Lee2017.pdf

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart Online First, published on October 27, 2017 as 10.1136/heartjnl-2017-312084 Heart failure and cardiomyopathies

Original research article

Hyponatraemia and its prognosis in acute heart failure is related to right ventricular dysfunction Heesun Lee,1 Sang Eun Lee,2 Chan Soon Park,3 Jin Joo Park,4 Ga Yeon Lee,5 Min-Seok Kim,2 Jin-Oh Choi,5 Hyun-jai Cho,3 Hae-Young Lee,3 Dong-Ju Choi,4 Eun-Seok Jeon,5 Jae-Joong Kim,2 Byung-Hee Oh3 ►► Additional material is published online only. To view please visit the journal online (http://d​ x.​doi.o​ rg/​10.​1136/​ heartjnl-​2017-​312084). 1

Division of Cardiology, Healthcare System Gangnam Center, Seoul National University Hospital, Seoul, Korea 2 Department of Cardiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea 3 Division of Cardiology, Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea 4 Department of Cardiology, Cardiovascular Center, Seoul National University Bundang Hospital, Seongnam, Korea 5 Division of Cardiology, Department of Internal Medicine, Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea Correspondence to Dr Sang Eun Lee, Department of Cardiology, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpagu, Seoul 138-736, Korea; ​ sangeunlee.​md@​gmail.c​ om Received 30 June 2017 Revised 5 October 2017 Accepted 5 October 2017

ABSTRACT Objectives  Hyponatraemia is a well-known predictor of clinical outcomes in heart failure (HF). However, the mechanism remains poorly understood. Previous reports suggest that hyponatraemia is related to right HF. We sought to evaluate the association between right ventricular (RV) dysfunction and hyponatraemia, and the impact of this relationship on the prognosis of patients with acute heart failure (AHF). Methods  This is a nested case–control study of the Korean Acute Heart Failure registry. Among 2935 AHF patients enrolled prospectively and consecutively at four tertiary hospitals in Korea from 2011 to 2014, 116 patients with severe persistent hyponatraemia, defined as serum sodium level <130 mmol/L at admission and <135 mmol/L before discharge, were matched with 232 controls, based on propensity scores for hyponatraemia. RV function was assessed with fractional area change (FAC) by echocardiography. Results  RV dysfunction (FAC <35%) was more prevalent in patients with severe persistent hyponatraemia than in those without (81.0% vs 33.6%, p<0.001). Hyponatraemia was strongly associated with RV dysfunction (adjusted OR 8.00, 95% CI 4.50 to 14.22, p<0.001), but not with left ventricular dysfunction (adjusted OR 1.21, 95% CI 0.74 to 1.50, p=0.308). RV dysfunction was an independent predictor of all-cause mortality, after adjustment for hyponatraemia (adjusted HR 2.20, 95% CI 1.53 to 3.15, p<0.001), while hyponatraemia was not (adjusted HR 1.33, 95% CI 0.94 to 1.87, p=0.108). Conclusions  In patients with AHF, hyponatraemia was more common with RV dysfunction. RV dysfunction, rather than hyponatraemia, was more significantly related with patients’ prognosis. Thus, the utility of RV dysfunction instead of hyponatraemia per se should be considered in HF risk models. Trial registration number  Korean Acute Heart Failure registryNCT01389843;Results.

Introduction

To cite: Lee H, Lee SE, Park CS, et al. Heart Published Online First: [please include Day Month Year]. doi:10.1136/ heartjnl-2017-312084

Hyponatraemia, the most common electrolyte imbalance in heart failure (HF), occurs in 20%–25% of patients1 2; it is as a surrogate marker of advanced disease and a significant predictor of poor outcomes in acute and chronic HF.1–5 Although the non-osmotic release of arginine vasopressin (AVP) due to arterial underfilling may be associated with hyponatraemia, the mechanism underpinning deleterious effects and the clinical features of hyponatraemia in

HF remain unknown.1 3 5 In our previous study of patients hospitalised for acute heart failure (AHF), an improvement of hyponatraemia was not associated with a better clinical outcome.6 Further, vasopressin antagonist therapy fails to improve survival in HF patients.7 These findings suggest that increased release of AVP and consequent hyponatraemia may not determine a poor prognosis in HF; rather, the causative factor responsible for both AVP and hyponatraemia is likely implicated. Right ventricular (RV) function is a strong predictor of mortality in various cardiac diseases, including HF.8 9 Several prior studies have propounded an association between hyponatraemia and RV function. An experiment in dogs showed indicated that hyponatraemia predominated in those with right HF.10 Two prospective studies with pulmonary arterial hypertension demonstrated similar results, noting that hyponatraemia was associated with right HF.11 12 Similarly, in the Korean Acute Heart Failure (KorAHF) registry, the prevalence of hyponatraemia was significantly higher in patients with isolated right HF than in those with other HF subtypes (27.1% vs 20.9%, p=0.002).13 Here, we hypothesised that hyponatraemia might be associated with RV dysfunction in patients with AHF, which accounts for the poor prognosis.

Methods Study population

This was a nested case–control study from the KorAHF registry, which is a prospective multicentre cohort study designed to describe the characteristics and clinical outcomes of patients hospitalised for AHF in Korea.13 14 We enrolled 2935 patients (52.2%) from four hospitals whose echocardiographic data were available. Among them, 122 patients (4.2%) showed severe persistent hyponatraemia, defined as serum sodium concentration (sNa) <130 mmol/L at admission and <135 mmol/L before discharge. After excluding six patients on dialysis, 116 patients were allocated as cases. To assess the independent relationship between hyponatraemia and RV dysfunction by reducing selection bias and statistical inferences of confounders related with hyponatraemia, we matched controls without hyponatraemia to cases according to a propensity score (PS). The PS, as a predicted probability for severe persistent hyponatraemia, was derived for each patient with logistic regression.15 Twelve variables that significantly differed included

Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

  1

Copyright Article author (or their employer) 2017. Produced by BMJ Publishing Group Ltd (& BCS) under licence.

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies

Figure 1  Flow chart of the study population. KorAHF, Korean Acute Heart Failure. to estimate the PS: alcohol, body mass index, history of hypertension, de novo HF, systolic and diastolic blood pressure, pulse rate, haemoglobin, renin-angiotensin system (RAS) blockers, beta-blockers, aldosterone antagonists and loop diuretics. We used a 1:2 nearest-neighbour-matching method without replacement, with no interactions included.16 Finally, we matched 116 patients with severe persistent hyponatraemia and 232 controls on the logit of PSs (figure 1). The protocol conformed to the Declaration of Helsinki, and was approved by the institutional review board of all hospitals with a waiver of written informed consent.

Clinical data and follow-up

We collected clinical information, laboratory results, electrocardiography, echocardiography and outcomes at admission, discharge and during follow-up (30 days, 90 days, 180 days, and 1 to 5 years, annually). Serial sNa values were collected according to the attending physician’s clinical decision, without a fixed interval. Detailed information regarding data collection has been described previously.13 14 All information was ascertained from patients by attending physicians. The mortality data were collected from national insurance data or death records.

Echocardiographic evaluation

All patients underwent a comprehensive echocardiographic examination at admission by certified cardiologists using commercially available equipment with second-harmonic imaging and a 3.5 MHz transducer. Echocardiographic evaluation was conducted according to current guidelines.16 17 Left ventricular ejection fraction (LVEF) was calculated from apical fourchamber and two-chamber views using the modified Simpson’s biplane method. If technically impossible, EF was evaluated with M-mode or visual estimation by experienced cardiologists. RV function was evaluated by the fractional area change (FAC) from an RV-optimised apical four-chamber view by tracing endocardial borders at end-diastole and end-systole.17 RV dysfunction was defined as FAC <35%. To minimise measurement bias, RV FAC was measured by two independent cardiologists, with final values decided by agreement. Tricuspid regurgitation (TR) was graded as 0 (nil) to 4 (severe), based on synthetic information regarding valve morphology, vena contracta width, proximal flow convergence radius and hepatic venous flow pattern.17 The diameter of the inferior vena cava (IVC) and its respiratory variation were measured 10–20 mm from the junction with the right atrium in the subcostal view, with an IVC size ≥20 mm and failure to collapse ≥50% during inspiration defined as dilatation and plethora, respectively.18 Pulmonary artery systolic pressure 2

(PASP) was estimated by adding right atrial pressure to peak trans-tricuspid pressure gradient at systole.19

Reproducibility

Intraobserver and interobserver variabilities for RV FAC value were assessed. Measurement of RV FAC was repeated twice within 2 weeks to assess the intraobserver variability. Interobserver variability was evaluated by comparing the value from two separate observers. Reproducibility was assessed using Bland-Altman analysis and intraclass correlation coefficient.

Statistical analysis

Descriptive statistics were used for clinical and echocardiographic characteristics, prevalence of RV dysfunction and death according to hyponatraemia. A p<0.05 was considered statistically significant. Data were presented as numbers and percentages for categorical variables, and mean±SD for continuous variables. Differences between continuous variables were compared using Student’s t-test or Wilcoxon rank sum test for independent samples, while those between categorical variables were analysed by the χ2 test or Fisher’s exact test, as appropriate. PS matching was performed to adjust the differences in baseline characteristics of the study population. All of the standardised differences for each of the baseline variables were <0.10 after matching. The predictive ability of our PS matching was assessed by the c-statistic (0.658) and p value of Hosmer and Lemeshow Goodness-of-Fit Test was 0.872, indicating good discrimination between the groups (χ2=3.835, df=8). In PS-matched cohort, associations of echocardiographic parameters with severe persistent hyponatraemia were derived from univariable and multivariable analyses with stepwise selection, and expressed as OR and corresponding 95% CI. Survival was determined using Kaplan-Meier analysis with the generalised Wilcoxon test to give greater weight to survival differences occurring at the early period of the study, since the proportional hazard assumption was not satisfied (p=0.034). Cox regression analyses were employed to determine significant predictors of survival during follow-up, including severe persistent hyponatraemia and RV dysfunction. The risk of all-cause death was expressed as an HR and corresponding 95% CI from univariable and multivariable analyses with stepwise selection. Sequential Cox analysis using four nested models ((1) significant clinical risk factors according to univariable analysis; (2) clinical risk factors+severe persistent hyponatraemia; (3) clinical risk factors+severe persistent hyponatraemia+left  ventricular (LV)  dysfunction; and (4) clinical risk factors+severe persistent hyponatraemia+RV dysfunction) was used to evaluate independent predictors and to compare Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies Table 1  Baseline characteristics in total patients before propensity score matching Total patients (n=2935)

Patients with severe persistent hyponatraemia (n=122)

Patients without severe persistent hyponatraemia (n=2813)

 Age, year

67.8±14.9

70.1±13.8

67.7±14.9

0.062

  ≥65  years (%)

1929 (65.7)

88 (72.1)

1841 (65.4)

0.128

 Male (%)

1634 (55.7)

70 (57.4)

1564 (55.6)

0.699

 Current smoking (%)

521 (17.8)

15 (12.3)

506 (18.0)

0.107

 Alcohol (%)

1261 (43.0)

39 (32.0)

1222 (43.4)

0.012

 BMI, kg/m2

23.4±4.1

22.5±3.8

23.4±4.1

0.010

 Hypertension (%)

1719 (58.6)

54 (44.3)

1665 (59.2)

0.001

 Diabetes mellitus (%)

1051 (35.8)

41 (33.6)

1010 (35.9)

0.604

 Chronic kidney disease (%)

465 (15.8)

18 (15.5)

447 (15.9)

0.917

 Ischaemic heart disease (%)

843 (28.7)

34 (27.9)

809 (28.8)

0.222

 Atrial fibrillation (%)

940 (32.0)

48 (39.3)

892 (31.7)

0.114

 Chronic obstructive lung disease (%)

477 (16.3)

19 (15.6)

458 (16.3)

0.836

878 (29.9)

37 (30.3)

841 (29.9)

0.919 0.002

Variables

p

Demographics

Medical history

Aetiology  Ischaemic (%) Clinical status on admission  De novo heart failure (%)

1432 (48.8)

43 (35.2)

1389 (49.4)

 Systolic BP, mm Hg

130.7±44.2

118.3±28.2

131.2±44.7

0.002

 Diastolic BP, mm Hg

79.0±39.2

70.1±17.7

79.3±39.8

<0.001

 Pulse rate, beats/min

92.5±35.4

83.5±24.1

92.9±35.7

<0.001

 NYHA class ≥III (%)

2526 (86.1)

104 (85.2)

2422 (86.1)

0.790

 LV ejection fraction, %

37.8±15.5

38.3±16.5

37.8±15.5

0.708

  <40%

1253 (42.7)

57 (46.7)

1196 (42.5)

0.358

 Hb, g/dL

12.4±2.3

11.9±2.4

12.4±2.3

0.010

 sCr, mg/dL

1.6±1.7

1.7±1.7

1.6±1.7

0.353 0.034

Evidence-based medical therapy before admission  RAS blocker (%)

1243 (42.4)

63 (51.6)

1180 (41.9)

 Beta-blocker (%)

940 (32.0)

29 (23.8)

911 (32.4)

0.046

 Aldosterone antagonist (%)

579 (19.7)

59 (48.4)

520 (18.5)

<0.001

 Loop diuretics (%)

2660 (90.6)

116 (95.1)

2544 (90.4)

0.042

 Nitrate (%)

1939 (66.1)

75 (61.5)

1864 (66.3)

0.274

 Amiodarone (%)

525 (17.9)

26 (21.3)

499 (17.7)

0.313

 Digitalis (%)

1032 (35.2)

50 (41.0)

982 (34.9)

0.169

714 (24.3)

59 (48.4)

655 (23.3)

<0.001

Outcome  1-year mortality (%)

BMI, body mass index; BP, blood pressure; Hb, haemoglobin; LV, left ventricular; NYHA, New York Heart Association; RAS, renin angiotensin system; sCr, serum creatinine.

the prognostic value among hyponatraemia per se, LV dysfunction and RV dysfunction. Subgroup analysis according to LVEF was performed to determine the effect of LV function on the relationship between hyponatraemia and RV dysfunction. All analyses were performed using SPSS V.22.0 (SPSS, Chicago, Illinois, USA) and the MatchIt R package (R Development Core Team, Ho et al, 2011).

Results Baseline clinical characteristics

Among all patients (n=2935, age 67.8±14.9 years, male 55.7%), hyponatraemia (sNa <135 mmol/L) was present in 697 (23.7%) patients, of which 122 (4.2%) had severe persistent hyponatraemia. Patients with severe persistent hyponatraemia (n=122, age 70.1±13.8 years, male 57.4%) were older, thinner, had a higher prevalence of alcohol history, hypertension and use of evidence-based medication and lower systolic/diastolic pressures and haemoglobin. Such patients exhibited higher 1-year mortality than those without (table 1). After excluding dialysis Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

patients, 116 patients with severe persistent hyponatraemia were compared with 232 PS-matched equivalents. No differences in baseline characteristics were noted. Matched patients (n=348; 116 with persistent hyponatraemia and 232 controls) had a mean age of 69.5±14.3 years, and 57.8% were male. The percentages of hypertension, diabetes mellitus and atrial fibrillation were 45.4%, 34.8% and 39.1%, respectively. The most frequent cause of HF was ischaemia (40.0%). Approximately one-third of these patients presented with de novo HF (37.4%), and half were classified as New York Heart Association (NYHA) class IV (48.0%). Prior evidence-based medical therapies included RAS blockers (50.0%), beta-blockers (23.6%) and aldosterone antagonists (44.8%) (table 2).

Echocardiographic characteristics

The mean LV and RV systolic function, indicated by LVEF and RV FAC, were 36.3% and 34.1%, respectively. RV FAC was significantly lower in patients with severe persistent hyponatraemia (28.2% vs 37.1%, p<0.001), whereas LVEF did 3

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies Table 2  Baseline characteristics in propensity score-matched patients Total patients (n=348)

Patients with severe persistent hyponatraemia (n=116)

 Age, year

69.5±14.3

70.0±14.0

69.3±14.5

0.663

  ≥65  years (%)

240 (69.0)

84 (72.4)

156 (67.2)

0.326

 Male (%)

201 (57.8)

68 (58.6)

133 (57.3)

0.818

 Current smoking (%)

49 (14.1)

13 (11.2)

36 (15.5)

0.276

 Alcohol (%)

110 (31.6)

37 (31.9)

73 (31.5)

0.935

 BMI, kg/m2

22.6±3.7

22.6±3.8

22.6±3.6

0.815

  ≥25  kg/m2 (%)

78 (22.4)

26 (22.4)

52 (22.4)

0.999

 Hypertension (%)

158 (45.4)

53 (45.7)

105 (45.3)

0.939

 Diabetes mellitus (%)

121 (34.8)

40 (34.5)

81 (34.9)

0.937

 Chronic kidney disease (%)

55 (15.8)

18 (15.5)

37 (15.9)

0.917

 Ischaemic heart disease (%)

80 (23.0)

29 (25.0)

51 (22.1)

0.542

 Atrial fibrillation (%)

136 (39.1)

48 (41.4)

88 (37.9)

0.534

 Chronic obstructive lung disease (%)

33 (9.5)

12 (10.3)

21 (9.1)

0.698

139 (40.0)

39 (33.6)

100 (43.1)

0.069

Variables

Patients without severe persistent hyponatraemia (n=232)

p

Demographics

Medical history

Aetiology  Ischaemic (%) Clinical status on admission  De novo heart failure (%)

130 (37.4)

43 (37.1)

87 (37.5)

0.938

 Systolic BP, mm Hg

117.9±27.4

119.3±28.3

117.1±27.0

0.220

 Diastolic BP, mm Hg

71.2±17.4

70.1±17.9

71.7±17.1

0.578

 Pulse rate, beats/min

87.9±25.8

83.9±24.0

89.9±26.5

0.185

 NYHA class≥III (%)

301 (86.5)

100 (86.2)

201 (86.6)

0.912

 Hb, g/dL

11.8±2.4

11.9±2.3

11.8±2.5

0.900

 sCr, mg/dL

1.7±1.9

1.7±1.7

1.7±2.1

0.941

Evidence-based medical therapy before admission  RAS blocker (%)

174 (50.0)

58 (50.0)

116 (50.0)

0.999

 Beta-blocker (%)

82 (23.6)

26 (22.4)

56 (24.1)

0.721

 Aldosterone antagonist (%)

156 (44.8)

54 (46.4)

102 (44.2)

0.672

 Loop diuretics (%)

323 (92.8)

110 (94.8)

213 (91.8)

0.304

 Nitrate (%)

215 (61.8)

72 (62.1)

143 (61.6)

0.938

 Amiodarone (%)

82 (23.6)

25 (21.6)

57 (24.6)

0.532

 Digitalis (%)

148 (42.5)

49 (42.2)

99 (42.7)

0.999

110 (31.6)

55 (47.4)

55 (23.7)

<0.001

Outcome  1-year mortality (%)

BMI, body mass index; BP, blood pressure; Hb, haemoglobin; LV, left ventricular; NYHA, New York Heart Association; RAS, renin angiotensin system; sCr, serum creatinine.

not differ (38.1% vs 35.5%, p=0.149). In addition, patients with severe persistent hyponatraemia showed significant TR (33.6% vs 22.0%, p=0.019), an estimated PASP of ≥40 mm Hg (65.5% vs 54.7%, p=0.035), and IVC plethora (47.0% vs 25.2%, p<0.001) more frequently (table 3). The measurement errors for RV FAC were small (differences within and between observers, –0.12%±3.95% and 0.10%±4.15%) (online supplementary figure). Reproducibility for RV FAC was good (intraclass correlation coefficients were 0.89 (intraobserver) and 0.80 (interobserver); p<0.001).

Severe persistent hyponatraemia and RV systolic function

RV dysfunction was observed in 81.0% (94/116) of patients with severe persistent hyponatraemia, and 33.6% (78/232) of patients without (χ2=69.6, p<0.001) (table 3). The strongest association of severe persistent hyponatraemia was with RV dysfunction (unadjusted OR 8.43, 95% CI 4.93 to 14.45, p<0.001); in contrast, no significant relationship was noted with LV dysfunction (unadjusted OR 1.22, 95% CI 0.96 to 1.51, p=0.091). Severe persistent hyponatraemia was associated with TR ≥grade 4

3 (unadjusted OR 1.80, 95% CI 1.10 to 2.95, p=0.020), IVC plethora (unadjusted OR 2.63, 95% CI 1.64 to 4.21, p<0.001) and a PASP of ≥40 mm Hg (unadjusted OR 1.57, 95% CI 1.00 to 2.49, p=0.049), suggesting a close correlation with RV dysfunction. In the multivariable logistic model, RV dysfunction, IVC plethora and TR ≥grade 3 maintained independent associations with severe persistent hyponatraemia, while LV dysfunction was not statistically significant (table 4).

Survival according to severe persistent hyponatraemia and RV systolic function During follow-up (median 24 months, IQR 6–36 months), 160 patients (46.0%) died in the matched cohort, two-thirds of which occurred within 1 year (n=110, 31.6%). Sixty-nine deaths during follow-up occurred in patients with severe persistent hyponatraemia, predominating in the subgroup with concomitant RV dysfunction (61/94 with RV dysfunction vs 8/22 without RV dysfunction). In univariable Cox regression analysis, ageing, anaemia with haemoglobin <11  g/dL, Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies Table 3  Echocardiographic characteristics in propensity score-matched patients Variables

Total patients (n=348)

Patients with severe persistent hyponatraemia (n=116)

Patients without severe persistent hyponatraemia (n=232)

LV ejection fraction, %

36.3±15.9

38.1±16.8

35.5±15.3

 <40% (%)

224 (64.4)

68 (58.6)

156 (67.8)

0.091

RVEDA, cm2

28.5±12.1

31.8±12.2

33.0±12.1

0.383

RVESA, cm2

19.1±9.3

24.3±10.0

22.5±8.7

0.088

RV fractional area change, %

34.1±10.4

28.2±10.2

37.1±9.1

<0.001

 <35%

172 (49.7)

94 (81.0)

78 (33.6)

<0.001

Tricupid regurgitation, grade

1.9±1.0

2.1±1.0

1.8±0.9

0.002

 ≥grade 3 (%)

90 (25.9)

39 (33.6)

51 (22.0)

0.019

Estimated PASP, mmHg

45.0±16.1

47.6±17.2

43.5±15.5

0.030 0.035

p 0.149

 ≥40 mm Hg (%)

203 (58.3)

76 (65.5)

127 (54.7)

Size of inferior vena cava (IVC), mm

21.5±5.9

21.9±5.3

21.2±6.2

0.319

IVC dilatation≥20 mm (%)

82 (23.6)

33 (28.7)

49 (21.8)

0.181

IVC plethora (%)

112 (32.2)

54 (47.0)

58 (25.2)

<0.001

LV, left ventricular; PASP, pulmonary artery systolic pressure; RV, right ventricle; RVEDA, right ventricular end-diastolic area; RVESA, right ventricular end-systolic area.

severe persistent hyponatraemia, prior use of beta-blockers and decreasing RV FAC were significant predictors for all-cause death (online supplementary table 1). Severe persistent hyponatraemia increased the risk of death by approximately 90% (unadjusted HR 1.91, p<0.001). RV dysfunction showed more than twofold hazard increase for death (unadjusted HR 2.28, p<0.001), whereas LV dysfunction failed to show any such association (unadjusted HR 1.02, p=0.700). In multivariable Cox models, severe persistent hyponatraemia was independently associated with poor clinical outcome (adjusted HR 1.86, 95% CI 1.35 to 2.54, p<0.001) after adjusting for age, haemoglobin <11 g/ dL and prior use of beta-blockers (model 2). The independent association between death and severe persistent hyponatraemia disappeared following inclusion of RV dysfunction in the model (model 4); conversely, statistical significance for severe persistent hyponatraemia remained regardless of LV dysfunction (model 3). RV dysfunction exhibited the strongest prognostic value for death, even after adjustment for other significant predictors, including severe persistent hyponatraemia (adjusted HR 2.20, 95% CI 1.53 to 3.15, p<0.001) (table 5). When we categorised patients into four groups according to sNa and RV FAC, clinical outcome was dependent on the presence of RV dysfunction. Kaplan-Meier survival curves revealed patients with severe persistent hyponatraemia and RV dysfunction had the worst prognosis, followed by those with RV dysfunction without hyponatraemia and those without RV dysfunction (p<0.001) (figure 2A). However, hyponatraemia was not associated with the clinical prognosis in patients without RV dysfunction. RV dysfunction was consistently associated with worse survival in

patients regardless of hyponatraemia (p=0.014 and p=0.004) (figure 2B, C).

Subgroup analysis according to LV systolic function

To assess the influence of LV function on developing severe persistent hyponatraemia in patients with AHF, subgroup analysis according to LVEF (≤40% and>40%) was performed (online supplementary table 2). No difference in the prevalence of severe persistent hyponatraemia was observed (33.5% vs 39.3%, p=0.096). Severe persistent hyponatraemia was strongly associated with RV dysfunction regardless of LVEF (≤40%: adjusted OR 7.15, 95% CI 3.55 to 14.41, p<0.001;>40%: adjusted OR 11.59, 95% CI 7.81 to 19.67, p<0.001). It alludes that RV function, rather than LV function, is significant in the development of hyponatraemia. Table 5  Multivariable Cox regression models for all-cause death Models

Adjusted HR (95% CI)

p

Model 1  Age

1.02 (1.01 to 1.03)

0.005

 Hb at admission <11 g/dL

1.49 (1.09 to 2.04)

0.015

 Prior use of beta-blocker

0.64 (0.42 to 0.96)

0.030

 Age

1.02 (1.01 to 1.03)

0.006

 Hb at admission <11 g/dL

1.45 (1.03 to 1.94)

0.032

 Prior use of beta-blocker

0.63 (0.42 to 0.94)

0.025

 Severe persistent hyponatraemia

1.86 (1.35 to 2.54)

<0.001

 Age

1.02 (1.01 to 1.03)

0.004

 Hb at admission <11 g/dL

1.42 (1.03 to 1.95)

0.023

 Prior use of beta-blocker

0.63 (0.42 to 0.94)

0.026

 Severe persistent hyponatraemia

1.86 (1.36 to 2.55)

<0.001

 LVEF <40%

1.26 (0.90 to 1.76)

0.175

 Age

1.02 (1.01 to 1.03)

0.001

 Hb at admission <11 g/dL

1.56 (1.13 to 2.15)

0.006 0.104

Model 2

Model 3

Table 4  Univariable and multivariable analysis for the association with severe persistent hyponatraemia Variables LV ejection fraction<40%

Unadjusted OR (95% CI) 1.22 (0.96 to 1.51)

Adjusted OR (95% CI)

p 0.091

1.21 (0.74 to 1.50)

p 0.308

Model 4

RV FAC<35%

8.43 (4.93 to 14.45)

<0.001

IVC dilatation

1.50 (0.90 to 2.50)

0.124

8.00 (4.50 to 14.22)

<0.001

IVC plethora

2.63 (1.64 to 4.21)

<0.001

1.52 (1.08 to 2.64)

0.033

 Prior use of beta-blocker

0.71 (0.47 to 1.07)

TR of ≥grade 3

1.80 (1.10 to 2.95)

0.020

1.14 (1.04 to 2.04)

0.046

 Severe persistent hyponatraemia

1.33 (0.94 to 1.87)

0.108

PASP≥40 mm Hg 1.57 (1.00 to 2.49)

0.049

1.18 (0.68 to 2.04)

0.558

 RV FAC<35%

2.20 (1.53 to 3.15)

<0.001

–

–

LV, left ventricular; FAC, fractional area change; IVC, inferior vena cava; PASP, pulmonary artery systolic pressure; RV, right ventricle; TR, tricuspid regurgitation.

Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Hb, haemoglobin; FAC, fractional area change;LVEF, left ventricular ejection fraction; RV, right ventricular.

5

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies

Figure 2  Kaplan-Meier survival curves according to combination of serum sodium and right ventricular (RV) fractional area change (FAC). (A) Curves stratified according to presence of persistent hyponatraemia or RV dysfunction (RVD). Patients with persistent hyponatraemia/with RVD had the worst prognosis of all subgroups. (B) In patients with persistent hyponatraemia, RVD could stratify the risk of event-free survival. (C) In patients without persistent hyponatraemia, RVD could also provide risk stratification for long-term clinical outcome. 

Discussion

The main findings of this study are as follows: (1) in AHF, severe persistent hyponatraemia was significantly and independently associated with RV dysfunction; (2) both severe persistent hyponatraemia and RV dysfunction were significant predictors of all-cause mortality during follow-up; (3) severe persistent hyponatraemia was not independently associated with death after adjusting for RV dysfunction, but maintained independence regardless of LV dysfunction; and (4) RV dysfunction improved risk prediction in patients with and without severe persistent hyponatraemia. 6

RV dysfunction as a key feature of hyponatraemia in HF

Although prior studies have evaluated hyponatraemia in HF patients,1 3–7 the nature of hyponatraemia and subsequent poor prognosis in HF remains unknown.20 Indeed, LV dysfunction alone cannot account for the aetiopathogenesis or clinical features of hyponatraemia.20–22 In an American cohort of 8862 ambulatory patients with HF (6185 with reduced LVEF and 2704 with preserved LVEF), hyponatraemia was prevalent at a similar frequency (>10%) and was an independent predictor in both patients regardless of LVEF.21 Rusinaru et al22 reported that hyponatraemia on admission was observed in 25.4% of patients with HF with preserved EF, an Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies incidence at least equivalent to that observed in large registries of HF with reduced EF.1 2 In agreement with previous findings, our analysis demonstrated that the prevalence of severe persistent hyponatraemia did not differ between patients with LVEF <40% and ≥40% (3.9% vs 4.5%; p=0.400). Additionally, LV dysfunction was not significantly associated with severe persistent hyponatraemia in the PS matched cohort (p=0.308). In contrast, RV dysfunction was highly prevalent, and had more than eightfold increased risk in patients with severe persistent hyponatraemia. Furthermore, severe persistent hyponatraemia was associated with IVC plethora and significant TR, congruent with RV dysfunction. These results provide the first evidence of a strong association between hyponatraemia and RV function in HF and a different viewpoint on the pathophysiological mechanism of hyponatraemia in HF.

Mechanism of hyponatraemia and RV dysfunction in HF

Though the pathogenesis of hyponatraemia in HF is multifactorial, it is generally assumed that hyponatraemia develops due to non-osmotic release of AVP resulting from a decrease in arterial baroreceptor stretch during low cardiac output in HF.23 24 However, low cardiac output can develop earlier in cases of RV dysfunction, due to decreased LV preload and ventricular interdependence, while LV dysfunction can mitigate the low cardiac output phenomenon by compensatory chamber dilatation.25 This is prominent in cardiogenic shock from RV infarction. Non-osmotic AVP can be also released by systemic arterial vasodilation during portal hypertension, which is frequently caused by liver cirrhosis but also occasionally by right HF.24 26 In some patients with severe and progressive RV failure, pulmonary arterial pressure and left atrial (LA) pressure may decrease as a consequence of small RV stroke volume.25 Baroreceptors in the LA inhibit AVP release and encourage water excretion in response to any increase in atrial pressure, a phenomenon known as the Henry-Gauer reflex.27 Theoretically, low LA pressure induced by severe RV dysfunction could weaken the HenryGauer reflex, and consequently disturb inhibition of AVP release. However, since the role of the RV in HF has been relatively overlooked until recently,25 further study regarding RV dysfunction and hyponatraemia is needed.

RV dysfunction as a key player of poor prognosis

RV dysfunction was the strongest independent predictor of longterm prognosis regardless of sNa status in this study, while the predictive power of hyponatraemia was attenuated by adjustment with RV dysfunction. In the multivariable Cox regression model, a 10% decrease in RV FAC was associated with >30% increase in the risk of death; indeed, only RV dysfunction significantly increased long-term mortality more than twofold, even after adjustment for hyponatraemia. Such results are consistent with previous studies demonstrating the independent prognostic value of RV function in advanced HF.9 28–30 Di Salvo et al noted that RV FAC ≥35% was a more powerful predictor of survival in HF than LVEF or exercise capacity.30 In prior research by our group,6 improvement of hyponatraemia during hospitalisation revealed no significant association with a better prognosis in patients with AHF, except for those having elevated blood urea nitrogen (BUN). Hyponatraemia may indirectly reflect the severity of RV dysfunction in patients with AHF, given that BUN elevation could indicate systemic congestion. This implies that the role of hyponatraemia as a predictor of longterm poor prognosis might derive from concomitant RV dysfunction and it is necessary to consider close monitoring and intensive treatment in AHF patients with RV dysfunction. Furthermore, RV dysfunction might be a therapeutic target in HF patients with hyponatraemia. Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Study limitations

First, as this study was an observational study from a prospective AHF registry, unmeasured confounding variables may have been present. Second, our definition of severe persistent hyponatraemia was arbitrary; however, severe persistent hyponatraemia as a metric is more suitable for assessment of relationships between RV dysfunction and hyponatraemia than instantaneous measurements since sNa can change temporally and dynamically with medication and fluid changes, rather than disease activity. Third, RV function was evaluated by 2D FAC only, which could be insufficient to accurately assess RV. However, as RV-focused apical four-chamber view was successfully obtained in all participants, and interobserver variability for RV FAC was good. Lastly, the study population was enrolled from an AHF registry in Korea; therefore, further international studies are needed to validate and generalise our results.

Key messages What is already known on this subject? ►► Hyponatraemia is the most common electrolyte imbalance in

heart failure (HF). It is one of the strongest prognostic factors of poor clinical outcome in patients with HF. ►► Previous reports have suggested that hyponatraemia is related with right HF. What might this study add?

►► This study provides the first evidence that the development of

hyponatraemia by right ventricular (RV) dysfunction (adjusted OR 8.00, 95% CI 4.50 to 14.22, p<0.001) in acute HF. Furthermore, RV dysfunction was an independent predictor of all-cause mortality (adjusted HR 2.20, 95% CI 1.53 to 3.15, p<0.001), while hyponatraemia was not (adjusted HR 1.33, 95% CI 0.94 to 1.87, p=0.108).

How might this impact on clinical practice?

►► This study helps us understand the role of RV dysfunction in

the development of hyponatraemia and its clinical relevance in HF. These findings can potentially improve HF management and clinical outcomes via identification of a novel therapeutic target. It also highlights the importance of considering the utility of RV dysfunction, instead of hyponatraemia, in HF risk models.

Contributors  Conception and design, HL, SEL, HJC, HYL, BHO; data acquisition, HL, SEL, CSP, JJP, GYL, MK; data analysis and interpretation, HL, SEL, CSP; statistical analysis, HL, SEL; drafting and finalising the article, HL, SEL, BHO; critical revision of the article for important intellectual content, HL, SEL, JJP, GYL, HJC, HYL, DJC, JOC, ESJ, JJK. Funding  This work was supported by grants from the Research of Korea Centres for Disease Controland Prevention [2010-E63003-00, 2011-E63002-00, 2012-E6300500, 2013-E63003-00,2013-E63003-01, 2013-E63003-02, and 2016-ER6303-00]. Competing interests  None declared. Patient consent  Obtained. Ethics approval  Seoul National University Hospital, Asan Medical Center, Seoul National University Bundang Hospital, Samsung Medical Center. Provenance and peer review  Not commissioned; externally peer reviewed. © Article author(s) (or their employer(s) unless otherwise stated in the text of the article) 2017. All rights reserved. No commercial use is permitted unless otherwise expressly granted.

References

1 Gheorghiade M, Abraham WT, Albert NM, et al. Relationship between admission serum sodium concentration and clinical outcomes in patients hospitalized for heart failure: an analysis from the OPTIMIZE-HF registry. Eur Heart J 2007;28:980–8.

7

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Heart failure and cardiomyopathies 2 Gheorghiade M, Rossi JS, Cotts W, et al. Characterization and prognostic value of persistent hyponatremia in patients with severe heart failure in the ESCAPE Trial. Arch Intern Med 2007;167:1998–2005. 3 Sato N, Gheorghiade M, Kajimoto K, et al. Hyponatremia and in-hospital mortality in patients admitted for heart failure (from the ATTEND registry). Am J Cardiol 2013;111:1019–25. 4 Kearney MT, Fox KA, Lee AJ, et al. Predicting death due to progressive heart failure in patients with mild-to-moderate chronic heart failure. J Am Coll Cardiol 2002;40:1801–8. 5 Goldberg A, Hammerman H, Petcherski S, et al. Hyponatremia and long-term mortality in survivors of acute ST-elevation myocardial infarction. Arch Intern Med 2006;166:781–6. 6 Lee SE, Choi DJ, Yoon CH, et al. Improvement of hyponatraemia during hospitalisation for acute heart failure is not associated with improvement of prognosis: an analysis from the Korean Heart Failure (KorHF) registry. Heart 2012;98:1798–804. 7 Konstam MA, Gheorghiade M, Burnett JC, et al. Effects of oral tolvaptan in patients hospitalized for worsening heart failure: the EVEREST Outcome Trial. JAMA 2007;297:1319–31. 8 Haddad F, Doyle R, Murphy DJ, et al. Right ventricular function in cardiovascular disease, part II: pathophysiology, clinical importance, and management of right ventricular failure. Circulation 2008;117:1717–31. 9 Bleasdale RA, Frenneaux MP. Prognostic importance of right ventricular dysfunction. Heart 2002;88:323–4. 1 0 Simon EE. Hyponatremia: Evaluation and treatment. New York: Springer, 2013. 11 Forfia PR, Mathai SC, Fisher MR, et al. Hyponatremia predicts right heart failure and poor survival in pulmonary arterial hypertension. Am J Respir Crit Care Med 2008;177:1364–9. 12 Haddad F, Peterson T, Fuh E, et al. Characteristics and outcome after hospitalization for acute right heart failure in patients with pulmonary arterial hypertension. Circ Heart Fail 2011;4:692–9. 13 Lee SE, Lee HY, Cho HJ, et al. Clinical Characteristics and Outcome of Acute Heart Failure in Korea: results from the Korean Acute Heart Failure Registry (KorAHF). Korean Circ J 2017;47:341–53. 14 Lee SE, Cho HJ, Lee HY, et al. A multicentre cohort study of acute heart failure syndromes in Korea: rationale, design, and interim observations of the Korean Acute Heart Failure (KorAHF) registry. Eur J Heart Fail 2014;16:700–8. 15 Sjölander A. Propensity scores and M-structures. Stat Med 2009;28:1416–20. 16 Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the american society of echocardiography and the european association of cardiovascular imaging. J Am Soc Echocardiogr 2015;39:e14.

8

17 Lancellotti P, Tribouilloy C, Hagendorff A, et al. Recommendations for the echocardiographic assessment of native valvular regurgitation: an executive summary from the european association of cardiovascular imaging. Eur Heart J Cardiovasc Imaging 2013;14:611–44. 18 Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the american society of echocardiography endorsed by the european association of echocardiography, a registered branch of the european society of cardiology, and the canadian society of echocardiography. J Am Soc Echocardiogr 2010;23:685–713. 19 Galiè N, Hoeper MM, Humbert M, et al. Guidelines for the diagnosis and treatment of pulmonary hypertension: the Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS), endorsed by the International Society of Heart and Lung Transplantation (ISHLT). Eur Heart J 2009;30:2493–537. 20 Bettari L, Fiuzat M, Felker GM, et al. Significance of hyponatremia in heart failure. Heart Fail Rev 2012;17:17–26. 21 Bavishi C, Ather S, Bambhroliya A, et al. Prognostic significance of hyponatremia among ambulatory patients with heart failure and preserved and reduced ejection fractions. Am J Cardiol 2014;113:1834–8. 22 Rusinaru D, Buiciuc O, Leborgne L, et al. Relation of serum sodium level to long-term outcome after a first hospitalization for heart failure with preserved ejection fraction. Am J Cardiol 2009;103:405–10. 23 Farmakis D, Filippatos G, Parissis J, et al. Hyponatremia in heart failure. Heart Fail Rev 2009;14:59–63. 24 Bansal S, Lindenfeld J, Schrier RW. Sodium retention in heart failure and cirrhosis: potential role of natriuretic doses of mineralocorticoid antagonist? Circ Heart Fail 2009;2:370–6. 25 Friedberg MK, Redington AN. Right versus left ventricular failure. Circulation 2014;129:1033–44. 26 Bloom S, Kemp W, Lubel J. Portal hypertension: pathophysiology, diagnosis and management. Intern Med J 2015;45:16–26. 27 Schrier RW. Role of diminished renal function in cardiovascular mortality: marker or pathogenetic factor? J Am Coll Cardiol 2006;47:1–8. 28 Ghio S, Gavazzi A, Campana C, et al. Independent and additive prognostic value of right ventricular systolic function and pulmonary artery pressure in patients with chronic heart failure. J Am Coll Cardiol 2001;37:183–8. 29 Kjaergaard J, Akkan D, Iversen KK, et al. Right ventricular dysfunction as an independent predictor of short- and long-term mortality in patients with heart failure. Eur J Heart Fail 2007;9:610–6. 30 Di Salvo TG, Mathier M, Semigran MJ, et al. Preserved right ventricular ejection fraction predicts exercise capacity and survival in advanced heart failure. J Am Coll Cardiol 1995;25:1143–53.

Lee H, et al. Heart 2017;0:1–8. doi:10.1136/heartjnl-2017-312084

Downloaded from http://heart.bmj.com/ on October 29, 2017 - Published by group.bmj.com

Hyponatraemia and its prognosis in acute heart failure is related to right ventricular dysfunction Heesun Lee, Sang Eun Lee, Chan Soon Park, Jin Joo Park, Ga Yeon Lee, Min-Seok Kim, Jin-Oh Choi, Hyun-jai Cho, Hae-Young Lee, Dong-Ju Choi, Eun-Seok Jeon, Jae-Joong Kim and Byung-Hee Oh Heart published online October 27, 2017

Updated information and services can be found at: http://heart.bmj.com/content/early/2017/10/27/heartjnl-2017-312084

These include:

References Email alerting service

This article cites 29 articles, 10 of which you can access for free at: http://heart.bmj.com/content/early/2017/10/27/heartjnl-2017-312084 #BIBL Receive free email alerts when new articles cite this article. Sign up in the box at the top right corner of the online article.

Notes

To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions To order reprints go to: http://journals.bmj.com/cgi/reprintform To subscribe to BMJ go to: http://group.bmj.com/subscribe/