Reversible De Novo Left Ventricular Trabeculations in Pregnant WomenCLINICAL PERSPECTIVE
Implications for the Diagnosis of Left Ventricular Noncompaction in Low-Risk Populations
Background—Patients with heart failure and chronic anemia frequently demonstrate left ventricular (LV) trabeculations, which may be compatible with the diagnosis of LV noncompaction. We used the pregnancy model, which is characterized by a reversible increase in cardiac preload and other changes in cardiac function, to assess the development of de novo LV trabeculations in women with morphologically normal hearts.
Methods and Results—One hundred two primigravida pregnant women were evaluated longitudinally with a series of echocardiograms in the first trimester, in the third trimester, and postpartum. Echocardiograms were analyzed according to established guidelines. Increased LV trabeculations and the presence of LV noncompaction were based on established criteria. Pregnancy was associated with an increased heart rate, stroke volume, and cardiac output, as well as increased LV volume and mass. During pregnancy, 26 women (25.4%) developed increased trabeculations. Eight women showed sufficient trabeculations to fulfill criteria for LV noncompaction. During the postpartum follow-up period of 24±3 months, 19 women (73%) demonstrated complete resolution of trabeculations, and 5 showed a marked reduction in the trabeculated layer.
Conclusions—Pregnancy induces de novo LV trabeculations in a significant proportion of women. The results suggest that LV trabeculations occur in response to increased LV loading conditions or other physiological responses to pregnancy and are not specific for LV noncompaction. These factors should be considered in the assessment of individuals with LV trabeculations outside the context of symptoms of heart failure or familial cardiomyopathy.
- ethnic groups
- isolated noncompaction of the ventricular myocardium
Left ventricular (LV) noncompaction (LVNC) cardiomyopathy is an unclassified primary cardiomyopathy characterized by increased myocardial trabeculations and recesses. The precise stage of development and the natural history of the condition are not fully understood; however, preliminary data are indicative of a morphologically and clinically heterogeneous disorder. Some individuals present with overt heart failure, fatal arrhythmias, and thromboembolic events, and others remain asymptomatic.1 Echocardiographic criteria to facilitate the identification and assessment of LVNC rely on the presence of LV myocardial trabeculations and a 2-layer distinction between compacted and noncompacted myocardium (Figure 1).2–5
Editorial see p 453
Clinical Perspective on p 483
The past 2 decades have witnessed significant advances in tissue harmonics, which have enabled detailed assessment of the ventricular myocardium. Such progress in image resolution has coincided with an increasing number of scientific reports relating to LVNC. The initial prevalence of LVNC was estimated to be <0.3%.2,6 However, recent studies in patients with heart failure have reported a high proportion (almost 35%) of patients with myocardial trabeculations, with nearly 25% fulfilling diagnostic criteria for LVNC regardless of the criteria used.7 A study of >1000 athletes demonstrated that 18% exhibited increased LV trabeculations and 8% fulfilled echocardiographic criteria for LVNC.8 Patients with sickle cell anemia also reveal a high (28.3%) prevalence of increased LV trabeculations.9 A common theme among all 3 cohorts is the presence of an increased cardiac preload. Given the relatively low prevalence of the other primary cardiomyopathies in the general population, it is unlikely that all such individuals suffer from LVNC. Indeed, the most probable explanation for the increased trabeculations and recesses observed in these states is a physiologically determined epiphenomenon in response to a chronic increase in cardiac preload.
Pregnancy is associated with doubling of maternal cardiac preload for at least 4 months.10 Based on the cardiovascular changes observed in pregnancy, a longitudinal pregnancy model was applied to evaluate whether increased LV trabeculations observed in the aforementioned situations are a cardiac response to increased preload rather than LVNC. We hypothesized that women with a morphologically normal LV myocardium may develop LV trabeculations during pregnancy and that such changes would resolve within a few weeks of delivery, after normalization of blood volume.
We conducted a prospective longitudinal echocardiographic study of 102 primigravida pregnant women (66 white, 36 black) between 2010 and 2013. The women were recruited from 2 specialized antenatal units in south London (University Hospital Lewisham and St. George’s Hospital NHS Trust). All women provided consent and were evaluated by a self-reported health questionnaire, which required information relating to the presence of cardiac symptoms, drug history, previous miscarriages, and family history of premature cardiovascular disease or sudden cardiac death; physical examination; urine dipstick; 12-lead ECG; and 2-dimensional echocardiography. The women’s ethnicity was self-assigned. All women underwent repeat 12-lead ECG and 2-dimensional echocardiography between 28 to 36 weeks of gestation and 3 to 12 months postpartum.
Selection criteria for the study included age from 18 to 35 years, primigravida status, singleton pregnancy, black or white ethnicity, absence of cardiac symptoms, a structurally normal heart on echocardiography without any evidence of increased LV trabeculations, and the ability to perform a full range of high-quality images with clear definition of the endocardium in the standard planes in the third trimester of pregnancy.
Two-dimensional echocardiography was performed by a cardiologist with accreditation in echocardiography (S.G.) with either the Philips Sonos 7500 or Philips iE33. Standard cardiac views were obtained and analyzed according to protocols specified by the American Society of Echocardiography.13 All measurements were recorded as absolute values. Cardiac dimensions and volumes were assessed as previously described.8 LV mass was calculated with the formula of Devereux and indexed for body surface area.14 Assessment of diastolic function included pulsed-wave Doppler across the mitral valve and tissue Doppler velocity imaging of the septal and lateral mitral valve annuli in the apical 4-chamber view. Stroke volume was calculated as the product of aortic Doppler flow velocity time integral and cross-sectional area of the LV outflow tract.15 Cardiac output was calculated as the product of stroke volume and heart rate derived from ECG monitoring.15 Mean arterial blood pressure was calculated as diastolic blood pressure minus systolic blood pressure divided by diastolic blood pressure. Total vascular resistance was calculated (in dynes·s·cm−5) with the standard formula: Total vascular resistance equals (mean arterial blood pressure [mm Hg] divided by cardiac output [L/min]) times 80.16
All echocardiographic studies were saved to compact disks as numeric files to generate anonymity. Assessments for increased LV trabeculations and cardiac measurements were repeated independently by an accredited cardiac physiologist (M.R.) who was blinded to the identity, ethnicity, and gestational stage of the women.
The original numeric files (generated and labeled by S.G.) were shuffled and renumbered before being presented back to that investigator to conduct the intraobserver variability assessment to ensure that she was also blinded to the identity, ethnicity, and gestational stage of the women during the second analysis. Discrepancies between the 2 investigators (M.P. and S.G.) relating to increased LV trabeculations or criteria for LVNC during interobserver analysis were referred to the Director of Echocardiography (R.S.) and the Director of Inherited Cardiac Diseases (S.S.) at St. George’s Hospital. Both of these investigators were blinded to the identity, ethnicity, and gestational age of the women when attempting to reach a final consensus decision.
Determination and Definition of Increased LV Trabeculation
LV trabeculations were assessed in the short-axis views at the level of the mitral valve, papillary muscle, and apex, as well as the 4-chamber and 2-chamber views. Care was taken to use only these standard views for every echocardiogram.
Myocardial trabeculations were defined as localized protrusions of the endocardial surface ≥3 mm in diameter8,17 associated with intertrabecular recesses on 2-dimensional echocardiography (Figure 1). The definition of increased LV trabeculations was derived from a randomly selected asymptomatic and normotensive (blood pressure <120/80 mm Hg) control population of 138 healthy nonpregnant females (91 white, 57 black) recruited from a population-screening program offered by the charity Cardiac Risk in the Young.18 The age and size of the control subjects were similar to those of the pregnancy cohort (mean age, 30.0±5.9 versus 30.7±4.4 years, P=0.309; body surface area, 1.7±06 versus 1.8±0.3 m2; P=0.122).
Increased LV trabeculation was defined as ≥95th percentile of the distribution of trabeculations in the population of women, which corresponded to 3 trabeculations. This number is identical to a previous description of increased LV trabeculations derived from a large cohort attending for echocardiography.17 Color-flow Doppler using a Nyquist limit of 20 to 30 cm/s was applied when numerous small cavities within the endocardial surface were identified in a 2-layered structure.
The echocardiographic criteria by Chin et al2 and Jenni et al3 were used to define LVNC (Figure 1). The distribution of trabeculations was determined with the use of the 16-segment model recommended by the American Society of Echocardiography.13
The National Research Ethics Service, East London 2 Research Ethics Committee granted ethical approval in the UK.
Statistical analyses were performed with SPSS version 18.0 (SPSS, Inc, Chicago IL). Variables were tested for normality with the Kolmogorov-Smirnov test. Values are expressed as either mean±SD or percentages as appropriate. Differences within group means were compared by use of the paired Student t test or the Wilcoxon test, and differences between group means were compared by use of independent t tests or Mann-Whitney U tests (for normally and nonnormally distributed variables, respectively). The χ2 test or Fisher exact test was used as appropriate to test group differences of proportions. Repeated-measures ANOVA was performed to evaluate changes in the clinical and echocardiographic parameters during pregnancy and the postpartum period. Logistic regression was used to determine multiple-adjusted risk for the presence of increased (≥3) trabeculations during pregnancy in relation to ethnicity, age, body mass index, heart rate, blood pressure, LV stroke volume, LV cavity and volume size, and LV mass. Statistical significance was defined as a 2-tailed value of P<0.05 throughout.
Reproducibility of LV and RV measurements was assessed with intraclass correlation coefficient analysis and reported as intraclass correlation coefficient (95% confidence interval). Averaged measures were used as an index for the reliability of 2 different rates. Intraobserver and interobserver agreement for the presence of LV trabeculations was assessed by κ statistics.
Pregnant women had a mean age of 30.7±4.4 years and a body surface area of 1.8±0.3 m2. The majority (64.7%) were white. None of the women had a family history of cardiomyopathy or premature (<40 years of age) sudden cardiac death. All women were normotensive (blood pressure ≤120/80 mm Hg), and none revealed proteinuria.
Hemodynamic and Structural Changes During Pregnancy
The progression of pregnancy from the first trimester to the third trimester was accompanied by increases in heart rate, stroke volume, and cardiac output and a reduction in total vascular resistance. Advancing pregnancy was also associated with a modest (11%–12%) increase in LV volume and LV mass (Table 1).
Pregnant Women With Increased LV Trabeculations
Twenty-five percent of pregnant women developed increased LV trabeculations (Figure 2), which were more common in black women than in white women (47.2% versus 13.6%; P=0.0003). Of the total pregnancy cohort, 10 women (9.8%) fulfilled the Jenni et al criteria, 19 (18.6%) fulfilled the Chin et al criteria, and 8 (7.8%) fulfilled both criteria for LVNC (Figure 3). Women fulfilling both criteria did not show any significant ethnic predilection (black: n=4, 11.1%; white: n=4, 6.1%; P=0.459). During pregnancy, the Chin et al X/Y ratio at the site of maximal wall thickness averaged 0.43±0.07 (range, 0.35–0.5), and the Jenni et al ratio of noncompacted to compacted myocardial at the site of maximal wall thickness averaged 2.2±0.6 (range, 2.1–3.4).
LV trabeculations were distributed predominantly in the anterolateral territory (n=14), followed by the inferolateral (n=7), inferior (n=7), apical lateral (n=7), and apical inferior (n=6) regions.
Differences in ECG and Echocardiographic Parameters in Pregnant Women Exhibiting Increased LV Trabeculations and Those With Normal Cardiac Morphology
Pregnant women who developed LV trabeculations did not differ from those without LV trabeculations in terms of electrocardiographic abnormalities. One pregnant woman with LV trabeculations showed minor T-wave inversion in leads V1 and V2 on the baseline ECG, as did 1 woman with normal myocardial morphology, which persisted during pregnancy and for 12 months postpartum. There were no significant differences in wall thickness, cavity size, indexes of systolic and diastolic function, or stroke volume between the women in the 2 groups (Table 2). None of the women in the study exhibited abnormal systolic or diastolic function.
Determinants of Increased LV Trabeculations
There was no significant association between increased LV trabeculations and age, body mass index, systolic blood pressure, LV cavity dimension, stroke volume, or LV mass.
Multiple-adjusted binary logistic regression demonstrated that ethnicity was the only independent predictor for the presence of increased (≥3) trabeculations during pregnancy. Black women were almost 3 times more likely to develop increased LV trabeculations than white women during pregnancy after adjustment for age, body mass index, systolic blood pressure, LV cavity dimension, LV stroke volume, and LV mass. (odds ratio, 2.7; 95% confidence interval, 1.1–6.7; P=0.034).
The Postpartum Assessment in Pregnant Women
In the postpartum period, 18 women (69.2%) showed complete resolution of LV trabeculations over a mean duration time of 8.1±4.2 months (Figure 2). Seven women (27%) continued to display LV trabeculations, which did not show any ethnicity predilection (black, n=2; white, n=5; P=0.202). The 7 women who continued to exhibit trabeculations did not reveal any differences in the number of trabeculated segments during or after pregnancy; LV trabeculations were distributed primarily in the apical inferior (n=5) and apical lateral (n=5) territories, followed by the midcavity region (n=3).
All 7 women were followed up for an additional 24±3 months and remained asymptomatic with blood pressures <120/80 mm Hg. One woman became pregnant again and was excluded from this analysis. At repeat echocardiographic assessment, 5 women continued to exhibit LV trabeculations, and 1 woman showed complete regression of the myocardial anomaly. In the 5 women with persisting trabeculations, the noncompacted and compacted layers progressively decreased over time according to 1 or both sets of criteria (Table 3). All 5 women continued to reveal normal LV systolic and diastolic functions.
Reproducibility of LV Measurements and Increased LV Trabeculations
LV measurements were highly reproducible at the interobserver level (Table 4). There were no cases of disagreement with the independent reviewer with respect to increased LV trabeculations during the first-trimester scans. There were 2 cases of disagreement with respect to increased LV trabeculations during the third trimester, translating to κ=0.87 (P<0.001). The κ values for the disparity in measurements relating to the Chin et al criteria and Jenni et al criteria were 0.62 (P=0.01) and 0.29 (P=0.15), respectively. In the postpartum scans, there was 1 case in which the independent reviewer disagreed about the increased LV trabeculations, resulting in κ=0.93 (P<0.001). The κ values for disparity with respect to the Chin et al criteria and Jenni et al criteria were 0.61 (P=0.02) and 0.33 (P=0.173), respectively.
There were only 2 cases of disparity with respect to increased LV trabeculations for third-trimester scans during reanalysis by first author (S.G.), translating to κ=0.93 (P<0.001). There was no disparity for first-trimester or postpartum scans.
The clinical entity LVNC is established in the pediatric population and is usually associated with other congenital cardiac abnormalities that increase cardiac preload.2 The issue of adult LVNC, which was first described by Engberding and Bender19 in 1984, is less clear, particularly because current diagnostic echocardiographic and magnetic resonance imaging (MRI)–derived criteria for LVNC are based on small cohorts and have not been validated prospectively in large populations.20 This study revealed that de novo LV trabeculations occurred in 25% of pregnant women. The majority of women (69.2%) with trabeculations demonstrated complete resolution, and another 12% showed gradual regression toward normal morphology over a 2-year follow-up period.
The absence of symptoms or echocardiographic features of LV dysfunction suggests that it is highly unlikely that such a large proportion of women with trabeculations suffered from an incomplete phenotypic manifestation of LVNC, particularly because the LV was morphologically normal at the time of recruitment into the study. We have previously reported increased LV trabeculations among a large cohort of athletes and patients with sickle cell anemia; 8% of both cohorts fulfilled criteria for LVNC.8 A study by Kohli et al7 reported that 24% of heart failure patients exhibited features consistent with LVNC. We believe that the common factor in all 3 situations is an increased cardiac preload. All of the aforementioned studies were cross-sectional; therefore, it has been impossible to ascertain whether increased LV trabeculations are a de novo response to an increased cardiac preload, are an exaggeration of anomalies that were already present as a result of incomplete manifestation of LVNC, or form part of a normal morphological spectrum in some healthy individuals.
We postulated that LV trabeculations represented a normal response to an increased cardiac preload in some individuals. Pregnancy is an attractive natural physiological model for testing the hypothesis in human subjects because it is associated with a 50% increase in plasma volume, stroke volume, and cardiac output. The demonstration that pregnancy induces de novo LV trabeculations in women with normal myocardial morphology indicates that LV trabeculae are an epiphenomenon of increased preload in several disease states, including heart failure7 and chronic anemia,9 but also physiological states such as high-level athletic training.8
The study revealed that the progression of pregnancy from the first to the third trimester was associated with a reversible increase in stroke volume, cardiac output, LV volumes, and LV mass. Although it is possible that LV trabeculations may simply have been more visible in those women with the greatest increases in LV volume, we did not identify any differences in LV volumes or stroke volume in women who developed trabeculations compared with women with normal morphology. Therefore, it is possible that in some individuals, a preliminary adaptive response to wall tension consists of a modest increase in LV volume and mass and the development of trabeculations. We suspect that these changes are likely to be more obvious in individuals who have been subject to a longer preload stress such as athletic training, chronic anemia, and heart failure.
The overall number of women fulfilling criteria for LVNC was small (n=8); therefore, we were unable to discern any statistical differences between the 2 ethnic groups. However, our analysis revealed that black ethnicity was an important determinant of increased LV trabeculations in this study. These observations concur with previous studies reporting an increased prevalence of both LV trabeculations and criteria for LVNC in black patients with heart failure7 and black athletes compared with whites.8 The higher prevalence of LV trabeculations during pregnancy in black women is probably synonymous with the greater magnitude of LV hypertrophy resulting from the increased LV afterload in hypertensive black patients21 and suggests that some confounding ethnicity-specific genetic factors could be involved in the exaggeration of this phenomenon.
It is important to emphasize that the temporal pattern of de novo development of LV trabeculations as pregnancy progressed and subsequent resolution in the postpartum period were observed in both ethnic groups. These findings indicate that LV trabeculations are part of a physiological spectrum of cardiac adaptation in both black and white patients. Therefore, such morphology should not be considered specific for the diagnosis of LVNC in either of these ethnic groups in the absence of cardiac symptoms, familial disease, LV dysfunction, or ventricular tachycardia.
This study does not necessarily support the notion that most patients with heart failure and increased LV trabeculations probably do not have genuine LVNC. However, it provides an alternative and adequate explanation for the high prevalence of LV trabeculations observed in low-risk populations such as athletes8 and those attending for echocardiograms in the absence of any suspicious features of heart failure.22 Such cases are an increasingly frequent source of diagnostic dilemmas and may occasionally lead to an erroneous life-long diagnosis. Our observations have important implications for the diagnosis and management of LVNC in asymptomatic individuals harboring LV trabeculations outside the context of familial heart failure or sudden cardiac death. The findings suggest that a more thorough assessment of current diagnostic criteria and entities relating to adult LVNC is crucial for a better understanding of this currently unclassified cardiomyopathy. The low interobserver κ values (0.29) for the Jenni et al criteria, which are measured in end systole, compared with the Chin et al criteria (0.62), which are measured in end diastole, suggest that future echocardiographic LVNC criteria will be more reproducible if LV measurements are made in diastole.
Given the safety concerns surrounding cardiac MRI in the first trimester, the conclusions from this study were based solely on echocardiographic assessment, although cardiac MRI may be considered the gold-standard assessment for myocardial trabeculations.23 We suspect that the prevalence of de novo trabeculations would have been even higher with cardiac MRI and would have helped to strengthen our findings. We have previously published a study revealing a high prevalence of LV trabeculations in athletes8 in which athletes fulfilling echocardiographic criteria for LVNC2,3 also revealed cardiac MRI features of the disorder.24
During the assessment for trabeculations, we attempted to be as meticulous as possible to make measurements in the same standard images each time. It is possible that our short-axis measurements when transcending from the level of the papillary muscles toward the LV apex may have varied by a millimeter or two between serial echocardiograms in any given pregnant woman. Therefore, we may have failed to detect trabeculations in some women as a result of these inherent practical limitations of echocardiography. Nevertheless, this limitation does not deter from the fact that 1 in 4 women with completely normal LV morphology went on to develop a significant number of trabeculations as pregnancy progressed.
The potential impact of high concentrations of estrogen on the development of trabeculations cannot be excluded with certainty; however, a high proportion of LV trabeculations and noncompaction criteria are also observed in men with heart failure and sickle cell anemia and male athletes, suggesting that an increased cardiac preload is the most plausible explanation for this epiphenomenon. Evaluation of first-degree relatives of women in whom increased LV trabeculations persisted was not conducted but may have provided valuable information for establishing whether increased LV trabeculations have a genetic predisposition.
Pregnancy induces de novo LV trabeculations in a significant proportion of women. Our observation provides an explanation for the high prevalence of increased trabeculations identified in low-risk individuals subjected to increased cardiac loading conditions. These factors should be taken into consideration in asymptomatic individuals with LV trabeculations and those fulfilling echocardiographic criteria for LVNC outside the context of symptoms of heart failure or familial cardiomyopathy.
We are grateful to Cardiac Risk in the Young for providing the portable echocardiography equipment and ECG machines used for the study in the United Kingdom. We also acknowledge Sandra Linton and Sayli Nirkhee, who assisted in the recruitment of pregnant women.
Source of Funding
Drs Gati, Zaidi, and Sharma and M. Reed were funded by research grants from Cardiac Risk in the Young.
- Received January 8, 2014.
- Accepted May 30, 2014.
- © 2014 American Heart Association, Inc.
- Oechslin E,
- Jenni R
- Chin TK,
- Perloff JK,
- Williams RG,
- Jue K,
- Mohrmann R
- Jenni R,
- Oechslin E,
- Schneider J,
- Attenhofer Jost C,
- Kaufmann PA
- Williams T,
- Machann W,
- Kühler L,
- Hamm H,
- Müller-Höcker J,
- Zimmer M,
- Ertl G,
- Ritter O,
- Beer M,
- Schönberger J
- Kohli SK,
- Pantazis AA,
- Shah JS,
- Adeyemi B,
- Jackson G,
- McKenna WJ,
- Sharma S,
- Elliott PM
- Gati S,
- Chandra N,
- Bennett RL,
- Reed M,
- Kervio G,
- Panoulas VF,
- Ghani S,
- Sheikh N,
- Zaidi A,
- Wilson M,
- Wilson M,
- Papadakis M,
- Carré F,
- Sharma S
- Friedmann H
- Cerqueira MD,
- Weissman NJ,
- Dilsizian V,
- Jacobs AK,
- Kaul S,
- Laskey WK,
- Pennell DJ,
- Rumberger JA,
- Ryan T,
- Verani MS
- Devereux RB
- Lewis JF,
- Kuo LC,
- Nelson JG,
- Limacher MC,
- Quinones MA
- 18.↵Cardiac Risk in the Young. www.c-r-y.org.uk. Accessed May 29, 2014.
- Thavendiranathan P,
- Dahiya A,
- Phelan D,
- Desai MY,
- Tang WH
Left ventricular (LV) noncompaction (LVNC) is a relatively novel cardiomyopathy that is characterized by increased LV trabeculations and 2-layered myocardial structure comprising an outer compacted and inner noncompacted layer. Major complications include systolic heart failure, ventricular tachycardia, and systemic thromboembolism. Advances in tissue harmonics have enabled more detailed assessment of the LV myocardium and have been paralleled by a significantly large proportion of individuals with heart failure and chronic anemia, as well as those engaged in athletic training fulfilling echocardiographic criteria for LVNC. A chronic increase in cardiac preload is common to all these conditions and raises the possibility that increased LV trabeculations in such cases may represent a morphological epiphenomenon of increased cardiac preload compared with LVNC. This study used a pregnancy model, which is associated with reversible 2-fold increase in blood volume, to assess the development of de novo LV trabeculations in 102 primigravida women with a morphologically normal LV. During pregnancy, 26 women (25.4%) developed de novo trabeculations, and 8 women fulfilled the criteria for LVNC. Over a 2-year follow-up period, almost three quarters of the women showed complete regression of LV trabeculations, and the remainder showed a significant reduction in the trabeculated layer. Women with trabeculations showed persevered LV function throughout. This study demonstrates that LV trabeculations occur in response to increased LV preload and are not specific for LVNC. This factor should be considered in the assessment of individuals with increased LV trabeculations outside the context of symptoms of heart failure or familial cardiomyopathy.