Should we resurrect acetazolamide as a diuretic for congestion due to heart failure?

Although there have been enormous advances in disease-modifying therapies for some forms of heart failure (HF) in recent decades,1 treatment of patients admitted to hospital with HF, the majority of whom have worsening peripheral oedema due to systemic venous congestion, has changed little in the last 60 years.2 The mainstay of treatment remains administration of intravenous (IV) furosemide. Other approaches, including inotropic agents, aiming to increase cardiac output, and vasodilator agents, aiming to reduce pre-load and afterload, appear useless or harmful.3,4 Accordingly, the results of the Acetazolamide in Acute Decompensated Heart Failure with Volume Overload (ADVOR) trial are most welcome.5 In the lumen of the proximal convoluted tubule (PCT), free hydrogen (H+) and bicarbonate (HCO3−) are converted to carbon dioxide (CO2) and water (H2O) by membrane-bound carbonic anhydrase; H2O and CO2 then diffuse into the cell where they are converted back into H+ and HCO3− by intracellular carbonic anhydrase (Figure 1A). Intracellular H+ is then exchanged for sodium in the PCT lumen via Na+/H+ co-transporters; HCO3− returns to the circulation via Na+/HCO3− and Cl−/HCO3− anti-porters. Acetazolamide inhibits carbonic anhydrase at both sites, reducing production of H+ and, thus sodium and HCO3− reabsorption, which causes natriuresis and diuresis and prevents or corrects systemic alkalosis.6 Sites of sodium reabsorption along the renal tubule and sites of action of natriuretic and aquaretic agents. Alongside mercurial diuretics, acetazolamide was commonly used to treat congestive HF prior to the advent of more potent loop diuretics in the 1960s but subsequently fell into disuse.7 However, most renal sodium absorption occurs in the PCT, upstream from the site of action of loop diuretics, making PCT sodium absorption an interesting therapeutic target. The ADVOR trial investigated the effects of IV acetazolamide or placebo, in addition to IV furosemide, in patients hospitalized with decompensated chronic HF.6 Of 2915 patients screened, 519 were randomized to either 500 mg of IV acetazolamide or matching placebo for at least 3 days. Furosemide was given as a 120 mg IV bolus on Day 1 and as approximately 60 mg IV boluses twice daily on Days 2 and 3.5 Patients taking sodium-glucose co-transporter 2 inhibitors (SGLT2i) were excluded because these agents also have effects on the PCT and were not an established treatment for HF when the trial was designed. Patients were allocated a congestion severity score ranging from 0 to 10, with up to 4 points for peripheral oedema and a maximum of 3 points each for pleural effusion and ascites. The primary outcome was ‘successful decongestion’ defined as the absence of congestion other than a trace of peripheral oedema (score 0 or 1) after 72 h of treatment without the need for further intensification of diuretic therapy. A key secondary outcome was a composite of HF re-hospitalization or death at 3 months.5 The mean age of trial participants was 78 years, 60% were men, 43% had a left ventricular ejection fraction (LVEF) of ≤40%, mean systolic blood pressure was 127 mmHg, median estimated glomerular filtration rate was 39 mL/min/1.73m2, and median N-terminal pro-B-type natriuretic peptide (NTproBNP) was 6173 ng/L. Most patients had peripheral oedema, which was to the knee or above (≥3 points on the congestion score) in 77%, 15% had obvious pleural effusions and 5% had clinically obvious ascites. The median initial congestion score was four. Most patients (82%) were receiving oral bumetanide (median dose 2 mg/day) rather than furosemide prior to admission. Only 6% were receiving digoxin.5 By Day 2, acetazolamide had exerted a greater diuresis (+0.5 L) and natriuresis (+98 mmol) and prevented the rise in serum bicarbonate observed with furosemide alone. A higher proportion of patients met the primary endpoint compared with placebo after 3 days of treatment (42% vs. 31%; P < 0.001), a difference that persisted until discharge (79% vs. 63%). By Day 3, the mean congestion score fell on placebo from ∼4.4 at baseline to 2.8 compared with ∼4.1 down to 2.1 on acetazolamide, and patients assigned to acetazolamide were administered less IV furosemide, either because they were less congested or because they had a greater decline in blood pressure (up to −10 mmHg) or renal function (up to −4 mL/min/1.73m2) By Day 4, quality-of-life scores improved similarly for placebo and acetazolamide but were not measured thereafter. Median length of stay was 1 day shorter for those assigned to acetazolamide (9 vs. 10 days), but, by 3 months, there was no difference in the composite of HF re-hospitalization or death.5 In a pre-specified sub-group analysis, patients receiving higher doses of oral diuretics prior to admission who, by design, received >120 mg IV furosemide, appeared to have little benefit [risk ratio [RR] for the primary endpoint 1.08 (95% confidence interval (CI): 0.76–1.55)] compared with those receiving ≤120 mg IV furosemide [RR: 1.78 (95% CI: 1.33–2.36)]. Clearly, the ADVOR trial has rejuvenated interest in diuretic strategies for managing congestion. However, few cardiologists have experience of using IV acetazolamide, and the treatment landscape has changed radically with the advent of SGLT2i, which have diuretic properties mediated by effects on the PCT and also reduce morbidity and mortality for patients with chronic HF.8 Whether acetazolamide is effective in addition to SGLT2i is uncertain. Higher doses of loop diuretic or the addition of thiazides can also increase diuresis and may be valid alternatives to acetazolamide. We do not know if a strategy of loop diuretic sparing by sequential nephron blockade is safer or more effective than simply giving larger doses of loop diuretics. Choosing relevant, clinically robust endpoints for trials of decompensated HF is problematic.9 Ultimately, rapid relief of symptoms and being ‘alive and well’ at some future point are likely to be most important for patients (Figure 1B). The primary endpoint of ADVOR was a reduction in signs (rather than symptoms) of congestion within 72 h, which is a reasonable short-term goal but only a stepping stone to addressing real needs.9 The use of a composite score including peripheral oedema, pleural effusion, and ascites makes it difficult to understand what actually improved. Measuring diuresis is notoriously difficult; change in weight should have been recorded, which is generally reliable if done using a standardized protocol. Net H2O loss after 48 h of acetazolamide compared with placebo was only 500 mL (95% CI 200–800 mL).5 This is about half of the difference observed between lower and higher doses of IV furosemide in the DOSE trial. The median total doses of furosemide over 72 h for the high- and low-dose strategies in the DOSE trial were 773 and 358 mg, respectively, compared with ∼360 mg in ADVOR.10 The median length of hospital admission in the acetazolamide arm was 9 days despite nearly half of the patients being decongested by Day 3,5 suggesting that length of stay is often driven by factors other than the severity of congestion such as co-morbid conditions and the need for and availability of carer support at home. Many patients remained congested at discharge and were not receiving renin–angiotensin–aldosterone blockers, although this may reflect weaker evidence of benefit for such agents when LVEF is >40%.5 The benefits of any short-term intervention for a chronic and progressive disease are likely to be limited. Better short-term diuresis could even be detrimental if it leads to underestimation of diuretic requirements at discharge or impedes the use of disease-modifying therapies. Acetazolamide did not reduce re-hospitalization or death,5 but it is unlikely that any treatment for HF administered for only 3 days could have such an effect. There are plausible mechanisms by which acetazolamide could improve HF outcomes. Loop diuretics increase urinary acidity, which may impair their action and that of other diuretics. The subsequent metabolic alkalosis may drive renin angiotensin aldosterone system activation, diuretic resistance, arrhythmias, and impair cell function globally, including myocardial contractility.6 Acetazolamide increases urinary bicarbonate concentration and improves metabolic alkalosis, which should be beneficial, but can cause hypokalaemia and hyponatraemia and may have troublesome (e.g. taste disturbance and paraesthesia) and occasionally serious (Stevens–Johnson syndrome) side-effects in the longer term. In conclusion, short-term administration of IV acetazolamide with IV furosemide in the absence of an SGLT2i improves signs of congestion in the short term. Whether acetazolamide is effective in addition to an SGLT2i needs to be investigated, as does the safety and efficacy of longer-term administration of oral acetazolamide. How the diuretic effect of acetazolamide compares with that of a thiazide diuretic, arginine vasopressin antagonist, digoxin, or, indeed, higher doses of loop diuretic is unknown. The ADVOR trial informs the design and need for further clinical trials but does not provide enough evidence to change current clinical practice. Cleland is supported by a British Heart Foundation through a Centre of Research Excellence award (grant number RE/18/6/34217). Biography: Joe Cuthbert is a Senior Lecturer at the Hull York Medical School (HYMS) and an honorary consultant cardiologist at Hull University Teaching Hospital NHS Trust. His research work focusses on the use of diuretics in patients with heart failure and has published widely in the field. He is Chief Investigator for the ADA-HF trial of acetazolamide in patients admitted to hospital with HF recruiting in Hull. He is involved in organising and delivering cardiovascular aspects of the undergraduate curriculum at HYMS and is the lead author of the heart failure online learning modules for the British Journal of Cardiology. Biography: John Cleland is Professor of Cardiology at the British Heart Foundation Centre of Research Excellence. School of Cardiovascular and Metabolic Health, University of Glasgow, United Kingdom. Previous appointments include Professor of Cardiology at the University of Hull and at Imperial College London. He helped found the European Society of Cardiology's (ESC) Task Force on Heart Failure and the British Society of Heart Failure. He founded the European Journal of Heart Failure and wrote the first ESC guideline (on the diagnosis of heart failure) and has contributed to many others. Special interests are diagnostic criteria for and epidemiology of heart failure (a cardio-renal syndrome) and randomised trials for preventing and treating heart failure. He is a Clarivate highly-cited researcher. He has authored >1,000 peer-reviewed papers and reviews. https://orcid.org/0000-0002-1471-7016