Comparison of prognostic variables in children and adults with Fontan circulation

International Journal of Cardiology 173 (2014) 277–283

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Comparison of prognostic variables in children and adults with Fontan circulation Hideo Ohuchi ⁎, Kenji Yasuda, Aya Miyazaki, Toru Iwasa, Heima Sakaguchi, Ono Shin, Masanori Mizuno, Jun Negishi, Kanae Noritake, Osamu Yamada Department of Pediatric Cardiology, National Cerebral and Cardiovascular Center, 5-7-1, Fujishiro-dai, Suita, Osaka 565–8565, Japan

a r t i c l e

i n f o

Article history: Received 15 September 2013 Received in revised form 1 February 2014 Accepted 2 March 2014 Available online 5 March 2014 Keywords: Fontan Prognosis Child Adult Non-cardiac

a b s t r a c t Background: Non-cardiac complications, such as hepato-renal and metabolic problems, are emerging late after the Fontan operation due to its unique hemodynamics. Consequently, associations between clinical variables and postoperative outcome may change during the prolonged postoperative course. Methods and results: To determine if child and adult Fontan patients differ in the impact of cardiac and noncardiac variables on clinical outcome, we prospectively evaluated associations between hemodynamics, neurohumoral factors, exercise variables, hepato-renal function and metabolic variables and unscheduled hospitalization, including death in 167 consecutive child and 116 adult Fontan patients. When compared with child patients, the adult patients showed higher rates of medications, lower cardiac index, higher values of natriuretic peptides, greater renal dysfunction, more cholestatic livers, and more impaired responses to exercise (p b 0.05–0.0001). During the follow-up of 3.7 ± 2.1 years, 64 clinical events (37 in adults), including 13 deaths, occurred. A high CVP and low arterial oxygen satutration strongly predicted the child events (p b 0.001), whereas these prognostic parameters were marginal in the adults. Instead, renal dysfunction and metabolic abnormality predicted adult events (p b 0.05). Neurohumoral activation, low albumin, hyponatremia, and impaired exercise variables equally predicted clinical events in child and adult Fontan patients. Conclusions: Distinctive differences in predictive value of clinical variables exist between child and adult Fontan patients. In addition to cardiac issues, we should consider non-cardiac determinents of clinical outcome to maximize our efforts to improve prognosis for adult Fontan survivors. © 2014 Elsevier Ireland Ltd. All rights reserved.

Owing to recent advances in surgical and perioperative management, most post-Fontan patients now reach adulthood [1]. Fontan patients encounter arrhythmias, heart failure, and other clinical events, including death, as they age because of multiple substrates for arrhythmias and the palliative nature of the hemodynamics [2,3]. Recent studies have also demonstrated glucose metabolic abnormalities along with of liver and kidney dysfunction [4–6]. Possibly these “non-cardiac” pathophysiologies adversely influence morbidity and mortality and to date many clinical variables have been evaluated as to how they might associate with morbidity and mortality in Fontan patients [7–12]. In addition to established clinical variables, “non-cardiac” pathophysiologies may be more significantly involved in deterioration in adult Fontans than in children. Furthermore, the established prognostic values of clinical variables, such as central venous pressure (CVP), may change over time in this situation. To confirm this hypothesis, we prospectively evaluated cardiac ⁎ Corresponding author. Tel.: +81 6 6833 501; fax: +81 6 6872 7486. E-mail address: [email protected] (H. Ohuchi).

http://dx.doi.org/10.1016/j.ijcard.2014.03.001 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

and non-cardiac clinical variables and compared their impacts on the morbidity and mortality in child and adult Fontan patients. 1. Methods 1.1. Subjects Since May 1980, 369 patients have undergone cardiac catheterization in our institute to evaluate their postoperative hemodynamics at least 6 months after the Fontan repair. Of those, 340 underwent their initial Fontan operation in our institution and the other 29 elsewhere. Our Fontan follow-up policy has included routine cardiac catheterization every 5 years post-operation to evaluate hemodynamics and exercise performance for all patients [11]. Among those patients, from January 2006 to December 2012, we prospectively studied 283 clinically stable Fontan patients (age 6 to 55 years) and they consisted of those ≥6 years old who were able to perform cardiopulmonary exercise testing (CPX). Clinical stability implied freedom from intravenous medications with no major change of oral medications and a postoperative follow-up of at least 3 months. Two-hundred and seventy patients had had a cavopulmonary connection (TCPC) and 13 an atriopulmonary connection (APC). We divided our patients into two groups according to age b18 years old (children, mean = 10 ± 4 years, n = 167) and age ≥18 years (adults, mean = 24 ± 6 years, n = 116). Our patients included 22 (14 in children) with protein

278

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

2.1. Neurohumoral activities

Table 1 Subject characteristics. Total n 283 Age (yr) 16 ± 8 Height (cm) 145 ± 21 Weight (kg) 40 ± 16 Body mass index (kg/m) 18 ± 4 Age at Fontan (yr) 5±6 Follow-up (yr) 9±8 SV type (LV/non-LV) 113/171 Diagnosis Heterotaxy 79 (28%) UVH 78 TA 62 DORV 49 MA 23 CAVC 18 PA 23 HLHS 10 Others 20 Type of repair APC 12 IAR 80 ECR 191 Previous or additional procedures at Fontan APS 160 PAB 78 Glenn 136 Fenestration 37 Medications (%) Diuretics 46 Anti-coagulant 77 ACEI/ARB 30 Beta blocker 20 Anti-arrhythmia 12

Adults

Children

116 24 ± 6 161 ± 8 53 ± 11 20 ± 4 9±8 15 ± 6 56/60

167 10 ± 4 133 ± 19 31 ± 13 16 ± 3 3±2 8±3 57/111

27 (23%) 32 34 16 12 4 11 0 7

52 (32%) 46 28 33 11 14 12 10 13

10 60 46

2 20 145

69 20 29 14

91 58 107 23

51 71 36 29 16

43 81 25 13 7

ACEI = angiotensin converting enzyme inhibitor, APC = atriopulmonary connection, APS = aortopulmonary shunt, ARB = angiotensin receptor blocker, AVVP/R = atrioventricular valve plasty/replacement, CAVC = common atrioventricular canal, DORV = double outlet right ventricle; ECR = extracardiac rerouting, HLHS = hypolastic left heart syndrome, IAR = intraatrial rerouting, LV = left ventricle, MA = mitral valve atresia, PA = pulmonary valve atresia, PAB = pulmonary artery banding, SV = systemic ventricle, TA = tricuspid.

losing enteropathy. Medications and history of surgical procedures are described in Table 1. The study protocol was approved by the Ethics Committee of the National Cerebral and Cardiovascular Center.

2. Assessment of hemodynamics, systemic ventricular and atrioventricular valvular functions

After at least 15 min supine rest, the plasma levels of norepinephrine (NE), brain natriuretic peptides (BNP) and renin activity (PRA) were determined in 283, 283, and 260 Fontan patients, respectively [13–15]. 2.2. Serum lipids, glucose, metabolic and biochemical variables Biochemical variables included plasma albumin (g/dl), sodium (Na+: mEq/l), creatinine, and blood urea nitrogen (mg/dl), liver enzymes, and total cholesterol (mg/dl). Glucose metabolic variables included fasting plasma levels of glucose (mg/dl) and insulin (μU/ml) and homeostasis model assessment (HOMA-IR) was used to assess insulin resistance [16]. 2.3. Pulmonary function tests We measured vital capacity (VC; l) and percent forced expiratory volume in 1 s (FEV1) in 110 child and 104 adult patients (Spirosift, SP-600, Fukuda Denshi, Tokyo) and VC was calculated as the percentage of the body height predicted normal value for Japanese children and adults [17]. 2.4. Exercise protocol Two hundred and sixty seven patients (152 children and 115 adults) underwent symptom-limited CPX on a treadmill [18] and peak oxygen uptake (VO2) (ml/kg/min) was measured and calculated as the percentage of body weight predicted normal value for our institute. We used a twelve lead ECG to determine heart rate. Ventilation and gas exchange were measured using a breath-by-breath method. Minute ventilation versus carbon dioxide production slope (VE–VCO2 slope) was determined from the start of ramp exercise to the respiratory compensatory point and expressed as the percentage of our normal values. 2.4.1. Prospective clinical events Clinical events were defined as events requiring unscheduled hospitalizations (USHs) for management and included arrhythmias, heart failure, thromboembolisms, emergency catheter and/or surgical interventions and death from all causes. Protein losing enteropathy [6], including its relapse, and renal failure were also included. Heart failure was diagnosed if there was at least one of the following: orthopnea, nocturnal dyspnea, pulmonary edema, increasing peripheral edema, or radiological signs [19]. Hemoptysis and infectious endocarditis were also included as a cause of USH. 2.5. Statistical analysis

Cardiac catheterization was performed in 165 child and 114 adult patients within 1 week of CPX. We measured pressures in the cardiac chambers and great vessels, estimated oxygen consumption from the age, sex, and heart rate, and measured cardiac index (l/min/m2) using the Fick principle assuming equal right and left pulmonary arterial saturations in patients with either a Glenn or a TCPC because it is difficult to measure accurate flow distribution in the bilateral pulmonary arteries. Ventricular morphology was determined by echocardiography and/or cine-ventriculography and patients were divided into 3 groups [12], i.e., those with 1) a dominant left ventricle (LV) with or without a rudimentary right ventricle; 2) presence of both right and left ventricles; and 3) a dominant right ventricle with or without a rudimentary LV. In this study, each group consisted of 74, 35, and 58 children and 49, 10, and 57 adults, respectively. We used Simpson's rule to estimate morphological right and left ventricular volumes. End-diastolic ventricular volume was divided by body surface area to obtain end-diastolic volume index (EDVI) and systemic ventricular ejection fraction (EF) was calculated. The atrioventricular valve regurgitation was estimated by color flow mapping and graded as; none to mild, moderate, or severe.

Differences in continuous clinical variables were evaluated using a Student t test between the child and adult Fontan groups. Comparisons of the prevalence of medications were evaluated with a chi-square test. We used Cox's proportional hazards model to predict the associations between the clinical variables and USHs, including all cause mortality. The variables that proved to be significant predictors of the outcome in the univariate analysis (p b 0.05) were included in the multivariate analysis to determine the independent predictors. In assessing the impact of systemic ventricular function, i.e., EF and EDVI, on the prognosis with Cox's model, the hazard ratios (HRs) were computed for intervals of 10 (%) for the EF and 20 (ml/m2) for the EDVI to minimize the error of volume estimation with Simpson's rule. Sytemic ventricular type was divided into two categories, LV type and non-LV type. A free status from USHs was estimated using the Kaplan–Meier method and the differences in the event free status between the groups were assessed using log rank tests. A p value of b 0.05 was considered statistically significant. Data are expressed as the mean ± SD. Analyses were performed with the software JMP 10 pro (SAS Institute, Cary, NC, USA).

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

279

significant reduction of FEV1 (p b 0.0001) with a tendency to a lower VC in the adults (p = 0.064).

3. Results 3.1. Patient characteristics

3.4. Biochemical markers, including liver function and metabolic variables Age at the time of Fontan operation was older and the follow-up period was longer in the adults (p b 0.0001). The adult group had a higher percentage of APC patients (p b 0.001) and a higher rate of medications for renin–angiotensin–aldosterone system blockers, ß receptor drugs, and anti-arrhythmic drugs were also higher in the adults (Table 1, p b 0.05–0.001). 3.2. Hemodynamics, neurohormons, and renal function In the adults, EF, cardiac index, and systemic blood pressure were lower and CVP tended to be lower (Table 2). In the analysis of TCPC patients, the adults showed greater EDVI, lower EF and cardiac index (p b 0.05–0.01), while the CVP tended to be lower (p = 0.053). Although the plasma BNP level was higher in the adults, there were no differences in NE or PRA. When the TCPC patients were analyzed separately, the BNP level was higher in the adults (p b 0.01). Plasma level of creatinine was higher in the adults. The same trend was seen in the analyses of TCPC patients.

The adults showed higher levels of hemoglobin, total protein, total bilirubin, and gamma-glutamyltransferase, indicating more cholestatic liver abnormalities when compared with the children (Table 2). The TCPC patients showed the same trends. There were no differences in the plasma sodium concentration or HOMA-IR between the child and adult patients, while the total cholesterol level tended to be higher in the adults (Table 2) and the same trends were seen in the TCPC patients. 3.5. Predictors of USHs and mortality

3.3. CPX-derived variables and pulmonary function

During the follow-up of 3.7 ± 2.1 years, our patients experienced 64 USHs and eventually 13 died (6 in adults: parenthesis for adults). Breakdown of the first USHs was arrhythmias in 18 (10), heart failure in 14 (8), relapse of protein losing enteropathy in 11 (6), re-operation in 9 (4), hemoptysis in 5 (3), stroke and death in 2 each (2 and 1, respectively), and others in 3 (3). Because of the small number of deaths, the predictors were analyzed in all Fontan patients (Table 3) and predictors of the USHs were analyzed separately in the children and adults, and the HRs were compared (Table 4).

The standardized peak VO2, and VE/VCO2 slope, and peak heart rate in the adults were more impaired than in the children (Table 2). Similar trends were seen in the analyses of TCPC patients. Standardized VC and FEV1 were lower in the adults (Table 2) and the same trends were seen in the analyses of TCPC patients, showing a

3.5.1. Predictors of mortality On the univariate model, the predictors were summarized in Table 3. Among those, the multivariate Cox model (n = 194) revealed that BNP (hazard ratio [HR]: 1.13 per 10 pg/ml, 95% confidence interval [CI]: 1.02–1.25, p = 0.0236), NE (HR: 1.93 per 100 pg/ml, 95%CI: 1.16–

Table 2 Comparison of hemodynamics, neurohumoral factors, renal function, exercise variables, and pulmonary function between child and adult Fontan patients. Clinical variables

Total

Adults

Children

Hemodynamics CVP (mm Hg) SV end-diastolic volume index (ml/m2) SV ejection fraction (%) AVVR grade Nmoderate (%) Cardiac index (l/min/m2) Arterial oxygen saturation (%) Neurohumoral factors BNP (pg/ml) Norepinephrine (pg/ml) Renin activity (ng/ml/h) Renal function Blood urea nitrogen (mg/dl) Creatinine (mg/dl) Exercise variables Peak heart rate (bpm) Peak VO2(% of normal) VE/VCO2 slope (% of normal) Pulmonary function Vital capacity (% of normal) Forced expired volume in 1 s (%) Hematologic variables Hemoglobin (g/dl) Electrolytes Na (meq/l) Liver variables Albumin (g/dl) Total bilirubin (U/l) ALT (U/l) GGT (U/l) Glucose metabolism HOMA-IR Lipid status Total cholesterol (mg/dl)

279 11 ± 3 78 ± 24 56 ± 10 26 (9) 2.9 ± 0.8 94 ± 3 283 33 ± 55 428 ± 219 14 ± 17 282 13 ± 3 0.56 ± 0.16 273 145 ± 25 57 ± 13 131 ± 24 214 78 ± 15 88 ± 9

114 10 ± 2 81 ± 26 55 ± 10 13 (11) 2.6 ± 0.7 94 ± 3 116 48 ± 70 424 ± 193 13 ± 18 115 13 ± 4 0.66 ± 0.14 114 139 ± 27 53 ± 12 137 ± 26 110 76 ± 16 87 ± 6

165 11 ± 3 76 ± 22 58 ± 9 13 (8) 3.1 ± 0.8 94 ± 3 167 22 ± 38 431 ± 239 14 ± 16 167 13 ± 3 0.49 ± 0.14 159 149 ± 23 61 ± 13 127 ± 22 104 80 ± 13 90 ± 6

14.5 ± 1.7

14.9 ± 1.9

14.1 ± 1.3

b0.0001

139 ± 2

139 ± 2

139 ± 2

0.2605

4.4 ± 0.4 1.0 ± 0.6 24 ± 14 74 ± 59

4.4 ± 0.5 1.2 ± 0.8 25 ± 15 87 ± 74 116 1.6 ± 1.2 116 139 ± 26

4.4 ± 0.4 0.8 ± 0.4 24 ± 14 64 ± 42 124 1.6 ± 1.3 154 133 ± 21

0.9414 b0.0001 0.5127 0.0017

1.6 ± 1.3 134 ± 23

p 0.0785 0.079 0.0126 0.3269 b0.0001 0.4813 b0.0001 0.7796 0.7453 0.4973 b0.0001 0.0011 b0.0001 0.0008 0.038 0.0003

0.921 0.0533

ALT = alanine aminotransferase, ANP = atrial natriuretic peptide, AVVR = atrioventricular valve regurgitation, BNP = brain natriuretic peptide, CVP = central venous pressure, GGT = gamma-glutamyltransferase, HOMA-IR = homeostasis model assessment of insulin resistance, SV = systemic ventricle, VO2 = oxygen uptake, VE/VCO2 slope = minute ventilation vs. carbon dioxide production slope. Values are mean ± SD. P values indicate statistical significance between variables of adults vs. children.

280

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

Table 3 Univariate predictors of unscheduled hospitalization and mortality in all Fontan patients. Unscheduled hospitalization

Patient characteristics Age (yr) Male Age at first Fontan (yr) Non-LV SV Heterotaxy syndrome Protein losing enteropathy Hemodynamics CVP (per 1 mm Hg) EDVI (per 20 ml/m2) EF (per 10%) AVVR grade Nmoderate Cardiac index (l/min/m2) Arterial oxygen saturation (per 1%) Renal function Blood urea nitrogen (mg/dl) Creatinine (per 0.1 mg/dl) Exercise variables Peak heart rate (per 10 bpm) Peak VO2(% of normal, per 5 %) Pred-VE/VCO2 slope (per 5%) Pulmonary function Pred-vital capacity (per 10%) Forced expired volume in 1 s (per 5%) Neurohumoral factors BNP (per 10 pg/ml) Norepinephrine (per 100 pg/ml) Renin activity (ng/ml/h) Hematologic variables Hemoglobin (g/dl) Electrolytes Na (meq/l) Liver variables Albumin (per 0.1 g/dl) Total bilirubin (U/l) ALT (U/l) GGT (per 10 U/l) Glucose metabolism HOMA-IR Lipid status Total cholesterol (per 10 mg/dl)

Mortality

HR

95% CI

P value

HR

95% CI

P value

1.05 1.15 1.05 1.58 1.61 10.4

1.03–1.08 0.70–1.95 1.02–1.08 0.94–2.77 0.96–2.65 5.90–176

b0.0001 0.5880 0.0011 0.0861 0.0683 b0.0001

1.06 1.08 1.06 1.04 1.54 5.81

1.00–1.11 0.33–3.24 0.99–1.11 0.31–3.13 0.47–4.63 1.75–17.5

0.0392 0.8946 0.0779 0.9434 0.4548 0.0059

1.32 1.12 0.97 2.83 0.82 0.87

1.19–1.46 0.91–1.35 0.75–1.25 1.47–5.04 0.58–114 0.82–0.92

b0.0001 0.2612 0.793 0.0027 0.2478 b0.0001

1.56 1.25 0.82 1.51 0.99 0.88

1.27–1.90 0.78–1.83 0.46–1.78 0.23–5.63 0.45–1.91 0.83–0.94

b0.0001 0.3312 0.5015 0.6124 0.9792 0.0017

1.09 1.23

1.02–1.16 1.07–1.40

0.0141 0.0044

1.09 1.23

0.95–1.24 0.88–1.65

0.2139 0.2130

0.74 0.60 1.13

0.67–0.81 0.53–0.68 1.09–1.17

b0.0001 b0.0001 b0.0001

0.71 0.56 1.16

0.57–0.87 0.41–0.73 1.08–1.24

0.0012 b0.0001 0.0002

0.68 0.86

0.55–0.83 0.70–1.06

0.0002 0.1487

0.97 0.57

0.64–1.51 0.35–0.91

0.8969 0.0182

1.09 1.15 1.03

1.06–1.12 1.05–1.26 1.02–1.04

b0.0001 0.0049 b0.0001

1.07 1.36 1.03

1.02–1.10 1.16–1.57 1.01–1.04

0.009 0.0005 0.0044

1.13

0.97–1.30

0.1216

1.14

0.84–151

0.402

0.72

0.65–0.80

b0.0001

0.72

0.62–0.86

0.0009

0.87 1.41 1.01 1.03

0.84–0.91 1.12–1.67 0.99–1.02 1.00–1.06

b0.0001 0.0071 0.1782 0.0574

0.86 1.56 0.98 0.94

0.79–0.96 1.06–1.99 0.92–1.02 0.79–1.05

0.0068 0.0297 0.443 0.3944

1.05

0.85–1.24

0.6134

0.90

0.50–1.27

0.6309

1.00

0.90–1.12

0.9484

1.00

0.97–1.02

0.7187

Abbreviations are same as in the Tables 1 and 2. CI = confidence interval, HR = hazard ratio.

4.07, p = 0.0093) and FEV1 (HR: 0.33 per 5 %, 95%CI: 0.10–0.81, p = 0.0149) were the independent predictors.

1 ng/ml/h, 95%CI: 1.00–1.06, p = 0.0417) independently predicted the USHs.

3.5.2. Predictors of USHs in all Fontan patients On the univariate Cox model, the predictors were summarized in Table 3. Among those, the multivariate Cox model (n = 192) revealed that age (HR: 1.08 per 1 year, 95%CI: 1.00–1.15, p = 0.0422), CVP (HR: 1.22 per 1 mm Hg, 95%CI: 1.03–1.45, p = 0.0186) and BNP (HR: 1.06 per 10 pg/ml, 95%CI: 1.00–1.10, p = 0.0451) independently predicted the USHs.

3.6. Comparison of USH predictors between child and adult Fontan patients

3.5.3. Predictors of USHs in child patients On the univariate Cox model, many clinical variables were the predictors (Table 4, Fig. 1). The multivariate model (n = 104) revealed that heterotaxy (HR: 3.90, 95%CI: 1.02–17.2, p = 0.0469), CVP (HR: 1.55 per 1 mmHg, 95%CI: 1.15–2.20, p = 0.0034), peak VO2 (HR: 0.55 per 5%, 95%CI: 0.31–0.88, p = 0.0109), albumin (HR: 0.87 per 0.1 g/dl, 95%CI: 0.76–1.00, p = 0.0486), and HOMA-IR (HR: 0.48, 95%CI: 0.14– 0.97, p = 0.0374) independently predicted the USHs.

Table 4 also summarizes comparisons (interactions) of the predictors between child and adult patients and some of them are also shown in Fig. 1. Significant predictors only in the children were non-LV, heterotaxy, arterial oxygen saturation, and gamma-glutamyltransferase, while the predictors only in adult patients were age, male gender, late Fontan operation, EDVI, atrioventricular valve regurgitation ≥moderate, blood urea nitogen, creatinine, NE, and VC. Predictive power of the CVP and hyponatremia was greater in the children than in the adults (p b 0.05). Interestingly, an opposite direction of the HR was observed for the HOMA-IR (p b 0.05). As for the predictive value of the other significant variables, protein losing enteropathy, most CPX-derived variables and neurohormonal factors had the same impact on the USHs in child and adult patients. 4. Discussion

3.5.4. Predictors of USHs in adult patients On the univariate Cox model, many clinical variables also predicted the USHs (Table 4, Fig. 1). The multivariate model (n = 106) revealed that age (HR: 1.12, 95%CI: 1.00–1.25, p = 0.0455), BNP (HR: 1.08 per 10 pg/ml, 95%CI: 1.01–1.14, p = 0.026), and PRA (HR: 1.03 per

We found that in Fontan patients, 1) BNP, NE and FEV1 independently predicted the mortality, and age, CVP, and BNP independently predicted the USHs. Thus, BNP could be the most valuable predictor of morbidity and mortality in Fontan patients irrespective of age and

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

281

Table 4 Comparison of hazard ratios for unscheduled hospitalization between child and adult Fontan patients. Unscheduled hospitalization Adults

Patient characteristics Age (yr) Male Age at first Fontan (yr) Non-LV SV Heterotaxy syndrome Protein losing enteropathy Hemodynamics CVP (per 1 mmHg) EDVI (per 20 ml/m2) EF (per 10%) AVVR grade Nmoderate Cardiac index (l/min/m2) Arterial oxygen saturation (per 1%) Renal function Blood urea nitrogen (mg/dl) Creatinine (per 0.1 mg/dl) Exercise variables Peak heart rate (per 10 bpm) Pred-Peak VO2 (per 5 %) VE/VCO2 slope (per 1) Pulmonary function Pred-vital capacity (per 10%) Forced expired volume in 1 s (per 5%) Neurohumoral factors BNP (per 10 pg/ml) Norepinephrine (per 100 pg/ml) Renin activity (ng/ml/h) Hematologic variables Hemoglobin (g/dl) Electrolytes Na (meq/l) Liver variables Albumin (per 0.1 g/dl) Total bilirubin (U/l) ALT (U/l) GGT (per 10 U/l) Glucose metabolism HOMA-IR Lipid status Total cholesterol (per 10 mg/dl)

Interaction

Children

HR

95% CI

P value

P value

HR

95% CI

P value

1.04 2.05 1.04 1.11 1.01 8.22

1.00 – 1.08 1.05 – 4.23 1.00 – 1.07 0.58 – 2.18 0.47 – 2.03 3.36 – 18.3

0.0479 0.0360 0.0356 0.7489 0.9705 b 0.0001

0.7276 0.0503 0.5385 0.0286 0.0519 0.1190

1.06 0.74 0.97 4.41 2.88 16.33

0.96 – 1.18 0.34 – 1.64 0.81 – 1.13 1.54 – 18.5 1.35 – 6.27 7.25 – 36.6

0.2297 0.4425 0.8687 0.0037 0.0066 b 0.0001

1.19 1.26 1.11 2.78 1.15 0.91

1.03 – 1.37 1.01 – 1.57 0.80 – 1.54 1.31 – 5.92 0.75 – 1.77 0.82 – 1.00

0.0174 0.0430 0.5304 0.0078 0.5171 0.0600

0.0121 0.1128 0.4815 0.7972 0.1826 0.1644

1.49 0.88 0.92 2.23 0.78 0.85

1.29 – 1.72 0.59 – 1.23 0.61 – 1.41 0.65 – 5.80 0.42 – 1.31 0.79 – 0.91

b 0.0001 0.4817 0.7134 0.1787 0.3624 0.0002

1.10 1.24

1.03 – 1.17 1.03 – 1.50

0.0075 0.0152

0.3302 0.2242

1.02 1.00

0.90 – 1.15 0.75 – 1.29

0.7247 0.9787

0.78 0.65 1.08

0.69 – 0.88 0.55 – 0.76 1.04 – 1.12

0.0001 b0.0001 0.0001

0.4075 0.1569 0.1311

0.73 0.55 1.14

0.63 – 0.86 0.44 – 0.70 1.06 – 1.21

0.0002 b 0.0001 0.0002

0.71 0.93

0.57 – 0.90 0.72 – 1.20

0.0044 0.551

0.9797 0.6487

0.72 1.04

0.47 – 1.11 0.66 – 1.72

0.1370 0.8842

1.09 1.23 1.03

1.05 – 1.13 1.05 – 1.43 1.02 – 1.05

0.0001 0.0099 b0.0001

0.8259 0.4768 0.6897

1.08 1.13 1.03

1.02 – 1.12 0.98 – 1.27 1.01 – 1.04

0.0118 0.0811 0.0003

1.08

0.91 – 1.27

0.3807

0.6722

1.00

0.75 – 1.33

0.9943

0.77

0.69 – 0.87

b 0.0001

0.0203

0.59

0.47 – 0.74

b 0.0001

0.90 1.29 1.00 1.01

0.86 – 0.95 0.98 – 1.56 0.97 – 1.02 0.96 – 1.04

0.0002 0.0694 0.7408 0.7223

0.069 0.607 0.0836 0.0247

0.83 1.66 1.02 1.11

0.76 – 0.90 0.58 – 4.17 1.00 – 1.03 1.02 – 1.18

b 0.0001 0.3270 0.0541 0.0138

1.27

1.01 – 1.53

0.041

0.0047

0.57

0.28 – 0.98

0.0409

0.92

0.80 – 1.05

0.2410

0.1623

1.07

0.90 – 1.27

0.4382

Abbreviations are same as in the Tables 1, 2, and 3.

follow-up duration. 2) There were some discrepancies in the clinical variables of morbidity in terms of the predictive value and power. For instance, ventricular morphology, heterotaxy, and CVP, i.e., diagnosis itself and hemodynamics, had a greater impact on the prognosis in child Fontan patients. On the other hand, renal dysfunction and male gender, i.e., non-cardiac variables, had a significant prognostic impact in adult Fontan patients. Interestingly, HOMA-IR in the adults had the opposite prognostic values compared with those in the children. Finally, 3) some established prognostic variables, such as protein losing enteropathy, most CPX-derived variables and neurohormonal factors, had a similar impact on the morbidity. 4.0.1. Hemodynamics In child patients, hemodynamic variables, such as non-LV, heterotaxy, CVP, and arterial oxygen saturation were associated with USHs, indicating that management strategies focusing on impaired hemodynamics lead to better outcome in child patients. On the other hand, hemodynamic variables in adult patients may have less prognostic impact when compared with those in child patients. In fact, the prognostic values of CVP and arterial saturation were marginal and, instead, age and PRA along with BNP had independent predictive values in adult patients. These transitions of prognostic variables imply that adult Fontan with originally better hemodynamics and/or after successful hemodynamic management could have survived. Therefore, there may be

little hemodynamic variation in the late phase (adult) of Fontan circulation and non-cardiac problems, such as renal dysfunction and glucose metabolic abnormality, may emerge as prognostic variables just as described in adult heart failure patients. In addition, interestingly, we might need to keep in mind that ventricular function (EDVI and EF) and cardiac out had little impact on the prognosis in both child and adult Fontan patients, suggesting the unique pathophysiology. 4.0.2. Neurohormonal activities Most TCPC patients have a normal range of BNP and our median value was 15.5 pg/ml. However, a high level of BNP is associated with high morbidity and mortality [20]. Our finding of a significant association between BNP and prognosis supports the concept. The prognostic value of NE has also been implicated [21] and we confirmed its value in adult patients, including TCPC patients. In addition, we, for the first time, demonstrated the prognostic value of PRA in Fontan patients, implying important pathophysiologic influences of activated sympathetic activity and renin–angiotensin–aldosterone system on worsening of the Fontan circulation, especially in adults. 4.0.3. CPX-derived variables and pulmonary function All CPX-derived variables have a prognostic value in adult Fontan patients [10,22] and our present study expands the idea into child Fontan patients. In our adult patients, VC predicted USHs and more

282

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

Children Ventricular Morphology

Free Rate (%)

100

100

80

80

60

(p = 0.0117)

40 LV

20

non-LV

40 LV

Free Rate (%)

80

80

(p = 0.0041) 40 Sat

20

93

(ns)

Sat < 93

0

Free Rate (%)

80

(ns)

40 NE

392

NE < 392

0

0

10 20 30 40 50 60 70 80 90

VC

81%

10 20 30 40 50 60 70 80 90

Follow-Up (months)

0

10 20 30 40 50 60 70 80 90

80

80

60

60

40

40

Cr

20

NE

0.46

0

10 20 30 40 50 60 70 80 90

0

MOMA IR

60

20

100

80

80

60

60

40

40 82%

VC < 82%

0

10 20 30 40 50 60 70 80 90

Follow-Up (months)

10 20 30 40 50 60 70 80 90

Free Rate (%)

80

VC

0.59

Cr < 0.59 0

100

20

Cr

0

10 20 30 40 50 60 70 80 90

Free Rate (%) (p = 0.0029)

(p = 0.0019)

40

Cr

Cr < 0.46

NE < 512

0

(ns)

20

512

0 0

10 20 30 40 50 60 70 80 90

16

BUN < 16

100

VC < 81%

0

BUN

80

80

(ns)

20 0

0

Free Rate (%)

40

13

Free Rate (%)

100

20

40

BUN

Free Rate (%)

100

60

60

100

10 20 30 40 50 60 70 80 90

Free Rate (%)

(p = 0.0024)

BUN < 13

(p = 0.0002)

0 0

(ns)

0

100

20

80

20

Free Rate (%)

100

60

80

Sat < 95

10 20 30 40 50 60 70 80 90

10 20 30 40 50 60 70 80 90

Free Rate (%)

40

0

0

0

100

40 95

10 20 30 40 50 60 70 80 90

Free Rate (%)

60

Sat

Male 0

100

60

20

Female

20

Male 0

BUN

60 40

Female

0

Free Rate (%) 100

(ns)

20

10 20 30 40 50 60 70 80 90

(p = 0.0368)

80

non-LV

0

100

60

Vital Capacity

60

40

10 20 30 40 50 60 70 80 90

100

80

60

Adults Free Rate (%)

Gender 100

0 0

NE

Children Free Rate (%)

(ns)

20

0

Arterial Saturation

Adults

Free Rate (%)

(p = 0.0731)

(p = 0.0223)

60 40

20

HOMA 1.46 HOMA < 1.46

0

20

HOMA

0

10 20 30 40 50 60 70 80 90

Follow-Up (months)

1.76

HOMA < 1.76

0 0

10 20 30 40 50 60 70 80 90

Follow-Up (months)

Fig. 1. Comparisons of Kaplan–Meier curves between child and adult Fontan patients on the basis of each cutoff value according to the receiver operator characteristic curve. If a variable has no prognostic value, the Kaplan–Meier curve was created using the median value. BUN = blood urea nitrogen, Cr = creatinine, NE = plasma norepinephrine, HOMA IR = homeostasis model assessment of insulin resistance, LV = left ventricle, Sat = arterial oxygen saturation, VC = vital capacity.

interestingly FEV1 independently predicted mortality. Because none of our patients showed obstructive ventilatory impairment, a strong association between FEV1 and mortality may emphasize the importance of the respiratory system for the Fontan circulation [23]. In fact, FEV1 is one of the independent predictors of mortality in adults without cardiac disease although the mechanisms remain unclear [24]. 4.0.4. Liver and renal function Liver and renal dysfunctions have a prognostic value in heart failure patients [25,26]. In Fontan patients, liver problems as well as renal dysfunction have been well described [27,28]; however, their prognostic values remain unknown. We, for the first time, found that lower albumin and higher total bilirubin levels predicted USHs and mortality. In addition, high levels of liver enzymes were also predictive of USHs in child patients, implying the reflection of impaired hemodynamics, such as high CVP. On the other hand, the renal variables had a prognostic value for USHs, indicating the importance of maintaining renal function if the outcome in adult Fontan survivors is to improve. 4.0.5. Metabolic variables Glucose metabolic abnormalities have strong prognostic value in adults with cardiovascular disease [29]. Recent studies demonstrated that adults with complex congenital heart disease have a high prevalence of abnormal glucose metabolic abnormalities and measurement

might be of prognostic value [6]. We reconfirm their prognostic value in adult Fontan patients. More interestingly, our child Fontan patients with a high HOMA-IR had encountered fewer USHs. Although the reason is unclear, important physiological roles for insulin signaling for determination of cardiac size may exist and regulation of energy preference from glucose to fatty acid oxidation during developmental phase of life may explain the potential benefit of hyperinsulinemia during childhood with Fontan circulation [30]. Adults with Fontan circulation, on the other hand, hyperinsulinemia, i.e., insulin resistance, may reflect endothelial dysfunction [31,32]. 4.0.6. Other variables Hyponatremia is one of common features in many pathologic conditions, including heart failure, and has prognostic value [33]. Similar stories are applicable to adults with congenital heart disease and Fontan patients [34] and our present study reconfirm this idea. On the other hand, anemia did not predict prognosis in our Fontan patients although adults with heart failure as well as congenital heart disease with anemia have a poor prognosis [35]. 4.0.7. Study limitations First, this is a single center study and the follow-up period was short although the number of patients was relatively large. Therefore, the number of deaths is too small to determine the predictors in each

H. Ohuchi et al. / International Journal of Cardiology 173 (2014) 277–283

group of children and adults. However, as described in the methods, our study is prospective and has little selection bias. So, we expect that additional future follow-up data will reconfirm the present findings. Second, our estimation of ventricular volume, especially non-LV morphology ventricle, with cine-ventriculography had a significant limitation. Finally, we did not evaluate detailed associations between Fontan pathophysiologies and each clinical variable because our main purpose in the present study was to compare the prognostic value of each clinical variable between child and adult Fontan patients. Acknowledgment The authors of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology. References [1] Warnes CA, Liberthson R, Danielson GK, et al. Task force 1: the changing profile of congenital heart disease in adult life. J Am Coll Cardiol 2001;37:1170–5. [2] Motoki N, Ohuchi H, Miyazaki A, et al. Clinical profiles of adult patients with single ventricular physiology. Circ J 2009;73:1711–6. [3] Khairy P, Fernandes SM, Mayer Jr JE, et al. Long-term survival, modes of death, and predictors of mortality in patients with Fontan surgery. Circulation 2008;117:85–92. [4] Johnson JA, Cetta F, Graham RP, et al. Identifying predictors of hepatic disease in patients after the Fontan operation: A postmortem analysis. J Thorac Cardiovasc Surg 2013;146:140–5. [5] Dimopoulos K, Diller GP, Koltsida E, et al. Prevalence, predictors, and prognostic value of renal dysfunction in adults with congenital heart disease. Circulation 2008;117:2320–8. [6] Ohuchi H, Miyamoto Y, Yamamoto M, et al. High prevalence of abnormal glucose metabolism in young adult patients with complex congenital heart disease. Am Heart J 2009;158:30–9. [7] Ohuchi H, Kagisaki K, Miyazaki A, et al. Impact of the evolution of the Fontan operation on early and late mortality: a single-center experience of 405 patients over 3 decades. Ann Thorac Surg 2011;92:1457–66. [8] Ohuchi H, Negishi J, Miyake A, et al. Long-term prognostic value of cardiac autonomic nervous activity in postoperative patients with congenital heart disease. Int J Cardiol 2011;151:296–302. [9] Fernandes SM, Alexander ME, Graham DA, et al. Exercise testing identifies patients at increased risk for morbidity and mortality following Fontan surgery. Congenit Heart Dis 2011;6:294–303. [10] Diller GP, Giardini A, Dimopoulos K, et al. Predictors of morbidity and mortality in contemporary Fontan patients: results from a multicenter study including cardiopulmonary exercise testing in 321 patients. Eur Heart J 2010;31:3073–83. [11] Ohuchi H, Ono S, Tanabe Y, et al. Long-term serial aerobic exercise capacity and hemodynamic properties in clinically and hemodynamically good, "excellent", Fontan survivors. Circ J 2012;76:195–203. [12] Ohuchi H, Yasuda K, Hasegawa S, et al. Influence of ventricular morphology on aerobic exercise capacity in patients after the Fontan operation. J Am Coll Cardiol 2001;37:1967–74. [13] Mori K. Automated measurement of catecholamines in urine, plasma and hemogenates by high-performance liquid chromatography with fluorometric detection. J Chromatogr 1981;218:631–7.

283

[14] Kono M, Yamauchi A, Tsuji T, et al. An immunoradiometric assay for brain natriuretic peptide in human plasma. Kaku Igaku 1993;13:2–7. [15] Ikeda I, Iinuma K, Takai M, et al. Measurement of plasma renin activity by a simple solid phase radioimmunoassay. J Clin Endocrinol Metab 1982;54:423–8. [16] Matthews DR, Hosker JP, Rudenski AS, et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–9. [17] Yamaguchi I, Morishita M. New predictive formulas of pulmoanry function in Japanese children. Ann Jpn Respir Soc 1998;36:665–71. [18] Ohuchi H, Nakajima T, Kawade M, et al. Measurement and validity of the ventilatory threshold in patients with congenital heart disease. Pediatr Cardiol 1996;17:7–14. [19] Solomon SD, Wang D, Finn P, et al. Effect of candesartan on cause-specific mortality in heart failure patients: the Candesartan in Heart failure Assessment of Reduction in Mortality and morbidity (CHARM) program. Circulation 2004;110:2180–3. [20] Koch AM, Zink S, Singer H, et al. B-type natriuretic peptide levels in patients with functionally univentricular hearts after total cavopulmonary connection. Eur J Heart Fail 2008;10:60–2. [21] Inai K, Nakanishi T, Nakazawa M. Clinical correlation and prognostic predictive value of neurohumoral factors in patients late after the Fontan operation. Am Heart J 2005;150:588–94. [22] Fernandes SM, Alexander ME, Graham DA, et al. Exercise testing identifies patients at increased risk for morbidity and mortality following Fontan surgery. Congenit Heart Dis 2011;6:294–303. [23] Shafer KM, Garcia JA, Babb TG, et al. The importance of the muscle and ventilatory blood pumps during exercise in patients without a subpulmonary ventricle (Fontan operation). J Am Coll Cardiol 2012;60:2115–21. [24] Guerra S, Sherrill DL, Venker C, et al. Morbidity and mortality associated with the restrictive spirometric pattern: a longitudinal study. Thorax 2010;65:499–504. [25] Poelzl G, Ess M, Mussner-Seeber C, et al. Liver dysfunction in chronic heart failure: prevalence, characteristics and prognostic significance. Eur J Clin Invest 2012;42:153–63. [26] Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296–305. [27] Dimopoulos K, Diller GP, Koltsida E, et al. Prevalence, predictors, and prognostic value of renal dysfunction in adults with congenital heart disease. Circulation 2008;117:2320–8. [28] Baek JS, Bae EJ, Ko JS, et al. Late hepatic complications after Fontan operation; noninvasive markers of hepatic fibrosis and risk factors. Heart 2010;96:1750–5. [29] Matsue Y, Suzuki M, Nakamura R, et al. Prevalence and prognostic implications of pre-diabetic state in patients with heart failure. Circ J 2011;75:2833–9. [30] Belke DD, Betuing S, Tuttle MJ, et al. Insulin signaling coordinately regulates cardiac size, metabolism, and contractile protein isoform expression. J Clin Invest 2002;109:629–39. [31] Kim JA, Montagnani M, Koh KK, et al. Reciprocal relationships between insulin resistance and endothelial dysfunction: molecular and pathophysiological mechanisms. Circulation 2006;113:1888–904. [32] Jin SM, Noh CI, Bae EJ, et al. Impaired vascular function in patients with Fontan circulation. Int J Cardiol 2007;120:221–6. [33] Dimopoulos K, Diller GP, Petraco R, et al. Hyponatraemia: a strong predictor of mortality in adults with congenital heart disease. Eur Heart J 2010;31:595–601. [34] Ohuchi H, Negishi J, Ono S, et al. Hyponatremia and its association with the neurohormonal activity and adverse clinical events in children and young adult patients after the Fontan operation. Congenit Heart Dis 2011;6:304–12. [35] Dimopoulos K, Diller GP, Giannakoulas G, et al. Anemia in adults with congenital heart disease relates to adverse outcome. J Am Coll Cardiol 2009;54:2093–100.