Article Text

Download PDFPDF

Body volume is the major determinant of worsening renal function in acutely decompensated heart failure with reduced left ventricular ejection fraction
Free
  1. Mei Han Ho1,
  2. Duo Huang1,
  3. Chi-Wai Ho1,
  4. Ming-Liang Zuo2,
  5. An-Guo Luo2,
  6. Emmanuel Cheung1,
  7. Mi Zhou1,
  8. Yangyang Cheng1,
  9. Mingya Liu3,
  10. Kai-Hang Yiu1,3,
  11. Chu Pak Lau1,
  12. Pauline Yeung4,
  13. Wen Sheng Yue5,
  14. Li-Xue Yin2,
  15. Hung Fat Tse1,
  16. Wei Jiang6,
  17. Zhen Lei6,
  18. Xin-Li Li7,
  19. M Cowie8,
  20. Chung Wah Siu1,3
  1. 1 Cardiology Division, Department of Medicine, University of Hong Kong, Hong Kong, Hong Kong
  2. 2 Department of Echocardiography & Non-invasive Cardiology Laboratory, Sichuan Academy of Medical Sciences & Sichuan Provincial People’s Hospital, Chengdu, China
  3. 3 Cardiology Division,Department of Medicine, University of Hong Kong-Shenzhen Hospital, Shenzhen, China
  4. 4 Respiratory Division, Department of Medicine, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China
  5. 5 Medical Imaging Key Laboratory, Affiliated Hospital of North Sichuan Medical College, Nanchong, Sichuan, China
  6. 6 Department of medicine, Union Shenzhen Hospital, Huazhong University of Science and Technology, Shenzhen, China
  7. 7 Department of Cardiology, The First Affiliated Hospital of Nanjing Medical University, NanJing, China
  8. 8 Imperial College London, Royal Brompton Hospital, London, UK
  1. Correspondence to Professor Chung Wah Siu, Medicine, Hong Kong, China; cwdsiu{at}yahoo.com.hk

Abstract

Aims Little is known about the relative importance of body volume and haemodynamic parameters in the development of worsening of renal function in acutely decompensated heart failure (ADHF). To study the relationship between haemodynamic parameters, body water content and worsening of renal function in patients with heart failure with reduced ejection fraction (HFrEF) hospitalised for ADHF.

Methods and results This prospective observational study involved 51 consecutive patients with HFrEF (age: 73±14 years, male: 60%, left ventricular ejection fraction: 33.3%±9.9%) hospitalised for ADHF. Echocardiographic-determined haemodynamic parameters and body volume determined using a bioelectric impedance analyser were serially obtained. All patients received intravenous furosemide 160 mg/day for 3 days. There was a mean weight loss of 3.95±2.82 kg (p<0.01), and brain natriuretic peptide (BNP) reduced from 1380±901 pg/mL to 797±738 pg/mL (p<0.01). Nonetheless serum creatinine (SCr) increased from 134±46 μmol/L to 151±53 μmol/L (p<0.01), and 35% of patients developed worsening of renal function. The change in SCr was positively correlated with age (r=0.34, p=0.017); and negatively with the ratio of extracellular water to total body water, a parameter of body volume status (r=−0.58, p<0.001); E:E’ ratio (r=−0.36, p=0.01); right ventricular systolic pressure (r=−0.40, p=0.009); and BNP (r=−0.40, p=0.004). Counterintuitively, no correlation was observed between SCr and cardiac output, or total peripheral vascular resistance. Regression analysis revealed that normal body volume and lower BNP independently predicted worsening of renal function.

Conclusions Normal body volume and lower serum BNP on admission were associated with worsening of renal function in patients with HFrEF hospitalised for ADHF.

  • heart failure
  • ischaemic heart disease
  • acute renal failure

Data availability statement

Data are available from the corresponding author

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Heart failure (HF) has become globally epidemic in the realm of cardiovascular medicine.1 The prevalence of HF ranges from 1% to 2% of the adult population with an incidence that varies between 0.6 and 10 per 1000 population per year.2–8 Over the past three decades, advances in pharmacological and device therapies have substantially improved the prognosis of ambulatory patients with chronic HF with reduced left ventricular ejection fraction (LVEF).9 In stark contrast, the management of patients hospitalised for acutely decompensated heart failure (ADHF) has remained suboptimal and largely unchanged.9 For the vast majority of patients hospitalised for ADHF, the main symptom is related to pulmonary and/or systemic congestion. An intravenous loop diuretic is prescribed almost universally to patients hospitalised for ADHF, and is recommended as the first-line therapy by international guidelines.9 Nonetheless during the treatment course of ADHF, worsening of renal function10 11 occurs in up to 39% of patients and is associated with prolonged hospitalisation and high in-hospital mortality and overall mortality.11 12

Traditionally, worsening of renal function in ADHF is thought to be the result of a reduced cardiac output that diminishes renal perfusion leading to activation of the renin–angiotensin–aldosterone system and consequent renal dysfunction (arterial hypoperfusion hypothesis).13 Nonetheless, recent studies have convincingly demonstrated that the reduced cardiac output is not the major driving force for worsening renal function in ADHF.14 Instead, increased vascular tone and increased body water in addition to reduced cardiac performance are the three main pathophysiological abnormalities encountered in ADHF. Little is known about the relative importance of these parameters in ADHF, particularly in relation to the development of worsening of renal function. In the present study, we sought to investigate the relationship between different echocardiographic-determined haemodynamic parameters, body volume determined using a bioelectric impedance analyser, and the change to serum creatinine in patients with heart failure with reduced ejection fraction (HFrEF) hospitalised for ADHF and treated with a fixed dose of intravenous furosemide.

Methods

Study participants

We recruited patients with HFrEF hospitalised at Queen Mary Hospital, Hong Kong. The diagnosis of ADHF was based on (1) the presence of symptoms and signs of HF, defined according to a Framingham Heart Failure Score ≥2 at the time of admission (figure 1)15; (2) radiographic evidence of pulmonary congestion on chest radiography; and (3) elevated serum concentration of brain-type natriuretic peptide (BNP) >100 pg/mL. The diagnosis of ADHF was only be made when the three conditions were met. Patients were excluded if they had had acute coronary syndrome (ACS) within the last 4 weeks, complex congenital heart disease, significant valvular stenosis, systolic blood pressure ≤100 mm Hg, fast atrial fibrillation (AF) with ventricular rate >100 beats/min on admission, ventricular fibrillation or sustained ventricular tachycardia, third-degree or Mobitz type II heart block without a permanent pacemaker, or renal impairment with serum creatinine ≥250 μmol/L.

Figure 1

Flow of the study cohort. ACS, acute coronary syndrome; ADHF, acutely decompensated heart failure; AF, atrial fibrillation; BNP, brain natriuretic peptide; JVP, jugular venous pressure; LVEF, left ventricular ejection fraction; PND: paroxysmal nocturnal dyspnoea; S3: third heart sound.

Study design

This was a single-centre, prospective, observational study. Demographics and cardiovascular comorbidities were recorded. Quantitative transthoracic echocardiography and measurement of body water content as determined by bioelectric impedance analysis were performed daily after admission for three consecutive days. Urine output, body weight and serum creatinine were likewise assessed daily. In addition, serum BNP and serum neutrophil gelatinase-associated lipocalin (NGAL), a biomarker for acute kidney injury, were measured on the first and third day of admission using a point-of-care immunofluorescence essay (Alere Triage CardioRenal Panel, Alere, USA).

Echocardiographic examination

Detailed quantitative transthoracic echocardiography was performed as previously described.16 17 Two-dimensional, M-mode, Doppler flow and tissue Doppler imaging studies were performed in all subjects using the Vivid 9 ultrasound system (GE Healthcare, USA) with a 3.5 MHz transducer.18 LVEF was determined using the biplane Simpson’s method. A standard Doppler echocardiographic method was used to estimate cardiac output.19 To calculate cardiac output, an average of five consecutive heartbeats during sinus rhythm and 13 beats during AF was obtained.20 Blood pressure was simultaneously measured during determination of cardiac output with the use of a calibrated semiautomatic device (Dinamap 1846XT, Critikon Corporation, Tampa, Florida, USA). Total peripheral resistance was derived using the following formula:

Embedded Image

The ratio of early transmitral inflow velocity (E) to early diastolic mitral annular velocity (E’) was determined by pulsed-wave Doppler and tissue Doppler imaging technique.21 Right ventricular systolic pressure (RVSP) was estimated from the velocity of tricuspid regurgitation using continuous-wave Doppler echocardiography with addition of right atrial pressure.22

Body composition measurement

Body composition parameters including total body water and segmental water were measured using the InBody 430 analyzer (Biospace Co., Seoul, Korea) with patients in a supine or sitting position. The body composition analyser used a tetrapolar, eight-point tactile electrode system that separately measured impedance of the arms, trunk and legs at a frequency of 1 kH to 1 MHz to determine total body water, segmental water, limb-to-truck segmental water ratio, and ratio of extracellular water to total body water (ECW/TBW). Patients were considered hypervolaemic when the ratio of ECW/TBW was ≥0.4 according to the manufacturer’s manual.23 24

Statistical analysis

Continuous variables are expressed as mean±SD and categorical variables as frequencies and percentages. Statistical comparisons were performed using Student’s t-test, χ2 test and Fisher exact test as appropriate. The relationship between baseline variables and change in renal function was analysed by bivariate correlation. Multivariable logistic regression was performed to investigate the relationship between baseline characteristics and occurrence of cardiorenal syndrome. Variables with number of events fewer than 10 were not included. To avoid multicollinearity between the univariate predictors, a correlation coefficient r<0.7 was used. Remaining variables with a p value<0.10 as determined by the univariate analyses were added into a stepwise backward multivariable regression model. Receiver operating characteristic curve analysis was used to assess the predictive value of baseline variables for cardiorenal syndrome. A p value<0.05 was considered statistically significant. Data analysis was performed using IBM SPSS Statistics 21 (SPSS).

Results

During the study period, 229 patients were hospitalised with typical symptoms and signs, and radiographic features of ADHF. Of these, 170 patients (74.2%) were confirmed to have ADHF with serum BNP level >100 pg/mL. Among these 170 patients with ADHF, only 59 (34.7%) had a LVEF<50%. Six patients were excluded from the final analysis because of concomitant medical conditions: three with renal impairment and SCr ≥250 μmol/L, two with recent ACS, and one with fast AF. Another two patients discharged before the completion of all assessments were also excluded. The final analysis included 51 patients with ADHF and reduced LVEF, and all were in New York Heart Association Function Class III (figure 1).

Table 1 summarises their baseline demographics. The mean age was 73±14 years and 60.8% were men. The majority of patients had hypertension (66.7%), coronary artery disease (52.9%), diabetes mellitus (52.9%) and/or AF (52.9%). The mean BNP was 1380±901 pg/mL. Prior to the index admission, over half of all patients (58.8%) were prescribed a beta-blocker, 29 (56.9%) received ACE inhibitor or angiotensin receptor blocker, and 15 (29.4%) a mineralocorticoid receptor blocker. In addition, 30 patients (58.8%) received regular frusemide prior to the index admission with the mean daily dose of 64.7±44.2 mg. Table 2 summarises the baseline haemodynamic and body water parameters of the study population. On admission, the mean systolic pressure and diastolic pressure were 130±20 mm Hg and 80±14 mm Hg, respectively. The mean heart rate was 83±21 beats/min and mean LVEF was 33.3%±9.9%. Body composition analysis revealed that 37 patients (72.5%) were hypervolaemic as determined by the ratio of ECW/TBW with the cut-off at 0.4.

Table 1

Baseline characteristics of the study population

Table 2

Baseline haemodynamic and body water parameters

After admission, all patients received intravenous furosemide at a total daily dose of 160 mg for the first 3 days of hospitalisation. During this time, there was a mean weight loss of 3.95±2.82 kg (from 64.0±11.2 kg on day 1 to 60.0±11.1 kg on day 3, p<0.01) (figure 2A). Body composition analysis revealed a concomitant reduction in mean total body water from 33.6±7.8 L to 30.1±7.0 L (p<0.001), segmental body water (limbs and trunk), and a reduced ratio of ECW/TBW from 0.41±0.02 to 0.40±0.01 (p<0.001) (figure 2B). Only 20 patients (39%) remained hypervolaemic as determined using the ratio of ECW/TBW on day 3 compared with 37 patients (72.5%) on day 1 (p=0.003). At the same time, there was a reduction in mean arterial pressure from 94±16 mm Hg to 84±14 mm Hg (p<0.01), a progressive reduction in E/E’, a surrogate of pulmonary capillary wedge pressure, from 18.1±7.9 to 15.9±6.7 (p<0.01), and RVSP, a surrogate of pulmonary pressure from 52.7±19.5 mm Hg to 36.5±13.7 mm Hg (p<0.01). Interestingly, there were no statistically significant changes to cardiac output or total vascular resistance (figure 2C). Although mean serum BNP level fell from 1380±901 pg/mL on day 1 to 797±738 pg/mL on day 3 (p<0.01), there was a concomitant rise in mean serum creatinine from 134±46 μmol/L on day 1 to 151±53 μmol/L on day 3 (p<0.01) with estimated glomerular filtration rate reduced from 47.8±18.3 mL/min to 40.8±14.5 mL/min (p<0.01), together with an increase in serum NGAL concentration from 176±100 ng/mL to 230±163 ng/mL during the same period of time. (figure 2D) From days 1 to 3, the mean daily net fluid output decreased from 1240±1280 mL/day to 752±817 mL/day (p<0.05), indicating a progressive reduction in diuresis (figure 2A).

Figure 2

Serial changes in (A) body weight and daily net fluid loss, (B) body water composition, (C) haemodynamic parameters, and (D) biochemical markers of the study population during the 3 days hospitalisation period. BNP, brain natriuretic peptide; CO, cardiac output; ECW:TBW, ratio of extracellular water to total body water; MAP, mean arterial pressure; NGAL, neutrophil gelatinase-associated lipocalin; RVSP, right ventricular systolic pressure; TPR, total peripheral resistance.

To analyse the mechanisms of worsening renal function, we examined the correlations of various parameters with change in serum creatinine. Change in serum creatinine was correlated positively with increasing age (r=0.34, p=0.017); and negatively with E:E’ (r=−0.36, p=0.01), RVSP (r=−0.40, p=0.009), serum BNP concentration (r=−0.40, p=0.004), and ECW/TBW ratio (r=−0.58, p<0.001) (figure 3). Of these, the ratio of ECW/TBW had the strongest correlation. Counterintuitively, haemodynamic parameters including mean arterial pressure, cardiac output, and total peripheral resistance were not correlated with change in SCr (figure 3). Likewise, neither serum creatinine at baseline nor NGAL concentration was correlated with change in serum creatinine.

Figure 3

Relationship between change in serum creatinine concentration during the first 3 days and (A) age, (B) left ventricular ejection fraction, (C) mean arterial pressure, (D) cardiac output, (E) total peripheral resistance, (F) E:E’ ratio, (G) right ventricular systolic pressure, (H) ratio of extracellular water to total body water, (I) serum brain natriuretic peptide concentration, (J) serum creatinine concentration on day 1, and (K) serum neutrophil gelatinase-associated lipocalin concentration.

To further evaluate the relationship of change in renal function with other potential clinical, haemodynamic and biochemical determinants, patients were categorised according to the status of worsening of renal function. Worsening of renal function was defined by a rise of SCr >26.5 µmol/L from baseline to the third admission day. Altogether, 18 patients (35.3%) developed worsening of renal function. As shown in table 1, there were no statistically significant differences in comorbidities between patients with worsening of renal function and those without. Baseline cardiac output, blood pressure, total peripheral resistance also did not differ significantly between patients with and without worsening of renal function although those with worsening of renal function had a lower baseline serum BNP concentration than those without (892±613 pg/mL vs 1647±928 pg/mL, p=0.003), lower baseline E:E’ (14.2±6.1 vs 20.2±8.0, p=0.008), and a smaller ratio of ECW/TBW (0.399±0.012 vs 0.412±0.014, p=0.002). (table 2) Nonetheless there was no statistically significant difference in net daily fluid output (1270±1039 mL/day vs 1220±1438 mL/day, p=0.908) or total weight lost (3.0±2.6 kg vs 4.5±2.8 kg, p=0.061).

On univariate analysis, high E/E’, hypervolaemia defined as the ratio of ECW/TBW≥0.4, and a higher BNP were negatively associated with the occurrence of worsening of renal function. On multivariate analysis, only hypervolaemia (HR: 0.12, 95% CI 0.02 to 0.92) and high BNP (HR: 0.99, 95% CI 0.99 to 1.00) remained protective of the occurrence of worsening of renal function (table 3). When categorising patients with ADHF according to their body volume status using the ratio of ECW/TBW≥0.4, 64.3% of patients with euvolaemic ADHF developed worsening of renal function in stark contrast to 24.3% of those with hypervolaemic ADHF.

Table 3

Association between baseline factors and cardiorenal syndrome

Furthermore, while preadmission HF medications remained unchanged during the 3 days study period, 2 out 29 patients previously receiving ACEI/ARB had ACEI/ARB and MRA discontinued because of worsening renal function on hospital discharge. In addition, 4 out of 30 patients previously on betablocker had betablocker discontinued due to lowish of heart rate or blood pressure on discharge. On the other hand, ACEI/ARB, MRA, and beta-blocker were initiated in 8 (15.6%), 10 (19.6%) and 9 (17.6%) patients on discharge, respectively.

Discussion

In the present study, we investigated the occurrence of worsening of renal function and its determinants in a cohort of patients with HFrEF hospitalised for ADHF. First, patients with HFrEF hospitalised for ADHF were at high risk of worsening renal function, and up to 35% developed worsening of renal function during the course of treatment. Second, we showed that change in serum creatinine concentration was correlated negatively with the ratio of ECW/TBW, a parameter of body volume status; E:E’ ratio, a surrogate of pulmonary capillary wedge pressure; RVSP; and serum BNP concentration. Of these, the ratio of ECW/TBW had the strongest correlation with change in serum creatinine concentration. Further regression analysis consistently revealed that a small ratio of ECW/TBW, that is, euvolaemia, and a lower serum BNP concentration at baseline were independently predictive of the development of worsening renal function. Counterintuitively, there was a total lack of correlation between cardiac output, total peripheral vascular resistance, and baseline renal function with worsening renal function.

Worsening of renal function is common in patients with HFrEF who are hospitalised with ADHF. In the contemporary Diuretic Optimisation Strategies Evaluation trial that evaluated different strategies for diuretic dosing and delivery in the setting of ADHF, 18.3% patients hospitalised with ADHF developed worsening of renal function within the first 72 hours of hospitalisation.25 On the contrary, in a real-life observational registry consisting of 1002 patients hospitalised for ADHF, more than twice as many (39%) developed worsening of renal function during the index hospitalisation.10 11 Similarly, in the present study, 35.3% of patients with HFrEF hospitalised for ADHF had worsening of renal function within 72 hours. Importantly, worsening of renal function during the treatment course of ADHF has been demonstrated to be associated with poor outcomes including prolonged hospitalisation, in-hospital mortality and overall mortality.11 12

Traditionally, it is believed that reduced stroke volume and cardiac output in the setting of ADHF directly contribute to worsening of renal function and consequent cardiorenal syndrome as a result of inadequate renal perfusion (arterial hypoperfusion hypothesis).13 Nonetheless, only 1%–2% patients with ADHF have a genuinely low cardiac output that leads to end-organ hypoperfusion. In the vast majority of patients with HFrEF hospitalised for ADHF, the main symptom is related to pulmonary and/or systemic congestion. There is an increasing body of evidence that apart from patients with cardiogenic shock, reduced cardiac output is only one of many underlying factors that contribute to worsening renal function in patients hospitalised with ADHF and only congestive symptoms. A recent study of 575 patients from the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial and registry, who were hospitalised with ADHF and who underwent pulmonary artery catheterisation to guide therapy, has convincingly demonstrated that reduced cardiac index (cardiac output) is not the major drive for worsening renal function in the setting of ADHF.14 These findings are in accordance with an absent correlation between cardiac output at baseline and change in serum creatinine concentration in the present study. Intriguingly, over the first 72 hours of hospitalisation, although clinical parameters including body weight, E/E’, a surrogate of pulmonary capillary wedge pressure, RVSP, a surrogate of systemic congestion, and serum BNP concentration progressively improved, cardiac output remained largely unchanged, suggesting that cardiac output is not the prime therapeutic target in the management of patients hospitalised for ADHF without cardiogenic shock.

Alternatively, systemic venous congestion, which increases pressure in the Bowman’s capsule and tubular system of the kidney and thereby opposes filtration, has been increasingly recognised as another important factor that contributes to impaired renal function in patients with HF. However, published data concerning the relationship between systemic venous congestion and worsening of renal function among patients hospitalised for ADHF appear conflicting.26 27 For instance, in a study involving 145 patients hospitalised for AHDF with a pulmonary artery catheter used to guide intensive medical therapy, those with greater central venous pressure on admission were more likely to develop worsening of renal function during the index hospitalisation.27 On the contrary, in the ESCAPE trial, no correlation was observed between baseline right atrial pressure obtained through pulmonary artery catheterisation and worsening renal function in patients hospitalised for ADHF.26 One of the possible explanations for the discrepancy would be related to the heterogeneity of dosage of intravenous frusemide in these studies. Indeed, there has been a wide variation of the reported dosage of intravenous frusemide for hospitalised patients with ADHF from 80 mg/day to up to 320 mg/day.28 29 In order to minimise the possible confounding effects of frusemide dosage, in the present study, all study participants received the same total intravenous dose of frusemide (160 mg/day) for the first 3 hospitalisation days. However, it should be emphasised that the more recent international guideline for the management of acute HF recommends the initial dosage of intravenous diuretic to be determined based on the preceding dosage of oral diuretics.30

In the present study, patients with smaller E:E’ ratio reflecting a lower pulmonary capillary wedge pressure, lower RVSP representing a lower central venous pressure and euvolaemia as determined by bioelectric impedance measurement at baseline were associated with the occurrence of cardiorenal syndrome in univariate analysis. In regression analysis, only euvolaemia, not E/E’ or RVSP, was independently associated with the occurrence of worsening renal function during the treatment course. This may be at least partly explained by the heterogeneity of the study populations as well as the acute ADHF management. In fact, international guidelines recommend assessing patients with ADHF based on a bedside evaluation of congestion (dry or wet) and perfusion (warm or cold), and classifying them into four main haemodynamic subtypes: (1) ‘wet and warm’, (2) ‘wet and cold’, (3) ‘dry and cold’, or (4) ‘dry and warm’ to guide acute treatment.9 Patients with ADHF classified as wet and warm, the most common haemodynamic profile subtype, will be further classified into ‘vascular type’, for which vasodilator is the preferred first-line treatment; or ‘cardiac type’ for which diuretic is the preferred first-line treatment.9 As shown in our study, patients with ADHF and apparently less severe volume overload and lower BNP were more likely to develop worsening of renal function during the first 72 hours when given the same dose of diuretic therapy as patients with genuine hypervolaemia. This observation may be due to the lower susceptibility to overdiuresis of more hypervolaemic patients with ADHF. Nonetheless although haemodynamic profiling is the key to successful management of ADHF and concurs with our findings in the present study, clinical assessment of volume status in ADHF remains largely qualitative rather than quantitative, requiring good clinical skills and experience. In real clinical practice, current clinical evaluation based on physical examination findings and radiographic signs lacks the required sensitivity or specificity to quantify the degree of congestion or volume status of individual patients with ADHF.31–34 In the setting of ADHF, the assessment of haemodynamic status has long been overshadowed by measurement of pressure rather than volume. It is because pressure can be readily measured either invasively (directly by pressure catheter) or non-invasively (indirectly by flow velocity). However, central venous pressure and pulmonary capillary wedge pressure, although often considered as the ‘gold standard’ for evaluating haemodynamic congestion,35 do not correlate with circulating blood volume and body volume. In fact, elevated pulmonary capillary wedge pressure and central venous pressure can be a consequence of decreased cardiac pump function, increased pulmonary arterial resistance, increased intrathoracic pressure, and/or venoconstriction secondary to increased sympathetic discharge and/or activation of the renin–angiotensin system without a real fluid overload or hypervolaemia. In the other words, congestion is not synonymous with hypervolaemia, but often congestion is accompanied by hypervolaemia. Likewise increased filling pressure is not always associated with hypervolaemia; aggressive diuretic prescription in patients with ADHF with apparently high filling pressure (raised central venous pressure and pulmonary oedema) but with relatively normal body volume may result in worsening of renal function. As shown in the present study, compared with patients with ADHF with hypervolaemia, patients with ADHF in a euvolaemic state were more likely to develop worsening of renal function when given the same dose of intravenous diuretic. Furthermore, in the regression analysis, only volume status at baseline as determined using bioelectric impedance measurement, and not E:E’ or RVSP, predicted worsening of renal function and cardiorenal syndrome. Accurately differentiating patients with HFrEF hospitalised with ADHF with true volume overload (hypervolaemic ADHF) from those with normal volume status (euvolaemic ADHF) is thus of clinical importance in guiding pharmacological treatment to avoid worsening of renal function. Recent developments in bioimpedance spectroscopy have made possible a more accurate assessment of body fluid status, and has been implemented in clinical fluid management, for example, in patients receiving dialysis.23

Last but not least, it has recently been highlighted in a review article that worsening of renal function may not necessarily translate into poor prognosis in patients with ADHF.36 For instance, increased in serum creatinine concentration as a result of haemoconcentration, and complete decongestion may indicate clinical improvement (pseudoworsening of renal function) rather than deterioration as in genuine worsening of renal function.36 This suggests worsening of renal function per se may not be a good surrogate endpoint to indicate long-term outcomes. Furthermore, this highlights the heterogeneity of both ADHF phenotypes as well as the clinical responses to diuretic, which necessitates proper differentiation in order to guide the treatment.

Our findings may have several important clinical implications. First, patients with HFrEF hospitalised with ADHF are at high risk of worsening renal function. Second, as body volume is the main determinant of the development of worsening renal function in ADHF, incorporating bioimpedance measurement in regular clinical practice to further subclassify patients with ADHF of ‘wet and warm’ haemodynamic subtype into hypervolaemic ADHF or euvolaemic ADHF may help reduce the risk of worsening renal function by guiding therapeutic strategy. While conceptually logical, the ‘ideal’ management strategies for individual volume subtypes have yet to be established and are challenging. This is particularly true for those with euvolumic ADHF. It remains uncertain whether vasodilators, either arterial or venous, should be the first-line treatment in patients with euvolumic ADHF. Even in the setting of hypervolemic ADHF, in which a therapeutic strategy to remove excessive body fluid appears to be logical and appropriate, the best mode (diuretic therapy or ultrafiltration) and optimal rate of fluid removal remain to be determined. Nonetheless, further classification of heterogeneous patients with ADHF into different volume subgroups with prognostic implications represents an important next step in the development of ADHF therapeutic strategies (figure 4).

Figure 4

Proposed haemodynamic classification of patients with acutely decompensated heart failure (ADHF) based on clinical profiles.

Limitations

This observational study had some limitations. First, although the definition of worsening of renal function, an absolute rise of serum creatinine >26.5 μmol/L, was frequently used by various investigators,10 11 37 38 and was demonstrated to be predictive of a poor outcome in patients with ADHF, there is no universally accepted cut-off value for the definition. Second, in the present study, the proportion of patients receiving guideline recommended HF medications prior to the hospitalisation was low, thereby the result may not be generalised to other patient population. In addition, it should be emphasised that the study included both patients receiving regular oral frusemide prior to the index admission and not, but all received the same dosage of intravenous frusemide, which was in contrast to the current international guideline for acute HF.30 Third, the present study was performed before the newly defined HF with mid-range EF (LVEF between 40% and 49%) widely advocated. Forth, while we showed that change in serum creatinine concentration was correlated with various parameters such as body volume status, E/E’, RVSP and serum BNP concentration potentially facilitating categorising patients with ADHF into different subtypes, the correlations were only modest with r-values less than 0.7. Finally, haemodynamic parameters were based on estimates from quantitative echocardiographic assessment instead of invasive haemodynamic assessments so precise measurements of cardiac output, total peripheral vascular resistance, and left ventricular filling pressure were not obtained. Nonetheless, a good correlation between haemodynamic measurements determined by Doppler echocardiography and invasive cardiac catheterisation in patients with ADHF has been previously established.39 The echocardiographic approach is more compatible with and applicable to real-world clinical practice because routine invasive haemodynamic monitoring is not appropriate in the vast majority of patients with ADHF.

Conclusions

Taken collectively, worsening renal function is a frequent complication in hospitalised ADHF in patients with HFrEF. Body volume is the most important determinant for the development of worsening renal function in these patients. Incorporation of body volume measurement into clinical practice may facilitate selection of therapeutic strategies in patients with HFrEF who are hospitalised for ADHF.

Current research questions

  • What is the prevalence of worsening renal function in patients with HFrEF hospitalised with acute decompensated heart failure?

  • What is/are the clinical, biochemical and/or haemodynamic determinants of worsening renal function in the setting of acute decompensated heart failure in patients with HFrEF?

  • How to stratified HFrEF with acute decompensated heart failure in order to prevent worsening of renal function?

Main messages

  • Patients with heart failure with reduced ejection fraction (HFrEF) hospitalised for acute decompensated heart failure were at high risk of worsening renal function, and up to 35% developed worsening of renal function during the course of treatment.

  • Euvolaemia and a lower serum brain-type natriuretic peptide concentration on admission were independently predictive of the development of worsening renal function.

  • Counterintuitively, there was a total lack of correlation between cardiac output, total peripheral vascular resistance, and baseline renal function with worsening renal function.

What is already known on the subject?

  • Although intravenous loop diuretic is prescribed to patients hospitalized for ADHF, worsening of renal function occurs in up to 39% of patients and is associated with prolonged hospitalization and high in-hospital mortality, and overall mortality. Increased vascular tone and increased body water in addition to reduced cardiac performance are the three main pathophysiologic abnormalities for worsening renal function in ADHF.

Data availability statement

Data are available from the corresponding author

Ethics statements

Patient consent for publication

Ethics approval

The study complies with the principles outlined in the Declaration of Helsinki, that the local Institutional Review Board has approved the research protocol and that informed consent has been obtained from all participants.

References

Footnotes

  • MHH, DH, C-WH and M-LZ contributed equally.

  • Correction notice This article has been corrected since it first published. The provenance and peer review statement has been included.

  • Contributors M-HH: conception, design, patient recruitment and assessment, data analysis, manuscript drafting and review. DH: conception, design, echocardiographic assessment, data analysis, manuscript review. C-WH and EC: conception, design, patient recruitment and assessment, data analysis, manuscript review. M-LZ and A-GL: conception, design, echocardiographic assessment, data analysis, manuscript review. MZ and YC: echocardiographic assessment, data analysis, manuscript review. ML, K-HY, PY, WSY, L-XY, HFT, WJ, ZL, X-LL and MC: data review and manuscript review. CPL: conception, design, and manuscript review. CWS: conception, design, echocardiographic assessment, data analysis, manuscript review and final edition.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.