Author + information
- Received October 15, 2009
- Revision received January 19, 2010
- Accepted February 15, 2010
- Published online June 15, 2010.
- Edgar Jaeggi, MD*,§,* (, )
- Carl Laskin, MD†,
- Robert Hamilton, MD*,§,
- John Kingdom, MD† and
- Earl Silverman, MD†,‡
- ↵*Reprint requests and correspondence:
Dr. Edgar Jaeggi, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada
Objectives The purpose of this study was to determine whether cardiac complications of neonatal lupus erythematosus (NLE) are related to maternal anti-Ro and anti-La autoantibody-levels.
Background Autoantibody-positive mothers are frequently referred for serial echocardiography because of the elevated fetal risk of developing immune-mediated heart block. Little is known why only some and not all offspring are affected.
Methods All cases referred since 2000 for serial fetal echocardiography or cardiac complications related to maternal antibodies were included. Patients without cardiac NLE (group 1) and with cardiac NLE (group 2) were compared. Antibody levels were measured by enzyme-linked immunosorbent assay with a cutoff value of 8 U/ml for a positive test result.
Results Group 1 included 146 serially screened fetuses with normal pregnancy outcomes. Group 2 consisted of 40 fetuses/neonates with a diagnosis of heart block or endocardial fibroelastosis or both, and included 4 fetuses diagnosed during serial screening. All cardiac complications were associated with moderate (≥50 U/ml; 15%) or high (≥100 U/ml; 85%) maternal anti-Ro levels, independently of anti-La antibody titres. The event rate of complete heart block was 5% for prospectively screened fetuses with Ro-values ≥50 U/ml (odds ratio: 7.8) and 0% for fetuses with lower titres (p < 0.0001). Infants with pre-natal exposure to high-titre anti-La levels ≥100 U/ml were the most likely to have noncardiac features of NLE (event rate: 57%; odds ratio: 4.7).
Conclusions Our findings support that the amount of maternal antibodies, rather than their presence, is associated with fetal tissue injury. As anti-Ro levels correlate with the risk of cardiac complications, serial echocardiography should be limited to women with high anti-Ro-titres.
Isolated congenital complete atrioventricular block (CAVB) has been recognized as a distinct clinical entity for 100 years. It was only in the 1950s that it was recognized that maternal autoantibodies will cross the placenta and that fetuses of mothers with an autoimmune disease may develop CAVB (1,2). It took another 20 years before this association became well recognized (3–5) and until the early 1980s that studies documented a close association between maternal anti-Ro/SSA and anti-La/SSB antibodies and congenital heart block (6,7). It was subsequently shown that the highest relative risks of CAVB were seen in offspring of mothers with antibodies against 52-kD Ro and 48-kD La proteins (8,9).
The current hypothesis is that maternal antibodies with specificities to the Ro-particle cross the placenta and likely initiate the inflammation of the atrioventricular (AV) node and the myocardium in the susceptible fetus. Subsequent replacement of inflamed tissue with fibrosis will lead to heart block, myocardial dysfunction, and/or endocardial fibroelastosis (EFE) (10,11). The most common time of antibody-mediated fetal cardiac manifestations to be initially detected is between 20 and 24 gestational weeks (12). Other manifestations of the so-called neonatal lupus erythematosus syndrome (NLE) are frequently observed after birth and include cutaneous rash, thrombocytopenia, leukopenia, anemia, and liver dysfunction (13). Unlike CAVB, the noncardiac symptoms of NLE usually resolve within a few months after birth, coincident with the clearance of the maternal antibodies from the child's circulation.
It is unclear why NLE develops in only some but not all antibody-exposed fetuses. In prospectively examined pregnancies of mothers with anti-Ro and anti-La antibodies, the reported prevalence of fetal CAVB is 1% to 5%, which rises to 6% to 25% for those with a previously affected child with heart block (14–17). Because of the fetal risk of acquiring potentially life-threatening, irreversible heart disease and the perception that this may be preventable if detected and treated at an early stage of AV nodal disease, it is recommended that antibody-positive mothers should be referred for close fetal echocardiographic surveillance beginning in the early second trimester (18–20).
The aim of this study was to determine if maternal anti-Ro and -La antibody levels rather than their presence could be used to more accurately predict fetuses at risk of immune-mediated cardiac complications.
The Hospital for Sick Children provides the exclusive tertiary cardiac care for approximately 80,000 live births per year. That includes the only program offering serial fetal echocardiography for women with anti-Ro/-La antibodies in our catchment area. The institutional research ethics board approved this prospective observational single center study.
The study cohorts consisted of 186 cases exposed to maternal anti-Ro and -La antibodies referred either: 1) with a diagnosis of immune-mediated cardiac complications (symptomatic cohort; n = 40); or 2) for serial echocardiography with a normal pregnancy outcome (asymptomatic cohort; n = 146) between January 2000 and December 2008. The presence of these antibodies was confirmed in a commercial and/or institutional laboratory before the referral for serial screening. Antibody-levels were determined in our laboratory at the time of referral and, when possible, in live-born infants by solid-phase enzyme-linked immunosorbent assay (ELISA) using recombinant human 60-kD Ro, 52-kD Ro, and 48-kD La proteins (Phadia GmbH, Freiburg, Germany). The cutoff for a positive test was 8 U/ml. Antibody levels were graded as negative (<8 U/ml), low positive (8 to 49 U/ml), moderate positive (50 to 99 U/ml), and high positive (≥100 U/ml). Our ELISA test did not allow the quantification of titres >100 U/ml.
Cohort with cardiac NLE
Heart block, myocardial dysfunction, pericardial effusion, and EFE were considered as cardiac manifestations of NLE. We prospectively collected the clinical and outcome data of 39 mothers and their 40 offspring with a fetal (n = 34) or neonatal (n = 6) diagnosis of immune-mediated heart disease. They included 4 fetal cases with a primary referral for serial echocardiographic surveillance with a diagnosis of first-degree (n = 1) or third-degree (n = 3) AV block.
Cohort without cardiac NLE
This cohort consisted of 125 mothers with 146 prospectively evaluated fetuses with a normal pregnancy outcome. Mothers were mainly referred on the basis of a diagnosed connective tissue disease (81%), a previous child with CAVB (6%), or the presence of autoantibodies in an otherwise healthy woman (13%). Mothers with no pregnancy history of cardiac NLE were invited to undergo weekly fetal echocardiograms between 19 and 24 weeks. If a previous child had CAVB, we recommended 1- to 2-weekly echocardiograms until 32 weeks. Informed consent was obtained before enrollment. Our protocol also called for post-natal follow-up examinations of the child by a rheumatologist (E.S.) beginning at 4 to 6 weeks of age. This included physical examinations, a 12-lead electrocardiogram, and serological, hematological, and biochemistry tests. Post-natal examinations and electrocardiograms were obtained in 111 and 116 cases, respectively, within the recommended time frame. The findings were compared with the age-matched normal values.
Echocardiography with Philips iU22/iE33 (Philips ATL, Bothwell, Washington) and GE Vivid-7 (GE Medical Systems, Horten, Norway) ultrasound systems entailed 2-dimensional, Doppler, and M-mode imaging to detect cardiovascular abnormalities. Collected data included atrial and ventricular rates, and in cases with 1:1 AV conduction, the Doppler-derived AV durations (21). Ventricular contractility was assessed quantitatively by M-mode echocardiography and after birth by left ventricular ejection fraction.
Study definitions and treatment
Abnormalities were classified as bradycardia (fetal heart rate <120 beats/min), AV block, reduced fractional shortening <28%, EFE, and/or effusions in body cavities. First-degree AV block was defined as persistent PR prolongation >2 z-scores above the age-specific normal mean values (21,22). CAVB implies complete absence of electrical AV conduction. The EFE was identified by echocardiography as areas of increased brightness of endocardial surfaces, and if possible, confirmed by necropsy (23). It was graded as severe if there was extensive ventricular involvement.
Transplacental anti-inflammatory treatment as per our institutional protocol (23) was used if 1 of the following was detected: CAVB, reduced contractility, EFE, or effusions. If EFE was detected, intravenous immunoglobulin (IVIG) was also given to the mother (70 g every 2 to 3 weeks to delivery) and to the newborn (single dose of 2 g/kg).
Results are expressed as median (range), mean ± SD, or frequencies, as appropriate. Comparisons between groups were performed with the chi-square test for categorical variables and the Student ttest for continuous variables. The odds of NLE were calculated in the cohort of prospectively examined fetuses with different levels of antibody exposure. Statistical analyses were performed with Prism 5 (GraphPad Software, San Diego, California).
To determine the role of maternal antibody specificity in NLE, we compared the cohorts with and without cardiac complications (Table 1).All mothers in this study had detectable anti-Ro antibodies, and approximately one-third of them also had anti-La antibodies. Mothers with a normal pregnancy outcomes were mainly referred by their rheumatologist for serial echocardiography. When we examined maternal health, as may be expected, mothers of children with cardiac involvement were less likely to have had a connective tissue disease than were mothers of children without cardiac involvement (15% vs. 81%; p < 0.0001).
Clinical features of the cohort with cardiac NLE
Table 2shows the characteristics, management, and outcome of 40 consecutive cases with a fetal diagnosis (n = 34; cases 1 to 34) or a neonatal diagnosis (n = 6; cases 35 to 40) of cardiac NLE. Cases 1 to 4 were part of a prospective study and had initially been referred for serial fetal echocardiography in a mother known to have anti-Ro antibodies before the detection of heart block. Abnormalities were first diagnosed in-utero at a median of 22.5 (18 to 37) gestational weeks or immediately after the delivery. Heart block was present in 37 of 40 (93%) cases, and the other 3 had either isolated pericardial effusion (case 21), EFE (case 23), or EFE with hydrops (case 31). Of 37 cases with heart block, it was complete in 35 cases, first degree in 2 cases, and associated with EFE in 10 cases. All 39 mothers had anti-Ro antibody levels ≥50 U/ml, but only 6 had connective tissue disease, and 1 woman had a previous child with CAVB. Interestingly, of 8 fetuses/infants with both anti-Ro and anti-La antibody levels ≥50 U/ml, none had EFE, fetal hydrops, reduced ventricular function, and/or died, whereas that was the case in 12 (38%) of 32 cases with exposure only to high anti-Ro levels (p = 0.08).
Six of the 34 pre-natally detected cases did not receive intrauterine treatment because the referral occurred late in gestation (case 12), the fetus had died (case 31), treatment was declined (case 28), the pregnancy was terminated (cases 2 and 10), or mild first-degree AV block was present and did not progress (case 4). The other 28 pregnancies were treated, including 3 cases with cardiac disease other than CAVB. Case 21 received corticosteroids as a result of treatment for its twin (case 20): the coincidental finding of pericardial effusion resolved within a week of treatment, whereas CAVB persisted in the sibling. Case 22 presented with bright areas in both ventricles that almost completely resolved with transplacental treatment. Case 23 had diffuse EFE of all papillary muscles, reduced biventricular contractility, and first-degree AV block with a significantly prolonged AV interval of 250 ms (z-score +14) by superior vena cava/aorta pulse-wave Doppler. On treatment, the cardiac function normalized within 1 week, the EFE improved, and the neonatal electrocardiogram confirmed first-degree AV block with a PR duration of 250 ms. Six pregnancies were referred perinatally and did not receive transplacental treatment (cases 35 to 40).
Outcome of the cohort without cardiac NLE
A median of 4 (2–12) echocardiograms were performed in 146 asymptomatic fetuses between 19 weeks (17 to 23 weeks) and 24 weeks (23 to 35 weeks). None experienced AV block, bradycardia, valvar regurgitation, EFE, cardiac dysfunction, and/or body effusions during the observation period. Electrocardiograms were obtained in 116 cases at 52 days (5 to 232 days) of life, showing sinus rhythm (153 ± 16 beats/min) with normal AV conduction (97 ± 10 ms). Seventeen (15%) had transient QTc prolongation between 440 and 470 ms. None was treated with a beta-blocker, and no sudden infant death syndrome occurred. Of 30 children with no post-natal examinations at our center, all had normal heart rates at birth and are reportedly healthy according to their physicians. That likely precludes major undetected cardiac complications. A post-natal rheumatology evaluation was obtained for 111 infants, and 30 (27%) had transient hematological abnormalities (n = 14; 13%), abnormal liver function tests (n = 4; 4%), and/or a skin rash (n = 17; 15%). Table 3shows the relationship between maternal antibodies and noncardiac NLE.
Role of antibody titres on the development of NLE
All mothers of children with cardiac involvement had moderate to high anti-Ro antibody levels (≥50 U/ml) as compared with only 44% of mothers of healthy infants (p < 0.0001). When we examined only the prospectively followed pregnancies, the event rate of CAVB was 5% (3 of 59 cases) for fetuses with exposure to anti-Ro levels ≥50 U/ml (odds ratio [OR]: 7.8; range 0.4 to 159) and 0% for those exposed to lower Ro-titres. Conversely, infants of anti-La antibody-negative mothers were the least likely to have post-natal features of noncardiac NLE, with an event rate of 19% (13 of 69 cases) and an OR of 0.34 (range 0.13 to 0.72). In contrast, 57% of infants with pre-natal exposure to anti-La levels ≥100 U/ml had noncardiac NLE (OR: 4.7).
Despite the consistent finding that only a small minority of antibody-exposed fetuses will eventually have clinically relevant cardiac complications, there is the consensus that anti-Ro/La antibodies are relevant markers of an increased fetal CAVB risk and, therefore, that pregnant women carrying these antibodies should undergo serial fetal echocardiography beginning in mid-gestation (18–20). Nevertheless, indiscriminate screening of all antibody-positive pregnant women is probably neither practical nor justified. As confirmed in this study, most cases of cardiac NLE are detected in asymptomatic women without connective tissue disease (85%) and without a history of a previous child with heart block (98%). Because the search for Ro- and La-antibodies is not integrated in pregnancy screening, serial echocardiography of “at risk” pregnancies will, therefore, not detect the majority of the fetuses destined to have cardiac disease before its development.
Furthermore, autoantibodies are relatively common in childbearing women. In an earlier prospective study on 15,114 unselected pregnant women in Ontario, we found Ro and La antibodies in 2.5% of maternal sera at around 16 weeks of gestation (E. Silverman, L. Hornberger, R.M. Hamilton, unpublished data, July 2002). In our catchment area, there are approximately 80,000 live births per year. If the cardiac NLE risk of the estimated 2,000 anti-Ro/La antibody-positive mothers is ≥1%, as suggested by different authors (14–17), then we would expect the referral of at least 20 fetuses/neonates per year with antibody-associated CAVB in our area. Yet, we regularly encounter 4 or 5 new cases per year, which is similar to the estimation 4 decades ago that about 1 of every 20,000 live births results in a baby with CAVB (24). Given that the average risk of an antibody-positive woman for fetal CAVB is perhaps only about 0.2%, serial fetal echocardiography appears not indicated for many antibody-positive woman. The aim of this study was therefore to identify a more predictive risk profile for the development of cardiac NLE, using commercially available measurements of anti-Ro and anti-La antibodies.
Previous studies have shown that all cases of cardiac NLE are associated with antibodies to at least 1 of the 52- and 60-kD Ro-particles and/or to the 48-kD La-protein (8,9). More recently, the response to peptides within the Ro- and La-proteins has also been examined (9,25,26). While this approach has provided new insights into the reasons for the low prevalence of cardiac NLE, it has little clinical utility as these assays are available in research facilities only. In this study, using commercially available ELISA, we found not only that the risk of NLE significantly correlated with maternal antibody-levels but also demonstrated a cutoff Ro-titre, below which we did not observe the development of cardiac NLE. Antibody-related cardiac complications occurred exclusively when a fetus was exposed to anti-Ro antibody levels ≥50 U/ml, irrespective of anti-La antibody levels. Nearly 60% of serially screened mothers in this study had Ro-titres <50 U/ml with, in hindsight, no risk of major antibody-related cardiac complications.
Although high levels of anti-Ro antibodies appear necessary, they are not sufficient to explain the development of cardiac NLE because the majority of women with high anti-Ro levels have normal pregnancy outcomes. In 5% (3 of 59) of all prospectively screened pregnancies in this study, exposure to moderate or high anti-Ro antibody levels resulted in a child with CAVB. Of mothers with a previous child with CAVB, the recurrence rate was 11% (1 of 9 pregnancies), which is consistent with previous reports (12,14). All women with a previously affected child in this study also had high anti-Ro antibody levels >100 U/ml, which may be part of the increased recurrence risk. If symptomatic fetuses were also exposed to high La-antibody levels, however, they appeared to be less severely affected as none had concomitant myocardial inflammation and EFE as compared with 38% of fetuses with exposure only to high anti-Ro levels. The importance of a high anti-La antibody level as a marker of a more benign disease pattern needs to be further addressed in a large prospective study. Finally, high-titre anti-La antibody exposure increased the child's risk of antibody-mediated noncardiac abnormalities developing after birth, although these antibodies are probably not causal because only 60% of mothers of affected infants had anti-La antibodies. This association is consistent with previous studies, including our own, which showed that in children with NLE, high-titre anti-La antibodies were more likely to be associated with cutaneous NLE (9,27,28).
The finding of a strong association between anti-Ro antibody levels and cardiac NLE supports the hypothesis that lowering or eliminating autoantibodies may prevent such complications. Some investigators have used plasmapheresis to eliminate the harmful antibodies in high-risk pregnancies. However, despite the demonstration that plasmapheresis can decrease anti-Ro antibody levels, this therapy was associated with a 7% to 14% rate of CAVB in the largest reported studies (29,30). More recently, it has been suggested that the repeated maternal administration of IVIG could prevent tissue damage by decreasing the transplacental passage and by increasing the catabolism of the maternal anti-Ro antibodies, as well as by modulating inhibitory signaling on macrophages with consequent reduction of inflammation and fibrosis. Although in a murine model the administration of IVIG inhibited the transfer of autoantibodies across the placenta and their subsequent deposition in the fetal heart (31), a small study in humans showed that, in 2 of 8 offspring, CAVB or cutaneous NLE still developed (32). A larger prospective study of 3-weekly maternal administration of IVIG between 12 and 24 gestational weeks to prevent cardiac NLE is under way.
The cutoff levels used in this study are only valid for the commercially available ELISA kit used in this study. The method is highly sensitive and quantitative to recognize Ro 60-kDa and Ro 52-kDa particles but does not distinguish among these specificities.
In summary, our study supports the concept that the development of immune-mediated CAVB and EFE is likely linked to the fetal exposure to high titres of anti-Ro antibodies, whereas the presence of anti-La antibodies may not be required. However, our study does not provide direct evidence that anti-Ro antibodies are the pathogenic antibodies but, rather, just demonstrates association. We have again shown that most mothers who have a pregnancy complicated by significant fetal cardiac involvement are not identified before the pregnancy because they do not a have connective tissue disease and thus are not referred for prospective fetal evaluation. Consequently, to improve the yield of early detection before CAVB, we agree with a previous suggestion that a successful program to prevent severe cardiac complications would require the screening for anti-Ro antibodies of all pregnant women irrespective of their symptoms or clinical status (33).
Based on our experience, we suggest that the measurement of anti-Ro antibody levels should be performed rather than simply test for their presence and that fetal echocardiography should be reserved for pregnant women with high levels. However, individual risk assessment based on antibody titres may help alleviate parental anxiety of having a child at risk of an irreversible cardiac condition when anti-Ro antibody-levels are low. The utility of this approach needs confirmation in a large prospective cohort study.
Dr. Silverman is on the advisory board of Abbott Canada and Amgen/Wyeth Canada.
- Abbreviations and Acronyms
- complete atrioventricular block
- endocardial fibroelastosis
- enzyme-linked immunosorbent assay
- neonatal lupus erythematosus
- odds ratio
- Received October 15, 2009.
- Revision received January 19, 2010.
- Accepted February 15, 2010.
- American College of Cardiology Foundation
- Hogg G.R.
- McCue C.M.,
- Manatakas M.E.,
- Tingelstad J.B.,
- Ruddy S.
- Esscher E.,
- Scott J.S.
- Schmidt K.G.,
- Ulmer H.E.,
- Silverman N.H.,
- Kleinman C.S.,
- Copel J.A.
- Nield L.E.,
- Silverman E.D.,
- Taylor G.P.,
- et al.
- Buyon J.P.,
- Hiebert R.,
- Copel J.,
- et al.
- Costedoat-Chalumeau N.,
- Amoura Z.,
- Lupoglazoff J.M.,
- et al.
- Friedman D.M.,
- Kim M.Y.,
- Copel J.A.,
- et al.,
- PRIDE Investigators
- Rein A.J.,
- Mevorach D.,
- Perles Z.,
- et al.
- Nii M.,
- Hamilton R.M.,
- Fenwick L.,
- Kingdom J.C.,
- Roman K.S.,
- Jaeggi E.T.
- Jaeggi E.T.,
- Fouron J.C.,
- Silverman E.D.,
- Ryan G.,
- Smallhorn J.,
- Hornberger L.K.
- Kaaja R.,
- Julkunen H.