Although reports of additional patients with congenital heart block reinforced the relation of the neonatal lupus syndrome with antibodies to the 52-kd and 60-kd SSA/Ro proteins and to the 48-kd SSB/La proteins, the associated maternal clinical status and prognosis remained poorly defined. In 1987, McCune and colleagues [16] assessed the health status of mothers and children with transient and permanent manifestations of the neonatal lupus syndrome in a follow-up study of 21 families. The average follow-up was 4.5 years (range, 0.25 to 9.5 years). All mothers eventually developed symptoms of a rheumatic disease. Fourteen children had congenital heart block. Three died in the neonatal period and 5 of the living children required pacemakers. Of 13 children born after the child with the neonatal lupus syndrome, 2 had congenital heart block and 1 had cutaneous disease.

In the last several years, our laboratory evaluated the sera of 60 mothers whose children had congenital heart block and 5 women currently carrying fetuses with heart block, who were from the continental United States, Canada, and Europe [17, 18]. Given the limited clinical data available, a follow-up study of this extensive group was initiated to address the long-term outcome of these mothers and their children. Moreover, this study evaluated whether maternal autoantibodies to the 52-kd SSA/Ro, 60-kd SSA/Ro, or 48-kd SSB/La proteins (as determined by enzyme-linked immunosorbent assay [ELISA] or immunoblot) were associated with the clinical prognosis of the mother or her child.

Methods

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Collection of Clinical Data

Since 1985, sera from 60 mothers of children with congenital heart block and from 5 women currently carrying fetuses with heart block were analyzed at the Hospital for Joint Diseases, New York University School of Medicine. Several women were followed at the Hospital for Joint Diseases or were referred for their care after congenital heart block was diagnosed. In other cases, we were contacted by the women's physicians because of our known interest in this condition. Two serum samples were obtained early in pregnancies that were later complicated by congenital heart block. Other mothers were referred through the Sjogren Syndrome Foundation or the Lupus Foundation of America. The affected births occurred during 1970 to 1992. Any woman whose serum contained anti-SSA/Ro antibodies alone or in association with anti-SSB/La antibodies (as shown in our laboratory) and whose child (or fetus) had congenital heart block was initially included in the study.

Each woman was sent an explanatory letter and a questionnaire. We also attempted to contact all mothers by telephone to confirm and expand information obtained in the questionnaire. Information was sought about the mother's health at the time of birth of the affected child and her health during the ensuing years. Emphasis was placed on organ system involvement characteristic of a rheumatic disease. Specifically included were questions seeking evidence of dry eyes or mouth, cutaneous lesions, sun sensitivity, cardiac or pulmonary disease, arthritic complaints, and renal or central nervous system disease. Medications used during the affected pregnancy were identified as were any intercurrent illnesses. A family history of rheumatic tissue disease or cardiac abnormalities was also sought. Information about subsequent pregnancies was obtained as was medical and developmental history of the affected child(ren), specifically the presence of heart failure and pacemaker placement.

Further information was obtained from the attending rheumatologist, primary care physician, or pediatrician; available medical charts were reviewed. Eight of the 65 women could not be contacted by mail or telephone and were excluded. Sufficient information to confidently describe initial maternal characteristics was obtained for 57 women; follow-up status could be assessed in 52 of these women (5 are currently pregnant with affected fetuses) and in 55 affected children. In 8 of these 52 women, a telephone interview was not done because the woman had moved and could not be located or because either the woman or her physician was hesitant to release the telephone number. In all these patients, the information was confirmed by the primary physician or the patient's medical records or both and was thus included in the study.

Mothers classified as having systemic lupus erythematosus met at least four of the criteria from the American College of Rheumatology (formerly the American Rheumatism Association) [19]. Mothers were categorized as having the Sjogren syndrome if sicca complaints were present in the absence of other characteristics of a rheumatic disease whether or not objective documentation existed of salivary or lacrimal gland hypofunction or of lymphocytic infiltration in these glands. None of these mothers met criteria for other rheumatic diseases. Undifferentiated autoimmune syndromes were diagnosed in those patients with features of a rheumatic disease who did not have prominent sicca complaints and did not meet criteria for systemic lupus erythematosus.

Characterization of Maternal Antibody Responses to SSA/Ro and SSB/La

Enzyme-linked immunosorbent assays were done as described using commercially purified SSA/Ro from bovine spleen and SSB/La antigens from rabbit thymus (Immunovision, Springdale, Arkansas) [17]. SSA/Ro at 5 µg/mL or SSB/La at 10 µg/mL in 0.05 M sodium bicarbonate with sodium carbonate at pH 9.2 were incubated in 96-well ELISA plates overnight at 4 °C. The ELISA plates were blocked with 1% bovine serum albumin in phosphate-buffered saline with 0.05% Tween, were incubated with sera diluted 1:1000 in phosphate-buffered saline with 0.05% Tween followed by incubation with F(ab')₂ goat anti-human IgG alkaline phosphatase conjugate (Sigma St. Louis, Missouri), and were developed with disodium p-nitrophenyl phosphate in diethanolamine buffer. Results are expressed as the optical absorbance at 405 nm minus that of the reagent blank. A result was considered positive if it was greater than 2 SD above the value obtained for normal healthy donors.

Immunoblot was done as described [20]. Briefly, the cell line MOLT4 was used as the source of antigen. Proteins were separated by electrophoresis and were electrotransferred to nitrocellulose. The nitrocellulose strips were incubated in phosphate-buffered saline containing 0.05% Tween 20 and 3% nonfat milk followed by incubation with a 1:100 dilution of sera. Protein A-(Iodine-125) was used to detect the bound human immunoglobulins.

Statistical Analysis

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Comparisons between groups were calculated using the Fisher exact test. P values (two-tailed) less than 0.05 were considered to be significant.

Results

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Demographic Characteristics and Clinical Status of the Women

Thirty-nine (68%) of the 57 women were white, 12 (21%) were black, and 6 (11%) were Hispanic. Ages ranged from 19 to 38 years. Most lived in the eastern United States, although many were from other areas of the country. One woman was from Canada, and 10 were from northern Europe.

At the time of identification of heart block either at birth or in utero, 23 (40%) of the 57 mothers were asymptomatic (mean age, 27.1 years), 15 (26%) had systemic lupus erythematosus (mean age, 30.1 years), 11 (19%) had an undifferentiated autoimmune syndrome (mean age, 27.7 years), and 8 (14%) had the Sjogren syndrome (mean age, 27.7 years). The specific clinical characteristics of all mothers with an undifferentiated autoimmune syndrome (initially or on follow-up) are included in Table 1 and Table 2. Of the 10 patients categorized with the Sjogren syndrome (initially or on follow-up), 6 had symptomatic dry eyes and mouth in the absence of objective criteria, 2 had dry eyes and mouth and an abnormal salivary gland biopsy, 1 had dry eyes and mouth and a positive Schirmer test, and 1 had dry eyes and mouth and keratoconjunctivitis sicca. Unlike for systemic lupus erythematosus, no universally accepted diagnostic criteria exist for the Sjogren syndrome, although several have been proposed [21-23].

Medication Use by the Mothers

Nine mothers took medications other than nutritional supplements during the affected pregnancies before the identification of congenital heart block. Four were taking prednisone (10 to 40 mg/d) for underlying clinical disease activity; one of these mothers was also taking hydroxychloroquine; two others were receiving anticoagulant agents (subcutaneous heparin) from the time of confirmation of pregnancy until just before delivery (one of whom had been taking oral warfarin early in the first trimester); another woman was prescribed an undetermined medication for protracted vomiting; another was receiving monthly benzathine penicillin injections; and one woman used miconazole vaginal suppositories. After the intrauterine diagnosis of heart block, seven mothers received dexamethasone at doses between 4 and 9 mg/d, three others received prednisone at doses between 40 and 60 mg/d, another received 60 mg of prednisone every other day along with weekly intramuscular injections of 12 mg of betamethasone, and another received only weekly injections of 12 mg of betamethasone. The prednisone dose of another woman was increased from 10 to 30 mg/d after the identification of heart block at 24 weeks of gestation. One woman given dexamethasone was also treated with plasmapheresis [24].

Maternal Clinical Status after Birth of the Affected Child

The time of follow-up for the 52 mothers ranged from 1 month to 20 years (mean, 5 years; median, 3.7 years). During follow-up, 11 mothers who were initially asymptomatic and 1 mother with the Sjogren syndrome had a change in clinical status (Table 2). The rate at which asymptomatic mothers became symptomatic was 0.15 per patient-year of follow-up. The mean and median times to development of symptoms were 2.6 and 1.5 years, respectively, after the affected birth. The total mean and median follow-up times of asymptomatic women whose status changed were 6.2 and 2.2 years, respectively. Eleven women remained asymptomatic during up to 16 years of follow-up (median, 1.7 years). One asymptomatic woman, still carrying her affected fetus, is not included in these calculations. One mother initially categorized as having the Sjogren syndrome fulfilled sufficient criteria for systemic lupus erythematosus when she developed nephritis 4 years later. None of the mothers initially categorized with an undifferentiated autoimmune syndrome developed new manifestations of disease during a median follow-up of 4.8 years. Notably, most of the mothers with systemic lupus erythematosus at the time of the affected pregnancy have not had progression of their disease. Only one woman developed involvement of the central nervous system, and none developed end-stage renal disease. The one maternal death was caused by myocardial infarction. None of the mothers had heart block.

Characteristics of the Children

Thirty-three of 61 offspring (55 born and 6 in utero) were boys (54%; 95% CI, 41% to 57%) and 28 (46%) were girls. One mother who previously gave birth to a child with congenital heart block is currently pregnant with a second affected child. Children with congenital heart block followed and preceded the birth of an unaffected sibling. Twenty-nine (48%) children were first pregnancies, 21 (34%) were second pregnancies, 4 (7%) were third pregnancies, 6 (10%) were fourth pregnancies, and 1 (2%) was a fifth pregnancy. Birth dates and estimated dates of conception were randomly distributed throughout the year. Two children were the result of twin pregnancies, one monozygotic and one dizygotic. Both sets were discordant for congenital heart block.

Heart block was identified in most of the fetuses before birth. In 30 (49%) of 61 pregnancies, the diagnosis was made between 16 to 24 gestational weeks; in 12 (20%) pregnancies, between 25 to 28 gestational weeks; in 10 (16%) pregnancies, between 29 to 37 weeks; and in 5 (8%) pregnancies, between 38 and 40 weeks. In four patients, this information was not available. Because congenital heart block is frequently detected by ultrasonography, it is not surprising that six of the nine patients in whom heart block was first noted at birth occurred more than 8 years ago when prenatal care may not have routinely included this measure of fetal surveillance. In the three recent children with congenital heart block who were identified at birth, fetal ultrasound had not been done after the 18th week of gestation.

Outcome of the Children

Seventeen (31%) of the 55 children died, including one stillbirth (heart block diagnosed by fetal echocardiogram 12 weeks before death). These deaths included nine boys and eight girls. Twelve (71%) of the 17 deaths occurred in the first month of life (Figure 1); the latest death occurred at 2.9 years of age. Of the 12 early deaths, 4 resulted from congestive heart failure, 3 from multiorgan failure, 2 from respiratory failure, and 2 from unknown causes. Of the 5 deaths in older children, 3 resulted from congestive heart failure, 1 from pacemaker failure, and 1 from multiple complications of pacemaker placement on the first day of life, including cardiac arrest and bilateral phrenic nerve paralysis.

View larger version (12K): [in this window] [in a new window]   Figure 1. Timing of pacemaker implantation and death for 55 children with congenital heart block. The x-axis represents various time periods from birth, and N represents the number of children remaining alive for any part of these designated periods. Thus, the decrease in N along the x-axis represents children not yet old enough to be included in the next age range and deaths that occurred in the previous age range. The y-axis represents the number of children who died and those who received pacemakers during these designated periods. For example, 12 children were followed for at least 5 years from birth and are included in the age range "5 to 10" years. None died and one had a pacemaker implanted for the first time, although some children needed pacemakers at earlier ages.

Thirty-seven (67%) of the 55 children required pacemakers. Twenty-seven (73%) received pacemakers within the first 3 months of postnatal life (Figure 1). Of the 13 children older than 5 years, only 4 (31%) do not have pacemakers, including 1 woman who is now 20 years old. Of 7 children now more than 10 years old, none has developed a rheumatic disease. For the seven mothers who received dexamethasone after the diagnosis of heart block in utero, 1 pregnancy has not yet come to term, 1 child died at 19 months of age, and 5 children are alive at ages of 0.4 to 1.7 years (2 with pacemakers). Of the two mothers treated during their affected pregnancy with intramuscular betamethasone, both children are alive with pacemakers (one is 6 years old and the other is 6 months old); neither child shows signs of congestive heart failure.

Subsequent Pregnancies

Twenty-two women have had 25 subsequent pregnancies; 4 (16%), each from a different mother, have resulted in a recurrence of congenital heart block. Three subsequent children without congenital heart block had rashes related to neonatal lupus syndromes. Another child died of hepatic dysfunction at 2 months of age; autopsy findings were consistent with neonatal hemochromatosis [25]. One mother gave birth to a second child with congenital heart block (died 1 hour after birth) despite taking 30 mg of prednisone daily from 9 weeks of gestation to delivery.

Maternal Autoantibodies and Outcome of Mother and Child

Using ELISA, 100% of the sera from 52 women who gave birth reacted with SSA/Ro and 40 (77%) reacted with SSB/La. Using immunoblot analysis, 29 (56%) reacted with the 60-kd SSA/Ro protein, 41 (79%) with the 52-kd SSA/Ro protein, and 44 (85%) with the 48-kd SSB/La protein. No association was noted between the presence of antibodies to the 48-kd SSB/La, 52-kd SSA/Ro, or 60-kd SSA/Ro antigens as measured by immunoblot and the final maternal clinical status (P > 0.2 for all comparisons). In addition, the antibody profile was similar between asymptomatic mothers who progressed to rheumatic diseases and those who remained asymptomatic (P > 0.2).

As assessed by immunoblot, the serum samples from all 16 mothers whose children died had antibodies to the 48-kd SSB/La protein compared with 28 (78%) serum samples from 36 mothers of surviving children (P = 0.089) (Table 3). In contrast, 10 (63%) serum samples from mothers of children who died contained antibodies to the 52-kd SSA/Ro protein compared with 31 (86%) of mothers whose children survived (P = 0.073). Fifty-six percent of the sera in both groups contained antibodies to the 60-kd SSA/Ro protein (P = 1.0). In 58% of patients, the serum sample was obtained during the affected pregnancy.

Discussion

Top Methods Statistical Analysis Results Discussion Author & Article Info References

This is the largest follow-up study of mothers and their children with autoantibody-associated congenital heart block. Such information is critical for counseling families regarding mother's and children's prognosis, discussing expectations for future child bearing, and determining possible prophylactic or therapeutic modalities or both. The women included represent a diverse ethnic and socioeconomic group, and the results obtained should be generalizable to mothers of children with anti-SSA/Ro antibodies alone or in association with anti-SSB/La antibodies whose children have congenital heart block. Most women did not have systemic lupus erythematosus at the time of birth of their affected offspring but were asymptomatic.

Nearly half of the mothers who were initially asymptomatic and were identified only by the birth of a child with congenital heart block developed rheumatic complaints or fulfilled diagnostic criteria for a rheumatic disease. The other asymptomatic mothers remained entirely healthy at the time of last follow-up. One woman has remained asymptomatic for 16 years. Thus, the onset of disease may not be universal in these mothers. Because the median length of follow-up was only 1.7 years in mothers who remained asymptomatic, it is possible that other mothers will eventually develop symptoms. However, in the present series, those women who became symptomatic usually did so soon after the affected birth (median, 1.5 years). In addition, it is intriguing and reassuring that, except for two mothers, those with the initial diagnosis of systemic lupus erythematosus have not had progression with major organ involvement and those mothers with the undifferentiated autoimmune syndromes have not progressed to defined rheumatic diseases. Most of the mothers had antibodies to both SSA/Ro and SSB/La. This is of interest with regard to systemic lupus erythematosus because such patients apparently have less renal disease than those with only anti-SSA/Ro antibodies [26]. Despite this encouraging data, caution about long-term prognosis is warranted based on the strong immunogenetic similarities between mothers of children with congenital heart block and patients with overlap syndromes (features of the Sjogren syndrome and systemic lupus erythematosus) who are reported to have progression of their systemic disease when followed for 30 to 40 years [27].

Death in the children was associated with maternal 48-kd anti-SSB/La antibodies, whereas survival was associated with maternal 52-kd anti-SSA/Ro antibodies, although these results were not statistically significant. Arguably, the frequent finding of both antibodies in both groups of mothers decreases the value of such determinations in predicting the outcome of the child. Antibody reactivities to the 60-kd SSA/Ro component did not distinguish children who died from those who survived. Because not all samples were obtained during the affected pregnancy, the antibodies identified may not have been those to which the fetus was exposed. However, antibodies to either SSA/Ro or SSB/La remain relatively constant, are not associated with fluctuations in autoimmune disease activity, occur in asymptomatic persons, and are not appreciably affected by glucocorticoids [28]. The antibody profiles did not predict either the final maternal clinical status or the occurrence of disease in the asymptomatic mothers.

Congenital heart block is a model of passively acquired autoimmunity [29], and the elucidation of the mechanisms involved might be generalizable to the understanding of other immune-mediated tissue injury. Although the target antigens of anti-SSA/Ro and anti-SSB/La antibodies are ribonucleoproteins normally sequestered intracellularly, and thus inaccessible, several lines of evidence support the hypothesis that fetal tissue damage is mediated by these autoantibodies. Isolated congenital heart block usually develops in a previously normal heart between 16 and 24 weeks of gestation, coincident with the increased transfer of maternal IgG antibodies into the fetal circulation [14, 24]. SSA/Ro and SSB/La antigens have been identified in fetal cardiac tissue, including the conducting system [17, 30]. One study [31] found that a 23-week-old fetal heart contained a greater quantity of SSA/Ro per milligram of protein than did 18- to 22-week-old hearts or an adult heart, thus supporting the possible preferential vulnerability of the fetal heart compared with the adult heart. Horsfall and colleagues [32] eluted anti-SSB/La antibodies from the cardiac tissue of a child with fatal congenital heart block. Antibodies to SSA/Ro altered the plateau phase of repolarization of the action potential of neonatal, but not adult rabbit ventricular papillary muscles [33].

The presence of SSA/Ro and SSB/La antibodies in maternal blood is not sufficient to cause congenital heart block. A second event, perhaps one that makes the target antigens accessible to the autoantibodies at a crucial phase of development, may be necessary for cardiac damage to occur. A viral infection could be a candidate for this second event. Adenovirus infection of human epithelial cells promotes translocation of SSB/La to the cell surface [34]. Although no specific pattern of viral illness was noted during the affected pregnancies reported here, we cannot rule out the possibility of subclinical infection. However, a seasonal variation in the dates of birth or conception of these children as a marker for a seasonal viral infection was not identified.

Substantial morbidity and mortality is associated with congenital heart block. Pacemakers were implanted in two thirds of the children, often in the first months of life. Indications for placement of pacemakers in children include daytime heart rates averaging less than 50 beats/min (especially when associated with episodes of junctional exit block or a flat junctional response to exercise [35]), a prolonged QT interval [36], and a wide QRS complex [37]. In this study, the mortality rate associated with isolated congenital heart block was 31%. Two previous studies [38, 39] have shown substantially lower mortality rates in children with congenital heart block in the absence of structural abnormalities. Several explanations are plausible. In our study, an accrual bias could account for an enrichment with children having more severe disease, but there is no obvious reason to believe that this is the case. The other studies [38, 39] have included children diagnosed months to years after birth, and those children with heart block dying soon after birth may have been missed. Thus, including children with late diagnoses could have led to lower estimates of mortality. Further, the largest multicenter study [38] assessing outcome of these children predated recognition of the autoantibody association with congenital heart block, which may well represent a distinct pathogenesis. Because the presence of autoantibodies was a prerequisite for inclusion in the present study, the groups may not be comparable. Indeed, in another study [40], two of six children with congenital heart block born to mothers known to have anti-SSA/Ro antibodies died, a rate similar to that reported here.

Because fetal arrhythmias can now be evaluated prenatally using echocardiographic and Doppler ultrasound techniques, atrioventricular block is more frequently identified during the fetal period. Auscultation may be misleading because the atrial rate may be mistaken for the ventricular rate [41, 42]. In our experience and that of others, bradycardia, as determined by fetal echocardiography, has been noted after the fetus had an initially normal heart rate [24, 43]. As previously reported in a few fetuses and confirmed in this larger study [15, 44-47], congenital heart block is most often detected before the 30th week of gestation. Fetal echocardiogram is essential to follow the course of disease and may suggest the presence of an associated myocarditis by finding decreased contractility and secondary changes such as an increase in cardiac size, pericardial effusions, and tricuspid regurgitation. Serial fetal echocardiography at 16, 18, 22, and 24 weeks of gestation may be reasonable for all women at risk for a child with congenital heart block; more frequent echocardiography may be reasonable if abnormalities are detected. Dexamethasone therapy has been administered to women carrying fetuses with heart block and myocarditis in an attempt to diminish the inflammatory insult to the fetus. Resolution of pleuropericardial effusions and ascites occurred during dexamethasone administration [24, 44]. Dexamethasone and betamethasone have been recommended rather than prednisone because these fluorinated steroids are not metabolized by the placenta and is available to the fetus in an active form [48, 49].

Recurrence of congenital heart block in subsequent pregnancies was uncommon. To date, no definitive evidence exists that any prophylactic intervention, including plasmapheresis to lower the maternal antibody titer or glucocorticoids to diminish the inflammatory insult, prevents the development of congenital heart block. In one of the recurrences reported in this study, the second child with congenital heart block was born to a mother who took 30 mg of prednisone daily from the ninth week of gestation through delivery. Given the available data, prospective management other than serial echocardiograms is not justified. No gender-based difference in the frequency or prognosis of congenital heart block was found. Data [50] support an increased rate of skin manifestations of neonatal lupus syndromes in girls compared with boys. Several reports [51-53] describe rheumatic diseases developing in children with neonatal lupus syndromes. None of the children in the present study has developed a rheumatic disease, including one who is now 20 years old.

Although many mothers of babies with congenital heart block are asymptomatic, women with rheumatic diseases are more likely to have this complication. It has been estimated [38] that only 1 of every 20 000 live births results in a baby with isolated congenital heart block. In one retrospective study, however, the rate was 7 of 259 live births of mothers with systemic lupus erythematosus and in those with anti-SSA/Ro antibodies, the rate was 6 of 79 [40]. Although others [54] have found substantially lower rates, selected groups of women may be at greater risk for this pregnancy complication [18].

This study shows that autoantibody-associated congenital heart block has substantial morbidity and mortality, with a death rate of nearly one third and a requirement for permanent cardiac pacing in most surviving children. Death may be more common in those children whose mothers have antibodies to SSB/La. Congenital heart block occurs equally in both sexes, and gender does not affect outcome. To date, the recurrence rate is 16% (4 of 25 pregnancies). More mothers are initially asymptomatic than in any other single clinical category and many, but not all, of these mothers progress to a rheumatic disease.

Addendum

Since the submission of our manuscript, a 26th subsequent pregnancy has resulted in a healthy baby; the recurrence rate is now 15% (4 of 26). The health status of two mothers has changed; one has progressed in 4 months from an undifferentiated autoimmune syndrome to systemic lupus erythematosus (lymphopenia, pleuritis, anti-DNA, proteinuria), and a second mother, initially asymptomatic, developed leukopenia, arthralgias, and painless mouth ulcers during a subsequent pregnancy.