Author + information
- Received November 11, 2013
- Revision received February 5, 2014
- Accepted February 25, 2014
- Published online August 1, 2014.
- Roland Weber, MD∗,
- Paul Kantor, MD†,
- David Chitayat, MD‡,
- Mark K. Friedberg, MD∗,
- Fraser Golding, MD∗,
- Luc Mertens, MD, PhD∗,
- Lynne E. Nield, MD∗,
- Greg Ryan, MB§,
- Mike Seed, MD∗,
- Shi-Joon Yoo, MD∗,
- Cedric Manlhiot, BSc∗ and
- Edgar Jaeggi, MD∗∗ ()
- ∗Fetal Cardiac Program, Labatt Family Heart Center, The Hospital for Sick Children, Toronto, Ontario, Canada
- †Heart Failure Program, Labatt Family Heart Center, The Hospital for Sick Children, Toronto, Ontario, Canada
- ‡Prenatal Diagnosis and Medical Genetics Programs, Mount Sinai Hospital; University of Toronto, Toronto, Ontario, Canada
- §Fetal Medicine Unit, Mount Sinai Hospital; University of Toronto, Toronto, Ontario, Canada
- ↵∗Reprint requests and correspondence:
Dr. Edgar Jaeggi, Fetal Cardiac Program, Labatt Family Heart Centre, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada.
Objectives The purpose of this study was to determine the phenotypic presentation, causes, and outcome of fetal cardiomyopathy (CM) and to identify early predictors of outcome.
Background Although prenatal diagnosis is possible, there is a paucity of information about fetal CM.
Methods This was a retrospective review of 61 consecutive fetal cases with a diagnosis of CM at a single center between 2000 and 2012.
Results Nonhypertrophic CM (NHCM) was diagnosed in 40 and hypertrophic CM (HCM) in 21 fetuses at 24.7 ± 5.7 gestational weeks. Etiologies included familial (13%), inflammatory (15%), and genetic-metabolic (28%) disorders, whereas 44% were idiopathic. The pregnancy was terminated in 13 of 61 cases (21%). Transplantation-free survival from diagnosis to 1 month and 1 year of life for actively managed patients was better in those with NHCM (n = 31; 58% and 58%, respectively) compared with those with HCM (n = 17; 35% and 18%, respectively; hazard ratio [HR]: 0.44; 95% confidence interval [CI]: 0.12 to 0.72; p = 0.007). Baseline echocardiographic variables associated with mortality in actively managed patients included ventricular septal thickness (HR: 1.21 per z-score increment; 95% CI: 1.07 to 1.36; p = 0.002), cardiothoracic area ratio (HR: 1.06 per percent increment; 95% CI: 1.02 to 1.10; p = 0.006), ≥3 abnormal diastolic Doppler flow indexes (HR: 1.44; 95% CI: 1.07 to 1.95; p = 0.02), gestational age at CM diagnosis (HR: 0.91 per week increment; 95% CI: 0.83 to 0.99; p = 0.03), and, for fetuses in sinus rhythm, a lower cardiovascular profile score (HR: 1.45 per point decrease; 95% CI: 1.16 to 1.79; p = 0.001).
Conclusions Fetal CM originates from a broad spectrum of etiologies and is associated with substantial mortality. Early echocardiographic findings appear useful in predicting adverse perinatal outcomes.
Cardiomyopathies (CMs) encompass a spectrum of heart muscle disorders that affect cardiac filling, contraction, or both, in the absence of correctible anatomic and/or hemodynamic abnormalities (1). Most children present with a dilated or hypertrophic phenotype (2–4) without an identifiable genetic, familial, infectious, or metabolic cause (5,6). CM is the most common indication for cardiac transplantation in infants (1,7). The condition is rarely diagnosed prenatally, and there is little knowledge of the disease spectrum and outcome when detected prenatally. In a study that predated this research, Pedra et al. (8) reported 55 fetuses with a hypertrophic (n = 33) or dilated (n = 22) phenotype, diagnosed at The Hospital for Sick Children in Toronto, Ontario, Canada between 1990 and 1999. The hypertrophic phenotype was predominantly (76%) related to maternal diabetes and twin-twin transfusion syndrome. Cardiac pathology secondary to these conditions is often reversible (9,10), with a substantially better long-term prognosis compared with primary CM. Accordingly, the purpose of this single-center cohort study was to assess the disease pattern and outcome of disorders in which the primary pathology is the fetal myocardium and to determine epidemiological and hemodynamic markers associated with adverse outcomes.
The Research Ethics Board of the Hospital for Sick Children approved this retrospective study.
The Hospital for Sick Children is the exclusive tertiary perinatal cardiac care provider for a population with 80,000 live births per year. Of 8,506 pregnancy referrals to the Fetal Cardiac Program between January 2000 and June 2012, 2,426 were affected by fetal heart disease. These included 61 fetuses (2.5%) with myocardial disease unrelated to structural heart disease, tachyarrhythmia, abnormal cardiac loading, ischemia, or maternal diabetes. After echocardiographic diagnosis of fetal CM, a comprehensive evaluation by the High-Risk Pregnancy Program and, after birth, by the Heart Failure Program was arranged. The diagnostic workup included genetic counseling, virology (polymerase chain reaction, TORCH [toxoplasmosis, other infections, rubella, cytomegalovirus, and herpes simplex virus] serology), metabolic screening, karyotype, pan-cardiomyopathy gene-panel screening, microarray, and, if applicable, invasive or post-mortem specialized testing. Echocardiograms were offered to first-degree relatives in whom familial CM was a possibility.
Patient information was systematically reviewed including demographic factors, tests, and outcomes to December 2012. All patients underwent detailed 2-dimensional, M-mode, and Doppler echocardiography to determine cardiovascular anatomy and function. Ventricular dimensions were obtained in the cardiac 4-chamber view from M-mode and 2-dimensional recordings. Offline measurements were made by a single investigator (E.J.) and the mean of 3 consecutive measurements were compared with institutional reference data (11). Findings considered abnormal included ventricular shortening fraction <28%, left ventricular (LV) isovolumic relaxation time >43 ms, LV myocardial performance index (MPI) >0.48, ventricular end-diastolic dimensions >2 z-scores, cardiothoracic area ratio (CTR) >35%, and more than mild valvar regurgitation (11–14). For fetuses with normal sinus rhythm, the severity of heart failure was quantified using the cardiovascular profile score (CVPS) (Table 1) with 2 modifications to the original scoring system (replacement of “skin edema” with “fetal hydrops” and elimination of “tricuspid valve dP/dt,” which had not been routinely measured, as 2-point criteria) (12).
Interrater agreement was assessed on 10 randomly selected fetal studies for the following variables: CVPS, CTR, systolic and diastolic ventricular diameters, ventricular septal wall thickness, isovolumic relaxation time, and MPI. Interrater bias was nonstatistically significant for all parameters. Interrater correlation was statistically significant (r > 0.8, p < 0.01) for all parameters with the exception of LV MPI (r = 0.55, p > 0.05). These data confirm that interrater reproducibility was excellent for most parameters.
CM was divided into 2 anatomic phenotypes, depending on the presence or absence of myocardial hypertrophy at final assessment. This approach was selected to allow for phenotypic crossover that can occur during fetal life. Hypertrophic cardiomyopathy (HCM) demonstrated inappropriate ventricular hypertrophy and was defined by diastolic ventricular wall thickness >2 z-scores at the last echocardiogram or at autopsy (15,16). Nonhypertrophic cardiomyopathy (NHCM) was defined by cardiac dysfunction in the absence of myocardial hypertrophy at any stage and included dilated and nondilated phenotypes (17,18). Dilated NHCM was defined by ventricular enlargement >2 z-scores of 1 or both ventricles. LV noncompaction was diagnosed on the basis of prominent trabeculations and multiple deep recesses at the ventricular apex (19,20). Fibrosis and calcification were identified as areas of persistently echogenic endomyocardium (8,16,21). Fetal hydrops was defined by ≥2 sites of fluid collections. Diastolic dysfunction was defined by ≥3 of 5 abnormal echocardiographic markers: monophasic tricuspid flow; monophasic mitral flow; pulmonary venous flow reversal during atrial systole; absent/reversed ductus venosus flow during atrial systole; and umbilical vein pulsations. Umbilical vein pulsation was the only variable used to diagnose diastolic dysfunction in cases of nonsinus rhythm (8).
Each condition was also classified by etiology as follows: 1) genetic-metabolic related to chromosomal disorder, inborn error of metabolism, or first-degree family members with CM; 2) inflammatory secondary to post-infectious or antibody-mediated myocardial damage; and 3) idiopathic if no etiology was established.
Descriptive statistics for both the entire cohort and specific subgroups are expressed as mean ± SD for continuous variables and frequencies for categorical variables. When available, anatomic measures were converted to gestational age or body surface area–based z-score. Statistical significance of the difference in patient characteristics and baseline echocardiographic findings between NHCM and HCM (Table 2) and between survivors and nonsurvivors (Table 3) was assessed using the Fisher exact test and Student t test assuming unequal variance between groups. In addition, in assessing differences between survivors and nonsurvivors (Table 3), freedom from the composite endpoints of death or transplantation after birth was modeled using Kaplan-Meier analysis with Cox proportional regression analysis to determine the significance of the between-group differences (PROC PHREG on the SAS system). Findings from this analysis are reported as hazard ratio (HR) with 95% confidence interval (CI). From the univariable regression models described, multivariable models were created. In an initial step, all potentially associated factors were assessed individually. Those associations with a univariable p value <0.20 were then entered into a multivariable regression analysis with backward selection of variables to obtain a final model. This was repeated for the subgroup of patients in sinus rhythm separately. Underlying assumptions of all Cox proportional hazard regression models were checked and satisfactory. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina).
Over a period of 12 years, we diagnosed 61 fetuses with CM as the primary abnormality, suggesting an incidence of 6.2 (95% CI: 3.8 to 8.5) per 100,000. Referral indications at a median of 23 (range, 15 to 41) gestational weeks included suspected cardiac (n = 35) or noncardiac (n = 21) anomalies on the basis of abnormal obstetrical ultrasound examinations or first-degree family members with CM (n = 5).
The results of the baseline fetal cardiac examination are summarized in Table 2. Similar findings between HCM (n = 21) and NHCM (n = 40) groups included age at diagnosis, prevalence of fetal hydrops, diastolic function indexes, and CVPS. Fetuses with NHCM had lower ventricular shortening fractions and larger end-diastolic RV dimensions compared with those with HCM. In 17 of 21 (81%) HCM cases, myocardial hypertrophy was present on the first echocardiogram. Exceptions included 1 fetus each with Noonan syndrome, Hurler syndrome, congenital glycosylation disorder, and familial HCM, all of whom initially presented with marked isolated diastolic dysfunction followed by development of myocardial hypertrophy before (n = 2) or early after (n = 2) birth.
Etiologies, spectrum of abnormalities, and outcome of NHCM and HCM are shown in Tables 4 and 5⇓, respectively. In 27 of 61 (44%) cases, no disease mechanism was established. Lack of a disease etiology was associated with increased odds of in utero demise (n = 17) or pregnancy termination (n = 13) (18 of 27 [67%] mortality in patients with no CM etiology vs. 12 of 34 (35%) mortality with CM etiology; OR: 3.7; CI: 1.1 to 12.4; p = 0.02). Genetic-metabolic, post-infectious, and idiopathic disorders were often associated with noncardiac abnormalities, most commonly of the brain (e.g., global developmental delay, autism, sensory-neural hearing loss, brain calcifications).
Among the 40 fetuses with the nonhypertrophic phenotype, 13 of 34 (38%) without a rhythm disorder and 2 of 6 (33%) with persistent fetal bradycardia due to complete heart block (n = 5) or atrial standstill (n = 1), respectively, had increased ventricular dimensions on the baseline echocardiogram. Functional abnormalities included reduced shortening fraction of 1 (n = 5) or both (n = 22) ventricles, ≥1 diastolic abnormality (n = 27), and more than mild tricuspid regurgitation (n = 17), mitral (n = 1) regurgitation, or both (n = 9). Other cardiac diagnoses included endocardial fibroelastosis (n = 12), myocardial calcifications (n = 2), first- (n = 2) or third- (n = 5) degree heart block, and LV noncompaction (n = 8). Etiologies were highly variable and included rare genetic-metabolic, familial, and inflammatory disorders (Table 4).
The atrial myocardium was the primary disease site in 2 (5%) fetuses. The first fetus had atrial standstill, bradycardia, and fetal hydrops at 26 weeks’ gestation. Post-mortem examination at 30 weeks’ gestation revealed complete loss of atrial myocytes secondary to a novel sarcolipin mutation (22). The second fetus had massive right atrial enlargement in the absence of other structural cardiovascular abnormalities. The perinatal outcome was complicated by pericardial effusion, atrial flutter, and respiratory distress due to mechanical airway obstruction from the enlarged atrium. The right atrial wall was partially excised in the neonatal period. Atrial histology showed focal complete loss of atrial myocytes. This now 4-year-old child is clinically well without signs of ventricular involvement.
The ventricular myocardium was predominantly affected in 38 fetuses (95%). In 18, the pregnancy was terminated (n = 9; 23%) or ended with fetal demise (n = 9; 23%). Of the 20 live births, 14 (70%) had significant dilation of 1 (n = 12) or both (n = 2) ventricles at the most recent echocardiogram. Moreover, 4 of 17 (24%) with >1 year of post-natal follow-up have neurocognitive abnormalities in association with metabolic or syndromic disorders.
Transplacental dexamethasone and immunoglobulin were routinely used to treat antibody-mediated endocardial fibroelastosis (n = 6) (23,24). Post-natal surgery was required in 6 cases, including a permanent pacemaker for bradycardia (n = 3), right atrial reduction surgery (n = 1), infant heart transplantation for biventricular noncompaction (n = 1), and repair of ruptured tricuspid valve secondary to antibody-mediated fibroelastosis (n = 1) (25).
In all 21 cases, myocardial hypertrophy eventually affected both ventricles. The main associations were with homozygous alpha-thalassemia (n = 5) and Noonan syndrome (n = 2). Other conditions were observed in isolation (Table 5). Four pregnancies (19%) were terminated, whereas 8 fetuses (38%) died in utero. Repeated fetal transfusions were used to treat alpha-thalassemia, but ultimately only 1 of 5 (20%) survived. Of 9 pregnancy survivors, 2 had post-natal interventions. The first fetus with Noonan syndrome underwent heart transplantation in infancy because of progressive ventricular hypertrophy and heart failure. The second with Hurler syndrome died of complications of bone marrow transplantation. Finally, a child of consanguineous parents with a previously affected sibling died of heart failure while awaiting heart transplantation.
Figure 1 shows Kaplan-Meier estimates of freedom from death and transplantation. When compared with NHCM, HCM was associated with significantly worse odds of fetal survival (OR: 2.3; 95% CI: 1.4 to 8.4; p = 0.007) and a significantly worse survival rate from birth (HR: 4.5; 95% CI: 1.9 to 27.9; p = 0.004), respectively. Table 3 compares characteristics among survivors and nonsurvivors from fetal diagnosis to infancy. Figure 2 illustrates the distribution of CVPS at the time of a fetal baseline echocardiogram and outcome at 1 month and 1 year of life. Multivariable analysis including all actively managed CM patients demonstrates significant associations between mortality and ventricular septal thickness (HR: 1.21 per z-score increment; 95% CI: 1.07 to 1.36; p = 0.002), CTR (HR: 1.06 per percent increment; 95% CI: 1.02 to 1.1; p = 0.006), diastolic dysfunction (HR: 1.44; 95% CI: 1.07 to 1.95; p = 0.02), and gestational age at CM diagnosis (HR per week increment: 0.91; 95% CI: 0.83 to 0.99; p = 0.03). If patients not in sinus rhythm are removed from the model, the only factor associated with increased hazard of mortality is a lower CVPS score (HR: 1.45 per point decrease; 95% CI: 1.16 to 1.79; p = 0.001). Multivariate subgroup analysis for fetal NHCM also shows similar association between death and lower CVPS (1.81 per point decrease; 95% CI: 1.09 to 3.00; p = 0.02).
This study documents the disease pattern of fetal CM in a well-defined population, the largest published to date. Echocardiography, by providing an accurate demonstration of cardiac morphology and function, was shown to be an important tool to predict fetal CM outcomes. The CVPS was demonstrated here for the first time to have a strong prognostic value. A CVPS ≤6 was associated with the highest risk of death or need for heart transplantation of any of the tested variables.
The incidence of prenatally diagnosed CM (6.2 of 100,000) in our population is comparable to that reported in children younger than 1 year of age in Finland (4.1 of 100,000), Australia (7.8 of 100,000), and the United States (8.34 of 100,000) and higher than among older children and adults (2–4). Similar to the post-natal experience (6), we found that prenatally diagnosed CM represents a heterogeneous group of predominantly rare disorders. Moreover, diseases with similar phenotypes have different etiologies and commonly involve other organs. Moreover, conditions with similar etiology may present and evolve differently, which makes parental counseling very challenging.
The diagnostic workup for CM is relatively invasive and not readily obtainable during pregnancy. In many cases, the etiology remains unknown despite comprehensive workups. In this series, a causal diagnosis was ascertained in 40% of fetal nonsurvivors but in 71% of live births. In pediatric series, only 33% to 43% had a known cause of CM (3,4,6). With the increasing use of microarray and whole-exome sequencing techniques, this percentage is likely to increase.
An important finding of this study is that despite its etiological heterogeneity, fetal CM prognosis is closely associated with the phenotypic and functional presentation. However, we found that conventional distinctions between hypertrophic, dilated, restrictive, and unclassified CM subtypes, frequently used in post-natal CM (15), were difficult to apply to the fetus. In infancy, ventricles are typically either hypertrophic or dilated, with the left ventricle predominantly affected. In fetal life, typically both ventricles were either hypertrophied or had normal wall thickness, whereas significant ventricular dilation was uncommon early in the disease. Restrictive diastolic filling occurred with all CM phenotypes. Therefore, we elected to simplify classification to hypertrophic or nonhypertrophic phenotypes. Although others have defined dilated CM to include restrictive and unclassified forms of fetal NHCM (8,17,18), we believe that our classification into 1 of 2 mutually exclusive phenotypes on the basis of myocardial wall thickness is phenotypically more precise, easily applied, reproducible, and of prognostic relevance. Moreover, it appears to have reasonable fidelity because the phenotypic case classification changed in only 4 patients (6%) over time, these being fetuses presenting with isolated diastolic dysfunction and in whom later phenotypic HCM developed.
There are few previous reports on fetal “dilated” CM. The largest by Sivasankaran et al. (26) included 36 of 50 fetuses with ventricular dilation secondary to genetic-metabolic disorders (n = 11), viral infections (n = 11), LV noncompaction (n = 1), and idiopathic (n = 13; 36%) causes. Two-thirds had fetal hydrops. Excluding the remaining 14 fetuses with cardiac dysfunction secondary to vascular obstruction, renal disease, or anemia, the overall survival rate was 15% with an undefined follow-up period. Pedra et al. (8) documented 22 fetal NHCM cases related to CMV infection (n = 2), maternal autoantibodies (n = 6), familial CM (n = 5), or unknown etiology (n = 9). Only 18% in which the pregnancy was continued survived to infancy. Similarly, only 1 of 6 fetuses with NHCM survived to childhood in the Yinon et al. series (18). A common feature of previous fetal CM reports has been the extremely poor outcome. However, this may not be true for patients in whom an identifiable cause is demonstrated. We reported improved outcomes of antibody-mediated CM with perinatal anti-inflammatory steroid and immunoglobulin therapy compared with untreated patients (24,27). This and the increased detection of less severely affected fetuses by NHCM may explain the better survival rates compared with previous studies. Freedom from death or transplantation of fetal survivors with NHCM was 85% and 75% at 1 and 5 years, respectively, comparing also more favorably with outcomes of children with dilated CM with transplantation-free survival rates of 75% and 63% at 1 and 5 years, respectively (28).
In contrast to NHCM, we found that fetal HCM was associated with a dismal rate (12%) of transplantation-free survival from diagnosis to school age, making this an even worse condition in terms of survival than any other form of severe congenital heart disease (29). Importantly, none of the survivors beyond the first year of life had myocardial wall thickness z-scores >2.5 at the baseline examination, suggesting that restrictive diastolic filling is not tolerated by the fetus and infant with more severe HCM. Data regarding pre-natal diagnosis and postnatal outcome of HCM are scarce: of 8 earlier fetal cases with nonpregnancy-related HCM, Pedra et al. (8) encountered a single survivor beyond the neonatal period. As in the current study, most of these cases were attributed to alpha-thalassemia, Noonan syndrome, or an unidentified etiology. Similarly, several pediatric publications have indicated that HCM presenting in infancy carries a worse prognosis than in older children due to its frequent association with intractable heart failure (30–32).
Ventricular noncompaction has been recognized as a pre-natally detectable entity, especially in the past decade, typically in association with structural heart disease, fetal hydrops, and demise (19,26,33,34). In our series, 5 of 8 patients with isolated ventricular noncompaction are currently alive, 1 of whom underwent heart transplantation in infancy. In 1 patient, severe fetal hydrops resolved spontaneously in late gestation. Two of the mothers in this subgroup were also affected by CM, suggesting a dominant inheritance pattern in some families.
This was a retrospective review, and some eligible cases of primary CM may not have been documented. A complete diagnostic workup was not always possible in cases of pre-natal demise. Finally, fetal CM represents a mixed group of rare disorders, and the relatively small number of fetuses with each condition precluded an etiology-specific statistical analysis.
Fetal CM is an etiologically highly diverse disease with cardiac function and morphometry-specific outcomes. We propose a simple classification on the basis of a hypertrophic phenotype that seems to correlate well with prognosis. Ventricular septal thickness, CTR, diastolic dysfunction, gestational age, and CVPS are independently associated with mortality. Fetuses with HCM had a very high risk of perinatal demise. In contrast, those fetuses with a fetal diagnosis of NHCM had better outcomes than previously reported. Involvement of other organs was common in fetuses with genetic, metabolic, post-infectious, or syndromic disorders and may substantially affect the clinical outcome of survivors.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- cardiothoracic area ratio
- cardiovascular profile score
- hypertrophic cardiomyopathy
- hazard ratio
- left ventricular
- myocardial performance index
- nonhypertrophic cardiomyopathy
- Received November 11, 2013.
- Revision received February 5, 2014.
- Accepted February 25, 2014.
- American College of Cardiology Foundation
- Arola A.,
- Jokinen E.,
- Ruuskanen O.,
- et al.
- Cox G.F.,
- Sleeper L.A.,
- Lowe A.M.,
- et al.
- Pedra S.R.,
- Smallhorn J.F.,
- Ryan G.,
- et al.
- Huhta J.C.
- Richardson P.,
- McKenna W.,
- Bristow M.,
- et al.
- Elliott P.,
- Andersson B.,
- Arbustini E.,
- et al.
- Kohli S.K.,
- Pantazis A.A.,
- Shah J.S.,
- et al.
- Jaeggi E.T.,
- Fouron J.C.,
- Silverman E.D.,
- Ryan G.,
- Smallhorn J.,
- Hornberger L.K.
- Trucco S.M.,
- Jaeggi E.,
- Cuneo B.,
- et al.
- Nield L.E.,
- Silverman E.D.,
- Taylor G.P.,
- et al.
- Daubeney P.E.,
- Nugent A.W.,
- Chondros P.,
- et al.
- Maron B.J.,
- Tajik A.J.,
- Ruttenberg H.D.,
- et al.
- Skinner J.R.,
- Manzoor A.,
- Hayes A.M.,
- Joffe H.S.,
- Martin R.P.
- Colan S.D.,
- Lipshultz S.E.,
- Lowe A.M.,
- et al.