Rheumatology 2001;40:565±573
Paediatric Rheumatology/Series Editor: P. Woo
Anticardiolipin and anti-b2 glycoprotein I
antibodies in sera of 61 apparently healthy
children at regular preventive visits
T. AvcÏin, A. AmbrozÏ icÏ1, M. Kuhar, T. Kveder1 and B. Rozman1
Department of Paediatrics and 1Department of Rheumatology,
University Medical Centre, Ljubljana, Slovenia
Abstract
Objectives. To determine anticardiolipin (aCL) and anti-b2 glycoprotein I antibodies
(anti-b2GPI) in apparently healthy children and express the cut-off levels in concentrations of
monoclonal antibodies, and to compare the mean values and frequencies of aCL and anti-b2GPI
in children with those in blood donors.
Methods. Blood samples were collected from 29 preschool children and 32 adolescents during
their routine preventive follow-up visits. The control group consisted of 52 blood donors. aCL
and anti-b2GPI were assayed by an ELISA method. Two monoclonal b2GPI-dependent aCL
(HCAL and EY2C9) were used as calibrators.
Results. The estimated cut-off values for immunoglobulin G (IgG) and immunoglobulin M
(IgM) aCL, expressed in concentrations of monoclonal antibodies and standardized
international units (GPLuMPL units), were 13.9 nguml (7.6 GPL) and 33.1 nguml (3.3 MPL)
for preschool children, 13.5 nguml (7.2 GPL) and 36.9 nguml (4.0 MPL) for adolescents, and
14.4 nguml (8.0 GPL) and 42.6 nguml (5.1 MPL) for blood donors. No statistically signi®cant
differences in the mean values for IgG and IgM aCL were found between the age groups. The
mean value of IgA aCL was signi®cantly higher in blood donors than in preschool children and
adolescents (P < 0.037 and P < 0.025 respectively). Seven (11.4%) of 61 apparently healthy
children had low positive values for aCL (IgG for all seven). The estimated cut-off values for IgG
and IgM anti-b2GPI were 4.2 and 13.1 nguml respectively for preschool children, 3.2 and
13.1 nguml for adolescents, and 2.9 and 20.5 nguml for blood donors. The mean value for IgG
anti-b2GPI was found to be higher in preschool children than in adolescents and blood donors
(P < 0.0001 and P < 0.0001). The mean values for IgM and IgA anti-b2GPI were higher in blood
donors than in preschool children (IgM, P < 0.007; IgA, P < 0.0001) and adolescents (IgM,
P < 0.01; IgA, P < 0.0001). Four (6.6%) of 61 apparently healthy children had positive values
for anti-b2GPI (two for IgG and two for IgA).
Conclusions. This is the ®rst report in which the cut-off values for aCL and anti-b2GPI in
children are expressed in concentrations of monoclonal antibodies. Low titres of aCL, which
were identi®ed frequently in apparently healthy children, were hypothesized to be the result of
previous infections. The high mean value of IgG anti-b2GPI observed in preschool children was
an unexpected result of the study and might indicate a default response to nutritional exposure to
b2GPI in this age group.
KEY WORDS: Antiphospholipid antibodies, Anticardiolipin antibodies, Anti-b2 glycoprotein I
antibodies, Healthy children.
Antiphospholipid antibodies (aPL) have been well
recognized in patients with autoimmune diseases w1±3x
as well as in association with various infections w4x.
They have also been described, but less commonly,
in individuals without any underlying disorders w5±7x.
As with most other autoantibodies, an increased frequency of aPL has been reported in the healthy elderly
population compared with healthy adults w5, 8x. As high
levels of aPL in the elderly population are not usually
associated with any clinical disease, it has been generally
Submitted 25 April 2000; revised version accepted 14 November
2000.
Correspondence to: T. AvcÏin, Division of Allergy, Immunology and
Rheumatology, Department of Paediatrics, University Medical Centre,
Vrazov trg 1, SI-1000 Ljubljana, Slovenia.
565
ß 2001 British Society for Rheumatology
566
T. AvcÏin et al.
assumed that this autoimmune phenomenon might be
a manifestation of age-related senescence of the immune
system w5, 8x.
In recent years, the clinical features of aPL have been
recognized increasingly in children, and a growing
number of reports have provided information on the
frequency of aPL in childhood disorders w9, 10x. However, there are few data available on the normal values
and the frequency of aPL in healthy children.
Among the best studied and clinically most widely
accepted aPL are anticardiolipin (aCL) antibodies.
Several studies have investigated aCL in childhood
disorders, but in most cases the control group for the
normal population was poorly de®ned. Some studies
have used healthy adult individuals as the control
group w11±13x while others have used apparently healthy
children w14±20x. In most cases, the source of apparently
healthy children was not stated clearly w15, 18x, while
sometimes sera from children hospitalized for minor
surgical or psychosomatic problems were used w14, 16,
17, 19x. There were also substantial differences in the
methods used for enzyme-linked immunosorbent assay
(ELISA) and in the methods of statistical analysis used
for de®ning cut-off points for immunoglobulin G (IgG)
and immunoglobulin M (IgM) aCL. Consequently, no
accurate information has been produced on normal
ranges of IgG and IgM aCL in children and comparisons
between studies have not been possible.
Recently, it has become clear that most autoimmune
aPL are directed against b2 glycoprotein I (b2GPI)
w21±23x, and anti-b2GPI antibodies (anti-b2GPI) have
been associated with the clinical features of the antiphospholipid syndrome w24, 25x. Because aPL induced
by infections do not require any plasma protein for
binding to negatively charged phospholipids, the detection of anti-b2GPI might also be a useful marker for
discriminating between presumably pathogenic autoimmune and post-infectious aPL w26, 27x. Furthermore,
human and chimaeric monoclonal antibodies against
b2GPI have been developed which could serve as the
standard for aCL and anti-b2GPI ELISA w28, 29x. To
our knowledge, no study has addressed the presence
of anti-b2GPI in healthy children or has used concentrations of monoclonal antibodies to express aCL and
anti-b2GPI values in children.
In our study, the values of aCL and anti-b2GPI
isotypes in apparently healthy children were examined
and the cut-off levels were expressed in standardized
international units (GPLuMPL units) and concentrations
of monoclonal antibodies. In addition, the mean values
and frequencies of aCL and anti-b2GPI in children were
compared with those in blood donors.
Patients and methods
Participants
Blood samples were collected from 61 apparently
healthy children during their routine preventive
follow-up visits in community health centres between
March and July 1999. The children included in the study
did not have clinical infections at the time of examination or any documented serious medical problems. On
the basis of their age they were divided into two groups:
group I consisted of preschool children (15 boys and
14 girls, mean age 5.0 yr, range 4.9±5.3) and group II
consisted of adolescents (10 boys and 22 girls, mean age
13.5 yr, range 13.0±14.2).
Blood was drawn by clean venipuncture and collected
into glass tubes. Informed consent for drawing extra
blood at the time of routine venipuncture was obtained
from the parents of all participating children. As the
control, we used sera from 52 healthy blood donors
(31 men and 21 women, mean age 34.0 yr, range 18.0±
65.0). All sera were separated within 2 h after collection,
aliquoted and stored at 2308C until assayed. The study
was approved by the Ethics Committee of the Slovenian
Ministry of Health.
Detection of aCL antibodies
aCL antibodies were detected by a slightly modi®ed aCL
ELISA, described previously w30±32x. Microtitre plates
(Costar Medium Binding EIAuRIA plates; Costar,
Cambridge, MA, USA) were coated with 40 mluwell of
cardiolipin (Sigma, St Louis, MO, USA) in ethanol and
allowed to evaporate at 48C overnight. After incubation with 120 mluwell of 10% fetal bovine serum (Sigma)
in phosphate-buffered saline, pH 7.4 (FBS±PBS) for 1 h
at room temperature (22±268C) the plates were washed
once with 300 mluwell of PBS. Then 100 mluwell of
standards and serum samples diluted 1 : 100 in 10%
FBS±PBS were added in duplicate and the plates were
incubated for 2.5 h at room temperature. The plates
were washed four times with PBS, and 100 mluwell of
alkaline phosphatase-conjugated goat anti-human IgG,
IgM or IgA (ACSC, Westbury, NY, USA) diluted in
10% FBS±PBS was added. After 1 h of incubation at
room temperature, the plates were washed four times
with PBS and 100 mluwell of p-nitrophenyl phosphate
(Sigma) dissolved at 1 mguml in 1 M diethanolamine
buffer (pH 9.8) was added. Optical density at 405 nm
(OD) was measured ®rst after 10 min and then every
3 min by a microtitre plate reader Rainbow Spectra
Thermo (Tecan Austria GmbH, GroÈdiguSalzburg,
Austria) vs a reagent blank until optimal ®tting to
predicted OD of standards was obtained. To prevent
the in¯uence of interassay variation, sera from all three
groups were selected randomly for each run. The plate
positioning of sera was also changed for each run to
prevent possible edge effects affecting the ®nal results.
Detection of anti-b2GPI antibodies
b2GPI was puri®ed by the perchloric acid precipitation
method using the same principles as described previously w33x. A modi®ed anti-b2GPI ELISA was
developed on the basis of a method described previously w34±36x and our preliminary results. Microtitre
plates (Costar High Binding EIAuRIA plates) were
coated with 50 mluwell of b2GPI dissolved at 10 mguml
in PBS for 2 h at room temperature. After one wash with
aCL and anti-b2GPI in healthy children
567
FIG. 1. Comparison of IgG, IgM and IgA aCL antibody mOD values in preschool children (group I), adolescents (group II) and
blood donors (group III). The data are means of results from two consecutive tests for each participant. The horizontal lines
indicate means of pooled data for each age group.
200 ml of PBS containing 0.05% Tween 20 (PBS±Tween),
50 mluwell of standards and serum samples diluted 1 : 100
in PBS±Tween were applied in duplicate and incubated
for 30 min at room temperature. After four washes,
50 mluwell of alkaline phosphatase-conjugated goat antihuman IgG, IgM or IgA (ACSC) diluted in PBS was
added. After 30 min of incubation at room temperature
and four washes, 100 mluwell of substrate was added. OD
was measured as described for the aCL ELISA.
Monoclonal antibodies
Two monoclonal b2GPI-dependent aCL were used as
calibrators: (i) the chimaeric IgG monoclonal antibody
HCAL, consisting of human k and c1 constant regions
and variable regions from the mouse monoclonal
b2GPI-dependent aCL WBCAL-1 w29, 37x; and (ii) the
IgM monoclonal antibody EY2C9, derived from a
patient with antiphospholipid syndrome w28x. Monoclonal antibodies were aliquoted in amounts suf®cient
for one assay and frozen at 2208C. Monoclonal antibodies in appropriate concentrations were used to create
reference curves. Calibrations according to original
KAPS (Kingston Anti-phospholipid Antibody Study)
standards w38, 39x and calibrations according to monoclonal antibodies were performed simultaneously in the
same plates as the tested sera, to allow more accurate
comparisons.
Statistical analysis
The distributions of OD and log OD were assessed by
the use of histograms. Because of the skewed frequency
distribution of OD in the aCL and anti-b2GPI ELISAs,
log-transformed data were used in the statistical test
procedures. Student's t-test was used to compare the
mean values and Fischer's exact test was used where
appropriate. The statistical signi®cance of intergroup
frequency rate was determined by x2 analysis. All
statistical analyses were performed using the program
MS Excel 6.0 for Windows. Differences were considered
statistically signi®cant whenever P < 0.05.
FIG. 2. Frequency distributions of log OD for aCL antibody
isotypes in preschool children (group I), adolescents (group II)
and blood donors (group III).
Results
Distributions of aCL
The cumulative results of aCL assays in all of the
participants studied are shown in Fig. 1. The frequency
distributions of OD for IgG, IgM and IgA aCL isotypes in the three age groups demonstrated a nonsymmetrical, positively skewed main peak with some
outliers. For IgG aCL, the data were slightly more
dispersed than for the other two isotypes. Additionally,
in group II there was a trend towards a bimodal
distribution of the OD for IgG aCL. To reduce the
skewness of the main peaks, OD measurements were
transformed to the log scale. The frequency distributions of log OD for aCL were nearly symmetrical,
showing less skewness than the distributions for the
untransformed measurements (Fig. 2).
T. AvcÏin et al.
568
TABLE 1. Estimated cut-off points for aCL antibodies in groups I, II and III
IgG
95th percentile (OD)
Mean+2 S.D. (OD)
GPLuMPL units
mAb (nguml)
IgM
IgA
I
II
III
I
II
III
I
II
III
0.114
0.156
7.6
13.9
0.100
0.151
7.2
13.5
0.133
0.158
8.0
14.4
0.064
0.083
3.3
33.1
0.071
0.095
4.0
36.9
0.084
0.112
5.1
42.6
0.030
0.048
±
±
0.034
0.051
±
±
0.041
0.056
±
±
mAb, concentration of monoclonal antibodies HCALuEY2C9.
FIG. 3. (A) Relationships between mOD and concentrations of HCALuEY2C9 in the aCL antibody assay. (B) Relationships
between GPLuMPL units and concentrations of HCAL/EY2C9 in aCL antibody assay. The concentration of HCAL can be
converted to GPL units with the formula GPL units=0.978 3 wconcentration of HCAL (nguml)x 25.542 (R2=0.999),2 and the
concentration of EY2C9 to MPL units with the formula MPL units=0.012 3 wconcentration of EY2C9 (nguml)x1.612 (R =0.999).
Estimation of cut-off points and comparison of mean
values for aCL between age groups
Because the frequency distributions for the log OD were
nearly symmetrical, the cut-off points were de®ned as
the mean value plus two standard deviations for the log
OD values. Corresponding values for the OD were then
computed by taking the anti-logarithm of the values for
the log OD. The estimated cut-off points for aCL
expressed as OD were transformed into GPLuMPL
KAPS units and concentrations of monoclonal antibodies (Table 1). The relationships between the mOD
(1023 OD), GPLuMPL units and appropriate concentrations of HCALuEY2C9 are presented in Fig. 3,
the legend of which gives equations for converting the
concentration of HCAL into GPL units and the
concentration of EY2C9 into MPL units.
Mean aCL values obtained from the log-transformed
data were also compared between different age groups.
For IgG and IgM aCL there were no statistically
signi®cant differences in the mean values of antibodies
between age groups. However, the mean value of IgA
aCL was signi®cantly higher in group III than in groups
I and II (P < 0.037 and P < 0.025 respectively).
Frequency of elevated aCL
According to our cut-off points, a total of 7u61 (11.4%)
apparently healthy children (groups I and II combined)
had low positive values for aCL. They were all positive
for only IgG aCL, and no child had elevated IgM or
IgA aCL. In the blood donors, the overall frequency
of aCL positivity was 9.6% (5u52); three (5.8%) aCLpositive donors had a low titre of IgG and two (3.8%)
had a low titre of IgA aCL antibodies. The differences
in frequency between apparently healthy children and
blood donors were not statistically signi®cant for either
aCL isotype.
Distributions of anti-b2GPI
The cumulative results of the anti-b2GPI assays in all
participants studied are shown in Fig. 4. The frequency
distributions of OD for anti-b2GPI were more uniform
than for aCL. Each of the three isotypes of anti-b2GPI
demonstrated one asymmetrical, positively skewed
main peak with some outliers, and the same pattern
was observed in all three age groups. As with aCL,
the frequency distributions of log OD for anti-b2GPI
were nearly symmetrical, showing less skewness than
the distributions for untransformed measurements
(Fig. 5).
Estimation of cut-off points and comparison of mean
values for anti-b2GPI between age groups
The cut-off points for anti-b2GPI isotypes were determined by the same principle as for aCL. The estimated
values were expressed as OD and transformed into
concentrations of monoclonal antibodies (Table 2). The
relationships between OD and concentrations of HCAL
and EY2C9 are presented in Fig. 6.
Mean values for anti-b2GPI were also compared
between age groups. The mean value for IgG anti-b2GPI
aCL and anti-b2GPI in healthy children
569
TABLE 2. Estimated cut-off points for anti-b2 GPI antibodies in groups I, II and III
IgG
95th percentile
Mean+2 S.D. (OD)
mAb (nguml)
IgM
IgA
I
II
III
I
II
III
I
II
III
0.050
0.068
4.2
0.041
0.052
3.2
0.037
0.044
2.9
0.059
0.059
13.1
0.057
0.061
13.1
0.071
0.092
20.5
0.031
0.038
±
0.039
0.040
±
0.043
0.057
±
mAb, concentration of monoclonal antibodies HCALuEY2C9.
FIG. 4. Comparison of IgG, IgM and IgA anti-b2GPI mOD values in preschool children (group I), adolescents (group II) and
blood donors (group III); the data are means of results of two consecutive tests for each participant. The horizontal lines indicate
means of pooled data for each age group.
was higher in group I than in groups II and III
(P < 0.0001 and P < 0.0001 respectively). In contrast,
the mean values for IgM and IgA anti-b2GPI were
higher in group III than in group I (IgM, P < 0.007;
IgA, P < 0.0001) and group II (IgM, P < 0.01; IgA,
P < 0.0001).
Frequency of elevated anti-b2GPI
A total of 4u61 (6.6%) apparently healthy children
(groups I and II combined) had positive values for antib2GPI. They were positive for either IgG (3.3%; 2u61)
or IgA (3.3%; 2u61) anti-b2GPI. The overall frequency
of anti-b2GPI-positive blood donors was 7.7% (4u52);
positive donors exhibited either IgG (1.9%; 1u52) or IgA
(5.8%; 3u52) anti-b2GPI. As with aCL, the intergroup
differences in frequency for all three anti-b2GPI isotypes
were not statistically signi®cant.
Comparison between aCL and anti-b2GPI assays
Among seven apparently healthy children who were
positive in the aCL assay, only one was also positive in
the anti-b2GPI assay; similarly, among ®ve aCL-positive
blood donors only one was also anti-b2GPI-positive. We
also found that, among four apparently healthy children
who were positive in the anti-b2GPI assay, only one was
also positive for aCL. Similarly, among four antib2GPI-positive blood donors only one was also positive
in the aCL assay.
FIG. 5. Frequency distributions of log OD for anti-b2GPI
isotypes in preschool children (group I), adolescents (group II)
and blood donors (group III).
Discussion
In the present study, all serum samples were taken from
apparently healthy children during their regular routine
preventive visits in community health centres. Their
health was assessed as part of a preventive visit before
570
T. AvcÏin et al.
FIG. 6. Relationships between mOD and concentrations of
HCAL and EY2C9 in the anti-b2GPI antibody assay. The
concentration of HCAL can be converted to mOD with the
formula mOD=16.941 3 wconcentration of HCAL (nguml)x0.935
(R2=0.999), and the concentration of EY2C9 to mOD
with the formula mOD=5.592 3 wconcentration of EY2C9
(nguml)x0.920 (R2=0.999).
blood was drawn, and those failing to pass this health
test were excluded from the study. Our results therefore
demonstrated unequivocally the presence of low levels
of aCL and anti-b2GPI in sera of apparently healthy
children.
To quantitate the risk associated with aCL in childhood disorders, it is essential to establish the normal
range in healthy children. There has been considerable
controversy over the most appropriate way of de®ning
`normal' and `abnormal' levels of aCL. Gattorno et al.
w18x considered an aCL test to be positive when the
values were at least two standard deviations above the
mean value for 25 healthy children. SiamopoulouMavridou et al. w15x considered the cut-off value to be
the mean plus three standard deviations of the 66 normal
sera obtained in children without rheumatic disease.
In three studies performed in the same laboratory w14,
16, 17x, the upper limits for aCL were set as the mean
plus ®ve standard deviations of 42 apparently healthy
children who were hospitalized for minor surgical
procedures or for bone marrow donation.
Our results demonstrate that, for aCL, the frequency
distributions of OD in apparently healthy children were
not normal but had a pronounced positive skew. Similar
asymmetrical distributions of OD for aCL and for
most other autoantibodies have been observed in the
healthy adult population w6, 40x. For this reason, nonparametric tests were preferred, and these gave better
estimates of normal ranges. Serra et al. w20x therefore
used the percentile principle and set the cut-off point as
the 95th percentile for 52 healthy children. However, to
obtain a more accurate estimate of the 95th percentile
requires a larger sample size than that used by Serra et al.
In this study, we used the method proposed by Verrier
Jones et al. w40x, in which OD measurements were
transformed to the log scale, thus substantially reducing the skewness. Although not Gaussian, the frequency distribution of log-transformed data was nearly
symmetrical and hence the cut-off point of the mean
value plus two standard deviations could be used as the
estimate of the 95th percentile. Our sample size (29
preschool children and 32 adolescents) seems a reasonable compromise to estimate the mean values and standard deviations accurately. Furthermore, such analysis of
log-transformed data appears to be more appropriate
than some of the non-parametric tests performed on
a small number of untransformed data. However, in
spite of the methodological differences discussed above,
our results were consistent with the values obtained by
Serra et al. w20x, who de®ned abnormal aCL values in
healthy children as being equal to or higher than 9 GPL
and 4 MPL. To our knowledge, the present study was
the ®rst in which the cut-off values for IgG and IgM
aCL in children were expressed in concentrations of
monoclonal antibodies, which could easily be compared
between different laboratories using the same detection
method. Monoclonal antibodies for IgA aCL were not
available; therefore the cut-off values for IgA aCL were
expressed only in OD. Mean values of IgG and IgM
aCL were comparable between different age groups and
only the mean value of IgA aCL was found to be
signi®cantly higher in blood donors than in preschool
children (P < 0.037) and adolescents (P < 0.025).
The frequency of low titres of aCL was high in
apparently healthy children, and was similar to values
obtained in healthy adults. Because aCL are also
expressed frequently in various viral and bacterial infections w4x that are commonly acquired in childhood, we
speculate that some positive aCL values in apparently
healthy children could be the result of previous infections. This hypothesis is supported by the observation
that six out of seven children positive in the aCL assay
were negative in the anti-b2GPI assay.
We also established the cut-off values for IgG, IgM
and IgA anti-b2GPI isotypes. Like those of aCL, the
distributions of OD for anti-b2GPI isotypes were positively skewed, and transforming the results to the log
scale substantially reduced the skewness. Therefore, the
same statistical test procedures as those used for
aCL were used to determine the cut-off values. The
estimated cut-off values for IgG and IgM anti-b2GPI
were expressed as HCAL and EY2C9 concentrations
while the cut-off values for IgA anti-b2GPI were
expressed only in OD for the same reasons as for aCL.
One of the most remarkable ®ndings of the present
study was that the mean value of IgG anti-b2GPI was
highest in preschool children and in this group it was
signi®cantly higher than in adolescents (P < 0.0001) and
blood donors (P < 0.0001). It appears, therefore, that
the mean value of IgG anti-b2GPI decreases after the
preschool period, in spite of the fact that the total IgG
level increases gradually and progressively w41x. The
relatively high mean value of IgG anti-b2GPI in preschool children might be related to several factors. First,
it might be in¯uenced by absorption of nutritional
b2GPI at gastrointestinal surfaces and might therefore
represent a possible default response to mucosal
challenge in infants. It is well known that b2GPI has
aCL and anti-b2GPI in healthy children
been remarkably conserved during the evolution of
animal species w42x, and thus it is possible that ingestion
of bovine or some other type of b2GPI found in different
milk or meat products could act as a peroral immunization agent and induce transitory production of antib2GPI in infants, in whom the intestinal mucosa is more
permissive for large molecules w43, 44x. Secondly, some
previous exposure to viral infection, together with
nutritional exposure to b2GPI, might have an adjuvant
effect on antibody production. An opposite trend to that
for IgG anti-b2GPI was observed for IgM and IgA
isotypes of anti-b2GPI, for which signi®cantly higher
mean values were obtained in blood donors than in
preschool children (IgM, P < 0.007; IgA, P < 0.0001)
and adolescents (IgM, P < 0.01; IgA, P < 0.0001).
These ®ndings were in accordance with the increasing
levels of total IgM and IgA after birth w41x.
The frequency of anti-b2GPI positivity in apparently
healthy children was lower than that of aCL. However,
among four anti-b2GPI-positive children only one was
simultaneously positive for both anti-b2GPI and aCL.
Three children clearly exhibited the presence of antib2GPI without aCL, which was surprising and could
indicate that anti-b2GPI might be present without aCL
in children. The explanation of this phenomenon could
be that the subset of sera positive only for anti-b2GPI
contain antibodies directed to epitopes expressed on
b2GPI when attached to a high-binding microtitre plate
but not when bound to cardiolipin w45x. In our study, the
ELISA for anti-b2GPI was performed without a
blocking step because of the relatively frequent occurrence of antibodies recognizing the most commonly used
blocking agents (bovine serum albumin, gelatin), giving
high background readings. On the other hand, sample
blank values obtained by our modi®cation of the antib2GPI ELISA were negligible (data not shown). In our
opinion, no additional interference effects were introduced into the assay system by using only PBS and
Tween.
To provide an additional illustration of the message
of this article, we report preliminary data on aCL and
anti-b2GPI titres in children with juvenile idiopathic
arthritis (unpublished data). According to our cut-off
points, slightly or moderately elevated aCL antibody
concentrations were found in 46.4% of 28 children with
juvenile idiopathic arthritis, and the level was usually
below 37 nguml HCAL (30 GPL) and 85 nguml EY2C9
(15 MPL). The frequency of anti-b2GPI in our children
with juvenile idiopathic arthritis was 10.7% and the
quantitative expression of anti-b2GPI was usually below
10 nguml HCAL and 20 nguml EY2C9.
In conclusion, in the present study the cut-off values
for aCL and anti-b2GPI were determined in apparently
healthy children and were expressed as concentrations of
HCAL and EY2C9 monoclonal antibodies. This could
facilitate clinical comparisons in childhood disorders
and enables comparison of data derived from different
laboratories. Low titres of aCL were common in
apparently healthy children; this might have been the
result of previous infections. Therefore, the clinical
571
signi®cance of low titres of aCL in childhood disorders
has to be judged critically. The higher mean value of
IgG anti-b2GPI observed in preschool children than in
adolescents and adults is an unexpected result of the
study. It has not been demonstrated previously and
might indicate a default response to nutritional exposure
to b2GPI in this age group.
Acknowledgements
We wish to thank Professor Takao Koike (Hokkaido
University School of Medicine, Sapporo, Japan) for
kindly providing the monoclonal antibodies HCAL and
EY2C9, Dr Ksenija Jelenc (Community Health Centre
BezÏigrad, Ljubljana, Slovenia) for help in collecting sera
from apparently healthy children and Assistant
Professor Borut BozÏicÏ for his critical reading of the
manuscript and for helpful suggestions. This study was
supported in part by a grant from The Slovenian
Ministry of Science and Technology (Grant J3 7924).
References
1. AlarcoÂn-Segovia D, Deleze M, Oria CV et al.
Antiphospholipid antibodies and the antiphospholipid
syndrome in systemic lupus erythematosus: a prospective
analysis of 500 consecutive patients. Medicine 1989;
68:353±65.
2. Love PE, Santoro SA. Antiphospholipid antibodies:
anticardiolipin and the lupus anticoagulant in systemic
lupus erythematosus (SLE) and in non-SLE disorders:
prevalence and clinical signi®cance. Ann Intern Med
1990;112:682±98.
3. Hughes GRV. The antiphospholipid syndrome: ten years
on. Lancet 1993;342:341±4.
4. Vaarala O, Palosuo T, Kleemola M, Aho K. Anticardiolipin response in acute infections. Clin Immunol
Immunopathol 1986;41:8±15.
5. Manoussakis MN, Tzioufas AG, Silis MP, Pange PJE,
Goudevenos J, Moutsopoulos HM. High prevalence of
anti-cardiolipin and other autoantibodies in a healthy
elderly population. Clin Exp Immunol 1987;69:557±65.
6. El-Roeiy A, Gleicher N. De®nition of normal autoantibody levels in an apparently healthy population.
Obstet Gynecol 1988;72:596±602.
7. Lockwood CJ, Romero R, Feinberg RF, Clyne LP,
Coster B, Hobbins JC. The prevalence and biologic
signi®cance of lupus anticoagulant and anticardiolipin
antibodies in a general obstetric population. Am J Obstet
Gynecol 1989;161:369±73.
8. Fields RA, Toubbeh H, Searles RP, Bankhurst AD.
The prevalence of anticardiolipin antibodies in a healthy
elderly population and its association with antinuclear
antibodies. J Rheumatol 1989;16:623±5.
9. Tucker LB. Antiphospholipid syndrome in childhood: The
great unknown. Lupus 1994;3:367±9.
10. Ravelli A, Martini A. Antiphospholipid antibody syndrome in pediatric patients. Rheum Dis Clin North Am
1997;23:657±76.
572
T. AvcÏin et al.
11. Shergy WJ, Kredich DW, Pisetsky DS. The relationship
of anticardiolipin antibodies to disease manifestations
in pediatric systemic lupus erythematosus. J Rheumatol
1988;15:1389±94.
12. Malleson PN, Fung MY, Petty RE, Mackinnon MJ,
Schroeder ML. Autoantibodies in chronic arthritis of
childhood: relations with each other and with histocompatibility antigens. Ann Rheum Dis 1992;51:1301±6.
13. Molta C, Meyer O, Dosquet C et al. Childhood-onset
systemic lupus erythematosus: antiphospholipid antibodies
in 37 patients and their ®rst-degree relatives. Pediatrics
1993;92:849±53.
14. Caporali R, Ravelli A, De Gennaro F, Neirotti G,
Montecucco C, Martini A. Prevalence of anticardiolipin
antibodies in juvenile arthritis. Ann Rheum Dis 1991;
50:599±601.
15. Siamopoulou-Mavridou A, Mavridis AK, Terzoglou C,
Filopoulou S, Athanasiadou S, Tzioufas AG. Autoantibodies in Greek juvenile chronic arthritis patients.
Clin Exp Rheumatol 1991;9:647±52.
16. Angelini L, Ravelli A, Caporali R, Rumi V, Nardocci N,
Martini A. High prevalence of antiphospholipid antibodies in children with idiopathic cerebral ischemia.
Pediatrics 1994;94:500±3.
17. Ravelli A, Caporali R, Di Fuccia G, Zonta L, Montecucco
C, Martini A. Anticardiolipin antibodies in pediatric systemic lupus erythematosus. Arch Pediatr Adolesc Med
1994;148:398±402.
18. Gattorno M, Buoncompagni A, Molinari AC et al.
Antiphospholipid antibodies in paediatric systemic lupus
erythematosus, juvenile chronic arthritis and overlap
syndromes: SLE patients with both lupus anticoagulant
and high-titre anticardiolipin antibodies are at risk for
clinical manifestations related to the antiphospholipid
syndrome. Br J Rheumatol 1995;34:873±81.
19. Kontiainen S, Miettinen A, SeppaÈlaÈ I, Verkasalo M,
MaÈenpaÈaÈ J. Antiphospholipid antibodies in children. Acta
Paediatr 1996;85:614±5.
20. Serra CRB, Rodrigues SH, Silva NP, Sztajnbok FR,
Andrade LEC. Clinical signi®cance of anticardiolipin
antibodies in juvenile idiopathic arthritis. Clin Exp
Rheumatol 1999;17:375±80.
21. McNeil HP, Simpson RJ, Chesterman CN, Krilis SA. Antiphospholipid antibodies are directed against a complex
antigen that includes a lipid-binding inhibitor of coagulation: b2-glycoprotein I (apolipoprotein h). Proc Natl Acad
Sci USA 1990;87:4120±4.
22. Galli M, Comfurius P, Maassen C et al. Anticardiolipin
antibodies (ACA) directed not to cardiolipin but to a
plasma protein cofactor. Lancet 1990;335:1544±7.
23. Matsuura E, Igarashi Y, Fujimoto M, Ichikawa K,
Koike T. Anticardiolipin cofactor(s) and differential
diagnosis of autoimmune disease. Lancet 1990;336:177±8.
24. Balestrieri G, Tincani A, Spatola L et al. Antib2-glycoprotein I antibodies: a marker of antiphospholipid
syndrome. Lupus 1995;4:122±30.
25. Cabiedes J, Cabral AR, AlarcoÂn-Segovia D. Clinical
manifestations of the antiphospholipid syndrome in
patients with systemic lupus erythematosus associate more
strongly with b2-glycoprotein I than with antiphospholipid
antibodies. J Rheumatol 1995;22:1899±906.
26. Hunt JE, McNeil HP, Morgan GJ, Crameri RM,
Krilis SA. A phospholipid-b2-glycoprotein I complex is
an antigen for anticardiolipin antibodies occurring in
autoimmune disease but not with infection. Lupus
1992;1:75±81.
27. McNally T, Purdy G, Mackie IJ, Machin SJ, Isenberg DA.
The use of an anti-b2-glycoprotein-I assay for discrimination between anticardiolipin antibodies associated with
infection and increased risk of thrombosis. Br J Haematol
1995;91:471±3.
28. Ichikawa K, Khamashta MA, Koike T, Matsuura E,
Hughes GRV. b2-Glycoprotein I reactivity of monoclonal
anticardiolipin antibodies from patients with the
antiphospholipid syndrome. Arthritis Rheum 1994;
37:1453±61.
29. Ichikawa K, Tsutsumi A, Atsumi T et al. A chimeric
antibody with the human c1 constant region as a putative
standard for assays to detect IgG b2-glycoprotein
I-dependent anticardiolipin and anti-b2-glycoprotein I
antibodies. Arthritis Rheum 1999;42:2461±70.
30. Loizou S, McCrea JD, Rudge AC, Reynolds R, Boyle CC,
Harris EN. Measurement of anti-cardiolipin antibodies by
an enzyme-linked immunosorbent assay (ELISA): standardization and quantitation of results. Clin Exp Immunol
1985;62:738±45.
31. Stegnar M, BozÏicÏ B, Peternel P, Kveder T, Vene N,
Rozman B. Prevalence of antiphospholipid antibodies
in deep vein thrombosis and their relationship to
blood coagulation and ®brinolysis. Thromb Res 1991;
63:433±43.
32. Roubey RAS, Maldonado MA, Byrd SN. Comparison
of an enzyme-linked immunosorbent assay for antibodies
to b2-glycoprotein I and conventional anticardiolipin
immunoassay. Arthritis Rheum 1996;39:1606±7.
33. Polz E, Wurm H, Kostner GM. Investigations on beta
2-glycoprotein-I in the rat: isolation from serum and
demonstration in lipoprotein density fractions. Int J
Biochem 1980;11:265±70.
34. Teixido M, Font J, Reverter JC, Cervera R, Tassies D,
Ingelmo M et al. Anti-b2-glycoprotein I antibodies:
A useful marker for the antiphospholipid syndrome.
Br J Rheumatol 1997;36:113±6.
35. Geurin J, Feighery C, Sin RB, Jackson J. Antibodies to
b2-glycoprotein IÐa speci®c marker for the antiphospholipid syndrome. Clin Exp Immunol 1997;109:304±9.
36. Erickson EN, Najmey SS, Keil LB, El-Kadi HS,
DeBari VA. Reference calibrators for IgG antibodies to
b2-glycoprotein I: preparation, properties and availability
to investigators. Clin Chem 1996;42:1116±7.
37. Hashimoto Y, Kawamura M, Ichikawa K, Suzuki T,
Sumida T, Yoshida S et al. Anticardiolipin antibodies in
NZW 3 BXSB F1 mice. A model of antiphospholipid
syndrome. J Immunol 1992;149:1063±8.
38. Harris EN, Gharavi AE, Patel SP, Hughes GRV.
Evaluation of the anti-cardiolipin antibody test: report of
an international workshop held 4 April 1986. Clin Exp
Immunol 1987;68:215±22.
39. Harris EN. The Second International Anti-Cardiolipin
Standardization
WorkshopuThe
Kingston
AntiPhospholipid Antibody Study (KAPS) group. Am J Clin
Pathol 1990;94:476±84.
40. Verrier Jones J, Eastwood BJ, Jones E, James H, Mansour
M. Antiphospholipid antibodies in a healthy population:
methods for estimating the distribution. J Rheumatol
1995;22:55±61.
41. Stiehm ER, Fudenberg HH. Serum levels of immune
globulins in health and disease: a survey. Pediatrics
1966;37:715±27.
42. Matsuura E, Igarashi M, Igarashi Y, Nagae H,
Ichikawa K, Yasuda T et al. Molecular de®nition
of human b2-glycoprotein I (b2-GPI) by cDNA
aCL and anti-b2GPI in healthy children
cloning and interspecies differences of b2-GPI in
alteration of anticardiolipin binding. Int Immunol
1991;3:1217±21.
43. Jakobsson I, Lindberg T, Lothe L, Axelsson I,
Benediktsson B. Human a-lactalbumin as a marker of
macromolecular absorption. Gut 1986;27:1029±34.
573
44. Kuitunen M, Savilahti E, Sarnesto A. Human a-lactalbumin
and bovine b-lactoglobulin absorption in infants. Allergy
1994;49:354±60.
45. Tincani A, Spatola L, Cinquini M, Cattaneo R, Meroni P,
Balestrieri G. Anti beta2-glycoprotein I antibodies: clinical
signi®cance. Lupus 1998;7(Suppl. 2):S107±9.