Prevalence and Thrombotic Risk A ssessment
of Anti- β
2
Glycoprotein I Domain I Antibodies :
A Systematic Review
Massimo Radin, MD
1
Irene Cecchi, MD
1
Dario Roccatello, MD
1
Pier Luigi Meroni, MD
2
Savino Sciascia, MD, PhD
1
1
Center of Research of I mmunopathology and Rare Diseases -
Coordinating Center of Piemonte and Valle d’Aosta Net work for Rare
Diseases, Department of Clini cal and Biological Sciences, and S CDU
Nephrology and Dialysis , Universit y of Turin and S. Giovanni Bosco
Hospital,Turin,Italy
2
Department of Clinical Sciences and Community Health, University
of M ilan, Laborator y of Immuno-Rhe umatology Research, Is tituto
Auxologico Italiano, Milan, Italy
Semin Thromb Hemost 2018;44:466–474.
Address for correspondence Savino Sciascia, MD, PhD, Center of
Research of I mmunopathology and Rare Diseases - Coordinating
Center of Piemonte and Valle d’Aosta Network for Rare Diseases,
Depar tment of Clinical and Biological Sciences, and SCDU Nephrolo gy
and Dialysis, University of Turin and S. Giovanni Bosco Hospital, Piazza
delDonatorediSangue3,10154,Turin,Italy
(e-mai l: savino.sciascia@unito.it).
Keywords
► antiphospholipid
syndro me
► antiphospholipid
antibodies
► β
2
glycoprotein I
domain
► non-criteria aPL
► thrombosis
Abstract
Background To date, the exact prevalence of anti-β
2
glycoprotein I domain I (anti-
β
2
GPI-DI) antibodies in patients with antiphospholipid syndrome (APS) and their role
when assessing thrombosis risk is uncertain.
Objectives To estimate the prevalence o f anti- β
2
GPI-DI in pati ents with APS and to
determine whether anti-β
2
GPI-DI-positive individuals are at greater risk of thrombosis,
as compared with individuals without anti-β
2
GPI-DI, by systematically reviewing the
literature.
Methods A detailed literature search was applied a priori to Ovid MEDLINE In-Process
and Other Non-Indexed Citation 1986 to present and to abstract s from the European
League Against Rheumatism (EUL A R) and American College of Rheumatology (ACR)/
Association for Rheumatology Health Professionals (ARHP ) Annual Meetings (2011–
2015).
Results A total of 11 studie s, including 1,585 patient s, we re analyzed. Patients were
distributed as follow : 1,218 patient s APS (45.4% anti-β
2
GPI-DI-positive; in more detail:
504 primar y APS [55.4% anti-β
2
GPI-DI-positive], 192 secondary APS [43.2% anti-β
2
GPI-
DI-positive], an d 522 not specified), 318 with systemic lupus erythematosus (SLE;
26.7% anti-β
2
GPI-DI-positive), 49 asymptomatic carriers of antiphospholipid antibodies
(aPL) (30.6% anti-β
2
GPI-DI-positive), and 1,859 healthy controls. When considering the
five st udies eligible for thrombotic risk assessme nt, four st udies found a significant
associat ion of anti-β
2
GPI-DI-positivity with thrombotic events, whereas one study
found no predictive correlation with thrombosis (overall odds ratio [OR] for pooled
data: 1.99; 95% confidence inter val [CI]: 1.52–2.6; p < 0.0001).
Conclusion We repor t an overall e stimated median preval ence of anti-β
2
GPI-DI
antibodies of 44.3% in patients with APS and/or SLE and a significantly higher
prevalence among patient s with APS compared with SLE alone. Anti-β
2
GPI-DI anti-
bodies might represent a promising tool when assessing thrombotic risk in patient s
with AP S.
published online
August 4, 2017
Issue Theme Recent Development s in
Antiphospholipid Antibodies and the
Antiphospholipid Syndrome; Guest
Editor: Rolf T. Urbanus, PhD.
Copyright © 2018 by Thieme Medical
Publishers, Inc., 333 Se venth Avenue,
New York, NY 10001, USA .
Tel: +1(212) 584-4662.
DOI https ://doi. org/
10.1055/s-0037-1603936.
ISSN 0094-6176.
466
Downloaded by: Università degli Studi di Torino. Copyrighted material.
The antiphospholipid syndrome (APS) is an autoimmune
disorder characterized by vascular thrombosis and/or pr eg-
nancy morbidity (miscarriages, fetal deaths, pre mature
births, etc.) associated with a persistent positivity for anti-
phospholipid ant ibodies (aPL). The current classification
criteria for APS include three laboratory tests: lupus antic-
oagulant (LA), anticardiolipin (aCL), and anti-β
2
glycopro-
tein-I (β
2
GPI). To prevent detection of transient antibodies,
tests must be positive on 2 occasions, at least 12 weeks
apart.
1
β
2
GPI is hypothesized to be the main antigenic target
for aPL, especially in vascular tissues after conformational
modification.
2
Although some evidence is available supporting a role for
anti-β
2
GPI antibodies in contributing to thrombosis and
pregnancy morbidity,
3,4
their pathogenicity remains a topic
of heated discussion. In fact, not all patients positive for the
presence of anti-β
2
GPI antibodies develop clinical aPL-re-
lated manifestations.
3
Besides, anti-β
2
GPI antibodies have
also been detected in a large ra nge of autoimmune diseases,
such as multiple sclerosis,
5
and nonautoimmune dise ases,
such as leprosy and in children with atopic de rmatitis,
6
with
a vast heterogeneity in terms of titers and persistence.
Thisheterogeneity in the pathogenic potential ofanti-β
2
GPI
antibodies might be ascribed to the molecular structure of
β
2
GPI, presenting multiple antigenic specificities that can be
targeted by different autoantibodies.
7
β
2
GPI has five homo-
logous domains (DI–DV), and recently, several studies have
focused on the epitope distribution of anti-β
2
GPI antibodies to
identify their clinical role.
8–10
The main epitope to have been
found to be associated with APS involves regions of DI,
9,11
and
growing evidence has resulted in the identification of DI as the
“immunodominant epitope.”
10,12,13
Both in vitro and in vivo
preliminary data are in line with this hypothesis, supporting a
role for anti-β
2
GPI-DI antibodies in the development of APS-
related clinical manifestations.
14,15
However, the clinical role of such antibodies is still
debated, and international criteria for derivation and con-
firmation of anti-β
2
GPI-DI antibodies are st ill lacking.
16
To
date, the exact prevalence of anti-β
2
GPI-DI antibodies in
patients with APS and the role of testing for ant i-β
2
GPI-DI
when assessing the thrombosis risk are unclear.
In thi s study, we aim to estimate the prevalence of anti-
β
2
GPI-DI in patient s w ith APS and to determine whether
anti-β
2
GPI-DI-positive individuals are at greater risk of
thrombosis when compared with individuals without anti-
β
2
GPI-DI by systematically reviewing the literature.
Methods
A detailed literature search has been developed a priori to
identify articles that reported findings from clinical and
laboratory studies that tested anti-β
2
GPI-DI antibodi es.
Key words and subjec t terms included are as follows:
(“beta 2-glycoprotein i”[MeSH Terms] OR “beta 2-glycopro-
tein I” [All Fields] OR “beta 2 glycoprotein 1”[All Fields]) AND
domain [All Fields]. The search strategy was applied to Ovid
MEDLINE In-Process and Other Non-Indexed Citat ion 1986 to
present. Abstracts from the European League Agai nst Rheu-
matism (EULAR) and American College of Rheumatology
(ACR)/Association for Rh eumatolog y Health Professionals
(ARHP) Annual Meetings (2011–2015) were screened and
included in the analysis when meeting the inclus ion criteria
and not replicating studies publishe d elsewhere.
Studies that met the criteria to evaluate th e prevalence of
anti-β
2
GPI-DI antibodies and their association with the
thrombotic risk in patients with APS and control populations
were systematically analyzed by two independen t reviewers
(M. R. and I. C.). Disagreements were resolved by consensus;
if consensus could not be achieved, a third party (S. S.) would
provide an assessment of eligibility. As the data on eligibility
were dichotomous (eligible: yes/no), interrater agreement at
both the title and abstract review stage and the full article
review stage was determined by calculation of Cohen’s kappa
coefficient (k ¼ 0.93).
17
Literature search strategy on the prevalence of positivity
for anti-β
2
GPI-DI antibodies is shown in ►Fig. 1.Weincluded
in our analysis only studies reporting: (1) clinical data
referring to aPL-related manifestations, (2) laboratory data
including aCL, LA, and/or anti-β
2
GPI testing, and (3) anti-
β
2
GPI-DI antibodies testing with detailed assay methodol-
ogy, isotype analyzed, defined cu toffs of positivity for anti-
β
2
GPI-DI antibodies. All published series including 10 or
more patients meeting the aforeme ntioned inclusion criteria
were recorded. Methods of enrollment were also analyzed.
Prevalence o f anti-β
2
GPI-DI antibodies was compared be-
tween popu lations by Fi sher’s exact test, two-tailed, whereas
results for association of positivity and risk of thrombosis
were compared using odds ratios reported in the studies
analyzed.
This study has been performed according to PRISMA
(Preferred Reporting Items for Systematic Reviews and
Meta-Ana lyses) guidelines,
18
and the relative checklist can
be supplied upon request.
Risk of Bias Assessment
Two reviewers (M. R. and S. S.) assessed the risk of bias of
individual studies using the Newcastle–Ottawa Scale (NOS)
for cohort studies, and the NOS for case–cont rol studies. The
NOS is a scoring tool used to assess quality of evidence and
risk of bias for nonrandomized studies included in m eta-
analyses.
19
Comment on Excluded Studies
A total of 13 st udies,
20–32
including a total of 6,169 patients
(comprising a total of 533 patients with APS and 68 with
systemic lupus erythematosus [S LE]), which included testing
of anti-β
2
GPI-DI antibodies in diverse cohorts of patients,
were excluded from the analysis of prevalence. The exclusion
of t hese studies was based on (1) the impossibility of extra-
polating clinical or laboratory data, in particular when
reporting the results of anti-β
2
GPI-DI positivit y,
23,26
(2)
lack of reported data on anti-β
2
GPI-DI prevalence,
25
(3) no
defined cutoffs of positivity,
27
and/or (4) testing results
expressed only in relation to the control groups.
26
Data summarizing the studies excluded from the analysis
are illustrated in
►Table 1.
Seminars in Thrombosis & Hemostasis Vol. 44 No. 5/2018
Prevalence of Antibodies Anti-β
2
GPI-DI in APS Patients Radin et al. 467
Downloaded by: Università degli Studi di Torino. Copyrighted material.
Table 1 Studies not included into the prevalence analysis and reason of exclusion
Study Year Tota l
number of
patients
APS SLE Reason of exclusion
Müller-Calleja et al
21
2016 4979 N/A N/A No clinical data of the patients analyzed
Manukyan et al
28
2016 4979 N/A N/A No clinical data of the patients analyzed
Roggenbuck et al
29
2016 61 61 N/A Impossibility of extrapolating prevalence of β
2
GPI-DI positivity
Pengo et al
22
2015 65 N/A N/A No clinical data of the patients analyzed
Willis and
Pierangeli
23
2015 72 72 N/A The study compared two assays for t he detection of anti-β
2
GPI-
DI antibodies, without describing patients’ positivity for the
test
Rodríguez-García
et al
30
2015 1404 1,404 N/A Systematic review
Meneghel et al
31
2015 88 88 N/A Impossibility of extrapolating prevalence of β
2
GPI-DI positivity
Despierres et al
24
2014 439 N/A 12 Clinical data were available for 371 patients, but only 12
patients with SLE were tested for β
2
GPI-DI IgA
Zohoury et al
26
2013 273 273 N/A In this study, β
2
GPI-DI positivity was only expressed as sensitivity,
specificity, and likelihood ratios compared with normal controls
Andreoli et al
25
2013 154 86 28 The results were considered as OD values, and the study used
the sa me cohort of pat ie nt s as in a study i ncl ude d in the a na lysis
Andreoli et al
20
2011 154 86 28 The study used the same cohor t of patients as a study included
in the analysis
de Laat et al
27
2007 33 16 N/A No clinical data of the patients analyzed, no clear cutof fs of
positivity
de Laat et al
32
2005 198 6 176 Impossibility of extrapolating prevalence of β
2
GPI-DI positivity
Abbreviations: β
2
GPI-DI; -β
2
glycoprotein I domain I; APS, antiphospholipid syndrome; N/A, not available; OD, optical density; SLE, systemic lupus
er ythe matosus.
Fig. 1 Literature search strategy on prevalence o f positivit y for anti-β
2
glycoprotein I do main I antibodies.
Seminars in Thrombosis & Hemostasis Vol. 44 No. 5/2018
Prevalence of Antibodies Anti-β
2
GPI-DI in APS Patients Radin et al.468
Downloaded by: Università degli Studi di Torino. Copyrighted material.
Statistical Analysis
A detailed stat istical analysis has been developed apriori.
Odds ratio s with 95% CI (OR [95%CI]) for arterial and/or
venous thrombosis were recorded. If not available, they
were cal culated, whenever possible, by means of contin-
genc y tables. In case–control and cross-sectional studies,
cont ingency tables wer e used to compare the proport ion of
anti-β
2
GPI-D I ant ibodies in patients with a nd without
thrombosis. In prospective st udies, contingency tables
were establish ed as previously repo rted.
33,34
Briefly, if
SLE was the enrolment criterion, the OR [95%CI] was
calculated by comparing the proportion of anti-β
2
GPI-DI
antibodies positivity in patients who did or did not develop
thrombosis. If thrombosis was the enrolment criterion, the
OR [95%CI] was calculated by comparing the proportion of
anti-β
2
GPI-D I an tibodies in pat ients with or without recur-
rent thrombosis during follow-up. If positivity for aPL was
the enrolment criterion, the OR [95% CI] was calc ulated by
compar ing the rates of thrombosis during follow-up of
patients grouped according to different antibody types
and titers.
Res ult s
A total of 11 stud ies,
9,35–44
including a total of 1,585 patients,
met the inclusio n criteria. Patie nts were distributed as
follow: 1,218 patients with APS (504 with primary APS
[PAPS], 192 with secondary APS [SAPS], and 522 not speci-
fied), 318 with SLE, 49 aPL asymptomatic car riers, a nd 1,859
healthy controls (HCs).
Of 11 studies, 9 differentiated patients with APS between
PAPS and SAPS, and only 5 out of 11 studies specified the
results of anti-β
2
GPI-DI antibodies positivity stratifying for
the presence of concomitant autoimmune diseases. Two
studies out of 11 did not specify if th e patients with APS
were PAPS or SAPS.
Detection of anti-β
2
GPI-DI antibodi es was performed
with an enzyme-linked immunosorbent assay (ELISA) in 5
out of 11 studies,
9,37,41,42,44
whereas 6 out of 11 studies used
a chemiluminescent imm unoassay (CIA).
35,36,38–40,43
All stu-
dies investigated the presence of immunoglobulin G (IgG)
isot ype, and one study
44
investigated the presence of IgM
and IgA anti-β
2
GPI-DI also.
The six studies that pe rformed testing with CIA used the
same cutoff of positivity (99th percentile: 20 chemilumines-
cence units [CU]). Regarding the studies t hat used ELISA,
three studies
9,37,41
expressed the cutoff of positivity as
mean þ 3 standard deviation (SD) of the HC as reference,
one study
44
used a cutoff at the 99th percentile of the HC, and
lastly one study
42
used a cutoff at the 95th percentile of the
HC.
Data summarizing the main characteristics of the tests
used to identify anti-β
2
GPI-DI antibodies are provided
in
►Table 2.
None of the studies investigated persistent pos itivity to
anti-β
2
GPI-DI antibodies.
Inclusio n criteria for 6 out of 11 studies
35,38–40,43,44
was
based on APS diagno sis according to Sydney revised Sapporo
guidelines.
1
Two out of 11 studies, although basing their
inclusion criteria on the APS diagnosis, did not specify i f the
APS diagnosis met the Sydney revised Sapporo guide-
lines.
36,41
One study out of 11 had as inclusion criteria the
diagnosis of SLE that met the ACR revised criteria.
37
Two
studies out of 11 had as inclusion criteria the presence of
anti-β
2
GPI antibodies detected at least twice at least 12
weeks apart.
9,42
Results of the cri tical appraisal of the included studies
according to NOS are shown i n
►Table 3.
The overall estimated median prevalence of an ti-β
2
GPI-DI
antibodies in patients with APS and/or SLE was 44.3% (range:
26.7–55.4%). When focusing on patients w ith APS (either
PAPS or SAPS), the estimated overall prevalence was high at
45.4%. Stratifying for diagnosis, a significantly higher pre-
valence of anti-β
2
GPI-DI antibodies was observed in patients
with APS (either PAPS or SAPS) than those with SLE without
previous history of thrombosis (45.4 vs. 26.7%; p < 0.0001).
Data r egarding the prevalence analysis are summarized
in
►Table 4.
When stratifying patients for inclusion criteria used to
enroll subjects in each study, a frequency of an ti-β
2
GPI-DI
positivity of 38.6% was found in those studies enrolling
patients with APS according to the Sapporo cri-
teria,
35,38–40,43,44
25.1% in studies includ ing patients with
SLE (according to ACR revised criteria),
37
40.3% in patients
with APS (without any details about Sydney or Sapporo
criteria),
36,41
and 57.9% i n those studies that aimed to assess
the frequency of anti-β
2
GPI-DI positivity in presently anti-
β
2
GPI positive patients.
9,42
As would be expected, and due to
selection bias, an overall statistically significant higher anti-
β
2
GPI-DI po sitivity was seen in the two studies that used as
inclusion criteria the presence of anti- β
2
GPI positivity de-
tected at least twice at least 12 weeks apart
9,42
when
compared with the other studies (
►Table 5).
Five out of 11 studies
9,35,36,38,42
were eligible for throm-
botic risk assessment analysis, including a total of 1,014
patients. Patients were distri buted as follow: 821 patients
with APS (358 with PAPS, 179 with SAPS, and 284 not
specified), 163 w ith SLE, 30 aPL asymptomatic car riers,
and 139 HC. Andreoli et al
42
included 100 HC base d on a
previous analysis to set the cutof fs at the 95th percentile
values for anti-β
2
GPI-DI IgG.
20
Three out of 5 studies analyzed IgG is otype with QUANTA
Flash β
2
GPI-DI CLIA assay and used th e same cutoff of
positivity that yielded a 99.5% specificity.
35,36,38
Two out of
5 studies analyzed IgG isotype with ELISA.
9,42
Four studies
reported a significant association between thrombotic
events and positive testing for anti-β
2
GPI-DI an tibo-
dies.
9,19,20,38
In detail, Zhang et al
35
when investigating the
associat ion between thrombotic events and ant i-β
2
GPI-DI
antibodies, reported an odds ratio (OR) of 3.27 (95%, CI 1.59–
6.71), Agmon-Levin et al
36
reported an OR of 2.54 (95%, CI
1.05–6.15) and de Laat et al
9
reported an OR of 3.5 (95%, CI
2.3–5.4). Mahler et al
38
reported an OR of 4 (95%, CI 1 .26–
12.6) with the original c utoff of 20 CU used in the other
studies as well. Furthermore, by optimizing the cutoff of
positivity to increase the likelihood ratio (LR)þ and OR
Seminars in Thrombosis & Hemostasis Vol. 44 No. 5/2018
Prevalence of Antibodies Anti-β
2
GPI-DI in APS Patients Radin et al. 469
Downloaded by: Università degli Studi di Torino. Copyrighted material.
Table 3 Results of the critical appraisal of the included studies according to Newcastle–Ottawa Scale
19
Study Newcastle–Ot tawa qualit y assessment scale
Selection Comparability Outcome
Zhang et al
35
$$$ $ $
Mahler et al
38
$$$ $ $
de Craemer and Devreese
43
$$$ $
Andreoli et al
42
$$ $ $
Cieśla et al
39
$$
Mondejar et al
40
$$$ $
Cousin s et al
44
$$
Agmon-Levin et al
36
$$$ $ $
Wahezi et al
37
$$
Hunt et al
41
$$ $
B. de Laat et al
9
$$ $
Table 2 Characteristics of the test used to identify anti-β
2
GPI-DI antibodies
Study Methodology Cutoff Notes
Zhang et al
35
CIA (QUANTA Flash
assay, Inova
Diagnostics)
20 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
Mahler et al
38
CIA (QUANTA Flash
assay, Inova
Diagnostics)
20 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
de Craemer and Devreese
43
CIA (QUANTA Flash
assay, Inova
Diagnostics)
20 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
Cieśla et al
39
CIA (QUANTA Flash
assay, Inova
Diagnostics)
19.9 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
Mondejar et al
40
CIA (QUANTA Flash
assay, Inova
Diagnostics)
20 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
Agmon-Levin et al
36
CIA (QUANTA Flash
assay, Inova
Diagnostics)
20 CU Recombinant DI coupled to paramagnetic
beads by the use of BIO -FLASH technology
Wahezi et al
37
ELISA in house Mean þ 3SD(HC) β
2
GPI coated on a hydrophobic plate and a
hydrophilic plate
a
Hunt et al
41
ELISA in house Mean þ 3SD(HC) β
2
GPI coated on a hydrophobic plate and a
hydrophilic plate
b
de Laat et al
9
ELISA in house OD (blank þ
three times SD
β
2
GPI coated on a hydrophobic plate and a
hydrophilic plate
c
Andreoli et al
42
ELISA developed by
Inova Diagnostics
95th percentile (HC) Recombinant DI coupled to ELISA plates
Cousin s et al
44
ELISA in house 99th percentile (HC) Binding to purified wild-type and mutant re-
combinant human DI; cCoat ed on Nickel che-
late-coated microwell plates; epitope R39–R43
Abbreviations: CU, chemioluminiscence units; DI, domain I; HC, healthy controls;ELISA,enzyme-linkedimmunosorbentassay;OD,opticaldensity;
SD, standard deviation; β
2
GPI, β
2
glycoprotein 1.
a
Epitope R39– R43.
b
Coated on hydrophobic microtiter plates, goat antihuman IgG alkaline phosphatase labeled antibody.
c
Coated on hydrophobic and hydrophilic pl ates, ep itope R39–R43, mono clonal mouse anti-domai n I antibody (mAb 3B7) .
Seminars in Thrombosis & Hemostasis Vol. 44 No. 5/2018
Prevalence of Antibodies Anti-β
2
GPI-DI in APS Patients Radin et al.470
Downloaded by: Università degli Studi di Torino. Copyrighted material.
