Discussion

In this large US-based registry of cardiac NL restricted to maternal anti-SSA/Ro exposure, there were 57 deaths (17.5%), with approximately one third of these offspring dying in utero. There was a significantly higher case fatality rate in minorities compared with whites. In addition, there was a trend toward increased mortality in children of mothers who had an established diagnosis of SLE and/or SS at the time of pregnancy that became significant in the multivariable analysis. Several fetal echocardiographic risk factors were associated with a statistically significant increase in mortality: the presence of hydrops, DCM, EFE, and valvular dysfunction. A higher ventricular nadir rate was protective. In multivariable analysis, the only significant echocardiographic predictors of mortality were hydrops and EFE. Separate analyses on fetuses dying in utero and children dying after birth revealed similar echocardiographic predictors of mortality.

In the analyses limited to in utero deaths, predictors of mortality, in addition to those identified by echocardiogram, included earlier gestational age at the time cardiac NL was first detected. This finding suggests that earlier injury results in more extensive damage to the cardiac structures. If an earlier event targets the majority of the fetal heart, it may result in a more severe lesion, such as cardiomyopathy. In a later event in which the exposed and vulnerable targets are restricted to the isolated conduction system tissues, the insult may not be lethal. The association of in utero death with more recent year of pregnancy, although modest, was initially counterintuitive. This may reflect increased awareness of anti-SSA/Ro–attributed cardiac disease that might otherwise have been categorized as an unexplained fetal demise. Additional predictors of in utero mortality included the use of fluorinated steroids and/or terbutaline. Each of these medications is generally prescribed in more severe cases, which is the likely reason for their association with mortality rather than direct causality. In the analysis limited to children who died after birth, as expected, preterm delivery was associated with an increased mortality. A later week of delivery and a higher ventricular nadir rate were protective, the latter remaining in the multivariable analysis of predictors for mortality in live-born children.

With regard to morbidity, pacemakers were required in the majority of children, with most paced in the neonatal period. Pacemaker placement was not associated with mortality in the overall analysis or live birth analysis, suggesting that pacemaker-induced cardiomyopathy may not be a common cause of mortality in this cohort.

Previous data on risk factors associated with mortality, although available, are limited by the heterogeneity of the cohorts, because not all children have documented exposure to maternal anti–SSA/Ro-SSB/La antibodies. Moreover, most reports reflect cases seen at tertiary medical centers; thus, data may be skewed toward the sickest patients. Impaired left ventricular function has been a common denominator associated with poor prognosis. In a study based in France, all 6 deaths in anti-SSA/Ro–exposed children with congenital heart block were associated with DCM.²² In an earlier US-based multicenter case series of congenital heart block, the development of cardiomyopathy after birth occurred in 16 fetuses; 4 died and 6 required transplantation.⁶ Antibodies were confirmed in 10 mothers. Recently, data from the Association for European Pediatric Cardiology identified impaired left ventricular function to be associated with mortality in 162 children with congenital heart block, of whom 80% were exposed to anti-SSA/SSB antibodies.²⁵ In our cohort, approximately half the patients with DCM died.

In recent years, EFE has emerged not only as an extension of the cardiac pathogenicity associated with anti-SSA/Ro, but also as a risk factor predicting mortality. Data from the RRNL confirm and extend these earlier observations in that half the patients with EFE died. Jaeggi et al²³ initially demonstrated the poor prognostic significance of EFE in 5 patients. In another study of 13 patients with congenital heart block exposed to anti-SSA/Ro antibodies who developed EFE, 11 (85%) either died or underwent cardiac transplantation.⁷ Emphasizing the injurious effect of anti-SSA/Ro on the endocardium per se, 2 fetuses with EFE absent any conduction abnormalities died, and 1 child with isolated EFE received a transplant.⁸ However, the severity of EFE was not confirmed in a recent case series of 5 children with isolated EFE in which 4 children were alive at 4 years, 3 of whom had normal heart function.⁹ In another study comprising 20 cases of anti-SSA–associated EFE, an 80% survival rate at a median follow-up of 3 years was observed in those treated with intravenous immunoglobulin and steroids at the time of diagnosis.³⁰

A novel finding not previously addressed by any study of cardiac NL was the higher case fatality rate in children born to nonwhite mothers (who make up 25% of the RRNL), an observation that was consistent across each of the racial/ethnic groups. Overall, nonwhites had a significantly higher case fatality rate than whites, although the association of race/ethnicity and mortality was not maintained in multivariable analyses. This was likely due to the observation that white fetuses were at lower risk of hydrops and EFE, variables highly predictive of mortality. One possible explanation is that more extensive cardiac injury occurs in minorities. Candidate neonatal factors could include genetics, a possibility difficult to explore given the low number of affected children in each minority group. To date, genetic studies have been limited to whites.³¹ The absence of racial/ethnic information in any of the published cohorts precludes comparison of these data with other studies. Further studies are needed to validate this association and to determine whether access to medical care accounts for this disparity.

Another previously unreported finding in this study is that associated hepatic/hematologic NL, but not cutaneous NL, is associated with an increased morbidity in live-born children. The liver abnormalities may have been due to hepatic congestion secondary to cardiac failure. However, the finding of a mononuclear infiltrate in several livers evaluated on autopsy supports an organ-specific inflammatory process similar to that proposed for the initial phase of cardiac injury. Cytopenias may add to the overall burden of disease by decreasing oxygenation of tissues, increasing the risk of bleeding and the risk of infection. Moreover, cytopenias may represent increased pathogenicity of the autoantibodies. In addition, mortality was associated with a maternal diagnosis of SLE or SS at the time of pregnancy. This observation was unexpected because it predicted that mothers with known rheumatic disease and the presence of autoantibodies before pregnancy might be more likely to have had optimal surveillance. Although medications such as nonfluorinated steroids, which would more often have been prescribed to patients with established rheumatic disease, might have explained the increased mortality rates, this was not found to be the case. Perhaps maternal illness per se conferred a less favorable in utero environment; it is generally accepted that SLE is associated with premature birth.³² Finally, the association of SLE and/or SS with mortality could possibly have been a reflection of maternal race/ethnicity because nonwhites may be more frequently represented in mothers with SLE.³³,³⁴ However, this was not observed; SLE mothers in this study were equally distributed between whites and nonwhites.

There are several limitations to this study, all of which are inherent in rare diseases. The low numbers of minorities make it difficult to discern why they have a higher case fatality rate. The race/ethnicity of the mother was used as a proxy for the child given the data available because the father's race/ethnicity was not a mandate of the RRNL and not systematically solicited. The low numbers of in utero deaths ¹⁸ limit the statistical power for related analyses. Although the data presented suggest that the probability of in utero death occurring before 20 weeks is zero, this observation may be misleading, because women enrolled in the registry are often totally asymptomatic and unaware of the presence of anti-SSA/Ro antibodies until detection of cardiac NL. Thus, obstetric evaluation beyond routine care may not have occurred before the early to mid second trimester. Moreover, death before 20 weeks may be due to many causes, and without proof of a cardiac disorder, unambiguous attribution to cardiac NL was not possible. The data in this study were largely collected in a retrospective manner, and in some pregnancies, not all of the data were available, which reduced the available sample size for the multivariable analyses. In addition, patients with available data may not be a random sample of the underlying study population, which could potentially lead to biased estimates of relative risk. The exact cause of the cardiac NL death was unknown in 7 cases; however, 5 deaths occurred within 6 months of birth, suggesting that cardiac NL was a contributing cause.

The significant influence of carditis on mortality is potentially biased because its diagnosis is dependent on histological evaluation of tissue. Predictably, biopsies were largely performed in the sickest of cases, the vast majority of whom had associated DCM and/or EFE on echocardiography. Of the 5 children with carditis who lived, 1 required cardiac transplantation. In addition, carditis was seen on several autopsies. The finding of a mononuclear infiltrate supports the hypothesis that an inflammatory process involving more than the conduction system contributes to the increase in mortality.