Serum Chloride and Heart Failure

Abstract

Despite significant advances in management, heart failure continues to impose a significant epidemiologic burden with high prevalence and mortality rates. For decades, sodium has been the serum electrolyte most commonly associated with outcomes; however, challenging the conventional paradigm of sodium’s influence, recent studies have identified a more prominent role in serum chloride in the pathophysiology of heart failure. More specifically, hypochloremia is associated with neurohumoral activation, diuretic resistance, and a worse prognosis in patients with heart failure. This review examines basic science, translational research, and clinical studies to better characterize the role of chloride in patients with heart failure and additionally discusses potential new therapies targeting chloride homeostasis that may impact the future of heart failure care.

Introduction

Medical history is rife with examples of novel theories that were rejected in favor of more conventional ideas. These include the laws of inheritance established by Gregor Mendel, the germ theory of disease proposed by Louis Pastuer, and perhaps none more famous than the antiseptic techniques pioneered by Ignaz Semmelweis, who was ultimately committed to an asylum in 1865. The role of chloride in heart failure has been met with similar skepticism in favor of the sodium theory. The prognostic importance of hyponatremia in patients with heart failure has been propagated for years, rendering chloride an afterthought and delegating the most abundant and vital extracellular anion to be routinely ignored by clinicians. The role of chloride has been minimized to passive maintenance of electroneutrality, with serum chloride levels generally only reported on hospital rounds for the purposes of calculating an anion gap. More recent research has garnered support for the “chloride theory” of heart failure, given the important role chloride plays in heart failure prognosis, neurohumoral activation, and diuretic resistance, concepts rooted in kidney physiology that are reviewed here.

Background

Chloride is the most abundant extracellular anion in the body and plays a critical role in the maintenance of acid-base balance through its inverse relationship with bicarbonate and in osmotic pressure homeostasis as the anionic counterpart to sodium. The kidney, along with the gastrointestinal tract, is the primary mediator of chloride homeostasis. As with most electrolyte handling, the proximal tubule is the primary site of chloride reabsorption via both paracellular and transcellular mechanisms and is responsible for the reabsorption of approximately 60% of filtered chloride. Additional sites of chloride reabsorption include the sodium/potassium/chloride cotransporter 2 (NKCC2) in the thick ascending limb of the loop of Henle, where around 15%-25% is reabsorbed, and the sodium/chloride cotransporter (NCC) in the distal tubule, where approximately 5% of chloride is reabsorbed.¹ The collecting duct plays a vital role in determining the final urine chloride concentration, most importantly by reabsorption via intercalated cells, primarily through the parallel actions of pendrin, a sodium independent Cl^-/HCO₃^- exchanger that has been proposed as a novel target for diuretic therapy,² and the sodium-dependent chloride-bicarbonate exchanger.³ Kidney handling of chloride throughout the nephron is depicted in Fig 1; however, further description of tubular anatomy and function is beyond the scope of this review and is discussed in detail elsewhere.⁴ Changes in chloride homeostasis have been linked to reduced glomerular filtration rate, which portends a worse prognosis in patients with heart failure.⁵ Hyperchloremia, which is distinct from hyperchloremic metabolic acidosis, where an increase in chloride is accompanied by a reciprocal decline in serum bicarbonate, can result from the loss of hypotonic fluid or infusion of chloride-rich solutions. This has been associated with an increased risk of mortality and acute kidney injury in critically ill patients, secondary to renal vasoconstriction and mediated by tubuloglomerular feedback.⁶ An increase in serum chloride may additionally be toxic to cells by increasing metabolic demands.⁷ For example, in rats, hyperchloremia has been shown to cause tubuloglomerular feedback-mediated reductions in the glomerular filtration rate that were independent of extracellular volume.⁸

Figure 1.

Kidney chloride handling.

Approximately 60% of filtered chloride is reabsorbed in the proximal tubule, primarily paracellularly though transcellular mechanisms exist as well (1). Transcellular transport occurs via chloride anion exchangers coupled with basolateral chloride channels. Acetazolamide inhibits proximal tubular sodium reabsorption and increases serum chloride levels indirectly via inhibition of carbonic anhydrase IV (not depicted). Filtered chloride (15%-25%) is reabsorbed in the thick ascending limb of the loop of Henle (2), primarily via the sodium/potassium/chloride cotransporter 2 (NKCC2), which is inhibited by loop diuretics. Approximately 5% of filtered chloride is reabsorbed in the distal tubule, primarily via a sodium/chloride cotransporter (NCC), which is the primary target of thiazide diuretics. WNK kinases play a vital role as chloride sensors, exerting control over both the NKCC2 and NCC channels. The collecting duct is the final step for regulation of urinary chloride concentration (4). In principle, cells’ electrogenic sodium reabsorption via eNaC drives chloride reabsorption, primarily via a paracellular route by generating a negative lumen potential. Transcellular chloride reabsorption occurs by coupling of pendrin with NDCBE in β-intercalated cells. Cl, chloride; eNaC, epithelial sodium channels; K, potassium; MRA, mineralocorticoid antagonist; Na, sodium; NDCBE, sodium-dependent chloride-bicarbonate exchanger; ROMK, renal outer medullary potassium channel; WNK, with-no-lysine protein kinase.

Specific to patients with heart failure, a post hoc analysis of the ROSE⁹ study found that changes in serum chloride levels between baseline and 72 hours demonstrated a strong, inverse correlation, with changes in cystatin C. The remainder of this review will focus primarily on the impact of hypochloremia in patients with heart failure.

Serum Chloride and Heart Failure Prognosis

Hypochloremia is frequently encountered in patients with heart failure, occurring with 2 distinct biochemical profiles: water excess and chloride depletion. The first is dilution (or water excess), which occurs secondary to nonosmotic release of arginine vasopressin. A murine model of heart failure demonstrated upregulated aquaporin-2 channels in the cortical collecting duct in the setting of decreased circulating volume and stimulation of carotid baroreceptors, which resulted in increased free water reabsorption.¹⁰ This is supported by studies in people with heart failure that have demonstrated increased arginine vasopressin levels in this setting.11, 12, 13 However, hypochloremia only partially overlaps with hyponatremia; therefore, the second etiology of hypochloremia is chloride depletion. This is commonly associated with the use of loop diuretics, which are ubiquitous in the treatment of heart failure and result in a disproportionately greater relative wasting of chloride as compared to sodium. Different loop diuretics may have differing effects. For example, bumetanide has been associated with more potent chlorouresis despite equipotent effects on sodium and potassium excretion as compared to furosemide.¹⁴ Figure 2 proposes possible mechanisms by which heart failure and associated heart failure treatment lead to hypochloremia.

Figure 2.

Relationship between heart failure and hypochloremia. Diagram proposing possible mechanisms by which heart failure and associated heart failure treatment leads to hypochloremia. The right side of the diagram depicts theoretical pathways through which hypochloremia leads to worse prognosis. The bottom half of the figure shows possible therapeutic interventions to raise serum chloride levels. AQ2, aquaporin 2; Cl, chloride; CLCNKA, chloride voltage-gated channel Ka; NCC, sodium/chloride cotransporter; RAAS, renin-angiotensin-aldosterone system; WNK1, with-no-lysine protein kinase 1.

Hypochloremia has also been associated with an increased risk of cardiovascular mortality in the general population, independent of well-established cardiovascular risk factors and levels of other electrolytes, including serum sodium.¹⁵ Similarly, recent studies have identified serum chloride levels as a stronger predictor of outcomes than serum sodium levels in chronic heart failure. A post hoc analysis of the BEST¹⁶ trial demonstrated an association of both hypochloremia and hyponatremia with baseline, 3-month, and 12-month mortality; however, in the multivariable model, only hypochloremia was associated with an increased risk of mortality, independent of serum sodium levels. Grodin et al¹⁷ described a 29% increased risk in 5-year mortality for each standard deviation decrease in serum chloride (∼4.1 mEq/L) on admission in patients with stable heart failure undergoing elective diagnostic coronary angiography, even after adjustment for serum sodium, medications, cardiorenal biomarkers, and functional status. Similar findings, associating low serum chloride levels with increased risk of heart failure hospitalization, cardiovascular death, and all-cause mortality, were seen in an analysis of the TOPCAT¹⁸ trial, which studied patients with heart failure with preserved ejection fraction. Additionally, hypochloremia was predictive of decreased 1-year survival in patients undergoing left ventricular assist device placement.¹⁹ Conversely, in the VICTORIA trial, there was a 21% decrease in the risk of heart failure hospitalization and cardiovascular death for each 5 mmol/L increase in serum chloride level.²⁰

Hypochloremia has also been implicated in poor outcomes in patients with acute decompensated heart failure. An assessment of over 1,300 patients admitted with heart failure, stratified by admission serum chloride and sodium levels, showed an association between hypochloremia and decreased survival over a 3-year follow-up, whereas hyponatremia had no impact on survival if chloride levels were normal.²¹ In a post hoc analysis of the ROSE-AHF study, lower baseline chloride levels were associated with worse diuretic efficiency, 60-day death, and rehospitalization as well as 180-day mortality, although the effect on mortality was attenuated after adjustment for prehospital loop diuretic dose. Interestingly, while serum chloride levels declined during hospitalization in the context of decongestive therapy, acute changes were not associated with poor outcomes in either unadjusted or adjusted analysis.⁹

The role of acute changes in chloride was similarly studied by Ter Maaten et al,²² who also showed no correlation with survival if hypochloremia resolved by day 14, whereas persistent or incident hypochloremia was independently associated with worse survival. This suggests that abnormalities in chronic chloride homeostasis are of far greater importance than acute perturbations. Additionally, the overall prognosis has been shown to be similar whether hypochloremia is secondary to dilution or chloride depletion.²³ Interestingly, recent genetic investigations into determinants of heart failure risk have pointed to a loss-of-function variant in the CLCNKA gene, associated with near 50% loss of function of a renal chloride channel that affects cardiorenal interactions through salt sensitivity, increasing heart failure risk.²⁴ These key studies are summarized in Table 1. Whether or not modulating serum chloride levels, independent of serum sodium, impacts clinical outcomes is not yet certain and is discussed in more detail below.

Table 1.

Key Findings From Studies Assessing Association Between Hypochloremia and Heart Failure Mortality

Study	Design	Patients (n)	Population	Findings
Grodin et al17	Prospective cohort	1,673	Chronic stable HF undergoing elective coronary angiography	Each standard deviation decrease in serum chloride (4.1 mEq/L) independently associated with a 29% increased risk of 5-y mortality (HR, 1.29; 95% CI, 1.12-1.4; P < 0.001) after multivariable adjustment
Testani et al16	Analysis of BEST	2,699	Chronic HF	Hypochloremia associated with increased risk of mortality (HR, 1.3; 95% CI, 1.18-1.42; P < 0.001) per standard deviation decrease, independent of serum sodium, which had no impact on risk of mortality when controlling for serum chloride
Grodin et al18	Analysis of TOPCAT	942	Chronic HFpEF	Each 4 mmol/L decrease in serum chloride level associated with increased risk of all cause and cardiovascular mortality (HR, 1.29; 95% CI, 1.02-1.62; P = 0.04) and (HR, 1.51; 95% CI, 1.11-2.06; P < 0.008), respectively
Cuthbert et al23	Registry cohort	4,705	Chronic HF	Decrease in serum chloride levels independently associated with increased mortality (HR, 1.04; 95% CI, 1.02-1.06; P < 0.001) and mortality or HF hospitalization (HR, 1.03; 95% CI, 1.02-1.05; P < 0.001) irrespective of dilution or depletion
Ter Maaten et al22	PROTECT RCT	1,960	Decompensated HFrEF	New or persistent hypochloremia at day 14 was associated with an increased risk of 180-d mortality (HR, 3.11; 95% CI, 2.17-4.46; P < 0.001); however, resolution of baseline hypochloremia at day 14 was not associated with mortality (P = 0.55)
Grodin et al18	Population based cohort	Main cohort 1,318 Validation cohort 876	Decompensated heart failure	Admission hypochloremia associated with increased risk of mortality. Each unit increase in chloride associated with 6% decreased mortality risk (HR, 0.94; 95% CI, 0.92-0.95; P < 0.001)
Grodin et al9	Analysis of ROSE	358	Decompensated heart failure	Admission hypochloremia increased risk of 60-d mortality, 60-d mortality, and hospitalization and 180 mortality by 14%, 10%, and 9%, respectively, which was mitigated after adjustment for prehospital loop diuretic dose. Change in serum chloride between baseline and 72 h inversely associated with cystatin C.
Mentz et al20	Analysis of VICTORIA	2,524	Chronic heart failure	Each 5 mmol/L increase in serum chloride reduced the risk of heart failure hospitalization or cardiovascular death by 21% (HR, 0.79; 95% CI, 0.74-0.85; P < 0.001)
Trovato et al19	Retrospective cohort	289	Patients who received a HeartMate II or HeartWare LVAD	The presence of hypochloremia at the time of LVAD placement was associated with decreased 1-y survival compared to patients with normal serum chloride P = 0.05

Abbreviations: CI, confidence interval; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HR, hazard ratio; LVAD, left ventricular assist device.

Chloride and Diuretic Resistance

Decongestion, or lack thereof, is an important predictor of patient outcomes,²⁵ even at the cost of worsening kidney function,26, 27, 28 rendering diuretic therapy a cornerstone in the management of patients with decompensated heart failure. Diuretic resistance, which lacks a uniform definition, is often described as a failure to decongest despite adequate and escalating doses of diuretics and is common among patients with heart failure.²⁹ While multiple mechanisms account for the development of diuretic resistance, distal sodium reabsorption through activation of thiazide-sensitive NCCs and intercalated cells via renal adaptation appears to be the primary driver in the majority of cases.³⁰ Supporting this, rats exposed to furosemide demonstrated significant hypertrophy of the distal tubule, resulting in a 3-fold increase in sodium reabsorption.³¹ Hypochloremia is associated with decreased diuretic efficiency, as defined by millimoles (mmol) of sodium excreted per doubling of loop diuretic dose, independent of serum sodium levels,³² a metric that is associated with poor long-term outcomes.³³ A post hoc analysis of the ROSE study demonstrated an association between lower admission serum chloride levels and decreased responsiveness to loop diuretics as well as a strong correlation with worse diuretic efficiency.⁹

Diuretic resistance may be explained by the interaction of chloride with a family of serine/threonine protein kinases with-no-lysine (WNK) that regulate both NCC and NKCC2. WNK4 exerts basal suppression of NCC channels, and this inhibition is removed by WNK1. Chloride-insensitive WNK4 mice developed hypertension and increased NCC activity.34, 35, 36, 37 Mice with missense mutations in WNK4 show increased NCC expression and distal tubule hyperplasia and nephron remodeling that is similar to adaptations after exposure to loop diuretics. Crystalline studies of the inactive form of WNK1 show that chloride stabilizes the inactive conformation by binding directly to the catalytic site. The absence of chloride results in autophosphorylation and activation of WNK1. An additional study demonstrated that NKCC2 expression is activated by chloride depletion in rats.³⁸ If these concepts are extrapolated to humans, the effect of WNK kinases on both NCC and NKCC2 channels, which are modulated by chloride, may provide a mechanism for the observed association between hypochloremia and diuretic resistance. In a small study, hypochloremic participants given loop diuretics showed decreased diuretic efficiency, defined as urinary sodium excretion per doubling of loop diuretic dose, compared with normochloremic participants, despite no difference in the quantity of diuretic excreted or difference in the fractional excretion of lithium, the latter being the gold standard for the estimation of proximal tubule sodium reabsorption,³⁹ thereby localizing diuretic resistance to the distal tubule.³²

An additional factor contributing to diuretic resistance is metabolic alkalosis, which is the most common acid-base abnormality in patients with heart failure. Metabolic alkalosis among patients admitted with acute decompensated heart failure is associated with more severe heart failure.⁴⁰ Moreover, metabolic alkalosis is associated with a 20% reduction in the natriuretic response to bumetanide,⁴¹ impairing decongestion efforts. Previously termed ‘contraction alkalosis,’ this terminology shifted to a ‘chloride depletion alkalosis’ after studies demonstrated that administration of chloride-containing solutions without correcting volume depletion corrected alkalosis.⁴² The most important regulator of correction is pendrin, which is a luminal Cl^-/HCO₃^- exchanger in the collecting duct. As such, it is possible that chloride depletion may result in insufficient substrate in the tubular lumen to exchange with bicarbonate, thus maintaining the alkalosis and decreasing diuretic efficacy.

Chloride and Neurohumoral Activation

A central tenet in the pathophysiology of heart failure is the impairment in ejection of blood or ventricular filling (as in the case of heart failure with preserved ejection fraction), resulting in decreased tissue perfusion. The upregulation of the renin-angiotensin-aldosterone system, in addition to the sympathetic nervous system, is a principal feature of the compensatory neurohumoral response. Although initially beneficial to restore tissue perfusion by increasing cardiac contractility, sodium and water retention, and vascular resistance, this becomes maladaptive over time resulting in deleterious effects, such as venous congestion and pathologic myocardial remodeling.⁴³ The crucial ability of various agents, such as angiotensin-converting enzyme inhibitors, β-blockers, and mineralocorticoid antagonists as part of guideline-directed medical therapy to inhibit distinct aspects of the renin-angiotensin-aldosterone system are central to their ability to slow heart failure progression and improve survival.⁴⁴

Chloride, not sodium, is the principal regulator of renin release from the juxtaglomerular apparatus via NKCC2 channels on the apical membrane of macula densa cells. Hypochloremia results in decreased chloride delivery to the macula densa and hence an increase in renin release. In rats fed low-sodium diets, the infusion of chloride salts, or without sodium, but not sodium salts in the absence of chloride, suppressed renin levels. Conversely, selective chloride depletion stimulates renin release, establishing the macula densa as a chloride sensor.^45,46 In patients with heart failure treated with loop diuretics, higher renin levels are found in those with hypochloremia as compared with those without hypochloremia, even after adjustment for sodium levels.³² Changes in serum chloride have been proposed as the principal regulator of changes in plasma volume, the renin-angiotensin-aldosterone system, and antidiuretic hormone systems in patients with heart failure under the “chloride theory” hypothesis.⁴⁷

Implications for Heart Failure Management

While the inverse relationship between serum chloride and mortality, independent of the effects of sodium, demonstrates its role as a prognostic marker, its status as a therapeutic target is not clear. Strategies targeting hyponatremia have failed to demonstrate improved outcomes.^48,49 Nevertheless, given the broad biological role chloride plays, including critical aspects relevant to heart failure management, such as neurohumoral activation and diuretic resistance, it is possible that strategies to maintain chloride homeostasis result in more favorable outcomes in patients with heart failure.

The use of hypertonic saline as an adjunctive therapy for diuretic resistance in acute decompensated heart failure has garnered significant interest. Although the mechanism remains unknown, various studies have shown promising results.50, 51, 52 The implication that the benefit of this paradoxical therapy is related to the infusion of sodium appears invalid, as the original Sicilian studies,^50,51 which measured 24-h urine excretion of sodium, revealed net-positive sodium balance post hypertonic saline administration. Instead, the benefit may be attributable to chloride repletion, although this is purely speculative and warrants further investigation. Similarly, sodium-restricted diets have failed to consistently demonstrate beneficial outcomes, resulting in level C evidence recommendations in those with heart failure.^53,54 In fact, several studies have identified worse outcomes when sodium intake is restricted.55, 56, 57 It is possible we need to shift the focus from sodium restriction to chloride depletion because sodium chloride provides greater than 90% of dietary sodium,⁵⁸ although the impact of dietary intake on chloride homeostasis is not clear.

The administration of lysine chloride has the advantage of administering chloride without acutely altering the sodium balance. The use of lysine chloride was initially trialed more than 60 years ago in patients with refractory congestion, resulting in enhanced natriuresis and weight loss.⁵⁹ In a small pilot study, lysine chloride administration to 10 patients increased renin levels, contrary to the initial hypothesis; however, other parameters of decongestion were improved, such as enhanced weight loss and markers of hemoconcentration.³² An ongoing clinical trial assessing the efficacy of lysine chloride administration in outpatients with stable heart failure with reduced ejection fraction may provide additional insights into the possible therapeutic potential of sodium-free chloride supplementation (ClinicalTrials.gov identifier NCT03440970).

Acetazolamide is a chloride-retaining diuretic, which may provide an attractive option to reduce the use of chloride-depleting diuretics such as loop and thiazide diuretics. Kataoka et al⁶⁰ utilized acetazolamide as add-on decongestion therapy for acute heart failure and demonstrated improvement in serum chloride concentrations within 10 days of administration, an effect that was sustained at 60 days. Although the mechanism for this is not known, one possibility may involve the inhibition of pendrin,⁶¹ resulting in decreased chloride excretion. A small prospective study of acetazolamide as an adjunctive therapy to loop diuretics in the treatment arm demonstrated similar natriuresis and decongestion metrics as compared to high-dose loop diuretics, indicating improved diuretic efficacy.⁶² Recently, the ADVOR⁶³ trial demonstrated increased natriuresis and improved clinical decongestion at 72 hours among patients with acute decompensated heart failure when acetazolamide was added to loop diuretic therapy as compared to placebo. Interestingly, the benefit of acetazolamide in this trial was shown to extend beyond the 72-hour decongestive period, which may be secondary to a beneficial impact on neurohumoral activation via chloride retention, although this is simply hypothesis generating.

Conclusion

An emerging body of literature has highlighted the importance of chloride as a prognostic marker in patients with heart failure. The role of hypochloremia with regard to neurohumoral activation and diuretic resistance is supported by plausible physiology and small investigational trials; however, the role of chloride as a therapeutic target remains unclear. Relatively novel interventions such as hypertonic saline and adjunctive decongestive therapy with acetazolamide have demonstrated successful outcomes and impact chloride balance; however, chloride levels were not a specified end point. The time is ripe for prospective studies further elucidating the impact of chloride modulation and clinical outcomes in patients with heart failure.