Long-Term Outcomes of Childhood Left Ventricular Non-Compaction

Cardiomyopathy:  Results from a National Population-Based Study

William Y. Shi, MBBS, PhD1,2,3; Margarita Moreno-Betancur, PhD2,3; Alan W. Nugent,

MBBS4; Michael Cheung, MBChB5; Steven Colan, MD6; Christian Turner, MBBS7; Gary

F. Sholler, MBBS8; Terry Robertson, MBBS9; Robert Justo MBBS9; Andrew Bullock,

MBBS10; Ingrid King2, Andrew M. Davis, MBBS5;2,3,5 Piers E. F. Daubeney, DM

MBBS11,12*; Robert G. Weintraub, MBBS2,3,5,*; for the National Australian Childhood

Cardiomyopathy Study

1. Department of Cardiac Surgery, Royal Children’s Hospital, Melbourne, AUSTRALIA

2. Murdoch Children’s Research Institute, Melbourne, AUSTRALIA

3. University of Melbourne, AUSTRALIA

4. Department of Pediatrics, University of Texas Southwestern, Dallas, TX, USA

5. Department of Cardiology, Royal Children’s Hospital, Melbourne, AUSTRALIA

6. Department of Cardiology, Boston Children’s Hospital, MA, USA

7. Department of Cardiology, Children’s Hospital at Westmead, Sydney, AUSTRALIA

8. Department of Cardiology, Women’s and Children’s Hospital, Adelaide, AUSTRALIA

9. Department of Cardiology, Mater Children’s Hospital, Brisbane, AUSTRALIA

10. Department of Cardiology, Princess Margaret Hospital, Perth, AUSTRALIA

11. Department of Paediatric Cardiology, Royal Brompton Hospital, London, UNITED

KINGDOM

12. National Heart and Lung Institute, Imperial College, London, UNITED KINGDOM

* Equal senior co-authors

Total word count: 3432 words

Corresponding author:Robert G. WeintraubDepartment of CardiologyRoyal Children’s HospitalMelbourne, Victoria, AUSTRALIARobert.Weintraub@rch.org.au

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

252627282930313233

ABSTRACT

Background: Long-term outcomes for childhood left ventricular non-compaction

(LVNC) are uncertain. We examined late outcomes for children with LVNC enrolled in a

national population-based study.

Methods: The National Australian Childhood Cardiomyopathy Study includes all

children in Australia with primary cardiomyopathy diagnosed <10 years of age between

1987 and 1996. Outcomes for LVNC subjects with a dilated phenotype (LVNC-D) were

compared to those with dilated cardiomyopathy (DCM). Propensity-score analysis was

used for risk factor adjustment.

Results: There were 29 subjects with LVNC (9.2% of all cardiomyopathy subjects) with

a mean annual incidence of newly diagnosed cases of 0.11 per 100,000 at-risk persons.

Congestive heart failure was the initial symptom in 24 (83%) of 29 subjects, and 27

(93%) had a dilated phenotype (LVNC-D). The median age at diagnosis was 0.3

(interquartile interval 0.08 – 1.3) years of age. The median (interquartile interval)

duration of follow-up was 6.8 (0.7-14.1) years for all subjects and 24.7 (23.3 – 27.7)

years for surviving subjects. Freedom from death or transplantation was 48% (95% CI 30

– 65%) at 10 years after diagnosis and 45% (95% CI 27-63%) at 15 years. By competing

risk analysis, 21% of LVNC subjects were alive with normal LV systolic function and

31% were alive with abnormal function at 15 years. Propensity-score matching between

LVNC-D and DCM subjects suggested a lower freedom from death/transplantation at 15

years after diagnosis in the LVNC-D subjects (LVNC-D: 46% (95% CI 26-66%) vs.

DCM: 70% (95% CI 42-97%), p=0.08). Using propensity-score inverse probability of

treatment weighted Cox regression, we found evidence that LVNC-D was associated with

a greater risk of death or transplantation (HR 2.3, 95% CI 1.4-3.8, p=0.0012).

Conclusions: Symptomatic children with LVNC usually present in early infancy with a

predominant dilated phenotype. Long-term outcomes are worse than for matched children

with DCM.

2

12

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

CLINICAL PERSPECTIVE

What is new?

Long-term outcomes for children with left ventricular non-compaction (LVNC)

are uncertain.

The NACCS uniquely represents the longest and most complete longitudinal

cohort study of childhood cardiomyopathy.

This population-based study defines the incidence, presentation and long-term

outcomes for LVNC diagnosed during childhood.

Risk-adjusted analyses in this study showed that subjects with LVNC and dilated

physiology tended to experience worse survival than matched subjects with

dilated cardiomyopathy, with a two-fold higher risk of death or transplantation.

Clinical implications:

Our findings support consideration of LVNC as a distinct cardiomyopathy

phenotype in pediatric patients.

Children with LVNC who develop heart failure at an early age may benefit from

specialist care delivered in paediatric heart failure centres

3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

INTRODUCTION

Left ventricular (LV) non-compaction (LVNC) is an increasingly recognized type of

cardiomyopathy (1) characterized by the presence of an extensive trabeculated

myocardium separated into 2 distinct compacted and non-compacted layers (2, 3). The

American Heart Association classifies it as a genetic cardiomyopathy caused by arrested

myocardial development (2, 4, 5) although on occasion cases have been reported to

develop postnatally in response to alterations of preload (6, 7). The European Society of

Cardiology lists LVNC as an unclassified cardiomyopathy (8).

In childhood LVNC, different phenotypes have been observed including dilated,

hypertrophic, restrictive, isolated (not associated with abnormal physiology) and that

associated with congenital heart disease, or in combination (1, 9). Whereas isolated

LVNC is thought to have a relatively good prognosis, (10), LVNC with a dilated

phenotype is associated with worse short-term outcomes (1, 11).

Long-term outcomes for children with LVNC remain uncertain. In addition, it is unclear

whether outcomes are determined solely by the extent and severity of cardiac

dysfunction, or whether the presence of LVNC confers additional adverse prognostic

information.

We examined late outcomes for children with LVNC enrolled in the National Australian

Childhood Cardiomyopathy Study (NACCS), a population-based, longitudinal cohort

study with long-term follow-up extending to 30 years (12-14)

4

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

METHODS

The NACCS study includes all children in Australia with primary cardiomyopathy who

were diagnosed at 0 - <10 years of age, between 1 January 1987 and 31 December 1996.

Local institutional ethics committee approval was obtained from participating centres.

The methodology has been detailed previously (12, 15, 16). In brief, study subjects were

enrolled through site visits undertaken from 1997-2007 by the same three investigators

visiting all 9 pediatric cardiac centers and an additional 12 hospitals caring for children

with cardiac conditions (Appendix 1). The data that support the findings of this study are

available from the corresponding author upon reasonable request.

Study subjects were identified from multiple sources, including local cardiology

databases, echocardiographic records and ICD-9 CM medical record codes. Children with

structural heart disease and progressive neuromuscular disorders were excluded.

Examination of centralized records compiled by the Australian Bureau of Statistics

indicated that there were no children with sudden death as their initial symptom who

were diagnosed at autopsy. Ethics committee approval was obtained from each

participating institution.

Prospective follow-up was arranged for any subjects who were not being followed

regularly. Study questionnaires were used to extract standardized clinical and

echocardiographic data obtained during follow-up.

Definitions

Cardiomyopathies were classified during the course of site visits according to the existing

World Health Organization cardiomyopathy classification (5) by a single observer after

reviewing all relevant investigations, including all available cardiac imaging(12).

Study subjects were classified based on echocardiographic findings as having LVNC if

there was the characteristic morphological appearance comprising: 1) multiple

trabeculations, 2) deep intertrabecular recesses seen on color flow and 3) a 2-layered

structure of the myocardium with ratio of non-compacted to compacted myocardium of

5

1

23

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

>2:1 in systole (17). An experienced pediatric pathologist with cardiac expertise

reviewed available pathological specimens, including cardiac histology.

A family history of cardiomyopathy was defined by the presence of at least one affected

first- or second-degree relative with LVNC identified from case notes, or from familial

screening.

The presence of congestive heart failure was based on signs and symptoms recorded by

the attending physician. Sudden death was defined as a sudden and unexpected death in a

child within 4 hours of new symptoms (14).

Serial echocardiograms at presentation, those closest to 3, 6, 12 months after diagnosis

and then at yearly intervals during follow-up were reviewed by a single observer and

measurements of LV dimensions, wall thickness and fractional shortening (FS) in those

without regional wall motion abnormalities were expressed as Z-scores based on body

surface area or age in the case of FS (18, 19).

LVNC with dilated pathophysiology (LVNC-D) was defined if the LV was dilated with

reduced systolic function (FS Z-score < -2.0 and/or LV ejection fraction (EF) < 45%).

Similar echocardiographic criteria were used to classify DCM subjects in the absence of

LVNC (13). Restrictive pathophysiology in LVNC subjects was diagnosed from

echocardiography and/or cardiac catheterization in subjects with impaired diastolic

function and normal LV size with preserved systolic function (12). A hypertrophic

phenotype was defined by otherwise unexplained LV free wall or septal hypertrophy

(diastolic wall thickness Z score >2) (15).

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics Version 23 and R

statistical package version 3.3.1. Incidence rates were calculated from age-specific

population at risk between 1987-97 (20, 21). Continuous variables were described as

median (interquartile interval) and compared, where relevant, using the Wilcoxon sum-

6

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

rank test. Categorical variables were described through proportions and compared using

Fisher’s exact test.

Analysis of long-term outcome

Our primary endpoint was the combined endpoint of death or transplantation from the

time of presentation. The endpoints of death and transplantation were combined into a

composite primary endpoint as cardiac transplantation was not widely utilized in the early

part of the study.

The Kaplan-Meier method was used to estimate long-term transplantation-free survival

probabilities, and the log-rank test was used for unadjusted comparison of these between

LVNC-D and DCM subjects.

Normalization of LV function in LVNC-D subjects was considered to have occurred on

the first occasion when the LV end-diastolic dimension (LVEDD) Z score was < 2.0 and

the LV FS Z score was > -2.0. Cumulative incidence curves were constructed for the

competing events of 1) normalization of LV function, 2) death or transplantation before

normalization and 3) remaining alive, transplant-free and without normalization. The

cumulative incidence curves were constructed so as to represent the first-encountered

competing event.

Due to the relatively low case numbers, only univariable Cox regression analysis was

performed to examine the association between variables and the combined outcome of

death or transplantation. Variables analyzed in a univariate fashion were: age at

presentation, gender, body surface area, presentation with congestive heart failure,

consanguinity, family history of cardiomyopathy, baseline LVEDD Z-score and baseline

FS Z-score.

The current value and the current rate of change of echocardiographic measurements

(LVEDD Z score and FS Z-score) were both analyzed as time-dependent variables. The

current rate of change was approximated as the difference between the two most recent

7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

measurements before the event divided by the time between the measurements.

Normalization of LV function was also analyzed as a time-dependent variable.

Propensity-score analysis

Cases with missing values were excluded from regression and propensity-score analyses.

Two separate propensity-score methods were used to risk-adjust for differences in clinical

profiles between LVNC-D and DCM subjects: Propensity-score matching and inverse

probability of treatment weighting. A propensity-score was generated for each case by

performing a logistic regression with LVNC-D binary indicator as the dependent

variable. Baseline clinical and investigative co-variables were used. These were: age at

presentation, gender, year of presentation, body surface area, congestive heart failure at

diagnosis, consanguinity, family history of cardiomyopathy, LVEDD Z-score and FS Z-

score on the first available echocardiogram.

For propensity-score matching, subjects were matched 1 to 1 on their propensity-score

without replacement using the greedy matching method with a caliper width equal to 0.2

of the standard deviation of the logit of the propensity-score. In the matched sample,

paired t-tests were used for continuous data, whereas McNemar’s test, which compares

discordance of 2 dichotomous outcomes, was used to compare categorical variables.

Similarly, long-term freedom from death/transplantation was compared using the test

proposed by Klein and Moeschberger (22).

For the propensity-score inverse probability of treatment weighting analysis (IPTW), the

contribution of each individual to Kaplan-Meier estimates or Cox regression is weighted

by the inverse of the probability that they belong to the corresponding group, which is

estimated from the propensity scores obtained from the aforementioned logistic

regression. Intuitively, these weights generate a pseudo-population in which both groups

to be compared are balanced with respect to the measured confounders. This enables the

comparison of these groups and estimation of the population-average (marginal) effect of

LVNC-D on the combined endpoint of death or transplantation, expressed as a hazard

ratio. The analysis was performed according to best practice recommendations. (23)

8

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

RESULTS

Incidence rate of LVNC

During the 10-year period there were 29 newly diagnosed cases of LVNC, comprising

9.2% of the NACCS study population. The mean annual at-risk population (number of

Australian children aged 0-10 years) during this period was 2,532,400 (20, 21) leading to

a mean annual incidence of newly diagnosed cases of 0.11 (95% CI 0.08-0.16) per

100,000 persons at risk. Table 1 shows mean annual incidence according to age at

presentation. The highest incidence was in the first year of life (0.83 (0.51-1.30) per

100,000) representing a more than 5-fold increase compared to subjects aged >1 year at

time of presentation. During this same time period there were 175 newly diagnosed cases

of DCM who did not have LVNC (ref 13)

Presentation and clinical characteristics

Table 2 shows demographics, clinical characteristics at presentation and principal

echocardiographic findings of those diagnosed with LVNC. Of the 29 cases, 27 had

dilated phenotype. One subject had restrictive phenotype and one subject with Barth

syndrome had isolated LVNC without associated cardiac dysfunction. Twenty of 29

(69%) of subjects were males. Associated anomalies in 9 subjects included Barth

syndrome in 7, left bronchial stenosis in 1 and bile duct hypoplasia requiring liver

transplantation in the remaining subject. Two first cousins, both with consanguineous

parents, were found to have a truncating mutation in the ALPK3 gene and one subject

with Barth syndrome had a Complex I respiratory chain enzyme deficiency.

Echocardiography was available in all 29 cases. Additional confirmation of LVNC was

available from characteristic findings on left ventricular angiography in 12 subjects and

direct examination of the heart (explant or autopsy) in 11.

The most common symptom at diagnosis was congestive cardiac failure in 24 (83%)

cases; two subjects were diagnosed on routine family screening, including 1 with Barth

syndrome who has never developed symptoms. The remaining 3 subjects had a diagnosis

of LVNC made on the basis of feeding difficulties, an abnormal chest X-ray and onset of

9

12

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

supraventricular tachycardia in a premature infant. Further clinical details are presented

in Table 2.

Light microscopic findings from autopsy, explantation at the time of transplantation or

endomyocardial biopsy were abnormal in 19/21 available cases and included myocyte

hypertrophy, nuclear irregularity and endocardial fibroelastosis.

Long-term outcomes

There were no subjects lost to follow-up. The median (interquartile interval) duration of

follow-up was 6.8 (0.7-21.7) years for all cases and 21.4 (11.9-24.5) years for survivors.

Of the 27 subjects with dilated phenotype, 19 died or underwent heart transplantation

(including 5 cases of sudden death). The subject with restrictive physiology died during

follow-up while the subject with Barth syndrome and isolated LVNC remains alive with

normal cardiac function 24 years after diagnosis. The two related subjects with an

ALPK3 mutations progressed from a dilated to a hypertrophic phenotype.

Normalization of left ventricular systolic function occurred in 8 of 29 (28%) subjects, all

with a dilated phenotype, including of 4 of 7 subjects (57%) with Barth syndrome and

both subjects who developed a hypertrophic phenotype. Two of 8 subjects who had

normalization of left ventricular function later died suddenly, two redeveloped a dilated

phenotype with one undergoing cardiac transplantation 14 years later. Two have normal

LV size but with reduced function at late follow-up, while the remaining two continue to

have normal LV size and systolic function at late follow-up.

Freedom from death or transplantation was 69% (95% CI 49- 83%) at 1 year, 52% (33-

68%) at 5 years, 48% (30-65%) at 10 years and 45% (27-63%) at 15 years. Figure 1

shows survival for all cases and Figure 2 shows the cumulative proportion of cases

experiencing each competing endpoint after presentation, for all 29 cases. At 15 years

after diagnosis, 21% of LVNC cases were alive with normal LV function while 31% were

alive without transplantation or normalization of LV function.

10

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

Of 8 subjects in the NACCS registry with Barth syndrome, 7 had LVNC. Of these, 2

died (both with sudden death), 1 underwent transplantation, 2 have normal LV systolic

function at latest follow-up, and 2 are alive with a persisting dilated phenotype. There

was a nominal association towards Barth syndrome patients having better survival at 15

years (Barth: 71% (39-100%) vs. non-Barth: 36% (16-57%), p=0.08).

Table 3 shows the association between clinical factors and the combined endpoint of

death or transplantation determined by univariable Cox regression. Favorable prognostic

variables comprised male gender, presence of a positive family history and a higher FS Z

score at any time during follow-up. A larger LVEDD Z score during follow-up was

associated with worse prognosis. Familial LVNC and male gender were unrelated to

survival in the absence of Barth syndrome.

Of the 29 cases of LVNC, long-term therapy with an angiotensin converting enzyme

inhibitor (ACEI) was used in 13, a beta-blocker in 6 and either an ACEI or beta-blocker

was used in 16.

Comparison between LVNC-D and DCM

Compared to the 175 cases with isolated DCM, cases with LVNC-D were more likely to

be male, and of lower weight at presentation. Baseline echocardiographic measurements

were similar (Table 4). Unadjusted freedom from death or transplantation was similar

between LVNC-D and DCM groups (Figure 3A).

Propensity-score matching yielded 24 matched pairs. The clinical and echocardiographic

characteristics amongst matched cases are presented in Table 5. Amongst matched pairs,

there was a nominal association (p=0.08) for subjects with LVNC to have worse

transplant-free survival compared to those with DCM at 15 years as shown in Figure 3B.

Long-term therapy with an ACEI or a beta-blocker was used in 12 of 24 (50%) matched

LVNC-D cases, and 11 the 24 (46%) of the matched DCM cases.

11

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

Propensity-score analysis using inverse probability of treatment weighted Cox regression

showed that LVNC was associated with a greater risk of death or transplantation (HR 2.3,

95% CI 1.4-3.8, p=0.0012) (Figure 3C).

12

1

2

3

4

DISCUSSION

The NACCS uniquely represents the longest and most complete longitudinal cohort study

of childhood cardiomyopathy with a median follow-up of survivors of nearly 25 years.

Ongoing follow-up of this cohort has provided novel information for long-term outcomes

for childhood cardiomyopathy (13, 14).

Our population-based study defines the incidence, presentation and long-term outcomes

for LVNC diagnosed during childhood. In particular, long-term outcomes for children

with LVNC associated with a dilated cardiomyopathy phenotype were worse than those

for matched children with DCM.

Incidence

Previously considered an uncommon cardiomyopathy type, LVNC accounted for nearly

10% of all childhood cardiomyopathy cases in the NACCS study. The incidence of newly

diagnosed cases in our study was 0.11/100,000 at risk subjects per year, with the highest

incidence being during the first year of life, similar to children with dilated and

hypertrophic cardiomyopathy from the NACCS registry. The overall proportion of cases

with LVNC in our study was similar to a single-center study from Texas Children’s

Hospital (24), and somewhat greater than the 4.8% reported from the North American

Pediatric Cardiomyopathy Registry (PCMR) (1). In another study from the UK, 3% of

children aged <16 years with myocardial disease resulting in new onset heart failure were

diagnosed with LVNC (25). The higher proportion of LVNC subjects in our study may

reflect systematic case classification by a single observer.

Inheritance

Multiple modes of LVNC inheritance have been described, X-linked recessive or

autosomal dominant are most common, while autosomal recessive and mitochondrial

inheritance have also been described. Our study provides further evidence for a genetic

basis for LVNC, with 31% of subjects having a positive family history of

cardiomyopathy, which is comparable to published reports of 16-44% (24, 26-29). A

13

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

significant proportion of these subjects had Barth syndrome, accounting for the male

preponderance.

Long-term outcomes for LVNC

In our study, long-term outcome was worse than previously described in some other

series, with only 45% of subjects alive and transplant-free 15 years after presentation.

Our findings contrast with a 10-year transplant-free survival of 60-86% in studies from

Toronto and Texas (24, 27), and 93% in a study from Japan (26). The higher mortality in

our population-based study most likely reflects a high proportion of sick, young infants,

many of whom died within one year after presentation. The low mortality in the Japanese

study may conversely reflect a high proportion of asymptomatic cases as a result of a

systematic childhood screening program in Japan.

Rates of death or transplantation for subjects in our study were also higher than those

reported by the PCMR (1). The PCMR study comprised children with variable LVNC

phenotypes, including a greater proportion without cardiac dysfunction. Outcomes for

children with LVNC and dilated physiology in the PCMR were also poor with a hazard

ration for death or transplantation of >6 compared to subjects with isolated LVNC.

In our study, risk factors on univariable regression analysis for death/transplantation

comprised females, sporadic LVNC (non-familial) and worse LV systolic function. The

sole case with restrictive cardiomyopathy died. As in other studies, severity of systolic

dysfunction was the most important predictor of survival at any time during follow-up.

The association towards better survival in Barth syndrome patients is unexpected. We

postulate that this may be related to more systematic screening and earlier diagnosis of

LVNC given the presence of an underlying syndrome.

Comparison with dilated cardiomyopathy

The 15-year transplant-free survival of 45% for LVNC subjects is worse than the 56%

20-year survival reported from the NACCS for children with dilated cardiomyopathy

14

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

(13). Propensity-score analyses in this study showed that subjects with LVNC and dilated

physiology tended to experience worse survival than matched subjects with dilated

cardiomyopathy, and that LVNC-D was associated with a two-fold higher risk of death or

transplantation.

The frequency of utilization of remodeling therapy was similar in both the LVNC-D and

DCM groups, suggesting that better outcomes for the matched DCM were not due to the

use of remodeling heart failure therapy. However our study was not designed to examine

the impact of any particular heart failure therapy.

Our findings contrast with studies of adult subjects with DCM and LVNC, such as

Amzulescu and colleagues (30), who examined 162 subjects with dilated cardiomyopathy

undergoing cardiac MRI. Those who met MRI criteria for LVNC had outcomes similar to

those of remaining subjects, leading the authors to argue against a non-compaction

phenotype being a more severe form of dilated cardiomyopathy. These contrasting

findings most likely reflect a different spectrum of etiologies for both conditions in adult

subjects.

Higher rates of death or transplantation in children with LVNC, when compared to those

with isolated DCM, may reflect a different genetic basis and in some cases the absence of

a reversible etiology such a lymphocytic myocarditis, which is present in a significant

proportion of children with DCM (12, 13). We also speculate that the spectrum of genetic

etiologies responsible for childhood LVNC may differ from those of affected adults,

leading to a more severe clinical course for those who develop heart failure at a young

age. Rates of sudden death in LVNC subjects were comparatively high and an undulating

phenotype or fluctuating LV systolic function (24) were observed in a significant

proportion of subjects. Taken together, these findings support consideration of LVNC as

a distinct cardiomyopathy phenotype in pediatric patients.

The use of circulatory support and adult-based heart failure therapies in children with

dilated cardiomyopathy has increased in recent years. Routine screening combined with

15

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

cascade genetic testing may contribute to earlier diagnosis of cardiomyopathy in at-risk

family members. These developments are most likely responsible for the improved

outcomes noted in children with DCM who are receiving care in specialized pediatric

heart failure centers in the past 3 decades (31). Children with LVNC, who develop

symptoms at an early age and have worse outcomes than those with DCM, might

therefore benefit similarly from specialist care delivered in high volume pediatric centers.

Limitations

Our population-based study had a high proportion of symptomatic young subjects with

dilated physiology. Outcomes from the present study cannot be extrapolated to children

who have LVNC with other phenotypes, and to those without symptoms diagnosed

during routine family screening, or as a result of genetic testing. Genetic and

mitochondrial testing were not routinely available for all subjects during the study period,

and the proportion of cases with one of these etiologies may therefore have been

underestimated. The limited number of cases precluded a multivariable analysis of risk

factors for death or transplantation. Furthermore, propensity-score analyses, as any other

type of analysis, are inherently unable to account for unmeasured variables, including

more sophisticated parameters of left ventricular function.

CONCLUSIONS

In our population-based study, LVNC accounted for 9.2% of all cardiomyopathies

diagnosed in Australia during the first decade of life. Most subjects were diagnosed

during infancy with a dilated phenotype, following symptoms of congestive heart failure.

Survival free from transplantation at 15 years after diagnosis was 45%. Transplant-free

survival was worse for subjects with the greatest degree of left ventricular systolic

dysfunction at any time during follow-up, and subjects with LVNC and a dilated

phenotype had worse outcomes than matched subjects with DCM.

Funding sources

Supported by Grant 98001 from Royal Children’s Hospital Research Foundation, Grants

G98M0159, G04M1586, G05M2151, G07M3180 from National Heart Foundation of

16

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

Australia, NACCS grant from Australia and New Zealand Children’s Heart Research

Centre. Dr William Shi is supported by the Royal Australasian College of Surgeons

Foundation for Surgery Peter King/Heart Foundation Research Scholarship in addition to

the University of Melbourne Viola Edith Reid and the RG and AU Meade Scholarships.

Dr Daubeney’s research was supported by the Biomedical Research Unit at the Royal

Brompton Hospital. The Murdoch Children’s Research Institute is supported by the

Victorian Government's Operational Infrastructure Support Program.

Disclosures

No conflict of interest disclosure

17

1

2

3

4

5

6

7

8

9

10

Table 1: Incidence of LVNC by age at presentation

Age at presentation (years) 0-<1yrs 1-<2yrs 2-<5yrs 5 -<10 yrs Total

Number of cases during 10-year study period (%) 21 (72.4) 3 (10.3) 4 (13.8) 1 (3.4) 29

Average annual at-risk population 253,954 253,685 761,737 1,262,971 2,532,347

Annual incidence rate/100,000 person-years at risk 0.83 0.12 0.053 0.0079 0.11

18

1

2

3

4

Table 2: Clinical and echocardiographic characteristics of all LVNC subjects

Clinical characteristics n=29

Age – median years (IQR)* 0.3 (0.1-1.3)

Male – n (%) 20 (69)

Year of presentation

1987-1991 17 (59)

1992-1996 12 (41)

Weight – median kg (IQ interval) 5.1 (3.7-9.0)

Dilated phenotype – n (%) 27 (93)

Parental consanguinity†– n (%) 4 (14)

Barth syndrome– n (%) 7 (24)

CHF at diagnosis – n (%) 24 (83)

Family history cardiomyopathy – n (%) 9 (31)

Baseline LVEDD Z-score – median (IQ interval)* 4.6 (0.2-6.0)

Baseline LVESD Z-score – median (IQ interval)* 7.0 (3.3-8.1)

Baseline FS Z-score – median (IQ interval)* -10.7 (-12.1 to -9.1)

Median (IQ interval) duration of follow-up in years 6.8 (0.7 – 24.0)Median (IQ interval) duration of follow-up in survivors in years 24.7 (23.3 – 27.7)

Number of subjects with death/transplant 20

Number of deaths 14

Number of transplants 6*On first available echocardiogram; BSA: body surface area; CHF: congestive heart failure; FS:

fractional shortening; IQ interquartile; LVEDD: left ventricular end-diastolic diameter; LVESD:

left ventricular end-systolic diameter; SD: standard deviation.

†There were 4 subjects with parental consanguinity including 2 siblings

At diagnosis, the median non-compacted-to-compacted myocardial thickness ratio in systole,

measured from echocardiography at the site of maximal involvement, was 2.6 (interquartile

interval 2.1-3.2). In diastole it was 2.3 (2.0-2.5).

Data were incomplete in 2 subjects, who were excluded from propensity-score analyses.

19

1

2

3

4

5

6

7

8

9

10

Table 3: Univariable predictors of death or transplantation after diagnosis of LVNC (n=29)

Variable HR 95% Confidence interval p value

Male* 0.39 0.15-1.00 0.049

Age at presentation (per year) 1.10 0.87-1.40 0.43

Age at presentation (per month) 1.01 0.99-1.03 0.44

Presentation with CHF 1.83 0.54-6.27 0.34

Family history of CMǂ 0.12 0.03-0.53 0.005

Consanguinity 0.61 0.14-2.67 0.51

Barth Syndrome 0.35 0.10-1.21 0.097

Baseline BSA 3.83 0.32-46.54 0.29

Baseline FS Z-score† 0.83 0.67-1.03 0.092

Baseline EDD Z-score† 1.05 0.89-1.24 0.55

LVNC ratio systole 2.33 0.88-6.14 0.088

LVNC ratio diastole 1.70 0.96-1.04 0.071

Current FS Z-score†‡ 0.78 0.66-0.91 0.002

Current EDD Z-score†‡ 1.25 1.02-1.54 0.03

Rate of change in FS Z-score†‡ 1.01 0.96-1.06 0.76* Values after exclusion of Barth syndrome subjects; Male: HR 0.28 (0.03-3.07), p=0.30; Family history: n/a† - per unit Z score‡ Variables treated as time-varying covariates in univariable Cox regression analysis

20

123

4567

Table 4: Comparison of clinical characteristics between LVNC with dilated

phenotype (LVNC-D) and those with isolated DCM.

LVNC-D DCM

Clinical characteristics n=27 n=175 p value

Age – median years (IQ interval) 0.3 (0.07-1.4) 0.6 (0.2-1.7) 0.26

Male – n (%) 19 (70) 77 (44) 0.013

Year of presentation

1987-1991 16 (59) 74 (42) -

1992-1996 11 (41) 101 (58) 0.14

Weight – median kg (IQ interval) 5.1 (3.6-9.0) 7.1 (4.3-11.3) 0.047

BSA (m2) – median (IQ interval) 0.3 (0.2-0.5) 0.4 (0.3-0.5) 0.061

Consanguinity 4 (15) 14 (8) 0.27

CHF at diagnosis 22 (81) 163 (93) 0.058

Family history of cardiomyopathy 8 (30) 26 (15) 0.092

Cardiothoracic ratio on CXR 66 (60-72) 65 (61-70) 0.75

Echocardiography

LVEDD Z-score – median (IQ interval) 4.8 (0.5-6.0) 4.4 (2.5-6.3) 0.50

LVESD Z-score – median (IQ interval) 7.2 (3.9-8.2) 6.5 (4.5-8.4) 0.75

FS Z-score – median (IQ interval) -11.2 (-12.1 to -9.9) -10.5 (-12.2 to -8.7) 0.46

BSA: body surface area; CHF: congestive heart failure; FS: fractional shortening; IQ

interquartile; LVEDD: left ventricular end diastolic diameter; LVESD: left ventricular

end systolic diameter; SD: standard deviation.

Data were missing in 2 LVNC subjects and 33 DCM subjects and these cases were

excluded from propensity-score analysis.

21

1

2

3

4

5

6

7

8

Table 5: Clinical characteristics of 24 propensity-score matched pairs.

LVNC-D DCM

Clinical characteristics n=24 n=24 p value

Age – median years (IQR) 0.3 (0.2-1.8) 0.5 (0.3-1.1) 0.88

Male – n (%) 17 (71) 16 (67) >0.99

Year of presentation

1987-1991 14 (58) 14 (58) -

1992-1996 10 (42) 10 (42) >0.99

Weight – median kg (IQR) 5.4 (4.0-9.5) 7.7 (5.6-9.5) 0.54

BSA (m2) – median (IQR) 0.30 (0.25-0.46) 0.39 (0.30-0.47) 0.55

Consanguinity 3 (13) 4 (17) >0.99

Presentation with CHF 18 (75) 19 (79) >0.99

Family history cardiomyopathy 8 (33) 7 (29) >0.99

Echocardiography

LVEDD Z-score – median (IQR) 4.8 (0.5-6.0) 4.0 (1.9-5.2) 0.86

LVESD Z-score – median (IQR) 7.2 (3.9-8.2) 6.1 (4.3-8.0) 0.92

FS Z-score – median (IQR) -11.2 (-12.1 to -9.9) -9.7 (-12.7 to -8.4) 0.48

BSA: body surface area; CHF: congestive heart failure; FS: fractional shortening;

LVEDD: left ventricular end-diastolic diameter; LVESD: left ventricular end-systolic

diameter; The c-statistic of the propensity-score model was 0.70.

22

1

2

3

4

5

FIGURE 1: Long-term freedom from death or transplantation of all 29 LVNC subjects

FIGURE 2: Cumulative incidence curve of competing outcomes after diagnosis of LVNC (n=29). Competing events: Normalization of LV function (blue), and death or transplantation before normalization (black). The probability of remaining alive, transplant-free and without normalization is also shown (red).

23

1

2

34567

89

10

1112

FIGURE 3: Comparison of probability of freedom from death/ transplantation at 15 years between subjects with LVNC-D (n=27) and DCM (n=175): A) Unadjusted (full sample); B) Adjusted using propensity-score matching (matched pairs); C) Adjusted using inverse probability of treatment weighted (full sample).

24

12345

67

REFERENCES

1. Jefferies JL., Wilkinson JD., Sleeper LA, Colan SD, Lu M, Pahl E, et al. Cardiomyopathy

Phenotypes and Outcomes for Children With Left Ventricular Myocardial Noncompaction:

Results From the Pediatric Cardiomyopathy Registry. J Card Fail. 2015;21(11):877-84.

2. Freedom RM., Yoo SJ, Perrin D, Taylor G, Petersen S, Anderson RH. The morphological

spectrum of ventricular noncompaction. Cardiol Young. 2005;15(4):345-64.

3. Lipshultz SE, Cochran TR, Briston DA., Brown SR, Sambatakos PJ, Miller TL., et al.

Pediatric cardiomyopathies: causes, epidemiology, clinical course, preventive strategies and

therapies. Future Cardiol. 2013;9(6):817-48.

4. Jenni R., Oechslin E. N., van der Loo B. Isolated ventricular non-compaction of the

myocardium in adults. Heart. 2007;93(1):11-5.

5. Maron BJ., Towbin JA., Thiene G, Antzelevitch C, Corrado D, Arnett D, et al.

Contemporary definitions and classification of the cardiomyopathies: an American Heart

Association Scientific Statement from the Council on Clinical Cardiology, Heart Failure and

Transplantation Committee; Quality of Care and Outcomes Research and Functional Genomics

and Translational Biology Interdisciplinary Working Groups; and Council on Epidemiology and

Prevention. Circulation. 2006;113(14):1807-16.

6. Finsterer J, Stöllberger C, Wegmann R, Janssend LA. Acquired left ventricular

hypertrabeculation/noncompaction in myotonic dystrophy type 1. Int J Cardiol. 2009;137(3):310-

13.

7. Gati S., Papadakis M., Papamichael N. D., Zaidi A., Sheikh N., Reed M., et al. Reversible

de novo left ventricular trabeculations in pregnant women: implications for the diagnosis of left

ventricular noncompaction in low-risk populations. Circulation. 2014;130(6):475-83.

8. Elliott P, Andersson B, Arbustini E, Bilinska Z, Cecchi F, Charron P, et al. Classification

of the cardiomyopathies: a position statement from the European Society Of Cardiology Working

Group on Myocardial and Pericardial Diseases. Eur Heart J. 2008;29:270-6.

9. Towbin JA. Left ventricular noncompaction: a new form of heart failure. Heart Fail Clin.

2010;6(4):453-69, viii.

25

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

10. Towbin JA, Lorts A, Jefferies JL. Left ventricular non-compaction cardiomyopathy. The

Lancet. 2015;386(9995):813-25.

11. Brescia ST., Rossano JW., Pignatelli R, Jefferies JL, Price JF, Decker JA., et al. Mortality

and sudden death in pediatric left ventricular noncompaction in a tertiary referral center.

Circulation. 2013;127(22):2202-8.

12. Nugent AW, Daubeney PE., Chondros P., Carlin JB., Cheung M., Wilkinson LC., et al.

The epidemiology of childhood cardiomyopathy in Australia. N Engl J Med. 2003;348(17):1639-

46.

13. Alexander PM., Daubeney PE., Nugent AW., Lee KJ, Turner C, Colan SD, et al. Long-

term outcomes of dilated cardiomyopathy diagnosed during childhood: results from a national

population-based study of childhood cardiomyopathy. Circulation. 2013;128(18):2039-46.

14. Bharucha T, Lee KJ., Daubeney PE., Nugent AW, Turner C, Sholler GF, et al. Sudden

death in childhood cardiomyopathy: results from a long-term national population-based study. J

Am Coll Cardiol. 2015;65(21):2302-10.

15. Nugent AW, Daubeney PE, Chondros P, Carlin JB., Colan SD., Cheung M., et al.

Clinical features and outcomes of childhood hypertrophic cardiomyopathy: results from a

national population-based study. Circulation. 2005;112(9):1332-8.

16. Daubeney PE, Nugent AW, Chondros P., Carlin JB., Colan SD, Cheung M., et al.

Clinical features and outcomes of childhood dilated cardiomyopathy: results from a national

population-based study. Circulation. 2006;114(24):2671-8.

17. Jenni R, Oechslin E, Schneider J, Attenhofer JC., Kaufmann PA. Echocardiographic and

pathoanatomical characteristics of isolated left ventricular non-compaction: a step towards

classification as a distinct cardiomyopathy. Heart. 2001;86(6):666-71.

18. Sluysmans T., Colan S. D. Theoretical and empirical derivation of cardiovascular

allometric relationships in children. J Appl Physiol. 2005;99(2):445-57.

19. Colan SD, Parness IA., Spevak PJ., Sanders SP. Developmental modulation of

myocardial mechanics: age- and growth-related alterations in afterload and contractility. J Am

Coll Cardiol. 1992;19(3):619-29.

26

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

20. Estimated resident population by sex and age, states and territories of Australia, June

1987 to June 1992. Australian Bureau of Statistics. Canberra: 1993 (catalog no. 3201.0).

21. Population by age and sex, Australian states and territories, June 1992 to June 1997.

Australian Bureau of Statistics. Canberra: 1997 (catalog no. 3201.0).

22. Austin PC. Propensity-score matching in the cardiovascular surgery literature from 2004

to 2006: a systematic review and suggestions for improvement. J Thorac Cardiovasc Surg.

2007;134:1128-35.

23. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of

treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in

observational studies. Stat Med. 2015;34:3661-79.

24. Pignatelli RH, McMahon CJ, Dreyer WJ, Denfield SW, Price J, Belmont JW, et al.

Clinical characterization of left ventricular noncompaction in children: a relatively common form

of cardiomyopathy. Circulation. 2003;108(21):2672-8.

25. Andrews RE., Fenton MJ., Ridout DA., Burch M. New-onset heart failure due to heart

muscle disease in childhood: a prospective study in the United kingdom and Ireland. Circulation.

2008;117(1):79-84.

26. Ichida F, Hamamichi Y, Miyawaki T, Ono Y, Kamiya T, Akagi T, et al. Clinical features

of isolated noncompaction of the ventricular myocardium: long-term clinical course,

hemodynamic properties, and genetic background. J Am Coll Cardiol. 1999;34(1):233-40.

27. Wald R, Veldtman G., Golding F., Kirsh J, McCrindle BW, Benson L. Determinants of

outcome in isolated ventricular noncompaction in childhood. Am J Cardiol. 2004;94(12):1581-4.

28. McMahon CJ., Pignatelli R. H., Nagueh S. F., Lee V. V., Vaughn W., Valdes S. O., et al.

Left ventricular non-compaction cardiomyopathy in children: characterisation of clinical status

using tissue Doppler-derived indices of left ventricular diastolic relaxation. Heart.

2007;93(6):676-81.

29. Saleeb SF, Margossian R, Spencer CT, Alexander ME, Smoot LB, Dorfman AL, et al.

Reproducibility of echocardiographic diagnosis of left ventricular noncompaction. J Am Soc

Echocardiogr. 2012;25:194-202.

27

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

30. Amzulescu MS, Rousseau MF, Ahn SA, Boileau L, de Meester de Ravenstein C,

Vancraeynest D, et al. Prognostic Impact of Hypertrabeculation and Noncompaction Phenotype in

Dilated Cardiomyopathy: A CMR Study. JACC Cardiovasc Imaging 2015;8:934-46.

31. Singh RK, Canter CE, Shi L, Colan SD, Dodd DA, Everitt MD, et al. Survival Without

Cardiac Transplantation Among Children With Dilated Cardiomyopathy. J Am Coll Cardiol

2017;70:2663-73.

28

1

2

3

4

5

6

7

8

APPENDIX

The National Australian Childhood Cardiomyopathy Study comprises the following

medical institutions and physicians:

- The Princess Margaret Hospital for Children, Perth:

o Dr L D’Orsogna, Dr A Bullock, Dr J Ramsey, Dr D Kothari

- The Women’s and Children’s Hospital Adelaide:

o Dr T Robinson, Dr M Richardson, Dr G Wheaton

- The Mater Children’s Hospital, Brisbane:

o Dr D Radford, Dr C Whight, Dr R Justo, Dr C Ward

- The Children’s Hospital at Westmead, Sydney:

o Dr G Sholler, Dr R Hawker, Dr S Cooper, Dr K Lau, Dr M Sherwood

- The Sydney Children’s Hospital:

o Dr O Jones,

- The John Hunter Children’s Hospital, Newcastle:

o Dr G Warner

- The Royal Children’s Hospital, Melbourne:

o Professor J L Wilkinson, Dr T Goh, Dr B Edis, Professor D Penny, Dr G

Lane

- The Monash Medical Centre, Melbourne:

o Professor S Menahem, Dr L Fong

- Other participating physicians include Dr C Semsarian, Dr A Gailbraith, Dr R

Jeremy, Dr R Fryda, Dr P Robinson and Dr L Lee

29

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24