Neuropsychological development of children born to patients with antiphospholipid syndrome

Authors


Abstract

Objective

To verify the occurrence of learning disabilities (LDs) in the offspring of women with primary antiphospholipid syndrome (APS) as a consequence of fetal exposure to maternal antiphospholipid antibodies (aPL), and to evaluate the impact of maternal chronic disease on children's development.

Methods

We studied 17 children of mothers with primary APS using a standardized intelligence test (Wechsler Intelligence Scale for Children, Revised), a specific LD battery of tests (Sartori, MT groups' test for reading ability, MT groups' test for math skills), and a questionnaire on behavioral and social characteristics (Child Behavior Checklist [CBCL]). Mothers were interviewed about their pregnancy and motherhood experience.

Results

All children had a normal intelligence level (full-scale intelligence quotient >85); 15 pregnancies occurred in mothers with IgG aPL. LDs were diagnosed in 4 children (26.7%), 2 boys and 2 girls. One of these children was born premature, with a brother also affected. Four children (26.7%) showed a higher risk to present problems on the CBCL total competence scale and 2 children (13.3%) on the CBCL total behavior scale. Two children were described as hyperactive (1 had an LD). All families had a good socioeconomic status and educational level.

Conclusion

Besides prematurity and genetic and environmental factors, the genesis of LDs may also include in utero exposure to aPL, in agreement with described experimental models and patients with systemic lupus erythematosus. Socioeconomic status does not seem to influence the occurrence of LDs. A long-term multidisciplinary followup may improve quality of life in patients with primary APS and their children.

INTRODUCTION

Primary antiphospholipid syndrome (APS) is an autoimmune disorder clinically characterized by venous or arterial thrombosis and recurrent fetal loss in the presence of stable positivity of at least 1 assay to detect antiphospholipid antibodies (aPL): lupus anticoagulant (LAC), anticardiolipin antibody (aCL), and anti–β2-glycoprotein I (anti-β2GPI) assays (1). The syndrome can occur within other autoimmune disorders or in otherwise healthy subjects. Pregnant women with primary APS can experience several adverse outcomes such as preeclampsia, growth restriction, fetal distress, premature delivery, miscarriage, and fetal loss (2). In the last decades we have assisted the improvement in the diagnosis and management of this disease, resulting in better care of pregnant women with primary APS and an increased possibility for these women to have children. Recently, interest in the long-term cognitive and psychological outcome of the offspring of mothers with autoimmune disease has grown, but only few works have been published; in particular, mainly children of mothers with systemic lupus erythematosus (SLE) have been investigated. Maternal autoantibodies, such as anti-Ro/SSA, are considered to be possibly responsible for altered functions occurring in the fetal brain, leading to the development of disorders such as dyslexia (3–5).

Dyslexia, which impairs the reading ability of school children with otherwise normal intelligence, is reported in up to 45% of male children of mothers with SLE (6), whereas it is observed in only 2.5–5% of healthy school-age children (7). Dyslexia seems to be much less frequent in female offspring of mothers with autoimmune disease (6, 8, 9). A possible explanation for this finding is the higher immune response observed in mothers of male fetuses, possibly causing fetus brain development impairment (10). Normal intelligence levels are reported in this population of children, leading to the assumption that mothers' autoimmunity does not affect offsprings' general cognitive capacity (5, 8). In a recent study, a higher incidence of learning disabilities (LDs) in the offspring of mothers with SLE was associated with maternal aPL (LAC and/or aCL) positivity (11). This finding, as well as the previously described animal models (12–14), supports the aPL influence on LD occurrence. Transplacental passage of aPL has been demonstrated in only 10.5% of newborns of mothers with primary APS, which is significantly lower than that observed for other autoantibodies (15). The presence of relatively large quantities of β2-glycoprotein I (β2GPI), the recognized target of most aPL, at the trophoblast surface may account for this finding (16). The low thrombosis rate reported in neonates of mothers with primary APS is consistent with this observation. The present study was the first to analyze the occurrence of LDs in the offspring of patients with primary APS.

PATIENTS AND METHODS

The mothers.

Between 1984 and 2003, we followed 82 patients with primary APS during their 122 pregnancies. Fifty-eight patients (71%) were diagnosed because of pregnancy losses, 9 (11%) because of thrombosis, and 15 (18%) because of both thrombosis and pregnancy losses. Of these mothers with primary APS, still regularly followed at Brescia Hospital (Rheumatology and Clinical Immunology Unit), 16 were selected during 2003 as possible participants in this study. According to the inclusion criteria, women had to live near Brescia Hospital and had to have at least 1 school-age (7–16 years) offspring without neurologic impairment. With the patients' permission, we sent them a letter explaining the project details and asking for their informed consent. Fourteen of the 16 mothers with primary APS (mean ± SD age 30.75 ± 2.35 years, mean ± SD disease duration at time of pregnancy 1.81 ± 1.6 years) agreed to enter the study. Ten of the 14 women (71.4%) reported miscarriages (range 1–4) and 4 reported thrombosis episodes. LAC was positive in 9 women (64.2%); IgG aCL were shown in 12 (85.7%), whereas 2 women (14.2%) had only IgM aCL; and 7 women (50%) had positive anti-β2GPI antibody assays.

Sample description.

Seventeen children (10 boys and 7 girls), including 3 brothers, were enrolled. At the time of the study, the mean ± SD age was of 11.74 ± 2.41 years (range 7.5–15.9 years). Among the 17 children, 15 had been exposed, at least theoretically, to maternal IgG aPL antibodies during pregnancy, whereas in 2 cases maternal aPL antibodies were of the IgM isotype, and therefore could not pass the placental barrier. Four children (23.5%) were classified as premature because they were born before the 37th week of gestation; 1 of these 4 was born at 32 weeks with a very low birth weight (<1,500 gm). The mean ± SD gestational age was 37.6 ± 2.55 weeks (range 32–40 weeks) and the mean ± SD birth weight was 2,854.7 ± 627.27 gm (range 1,500–3,820 gm). None of the children had a slow weight and statural development, or a retarded psychomotor or speech development. The mothers reported minor health problems during their children's life, namely, asthma (n = 1), allergy (n = 2), synovitis (n = 2), and headache (n = 2).

Children: neuropsychological evaluation.

We evaluated general cognitive capacities through the administration of the Wechsler Intelligence Scale for Children, Revised (WISC-R) (17). This instrument measures different aspects of intelligence using 10 subscales (plus 2 optional scales) divided into 2 groups relating to verbal skills (5 subscales: information, similarities, arithmetic, vocabulary, and comprehension), plus the optional digit span (measuring verbal intelligence quotient [VIQ]), and performance skills (5 subscales: picture completion, picture arrangement, block design, object assembly, and coding), plus the optional mazes (measuring performance intelligence quotient [PIQ]). The full-scale intelligence quotient (FSIQ) was also assessed. To test the children, we used the 10 subscales and the digit span optional scale for testing short-term memory and working memory.

To directly estimate the occurrence of LDs, we evaluated all the children using a specific LD battery of tests, standardized for Italian pupils, that comprised the Sartori test (18) (items 4 and 5 for reading abilities [words and nonsense words reading speed and accuracy] and items 10 and 11 for words and nonsense words writing skills); the MT groups' test for reading ability (19) for analyzing text reading speed and accuracy; and the MT groups' test for math skills (20, 21) for arithmetic abilities. An LD can be diagnosed when the results are 2 SDs below the mean observed in the general population of the same age; when the FSIQ is >85, indicating a normal general intelligence; and when there are no psychiatric, neurologic, or sensorial disturbances that may determine a nonspecific form of an LD. Neurologic impairment was excluded in the sample selection phase.

To identify eventual risk of psychopathologies, the Child Behavior Checklist (CBCL) (22) was administered. The CBCL is a self-administered questionnaire for parents that analyzes children's behavioral and social characteristics. A semistructured interview was also used to investigate parents' sociodemographic profiles. The neuropsychiatric expert asked parents about the pregnancy, the child's growth, and the impact of maternal illness on family life. Parents were also asked to describe their child's character traits and attitudes in order to record not only clinical data but also information about the family environment in which the child grew up.

Autoantibody determinations.

Patients' autoantibody profiles were determined by methods routinely performed in our laboratory. In particular, aCL and anti-β2GPI were detected by enzyme-linked immunosorbent assay as previously described (23, 24). LAC testing was performed according to international guidelines (25).

Statistical analysis.

Descriptive statistics were expressed as the mean ± SD; categorical variables were expressed as crude numbers and percentages.

RESULTS

We examined 17 offspring from mothers with primary APS. All children were born from Italian parents and were living with them at the time of the pregnancies; 2 couples divorced when their 3 children were between 3 and 5 years old. Six mothers (42.8%) of 7 children were housewives, and the remaining mothers were working (5 office workers, 1 contractor, 1 consultant, and 1 blue-collar worker). Of the fathers, 6 were office workers, 3 consultants, 3 contractors, 1 farmer, and 1 blue-collar worker.

Two children were born from patients with IgM aPL only, therefore they were not exposed to maternal antibodies because the IgM isotype impairs the transplacental passage and prevents contact with the fetus. Interestingly, these 2 children were found to have normal cognitive capacity, learning abilities, and behavioral attitude.

All 15 children enrolled were assessed by the WISC-R and showed a normal intelligence level with FSIQ >85 (mean ± SD score 104 ± 12.06, range 88–129, median 101). The 2 different aspects of the WISC-R (verbal and performance) were analyzed separately. The mean ± SD VIQ was 104.86 ± 14.91 (range 88–136, median 98) and the mean ± SD PIQ was 102.66 ± 11.01 (range 84–117, median 102). The mean IQ scores obtained in the verbal and performance subscales were above the cutoff of normal levels (>7) in all items, whereas the single scores for each child in the 11 subscales of the WISC-R were in some cases below the fixed limit (information and picture completion in 1 child, arithmetic and picture arrangement in 2 children, similarities in 3 children, object assembly and coding in 4 children, and comprehension in 7 children). The results of the profile analysis (VIQ, PIQ, and single subscales) are shown in Table 1.

Table 1. Wechsler Intelligence Scale for Children, Revised (WISC-R) results*
WISC-R subsetsMean ± SDRangeMedianNormal level
  • *

    The full-scale IQ comprises the verbal and performance IQ corrected by age. IQ = intelligence quotient.

Verbal IQ104.86 ± 14.9188–13698>70
 Information11.07 ± 2.348–1610>7
 Similarities10.27 ± 2.686–1610>7
 Arithmetic10.13 ± 3.685–1810>7
 Vocabulary13.47 ± 3.069–1913>7
 Comprehension9.13 ± 3.445–178>7
 Digit span9.07 ± 3.096–179>7
Performance IQ102.66 ± 11.0184–117102>70
 Picture complete10.46 ± 2.167–1510>7
 Picture arrangement9.6 ± 2.943–1410>7
 Block design11.88 ± 2.658–1911>7
 Object assembly9.26 ± 2.344–139>7
 Coding11.06 ± 3.325–1611>7
Full-scale IQ104.00 ± 12.0688–129101>70

The CBCL explores the presence of competence and behavioral problems. The mean scores from the present study are summarized in Table 2. For the total competence scale score corrected for age and sex, 4 offspring (26.7%) seemed to have a higher risk of present problems (1 on the activity subscale, 4 on the social subscale, and none on the school subscale). For the total behavior scale (always corrected for age and sex), 2 children (13.3%) had a score that indicated risk.

Table 2. Child Behavior Checklist (CBCL) results*
CBCL subsetsMean ± SDAt risk scores, no. (%)
  • *

    Evaluation of the score obtained for the CBCL was corrected for sex and age of the examined children.

Total competence scale16.44 ± 3.44 (26.7)
 Activity5.7 ± 1.91 (6.67)
 Social5.9 ± 2.24 (26.7)
 School4.9 ± 0.90 (0)
Total behavior scale20.07 ± 14.12 (13.3)
 Internalizing problems6 ± 5.92 (13.3)
  Retreat1.7 ± 1.90 (0)
  Somatic complaint1.1 ± 1.60 (0)
  Anxiety/depression3.4 ± 41 (6.67)
 Externalizing problems6.07 ± 4.531 (6.67)
  Social problems1.8 ± 1.90 (0)
  Thought problems0.1 ± 0.40 (0)
  Attention problems3.6 ± 2.92 (13.3)
  Delinquency1.2 ± 1.60 (0)
  Aggressiveness4.9 ± 4.10 (0)

The mothers' interviews pointed out some aspects of their offspring's personality traits; in general, no specific traits were overexpressed. It is relevant that 2 children were described as hyperactive.

Five mothers (35.7%) reported having anxiety problems during pregnancy. Miscarriages were strictly linked to the presence of anxiety traits: 10 women (71.42%) reported at least 1 pregnancy loss before a successful delivery, whereas 1 woman (7.14%) had delivered a child who died after 2 days of life. Another aspect related to anxiety during pregnancy was the worry of fetal disease transmission, which was present in 4 mothers (28.57%). Three women (21.42%) experienced postpartum depression.

The evaluation of LDs was performed with the Sartori test. A dyslexic syndrome, with the involvement of all 3 aspects (reading, writing, and arithmetic), was diagnosed in 2 boys and 1 girl (20%). One of these children, a boy age 14.7 years, had a moderate form of LD and showed good compensation strategies. In addition, one other girl showed arithmetic difficulties only, which is characteristic of a dyscalculia syndrome that can also be classified as an LD. Altogether, we identified 4 LD cases (26.7%). According to the available literature, the incidence of dyslexia in Italian school-age healthy children is 2.5–5% (7) and was found in 4% of 7,827 children in a large study performed in our town (Trebeschi, MD, Tiberti, MD, Parrinello, PhD: unpublished observations). Of the 4 children with LDs, 1 girl was born premature and was very small for gestational age (32 weeks of gestation; birth weight 1,500 gm); her brother, born at term, was also 1 of the 4 affected subjects.

As expected, lower scores on the arithmetic and coding subscales of the WISC-R were recorded in this subgroup, without a significant decrease in the FSIQ. In the subgroup of children with LDs, the mean ± SD FSIQ was 104.75 ± 12.84, the mean VIQ was 100.5 ± 11.03, and the mean PIQ was 103.0 ± 9.76, whereas in the other subjects the mean FSIQ was 106.82 ± 17.91, the mean VIQ was 106.45 ± 16.23, and the mean VIQ was 104.36 ± 13.22. WISC-R subscale results divided into 2 subgroups by the presence of a specific LD are illustrated in Figure 1. Interestingly, 1 of the 4 children with LDs was found to be at risk in the total competence scale of the CBCL, but no child had problems in the total behavioral scale (Table 3).

Figure 1.

Illustration of the different subscale scores of the Wechsler Intelligence Scale for Children, Revised divided according to the presence (solid diamonds) or absence (shaded diamonds) of a specific learning disability. Picture Compl. = picture completion; Picture Arrang. = picture arrangement.

Table 3. Child Behavior Checklist (CBCL) results, divided into 2 subgroups by the presence of a learning disability (LD)*
CBCL subsetsNon-LD (n = 11)LD (n = 4)
Mean ± SDAt risk scores, no. (%)Mean ± SDAt risk scores, no. (%)
  • *

    Evaluation of the score obtained for the CBCL was corrected for sex and age of the examined children.

Total competence scale16.54 ± 3.583 (27.3)12.08 ± 8.601 (25)
 Activity5.53 ± 2.141 (9.1)4.72 ± 3.240 (0)
 Social5.82 ± 2.133 (27.3)4.54 ± 3.911 (25)
 School5.20 ± 0.560 (0)2.82 ± 2.190 (0)
Total behavior scale18.55 ± 15.602 (18.2)19.25 ± 13.250 (0)
 Internalizing problems5.82 ± 6.812 (18.2)5.00 ± 3.370 (0)
  Retreat1.45 ± 2.070 (0)1.75 ± 1.500 (0)
  Somatic complaint1.18 ± 1.670 (0)0.75 ± 1.500 (0)
  Anxiety/depression3.36 ± 4.521 (9.1)2.75 ± 1.890 (0)
 Externalizing problems5.27 ± 4.821 (9.1)6.75 ± 4.570 (0)
  Social problems1.55 ± 1.860 (0)2.00 ± 2.160 (0)
  Thought problems0.18 ± 0.410 (0)0.00 ± 0.000 (0)
  Attention problems3.09 ± 2.551 (9.1)4.25 ± 4.351 (25)
  Delinquency1.00 ± 1.340 (0)1.50 ± 2.380 (0)
  Aggressiveness4.27 ± 4.340 (0)5.25 ± 4.110 (0)

The families of the affected children had good socioeconomic status and a fair educational level. All of the fathers and 3 mothers were office workers, and 1 mother was a housewife. Of the 4 with LDs, 1 boy (age 11.5 years) had divorced parents at the time of study entry, while 2 other children (ages 14.3 and 15.9 years) with divorced parents had good results for all tests.

According to the mothers' interviews, the children with LDs were described as having the following traits: 1 child was hypermature and sociable/extroverted, 1 was sociable/extroverted, 1 was introverted, and 1 was hyperactive and sociable/extroverted (Table 4), with an overall prevalence of children described as sociable/extroverted in the LD group compared with the non-LD group (3 of 4 versus 2 of 11). All of the mothers of children with LDs experienced 2 pregnancy losses, 1 was treated with corticosteroids during pregnancy, and none had postpartum depression. The mother of the girl with dyscalculia reported anxiety problems during pregnancy associated with the worry of fetal disease transmission (Table 4).

Table 4. Mothers' interview results*
 Non-LD (n = 11)LD (n = 4)
  • *

    Values are the number (percentage). Children's personality traits and mothers' psychological aspects related to pregnancy, divided into 2 subgroups by the presence of learning disabilities (LDs).

Child's personality traits  
 Hypermature3 (27.3)1 (25)
 Sociable/extrovert2 (18.2)3 (75)
 Obliging4 (36.3)0 (0)
 Introverted3 (27.3)1 (25)
 Hyperactive1 (9.1)1 (25)
Mothers' psychological aspects  related to pregnancy  
 Anxiety problems during pregnancy4 (36.3)1 (25)
 Worry of fetal disease transmission3 (27.3)1 (25)
 Postpartum depression3 (27.3)0 (0)
 At least 1 pregnancy loss6 (54.5)4 (100)

DISCUSSION

In the present study, the overall cognitive evaluation showed that all of the children exposed in utero to maternal aPL presented general cognitive capacities (FSIQ score) within the normal range. Therefore, we suggest that the aPL exposure during fetal life does not impair the global intelligence capacities. This finding is in agreement with the reported normal intelligence level shown in the offspring of patients with SLE, who are frequently aPL positive (7, 8, 11). In contrast to the known high rate of memory loss in patients with aPL (26), no problems in short-term memory were observed in the digit span subscales of the WISC-R.

The presence of behavioral and competence problems related to maternal disease was assessed by the CBCL, considering the parents' point of view. The average scores recorded were within the normal range for all items, while a minority of patients were found to be at risk only in a small number of items (Table 2). The relatively limited size of our sample and the absence of a generalized trait do not allow a general conclusion on the effect of mothers' primary APS on their children. Interestingly, however, the problems on the social/activity competence scale, recorded in 5 of the 15 children, are usually observed in children with chronic diseases (27, 28), as though the mother's long-lasting illness could influence children's social development.

In agreement with the literature, 4 mothers with primary APS (23.5%) reported being worried during pregnancy about losing their babies or transmitting their disease to their babies. Some of them expressed the need for psychological support, especially during gestation, when the disease was much more pervasive and difficult to tolerate. In fact, the experience of previous pregnancy losses, so diffuse in this group of patients, is often associated with an increased risk of depression and anxiety (29, 30).

Children were described by the mothers during the interviews as having a generally good psychological outcome without a shared personality trait; however, 2 of 15 children were described as hyperactive. According to the literature, the occurrence of attention deficit/hyperactivity disorder (ADHD) in mice is supposedly induced by the presence of aPL; on the contrary, in a study of 41 children affected by ADHD, no evidence of a relationship between this condition and aPL was reported (31).

There are various possible explanations for the presence of LDs in 4 of the 15 children. Exposure to aPL during fetal life, occurring when IgG antibodies were present in the mother, may have influenced the neuropsychological development. The low rate of aPL transplacental passage (15) supports the relatively low prevalence of LDs in the 15 children who were exposed. The potential pathogenic role of aPL antibodies in fetal brain neurodevelopment is shown by their capability to bind in vitro to brain tissue and brain endothelial cells (32–34), as well as by the recorded propensity for hyperactivity and anxiety in animal models after prolonged exposure (12–14). In addition, aPL antibodies were described as being associated with psychosis, behavioral changes, and memory loss in a patient series (26).

We identified LDs in 2 boys and 2 girls, indicating that the presence of aPL in maternal circulation may be a sex-independent risk factor. Two of the 4 affected children were brothers and this fact could be consistent with the hypothesis of a genetic influence on the occurrence of LDs (35). In fact, the risk of dyslexia is more than 3.5 times higher for siblings of affected children, as compared with the general population (36).

In our cohort, a girl with an LD was born prematurely (gestational age 32 weeks, birth weight 1,500 gm). In this case, prematurity may have been one of the factors concurring in the development of the LD. In fact, children who are born preterm are reported to have a 3–5-fold higher risk of having reading, spelling, mathematics, or writing disorders (37).

Three of the 4 children with LDs were described by their mothers as being sociable/extroverted, although in 1 of them the CBCL revealed social problems (Table 4). A possible interpretation of these apparently contrasting data could be that this child may try to exceed his social difficulties by being affable with peers, without being capable in actively searching for social occupations, such as sports or hobbies (behaviors measured by the CBCL competence scale). Interestingly, one of the children with LDs was described by his mother as being hyperactive. This observation is in agreement with the literature data on the high prevalence of LD and ADHD comorbidity (38). According to the literature, socioeconomic status seems to influence the occurrence of learning impairment (39). Our data do not allow verification of this hypothesis because family wealth and parents' educational levels were generally good in all parents of children entering the study.

In conclusion, the present study focused on school-age children of patients with primary APS. It is difficult to recruit subjects who satisfy these inclusion criteria because of the relatively rare prevalence of this disease and because only women with a very long followup can be diagnosed before pregnancy and have, at present, a school-age child. Even with the limitation of a small sample study, the conclusion we came to is that the LD occurrence may be related to the presence of aPL in the mother during pregnancy; these antibodies, interacting with other well-known risk factors such as prematurity and genetic and environmental factors, could contribute to the pathogenesis of this developmental disorder. In this case, early diagnosis would be worthwhile to help children overcome their difficulties during the school years. In addition, pregnancy represents a delicate period of life for every woman, and, in the case of mothers with primary APS, a threatening risk of fetal loss makes them more susceptible to anxiety and more in need of global care. To meet patients' needs, we think we should offer them a multidisciplinary team in which different experts may effectively cooperate to ameliorate patients' global quality of life, during pregnancy and child growth stages.

AUTHOR CONTRIBUTIONS

Dr. Tincani had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Nacinovich, Bomba, Nuzzo, Lojacono, Motta, Tincani.

Acquisition of data. Galli, Filippini, Nuzzo, Lojacono.

Analysis and interpretation of data. Nacinovich, Bomba, Nuzzo, Motta, Tincani.

Manuscript preparation. Nacinovich, Galli, Bomba, Lojacono, Tincani.

Statistical analysis. Bomba, Parrinello.

Ancillary