Etiology and Pathogenesis
Neonatal lupus erythematosus (NLE) is a disease of the developing fetus and neonate defined by characteristic clinical features in the presence of specific maternal autoantibodies. It is considered a model of passively acquired autoimmunity. The transplacental passage of these autoantibodies is necessary but not sufficient to cause the disease. The autoantibodies associated with NLE are directed against a group of small cytoplasmic and nuclear ribonucleoproteins (RNPs): Ro/SSA and La/SSSB. Rarely, cutaneous NLE is associated with isolated anti-U1RNP antibodies. The most common clinical manifestations of NLE are cardiac, dermatologic, and hepatic. The term neonatal lupus erythematosus is misleading as the affected child does not have systemic lupus erythematosus (SLE), and the mother is frequently healthy, without any symptoms of an autoimmune disease. We will first review the autoantigens and autoantibodies associated with NLE and the genetics of NLE. We will then examine the specific clinical features and proposed pathogenesis of these features.
Target Antigens of the Ro/La (RoRNP) System
The major candidate autoantigens in NLE are the Ro and La proteins, which are present in all cells. The first Ro protein identified was a 60-kD polypeptide (Ro60). Isolation and cloning of Ro60 identified a zinc finger and an RNA-binding protein consensus motif. Crystallographic studies demonstrated a ring-shaped protein with two overlapping RNA binding sites that may serve different functions depending on the cellular location of Ro60. One important function of Ro60 is to protect cells from damage from ultraviolet irradiation. In the nucleus, Ro60 likely plays a role in RNA quality control. Ro binds to a class of noncoding RNAs (YRNAs) on the outer surface of the ring. The binding to YRNAs allows for the translocation of Ro60 from the nucleus. This translocation occurs via the zipcode-binding protein ZBP1 and other proteins. During apoptosis YRNAs are required for translocation of Ro60 to the cell surface that may be pivotal to the formation of immune complexes on apoptotic cells and a Toll-like receptor–dependent proinflammatory cascade. Murine studies have suggested that Ro60 may protect against the development of autoantibodies by sequestering defective ribonucleoproteins.
The second RoRNP protein recognized was the 48-kD La protein. Although it may, at least transiently, be associated with Ro60, La does not share antigenic determinants with either Ro60 or Ro52. La is composed of at least two structural domains, each of which contains a distinct antigenic binding site. La is mainly found in the nucleus, and its nuclear localization is dependent on the C-terminus but not the N-terminus (RNP-consensus motif). Similar to Ro60, it can appear on the cell surface during cell stress or apoptosis, and therefore be a direct target for autoantibodies. La facilitates maturation and termination of RNA polymerase III transcripts including transfer RNA (tRNA), and directly binds multiple different RNAs and precursors to small nucleolar RNAs (SnRNAs) involved in ribosome biogenesis. La is required for embryogenesis and normal development.
The other recognized target of the autoimmune response is a 52-kD polypeptide, Ro52. Although two isoforms of Ro52 (52α and 52β) have been recognized, postnatally the 52α is the predominate form (now called Trim21). The 52β is derived from the splicing of exon 4 encoding aa168-245 that includes the leucine zipper. This results in a smaller protein with a predicted molecular weight of 45,000 and lacking a leucine zipper. The 52β transcript is the predominant transcript in the early second trimester when maternal antibodies begin to gain access to the fetus. However, by 18 weeks’ gestation, 52α becomes the predominant or sole transcript. This evidence is consistent with a role for 52β in the development of congenital heart block (CHB), but further proof is required. The remainder of the discussion will be about 52α/Trim21. The full-length protein 52α/Trim21 has three distinct domains: an N-terminal region rich in cysteine/histidine motifs that contains two distinct zinc fingers known as RING finger and B-box; a central region containing two coiled stations of 52α, one being a leucine zipper with potential for intramolecular dimerization, and the other a C-terminal ret finger protein (rpf)-like domain. Ro52 is an E3 ubiquitin ligase that catalyzes the ubiquitination of several proteins, including Ro52 itself. Other important functions of Trim21/Ro52 include regulation of proinflammatory cytokine production ; the production of type 1 interferon and cytokine production via its ubiquitination of interferon regulatory factors (IRFs) including IRF3, IRF5, IRF7, and IRF8 (reviewed ); and regulation of the innate response to intracellular double-stranded DNA.
Calcium (Ca + + ) channels are important in maintaining cardiac rhythm, and antibodies against these ion channels have been hypothesized to be important in CHB. Specifically L-type channels, α 1c and α 1D subunits, and T-type calcium channels, α 1G subunit, have been shown to be present in both adult and fetal conducting tissue. (See the section “ Calcium Channels .”)
The search for the pathogenic antibody specificities that lead to CHB has been attempted by many researchers. The following paragraphs review the history of these efforts. Unfortunately there are important limitations in interpreting many reports; these limitations include small numbers of patients and the failure to use only sera obtained during the pregnancy when comparing autoantibodies from mothers of affected and unaffected children. The problem with this latter approach is that the autoantibody repertoire can change over time.
The first antibodies to be recognized to be associated with the development of NLE were antibodies directed against Ro60 and La48. However, it soon became apparent that not all children born to mothers with anti-Ro60 antibodies developed NLE and that although a sensitive marker for the development of NLE, the presence of anti-Ro60 antibodies had poor specificity to identify at-risk pregnancies. Furthermore, the presence of these antibodies alone could not differentiate fetuses or infants at risk to different manifestations of NLE (such as cutaneous NLE [C-NLE]). In order to increase the specificity of anti-Ro antibodies as a biomarker for NLE, investigators used different immunologic methods to determine the autoantibody repertoire.
Initial studies used the relatively insensitive method of immunodiffusion assay, which was rapidly usurped with the advent of enzyme-linked immunoassays (ELISAs), using either affinity-purified or recombinant proteins, or an RNA immunoprecipitation assay. The latter assay is quite sensitive and can distinguish anti-Ro60 from anti-Ro52 antibodies. However, it is not commercially available and is only mentioned because it was used as a research tool in studies that helped determine the importance of the anti-Ro52 antibody response in the development of CHB. It is important to note which assay is used when analyzing results, as the type of protein used may give different results. For technical reasons all ELISAs for Ro52 use recombinant proteins, whereas assays for anti-La48 or anti-Ro60 antibodies may use either affinity-purified or recombinant proteins. Assays for anti-Ro antibodies that use affinity-purified proteins cannot identify anti-Ro52 antibodies.
The easiest and cheapest assay to screen pregnant women for the risks of delivering a child with NLE is the anti-Ro ELISA. This assay has excellent sensitivity at greater than 99% to identify all fetuses with CHB, but it has poor specificity. Therefore other autoantibodies have been examined to increase the specificity without decreasing the sensitivity. A prospective study of 146 mothers at risk to deliver a child with cardiac NLE showed that the titer of anti-Ro antibodies was important in determining the risk of developing cardiac NLE. Specifically, none of the mothers with low-titer anti-Ro60 antibodies (less than 50 IU; commercially available ELISA with maximum of 100 IU) delivered a child with NLE, whereas the pregnancies of 5% of mothers with anti-Ro60 antibody titers of 50 IU or more were complicated by fetal cardiac NLE. Based on these data, it was suggested that mothers with low levels of anti-Ro antibodies were at a very low risk for delivering a child with cardiac NLE and therefore may require less intensive screening. Subsequently, a validation cohort confirmed these findings.
One of the earliest studies suggested that the presence of both anti-Ro60 and anti-La48 antibodies was sensitive and more specific than the presence of anti-Ro60 antibodies alone for the risk of delivering a child with CHB. Following this observation, a study showed that antibodies directed against a small La polypeptide, named DD, were found only in the sera of mothers of children with NLE and not in sera from mothers of unaffected children. Although this finding was specific for NLE, it had only a 30% sensitivity because many mothers without anti-DD antibodies delivered children with NLE, and anti-DD antibodies were present in only 30% of children with NLE. Subsequent larger studies showed that anti-La48 antibodies were present in up to 50% of sera of mothers of children with CHB and that they had a lower sensitivity than either anti-Ro60 or anti-Ro52 antibodies for the risk of delivering a child with CHB. Therefore testing for the presence of anti-La48 antibodies alone is not a good screening tool to determine at risk pregnancies.
Antibodies to Ro52 are found in 70% to 80% of mothers of children with CHB and although less sensitive than anti-Ro60 antibodies, they are more specific for the development of CHB. Epitope mapping of the anti-Ro52 antibody response revealed an immunodominant region, spanning aa169-291, a region containing the leucine zipper (aa220-232) that was recognized by the majority of the sera from mothers of children with CHB. Examination of the fine specificity of the response to this region showed that the dominant antibody response in mothers of children with CHB was directed against a polypeptide consisting of aa200-239 (p200), whereas a reduced risk of CHB was found when the dominant response was directed against aa176-196 (p176) and aa197-232 (p197). These intriguing data led to replication studies to address the sensitivities and specificities of these antibodies in CHB. The first study showed that although the majority of the maternal sera with anti-Ro52 antibodies had anti-p200 antibodies, the presence of anti-p200 antibodies did not differentiate mothers of children with CHB, C-NLE, or healthy children. Specifically, anti-p200 antibodies were found in 77% of mothers of children with CHB, 80% of mothers of children with C-NLE, and 95% of mothers of unaffected children. The largest study, which was international in scope, found that the mean anti-p200 antibody level in sera from mothers of children with CHB or second-degree heart block (HB) was significantly higher than the mean level in mothers who had children with normal heart rates. Closer examination of one of the cohorts (pregnancies) showed the sera from mothers of children first-, second-, and third-degree HB (detected by fetal echocardiogram during pregnancy) had higher mean anti-p200 antibody levels than the mean anti-p200 antibody level seen in mothers of children without any degree of heart block. However, the first-degree HB resolved by birth.
The most recent study to address this issue used either maternal sera during the pregnancy or cord blood. It was again shown that the frequencies of anti-p200, anti-Ro52, anti-Ro60, and anti-La48 antibodies were not different between affected and unaffected neonates. However, in confirmation of a previous study, mean anti-Ro52 and anti-Ro60 titers were higher in pregnancies with a child with CHB and unaffected siblings of a child with CHB, as compared to those associated with well neonates who never had a sibling with CHB. Maternal anti-p200 antibodies, taken during the pregnancy, were more frequent in mothers who delivered a child with CHB (88%) as compared to mothers who had delivered an unaffected infant (67%). However, the frequency, at 88%, did not differ as compared to that of anti-p200 antibodies in mothers who had previously delivered a child with CHB but had an unaffected child during the study (97%). Overall, anti-p200 antibodies were more specific but less sensitive than anti-Ro60 or anti-Ro52 antibodies for the risk of delivering a child with CHB as compared to delivering an unaffected child.
Translational studies showed that anti-p200 antibodies and anti-Ro52 antibodies both bound the surface of nonpermeabilized apoptotic, but not healthy, human fetal cardiocytes, suggesting that they may be pathogenic in CHB. In vivo rodent models and in vitro culturing systems suggested that anti-p200 antibodies bind neonatal rodent cardiocytes and altered calcium homeostasis.
Anti-L-Type and Anti-T-Type Calcium Channel Antibodies
Anti-Ro/La antibodies were shown to react with, and alter the function of, α 1c and α 1D subunits of the L-type calcium channel and T-type calcium channels. The major target of the anti-T-type Ca + + channel antibodies was an epitope present on an extracellular loop of the channel.
Anti–5-Serotoninergic 5-Hydroxytryptamine A receptor (5-HT 4 R) Antibodies
The search for new cardiac targets led investigators to examine whether anti-Ro52 antibodies cross-react with 5-HT 4 R, which is present in human atrium, including the fetus, and may be important in atrial arrhythmias and fetal cardiac development. It was found that two peptides in the C terminus of Ro52—aa365-382 and aa380-396—shared some similarity with the 5-HT 4 R. Furthermore, the Ro52 365-382 peptide, recognized by sera from patients with SLE, cross-reacted with peptide aa165-185, derived from the second extracellular loop of the 5-HT 4 R. However, when sera from mothers of children with CHB were examined for the presence of anti-5-HT 4 R antibodies, they were found only in a minority of the sera, suggesting that these antibodies are not necessary for the development of the CHB in the majority of patients.
Calreticulin, a calcium-binding protein that is important in cardiac development, was initially thought to be part a Ro protein. However, later studies showed that calreticulin is not a part of the RoRNP although antibodies directed against calreticulin are present in the sera from patients with autoimmune diseases. Elevated anticalreticulin antibody levels have been reported in mothers of children with NLE compared with levels in healthy subjects but not when compared with healthy pregnant women. Currently, it is felt that anticalreticulin antibodies are not important in NLE.
Other antigens present on the fetal heart or placenta have been suggested to be important in the development of, or protection from, atrioventricular (AV) block. Specifically, anti-La, anti-calreticulin, and anti–ERV-3 antibodies can bind to the placenta or placental trophoblast or both. A subset of anti-La antibodies cross-react with laminin (present in the placenta). Binding of these autoantibodies to placental tissue may alter the autoantibody repertoire in the fetal circulation and therefore affect binding to fetal tissue. Laminin, calreticulin, or ERV-3 may be targets for maternal autoantibodies, and direct binding to these fetal cardiac proteins may initiate or potentiate inflammation in the fetal heart. Anti-laminin autoantibodies bind to fetal but not adult heart and cardiac tissue, and ERV-3 and laminin are maximally expressed between 11 and 17 weeks’ gestation.
Two other autoantibodies examined in NLE and CHB were anti-p57 recombinant protein and anti–α-fodrin antibodies. Anti-p57 antibodies were present in one third of sera of mothers of children with NLE, but were almost always associated with anti-Ro antibodies and did not add to sensitivity and specificity testing. Anti–α-fodrin antibodies, present in sera from patients with Sjögren syndrome (SS), were demonstrated to be present in the sera of mothers of children with NLE; however, measurement of these antibodies has not been pursued as they do not add to the sensitivity or specificity of autoantibody testing over routine anti-Ro and anti-La antibody testing.
As early as 1980 it was recognized that human leukocyte antigen (HLA)-DR and HLA-DQ genes were important in the production of anti-Ro/La antibodies. DR3 was shown be associated with anti-Ro52 and anti-La48, but not anti-Ro60 antibodies. This association was present in most ethnic backgrounds and differed between patients with primary SS and SLE. Further assessment showed that high-titer anti-Ro52 antibodies were associated with the DQw1/DQw2 heterozygote state. It is likely that extended haplotypes are more important than single loci in determining production of anti-Ro/La antibodies. In patients of most ethnic backgrounds, high levels of anti-Ro60 and anti-La48 antibodies were associated with all, or at least most, of the DRB1*0301, DQA1*0501, and DQB1*0201 extended haplotypes. If the extended haplotype was not present, then the DQA1*0501 allele was frequently found. Glutamine at position 34 of DQA1 and leucine at position 26 of DQB1 were shown to be associated with anti-Ro60/La antibodies, although the extended haplotype (in linkage disequilibrium) still had the strongest correlation. The genetic control of this response is further complicated by the demonstration that response to individual peptides may be under different genetic control distinct from that found for the response to the whole protein.
In mothers of children with NLE, the HLA profile more closely resembled that present in patients with primary SS than with SLE. In Japanese mothers, the production of anti-Ro/La antibodies was associated with the extended haplotypes DRB1*1101-DQA1*0501-DQB1*0301 and DRB1*08032-DQA1*0103-DQB1*0601, as well as the individual alleles DRB1*1101, DRB1*08032, and DQB1*0301. All of the anti-Ro/La antibody–positive mothers had DRB1 alleles that shared the same amino acid residues at positions 14-31 and 71 of DRB1 and were either homozygous or heterozygous at DQ6 and DQ3 alleles that shared the same amino acid residues at positions 27-36 and 71-77 of the hypervariable regions of DQB1. Individual manifestations of NLE may also be influenced by maternal HLA, as DRB1*1101-DQA1*0501-DQB1*0301 (DR5 haplotype) and individual class II alleles making up this haplotype, including DQA1 alleles with glutamine at position 34 of the first domain, were significantly associated with C-NLE but not CHB. DQB1*0602 (carried on DR2 haplotypes) was associated with CHB but not C-NLE. Studies in other racial or ethnic backgrounds also have supported a role for maternal HLA-DRB and DQ alleles in the predisposition to deliver a child with NLE, although the alleles differed between ethnicities. A few studies have suggested that maternal major HLA class I haplotypes may predispose the delivery of child with CHB. A review of HLA-DRB1, HLA-DQA1, and HLA-DQB1 loci in Japanese and Caucasian children with CHB showed that these alleles were frequently identical to those in their mothers. However, the specific alleles differed between ethnicities.
In the current genome-wide association study (GWAS) era, a GWAS of children of European ancestry showed that the majority of loci with the most significant associations with cardiac NLE were within the HLA region. However, no individual locus previously implicated in autoimmune diseases achieved genome-wide significance. The finding of the association of cardiac NLE and fetal genes in the HLA region was confirmed in a second study (also of European heritage). Using a family-based approach, these latter authors showed that HLA-DRB1*04 and HLA-Cw*05 variants were transmitted significantly more often to affected individuals, and HLA-DRB1*13 and HLA-Cw*06 variants were transmitted significantly less often to affected children.
Genetics of Children with NLE
Despite the strong association of maternal HLA antigens, the association of HLA genes in the offspring is weaker. Two small studies demonstrated that children with NLE tended to have the same DRB, DQA, and DQB genes as their mothers. One report suggested that DR3 (DRB1*03) in the fetus might protect against in utero death, whereas another suggested that DR2 may be protective (associated with maternal DR2-DRB1*02). A large cohort of children with NLE demonstrated that DQB and DRB genes were important in NLE.
Multiple genes outside the major histocompatibility complex (MHC) locus that alter the immune response can influence the onset, susceptibility, and progression of autoimmune diseases. The same is likely true in NLE, particularly with regard to amplifying the pathological cascade to scar (e.g., amplification of apoptosis, uptake of opsonized cardiocytes, and ultimately fibrosis). Support for the importance of non-MHC genes in the offspring is suggested by the observation that children with NLE were more likely to have a tumor necrosis factor-α (TNF-α) polymorphism associated with high TNF-α production. A tumor growth factor-β (TGF-β) gene polymorphism associated with increased TGF-β production was higher in children with CHB than controls although it did not differ between children with CHB and C-NLE.
A multiethnic cohort showed significantly higher case fatality rates in non-whites affected by CHB compared with whites, and whites were at a lower risk of fetal hydrops and endocardial fibroelastosis (EFE). It is not clear whether the difference in outcomes is the result of maternal or fetal genetics, environment or socioeconomic factors. Further studies are required to confirm these findings and to help determine the factors leading to these observations.
Other Genetic Contributions
An observation in 1992 suggested that fathers of children with NLE were more likely to carry C4 null alleles than the controls. The only reported study to show a paternal contribution to NLE demonstrated that there may be an increased paternal (but not maternal) HLA-DRB1*04 transmission to offspring with CHB than to unaffected siblings. An interesting observation was enriched transmission of the maternal risk alleles of TNF-α and C6orf10 (Refs. ) from grandmothers of children with NLE to the mothers of the children with NLE ; the significance is unknown.
All of the genetic studies suggest that maternal, paternal, and fetal HLA genes, and fetal non-HLA genes may be important factors that influence the development of NLE. However, all studies are of relatively small size and cannot explain the observation that monozygotic twins are frequently discordant for NLE. In addition, there is a lack of validation cohorts.
The most clinically significant manifestations of NLE are cardiac, specifically CHB. In most cases, the CHB is isolated, but it may be associated with other cardiac lesions, including ventricular septal defect or patent ductus arteriosus. The first reported case of CHB associated with maternal autoimmune disease (i.e., Mikulicz syndrome or SS) was published in 1901. However, it was not until the 1950s that it was generally recognized that autoantibodies in the mother were associated with NLE. It was another 20 to 30 years until the association with anti-Ro and anti-La antibodies was reported. Children with cardiac NLE can have first-, second-, or third-degree HB. The HB may be isolated or associated with myocarditis and/or EFE. Myocarditis and EFE associated with NLE can be seen in the absence of conduction abnormalities.
It has been estimated that CHB occurs in approximately 1 in 14,000 live births, of which at least 90% of the fetal and neonatal cases, with otherwise structurally normal hearts (isolated CHB), are related to the transplacental passage of maternal autoantibodies. These estimates are based on live births and likely underestimate the true incidence of congenital CHB, as they were performed at a time when fetal echocardiography was not available and more severe fetal cardiac NLE commonly resulted in intrauterine death. Most cases of CHB occur in fetuses of mothers without a diagnosed autoimmune disease. The initial demonstration of autoantibodies in these mothers is during pregnancy or after delivery of a child with NLE. It remains a matter of debate and ongoing research why only some and not all anti-Ro antibody–exposed fetuses acquire cardiac NLE. In prospectively examined pregnancies of mothers with anti-Ro antibodies and a known autoimmune disease, the reported incidence of fetal CHB was approximately 1% to 2%, which increased to approximately 15% for those with a previously affected child with CHB. Recent research suggests that the fetal exposure to high-titer anti-Ro antibodies, rather than the presence of these antibodies, is a requirement for the development of fetal cardiac NLE.
Etiology and Pathology of Cardiac NLE
The characteristic pathology of CHB is an absence or a degeneration of the AV node with replacement by fibrosis, calcification, or fatty tissue ( Fig. 25-1 ). The latter lesion is present in more than 80% of cases of CHB and is associated with maternal anti-Ro/La antibodies, whereas the absence of the AV node without scarring is likely a congenital dysplasia of the conducting system and not associated with maternal autoantibodies. We will focus on autoantibody-associated CHB.
In most autopsy cases of CHB there is not only evidence of degeneration of the AV node with replacement by fibrosis and calcification, but also an inflammatory cell infiltrate. Studies have shown a generalized myocardial inflammation with evidence of deposition immunoglobulin G (IgG), complement, and fibrin deposition on the fetal myocardium. As the fetus is unable to produce IgG, the demonstration of IgG on fetal myocardium implicates maternally derived autoantibodies as an important factor in the fetal pathology. One group was able to elute anti-Ro52 antibodies from the fetal heart, but not unaffected tissue of a fetus that died of CHB, which further implicated maternally derived autoantibodies in the pathogenesis of NLE. Lastly, maternal cells have been found in fetal hearts with CHB. Therefore the transplacental passage of IgG, and possibly maternal lymphocytes, is critical in initiating inflammation leading to CHB. However, it is likely that the fetal immune response to the maternally derived IgG present on the myocardium is important.
The inflammatory infiltrate can be extensive and frequently involves the more distal conducting system and other areas of the heart, including extensive scarring of the atrium and ventricles, and EFE. EFE, in association with an inflammatory infiltrate and maternal autoantibodies, has been found without CHB, suggesting that the primary target may be the myocardium and not the specialized conducting tissue. Immunohistological evaluation of hearts from fetuses that died with CHB has revealed increased apoptosis, macrophages in zones of fibrosis that colocalize with IgG and apoptotic cells, expression of TNF-α and TGF-β messenger RNA (mRNA) in these cells, and extensive collagen deposition in the conducting system. All of these findings are consistent with an inflammatory response initiated by maternal IgG and expanded by the fetal immune response.
The maternal component is the autoantibody that, by binding to its cognate antigen, initiates the first step to injury. It is logical to hypothesize that the target is a cardiac protein containing either a cross-reactive epitope recognized by anti-Ro/La antibodies or an epitope within the Ro protein itself. Ro itself is intracellular, and therefore either the antibody must be internalized before binding or Ro must be present on, or translocated to, the cell surface. However, a direct pathological consequence to cells by inhibiting function would predict an even higher recurrence rate of CHB in subsequent pregnancies than that observed unless the repertoire changes over time. What is accepted is that without this maternal component, CHB would not ensue, and even the presence of high-titer anti-Ro antibodies is not sufficient. Attempts to generate a robust reproducible animal model exploiting the potential pathogenicity of the antibody as an isolated factor have not met with success. Therefore the exact nature of the maternal component remains unresolved.
A role for the fetal immune response is based on the clinical observations that the rate of recurrence of CHB is 15% to 17% in pregnancies that occur after the birth of a child with CHB and that most studies of twins (and triplets) born to a mother with anti-Ro antibodies are discordant for NLE/CHB. Histological examination of the fetal heart has shown that in addition to IgG, there is evidence of IgM deposition (unable to cross the placenta) and a T-cell infiltrate. These autopsy observations support a role for a fetal component, and therefore differences in fetal response to the maternally derived IgG may be part of the explanation for discordant twins and the low recurrence rate.
Recent studies have demonstrated areas of apoptosis of cardiac myocytes. Physiological clearance of apoptotic cardiocytes by resident cardiocytes may be inhibited by maternal autoantibodies, resulting in accumulation of apoptotic cells, inflammation via the production of proinflammatory cytokines, and then scarring. Many of the cardiocytes had a myofibroblast-like phenotype suggesting that these cells were activated and involved in inducing (or perpetuating) the damage. It is difficult to directly implicate the pathogenic effect of autoantibodies in CHB as not only is the injury an infrequent event but the extent of injury varies among fetuses. Accordingly, CHB is likely to represent the sum of several maternal and fetal components.
These in vivo observations are supported by in vitro studies. The link between maternal autoantibodies and tissue injury is supported by the demonstration that apoptosis of fetal cardiocytes leads to expression of Ro/La on the cell surface. It was then shown that cardiac cells can phagocytose autologous apoptotic cardiocytes, and anti-Ro antibodies inhibit the ability of these cells to perform this clearance function. Therefore the potential perturbation of the normal remodeling of the developing heart by the presence of anti-Ro antibodies would allow for the uptake of these apoptotic cells by professional phagocytes (including macrophages). In vitro studies have shown that macrophages, when coincubated with apoptotic cardiocytes bound by Ro/La antibodies, can secrete proinflammatory and fibrosing cytokines, which could then lead to cardiac damage rather than remodeling. Murine studies showed that anti-Ro/La antibodies inhibited the clearance of apoptotic cardiocytes, allowing them to accumulate and thereby leading to an inflammatory reaction and scarring. Studies using human fetal myocytes have shown that the complex of autoantibodies with apoptotic cardiomyocytes leads to transdifferentiation of these cells to a myofibroblast phenotype and the production of the profibrotic cytokine, TGF-β.
The above-mentioned studies have demonstrated that intracellular Ro60, Ro52, and La48 can be translocated to the cell surface and become a target for autoantibodies. An alternative hypothesis has suggested that molecules constitutively present on the surface of cardiac tissue may be the target, and by binding to these antigens the autoantibodies perturb cardiac function. Cardiac calcium channels are present on the cell surface, and based on their importance in conduction, are obvious potential candidate targets. An overview of studies of in vitro and in vivo data of the targets of maternal autoantibodies in the pathogenesis of NLE is in the “Animal Models” section.
It is logical to assume that there are both fetal and maternal factors that are important in the development of NLE and CHB. In fact, there is increasing evidence of a genetic contribution to the risk, disease pattern, and outcome of NLE (as previous). Lastly, the fact that identical twins are more often discordant than concordant for CHB suggests an environmental factor present in utero might be a third factor. Environmental factors would be expected to amplify the injury in susceptible fetuses exposed to the transplacental passage of potentially pathogenic maternal antibodies.
Clinical Approach to CHB, Surveillance of At-Risk Pregnancy, and Treatment of HB
Outcome of cardiac NLE.
Risk factors identified to be significantly associated with perinatal death include fetal hydrops; ventricular rates of fewer than 50 to 55 beats per minute; impaired cardiac function/carditis/EFE/cardiomyopathy; and an earlier gestational age at CHB diagnosis. Most children with CHB require a permanent pacemaker during childhood and thus will carry lifelong risks for pacemaker-related morbidity and mortality. In addition, approximately 5% to 10% of children with CHB and normal cardiac function at birth will develop a dilated cardiomyopathy and/or EFE and cardiac death during childhood. Of note, the experience of five U.S. and Canadian centers demonstrated that the perinatal use of intravenous immunoglobulin (IVIG) and dexamethasone was associated with improved survival of patients affected by EFE and a lower rate of late-onset cardiomyopathy as compared with untreated patients.
Management of cardiac NLE.
While there is agreement that immune-mediated complications represent a spectrum of potentially life-threatening cardiac anomalies there is no consensus on the optimal perinatal management of cardiac NLE, at least in part as a result of the absence of evidence from randomized clinical trials.
Pregnancy management options range from no prenatal therapy; therapy targeting the most severely affected fetuses, e.g., with fetal hydrops and/or EFE; to the routine use of antiinflammatory therapy in fetuses with CHB. The rationale for therapy is based on the evidence that CHB carries a significant mortality risk as the fetus needs to overcome the sudden drop in ventricular rate, the loss of normal atrial systolic contribution to ventricular filling, and maternal antibody-triggered myocardial inflammation and/or fibrosis. Although the standard, echocardiography tends to underappreciate the true extent and severity of the cardiac pathology. The reason to treat a fetus with CHB is not to reestablish normal AV conduction but to improve outcome by mitigating cardiac inflammation and damage, and by increasing fetal cardiac output. The fluorinated steroids dexamethasone and betamethasone, unlike prednisone, are only minimally metabolized by the placenta and pass to the fetus, making them useful in treating fetal inflammation. Dexamethasone is often preferred as it is a single oral daily dose. Repeated transplacental steroid administration has been shown to improve or resolve incomplete fetal AV block, myocardial dysfunction, and pericardial effusions.
Concerns with the use of steroids to the fetus are related to neurological development, growth retardation, and oligohydramnios. These concerns are based mainly on animal studies. In humans, neuroimaging showed that multiple administration of dexamethasone before birth was associated with decreased cortical involution and brain surface area although the clinical significance was not reported. However, follow-up studies showed that by early to mid-childhood fetal exposure to fluorinated steroids had no effect on growth, or neurological or psychological development. Oligohydramnios resolves with either a reduction in the steroid or temporarily holding the medication. It was shown that roughly a quarter of children with CHB who were treated prenatally with dexamethasone were born with mild growth restriction, but there was catch-up postnatal growth. It is important to note that children with CHB, irrespective of prenatal therapy, had low weight until about 2 to 3 years of age and reached reference standards only at age 9 to 11 years of age. A study on neurodevelopment that reviewed that 16% of siblings with (n = 60) and without (n = 54) CHB born to anti-Ro antibody–positive mothers had impaired neurodevelopment at age 13 years. No negative effects on intelligence or neurodevelopment were found in two studies that examined cohorts of preschool- and/or school-age children with CHB who had been prenatally exposed to maternal anti-Ro antibodies with and without prolonged dexamethasone treatment.
Potential maternal glucocorticoid side effects include an increased susceptibility to infections, mood changes, weight gain, fluid retention, arterial hypertension, glucose intolerance, insomnia, hirsutism, striae, impaired wound healing, stomach irritation, headache, and, rarely, psychosis. A review of adverse effects associated with the use of antenatal steroids included oligohydramnios in 12% of pregnancies, maternal hypertension in 5%, insulin-dependent diabetes in 2%, and insomnia or mood changes in 7%. However, there was no control group of untreated CHB pregnancies.
The currently used dose of dexamethasone to treat cardiac NLE results in a daily fetal dexamethasone exposure that does not exceed 0.05 mg/kg (maximal 8 mg/day maternal dosage and a cord-to-maternal drug ratio of 30%). This treatment may be maintained over many weeks with tapering at the time of delivery. Significantly higher steroid dosages over a shorter period of time were used in animal studies, for prenatal human pulmonary maturation, and for the treatment of respiratory distress syndrome in premature newborns.
β 1 -Adrenergic stimulation with the bronchodilators salbutamol and terbutaline may be used to increase the fetal cardiac output by a change in heart rate and a decrease in systemic vascular resistance. The indications of chronic transplacental fetal β-adrenergic therapy are a fetal heart rate fewer than 50 beats per minute and/or significantly reduced cardiac contractility. When given orally to the mother, salbutamol (10 mg every 8 hours to a maximum dosage of 40 mg/day) or terbutaline (2.5 to 7.5 mg every 4 to 6 hours to a maximum dosage of 30 mg/day) typically increased the ventricular rate by 5 to 10 beats per minute. Side effects include tremor, palpitations, and sweating, which usually improve or resolve with continuation of therapy. More serious maternal adverse events or intolerable symptoms that required a change in drug treatment have not been reported with oral β-adrenergic stimulation. Nevertheless, β-agonists should be used cautiously in mothers with diabetes, hypertension, hyperthyroidism, and a history of seizures or tachyarrhythmias.
The last modality of therapy advocated by some investigators is the repeated use of IVIG as an adjunct to transplacental steroid treatment, particularly if there is evidence of myocardial inflammation and fibrosis. The prognosis for fetuses and infants with diffuse EFE is poor, with death or need for cardiac transplant in 85% of infants. In a retrospective review from four centers of 20 fetuses with EFE, perinatal use of both IVIG and dexamethasone was associated with an overall survival rate of 80%, all with normal systolic function. It is not clear whether it was the combination of both therapies or either one individually that was associated with this outcome as IVIG alone was not tried, whereas dexamethasone alone has been associated with improved fetal outcome. It has been proposed that IVIG blocks either the transplacental passage of maternal autoantibody and/or the binding of these autoantibodies to the fetal heart. This would result in decreased macrophages or T cells with decreased cytokine production and complement activation, leading to a reduction in myocardial damage. Immunomodulatory treatment with IVIG is usually well tolerated, but both the mother and fetus carry potential risks associated with the exposure to blood products. The benefit of the use of IVIG is not proven.
The guidelines used in Toronto include the routine initiation of oral dexamethasone therapy at the time of diagnosis of CHB (after confirmation of the presence of maternal anti-Ro antibodies) and maintained for the duration of the pregnancy. If the average fetal heart rate declines below 50 to 55 beats per minute or if the cardiac function is reduced, a β-adrenergic agent is added. Pregnancies are monitored weekly by a team that includes obstetrics, fetal cardiology, and rheumatology. Delivery, usually by cesarean section, is arranged between 37 and 38 weeks’ gestation with transfer to the intensive care unit for neonatal care. A study from two Canadian centers using these guidelines showed significant improvement in outcomes of fetal CHB as compared to pregnancies followed in these centers prior to the routine use (beginning in 1997) of the guidelines. Prior to 1997, pregnancies with isolated CHB were typically followed by serial echocardiography without transplacental fetal therapy. Live birth and 1-year survival of fetuses diagnosed between 1990 and 1997 were 80% and 47%, respectively, and improved to 95% for both between 1997 and 2003. Immune-mediated conditions that caused postnatal death or required cardiac transplantation were only observed in fetal survivors of untreated pregnancies. These findings suggest that prolonged administration of dexamethasone rendered a fetus with isolated CHB less likely to develop myocarditis, cardiomyopathy, and/or hydrops fetalis, thus improving overall outcome.
The treatment guidelines used in Toronto were adjusted in recent years to minimize the risk of oligohydramnios. Dexamethasone is started at 8 mg/day then reduced to 4 mg/day after 2 weeks, and then to 2 mg/day at 28 to 30 weeks’ gestation in uncomplicated cases. Maternal IVIG (70 g every 2 to 3 weeks) is added if significant EFE is detected. The experience of five U.S. and Canadian centers, including Toronto, that used the contemporary routine prenatal treatment approach demonstrated survival rates to birth and to 10 years of age were 98% and 91%, respectively, for 88 prenatally treated CHB cases. Only 1.1% developed a late-onset dilated cardiomyopathy as compared to historic rates of 10% to 15%. These survival rates compare favorably to the lower survival rates in studies that used no prenatal treatment strategies or targeted ones.
However, many centers do not advocate the use of any maternal therapy, fluorinated steroids, and/or IVIG for fetuses with CHB. The rationale for nontreatment includes (a) a significant proportion of fetuses with uncomplicated isolated CHB will survive without antiinflammatory treatment; (b) the potential risks to the fetus and mother, as described above; and (c) the lack of prospective randomized studies that indicate the beneficial effect of treatment.
Prevention of CHB.
Most cases of cardiac NLE are diagnosed in utero in pregnant women without a history of an autoimmune disease. Unfortunately, at this point, the AV node is usually already irreversibly damaged and replaced by fibrotic scar tissue; antiinflammatory treatment may not restore the fetal AV conduction. However, CHB may be preventable if the disease process is either abolished with preventive therapy (primary prevention) or is detected and treated at an early stage of AV nodal disease (secondary prevention). Strategies used for primary prevention CHB have included prophylactic therapy with plasmapheresis, corticosteroids, and/or IVIG. Most researchers do not advocate primary prevention.
Two large, controlled trials of anti-Ro antibody–positive mothers who had previously delivered a child with NLE failed to show a benefit of primary prevention with IVIG. In both studies mothers received IVIG at a dosage of 400 mg/kg/dose every 3 weeks from 12 to 24 weeks’ gestation. In one study 3 of 15 fetuses developed CHB as compared to 1 of 7 mothers who refused IVIG, whereas in the other study 3 of 19 fetuses developed CHB (all mothers received IVIG). IVIG was well tolerated, but there was no benefit as compared to the predicted 15% to 17% rate of CHB (historic controls). Primary prevention with fluorinated steroids has not been prospectively examined in large numbers of patients.
Most recently interest has focused on the potential beneficial effects of maternal use of antimalarials in decreasing the risk of delivering a child with CHB. The rationale for examining the use of these medications is based on the demonstration that (1) Ro60 can bind to single-stranded RNA (ssRNA), leading to Ro60-associated ssRNA complexes that can activate signaling through Toll-like receptor 7 (TLR7), and activation of TLR7 can lead to fibrosis ; and (2) that one of the mechanisms of action of hydroxychloroquine (HCQ) is to inhibit activation of intracellular TLRs, including TLR7. The hypothesis therefore was that Ro60-associated ssRNA complexes lead to abnormal TLR signaling and that maternal use of antimalarial medications, drugs that are known to cross the placenta, would decrease or prevent this signaling and therefore decrease the risk of scarring and CHB. A case-control multicenter study showed that 14% of mothers of fetuses with cardiac involvement used HCQ throughout the pregnancy as compared to 37% of mothers of fetuses without cardiac involvement, thereby giving an odds ratio of 0.46 (95% confidence interval [CI] 0.18-1.18; P = 0.10) that HCQ was protective. A second international study examining infants born to mothers with a history of delivering a child with CHB, showed that maternal use of HCQ throughout pregnancy may protect subsequent pregnancies from the development of CHB (odds ratio 0.23; 95% CI 0.06-0.92; P = 0.037). The most recent study using Bayesian analysis was a single-center inception cohort study that showed that the probability that antimalarial use was associated with a decreased of the development of CHB was greater than 99%.
Secondary CHB prevention refers to treatment early during the process prior to CHB. It requires that antibody-mediated CHB is the result of progression from first- and/or second-degree AV block to complete AV block rather than the sudden appearance of CHB directly from normal sinus rhythm. In addition, when the abnormality is detected, it must be reversible with treatment. The current evidence suggests that when incomplete HB is detected, treatment with steroids may prevent progression or lead to reversal. However, the diagnosis of incomplete AV block is rare, with fewer than 50 reported cases in the medical literature. Importantly, progression from first- to third-degree AV block has not been documented in the literature, including in patients who were entered in studies with weekly fetal echocardiograms and within days of a fetal echocardiogram with normal rhythm and no evidence of inflammation. The reason for these observations may be that CHB may develop too fast (within 1 day), progression is not staged, early stages are overlooked, or early treatment of the few suspected cases with first-degree AV block has prevented progression. If the earliest manifestation of evolving AV nodal disease is clinically silent prolongation in AV conduction (first-degree HB), one would predict that a fetal electrocardiogram (ECG), fetal magnetocardiogram (MCG), or fetal echocardiography would be able to demonstrate a first-degree block prior to the emergence of sustained fetal bradycardia. With this concept in mind, many centers have been offering serial echocardiographic assessment of the fetal AV conduction during the period of highest risk of HB (weeks 18 to 24 of gestation) as a strategy to prevent CHB.
Although likely the most sensitive tools to detect early conduction abnormalities, for various reasons, fetal ECG and fetal MCG are available only in very few centers and are predominantly used as research tools. Alternatively, M-mode, pulse-wave, and tissue Doppler echocardiographic techniques have been used to study the chronology of atrial and ventricular electrical events indirectly by their respective mechanical consequences. In particular, pulsed Doppler techniques have been found clinically useful to measure mechanical AV time intervals. It is, however, important to realize that reference values differ among modalities, and AV times physiologically prolong with gestational age. There have been four large prospective clinical studies that examined the feasibility and utility of serial fetal echocardiographic assessment of antibody-exposed fetuses to detect early, potentially treatable AV conduction disease. While studies demonstrated the feasibility of Doppler-derived AV time measurements, there remains significant controversy about the optimal fetal cardiac surveillance strategy, the clinical relevance of AV prolongation as predictor of emerging CHB, and the indications for in utero treatment. An important observation was that most serially followed fetuses had completely normal AV time intervals prior to the detection of CHB, although other findings suggestive of cardiac NLE (tricuspid regurgitation, atrial echodensity, pericardial effusion) were present in some of the fetuses. These observations suggest that CHB likely results from a rapidly evolving inflammatory process that may not necessarily manifest as a brief episode of incomplete AV block.
Currently there is significant controversy regarding the clinical relevance of AV prolongation, as defined by greater than 2 to 3 standard deviations for gestational age-matched controls, because in most studies this was a relatively frequent observation. In the earliest study, AV intervals were studied in 24 anti-Ro/La antibody–positive women between 18 and 24 weeks’ gestation in comparison with 284 pregnant women as controls. The authors found 8 of 24, or 33%, of the autoantibody-associated pregnancies were complicated by statistical prolongation of the AV times. It is important to note, however, that the AV prolongation was transient in all but two of eight pregnancies: one fetus progressed from a borderline prolonged AV interval (140 ms) to CHB within 6 days; the other fetus had second-degree block that reversed to first degree after dexamethasone therapy. The three more recent prospective studies reported AV prolongation with a greater than 2 z-scores in 9% and greater than 3 z-scores in 2% of serially assessed antibody-exposed fetuses. AV prolongation up to 6 z-scores resolved in all cases regardless of maternal steroid use. However, to date, in the few cases with first-degree AV block with AV prolongation greater than 90 ms (6 z-scores), all persisted despite steroids.
In summary, in pregnancies exposed to anti-Ro antibodies, the fetal echocardiographic Doppler finding of the mechanical heart rate (HR) interval prolongation may serve as a surrogate marker of subclinical disease. The two critical issues raised are (1) the clinical significance of a prolonged AV interval and (2) the biological implication with regard to tissue injury. An isolated prolongation of the AV interval may be transient (related to vagal tone or medication use) or reversible injury, or it may be permanent or progress to more marked delay as a result of physical injury to the specialized electrical pathway (e.g., because of inflammation or scarring). It may be that AV prolongation represents a variant of normal, and only in retrospect does it have clinical significance if it progresses to more advanced block. Given the identification of cases with new CHB and severe cardiomyopathy within 1 week of a normal echocardiogram, with the most frequent detection time being between 20 and 24 weeks’ gestation, it would seem appropriate to perform weekly or even shorter monitoring intervals between 18 and 24 weeks. The goal of this monitoring would be to identify a biomarker of reversible injury, such as AV interval prolongation, and myocardial echodensity. The task of identifying a biomarker of reversibility is particularly important as dexamethasone and betamethasone may have both maternal and fetal risks. Based on available evidence, the Toronto group currently limits the use of preventive transplacental treatment to those fetuses that either present with AV intervals that have greater than 6 z-scores, with second-degree AV block in association with persistent AV prolongation, or with additional signs of suggestive of cardiac inflammation and tissue damage, such as EFE, effusions, and sudden onset of valvular regurgitation. The potential drawback of this less aggressive approach is that if subtle prolongation of the AV interval does represent tissue injury, untreated CHB might evolve too rapidly in some fetuses and will go unnoticed by weekly screening.
Risk of delivery of a child with CHB.
While it is clear that children with NLE are born to mother with anti-Ro/La antibodies, it is important to estimate the risk of delivering a child with NLE from mothers with a rheumatic disease and from mothers who have previously delivered a child with NLE. Large series of pregnancies in women with SLE who had anti-Ro/La antibodies have suggested that the risk of delivering a child with NLE varied between 1% and 10%. The most recent, and larger, prospective studies have suggested that the risk of delivering a child with NLE in mothers with anti-Ro and/or anti-La antibodies was 1% to 2%. Large, prospective studies suggest that the risk of delivery of a child with CHB following the delivery of a child with CHB or C-NLE is approximately 15% to 17%.
Cutaneous Neonatal Lupus Erythematosus
The rash of C-NLE was first reported in 1954 in a child born to a mother with an autoimmune disease. It was not until 1981 that the association of C-NLE and maternal anti-Ro antibodies was described. Similar to mothers of children with CHB, mothers of children with C-NLE are usually clinically healthy despite the presence of circulating anti-Ro/La antibodies. It is likely that a rash is present in 15% to 25% of children with NLE, although it is difficult to determine the true percentage because the rash can be easily missed and spontaneously resolves. There is a female predominance with a female-to-male ratio of 2 : 1 to 3 : 1. The reason for the increased female incidence may be related to the fact that estrogens enhance surface expression of Ro and La proteins on keratinocytes. It should be noted, however, that in the Research Registry for Neonatal Lupus (RRNL) the female predominance was not as pronounced—55% were female and 45% were male.
The photosensitive nature of the cutaneous lesions has led investigators to examine the effect of ultraviolet irradiation on keratinocyte surface expression of Ro/La proteins. Clinically, the rash more closely resembles the lesions of subacute cutaneous lupus erythematosus (SCLE) than the malar rash of SLE ( Figs. 25-2, 25-3, and 25-4 ). Most patients with SCLE have anti-Ro60, anti-Ro52, and/or anti-La antibodies; the same antibodies are associated with NLE. The rash of C-NLE is less frequently malar, and the lesions are not indurated ; follicular plugging or dermal atrophy, typical of discoid lupus erythematosus, is rare. The rash may mimic Langerhans cell histiocytosis, resemble capillary malformations, present as targetoid lesions or bullous lesions. The face and scalp are the most commonly involved areas, but the rash may occur at any site, including the palms and soles ( Fig. 25-5 ). Commonly, it develops around the eyes in a raccoonlike distribution ( Fig. 25-6 ). The rash tends to consist of discrete, round, or elliptical plaques with a fine scale that has central clearing, and it tends to be papulosquamous (similar annular erythema). An infant may have one or both of these rashes. In North America, papulosquamous lesions are most common, whereas in Japan, annular erythema is more common. Bullous lesions may be seen, especially on the soles of the feet. The rash may resemble cutis marmorata telangiectatica congenita, capillary malformation, bullous impetigo, primary herpes simplex infection, and erythema multiforme.
The rash may be present at birth but more commonly develops within the first few weeks of life. The most common age of appearance is 6 weeks, but it may not be recognized until as late as 12 weeks. New lesions may appear for several months, but they rarely develop beyond 6 months, consistent with the disappearance of maternal antibodies from the infant’s circulation. The mean duration of the rash is 17 weeks. The lesions may be induced or exacerbated by sun exposure, but as the rash may be present at birth and may occur on the soles of the feet and diaper area, it is evident that sun exposure is not absolutely required for development of the rash. A rash that appears after phototherapy for neonatal jaundice has been reported, but it is uncommon.
The lesions of C-NLE are transient and usually resolve without scarring, although some mild atrophy may result. Cutaneous telangiectasias, beginning at 6 to 12 months, occur in approximately 10% of affected infants. The telangiectasias may occur in areas that were not initially involved and therefore are not just the result of healing of the rash. The most common area for telangiectasias is the temple near the hairline, an area not usually affected by the acute lesion ( Fig. 25-7 ). Telangiectasias tend to be bilateral. This lesion may be the presenting feature of C-NLE, although it is not clear in these instances whether the initial rash was missed or the telangiectasias had occurred de novo . There have been reports of telangiectasias with atrophy persisting into adolescence.
Most mothers of infants with C-NLE are anti-Ro antibody positive, often in combination with anti-La antibodies. The percentage of mothers with elevated anti-Ro antibody levels depends on the assay used. A few cases of C-NLE have been reported in association with antibodies to U1RNP in the absence of anti-Ro/La antibodies. One infant with C-NLE had anti-U1RNP but not anti-Ro or anti-La antibodies when tested by ELISA but identified by immunoblot.
Biopsies of the lesions of C-NLE demonstrate the typical histopathology of SCLE, which includes epidermal basal cell damage, a mild mononuclear cell dermal infiltrate, vacuolation of the basal layer, and epidermal colloid bodies. IgG, IgM, and complement are deposited usually at the dermoepidermal junction. There have been rare reports of neutrophilic as well as lymphocytic infiltrates, as well as nuclear dust and extravasated red blood cells.
Currently Recognized as Important
Anti-Ro60: Affinity-purified *
* Currently the most sensitive assay to detect all at-risk pregnancies for CHB but has low specificityor recombinant protein ELISA.
Anti-R052: Recombinant protein ELISA
Anti-La48: Affinity-purified or recombinant protein ELISA
† Associated with cutaneous NLE only
Possible Clinical Significance but Require Further Validation
Anti-Ro p200 peptide
Anti-T-type calcium channels (α 1G subunit)
Anti-L-type calcium channels (α 1D -Cav1.3 subunit)
Anti-La DD peptide
Antibodies Not Validated Following Initial Studies
Anti-serotoninergic 5-hydroxytryptamine (5-HT 4 ) receptor
|POLYCYCLIC LESION||ANNULAR ERYTHEMA|
|Urticaria||Erythema annulare centrifugum|
|Erythema marginatum||Familial annular erythema|
|Tinea corporis||Infantile epidermodysplastic erythema|
|Seborrheic dermatitis||Infection with Pityrosporum|
|Ichthyosiform genodermatosis||Annular erythema of infancy|
|Erythema gyratum atrophicans|
The usual approach to management of C-NLE is reassurance and observation as the natural history of the skin lesions is spontaneous resolution without scarring; therefore aggressive treatment is not indicated. However, topical application of a glucocorticoid cream may hasten the resolution of the lesions and be used for cosmetic reasons, although it is possible, but not proven, that steroid use increases the risk of developing telangiectasias. Some physicians have advocated the use of topical calcineurin inhibitors, but as in the instances of topical steroid use, there is no evidence that they alter the natural history or risk of telangiectasias. Telangiectasias can be treated with pulse dye laser therapy, although they may also spontaneously improve.
Hepatic dysfunction in NLE is characterized by abnormal levels of liver enzymes and hepatomegaly that were initially ascribed to congestive heart failure, intrauterine fetal hydrops, disseminated intravascular coagulation, or total parenteral nutrition. A large, unselected series demonstrated that liver involvement occurred in approximately 25% of all infants with NLE. Liver disease can present as an isolated disorder or in association with other manifestations of NLE. Usually, patients have mild hepatomegaly, with or without splenomegaly, and cholestasis with mildly to moderately elevated transaminase. Although a liver biopsy is usually not clinically indicated, histological abnormalities are similar to idiopathic neonatal giant cell hepatitis with mild bile duct obstruction, occasional giant cell transformation, and mild portal fibrosis ( Fig. 25-8 ). It is possible in this regard that idiopathic neonatal hepatitis may be another manifestation of NLE. Liver biopsies should be reserved for infants with clinical evidence of severe dysfunction or with persistent, moderate-to-severe dysfunction.