Recurrence rates of cardiac manifestations associated with neonatal lupus and maternal/fetal risk factors

Authors


Abstract

Objective

Identifying the frequency of recurrent cardiac manifestations of neonatal lupus (NL) in a second child is critical to understanding the pathogenesis of anti-SSA/Ro–mediated injury and would improve counseling strategies regarding future pregnancies and power the design of clinical prevention trials. Accordingly, this study was undertaken to address the recurrence rates of cardiac NL and associated risk factors in a large US-based cohort.

Methods

Families enrolled in the Research Registry for Neonatal Lupus were evaluated for rates of recurrence of cardiac NL and potential risk factors, with a focus on pregnancies immediately following the birth of an affected child.

Results

The overall rate of recurrence of cardiac NL in 161 pregnancies of 129 mothers with anti-SSA/Ro antibodies was 17.4% (95% confidence interval 11.1–23.6%). Analysis of the potential risk factors among 129 mothers with a pregnancy immediately following the birth of a child with cardiac NL showed that the maternal diagnosis was not associated with the outcome in a subsequent pregnancy. In this group, 23% of mothers who were either asymptomatic or had an undifferentiated autoimmune syndrome, compared with 14% of mothers with systemic lupus erythematosus or Sjögren's syndrome, had a second child with cardiac NL (P = 0.25). The recurrence rate was not statistically significantly different in mothers who had taken steroids compared with those who had not taken steroids (16% versus 21%; P = 0.78). The antibody status of the mother was not predictive of outcome in subsequent pregnancies. Moreover, death of the first child with cardiac NL was not predictive of recurrence of cardiac NL in a subsequent pregnancy (P = 0.31). The risk of cardiac NL was similar between male and female children (17.2% versus 18.3%; P = 1.0).

Conclusion

In this cohort, the overall recurrence rate for cardiac NL was 17%. The recurrence rate appeared to be unaffected by maternal health, use of steroids, antibody status, severity of cardiac disease in the first affected child, or sex of the subsequent child.

Neonatal lupus (NL) is a pathologic readout associated with the presence of maternal anti-SSA/Ro and/or anti-SSB/La autoantibodies but is a rare manifestation that affects only a minority of exposed offspring. Cardiac manifestations of NL are the most serious and can range from clinically benign first-degree heart block to third-degree (complete) heart block (the latter being most common) and, in some cases, can lead to fatal cardiomyopathy (1, 2). An intriguing aspect of cardiac NL is that it is an injury unique to some phase of development, since it has only once been reported as a manifestation in the maternal heart despite the presence of identical antibodies in the maternal circulation (3). Cardiac NL carries a substantial mortality (∼30%) and morbidity, with >60% of children requiring lifelong pacemakers (4, 5). Despite the attempt of large multicenter studies to forestall disease by careful monitoring of the fetal heart, irreversible complete heart block and extensive myocardial injury have been documented within 7 days of a documented normal rhythm and PR interval (6, 7). Moreover, postnatal incomplete heart block (electrocardiogram [EKG] identification of first- or second-degree heart block) can progress despite clearance of maternal antibodies from the neonatal circulation (8). In contrast to cardiac NL, the cutaneous manifestations of NL, despite being associated with the same maternal autoantibodies, are most common several weeks after birth and are transient (9).

Identification of isolated fetal conduction disease may be the first indication of previously unrecognized autoimmunity in the mother. It is this link that is the basis of the powerful association of anti-SSA/Ro and anti-SSB/La autoantibodies with cardiac disease. Since mothers with established systemic lupus erythematosus (SLE) or Sjögren's syndrome (SS) often have these reactivities, they comprise the clinical groups most frequently counseled regarding the risk of cardiac NL. Two prospective studies to date, one from Italy and the other US based, have identified a risk of cardiac NL of 2% in children of mothers who were previously known to have anti-SSA/Ro antibodies (3, 6). Anti-SSB/La antibodies may increase the risk to 5% (10). The specificity of maternal antibodies for amino acids 200–239 of the 52-component SSA/Ro has received considerable attention, although there have been some differences in the extent of added risk observed, depending on the assay used or the cohorts studied (11–13). Although these aggregate risks are still relatively low, at <10%, the burden of disease is high and mothers contemplating a subsequent pregnancy entreat their managing physicians for advice regarding recurrence rates.

Accordingly, the present study was undertaken to address the recurrence rates of cardiac NL and associated risk factors in a large, extensively characterized US-based cohort. Such information would be expected to improve pregnancy counseling of anti-SSA/Ro–anti-SSB/La–positive mothers, would provide power for the design of prevention trials, and would offer insight into the pathogenesis of cardiac injury, particularly with regard to a fetal genetic contribution. This study leveraged the data from the Research Registry for Neonatal Lupus (RRNL), established in 1994 (5), to accomplish these goals. One hundred twenty-nine families with mothers who had a pregnancy subsequent to the birth of a child with cardiac NL were evaluated for the overall recurrence rate of cardiac NL. In addition, factors that might amplify or decrease the effects of the “necessary” antibody (assuming that the overall profile of antibodies remained similar between pregnancies) were explored. These included maternal factors such as health status, medications, and antibody specificities and titer. Fetal risk factors included severity of cardiac disease in the first affected child, and sex of the subsequent child.

PATIENTS AND METHODS

Subjects.

As previously described, mothers enrolled in the RRNL must satisfy 2 requirements: 1) presence of antibodies to SSA/Ro or SSB/La (or, if only cutaneous manifestations are present, presence of anti-RNP is considered as an exception), and 2) having a child with any manifestation of NL (5) as verified by review of the medical records. The RRNL and the informed consent of the participants received the approval of the New York University School of Medicine Institutional Review Board. The enrollment period for this study extended from September 1994 to January 2009. However, a mother could be entered into the RRNL in 1994 after having had a child with NL years prior. Inclusion criteria for the present study were as follows: 1) enrollment in the RRNL, 2) documented maternal anti-SSA/Ro and/or anti-SSB/La antibodies, 3) a previous child with cardiac NL, and 4) a pregnancy immediately subsequent to the birth of a child with cardiac disease. Cardiac NL was defined herein as the presence of heart block (first-, second-, or third-degree) documented by EKG, echocardiogram, history of pacemaker, or statement in the medical record, and/or the presence of cardiac injury, which specifically included autopsy evidence of a mononuclear infiltrate in the endocardium, myocardium, and pericardium, and/or the presence on echocardiogram of endocardial fibroelastosis (EFE) that was always associated with cardiac dysfunction. A prolonged PR interval ≥150 msec, identified in utero, was only counted as indicative of cardiac NL if confirmed by the EKG at birth. Noncardiac NL cases were defined as children who were healthy or who had any manifestation of NL other than cardiac.

Maternal health status, ethnicity, and medications were based on telephone interviews and information obtained from the medical records, as well as from enrollment and followup questionnaires available in the RRNL. For the purposes of this study, the maternal health status at the time of the subsequent pregnancy following the birth of the child with cardiac NL was chosen.

Subsequent pregnancies.

An overview of the birth-order data on families who were enrolled in the RRNL and had at least 1 child with cardiac NL is presented in Table 1. In families with the birth of more than 1 child following the birth of a baby with cardiac NL, another affected child could have been the immediate next child or could have followed a healthy child (n = 3 families) or could be the third affected child (n = 2 families). For the purposes of calculating the overall recurrence rate, all pregnancies in 129 families in which there were any pregnancies subsequent to the birth of an initial child with cardiac NL (the index case) were included (n = 161 pregnancies). For the evaluation of potential risk factors of recurrence, data from only the pregnancy occurring immediately following the index case were analyzed (n = 129 pregnancies). Fetuses in which death secondary to heart block occurred, i.e., hydrops, were also included.

Table 1. Birth order and families enrolled in the RRNL*
Family size, birth-order case patternNo. of families of each sizeNo. of families exhibiting case pattern
  • *

    Children with cardiac neonatal lupus (NL), including fetal demises secondary to heart block, are represented by the solid circles, while noncardiac NL cases are represented by the open circles. The values shown are a summary of the entire Research Registry for Neonatal Lupus (RRNL) data with regard to family size and birth order.

  • Families with births subsequent to a child with cardiac NL (n = 129).

1-child56 
 • 56
2-child139 
 •○ 58
 ○• 69
 •• 12
3-child63 
 •○○ 11
 ○•○ 25
 ○○• 16
 ••○ 2
 ○•• 6
 ••• 2
 •○• 1
4-child14 
 •○○○ 1
 ○○•○ 3
 ○○○• 7
 ○○•• 1
 ○•○• 1
 •○•○ 1
5-child2 
 •○○○○ 2
6-child2 
 ○○○○•○ 1
 ○○•○○○ 1
8-child1 
 ○○○•○○○○ 1

Forty-five families were entered into the RRNL after the birth of a child with noncardiac NL subsequent to the index case. Ten families were entered into the RRNL after the birth of a child with cardiac NL subsequent to the index case, and these families were evaluated prospectively for all subsequent pregnancies. Seventy-two families were entered into the RRNL at the time of birth of the initial child with cardiac NL, and these families were also evaluated prospectively for all pregnancies subsequent to the index case.

Detection of antibodies to SSA/Ro and SSB/La proteins.

Determination of antibodies to SSA/Ro and SSB/La was done in the clinical immunology laboratory at the Hospital for Joint Diseases using a commercial enzyme-linked immunosorbent assay (ELISA) kit (Diamedix, Miami, FL). Reactivity to the 52-kd SSA/Ro, 60-kd SSA/Ro, or 48-kd SSB/La ribonucleoproteins was determined by ELISA using recombinant proteins and/or a sodium dodecyl sulfate Molt-4 immunoblot as previously described (14, 15).

Statistical analysis.

The overall rate of recurrence of cardiac NL was computed as the proportion of cardiac NL cases among all pregnancies following the birth of the initial child with cardiac NL (the index case). Because data from multiple pregnancies in the same subject were included in the estimate, the 95% confidence interval (95% CI) for the recurrence rate was computed using a calculation of the standard error based on the approach by Donald and Donner (16) for clustered binary data. For the analysis of the potential risk factors for recurrent cardiac NL, only data from the pregnancy immediately following the index case were used. The effects of maternal health status, steroid use, maternal antibody status (i.e., presence of anti-SSB/La antibodies), death of the first child with cardiac NL, and sex of the second child on the risk of a recurrence were evaluated with the Fisher's exact test. The Mann-Whitney test was used to compare the titers of anti-SSA/Ro antibodies in the mothers whose fetuses developed cardiac NL with those in the mothers whose fetuses had no disease. Two-sided P values less than 0.05 were considered statistically significant.

RESULTS

Outcome of pregnancies subsequent to the index case.

Over the 15-year study period, 129 of the 277 families currently enrolled in the RRNL (verified as comprising a mother with anti-SSA/Ro antibodies and a child with cardiac NL) included a pregnancy immediately subsequent to the birth of an initial child with cardiac NL. Seventy-nine percent of the mothers were Caucasian, 9% were African American, 6% were Asian, 6% were Hispanic, and only 1 mother (<1%) was American Indian.

The overall recurrence rate of cardiac NL among all pregnancies following the index case was 17.4% (95% CI 11.1–23.6%). As summarized in Table 2, of the 28 children with cardiac NL, 23 had either second- or third-degree heart block. Two children had first-degree heart block, diagnosed in utero in 1 child and at birth in the other child, and in both of these children, the first-degree block was confirmed by EKG and has persisted through 1 year of followup. One child had evidence of EFE in utero on echocardiogram and died postpartum after experiencing severe cardiac dysfunction. Two children had EFE that was detected on echocardiogram and has persisted over followup; 1 of these children remains alive at age 7 months and the other remains alive at age 2 years. The majority of cardiac NL cases (75%) were detected between 18 weeks and 25 weeks of gestation. The mortality rate was 21% (occurring in 2 children at birth, 3 within 3 months after birth, and 1 at age 5 years). Sixty-eight percent of the children required pacemakers.

Table 2. Specific cardiac manifestations, time of detection, and outcomes in the 28 subsequent children with cardiac NL*
 Type of cardiac NL
First-degree HBSecond- degree HBThird-degree HBEFE
  • *

    Values are the number (%) of children. NL = neonatal lupus; HB = heart block; EFE = endocardial fibroelastosis.

Death005 (18)1 (4)
Cardiac NL detected at 18–25 weeks of gestation01 (4)18 (64)2 (7)
Pacemaker0019 (68)0
Total21223

For the evaluation of potential risk factors for recurrent cardiac NL, only 1 pregnancy immediately following the index case in each family was analyzed. Thus, of the 129 pregnancies studied, 23 (18%) resulted in a child with cardiac NL. Twelve subsequent children (9%) had a rash and 94 subsequent children (73%) were healthy.

Maternal risk factors associated with pregnancy outcome.

Table 3 summarizes the results of association analyses of maternal risk factors. Maternal health status was determined at the time of birth of the child immediately following the index case. The results showed that pregnancy outcome was not associated with maternal diagnosis. Specifically, 14 (22.6%) of 62 mothers characterized as asymptomatic or having an undifferentiated autoimmune syndrome had a child with recurrent cardiac NL, compared with 9 (13.9%) of 65 mothers who were diagnosed as having SLE or SS (P = 0.25). For 2 mothers giving birth to a child with noncardiac NL, information regarding health status was unavailable.

Table 3. Association between maternal factors and pregnancy outcome*
 TotalSubsequent cardiac NLSubsequent noncardiac NLP
  • *

    Values are the number (%) of mothers. P values were determined by Fisher's exact test. NL = neonatal lupus; Asym/UAS = asymptomatic/undifferentiated autoimmune syndrome; SLE = systemic lupus erythematosus; SS = Sjögren's syndrome.

Maternal diagnosis    
 Asym/UAS6214 (22.6)48 (77.4)0.25
 SLE/SS659 (13.9)56 (86.1)
Maternal medications    
 Steroids254 (16)21 (84)0.78
 No steroids9119 (20.9)72 (79.1)
Maternal antibody status    
 Anti-Ro/La8415 (17.9)69 (82.1)1.0
 Anti-Ro458 (17.8)37 (82.2)
 Anti–Ro 52    
  Positive9914 (16.5)85 (83.5)1.0
  Negative910 (0)91 (100)

Since steroids, both fluorinated and nonfluorinated, have been considered in the prevention of cardiac NL (17–19), the use of prednisone and the use of dexamethasone were assessed. Only mothers who were taking steroids from the time of conception with continuation through at least 34 weeks of gestation were considered in the analysis. Although the records did not always provide an indication for prednisone, the use of prednisone was largely for maternal disease. The indication for dexamethasone was prophylaxis against recurrence of cardiac NL. Four (16%) of the 25 mothers taking steroids had a child with recurrent cardiac NL, compared with 19 (20.9%) of 91 mothers who were not taking steroids (P = 0.78). Three of 4 mothers with a child with recurrent cardiac NL who were taking steroids had received prednisone, while the other mother had received dexamethasone (dosage not stated). Seventeen of the mothers who had received steroids and had children with noncardiac NL were taking prednisone (mean dosage 23 mg/day) and 4 were taking dexamethasone (mean dosage 4 mg/day). For 13 mothers of children in the noncardiac NL group, information about medications was not available.

The maternal antibody status did not influence the recurrence rate. Specifically, 15 (17.9%) of the 84 mothers with antibodies to both SSA/Ro and SSB/La had a child with cardiac NL, compared with 8 (17.8%) of 45 mothers with antibodies to SSA/Ro only (P = 1.0). Only 1 mother had antibodies against SSB/La only, and she had 1 child with cardiac NL who died and a subsequent child who was healthy. Antibodies reactive with Ro 52 were almost universally present but were not associated with pregnancy outcome (P = 1.0). Furthermore, the titer of anti-SSA/Ro antibodies within 6 months of the subsequent pregnancy was not predictive of outcome (P = 0.287).

Fetal factors associated with pregnancy outcome.

As summarized in Table 4, the contribution of fetal factors to the risk of recurrence of cardiac NL was considered. Death of the first child with cardiac NL was used as a proxy to address whether severity of cardiac involvement in the index case influences the recurrence rate. Eight (23.5%) of the 34 initial children with cardiac NL who died were followed by a recurrence of cardiac NL in a subsequent child, compared with 15 (15.8%) of the 95 initial children with cardiac NL who remained alive (P = 0.31).

Table 4. Association between fetal factors and pregnancy outcome*
 TotalSubsequent cardiac NLSubsequent noncardiac NLP
  • *

    Values are the number (%) of children. P values were determined by Fisher's exact test. NL = neonatal lupus.

Index cardiac NL death    
 Yes348 (23.5)26 (76.5)0.31
 No9515 (15.8)80 (84.2)
Fetal sex    
 Female7113 (18.3)58 (81.7)1.0
 Male5810 (17.2)48 (82.8)

Fetal sex was evaluated with regard to its influence on recurrence and overall association with cardiac NL. Evaluation of the entire RRNL revealed that female sex was observed in 57% of the children with cardiac NL, 55% of the children with rash, and 49% of the completely unaffected children (P = 0.062 comparing children with cardiac NL with completely healthy children). Comparison of our data with the National Vital Statistics Report of 2003 (female-to-male ratio 1,000:1,048; http://www.cdc.gov/nchs/products/nvsr.htm) revealed significant skewing toward a predominance of females among the children with cardiac NL in the RRNL (P = 0.013), likely due to the larger sample size in that report. However, 13 (18.3%) of 71 girls born subsequent to the index case had cardiac NL, compared with 10 (17.2%) of 58 boys born subsequent to the index case (P = 1.0).

Outcome of 4 pregnancies immediately subsequent to the birth of 2 consecutive children with cardiac NL.

Four of the pregnancies recorded for the mothers included in this study immediately followed the consecutive births of 2 children who each had cardiac NL. For 2 of these mothers, the third child was also affected. For 1 mother, the third child had a rash, whereas the other mother had a third child who was healthy. Although not considered in the analyses herein, in another mother with 2 consecutive children with cardiac NL, the use of a surrogate carrier for a subsequent pregnancy resulted in an unaffected child.

DISCUSSION

Based on the data from this large US-based cohort, the overall recurrence rate of cardiac NL was 17%, and for the immediately subsequent pregnancy, the recurrence rate was 18%. This is a nearly 10-fold increased risk in an anti-SSA/Ro–positive mother who has not had a previous child with cardiac NL. The estimated recurrence rate of cardiac NL for a pregnancy subsequent to 2 consecutive children with cardiac NL was 50%. Maternal health status, use of steroids during pregnancy, antibody status, severity of disease in a first affected child, and sex of the second child were not predictors of the outcome of subsequent pregnancies.

Over the last decade, several groups have reported rates of recurrence of cardiac NL. One study from Finland assessed the recurrence rate in 33 pregnancies immediately subsequent to a child with complete heart block, and those investigators found that 6 (18%) had a recurrence of cardiac NL (20). However, that study was smaller and the presence of maternal anti-SSA/Ro–anti-SSB/La antibodies was not an inclusion criterion; consequently, women without documented anti-SSA/Ro antibodies were pooled with those who did have anti-SSA/Ro antibodies (20). This latter shortcoming may be critical, since the etiology of congenital conduction disease includes such entities as complex structural congenital heart diseases (heterotaxia and endocardial cushion defect) as well as long QT syndrome (21), neither of which would be applicable in mothers with autoantibodies who are seeking pregnancy counseling for the risk of cardiac NL. In a prospective followup study by Gladman et al in a cohort of women from Toronto, of 118 pregnancies in 105 anti-SSA/Ro–positive mothers, only 15 of these mothers had a previous child with NL (rash or cardiac NL not specified), and in 16 subsequent pregnancies, 1 (6%) resulted in a second child with cardiac NL (22). Another previous study, which was based on data from anti-SSA/Ro–positive mothers enrolled in the RRNL, evaluated the outcome of 87 pregnancies immediately following the birth of a child with complete heart block, and the recurrence rate was found to be 19.5% (23).

In the present cohort, the majority (75%) of the cardiac NL cases were detected in the fetal period, specifically, between 18 weeks and 25 weeks of gestation. This is consistent with the results reported in previous literature, which indicated that cardiac NL is usually diagnosed before 30 weeks of gestation, with a peak incidence between 20 weeks and 24 weeks of gestation (24, 25). Two of the 5 children with atrioventricular conduction abnormalities first identified after birth were noted to have first-degree block, 1 of whom was diagnosed despite having normal measurements of the PR interval throughout the pregnancy, and the other of whom was diagnosed in the early 1990s, at a time prior to the more widespread use of PR interval monitoring (6). The other 3 children with third-degree block were born before 1990, when considerably less information about the disease and fewer diagnostic tools were available. The overall mortality rate found in this latter group of children with recurrent complete heart block was 21%, which is consistent with the results in the literature (5, 24, 25). The frequency of a pacemaker implantation in the surviving newborns was likewise consistent with that in previous reports (4, 5, 24, 25).

Based on the literature to date (3, 6), only a minority (2%) of anti-SSA/Ro–positive mothers have given birth to a child with cardiac NL, suggesting that maternal autoantibodies, while necessary, are not sufficient to cause disease. Recurrence rates for siblings of affected infants are 3,000 times higher than the population prevalence estimated at 1:15,000 (26–28). However, if only anti-Ro–positive mothers are considered in the prevalence, this risk is still 10-fold higher, suggesting a genetic influence on susceptibility to cardiac NL in an anti-SSA/Ro environment. It is widely held that genes predispose individuals to, but do not solely account for, disease expression. Other factors, such as events in utero, can modulate the penetrance and severity of genetic predisposition. In the case of cardiac NL, the phenotype is well characterized and covers a limited range of presentation, thus conveying an advantage for studying a fetal genetic component. A candidate gene approach based on the histologic hallmark of disease, fibrosis of the conduction tissue, has focused on transforming growth factor β (TGFβ). The TGFβ polymorphism Leu10 (associated with increased fibrosis) was found to be significantly enriched in children with cardiac NL compared with unaffected offspring and controls (29).

Several limitations of this study should be noted. Although the present data have been obtained from a large cohort, the predominance of Caucasians restricts an extrapolation of the data to other races. The mothers included in this study are all enrolled in the RRNL, which specifically seeks families with NL. There may be a bias in attracting families with more than 1 affected child. However, even when only prospectively evaluated pregnancies are included, the recurrence rate remains considerable, at 16%.

The substantial morbidity and mortality associated with cardiac NL drive the search for preventive therapies. Given the low event rate of disease in anti-SSA/Ro–positive mothers overall, such therapy cannot be easily evaluated in mothers without previously affected children. Thus, the target for prevention studies should be the mothers with a prior affected child. Overall, it is expected that these data will serve as an important reference point for development of family counseling strategies, power calculations for evaluation of prevention studies, estimation of genetic burden, and basic science studies on the pathogenesis of cardiac NL.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Buyon 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 conception and design. Llanos, Izmirly, Clancy, Buyon.

Acquisition of data. Llanos, Izmirly, Katholi, Clancy, Friedman, Buyon.

Analysis and interpretation of data. Llanos, Izmirly, Clancy, Friedman, Kim, Buyon.

Acknowledgements

We would like to thank the mothers enrolled in the RRNL. We also thank Amy Lawless for her help in preparing the manuscript.

Ancillary