Abstract
Aims
Previous studies demonstrated the relationship between hypochloraemia and poor prognosis in patients hospitalized for acute heart failure (AHF). However, the usefulness of chloride in clinical practice remains uncertain, notably in very old patients with predominantly heart failure (HF) with preserved ejection fraction (HFpEF). We aimed to evaluate the prognostic impact of chloride in a cohort of very aged patients with AHF and the possible existence of different phenotypes of hypochloraemia with distinct clinical significance.
Methods and results
It was an observational study of 429 patients hospitalized for AHF in which chloraemia was measured. Two different phenotypes of hypochloraemia were identified by their relationship with estimated plasma volume status (ePVS) as a proxy of intravascular congestion. The endpoint of interest was time to all‐cause mortality and the composite of death and/or HF readmission. A multivariable Cox proportional hazard regression model was constructed to analyse the endpoints. The median age was 85 (78–92) years, 266 (62%) were women, and 80% had HFpEF. After multivariable analysis, chloraemia, but not natraemia, was associated with the risk of death and HF readmission in a U‐shaped pattern. The phenotype characterized by hypochloraemia and low ePVS (depletional) was associated with an increased risk of mortality when compared with patients with normochloraemia [hazard ratio (HR) 1.86, P = 0.008]. In contrast, hypochloraemia with high ePVS (dilutional) had no prognostic significance (HR 0.94, P = 0.855).
Conclusions
In very old patients hospitalized with AHF, plasma chloride was associated with the risk of death and HF readmission in a U‐shaped pattern and could potentially be used for congestion phenotyping.
Keywords: Acute heart failure, Elderly, Chloraemia, Prognosis, Phenotyping
Introduction
Most patients admitted for acute heart failure (AHF) display symptoms and signs of fluid overload requiring, in most cases, intravenous (IV) diuretic treatment. 1 The depletive treatment and the disease itself, through the activation of the renin–angiotensin system and vasopressin, induce deleterious electrolyte disturbances. 2 , 3 Sodium and potassium disturbances have drawn most of the attention of clinicians. 4 For example, hyponatraemia is reported to be detected in up to 20% of patients with AHF, proving to be correlated with increased in‐hospital and post‐discharge mortality. 5 , 6 Nevertheless, some recent reports have shown an inversely proportional association between chloride and mortality, independently from natraemia. 7 Additionally, chloride has been postulated as a marker of diuretic resistance. 8
Two mechanisms could explain the pathophysiology of hypochloraemia in heart failure (HF) 9 : (i) the renal loss of chloride resulting from intensive diuretic use (chloride depletion) and (ii) the hyper‐activation renin–angiotensin–aldosterone system (RAAS) or vasopressin secretion, both inducing increased water retention (chloride dilution). With this premise, we hypothesize that the prognostic role of hypochloraemia may be largely dependent on the aforementioned mechanisms (diluted or depleted) as it occurs with plasma sodium. 9
To our knowledge, no previous studies have analysed the prognostic impact of chloraemia across the two phenotypes (diluted or depleted) in elderly patients of European ancestry.
The aim of this study was to evaluate the association between chloraemia and the risk of long‐term mortality in an elderly European cohort of patients with AHF. Additionally, we investigated the prognostic value of the two phenotypes of hypochloraemia.
Methods
Study design and patients
This study was based on the UICA Registry, an ongoing single‐centre cohort registry of consecutive patients hospitalized for HF followed at the Heart Failure Unit (Unidad de Insuficiencia CArdiaca, UICA, in Spanish) of the Internal Medicine Department, at Hospital Universitario Ramon y Cajal, Madrid, Spain. The UICA Registry and the informed consent form were approved by the Research Ethics Committee of Hospital Universitario Ramón y Cajal (protocol reference: 052/18). In this study, we retrospectively analysed 429 consecutive patients admitted with a principal diagnosis of AHF, from June 2020 to March 2022.
Diagnosis of acute decompensated HF was defined as the rapid onset of symptoms and signs secondary to abnormal cardiac function and the presence of objective evidence of structural or functional abnormality of the heart at rest (such as cardiomegaly, third heart sound, cardiac murmur, abnormality of the echocardiogram, or raised natriuretic peptides). Information related to demography, medical history, physical examination, 12‐lead electrocardiogram (ECG), standard laboratory tests, echocardiographic data, and pharmacological therapies during hospitalization and discharge was routinely recorded using pre‐established electronic questionnaires. Personalized pharmacological treatments were given following established guidelines operating at the time the patient was recruited in the registry. This study conformed to the principles outlined in the Declaration of Helsinki and was approved by an institutional review committee.
Measurements and definitions of exposures
Serum chloride and sodium were measured within the first 48 h of admission. Hypochloraemia, normochloraemia, and hyperchloraemia were defined as chloride ≤96 mmol, 97–105 and >106 mmol/L, respectively. Hyponatraemia, normonatraemia, and hypernatraemia were defined as sodium values ≤134, 135–145, and >145 mmol/L, respectively, as described in the previous studies. 6 Worsening renal function (WRF) was defined as an increase in plasma creatinine of >0.5 mg/dL from admission in the emergency department to the next measurement. 10
Hypochloraemia phenotypes
Hypochloraemia phenotypes were defined as follows: (i) diluted: estimated plasma volume status (ePVS) >5.5 mL/g and (ii) depleted: ePVS < 5.5 mL/g. The ePVS was calculated through instantaneous estimation formula as described by Duarte (100 − haematocrit/haemoglobin). 11 , 12
Outcomes
The endpoint of interest was time to all‐cause mortality and the composite of death and/or HF readmission. The definition of readmission included any unplanned in‐hospital stay longer than 24 h requiring IV therapy.
Statistical analysis
Continuous variables were expressed as median [inter‐quartile range (IQR)]. Discrete variables were summarized as percentages. Baseline characteristics among hypochloraemia, normochloraemia, and hyperchloraemia were compared by ANOVA, Kruskal–Wallis, or χ 2 tests as appropriate. Survival analysis was performed using the Kaplan–Meier method and the log‐rank test to compare the survival curves between patients according to the chloride categories.
Univariate and multivariable Cox proportional hazard regression models were constructed to study the predictive value of chloride as a continuous and categorized (based on established categories for both endpoints). Estimates of risk were presented as hazard ratios (HRs) with their respective 95% confidence interval (CI).
All covariates shown in Table 1 were evaluated for predictive purposes. A final model was derived by using backward stepwise selection. Well‐established prognostic factors and potential confounders were included in the final model regardless of their P value. During this selection process, the linearity assumption for all continuous variables was simultaneously tested, and the variable transformed, if appropriate, with fractional polynomials. Covariates included in the final multivariate model were sex, age, estimated glomerular filtration rate (eGFR), left ventricular ejection fraction (LVEF), Charlson index, atrial fibrillation (AF), heart rate, furosemide equivalent dose (FED), systolic blood pressure, serum sodium, and haemoglobin. The shape of the curve was determined by modelling serum chloride with a fractional polynomial with 4 degrees of freedom [3 3]. In order to assess the association between serum chloride and cause‐specific mortality (cardiovascular vs. non‐cardiovascular), we fitted Fine–Gray subdistribution hazard models, 13 for both cardiovascular and non‐cardiovascular death, where other causes of death were treated as a competing risk.
Table 1.
Baseline characteristics
| Variables |
Total (n = 429) |
Hypochloraemia Cl < 97 (n = 125) |
Normochloraemia 97 < Cl < 106 (n = 257) |
Hyperchloraemia Cl > 105 (n = 47) |
P |
|---|---|---|---|---|---|
| Age | 85 (78–92) | 84 (76–92) | 85 (79–91) | 86 (78–93) | 0.118 |
| Women, n (%) | 266 (62) | 79 (63.2) | 163 (63.4) | 24 (51.1) | 0.261 |
| Medical history | |||||
| Arterial hypertension, n (%) | 379 (88.8) | 112 (89.6) | 227 (89) | 40 (85.1) | 0.693 |
| Diabetes mellitus, n (%) | 187 (43.6) | 58 (46.4) | 106 (41.2) | 23 (48.9) | 0.467 |
| Previous HF, n (%) | 340 (79.2) | 101 (80.8) | 199 (77.4) | 40 (85.1) | 0.432 |
| Atrial fibrillation, n (%) | 253 (59) | 69 (55.2) | 159 (61.9) | 25 (53.2) | 0.321 |
| HFpEF, n (%) | 343 (80) | 105 (84) | 202 (78.6) | 36 (76.6) | 0.386 |
| Coronary artery disease, n (%) | 130 (30.4) | 41 (33.1) | 79 (30.7) | 10 (21.7) | 0.357 |
| Chronic kidney disease, n (%) | 215 (50.2) | 53 (42.4) | 132 (51.6) | 30 (63.8) | 0.035 |
| COPD, n (%) | 59 (13.8) | 20 (16) | 33 (12.9) | 6 (12.8) | 0.702 |
| Dementia, n (%) | 54 (12.6) | 13 (10.4) | 29 (11.4) | 12 (25.5) | 0.018 |
| Valvular disease, n (%) | 165 (38.4) | 51 (40.8) | 98 (38.1) | 16 (34) | 0.712 |
| Chronic treatment | |||||
| ACE‐I, n (%) | 103 (24.2) | 26 (21) | 68 (26.7) | 9 (19.6) | 0.352 |
| ARB, n (%) | 118 (27.8) | 35 (28.2) | 74 (29.1) | 9 (19.6) | 0.409 |
| Beta‐blockers, n (%) | 229 (53.8) | 64 (51.2) | 141 (55.3) | 24 (52.2) | 0.734 |
| Loop diuretics, n (%) | 330 (77.6) | 97 (77.6) | 192 (75.6) | 41 (89.1) | 0.128 |
| Thiazides, n (%) | 45 (10.5) | 20 (16) | 24 (9.3) | 1 (2.1) | 0.020 |
| Spironolactone, n (%) | 61 (15) | 21 (17.8) | 36 (14.6) | 4 (9.5) | 0.420 |
| Statins, n (%) | 212 (49.8) | 61 (48.8) | 132 (51.8) | 19 (41.3) | 0.412 |
| Vital signs | |||||
| Heart rate (b.p.m.) | 82.5 (62–102) | 82.8 (63–103) | 83.3 (63–103) | 77.5 (58–106) | 0.169 |
| SBP (mmHg) | 136.2 (113–159) | 132 (110–152) | 138.3 (113–163) | 136 (119–153) | 0.052 |
| DBP (mmHg) | 74.8 (58–92) | 74.8 (58–92) | 75.3 (58–92) | 71.9 (59–85) | 0.738 |
| Clinical presentation | |||||
| Pleural effusion, n (%) | 178 (41.5) | 61 (48.8) | 92 (35.8) | 25 (53.2) | 0.012 |
| Oedema, n (%) | 292 (68.1) | 89 (71.2) | 173 (67.3) | 30 (63.8) | 0.601 |
| Echocardiography | |||||
| LVEF (%) | 59.5 (47–71) | 61.6 (49–73) | 58.8 (47–71) | 57.9 (46–70) | 0.087 |
| LAVI (mL/m2) | 49.3 (31–67) | 50.4 (32–68) | 48.7 (29–67) | 49.2 (34–64) | 0.780 |
| Septum (cm) | 1.21 (0.97–1.43) | 1.17 (0.94–1.4) | 1.23 (0.99–1.43) | 1.29 (1.02–1.56) | 0.051 |
| TAPSE (mm) | 2.0 (1.58–2.44) | 1.9 (1.49–2.37) | 2.0 (1.62–2.42) | 2.1 (1.63–2.69) | 0.104 |
| PASP (mmHg) | 49 (35–63) | 47 (33–61) | 48.9 (34–64) | 55 (46–64) | 0.110 |
| E/e′ | 16.3 (9–23) | 14.1 (9–19) | 17.4 (10–24) | 17.8 (11–25) | 0.005 |
| Laboratory | |||||
| ePVS (mL/g) | 5.7 (4.1–7.3) | 5.5 (3.9–7.1) | 5.7 (4.1–7.3) | 6.1 (4.5–7.7) | 0.022 |
| Haemoglobin (g/dL) | 11.6 (9.6–13.6) | 11.9 (9.9–13.9) | 11.6 (9.6–13.6) | 11.0 (9.1–13.9) | 0.035 |
| Haematocrit (%) | 37.1 (30.5–43.5) | 37.5 (30.5–44.5) | 37.1 (30.8–43.4) | 35.3 (29–41.6) | 0.505 |
| Plasma sodium (mmol/L) | 138.8 (133–145) | 135.8 (128–144) | 139.6 (135–143) | 142 (138–146) | 0.0001 |
| Plasma potassium (mmol/L) | 4.1 (3.5–4.7) | 4.0 (3.3–4.7) | 4.1 (3.5–4.7) | 4.2 (3.8–4.6) | 0.085 |
| Urea (mg/dL) | 81.7 (37–125) | 77.5 (29–125) | 79.8 (41–119) | 104.6 (51–157) | 0.003 |
| Creatinine (mg/dL) | 1.45 (0.73–2.17) | 1.30 (0.59–2.01) | 1.46 (0.81–2.11) | 1.76 (0.78–2.74) | 0.001 |
| eGFR (mL/min/1.73 m2) | 43.7 (30.4–62) | 49 (34.3–74.9) | 42.2 (30.7–59) | 37.7 (25–58.6) | 0.004 |
| eGFR < 60 mL/min (%) | 314 (73.2) | 82 (65.6) | 196 (76.3) | 36 (76.6) | 0.079 |
| BNP (pg/mL) | 556 (308–992) | 471 (240–851) | 622 (320–1011) | 654 (413–1549) | 0.0125 |
| CA125 (U/mL) | 58.2 (28–120) | 67.6 (31–125) | 54.7 (27–116) | 62.2 (26–133) | 0.301 |
| Albumin (g/dL) | 3.10 (2.66–3.54) | 3.10 (2.69–3.51) | 3.11 (2.64–3.58) | 3.08 (2.72–3.44) | 0.960 |
| Urine sodium (mEq/L) | 67.5 (33–101) | 59.7 (28–90) | 68.4 (33–101) | 84.4 (47–121) | 0.011 |
| Urine chloride (mEq/L) | 50.1 (17–83) | 39.3 (15–63) | 54.9 (19–91) | 54.7 (20–90) | 0.056 |
| Diuretic treatment at home | |||||
| Total dose furosemide (mg) | 52.3 (7–95) | 54.7 (7–100) | 50.4 (6–94) | 56.6 (12–80) | 0.712 |
| Events | |||||
| Readmissions, n (%) | 196 (45.9) | 53 (42.4) | 120 (47) | 23 (48.9) | 0.654 |
| Mortality, n (%) | 165 (38.5) | 50 (40) | 90 (35) | 25 (53.2) | 0.057 |
| CV mortality, n (%) | 97 (22.6) | 32 (25.6) | 55 (21.4) | 10 (21.3) | 0.637 |
| No CV mortality, n (%) | 68 (15.6) | 18 (14.4) | 35 (13.5) | 15 (31.9) | 0.005 |
ACE‐I, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; BNP, brain natriuretic peptide; CA125, cancer antigen 125; COPD, chronic obstructive pulmonary disease; CV, cardiovascular; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; ePVS, estimated plasma volume status; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; LAVI, left atrial volume index; LVEF, left ventricular ejection fraction; PASP, pulmonary arterial systolic pressure; SBP, systolic blood pressure; TAPSE, tricuspid annular plane systolic excursion.
There were missing data for BNP in 55 patients and CA125 in 82 patients.
Additionally, a Cox proportional hazard regression model was used for estimating the multivariate‐adjusted HR of different phenotypes of hypochloraemic patients (diluted vs. depleted) using the same set of covariates.
Association between the phenotype with worst prognosis and biochemical, pharmacological, and clinical variables was investigated through a multivariable logistic regression model.
A two‐sided P value <0.05 was considered statistically significant for all analyses. All analyses were performed using STATA 15.1.
Results
Baseline characteristics
The mean age was 85 (78–92) years, 266 (62%) patients were women, and 340 (79%) had a prior diagnosis of HF. The most prevalent comorbidities were hypertension (88%), AF (59%), chronic kidney disease (CKD) (50%), and diabetes mellitus (DM) (43%). Median (IQR) values of cancer antigen 125 (CA125) and BNP were 58.2 (28–120.2) U/mL and 556.4 (308–992.5) pg/mL, respectively. The baseline characteristics across chloride categories are presented in Table 1 . Patients in the upper category of chloride displayed a higher prevalence of CKD and dementia. They also showed higher values of urea, creatinine, sodium, and BNP (Table 1 ). Patients with hypochloraemia showed a higher prevalence of thiazide use, pleural effusion, hyponatraemia, lower creatinine, and lower prevalence of prior CKD.
Mortality risk
During a median (p25% to p75%) follow‐up of 238 (118–371) days, 165 (38.46%) patients died. There was an overall trend to higher mortality in patients with hyperchloraemia (Table 1 ). Kaplan–Meier plots showed divergent trajectories among categories, with a higher risk for hyperchloraemia followed by hypochloraemia and a lower risk for normochloraemia (Figure 1 ). Sodium categories were not related to higher mortality (Figure 2 ).
Figure 1.
Kaplan–Meier curves across chloride categories for all‐cause mortality.
Figure 2.
Kaplan–Meier curves across sodium categories for all‐cause mortality.
Under the multivariate setting, chloride remained associated with a higher risk of all‐cause mortality. This relationship showed a U‐shaped pattern (P < 0.0001). A marked increase in risk was observed both with values below the cut‐off point (97 mmol/L) and with values above the normal cut‐off point (105 mmol/L) (Figure 3 ). The shaded area represented the 95% CI and was centred at the median chloride in the sample (100 mmol/L). Thus, any portion of the curve above the y‐scale reference line of 1 is statistically significant.
Figure 3.
Predictive value of chloride for all‐cause mortality after multivariate analysis.
Under the same multivariate scenario, natraemia was not associated with a higher risk of this endpoint (P = 0.582) (Figure 4 ).
Figure 4.
Predictive value of sodium for all‐cause mortality after multivariate analysis.
Composite of mortality and/or heart failure readmission
A total of 215 (50.1%) patients presented the combined endpoint. As for mortality, there was a trend to more rates of the combined endpoint in those with hyperchloraemia category (Table 1 ). Kaplan–Meier plots showed a significant risk of this endpoint across the categories of chloraemia (Figure 5 ).
Figure 5.
Kaplan–Meier curves across chloride categories for composite of mortality and/or heart failure readmission.
After multivariate analysis, chloride remained associated with a higher risk of combined events (Figure 6 ). Similar to all‐cause mortality, this relationship showed a U‐shaped association (P = 0.003).
Figure 6.
Predictive value of chloride for composite of mortality and/or heart failure readmission after multivariate analysis.
Cardiovascular vs. non‐cardiovascular mortality
Adjusting for the same set of covariates and non‐cardiovascular mortality, we found that the excess of risk cardiovascular mortality was seen in those with lower chloride (P = 0.002) (Figure 7 ). Conversely, in the same multivariate setting and adjusting for cardiovascular mortality as a competing event, higher but not lower chloride was associated with a higher risk of non‐cardiovascular mortality (P = 0.039) (Figure 8 ).
Figure 7.
Predictive value of chloride for cardiovascular mortality after multivariate analysis using a competing event methodology.
Figure 8.
Predictive value of chloride for non‐cardiovascular mortality after multivariate analysis using a competing event methodology.
Hypochloraemic phenotypes
Among all 125 hypochloraemic patients, 74 (59%) had an ePVS < 5.5 mL/g and 51 (41%) had an ePVS > 5.5 mL/g. Hypochloraemic patients with haemoconcentration (ePVS < 5.5 mL/g) had an almost two‐fold increased risk of all‐cause mortality compared with normochloraemic patients (HR 1.86, 95% CI: 1.17–2.97, P = 0.008) (Figure 9 and Table 2 ).
Figure 9.
Cox multivariable proportional hazard model curve across hypochloraemic phenotype for all‐cause mortality. ePVS, estimated plasma volume status.
Table 2.
Cox multivariable proportional hazard model of serum chloride categories for the prediction of all‐cause mortality
| All‐cause mortality | ||
|---|---|---|
| HR (95% CI) | P value | |
| Normochloraemia | 1 | |
| Hypochloraemia and ePVS < 5.5 mL/g | 1.86 (1.17–2.97) | 0.008 |
| Hypochloraemia and ePVS > 5.5 mL/g | 0.94 (0.53–1.68) | 0.855 |
| Hyperchloraemia | 1.44 (0.90–2.29) | 0.123 |
CI, confidence interval; ePVS, estimated plasma volume status; HR, hazard ratio.
On the other end, hypochloraemic patients with haemodilution (ePVS > 5.5 mL/g) had no increased risk of all‐cause mortality (HR 0.94, 95% CI: 0.53–1.68, P = 0.855). Hyperchloraemic patients had a non‐significant trend to an increased risk of mortality (HR 1.44, 95% CI: 0.90–2.29, P = 0.123) (Figure 9 ).
We analysed the interaction between the different plasma volume phenotypes with the plasma chloride categories, obtaining a P for interaction = 0.084.
After multivariable logistic regression, the phenotype characterized by hypochloraemia and haemoconcentration was strongly associated with hyponatraemia [odds ratio (OR) 3.2, 95% CI: 1.73–5.94, P = 0.0001], worsening renal function (OR 3.8, 95% CI: 1.22–11.8, P = 0.021), and baseline creatinine <1.1 mg/dL (OR 2.76, 95% CI: 1.41–5.38, P = 0.003) and inversely correlated with previous CKD (OR 0.48, 95% CI: 0.25–0.91, P = 0.026). There was a tendency towards a positive association with previous thiazide treatment and pleural effusion that did not reach statistical significance (OR 1.95, P = 0.100 and OR 1.49, P = 0.151, respectively). No association was found with the previous mean dose of oral furosemide (OR 1.11, P = 0.737).
Discussion
In this study, which included a cohort of very old patients with predominantly HF with preserved ejection fraction (HFpEF), we found that chloraemia, and not natraemia, predicted the long‐term burden of all‐cause mortality and the combined of death/readmission for HF, by a U‐shaped curve. Additionally, the prognostic value of hypochloraemia was heterogeneous across ePVS, where patients with depleted hypochloraemia were associated with a higher risk of death.
Plasma chloride and prognosis
The prognostic value of plasma chloride was described before, but the studies were performed using plasma chloride at discharge 1 , 14 or admission, 7 , 8 in younger populations, 7 , 8 , 14 , 15 the majority with HF with reduced ejection fraction (HFrEF) 8 , 9 or chronic HF. 15 In the current study, we measured plasma chloride during the acute congestive phase (24–48 h from the admission) in a very old European cohort hospitalized in an Internal Medicine Department, with HFpEF mainly (80%).
One of the interesting features of the study is the U‐shaped curve describing the relation between chloride and mortality or a composite of mortality and/or HF readmission. This curve, not described before in other studies, suggests a higher risk in both hypochloraemia and hyperchloraemia. In other studies, an inverted J curve was observed, clearly showing a worse prognosis for hypochloraemia. 7 , 15 , 16 Our study shows an increased risk also for hyperchloraemia, not previously reported. The exact mechanisms by which hyperchloraemia contributes to increased mortality are not fully understood but might reflect patients with worse baseline risk profile. The deleterious prognostic effect could be explained by the relationship with metabolic acidosis, particularly in patients with CKD, 17 a subset of patients highly represented in this study. Additionally, the hyperchloraemic patients in our cohort had a higher prevalence of dementia, which increases the risk of hyperchloraemia, due to poorer hydration and weaker tolerance to high doses of diuretics.
The pathophysiology of chloride in HF is not fully understood. It is known that chloride plays a key role in the alterations of the RAAS, in the maintenance of the acid–base balance, and in the renal transporters on which diuretics act. 14 , 18 There are two main mechanisms that reduce chloride levels in HF: the use of diuretics and the activation of RAAS and/or vasopressin. 9 The hypochloraemia that we observed in our patients can be explained both using diuretics and by the activation of the RAAS and vasopressin. We attribute less importance to other drugs given that only 50% of the patients were taking angiotensin‐converting enzyme inhibitor (ACE‐I)/angiotensin receptor blocker (ARB) and 15% mineralocorticoid receptor antagonist (MRA) and none tolvaptan, unlike other studies where this percentage was higher. 14
Cardiovascular vs. non‐cardiovascular mortality
In our cohort, most patients have HFpEF, and in previous studies, mortality in these patients has been predominantly non‐cardiovascular. 19 After analysis, we have observed differences in the types of death based on chloride levels. Patients with hypochloraemia have a mortality rate of cardiovascular origin. In contrast, those with hyperchloraemia die from non‐cardiovascular causes. This result validates our observations that patients with hyperchloraemia have a high burden of comorbidities, and the development of hyperchloraemia in patients hospitalized for AHF represents a marker of frailty and impaired homeostasis, increasing the risk of mortality due to causes other than HF. In addition, the higher cardiovascular mortality observed in patients with hypochloraemia confirms its importance as a prognostic factor.
The importance of plasma volume status
The second interesting aspect that we highlight is the prognostic utility of hypochloraemic phenotypes. This is the first study that deeply investigates the prognostic value of different phenotypes within the category of hypochloraemic patients, analysing the possible relations with the intravascular congestion state. We discovered that not all patients with hypochloraemia have an increased risk of death when compared with normochloraemic patients.
Only patients with hypochloraemia and haemoconcentration, defined as an ePVS < 5.5 mL/g, have increased mortality. Patients with hypochloraemia and haemodilution have the same mortality as normochloraemic patients.
It is important to note that we are analysing the prognostic value of different phenotypes of hypochloraemia, measured at 24–48 h from the admission. In patients with an ePVS > 5.5 (haemodilution) at 24–48 h from admission, hypochloraemia is probably a marker of residual congestion, which suggests the possibility of improving and increasing the dose of diuretics to reach the complete decongestion. We speculate that hypochloraemia has no prognostic role in this case because other covariates reflecting residual congestion capture the predictive information.
The logistic regression analysis shows that hypochloraemia with haemoconcentration identifies a particular population with good renal function and consequently a good diuretic response in which a probably excessive use of loop diuretics during the first 24–48 h induces a diminished circulating blood volume, electrolytic depletion (hypochloraemia and hyponatraemia), and worsening renal function, increasing the risk of death.
Loop diuretics reduce circulating blood volume and intravascular congestion. Normally, during decongestion, the shift of fluid from extravascular spaces (lung, extremities, and abdomen) refills circulation. When decongestion occurs at an exaggerated velocity, because of an aggressive loop diuretic use, in less severe stages of congestion, the refill rate is overcome by rapid intravascular depletion resulting in diminished peripheral vascularization and possibly worsening renal function. 20 , 21 , 22 , 23
Boorsma et al. 20 have suggested that, once intravascular congestion is treated but extravascular congestion is still present, an aquaretic drug should be added, diminishing the loop diuretic dose. Aquaretic drugs, such as tolvaptan, promote free water excretion inducing an increase in plasma osmolality and favouring the translocation of fluid from extravascular to intravascular space. Sodium–glucose cotransporter‐2 (SGLT2) inhibitors too would be useful for the treatment of this phenotype of patients, characterized by low intravascular congestion and predominance of tissue congestion. 24 This would shift from a strategy aiming at increasing natriuresis to a strategy aiming at fluid redistribution.
We hypothesize that hypochloraemia with haemoconcentration is a marker of intravascular decongestion that identifies patients who could benefit from this shift in treatment strategy in case there is or persists tissue congestion. In these patients, the use of tolvaptan or SGLT2i, and the decrease in loop diuretic dose, would promote translocation of fluid from extravascular to intravascular space, increase circulating blood volume, restore kidney vascularization, and resolve hyponatraemia 25 and hypochloraemia.
A previous study showed an association between hypochloraemia, haemoconcentration, and good renal function but did not analyse different phenotypes. 14 Kataoka recently proposed an interesting theory aiming to address the relation between hypochloraemia and low ePVS: the organism would use an increase in chloride concentration, and consequently plasma tonicity, to preserve effective blood volume, avoiding arterial hypoperfusion. Patients with a low ePVS and hypochloraemia would represent a dysfunctional phenotype of patients unable to implement this mechanism. 26
Clinical implications
The present findings endorse the prognostic role of plasma chloride levels, being ominous in both hypochloraemia and hyperchloraemia predicting a higher risk of long‐term adverse events after admission for acute decompensated HF. Interestingly, the same prognostic role could not be found with plasma sodium. These results alert us to the importance of using plasma chloride values, an unknown or little‐used ion in HF, as a prognostic marker.
In addition, the finding of different phenotypes of hypochloraemia with different clinical consequences advocates for the necessity of using chloride in hospitalized patients as a guide in the optimization of diuretic treatment, with the aim of avoiding excessive diuretic treatments with furosemide/thiazide. This is of especially importance in those patients with a predominantly tissue congestion phenotype, with little intravascular congestion, characterized by hypochloraemia and haemoconcentration, advocating other strategies based on SGLT2i and tolvaptan, with the aim of minimizing adverse events. Undoubtedly, the assessment of the congestion phenotype must be based on a multiparametric approach, with physical examination, anamnesis, ultrasound, and biomarkers, 27 , 28 , 29 with chloride being a new tool, together with haemoconcentration, necessary for this phenotyping.
Limitations
Some important limitations need to be addressed: firstly, the observational design of a single‐centre study in which unmeasured confounders might be playing a crucial role. These findings were found in elderly population with predominantly HFpEF. The extrapolations of current findings to other HF populations require further confirmation. Secondly, the lack of information about the doses of IV furosemide the patients received on admission is an important limitation. Thirdly, the BNP and echocardiographic assessment was unavailable in all the patients. Finally, the cortisol and pH levels were unavailable, both of which can influence chlorine levels and may be confounding factors.
Conclusions
In very old patients hospitalized with AHF, plasma chloride was associated with the risk of death and HF readmission in a U‐shaped pattern. The prognostic implications of hypochloraemia may differ across plasma volume status. Further work should be performed to confirm these findings.
Conflict of interest
None declared.
Funding
This study has no funding.
Llàcer, P. , Croset, F. , Núñez, J. , Campos, J. , Fernández, C. , Fabregate, M. , Del Hoyo, B. , Ruiz, R. , López, G. , Tello, S. , Lorenzo, A. , Fernández, J. M. , and Manzano, L. (2023) Prognostic significance of plasma chloride in elderly patients hospitalized for acute heart failure. ESC Heart Failure, 10: 2637–2647. 10.1002/ehf2.14434.
References
- 1. Gheorghiade M, Pang PS. Acute heart failure syndromes. J Am Coll Cardiol. 2009; 53: 557–573. [DOI] [PubMed] [Google Scholar]
- 2. Rivera FB, Alfonso P, Golbin JM, Lo K, Lerma E, Volgman AS, Kazory A. The role of serum chloride in acute and chronic heart failure: a narrative review. Cardiorenal Med. 2021; 11: 87–98. [DOI] [PubMed] [Google Scholar]
- 3. Verbrugge FH, Steels P, Grieten L, Nijst P, Tang WH, Mullens W. Hyponatremia in acute decompensated heart failure: depletion versus dilution. J Am Coll Cardiol. 2015; 65: 480–492. [DOI] [PubMed] [Google Scholar]
- 4. McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, Burri H, Butler J, Čelutkienė J, Chioncel O, Cleland JGF, Coats AJS, Crespo‐Leiro MG, Farmakis D, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Francesco Piepoli M, Price S, Rosano GMC, Ruschitzka F, Kathrine Skibelund A, ESC Scientific Document Group . 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure: developed by the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure of the European Society of Cardiology (ESC) with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J. 2021; 42: 3599–3726. [DOI] [PubMed] [Google Scholar]
- 5. Gheorghiade M, OPTIMIZE‐HF Investigators and Coordinators . 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: 980–988. [DOI] [PubMed] [Google Scholar]
- 6. Shchekochikhin DY, Schrier RW, Lindenfeld J, Price LL, Jaber BL, Madias NE. Outcome differences in community‐ versus hospital‐acquired hyponatremia in patients with a diagnosis of heart failure. Circ Heart Fail. 2013; 6: 379–386. [DOI] [PubMed] [Google Scholar]
- 7. Grodin JL, Simon J, Hachamovitch R, Wu Y, Jackson G, Halkar M, Starling RC, Testani JM, Tang WHW. Prognostic role of serum chloride levels in acute decompensated heart failure. J Am Coll Cardiol. 2015; 66: 659–666. [DOI] [PubMed] [Google Scholar]
- 8. Ter Maaten JM, Damman K, Hanberg JS, Givertz MM, Metra M, O'Connor CM, Teerlink JR, Ponikowski P, Cotter G, Davison B, Cleland JG. Hypochloremia, diuretic resistance, and outcome in patients with acute heart failure. Circ Heart Fail. 2016; 9: e003109. [DOI] [PubMed] [Google Scholar]
- 9. Zandijk AJL, van Norel MR, Julius FEC, Sepehrvand N, Pannu N, Voors AA, Ezekowitz JA. Chloride in heart failure: the neglected electrolyte. JACC Heart Fail. 2021; 9: 904–915. [DOI] [PubMed] [Google Scholar]
- 10. Mullens W, Damman K, Testani JM, Martens P, Mueller C, Lassus J, Tang WHW, Skouri H, Verbrugge FH, Orso F, Hill L, Ural D, Lainscak M, Rossignol P, Metra M, Mebazaa A, Seferovic P, Ruschitzka F, Coats A. Evaluation of kidney function throughout the heart failure trajectory—a position statement from the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail. 2020; 22: 584–603. [DOI] [PubMed] [Google Scholar]
- 11. Huang CY, Lin TT, Wu YF, Chiang FT, Wu CK. Long‐term prognostic value of estimated plasma volume in heart failure with preserved ejection fraction. Sci Rep. 2019; 9: 14369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Núñez J, de la Espriella R, Rossignol P, Voors AA, Mullens W, Metra M, Chioncel O, Januzzi JL, Mueller C, Richards AM, de Boer RA, Thum T, Arfsten H, González A, Abdelhamid M, Adamopoulos S, Anker SD, Gal TB, Biegus J, Cohen‐Solal A, Böhm M, Emdin M, Jankowska EA, Gustafsson F, Hill L, Jaarsma T, Jhund PS, Lopatin Y, Lund LH, Milicic D, Moura B, Piepoli MF, Ponikowski P, Rakisheva A, Ristic A, Savarese G, Tocchetti CG, van Linthout S, Volterrani M, Seferovic P, Rosano G, Coats AJS, Bayes‐Genis A. Congestion in heart failure: a circulating biomarker‐based perspective. A review from the Biomarkers Working Group of the Heart Failure Association, European Society of Cardiology. Eur J Heart Fail. 2022; 24: 1751–1766. [DOI] [PubMed] [Google Scholar]
- 13. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999; 94: 496509. [Google Scholar]
- 14. Seo M, Watanabe T, Yamada T, Yano M, Hayashi T, Nakagawa A, Nakagawa Y, Tamaki S, Yasumura Y, Sotomi Y, Hikoso S, Nakatani D, Fukunami M, Sakata Y, Osaka CardioVascular Conference (OCVC)‐Heart Failure investigators . Prognostic significance of serum chloride level in heart failure patients with preserved ejection fraction. ESC Heart Fail. 2022; 9: 1444–1453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Grodin JL, Testani JM, Pandey A, Sambandam K, Drazner MH, Fang JC, Tang WHW. Perturbations in serum chloride homeostasis in heart failure with preserved ejection fraction: insights from TOPCAT. Eur J Heart Fail. 2018; 20: 1436–1443. [DOI] [PubMed] [Google Scholar]
- 16. Testani JM, Hanberg JS, Arroyo JP, Brisco MA, Ter Maaten JM, Wilson FP, Bellumkonda L, Jacoby D, Tang WW, Parikh CR. Hypochloraemia is strongly and independently associated with mortality in patients with chronic heart failure. Eur J Heart Fail. 2016; 18: 660–668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wallia R, Greenberg A, Piraino B, Mitro R, Puschett JB. Serum electrolyte patterns in end‐stage renal disease. Am J Kidney Dis. 1986; 8: 98–104. [DOI] [PubMed] [Google Scholar]
- 18. Subramanya AR, Yang CL, McCormick JA, Ellison DH. WNK kinases regulate sodium chloride and potassium transport by the aldosterone‐sensitive distal nephron. Kidney Int. 2006; 70: 630–634. [DOI] [PubMed] [Google Scholar]
- 19. Vaduganathan M, Patel RB, Michel A, Shah SJ, Senni M, Gheorghiade M, Butler J. Mode of death in heart failure with preserved ejection fraction. J Am Coll Cardiol. 2017; 69: 556–569. [DOI] [PubMed] [Google Scholar]
- 20. Boorsma EM, Ter Maaten JM, Damman K, Dinh W, Gustafsson F, Goldsmith S, Burkhoff D, Zannad F, Udelson JE, Voors AA. Congestion in heart failure: a contemporary look at physiology, diagnosis and treatment. Nat Rev Cardiol. 2020; 17: 641–655. [DOI] [PubMed] [Google Scholar]
- 21. Testani JM, Chen J, McCauley BD, Kimmel SE, Shannon RP. Potential effects of aggressive decongestion during the treatment of decompensated heart failure on renal function and survival. Circulation. 2010; 122: 265–272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Boyle A, Sobotka PA. Redefining the therapeutic objective in decompensated heart failure: hemoconcentration as a surrogate for plasma refill rate. J Card Fail. 2006; 12: 247–249. [DOI] [PubMed] [Google Scholar]
- 23. Dupont M, Mullens W, Finucan M, Taylor DO, Starling RC, Tang WH. Determinants of dynamic changes in serum creatinine in acute decompensated heart failure: the importance of blood pressure reduction during treatment. Eur J Heart Fail. 2013; 15: 433–440. [DOI] [PubMed] [Google Scholar]
- 24. Hallow KM, Helmlinger G, Greasley PJ, McMurray JJV, Boulton DW. Why do SGLT2 inhibitors reduce heart failure hospitalization? A differential volume regulation hypothesis. Diabetes Obes Metab. 2018; 20: 479–487. [DOI] [PubMed] [Google Scholar]
- 25. Felker GM, Mentz RJ, Cole RT, Adams KF, Egnaczyk GF, Fiuzat M, Patel CB, Echols M, Khouri MG, Tauras JM, Gupta D, Monds P, Roberts R, O'Connor CM. Efficacy and safety of tolvaptan in patients hospitalized with acute heart failure. J Am Coll Cardiol. 2017; 69: 1399–1406. [DOI] [PubMed] [Google Scholar]
- 26. Kataoka H. Mechanistic insights into chloride‐related heart failure progression according to the plasma volume status. ESC Heart Fail. 2022; 9: 2044–2048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Mullens W, Damman K, Harjola VP, Mebazaa A, Brunner‐La Rocca HP, Martens P, Testani JM, Tang WW, Orso F, Rossignol P, Metra M. The use of diuretics in heart failure with congestion—a position statement from the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail. 2019; 21: 137–155. [DOI] [PubMed] [Google Scholar]
- 28. De la Espriella R, Cobo M, Santas E, Verbrugge FH, Fudim M, Girerd N, Miñana G, Górriz JL, Bayés‐Genís A, Núñez J. Assessment of filling pressures and fluid overload in heart failure: an updated perspective. Rev Esp Cardiol (Engl Ed). 2022; S1885‐5857: 00207–00209. [DOI] [PubMed] [Google Scholar]
- 29. Verma S, McMurray JJV. SGLT2 inhibitors and mechanisms of cardiovascular benefit: a state‐of‐the‐art review. Diabetologia. 2018; 61: 2108–2117. [DOI] [PubMed] [Google Scholar]