- Email: [email protected]
Potential for and timing of recovery in children with dilated cardiomyopathy
International Journal of Cardiology 266 (2018) 162–166
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Potential for and timing of recovery in children with dilated cardiomyopathy Matthew J. Fenton ⁎, Philippa Horne, Jacob Simmonds, Sophie L. Neligan, Rachel E. Andrews, Michael Burch Cardiothoracic Unit, Great Ormond Street Hospital for Children NHS Foundation Trust, Great Ormond Street, London WC1N 3JH, UK
a r t i c l e
i n f o
Article history: Received 8 March 2017 Received in revised form 10 December 2017 Accepted 20 December 2017
Keywords: Pediatric Heart failure Heart transplant Morbidity Mortality
a b s t r a c t Objective: Understanding the clinical course and time-frame for recovery is helpful to guide management and counselling following a diagnosis of Dilated Cardiomyopathy (DCM). We aimed to document outcomes and time to recovery for a cohort of patients with a dilated cardiomyopathy phenotype. Methods: An observational cohort methodology was used to collect retrospective data from the departmental database for those identified with DCM. Data relating to mode of presentation, echocardiographic parameters, clinical management and outcome were collated and analysed. Predictors and time-scale for recovery were investigated and reported. Results: 209 new referrals were included within the time frame. 82 children median age 1.0 years (IQR 3.4) required intensive care (ICU) and their survival without death or transplant was 51% to one year and 45% to five years. 127 children presented to the pediatric heart failure clinic. Excluding 58 with neuromuscular disease, median age was 4.1 years (IQR 11.3) & survival without death or transplant 85% to 1 year and 50% to 5 years. NT-proBNP normalized in survivors before echocardiographic parameters. Predictors of recovery included younger age, female sex and smaller left ventricular end diastolic Z score on echocardiogram at presentation. Conclusion: Transplant-free survival to one year is significantly better for patients presenting to clinic, but longer-term survival is better amongst those presenting to ICU due to a late attrition in those with less severe heart failure at presentation. Falling NT-proBNP is the earliest marker of recovery. Recovery of cardiac function remains possible up to three years from presentation. © 2018 Elsevier B.V. All rights reserved.
1. Introduction Our understanding of the epidemiology of pediatric DCM is well served with good databases from the United States of America and Australia [1,2]. However, while there is striking similarity in incidence, despite variation in inclusion criteria, the data on the rates of death, transplant and recovery of function are discrepant between these 2 major registries. The highest risk period in both datasets is the first year after presentation and transplant-free survival is similar at 74% and 69% respectively [3,4]. However for children requiring hospitalization for a first episode of heart failure, a single center study from America has shown transplant-free survival to be only 46% at 1 year, whereas registry data from the UK showed a transplant-free 1 year survival of 66% [5,6]. While congestive heart failure is common in pediatric DCM and its presence increases the hazard for death, the severity has not been specifically addressed as a risk factor
⁎ Corresponding author. E-mail address: [email protected] (M.J. Fenton).
https://doi.org/10.1016/j.ijcard.2017.12.075 0167-5273/© 2018 Elsevier B.V. All rights reserved.
[4]. Recent Dutch registry data has shown transplant-free 1 year survival of 82% and suggested rather controversially “a conservative approach [to transplant listing], specifically in young children presenting with heart failure and requiring ICU admission, but not mechanical circulatory support, is justifiable” [7]. We aimed to provide further data on outcomes based on severity of presentation by using a surrogate evaluation relating to mode of presentation. We hypothesised that children presenting to an ambulatory clinic or low dependency inpatient setting versus those decompensated to the intensive care would have different outcomes in their clinical progress in the short and medium term from presentation. We were also interested in determining the outcomes for those patients who recovered following their initial presentation and whether any improvement was sustained. This detailed information from a single centre would be useful for pediatric cardiologists to be aware of when faced with families and children with new onset DCM. Therefore, we reviewed the data from a large pediatric heart failure program and compared the outcomes for children admitted to the ICU with those who presented less acutely, over a 5 year period. For those that survived without a transplant we investigated the proportion and timescale for recovery of clinical parameters with particular attention
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
to the role that routine measurement of NT-proBNP might add to our predictive ability. 2. Materials and methods Great Ormond Street Hospital is a tertiary and quaternary pediatric referral center accepting patients from district general hospitals in the South East of England, and from other pediatric cardiac units across the UK. This study complies with the declaration of Helsinki and received approval from the local research governance department. We evaluated all children ≤16 years with primary or secondary DCM who were referred to our unit between January 2008 and July 2013 inclusive. Follow up data were included if individuals had a minimum of two episodes of care. Follow up was censored at the most recent known point of follow up. A DCM phenotype was diagnosed on 2 dimensional echocardiography. All patients included in this study had a fractional shortening (FS) measurement of ≤26% at presentation and a left ventricular end diastolic dimension (LVEDD) Z score of N+2. Patients were divided into two groups according to severity of presenting features; those who required admission to the cardiac ICU, and those who were managed in the outpatient clinic or a low dependency ward setting. For all patients in addition to the echocardiographic measurements above we recorded the NYHA/Ross score, NT-proBNP levels, and for ward/clinic patients' medications used. Between-group comparisons were made using unpaired Student t-test or Mann–Whitney U test as appropriate to the underlying distribution of the variables. A curve demonstrating the proportions of patients who died, were transplanted or remained alive without transplant over time was created to graphically demonstrate the patients' outcomes. Survival curves using Kaplan–Meier methodology censored for patients with a combined end point of either death or transplant were created and censored at the time of most recent follow up. Survival outcomes comparing patients who presented to intensive care versus those who presented to a low dependency or ward environment were compared using the log-rank test. A series of univariate Cox regression analyses were then performed to identify the risk of either death or transplant attributable to individual variables. Variables with a p value b0.1 were included in the subsequent multivariate analysis. To provide clinical relevance a change in NT-proBNP of 1000 pg/ml was considered to be clinically relevant with the absolute value divided by 1000. For those children who remained event free, without death or transplant, we evaluated whether the shortening fraction (FS), the ventricular size (LVEDD Z score) or NT-proBNP returned to the normal range. Using time to recovery for each of these variables and defining overall recovery as return to normal of all three clinical parameters, Kaplan–Meier analysis was used to construct an inverse survival curve to graphically represent the probability of recovery and the recovery time for each of the variables as well as the composite end point of all three parameters to provide overall recovery. Cox regression analysis was used to identify individual variables that significantly predict recovery. Variables with a p value b0.1 were carried forward to a subsequent multivariate analysis. Sources of bias were considered due to patients being seen by different clinicians for follow up. Measurement bias on echocardiography was limited by use of a departmental protocol though studies were performed by multiple operators.
3. Results Within the five year timeframe there were 209 new referrals for pediatric patients with a DCM phenotype. Of these 82 children, median age 1.0 year (IQR 3.4), were admitted to the cardiac intensive care unit and 12 (15%) of these were clinically diagnosed with myocarditis. Cardiac biopsy is not routinely undertaken to diagnose myocarditis in the UK and the rationale for this has been described previously [6]. Mechanical support was needed with ECMO or Berlin Heart in 28 (34%) and transplant in 29 (35%). Survival without death or transplant was 51% to one year and 45% to five years. All the patients presenting to intensive care were in NYHA class III or IV and distributed across the NYHA classification for children in low dependency settings. Patient details are included in Table 1. During the same period 127 new referrals were seen in the pediatric heart failure clinic or admitted to a low dependency ward with AHA/ACC heart failure stages B and C. Of these 58 had neuromuscular disease with reduced FS and 3 died within a year of presentation. Neuromuscular patients were excluded from the subsequent analysis. Of the remaining 69 new patients with DCM seen in low dependency settings the median age was 4.1 (IQR 11.2) and 3 (4.3%) were clinically diagnosed with myocarditis. Oral anti-failure medication was used to support cardiac function. The following drugs were used at any time
163
point during follow up. Either angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB) were used in all patients (ACEI in 96%, ARB in 4%), Furosemide in 78%, Spironolactone in 74%, Warfarin in 9%, Aspirin in 54%, Betablocker (Carvedilol) in 65%, and Digoxin in 56%. 4 patients received Ivabradine, 2 patients Carnitine and 1 patient received Sildenafil. Survival without death or transplant for these patients was 85% to 1 year, and was 50% to 5 years. The different transplant-free survival between those children presenting to intensive care and those to a low dependency setting is shown in Fig. 1A. Assumptions for proportional hazards modelling were met with no significance demonstrated due to changes over time. Multivariate modelling for predictors of death or transplant for all patients despite their location of presentation is shown in Table 2. Older age, reduced FS at presentation and increased NT-proBNP level per 1000 unit increase were significant risk factors. LVEDD Z score at baseline was not a significant predictor of outcome. Whilst presenting to the intensive care unit was a significant risk factor for death or heart transplantation as part of univariate analysis there was no contribution to the multivariate model. For the 87 patients from both groups who survived without transplant, the time course of recovery is shown in Fig. 1B. NT-proBNP improved before FS, and LVEDD was the last to improve. Recovery was most rapid in the first year after presentation but continued over the subsequent 3 years before entering a plateau phase indicating recovery of biomedical and echocardiographic parameters in 60% of patients. Table 3 documents the predictors for recovery for children presenting in clinic, with young, female children with a smaller left ventricular end diastolic diameter Z score more likely to recover their parameters to within the normal range. 4. Discussion While it is clear that the first year after presentation poses the highest risk in pediatric DCM, this paper shows clear divergence in 1 year transplant-free survival for children with DCM presenting to the ward or clinic (85%) and those needing initial intensive care (51%) [5,6]. This distinction is useful additional information when counselling families, using the ‘route of care’ as a surrogate for disease severity stratifies the presentation of DCM and provides more granularity regarding prognosis for different modes of presentation. Unlike other cohorts our data includes NT-proBNP which whilst contributing to the model failed to produce a clinically important hazard ratio during univariate or multivariate analysis. However for those patients who recovered the NT-proBNP serial data was clinically useful. NT-proBNP is the first parameter for recovery and when measured serially seems to be an important clinical marker to guide therapy. Further long-term analysis will be useful to answer this question over time. Other predictors of
Table 1 Patient details.
Number presented Mean age at presentation in years (SD) Median age at presentation (IQR) Number of males (%) Mean NT-proBNP (SD) Mean LVEDD Z score (SD) Mean FS % (SD) Number required mechanical support (%) Number transplanted (%) Number died (%) Number in NYHA class I (%) Number in NYHA class II (%) Number in NYHA class III (%) Number in NYHA class IV (%) ⁎ P b 0.05.
ICU
Ward/Clinic
82 3.2 (4.5) 1.0 (3.4) 40 (49%) 19,460 (12,459) 5.3 (4.0) 13.3 (6.2) 28 (34%) 29 (35%) 14 (17%) 0 (0%) 0 (0%) 7 (9%) 75 (91%)
69 6.3 (5.6) 4.1 (11.2) 32 (46%) 5609 (8208) 5.6 (3.5) 18 (8.1) 8 (12%) 12 (17%) 9 (13%) 41 (59%) 12 (17%) 9 (13%) 7 (9%)
P value b0.001* b0.001* 0.87 b0.001* 0.57 b0.001* 0.001* 0.01* 0.49
164
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
data towards the more severe cases. Over a third of those admitted to the ICU received mechanical support. This is different from the Australian data which predated the Berlin Heart era where no patients had mechanical support, the North American data where 5% had mechanical support, and the more recent Dutch data which did incorporate the Berlin Heart era but only 8% had mechanical support [3,4,7]. Therefore our data appears to offer an opportunity to assess these more severely ill children. Rather intriguingly the survival data does show a better longer-term outcome for those who presented to ICU. A possible explanation for this is related to aetiology. Myocarditis is known to have a better outcome than idiopathic DCM, particularly if fulminant, and in our patients myocarditis was more common in the intensive care unit compared to the clinic. (15% vs 4%) [12,13]. This needs to be qualified over a longer period of time once outcomes after heart transplantation enter the follow-up period. Recently the Dutch registry paper suggested that a “conservative approach specifically in young children presenting with heart failure and requiring intensive care admission, but not mechanical circulatory support, is justifiable” [7]. This appears to be supported by transplant registry data, which shows that wait list mortality is low for pediatric DCM with the exception of those on ECMO, ventilation or who have arrhythmias [14]. In our experience the very sick patients still have a high mortality and without mechanical support and transplant this mortality would be greater, but perhaps those who survive ICU should be given adequate time to stabilise or even improve before listing for transplant. Whether this would extend to using Berlin Heart and other ventricular assist devices as bridges to recovery remains unclear. Data from the UK does show that Berlin Heart explantation is possible and often successful in myocarditis, but in DCM it is less likely to be associated with a good longer term outcome [15]. The finding of a higher mortality after transplant for myocarditis compared to DCM further supports a conservative approach to listing for transplant in myocarditis [16]. However, our data suggests that a conservative approach to listing for transplant may be an oversimplification. Data from Europe and America has shown that the only therapy that has improved survival for pediatric DCM is transplantation and that drug therapy has been disappointing [17–20]. There is a late deterioration even in children who have recovered function [9]. Our study shows a steady attrition in patients whose initial presentation was to the ward or clinic and clearly some patients with chronic heart failure will benefit from transplantation. This underscores the need to predict which patients will benefit from listing and detailed echocardiography, exercise test and blood biomarkers may be of help [10,21–23]. Although recovery of function is more common with myocarditis than idiopathic DCM, cases of idiopathic DCM may still recover function. The North American registry data has been re-analysed recently to assess recovery, with myocarditis and neuromuscular disease excluded [9]. In the remaining cohort of idiopathic DCM patients there was a recovery of function in 22% at 2 years. The Dutch data showed 38% of patients recovered function with similar numbers of myocarditis and idiopathic patients, with most recovering within a year [7]. However data from the Australian registry and a small UK study have shown recovery may continue over a long period [24].
Fig. 1. A) Survival curve for patients presenting to the clinic/ward or to intensive care; B) graph depicting recovery of cardiac parameters. FS, fractional shortening; LVEDD, left ventricular diastolic diameter; NT-proBNP, N-terminal pro-brain natriuretic peptide.
transplant-free survival are consistent with and strengthen the existing literature, namely age and FS measurement at presentation [8–11]. Recovery of function to normal occurred in survivors from both ICU and the clinic, and approximately 60% had both a normal echo and NT-proBNP after 3 years. Age and FS again predicted recovery. Another novel finding from our data is that survivors of new onset pediatric DCM do better in the medium term if their initial presentation was to the ICU. Presumably this phenomenon is related to the fact that more of the patients in intensive care will have received mechanical support and subsequent transplant, providing improved medium term survival. Our data demonstrates the continued attrition due to deterioration for children managed in an outpatient setting. The number of patients reported here over a 5-year period is greater than the complete national cohort from Holland over a similar period [7]. One benefit of a large single center study is that the clinical data can be collated in detail and can include consistent echocardiographic and biomarker data such as NT-proBNP. However, as we are a national referral unit for advanced heart failure, this will inevitably skew our Table 2 Analysis of predictors for either death or transplant. Risk factor
Male sex Age at presentation Baseline NT-proBNP Baseline LVEDD Z score Baseline FS% Presentation to ICU
Univariate analysis
Multivariate analysis (events = 47, n = 133)
HR
n
p
HR
Lower 95% confidence limit
Upper 95% CI
1.00 1.04 0.99 1.05 0.93 2.36
151 151 133 151 151 151
0.997 0.05 0.006 0.22 0.001 0.001
– 1.16 1.04 – 0.94 1.32
– 1.09 1.01 – 0.89 0.58
– 1.23 1.07 – 0.98 3.01
p 0.001 0.007 0.008 0.514
HR, hazard ratio; n, number of complete patient datasets included in analysis; NT-proBNP, N-terminal pro-brain natriuretic peptide; LVEDD, left ventricular end diastolic diameter; FS, fractional shortening; CI, confidence interval; ICU, intensive care unit.
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
165
Table 3 Analysis of predictors for recovery in clinic patients. Risk factor
Male sex Age at presentation Baseline NT-proBNP Baseline LVEDD Z score Baseline FS% Presentation to ICU
Univariate analysis
Multivariate analysis (events = 46, n = 86)
HR
n
p
HR
Lower 95% confidence limit
Upper 95% confidence limit
p
0.56 0.94 1.00 0.87 1.03 1.55
87 87 86 87 87 87
0.062 0.081 0.48 0.012 0.142 0.138
0.51 0.9 – 0.83 – –
0.28 0.83 – 0.75 – –
0.96 0.97 – 0.92 – –
0.035 0.005 – 0.001 – –
HR, hazard ratio; n, number of complete patient datasets included in analysis; NT-proBNP, N-terminal pro-brain natriuretic peptide; LVEDD, left ventricular end diastolic diameter; FS, fractional shortening; ICU, intensive care unit.
Our data offers new insight into the sensitivity and timescale for markers of recovery with distinct recovery curves for the 3 studied parameters. No prior data on NT-proBNP and recovery exists although its prognostic use in pediatric heart failure has been shown in America and Europe [22,23]. Our data demonstrates that in those recovering, NT-proBNP is the earliest marker to normalise, followed by FS and finally LVEDD. This emphasizes the importance of measuring serial NT-proBNP levels in these children. This data demonstrates that recovery can occur over a prolonged period of time after presentation and for those patients who remain relatively stable there is the potential for improvement beyond two years from presentation. This data supports prolonged outpatient management and suspending transplant listing in those patients who are clinically well enough to do so. 5. Limitations This is a single center study. Although our data may be helpful to those counselling patients and families it should be applied cautiously to other populations unless or until confirmed by registry data. The number of patients seen in this single centre study is unusually large. It reflects a far wider referral practice than just the city of London area and therefore our data cannot be used for population studies or incidence. Biopsies were not performed routinely but clinical diagnosis of myocarditis is accepted to be equally reliable in many countries, with similar outcomes in clinically suspected and biopsy-proven myocarditis [12]. The data is limited to a 5 year period, but this does reflect the current practice of mechanical support, encompassing the more widespread use of the Berlin Heart in the pediatric population. 6. Conclusions Our data is the first to demonstrate better 1-year outcome for children with DCM presenting with less severe heart failure, and supports more optimistic counselling regarding early outcomes in those presenting to the ward or clinic. However, there is an ongoing attrition in these patients and medium term survival was actually better in those with a more fulminant heart failure who required intensive care support at presentation. Recovery of function is possible after presentation with NT-proBNP demonstrating early improvement prior to echocardiographic parameters and female sex, younger age and a smaller left ventricle end diastolic diameter at presentation being predictive of recovery up to three years after diagnosis. Funding sources This study was supported by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children NHS Foundation Trust and University College London.
Conflict of interest No conflict of interest for any authors.
References [1] S.E. Lipshultz, L.A. Sleeper, J.A. Towbin, et al., The incidence of pediatric cardiomyopathy in two regions of the United States, N. Engl. J. Med. 348 (2003) 1647–1655. [2] A.W. Nugent, P.E.F. Daubeney, P. Chondros, et al., The epidemiology of childhood cardiomyopathy in Australia, N. Engl. J. Med. 348 (2003) 1639–1646. [3] P.E.F. Daubeney, A.W. Nugent, P. Chondros, et al., Clinical features and outcomes of childhood dilated cardiomyopathy: results from a national populationbased study, Circulation 114 (2006) 2671–2678. [4] J.A. Towbin, A.M. Lowe, S.D. Colan, et al., Incidence, causes, and outcomes of dilated cardiomyopathy in children, JAMA 296 (2006) 1867–1876. [5] S.A. Hollander, D. Bernstein, J. Yeh, et al., Outcomes of children following a first hospitalization for dilated cardiomyopathy, Circ. Heart Fail. 5 (2012) 437–443. [6] R.E. Andrews, M.J. Fenton, D.A. Ridout, et al., New-onset heart failure due to heart muscle disease in childhood: a prospective study in the United kingdom and Ireland, Circulation 117 (2008) 79–84. [7] Boer den SL, M. Lennie van Osch-Gevers, G. van Ingen, et al., Management of children with dilated cardiomyopathy in The Netherlands: Implications of a low early transplantation rate, J. Heart Lung Transplant. 34 (2015) 963–969. [8] T.P. Singh, L.A. Sleeper, S. Lipshultz, et al., Association of left ventricular dilation at listing for heart transplant with postlisting and early posttransplant mortality in children with dilated cardiomyopathy, Circ. Heart Fail. 2 (2009) 591–598. [9] M.D. Everitt, L.A. Sleeper, M. Lu, et al., Recovery of echocardiographic function in children with idiopathic dilated cardiomyopathy: results from the pediatric cardiomyopathy registry, J. Am. Coll. Cardiol. 63 (2014) 1405–1413. [10] K.M. Molina, P. Shrader, S.D. Colan, et al., Predictors of disease progression in pediatric dilated cardiomyopathy, Circ. Heart Fail. 6 (2013) 1214–1222. [11] P.M.A. Alexander, P.E.F. Daubeney, A.W. Nugent, et al., Long-term outcomes of dilated cardiomyopathy diagnosed during childhood: results from a national population-based study of childhood cardiomyopathy, Circulation 128 (2013) 2039–2046. [12] S.R. Foerster, C.E. Canter, A. Cinar, et al., Ventricular remodeling and survival are more favorable for myocarditis than for idiopathic dilated cardiomyopathy in childhood: an outcomes study from the Pediatric Cardiomyopathy Registry, Circ. Heart Fail. 3 (2010) 689–697. [13] R.E. McCarthy, J.P. Boehmer, R.H. Hruban, et al., Long-term outcome of fulminant myocarditis as compared with acute (nonfulminant) myocarditis, N. Engl. J. Med. 342 (2000) 690–695. [14] R. Kirk, D. Naftel, T.M. Hoffman, et al., Outcome of pediatric patients with dilated cardiomyopathy listed for transplant: a multi-institutional study, J. Heart Lung Transplant. 28 (2009) 1322–1328. [15] C.A. Irving, D.S. Crossland, S. Haynes, et al., Evolving experience with explantation from Berlin Heart EXCOR ventricular assist device support in children, J. Heart Lung Transplant. 33 (2014) 211–213. [16] B.A. Pietra, P.F. Kantor, H.L. Bartlett, et al., Early predictors of survival to and after heart transplantation in children with dilated cardiomyopathy, Circulation 126 (2012) 1079–1086. [17] A.-M. McMahon, C. Van Doorn, M. Burch, et al., Improved early outcome for end-stage dilated cardiomyopathy in children, J. Thorac. Cardiovasc. Surg. 126 (2003) 1781–1787. [18] A.E. Tsirka, K. Trinkaus, S.-C. Chen, et al., Improved outcomes of pediatric dilated cardiomyopathy with utilization of heart transplantation, J. Am. Coll. Cardiol. 44 (2004) 391–397. [19] R.E. Shaddy, M.M. Boucek, D.T. Hsu, et al., Carvedilol for children and adolescents with heart failure: a randomized controlled trial, JAMA 298 (2007) 1171–1179. [20] P.F. Kantor, J.R. Abraham, A.I. Dipchand, et al., The impact of changing medical therapy on transplantation-free survival in pediatric dilated cardiomyopathy, J. Am. Coll. Cardiol. 55 (2010) 1377–1384.
166
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
[21] A. Giardini, M. Fenton, R.E. Andrews, et al., Peak oxygen uptake correlates with survival without clinical deterioration in ambulatory children with dilated cardiomyopathy, Circulation 124 (2011) 1713–1718. [22] J.F. Price, A.K. Thomas, M. Grenier, et al., B-type natriuretic peptide predicts adverse cardiovascular events in pediatric outpatients with chronic left ventricular systolic dysfunction, Circulation 114 (2006) 1063–1069.
[23] J. Mangat, C. Carter, G. Riley, et al., The clinical utility of brain natriuretic peptide in paediatric left ventricular failure, Eur. J. Heart Fail. 11 (2009) 48–52. [24] J.J. O'Sullivan, S.L. Roche, D.S. Crossland, et al., Recovery of heart function in children with acute severe heart failure, Transplantation 85 (2008) 975–979.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Potential for and timing of recovery in children with dilated cardiomyopathy Matthew J. Fenton ⁎, Philippa Horne, Jacob Simmonds, Sophie L. Neligan, Rachel E. Andrews, Michael Burch Cardiothoracic Unit, Great Ormond Street Hospital for Children NHS Foundation Trust, Great Ormond Street, London WC1N 3JH, UK
a r t i c l e
i n f o
Article history: Received 8 March 2017 Received in revised form 10 December 2017 Accepted 20 December 2017
Keywords: Pediatric Heart failure Heart transplant Morbidity Mortality
a b s t r a c t Objective: Understanding the clinical course and time-frame for recovery is helpful to guide management and counselling following a diagnosis of Dilated Cardiomyopathy (DCM). We aimed to document outcomes and time to recovery for a cohort of patients with a dilated cardiomyopathy phenotype. Methods: An observational cohort methodology was used to collect retrospective data from the departmental database for those identified with DCM. Data relating to mode of presentation, echocardiographic parameters, clinical management and outcome were collated and analysed. Predictors and time-scale for recovery were investigated and reported. Results: 209 new referrals were included within the time frame. 82 children median age 1.0 years (IQR 3.4) required intensive care (ICU) and their survival without death or transplant was 51% to one year and 45% to five years. 127 children presented to the pediatric heart failure clinic. Excluding 58 with neuromuscular disease, median age was 4.1 years (IQR 11.3) & survival without death or transplant 85% to 1 year and 50% to 5 years. NT-proBNP normalized in survivors before echocardiographic parameters. Predictors of recovery included younger age, female sex and smaller left ventricular end diastolic Z score on echocardiogram at presentation. Conclusion: Transplant-free survival to one year is significantly better for patients presenting to clinic, but longer-term survival is better amongst those presenting to ICU due to a late attrition in those with less severe heart failure at presentation. Falling NT-proBNP is the earliest marker of recovery. Recovery of cardiac function remains possible up to three years from presentation. © 2018 Elsevier B.V. All rights reserved.
1. Introduction Our understanding of the epidemiology of pediatric DCM is well served with good databases from the United States of America and Australia [1,2]. However, while there is striking similarity in incidence, despite variation in inclusion criteria, the data on the rates of death, transplant and recovery of function are discrepant between these 2 major registries. The highest risk period in both datasets is the first year after presentation and transplant-free survival is similar at 74% and 69% respectively [3,4]. However for children requiring hospitalization for a first episode of heart failure, a single center study from America has shown transplant-free survival to be only 46% at 1 year, whereas registry data from the UK showed a transplant-free 1 year survival of 66% [5,6]. While congestive heart failure is common in pediatric DCM and its presence increases the hazard for death, the severity has not been specifically addressed as a risk factor
⁎ Corresponding author. E-mail address: [email protected] (M.J. Fenton).
https://doi.org/10.1016/j.ijcard.2017.12.075 0167-5273/© 2018 Elsevier B.V. All rights reserved.
[4]. Recent Dutch registry data has shown transplant-free 1 year survival of 82% and suggested rather controversially “a conservative approach [to transplant listing], specifically in young children presenting with heart failure and requiring ICU admission, but not mechanical circulatory support, is justifiable” [7]. We aimed to provide further data on outcomes based on severity of presentation by using a surrogate evaluation relating to mode of presentation. We hypothesised that children presenting to an ambulatory clinic or low dependency inpatient setting versus those decompensated to the intensive care would have different outcomes in their clinical progress in the short and medium term from presentation. We were also interested in determining the outcomes for those patients who recovered following their initial presentation and whether any improvement was sustained. This detailed information from a single centre would be useful for pediatric cardiologists to be aware of when faced with families and children with new onset DCM. Therefore, we reviewed the data from a large pediatric heart failure program and compared the outcomes for children admitted to the ICU with those who presented less acutely, over a 5 year period. For those that survived without a transplant we investigated the proportion and timescale for recovery of clinical parameters with particular attention
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
to the role that routine measurement of NT-proBNP might add to our predictive ability. 2. Materials and methods Great Ormond Street Hospital is a tertiary and quaternary pediatric referral center accepting patients from district general hospitals in the South East of England, and from other pediatric cardiac units across the UK. This study complies with the declaration of Helsinki and received approval from the local research governance department. We evaluated all children ≤16 years with primary or secondary DCM who were referred to our unit between January 2008 and July 2013 inclusive. Follow up data were included if individuals had a minimum of two episodes of care. Follow up was censored at the most recent known point of follow up. A DCM phenotype was diagnosed on 2 dimensional echocardiography. All patients included in this study had a fractional shortening (FS) measurement of ≤26% at presentation and a left ventricular end diastolic dimension (LVEDD) Z score of N+2. Patients were divided into two groups according to severity of presenting features; those who required admission to the cardiac ICU, and those who were managed in the outpatient clinic or a low dependency ward setting. For all patients in addition to the echocardiographic measurements above we recorded the NYHA/Ross score, NT-proBNP levels, and for ward/clinic patients' medications used. Between-group comparisons were made using unpaired Student t-test or Mann–Whitney U test as appropriate to the underlying distribution of the variables. A curve demonstrating the proportions of patients who died, were transplanted or remained alive without transplant over time was created to graphically demonstrate the patients' outcomes. Survival curves using Kaplan–Meier methodology censored for patients with a combined end point of either death or transplant were created and censored at the time of most recent follow up. Survival outcomes comparing patients who presented to intensive care versus those who presented to a low dependency or ward environment were compared using the log-rank test. A series of univariate Cox regression analyses were then performed to identify the risk of either death or transplant attributable to individual variables. Variables with a p value b0.1 were included in the subsequent multivariate analysis. To provide clinical relevance a change in NT-proBNP of 1000 pg/ml was considered to be clinically relevant with the absolute value divided by 1000. For those children who remained event free, without death or transplant, we evaluated whether the shortening fraction (FS), the ventricular size (LVEDD Z score) or NT-proBNP returned to the normal range. Using time to recovery for each of these variables and defining overall recovery as return to normal of all three clinical parameters, Kaplan–Meier analysis was used to construct an inverse survival curve to graphically represent the probability of recovery and the recovery time for each of the variables as well as the composite end point of all three parameters to provide overall recovery. Cox regression analysis was used to identify individual variables that significantly predict recovery. Variables with a p value b0.1 were carried forward to a subsequent multivariate analysis. Sources of bias were considered due to patients being seen by different clinicians for follow up. Measurement bias on echocardiography was limited by use of a departmental protocol though studies were performed by multiple operators.
3. Results Within the five year timeframe there were 209 new referrals for pediatric patients with a DCM phenotype. Of these 82 children, median age 1.0 year (IQR 3.4), were admitted to the cardiac intensive care unit and 12 (15%) of these were clinically diagnosed with myocarditis. Cardiac biopsy is not routinely undertaken to diagnose myocarditis in the UK and the rationale for this has been described previously [6]. Mechanical support was needed with ECMO or Berlin Heart in 28 (34%) and transplant in 29 (35%). Survival without death or transplant was 51% to one year and 45% to five years. All the patients presenting to intensive care were in NYHA class III or IV and distributed across the NYHA classification for children in low dependency settings. Patient details are included in Table 1. During the same period 127 new referrals were seen in the pediatric heart failure clinic or admitted to a low dependency ward with AHA/ACC heart failure stages B and C. Of these 58 had neuromuscular disease with reduced FS and 3 died within a year of presentation. Neuromuscular patients were excluded from the subsequent analysis. Of the remaining 69 new patients with DCM seen in low dependency settings the median age was 4.1 (IQR 11.2) and 3 (4.3%) were clinically diagnosed with myocarditis. Oral anti-failure medication was used to support cardiac function. The following drugs were used at any time
163
point during follow up. Either angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB) were used in all patients (ACEI in 96%, ARB in 4%), Furosemide in 78%, Spironolactone in 74%, Warfarin in 9%, Aspirin in 54%, Betablocker (Carvedilol) in 65%, and Digoxin in 56%. 4 patients received Ivabradine, 2 patients Carnitine and 1 patient received Sildenafil. Survival without death or transplant for these patients was 85% to 1 year, and was 50% to 5 years. The different transplant-free survival between those children presenting to intensive care and those to a low dependency setting is shown in Fig. 1A. Assumptions for proportional hazards modelling were met with no significance demonstrated due to changes over time. Multivariate modelling for predictors of death or transplant for all patients despite their location of presentation is shown in Table 2. Older age, reduced FS at presentation and increased NT-proBNP level per 1000 unit increase were significant risk factors. LVEDD Z score at baseline was not a significant predictor of outcome. Whilst presenting to the intensive care unit was a significant risk factor for death or heart transplantation as part of univariate analysis there was no contribution to the multivariate model. For the 87 patients from both groups who survived without transplant, the time course of recovery is shown in Fig. 1B. NT-proBNP improved before FS, and LVEDD was the last to improve. Recovery was most rapid in the first year after presentation but continued over the subsequent 3 years before entering a plateau phase indicating recovery of biomedical and echocardiographic parameters in 60% of patients. Table 3 documents the predictors for recovery for children presenting in clinic, with young, female children with a smaller left ventricular end diastolic diameter Z score more likely to recover their parameters to within the normal range. 4. Discussion While it is clear that the first year after presentation poses the highest risk in pediatric DCM, this paper shows clear divergence in 1 year transplant-free survival for children with DCM presenting to the ward or clinic (85%) and those needing initial intensive care (51%) [5,6]. This distinction is useful additional information when counselling families, using the ‘route of care’ as a surrogate for disease severity stratifies the presentation of DCM and provides more granularity regarding prognosis for different modes of presentation. Unlike other cohorts our data includes NT-proBNP which whilst contributing to the model failed to produce a clinically important hazard ratio during univariate or multivariate analysis. However for those patients who recovered the NT-proBNP serial data was clinically useful. NT-proBNP is the first parameter for recovery and when measured serially seems to be an important clinical marker to guide therapy. Further long-term analysis will be useful to answer this question over time. Other predictors of
Table 1 Patient details.
Number presented Mean age at presentation in years (SD) Median age at presentation (IQR) Number of males (%) Mean NT-proBNP (SD) Mean LVEDD Z score (SD) Mean FS % (SD) Number required mechanical support (%) Number transplanted (%) Number died (%) Number in NYHA class I (%) Number in NYHA class II (%) Number in NYHA class III (%) Number in NYHA class IV (%) ⁎ P b 0.05.
ICU
Ward/Clinic
82 3.2 (4.5) 1.0 (3.4) 40 (49%) 19,460 (12,459) 5.3 (4.0) 13.3 (6.2) 28 (34%) 29 (35%) 14 (17%) 0 (0%) 0 (0%) 7 (9%) 75 (91%)
69 6.3 (5.6) 4.1 (11.2) 32 (46%) 5609 (8208) 5.6 (3.5) 18 (8.1) 8 (12%) 12 (17%) 9 (13%) 41 (59%) 12 (17%) 9 (13%) 7 (9%)
P value b0.001* b0.001* 0.87 b0.001* 0.57 b0.001* 0.001* 0.01* 0.49
164
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
data towards the more severe cases. Over a third of those admitted to the ICU received mechanical support. This is different from the Australian data which predated the Berlin Heart era where no patients had mechanical support, the North American data where 5% had mechanical support, and the more recent Dutch data which did incorporate the Berlin Heart era but only 8% had mechanical support [3,4,7]. Therefore our data appears to offer an opportunity to assess these more severely ill children. Rather intriguingly the survival data does show a better longer-term outcome for those who presented to ICU. A possible explanation for this is related to aetiology. Myocarditis is known to have a better outcome than idiopathic DCM, particularly if fulminant, and in our patients myocarditis was more common in the intensive care unit compared to the clinic. (15% vs 4%) [12,13]. This needs to be qualified over a longer period of time once outcomes after heart transplantation enter the follow-up period. Recently the Dutch registry paper suggested that a “conservative approach specifically in young children presenting with heart failure and requiring intensive care admission, but not mechanical circulatory support, is justifiable” [7]. This appears to be supported by transplant registry data, which shows that wait list mortality is low for pediatric DCM with the exception of those on ECMO, ventilation or who have arrhythmias [14]. In our experience the very sick patients still have a high mortality and without mechanical support and transplant this mortality would be greater, but perhaps those who survive ICU should be given adequate time to stabilise or even improve before listing for transplant. Whether this would extend to using Berlin Heart and other ventricular assist devices as bridges to recovery remains unclear. Data from the UK does show that Berlin Heart explantation is possible and often successful in myocarditis, but in DCM it is less likely to be associated with a good longer term outcome [15]. The finding of a higher mortality after transplant for myocarditis compared to DCM further supports a conservative approach to listing for transplant in myocarditis [16]. However, our data suggests that a conservative approach to listing for transplant may be an oversimplification. Data from Europe and America has shown that the only therapy that has improved survival for pediatric DCM is transplantation and that drug therapy has been disappointing [17–20]. There is a late deterioration even in children who have recovered function [9]. Our study shows a steady attrition in patients whose initial presentation was to the ward or clinic and clearly some patients with chronic heart failure will benefit from transplantation. This underscores the need to predict which patients will benefit from listing and detailed echocardiography, exercise test and blood biomarkers may be of help [10,21–23]. Although recovery of function is more common with myocarditis than idiopathic DCM, cases of idiopathic DCM may still recover function. The North American registry data has been re-analysed recently to assess recovery, with myocarditis and neuromuscular disease excluded [9]. In the remaining cohort of idiopathic DCM patients there was a recovery of function in 22% at 2 years. The Dutch data showed 38% of patients recovered function with similar numbers of myocarditis and idiopathic patients, with most recovering within a year [7]. However data from the Australian registry and a small UK study have shown recovery may continue over a long period [24].
Fig. 1. A) Survival curve for patients presenting to the clinic/ward or to intensive care; B) graph depicting recovery of cardiac parameters. FS, fractional shortening; LVEDD, left ventricular diastolic diameter; NT-proBNP, N-terminal pro-brain natriuretic peptide.
transplant-free survival are consistent with and strengthen the existing literature, namely age and FS measurement at presentation [8–11]. Recovery of function to normal occurred in survivors from both ICU and the clinic, and approximately 60% had both a normal echo and NT-proBNP after 3 years. Age and FS again predicted recovery. Another novel finding from our data is that survivors of new onset pediatric DCM do better in the medium term if their initial presentation was to the ICU. Presumably this phenomenon is related to the fact that more of the patients in intensive care will have received mechanical support and subsequent transplant, providing improved medium term survival. Our data demonstrates the continued attrition due to deterioration for children managed in an outpatient setting. The number of patients reported here over a 5-year period is greater than the complete national cohort from Holland over a similar period [7]. One benefit of a large single center study is that the clinical data can be collated in detail and can include consistent echocardiographic and biomarker data such as NT-proBNP. However, as we are a national referral unit for advanced heart failure, this will inevitably skew our Table 2 Analysis of predictors for either death or transplant. Risk factor
Male sex Age at presentation Baseline NT-proBNP Baseline LVEDD Z score Baseline FS% Presentation to ICU
Univariate analysis
Multivariate analysis (events = 47, n = 133)
HR
n
p
HR
Lower 95% confidence limit
Upper 95% CI
1.00 1.04 0.99 1.05 0.93 2.36
151 151 133 151 151 151
0.997 0.05 0.006 0.22 0.001 0.001
– 1.16 1.04 – 0.94 1.32
– 1.09 1.01 – 0.89 0.58
– 1.23 1.07 – 0.98 3.01
p 0.001 0.007 0.008 0.514
HR, hazard ratio; n, number of complete patient datasets included in analysis; NT-proBNP, N-terminal pro-brain natriuretic peptide; LVEDD, left ventricular end diastolic diameter; FS, fractional shortening; CI, confidence interval; ICU, intensive care unit.
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
165
Table 3 Analysis of predictors for recovery in clinic patients. Risk factor
Male sex Age at presentation Baseline NT-proBNP Baseline LVEDD Z score Baseline FS% Presentation to ICU
Univariate analysis
Multivariate analysis (events = 46, n = 86)
HR
n
p
HR
Lower 95% confidence limit
Upper 95% confidence limit
p
0.56 0.94 1.00 0.87 1.03 1.55
87 87 86 87 87 87
0.062 0.081 0.48 0.012 0.142 0.138
0.51 0.9 – 0.83 – –
0.28 0.83 – 0.75 – –
0.96 0.97 – 0.92 – –
0.035 0.005 – 0.001 – –
HR, hazard ratio; n, number of complete patient datasets included in analysis; NT-proBNP, N-terminal pro-brain natriuretic peptide; LVEDD, left ventricular end diastolic diameter; FS, fractional shortening; ICU, intensive care unit.
Our data offers new insight into the sensitivity and timescale for markers of recovery with distinct recovery curves for the 3 studied parameters. No prior data on NT-proBNP and recovery exists although its prognostic use in pediatric heart failure has been shown in America and Europe [22,23]. Our data demonstrates that in those recovering, NT-proBNP is the earliest marker to normalise, followed by FS and finally LVEDD. This emphasizes the importance of measuring serial NT-proBNP levels in these children. This data demonstrates that recovery can occur over a prolonged period of time after presentation and for those patients who remain relatively stable there is the potential for improvement beyond two years from presentation. This data supports prolonged outpatient management and suspending transplant listing in those patients who are clinically well enough to do so. 5. Limitations This is a single center study. Although our data may be helpful to those counselling patients and families it should be applied cautiously to other populations unless or until confirmed by registry data. The number of patients seen in this single centre study is unusually large. It reflects a far wider referral practice than just the city of London area and therefore our data cannot be used for population studies or incidence. Biopsies were not performed routinely but clinical diagnosis of myocarditis is accepted to be equally reliable in many countries, with similar outcomes in clinically suspected and biopsy-proven myocarditis [12]. The data is limited to a 5 year period, but this does reflect the current practice of mechanical support, encompassing the more widespread use of the Berlin Heart in the pediatric population. 6. Conclusions Our data is the first to demonstrate better 1-year outcome for children with DCM presenting with less severe heart failure, and supports more optimistic counselling regarding early outcomes in those presenting to the ward or clinic. However, there is an ongoing attrition in these patients and medium term survival was actually better in those with a more fulminant heart failure who required intensive care support at presentation. Recovery of function is possible after presentation with NT-proBNP demonstrating early improvement prior to echocardiographic parameters and female sex, younger age and a smaller left ventricle end diastolic diameter at presentation being predictive of recovery up to three years after diagnosis. Funding sources This study was supported by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children NHS Foundation Trust and University College London.
Conflict of interest No conflict of interest for any authors.
References [1] S.E. Lipshultz, L.A. Sleeper, J.A. Towbin, et al., The incidence of pediatric cardiomyopathy in two regions of the United States, N. Engl. J. Med. 348 (2003) 1647–1655. [2] A.W. Nugent, P.E.F. Daubeney, P. Chondros, et al., The epidemiology of childhood cardiomyopathy in Australia, N. Engl. J. Med. 348 (2003) 1639–1646. [3] P.E.F. Daubeney, A.W. Nugent, P. Chondros, et al., Clinical features and outcomes of childhood dilated cardiomyopathy: results from a national populationbased study, Circulation 114 (2006) 2671–2678. [4] J.A. Towbin, A.M. Lowe, S.D. Colan, et al., Incidence, causes, and outcomes of dilated cardiomyopathy in children, JAMA 296 (2006) 1867–1876. [5] S.A. Hollander, D. Bernstein, J. Yeh, et al., Outcomes of children following a first hospitalization for dilated cardiomyopathy, Circ. Heart Fail. 5 (2012) 437–443. [6] R.E. Andrews, M.J. Fenton, D.A. Ridout, et al., New-onset heart failure due to heart muscle disease in childhood: a prospective study in the United kingdom and Ireland, Circulation 117 (2008) 79–84. [7] Boer den SL, M. Lennie van Osch-Gevers, G. van Ingen, et al., Management of children with dilated cardiomyopathy in The Netherlands: Implications of a low early transplantation rate, J. Heart Lung Transplant. 34 (2015) 963–969. [8] T.P. Singh, L.A. Sleeper, S. Lipshultz, et al., Association of left ventricular dilation at listing for heart transplant with postlisting and early posttransplant mortality in children with dilated cardiomyopathy, Circ. Heart Fail. 2 (2009) 591–598. [9] M.D. Everitt, L.A. Sleeper, M. Lu, et al., Recovery of echocardiographic function in children with idiopathic dilated cardiomyopathy: results from the pediatric cardiomyopathy registry, J. Am. Coll. Cardiol. 63 (2014) 1405–1413. [10] K.M. Molina, P. Shrader, S.D. Colan, et al., Predictors of disease progression in pediatric dilated cardiomyopathy, Circ. Heart Fail. 6 (2013) 1214–1222. [11] P.M.A. Alexander, P.E.F. Daubeney, A.W. Nugent, et al., Long-term outcomes of dilated cardiomyopathy diagnosed during childhood: results from a national population-based study of childhood cardiomyopathy, Circulation 128 (2013) 2039–2046. [12] S.R. Foerster, C.E. Canter, A. Cinar, et al., Ventricular remodeling and survival are more favorable for myocarditis than for idiopathic dilated cardiomyopathy in childhood: an outcomes study from the Pediatric Cardiomyopathy Registry, Circ. Heart Fail. 3 (2010) 689–697. [13] R.E. McCarthy, J.P. Boehmer, R.H. Hruban, et al., Long-term outcome of fulminant myocarditis as compared with acute (nonfulminant) myocarditis, N. Engl. J. Med. 342 (2000) 690–695. [14] R. Kirk, D. Naftel, T.M. Hoffman, et al., Outcome of pediatric patients with dilated cardiomyopathy listed for transplant: a multi-institutional study, J. Heart Lung Transplant. 28 (2009) 1322–1328. [15] C.A. Irving, D.S. Crossland, S. Haynes, et al., Evolving experience with explantation from Berlin Heart EXCOR ventricular assist device support in children, J. Heart Lung Transplant. 33 (2014) 211–213. [16] B.A. Pietra, P.F. Kantor, H.L. Bartlett, et al., Early predictors of survival to and after heart transplantation in children with dilated cardiomyopathy, Circulation 126 (2012) 1079–1086. [17] A.-M. McMahon, C. Van Doorn, M. Burch, et al., Improved early outcome for end-stage dilated cardiomyopathy in children, J. Thorac. Cardiovasc. Surg. 126 (2003) 1781–1787. [18] A.E. Tsirka, K. Trinkaus, S.-C. Chen, et al., Improved outcomes of pediatric dilated cardiomyopathy with utilization of heart transplantation, J. Am. Coll. Cardiol. 44 (2004) 391–397. [19] R.E. Shaddy, M.M. Boucek, D.T. Hsu, et al., Carvedilol for children and adolescents with heart failure: a randomized controlled trial, JAMA 298 (2007) 1171–1179. [20] P.F. Kantor, J.R. Abraham, A.I. Dipchand, et al., The impact of changing medical therapy on transplantation-free survival in pediatric dilated cardiomyopathy, J. Am. Coll. Cardiol. 55 (2010) 1377–1384.
166
M.J. Fenton et al. / International Journal of Cardiology 266 (2018) 162–166
[21] A. Giardini, M. Fenton, R.E. Andrews, et al., Peak oxygen uptake correlates with survival without clinical deterioration in ambulatory children with dilated cardiomyopathy, Circulation 124 (2011) 1713–1718. [22] J.F. Price, A.K. Thomas, M. Grenier, et al., B-type natriuretic peptide predicts adverse cardiovascular events in pediatric outpatients with chronic left ventricular systolic dysfunction, Circulation 114 (2006) 1063–1069.
[23] J. Mangat, C. Carter, G. Riley, et al., The clinical utility of brain natriuretic peptide in paediatric left ventricular failure, Eur. J. Heart Fail. 11 (2009) 48–52. [24] J.J. O'Sullivan, S.L. Roche, D.S. Crossland, et al., Recovery of heart function in children with acute severe heart failure, Transplantation 85 (2008) 975–979.