additionally, detection
of aPL is more than a diagnostic tool since the type of aPL and the level define the
risk of vascular and obstetric features in APS patients.[7]
[8] In particular, patients positive for the three groups of aPL (LA, aCL, aβ2GPI) or
double positive (aCL and aβ2GPI), and those with medium/high titer of aPL, have the
highest risk for adverse events.[9] All assays currently used for aPL detection show shortcomings, making the laboratory
diagnosis of APS a real challenge.[10]
[11]
[12] In this state-of-the-art article, insights on the clinical relevance and the interpretation
of aPL will be described, with focus on aCL and aβ2GPI detected by solid phase assays.
Type of Assays
aCL and aβ2GPI are measured by solid phase assays, a technique frequently used to
detect autoantibodies. It is important to understand the limitations of the various
techniques as well as their clinical relevance. This implies insight into the sensitivity
and specificity of these assays but also the relevance in the clinical context and
the relevance of the antibody level.
Already in 1985, the first ELISA (enzyme linked immunosorbent assay) for aCL was developed
and resulted in the first description of the so-called anticardiolipin syndrome, later
named APS with persistent positivity of LA and aCL as necessary conditions.[13] In the 1990s it was illustrated that aCL activity in most APS patients depends on
a co-factor protein identified as β2GPI.[13] In the revised Sapporo classification criteria, the Sydney criteria, the aβ2GPI
were included as separate laboratory criterion.[1]
The currently commercially available aCL assays coat the solid phase with cardiolipin
and β2GPI, which increases the specificity of the assay. The β2GPI dependency of the
aCL assays is mandatory to avoid detection of non-cofactor dependent aCL associated
with infection and drugs.[14] Limited data on the clinical relevance of β2GPI-independent aCL are available and
β2GPI binding antibodies are regarded as a main player in the pathophysiology of APS.[2]
By the introduction of the aβ2GPI assay into the criteria, it was thought that these
antibodies were more specific and more reliable for standardization, and the residual
role of the aCL was then debated. Soon, it became clear that aβ2GPI assays also suffer
from the same problems of standardization.[13] β2GPI is regarded as the main cofactor for aPL, and its pathogenic role has been
illustrated in in vitro experiments and animal models.[2] However, the clinical evidence of aβ2GPI is not that strong, as illustrated in a
recent review where the presence of aβ2GPI had only a weak independent association
with thrombosis and was, at best, inconsistently associated with obstetric complications.[15]
The methodologically correct aCL assays show a comparable sensitivity and specificity
to aβ2GPI assays and aCL and aβ2GPI levels are highly correlated.[14]
[16]
[17]
[18]
aβ2GPI antibodies bind to β2GPI complexed with cardiolipin in cardiolipin coated plates,
or directly to β2GPI in β2GPI coated plates. β2GPI-dependent antibodies are responsible
for positive results in the two solid-phase assays, aCL and aβ2GPI assays, both laboratory
classification criteria for APS.[1]
[3]
This can raise the question why we need both aCL and aβ2GPI assays. First, a meta-analysis
confirmed the role of aCL as a marker related with an increased risk of thrombosis.[19] Second, in APS patients we expect no single positive result for one or the other
assay. aCL measurement is of help in confirming the positivity of aβ2GPI of the same
isotype, and vice versa. Positive aCL in the absence of aβ2GPI should be interpreted
with caution. False positive and false negative results vary between assays and depend
on solid phase assay conditions from different manufacturers.[18] Isolated aCL are usually not associated with APS-related clinical symptoms. Due
to variability in immunological assays,[18] checking with another type of aβ2GPI assay can be useful to exclude or confirm the
β2GPI dependence of the aCL. If the aβ2GPI are negative, the aCL antibodies are either
co-factor independent or bind through cofactors different from β2GPI with unknown
significance.[3] Positive aβ2GPI with a negative aCL may include aβ2GPI directed against domain four
or five of the protein. These antibodies are usually positive in the aβ2GPI assays
that are coated with the whole β2GPI molecule, and give negative result for aCL.[20] Antibodies against domain four or five of the β2GPI are regarded non-pathogenic,
compared with the antibodies against the domain one of β2GPI (aDI) that are well correlated
with thrombosis and obstetric complications.[21]
[22] Regarding the assay limitations with possible discrepancies in aCL and aβ2GPI positivity,
the current laboratory criteria with allowance for only one positive laboratory criterion
sufficient for diagnosis of APS, may be questioned. To overcome over-diagnosis based
on single positivity of aCL or aβ2GPI, both aCL and β2GPI have diagnostic value.
Traditionally, aCL and aβ2GPI are detected by ELISA[1]; recently, however, automated platforms with variations of the solid phase (e.g.,
magnetic microparticles, microspheres) and various detection systems (e.g., chemiluminescence,
flow cytometry, multiplex systems) have been introduced into the market.[16]
[17]
[18] Automated systems have the advantage of the working conditions being more harmonized,
providing a strict protocol on how to perform the assay, which may reduce the inter-laboratory
variation for laboratories using the same system.[23] In general, ELISAs have shown large inter-laboratory variation and limited consensus
in external quality control programs.[24] Head-to-head comparison of different platforms show comparable sensitivity and specificity
for ELISA and automated systems[18] and correlation of titers between automated solid phase assays and ELISA is good.[17]
[18] The automated platforms perform aCL and aβ2GPI IgG/IgM assays more rapidly and in
a less labor-intensive fashion, providing up to four results at once instead of running
multiple ELISAs. However, the availability and regulatory approval of automated systems
requiring specific instrumentation may be a drawback for some geographical regions.
Interferences
Interfering factors of solid phase assays may include rheumatoid factor (RF), or substances
(hemoglobin, bilirubin, triglycerides) interfering with the absorbance signal measured
to determine the antibody level, which may relate to specific methodology. Manufacturers
should indicate the level of IgM RF as well as the concentrations of hemoglobin/bilirubin/triglycerides
and other interfering factors that may bias results.[4]
In contrast to LA detection that is prone to interference with the phospholipid-dependent
coagulation tests, aCL and aβ2GPI solid phase assays do not suffer from confounding
effects of anticoagulants or acute phase proteins that may otherwise interfere with
results.[5] Although LA is an important laboratory criterion having a strong correlation with
thrombosis and obstetric complications, the solid phase assays have the advantage
not giving false positive results when measured during anticoagulation therapy.[25]
The phospholipid-dependent tests used to detect LA (the activated partial thromboplastin
time, dilute Russel viper venom test, or silica clotting time) are prolonged by most
anticoagulants as well as by LA, thus making the interpretation of screening, mixing,
and confirmation steps very difficult.[5] APS patients are candidates for long-term anticoagulation and require laboratory
evaluation to assess the antibody profile. Despite that some strategies to overcome
interference of anticoagulants exist (neutralization of anticoagulants, alternative
tests), LA detection remains a challenge during anticoagulation therapy.[25] Although strongly discouraged to test in the acute phase or during anticoagulant
therapy, many patients are referred for laboratory testing at the start of, and during,
anticoagulation therapy.[5]
[25] One of the reasons here being that the antibody profile, and presence versus absence
of aPL, may influence the decision to discontinue or to extend anticoagulation duration,
and may define the choice of anticoagulant since vitamin K antagonists are preferred
over direct oral anticoagulants in triple positive patients.[26]
Agreement between Solid Phase Assays
Agreement between Solid Phase Assays
Variability between aPL detecting assays is hypothesized to result from preanalytical,
analytical, and postanalytical conditions, calibration and assay-specific issues.[12] Technical aspects on how to measure aPL with solid phase assays and how to interpret
results are addressed in guidance documents[4] but cannot prevent existing differences amongst the large variety of commercial
and in-house assays.
Best illustrated by external quality control programs,[24]
[27] this inter-assay variation remains a major concern. In a real-world setting on a
large cohort of APS and non-APS patient samples, we have recently shown that with
commercially available solid phase assays, even with eliminating inter-laboratory
and inter-operator variation, there is still variable detection of aCL and aβ2GPI
IgG/M between platforms.[18] In this study, a discrepancy of 36, 60, 53, and 36% for aCL IgG, aCL IgM, aβ2GPI
IgG, and aβ2GPI IgM positivity was observed, respectively. Detection of aCL IgG and
aβ2GPI IgM resulted in the best agreement. Remarkably, discrepant samples displayed
lower median aPL titers.[18] Nevertheless, these may impact on the identification of samples as negative or positive,
where results are around the cut-off values. Despite the discrepancies in positive/negative
results, this was not reflected in the clinical performance of the platforms by odd
ratios for thrombosis and/or pregnancy morbidity considering the results of all four
aPL together, being globally concordant among the tested platforms.[18] Therefore, if both IgG and IgM are needed, it is recommended to measure aCL and
aβ2GPI IgG and IgM within the same platform.[18]
Nevertheless, detection of patients positive for aCL or aβ2GPI is assay dependent
and this poor agreement between available platforms hampers a uniform classification
of positive aCL/aβ2GPI antibodies. When clinical suspicion is high for APS, and results
of aPL are not in line with what is expected, consideration of retesting with another
type of solid phase platform can be useful. A sample assigned positive in one assay
does not automatically test positive in the same type of assay from a different manufacturer
or in another laboratory. Clinicians are not always aware of the sensitivity and specificity
of laboratory tests. Test results should always be related to clinical symptoms and
an interaction of the laboratory and clinician is mandatory.[11]
Besides differences in agreement (positivity vs. negativity), also large variation
in titers has been described.[28] The differences in titer can be explained by the difference in calibration of the
assays. Results of aCL and aβ2GPI are expressed in arbitrary units and derived from
the assay-specific calibration curve converting the test signal into antibody units.[4] Manufacturers use a variety of calibrators, often using secondary standards and
working calibrators.[10]
[18] Ideally, these calibrators should be traceable to a primary standard. If aCL assays
are calibrated against the Harris standards, results are expressed in GPL or MPL,
1 GPL or MPL referring to 1 μg of IgG or IgM antibody. No reference material is available
for aβ2GPI, therefore, results are expressed in arbitrary units.[4]
Harris/Louisville standards and Koike monoclonal antibody standards were developed
to introduce an international standard for the aCL assays. Patient-derived material
such as Harris/Louisville standards is finite and may have batch-to-batch variation.
In contrast, production of monoclonal antibodies is indefinite and reproducible over
time, but have the disadvantage that these may not be representative for the reactivity
of patient's polyclonal aPL, and they are not always positive in all aPL assays. The
Koike standards are not commercially available anymore. Human monoclonal antibodies
derived from APS patients can offer an alternative.[10] Recently, a patient derived reference material for aβ2GPI has been developed but
is not available yet.[29] The lack of a “gold standard” reference material makes comparison of titers across
different platforms very difficult.
Criteria for Positive and Negative Results
Criteria for Positive and Negative Results
In the current Sydney classification criteria, only medium and high levels, i.e.,
greater than the 99th percentile or 40 units IgG or IgM for aCL and greater than the
99th percentile for aβ2GPI, are regarded as relevant.[1] Patients with definite APS usually have values of aCL far exceeding 40 GPL.[30] Lower levels of antibodies are observed in certain clinical settings, especially
pregnancy morbidity.[31] Awaiting more clinical studies and international consensus, the same threshold (99th percentile) for aCL and aβ2GPI is recommended for patients with thrombosis and patients
with obstetrical complications.
Although titers between systems correlate, the numerical values of the antibody titers
may differ.[17]
[18]
[28] Titers of aCL and aβ2GPI measured with chemiluminescence techniques are much higher
compared with ELISA.[28] The nonparametric 99th percentile cutoff appears to be more specific than the >40 GPL/MPL value. Because
of the variation in titers in different commercial kits, it is advised that each laboratory
uses the 99th percentile of a local normal population to determine the cutoff value, also for aCL.[4] However, as this requires in general a recommended minimum of 120 normal samples,
some laboratories are challenged to do so. Nevertheless, it seems nearly impossible
to recommend one numerical value (being >40 GPL/MPL[1]) as general criterion for positivity. So far, each test result above the cutoff
should be regarded as positive and reported quantitatively along with the local cutoff
value.[4] Calculation of an in-house cutoff value by the 99th percentile should be done, and to obtain a reliable cutoff value at least 120 normal
donors should be used.[4]
[32] Since the high number of normal donors is not feasible for every laboratory, an
alternative is the transference of the manufacturer's cutoff values, which is often
applied for the solid phase assays.[4] Before transference of the manufacturer's cutoff values a verification should be
performed on a cohort of 20 normal donors. Applying the CLSI C28-A3 guideline, manufacturer's
cutoffs may be accepted if no more than two out of 20 tested normal values fall outside
those of the manufacturer's reference range, if so, a second cohort of 20 normals
should be tested.[33] If more than 10% of the normal values are outside the manufacturer's reference range,
cutoff values should be calculated in-house. In addition, transference of manufacturer's
cutoff values assumes that these cutoffs are established by appropriate statistical
models using a sufficiently large donor population. Of note, most manufacturers suggest
in the package insert that in-house cutoff values should be calculated.
Medium and high aPL titers are considered to be more strongly correlated with clinical
outcomes of APS than low titers. Increasing likelihood ratios (LR) with increasing
autoantibody titers has already been reported for other autoimmune diseases.[34] Increasing LR for thrombotic and obstetric APS based for aCL IgG was observed with
increasing titers. The same trend was observed in the thrombotic test population for
aβ2GPI IgG, but less pronounced in the obstetric test population. For aCL IgM and
aβ2GPI IgM no prominent difference in LR was observed for diagnosis of thrombotic
or obstetric APS.[35]
This raises the issue of classification into low-medium-high positive. Qualitative
reporting of results categorizing aCL and aβ2GPI IgG/M titers in low, medium, and
high titers could increase harmonization of interpretation and therefore, could be
very useful for the clinician.[8]
[11] Even more if qualitative gradation of results is interchangeable between different
systems. An international multi-disciplinary initiative has been started to develop
new classification criteria to identify patients with high likelihood of APS for research
purposes, and recognize the difference in semiquantitative reporting between solid
phase platforms.[6] So far, no standardized method to define these ranges is available. Reports using
gradation of positivity by defining intervals are rare,[36] and did not sufficiently investigate whether this can be applied to different solid
phase systems. A recent study investigated whether traditionally established aCL and
aβ2GPI thresholds of 40 and 80 MPL/GPL are appropriate to categorize results as moderate
and high titers for ELISA and automated solid phase platforms.[35] Agreement for the 40/80 threshold and otherwise defined thresholds for low, medium,
and high categorization were studied. Threshold levels of 40/80 units showed poor
agreement between ELISA and automated platforms for classification into low-medium-high
positivity, especially for aCL and aβ2GPI IgG. Use of 40/80 units as medium-high thresholds
was acceptable for aCL and aβ2GPI IgG ELISA, confirming the Sydney criteria that were
based on standardized ELISA techniques.[1] Semiquantitative reporting of aCL and aβ2GPI IgM had less impact on increasing probability
for APS, for both ELISA and automated systems. The use of 40 and 80 units as medium
and high thresholds is acceptable for aCL IgG and aβ2GPI IgG ELISA but cannot be applied
to analytical solid phase platforms with chemiluminescence and multiplex flow immunoassay
methodology. Agreement for semiquantitative interpretation of aPL IgG between ELISA
and chemiluminescence and multiplex flow immunoassay techniques improved by defining
thresholds based on receiver operator curves. However, this is difficult and not feasible
for many laboratories. A more accessible method to define platform-specific thresholds,
is the calculation of thresholds following comparison of parallel measurement of monoclonal
antibodies by an automated solid phase platform and ELISA.[35]
Isotype of aCL and aβ2GPI
Isotype of aCL and aβ2GPI
For both aCL and aβ2GPI, IgG and IgM antibodies are included in the current criteria.[1]
[3] IgG aCL and aβ2GPI show a stronger association with clinical events and are often
associated with IgM positivity, as illustrated in a multicenter study and two recent
reviews.[37]
[38]
[39] This multicenter study comparing four solid phase platforms demonstrated that this
was independent of the platform used.[37] Isolated aCL or aβ2GPI IgM is rare in thrombotic APS and more frequent in obstetric
APS. Although, in contrast to thrombotic complications, in pregnancy morbidity aCL
and aβ2GPI IgM are independent risk factors. Data of this multicenter study support
testing for IgG and IgM, especially in women suspected of obstetric APS, aCL and aβ2GPI
IgM have an added value for diagnosis. In thrombotic patients suspected of APS, first
line testing for IgM has no added value for diagnosis, but can be used in risk stratification
since positivity of IgM, on top of IgG and LA, increases the risk.[37] A stepped approach starting with aCL and aβ2GPI IgG, and if positive also performing
IgM, can be a good option to estimate the risk profile. For non-criteria clinical
manifestations, the association is mainly shown for LA and aCL IgG, except for thrombocytopenia
that is also significantly associated with IgM.[40]
[41]
[42] Several studies confirmed that the presence of aCL and aβ2GPI of the same isotype
reinforces the clinical probability of APS.[30]
[43]
[44]
IgA aCL and aβ2GPI are not included in the current criteria for APS.[1]
[3] Many studies have illustrated the association between APS-related clinical symptoms
and the presence of aCL/aβ2GPI IgA, especially in SLE.[45] However, there is no strong evidence of the added value of IgA aPL[46]
[47] since isolated IgA is very rare, and IgA aPL are usually found in association with
IgG and/or IgM. The clinical significance is unknown as IgA aPL are in most cases
linked to non-criteria clinical manifestations of APS.[8]
[46]
[47]
[48] So far, there is not enough evidence to recommend testing for IgA aCL and/or IgA
aβ2GPI to increase the diagnostic accuracy of APS. aPL are included in the classification
criteria for SLE, including IgG, IgM and IgA aCL and aβ2GPI. Overall, aPL have a low
relative weight in the classification criteria for SLE, but no difference is made
between the isotype.[49]
Other Antiphospholipid Antibodies
Other Antiphospholipid Antibodies
Amongst the “non criteria” aPL, anti-phosphatidylserine/prothrombin antibodies (aPS/PT)
and antibodies toward the domain one of β2GPI (aDI) have been most frequently studied.
aDI antibodies are a subgroup of aβ2GPI antibodies, directed against the domain one
of the protein β2GPI that consists of five domains. A high association with thrombosis
has been shown for antibodies recognizing a specific epitope (G40-R43) within the
domain one of β2GPI.[50] However, the different methods to detect aDI IgG vary in specificity for this subgroup
of antibodies.[51] A commercial chemiluminescence-based assay (only available for IgG) has been evaluated
in many studies with various results regarding association with thrombosis and pregnancy
morbidity.[51] The original in-house ELISA test measured a more specific population of aDI directed
against the G40-R43 epitope, compared with the commercial aDI probably measuring all
aDI antibodies against any epitope on domain one of β2GPI.[22] No higher risk association for thrombosis was found when aDI was added to the current
aPL panel or when aβ2GPI was replaced by aDI.[22]
[52] The absence of an added value may result because of reduced exposure of the epitope
G40-R43 in the chemiluminescence assay.[51] One recent study indicated aDI as an independent risk factor.[53]
Clinical studies have illustrated that aDI are frequently present in triple positive
patients (positivity for LA, aCL, and aβ2GPI) known as high-risk patients. Titers
of aDI in these patient groups are higher compared with double positive patients (positive
for aCL and aβ2GPI).[22]
[54]
[55]
[56] The high correlation between aDI and triple positivity in obstetric and thrombotic
patients, confirm the higher risk for clinical events in APS.[22]
[54] However, there is not 100% overlap with aβ2GPI and up to 20% of the patients positive
for antibodies against the whole β2GPI molecule can test negative for specific aDI,
depending on the assay used for aβ2GPI.[22]
[57] Instead of being a replacement of aβ2GPI, aDI can be used as a confirmatory test
and is useful for proving the specificity of the aβ2GPI antibodies. Especially in
patients with an incomplete antibody profile defined as double positive patients (aCL
and aβ2GPI positive) or single LA or single aβ2GPI positive patients, aDI may confirm
the risk. Testing for aDI in these patients could confirm or exclude the association
of pathogenic aβ2GPI autoantibodies.[58]
aPS/PT antibodies seem to represent a strong risk factor for thrombosis, both arterial
and/or venous.[59] Persistently positivity for aPS/PT has been shown in patients and asymptomatic carriers
with triple positivity, a marker of clinical severity, making aPS/PT effective as
part of risk scoring.[60]
[61] Equal as for aDI, in triple positive patients, titers of aPS/PT are higher than
in double or single positive patients.[60]
[62]
[63] aPS/PT IgG are frequently present in APS patients.[61]
[64] aPS/PT of isotype IgM are frequently present in isolated LA (non-APS) patients,
as well as in asymptomatic carriers.[61]
[62]
[65] aPS/PT IgG and IgM are more frequently found in triple positive patients compared
with single LA positive patients.[63]
aPS/PT antibodies might be a surrogate for LA because of their strong correlation
with LA,[61]
[65]
[66] particular in conditions where LA assays show methodological shortcomings as in
anticoagulated patients, this might be helpful.[58]
[67] However, not all single LA positive patients are aPS/PT positive.[63]
[65] To replace LA with a suitable and similarly predictive marker that can be measured
with an immunological method needs further investigation.
Similarly, as for aCL and aβ2GPI antibodies, a standardized ELISA for aPS/PT is not
available and reference sera are lacking. Automated systems are not available yet,
and a limited number of commercial ELISAs for aPS/PT is available. A comparison of
a homemade and a commercial assay showed reproducible and accurate results.[68] Recently reported, a newly developed ELISA looks promising to detect pathogenic
antiprothrombin antibodies by coating the ELISA plate with a novel prothrombin variant.[69] Although searching for aPS/PT in daily practice is not recommended yet, in some
conditions aPS/PT might help, especially in confirming the risk. Adding aPS/PT to
the triple positive profile, further consolidates the diagnosis of APS. When the aPL
profile shows double positivity, positive aPS/PT may suggest a false negative LA,
and when aDI and aPS/PT are negative, this indicates a lower risk for thromboembolic
events.[58]
[Table 1] summarizes some of the potential uses of the non-criteria aPL tests.
Table 1
Use of non-criteria aPL tests
|
1. Scope of non-criteria aPL
|
|
aPL other than aCL and aβ2GPI are not included in the laboratory criteria for APS
• aPS/PT antibodies correlate with clinical symptoms of APS, are strongly associated
with LA and frequently present in APS patients.
• aDI β2GPI IgG antibodies correlate with clinical symptoms of APS and are associated
with triple positive high-risk population.
• aPS/PT and aDI β2GPI IgG antibodies confirm patients at risk.
|
|
2. Role of aPS/PT and aDI
|
|
• In triple positive patients (LA, aCL and aβ2GPI), positive aPS/PT consolidates
the diagnosis of APS.
• In double positive patients (aCL and aβ2GPI), positive aPS/PT may indicate a false
negative LA.
• Presence of aDI may indicate the presence of pathogenic aβ2GPI antibodies. When
aDI and aPS/PT are negative, this indicates a lower risk for thromboembolic events.
• If the only positive test is aβGPI antibodies aDI and anti-domain four or five
(4/5) antibodies (if available) should be analyzed; anti-domain 4/5 antibodies are
not associated with thromboembolic events.
• aPS/PT may be used in isolated LA patients to confirm the single LA positivity.
|
Abbreviations: aPL, antiphospholipid antibodies; APS, antiphospholipid syndrome; aCL,
anticardiolipin antibodies; aβ2GPI, anti-β2-glycoprotein 1 antibodies; aPS/PT, anti-phosphatidylserine/prothrombin
antibodies; aDI, antibodies toward the domain one of β2GPI (aDI).
Combination of aPL
Although it is sufficient to have one aPL positive to define or diagnose APS, guidelines
strongly advise to classify APS patients into categories according to type and number
of tests positive.[1]
[3] To assess the risk profile, results of all three groups of aPL, LA, aCL, and aβ2GPI
should always be interpreted together.[3]
[4]
[5] Therefore, all three tests should be performed preferable on the same sample. LA
testing with coagulation assays and aCL/aβ2GPI testing with solid phase assays tend
to be performed in different laboratory departments: the hemostasis/hematology laboratory
and the immunology/ biochemistry laboratory. Consequently, most of the laboratories
use serum to perform solid phase assays.[10] Both serum and citrated plasma can be used for aCL and aβ2GPI, on the condition
that assay specifications (including cutoff values) are validated for the corresponding
sample type.[4]
Combined positivity for LA, aCL, and aβ2GPI antibodies (i.e., triple positivity) has
been shown to be associated with a high risk of both a first event and recurrence,
and for a first event in asymptomatic carriers.[9]
[43]
[70]
[71]
[72] Double or triple positivity for aPL is a risk factor for future thrombotic events,
especially in individuals with an underlying autoimmune disease, whereas single positivity
does not seem to carry an elevated risk of thrombosis.[71] Compared with triple positives, the risk in double positives (aCL and aβ2GPI) is
slightly lower.[71] Triple positive patients usually have a persistent antibody profile on follow-up
testing after 12 weeks.[73]
[74] But also, double and single positive patients may be persistent positive, as illustrated
in a retrospective study illustrating no significant lower persistence in the single
positive patients (93.3%) compared with the double and triple positive patients (96.8
and 97.3%, respectively).[74]
Isolated aCL and aβ2GPI are regarded of less clinical relevance, either because of
the non-cofactor characteristics of aCL or aβ2GPI directed against epitopes not determining
LA activity.[14]
[20]
[75]
If inconsistencies are observed between aCL and aβ2GPI results, retesting with another
type of solid phase assay may be helpful when clinical suspicion is high for APS,
as well as testing for aPS/PT and aDI (see above).
Repeat Testing and Reporting of Results
Repeat Testing and Reporting of Results
If a test is positive on a first occasion, testing should be repeated after 12 weeks
to confirm the persistence, as aPL are persistent over time according to the classification
criteria.[1]
[3] Transient positive results are usually found in association with infections and
drugs[76]
[77] and are thought not to be of clinical significance, therefore re-testing was originally
meant to avoid over-diagnosis of APS patients who were not persistently positive.[3] Viruses can induce autoimmune diseases, in addition to genetic predisposition and
environmental factors, including formation of transient aPL. Particularly, coronaviruses
are mentioned among the viruses implicated in autoimmunity. Since the first description
of a COVID-19 infected patient who was positive for aPL, many case reports, case series,
and cross-sectional studies were published describing the association of aPL and COVID-19
infection. Interestingly, a high prevalence of aCL and aβ2GPI IgA has been described.
However, few studies investigated the persistence over time of the aPL, and the role
in thrombotic events was not consistent in published literature.[78] The type of aPL as well as the prevalence of different aPL over time vary, possibly
linked to the inflammatory phase of the disease. Routine screening for aPL in COVID-19
patients is not recommended, as their pathogenic role in thrombosis in these patients
is still unclear, although possible. Therefore, testing for aPL in COVID-19 patients
with thrombosis or underlying autoimmune conditions can be performed, as recommended
in the general guidelines for testing for aPL.[5] The presence of type and persistency of aPL would change the anticoagulant treatment
decisions, as it is in other thrombotic patients with aPL.[8] Also in these patients, it is mandatory to confirm the identified aPL after 3 months
whenever possible.
Limited information is available on antibody titer fluctuation over time. Decrease
in levels of aCL over time has been described in association with belimumab (anti
B-cell therapy), and decrease in aCL and aβ2GPI levels in patients treated with hydroxychloroquine
(antimalarial).[79]
[80] There is little information on the value or therapeutic consequences of retesting
persistently positive patients annually. This could be potentially useful as part
of risk assessment related to fluctuation of titers or change of antibody profile
over long-term period.[5]
Results of aCL and aβ2GPI around the cutoff value should be interpreted with care
and preferably repeated. Imprecision of the solid phase method should be considered
before classifying a sample as positive or negative, as an imprecision of 10% is acceptable,
this implies that a 10% difference around a value near the cutoff value may influence
the classification as positive or negative.[4]
Repeat testing of aCL and aβ2GPI, as well as LA, is necessary to confirm positivity
over time, in the context of patients with clinical criteria of APS, to manage these
patients adequately as APS patients regarding anticoagulation therapy.[3]
[8] As described above, aCL and aβ2GPI are also part of the classification criteria
of SLE, patients often presenting without the clinical features of APS.[49] Classification criteria for SLE do not specify that aPL should be persistent positive.
Although, to identify SLE patients at risk for thrombosis, it was demonstrated that
repeat testing with persistent positivity is associated with a higher risk, as demonstrated
for LA, an observation that can be probably extrapolated to aCL and aβ2GPI.
The report on aPL testing should include the recommendation of repeat testing after
12 weeks. Overall, a report with an explanation and an integrated interpretation on
all aPL results should be provided. Therefore, a close interaction between the clinical
pathologist (laboratory expert) and the clinician is necessary.[11] If insufficient clinical information is available, this should be mentioned on the
report possibly hampering a correct interpretation of results. In daily practice,
based on the clinical suspicion of APS, clinicians request testing for aPL. However,
they may not be aware of the limitations of the tests used for measurement of aPL
and the difficulties in interpretation. On the other hand, the clinical pathologist
in the laboratory may not be aware of the clinical symptoms or the use of anticoagulants,
the latter extremely important for the interpretation of LA. Information on the anticoagulation
status and the clinical symptoms of the patient is mandatory for good interpretation
of results and should be provided. In addition to the repeat testing after 12 weeks,
retesting should be suggested on the report, depending on the clinical symptoms (criteria
vs. non-criteria, place and type of thrombosis, underlying disease, aPL test results
not in line with the clinical picture) and the demographic characteristics of the
patient (age, other thrombotic risk factors, possible interference of anticoagulant
therapy), the antibody profile (all three groups of aPL should be considered, in incomplete
profiles aCL/aβ2GPI could be retested with another assay), discrepancies in aCL/aβ2GPI
(disagreement in positivity of aCL/aβ2GPI or isotype), borderline positive or negative
titer of aCL/aβ2GPI, and possible analytical interferences.
Besides discussion of the individual results of LA, aCL, and aβ2GPI in view of the
local cutoff values stated in the report, finally, an overall conclusion should be
made considering LA, aCL, and aβ2GPI results.
Conclusion
The detection of aPL by solid phase assays is an essential step in the diagnosis of
APS. The laboratory diagnosis of APS remains challenging. Progress has been made to
address some of the methodological challenges of aCL and aβ2GPI assays and in better
understanding their diagnostic role. Assays for all three groups of aPL, LA, β2GPI-dependent
aCL, aβ2GPI IgG and IgM should be performed to increase diagnostic utility, with an
integrated interpretation of all results. Positivity should be confirmed on a second
occasion after 12 weeks. In the interpretation of aPL results, antibody profiles help
identify patients at risk. All assays must be performed according to the guidelines
and only high titers (>99th percentile) should be considered. Differences in agreement between solid phase assays
remain, and differences in titer hamper the semiquantitative classification into low-medium-high
positivity.
The variety of available solid phase assays on the market may complicate the decision
of what solid phase assays to implement in the laboratory. The selection of the assays
should rely on some important minimal performance characteristics. Imprecision can
be assessed by internal quality control material or patient samples, and should be
<10%.[4] As no golden standard assay exists, the new assay should be validated from an analytical
and a clinical point of view. Agreement between assays should be explored as well
as the clinical performance assessing the association of an assay with thrombotic
or pregnancy complications.[10] Since an extensive clinical validation is not feasible for all laboratories, published
reports in peer reviewed journals may help, critically considering the study design.
Consulting external quality assessment exercise reports before decision making, may
demonstrate what the imprecision of the assays is, and how assays perform. A laboratory
with less experienced laboratory technologists may benefit from an automated system
instead of ELISA, having more harmonized protocols and being easier to perform. An
alternative can be sending the samples to a centralized fully equipped laboratory
with experience in aPL testing. Equally, if laboratories are not able to measure all
aPL, this issue can be resolved by sending samples to a reference laboratory performing
all aPL on the same sample, as well as offering the analysis of the non-criteria aPL,
of value in some special circumstances. Once implemented, laboratories should include
internal quality control material (commercial or patient based) for their assays and
should participate at external quality assessment exercises. By participating in such
schemes, laboratories can assess their performance against those of the peers.
The non-criteria aPL, such as antibodies against the domain one of β2GPI and aPS/PT
antibodies have their role in confirming the risk in APS, and can be useful, especially
in patients with incomplete antibody profiles.
[Table 2] summarizes the key messages on testing for aPL with solid phase assays.
Table 2
Key messages on measurement of aCL and aβ2GPI
|
1. Scope of aCL and aβ2GPI
|
|
• aCL and aβ2GPI IgG and IgM are two out of the three classification criteria for
APS, both measured with solid phase assays.
○ Identification of all three groups of aPL (LA, aCL, aβ2GPI) enables risk stratification.
○ Triple positive patients (LA, aCL and aβ2GPI positive) and double positive patients
(aCL and aβ2GPI positive) are at high risk to develop clinical symptoms associated
with APS.
○ Medium and high aPL IgG titers are more strongly correlated with clinical outcomes
of APS than low titers. IgM aPL titer correlates less with clinical outcome.
○ IgM aCL and/or aβ2GPI are significantly correlated with thrombosis and pregnancy
morbidity, but higher odd ratios are obtained for IgG compared with IgM positivity.
○ Presence of aCL and aβ2GPI of the same isotype reinforces the clinical probability
of APS.
○ Isolated aCL/aβ2GPI IgM is rare in thrombotic APS and is not an independent risk
factor. IgM positivity on top of LA and IgG increases the risk.
○ Isolated aCL/aβ2GPI IgM is an independent risk factor for obstetric APS.
○ All three parameters (LA, aCL, aβ2GPI) should be performed at once on the same
sample.
|
|
2. Analytical aspects for aCL and aβ2GPI
|
|
• ELISA or automated systems with difference in solid phase or detection method can
be used.
• A β2-dependent aCL assay is mandatory to avoid detection of non-cofactor dependent
aCL.
• aCL and aβ2GPI correlate and show comparable sensitivity and specificity for diagnosis
of APS.
• If aßGPI are negative and aCL positive, then aCL are antibodies recognizing proteins
different from β2GPI, whose significance is not known.
• Detection of aCL and aβ2GPI IgG/M varies between test platforms.
• Numerical values vary between test platforms, hampering a uniform semiquantitative
classification into low-medium-high titers. Local cutoff values (> 99th percentile) are recommended.
• It is recommended to measure aCL and aβ2GPI IgG and IgM within the same platform.
• When clinical suspicion is high for APS, and results of aPL are not in line with
what is expected, consideration of retesting with another type of solid phase platform
can be useful.
• In single aCL, single aβ2GPI or single LA positivity retesting of aCL and aβ2GPI
is advised, with consideration of another type of solid phase platform.
• Imprecision of the method of solid phase assays should be considered, especially
for results around the cut-off.
• Solid phase assays for aCL and aβ2GPI are not prone to interference of anticoagulant
therapy, but some interfering substances have been described, such as rheumatic factor,
bilirubin, hemoglobin or triglycerides.
|
|
3. Interpretation and report of results
|
|
• aCL and aβ2GPI assays should be interpreted together with coagulation LA tests
to assess the clinical significance.
• Individual results of aCL IgG/IgM and aβ2GPI IgG/IgM and the antibody profile should
be discussed.
• Results are interpreted according to the local cutoff values stated in the report.
aCL and aβ2GPI are reported with their titer. Each result above the local cut-off
should be considered as positive.
• Report with an explanation of the results should be given and interpretive comments
should be included (e.g., positive results of aCL or aβ2GPI need to be confirmed on
a second occasion after 12 wk to confirm persistent positivity).
• To prevent misdiagnosis the diagnostic workup for APS requires collaboration between
the clinician and the laboratory.
|
Abbreviations: aCL, anticardiolipin antibodies; aDI, antibodies toward the domain
one of β2GPI (aDI); aPL, antiphospholipid antibodies; aPS/PT, anti-phosphatidylserine/prothrombin
antibodies; APS, antiphospholipid syndrome; aβ2GPI: anti-β2-glycoprotein 1 antibodies;
LA, lupus anticoagulant.
