JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 81, NO.

20, 2023

ª 2023 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

PUBLISHED BY ELSEVIER

Natriuretic Response to Acetazolamide

in Patients With Acute Heart Failure
and Volume Overload
Frederik H. Verbrugge, MD, PHD, MS,a,b Pieter Martens, MD, PHD,c,d Jeroen Dauw, MD,c,e Petra Nijst, MD, PHD,c
Evelyne Meekers, MD,c,e Silvio Nunes Augusto, JR, MSC,d Jozine M. ter Maaten, MD, PHD,f Kevin Damman, MD, PHD,f
Gerasimos Filippatos, MD, PHD,g Johan Lassus, MD, PHD,h Alexandre Mebazaa, MD, PHD,i
Frank Ruschitzka, MD, PHD,j Matthias Dupont, MD,c Wilfried Mullens, MD, PHDc,e

ABSTRACT

BACKGROUND Acetazolamide facilitates decongestion in acute decompensated heart failure (ADHF).

OBJECTIVES This study sought to investigate the effect of acetazolamide on natriuresis in ADHF and its relationship
with outcomes.

METHODS Patients from the ADVOR (Acetazolamide in Decompensated Heart Failure with Volume Overload) trial with
complete data on urine output and urine sodium concentration (UNa) were analyzed. Predictors of natriuresis and its
relationship with the main trial endpoints were evaluated.

RESULTS This analysis included 462 of 519 patients (89%) from the ADVOR trial. During 2 days after randomization,
UNa was 92
25 mmol/L on average, and total natriuresis was 425
234 mmol. Allocation to acetazolamide strongly
and independently predicted natriuresis with a 16 mmol/L (19%) increase in UNa and 115 mmol (32%) greater total
natriuresis. Higher systolic blood pressure, better renal function, higher serum sodium levels, and male sex also
independently predicted both a higher UNa and greater total natriuresis. A stronger natriuretic response was associated
with faster and more complete relief of signs of volume overload, and this effect was already significant on the first
morning of assessment (P ¼ 0.022). A significant interaction was observed between the effect of allocation to
acetazolamide and UNa on decongestion (P ¼ 0.007). Stronger natriuresis with better decongestion translated into a
shorter hospital stay (P < 0.001). After multivariable adjustments, every 10 mmol/L UNa increase was independently
associated with a lower risk of all-cause death or heart failure readmission (HR: 0.92; 95% CI: 0.85-0.99).

CONCLUSIONS Increased natriuresis is strongly related to successful decongestion with acetazolamide in ADHF. UNa
may be an attractive measure of effective decongestion for future trials. (Acetazolamide in Decompensated Heart Failure
with Volume Overload [ADVOR]; NCT03505788) (J Am Coll Cardiol 2023;81:2013–2024) © 2023 by the American
College of Cardiology Foundation.

From the aCentre for Cardiovascular Diseases, University Hospital Brussels, Jette, Belgium; bFaculty of Medicine and Pharmacy,
c
Vrije Universiteit Brussel, Brussels, Belgium; Department of Cardiology, Ziekenhuis Oost-Limburg, Genk, Belgium;
d
Cardiovascular and Metabolic Sciences, Lerner Research Institute, Cleveland Clinic, Cleveland, Ohio, USA; eBiomedical Research
Listen to this manuscript’s Institute, Hasselt University, Diepenbeek, Belgium; fDepartment of Cardiology, University of Groningen, University Medical
audio summary by Center Groningen, Groningen, the Netherlands; gNational and Kapodistrian University of Athens, Athens University Hospital
Editor-in-Chief Attikon, Chaidari, Greece; hHeart and Lung Center, Helsinki University Central Hospital, Helsinki, Finland; iUniversité Paris Cité,
Dr Valentin Fuster on National Institute of Health and Medical Research MASCOT (Cardiovascular Markers in Situation of Stress), APHP (Public Assis-
www.jacc.org/journal/jacc. tance Hospital of Paris), Paris, France; and the jDepartment of Cardiology, University Heart Center, University Hospital Zurich and
University of Zurich, Zurich, Switzerland.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’
institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information,
visit the Author Center.

Manuscript received December 28, 2022; revised manuscript received March 3, 2023, accepted March 7, 2023.

ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2023.03.400

2014 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

I
ABBREVIATIONS ntravenous loop diuretic agents are rec- Products. In brief, ADVOR was a Belgian academic,
AND ACRONYMS ommended for all patients with heart multicenter, randomized, parallel-arm, double-blind,
failure (HF) and signs or symptoms of placebo-controlled trial that tested intravenous
ADHF = acute decompensated
heart failure
fluid overload, which is a Class I indication acetazolamide (500 mg once daily) on top of stan-
according to both the joint American Heart dardized loop diuretic agents (intravenous bolus of 2
FENa = fractional sodium
excretion Association/American College of Cardiology/ times the oral maintenance dose) in patients with
HF = heart failure Heart Failure Society of America guidelines ADHF and clinical signs of volume overload. On
NT-proBNP = N-terminal pro–
and the European Society of Cardiology randomization, all patients were receiving oral
B-type natriuretic peptide guidelines.1,2 Until recently, the DOSE-AHF maintenance loop diuretic agents at a dose $40 mg
UNa = urine sodium (Diuretic Optimization Strategies Evaluation furosemide equivalents for $1 month. Use of sodium-
concentration in Acute Heart Failure) was one of the few glucose cotransporter-2 inhibitors was not permitted.
large (n ¼ 308) randomized clinical trials with loop Allocation to acetazolamide was associated with a
diuretic agents in acute decompensated HF (ADHF).3 46.0% relative and 11.7% absolute increased chance
DOSE-AHF showed no significant difference in symp- of meeting the primary endpoint, which was suc-
tom relief with continuous vs bolus administration or cessful decongestion or the absence of any residual
high-dose vs low-dose administration of loop diuretic sign of volume overload other than trace edema no
agents. With high-dose loop diuretic agents (2.5 times later than the third morning after randomization,
the oral maintenance dose), still only 18% of patients including a cumulative urine output $3.5 L on the
were free of signs of fluid overload within 3 days. second morning. Patients treated with acetazolamide
had a higher urine output and natriuresis, showed
SEE PAGE 2025
faster relief of fluid overload signs, and were dis-
Given that congestion remains the main reason for charged on average 1 day earlier, significantly more
(re)hospitalization in patients with HF, more effective often without any residual sign of fluid overload.
decongestive strategies are needed, especially Occurrence of death or HF readmissions did not differ
because residual congestion has been repeatedly and significantly between both treatment arms over the
consistently linked to poor outcomes.4,5 The ADVOR 3-month follow-up period. The current analysis de-
(Acetazolamide in Decompensated Heart Failure with scribes the natriuretic response to acetazolamide,
Volume Overload; NCT03505788) trial chose suc- with a focus on its predictors and association with the
cessful decongestion, defined as the absence of any main trial endpoints.
clinical sign of fluid overload (ie, edema, pleural
STUDY GROUP. According to the ADVOR protocol, 2
effusion, or ascites) within 72 hours, as its primary
consecutive urine collections were performed after
endpoint. 6 In the ADVOR trial, allocation to intrave-
randomization: from first administration of the study
nous acetazolamide (500 mg once daily) compared
intervention to the moment of clinical assessment on
with placebo, on top of standardized high-dose
the next morning (ie, urine collection 1); and further
intravenous loop diuretic agents (2 times the oral
to the next clinical assessment on the second morning
maintenance dose), increased the chance of success-
after randomization (ie, urine collection 2). 6 Patients
ful decongestion relatively with 46% and absolutely
with results available for both urine volume and urine
with 11.7% (corresponding to a number needed to
sodium concentration (UNa) on both occasions were
treat of 8.5).7 Acetazolamide use was associated with
included in the current analysis. To avoid the inclu-
increased urine output and natriuresis, which may
sion of urine samples with unreliable UNa, those with
have mediated this result.
a urine output <500 mL over 24 hours were excluded.
The current analysis explores in detail the impact
of acetazolamide on different metrics of natriuresis ASSESSMENT OF NATRIURESIS. Three metrics of

and its associated clinical variables and effect on natriuresis were calculated for every timed urine
decongestion and clinical outcomes in ADHF. collection: 1) UNa (mmol/L); 2) fractional sodium
excretion (FENa) (%), calculated as the ratio of UNa
METHODS multiplied by serum creatinine over the product of
the urine creatinine and serum sodium concentra-
STUDY DESIGN. This is a prespecified analysis from tion; and 3) total sodium excretion (mmol), calculated
the ADVOR trial. Detailed methods and results of this as the product of UNa and urine output. On the basis
trial have been published previously. 6,7 The ADVOR of the volume and UNa of both urine samples, a
trial protocol was approved by the Institutional Re- weighted overall UNa was calculated for every sub-
view Board at every participating center and by the ject, whereby urine collections 1 and 2 were consid-
Belgian Federal Agency for Medicines and Health ered together as if they were 1 collection. On the basis
JACC VOL. 81, NO. 20, 2023 Verbrugge et al 2015
MAY 23, 2023:2013–2024 Acetazolamide in Heart Failure

of knowledge from observational studies, with the

T A B L E 1 Urine Output and Natriuresis During First 2 Days According to
evidence summarized in a position statement on Treatment Allocation
diuretic use in ADHF of the Heart Failure Association
Placebo Acetazolamide
of the European Society of Cardiology, patients were (n ¼ 231) (n ¼ 231) P Value
grouped according to weighted UNa: <50 mmol/L (ie, Urine collection 1
poor natriuretic response); 50 to 79 mmol/L Volume, mLa 2,820
1,382 3,438
1,570 <0.001
(ie, diminished natriuretic response); 80 to 99 mmol/L UNa, mmol/L 89
27 103
26 <0.001
(ie, good natriuretic response); or $100 mmol/L (ie, FENa, % 3.70
3.53 5.17
3.02 <0.001

excellent natriuretic response). 8-10 Total natriuresis, mmol 192
123 258
133 <0.001
Urine collection 2
ASSESSMENT OF FLUID OVERLOAD. The ADVOR Volume, mLa 2,167
1,078 2,300
980 0.168
trial used a dedicated congestion score to quantify UNa, mmol/L 80
26 91
25 <0.001

fluid overload in patients (Supplemental Figure 1).6 FENa, % 2.47
1.65 3.35
2.21 <0.001
Total natriuresis, mmol 183
131 218
126 0.003
This score ranges from 0 to 10 points, with 4 poten-
Total urine output, mL 4,166
1,789 4,689
1,667 0.001
tial points allocated to edema and 3 points each
Total natriuresis, mmol 375
229 476
229 <0.001
allocated to the presence of pleural effusion and as-
Averaged UNa, mmol/L 86
25 98
24 <0.001
cites. Both pleural effusion and ascites were manda-
tory to be confirmed through ultrasound and/or chest Values are mean
SD. aNormalized for a 24-hour collection period.
FENa ¼ fractional sodium excretion; UNa ¼ urine sodium concentration.
radiograph. Successful decongestion was defined as a
score of 0 to 1, indicating no more than trace edema
and no residual pleural effusion or ascites.
this report have not been adjusted for multiplicity,
STATISTICAL ANALYSIS. Continuous variables are and therefore inferences drawn from these statistics
expressed as mean
SD or median (IQR) and are may not be reproducible. Significance was always set
compared using analysis of variance or the Kruskal- at a 2-tailed probability level <0.05. All statistics were
Wallis H test, as appropriate. Categorical data are performed using SPSS software version 25 (IBM Corp)
expressed as percentages and are compared with and R software version 4.2.2 (R Foundation).
Pearson’s chi-square test. Correlations are described
with Pearson’s r. Predictors of the natriuretic RESULTS
response to diuretic therapy were assessed with a
linear mixed regression model with a random inter- STUDY GROUP. In the overall ADVOR group of 519

cept. Baseline variables associated with P < 0.2 were patients, 502 patients (96.7%) had complete urine
put in the model, which was adjusted for age, sex, collections. Twelve patients (2.3%) had a urine
and congestion score and included treatment alloca- sample <500 mL over 24 hours and were excluded
tion to acetazolamide or placebo as a fixed effect. (Supplemental Table 1). Full UNa results were lacking
Logarithmic transformation was performed for in another 28 patients (5.4%), resulting in a final
skewed variables. The effect of weighted UNa and study population of 462 patients or 89% of the overall
total sodium excretion on clinical decongestion was ADVOR group. A comparison with excluded patients
assessed with linear mixed-effect models, including is presented in Supplemental Table 2. Excluded pa-
the treatment allocation to acetazolamide vs placebo tients were more often women and less often had
and its interaction term with either UNa or total so- undergone cardiac resynchronization therapy.
dium excretion as fixed effects, with a random NATRIURESIS IN THE STUDY GROUP. Weighted UNa
intercept. Cumulative incidence curves were con- was 92
25 mmol/L (Supplemental Figure 2), and
structed according to the Kaplan-Meier method with total natriuresis was 425
234 mmol (Supplemental
the log-rank test used for comparison between Figure 3) in the study group over a total urine
groups. HRs and corresponding 95% CIs were calcu- collection period of 41
4 hours. Both UNa (98

lated with a Cox regression model adjusted for age, 24 mmol/L vs 86
25 mmol/L; P < 0.001) and total
sex, systolic blood pressure, ischemic origin of HF, natriuresis (476
229 mmol vs 375
229 mmol; P <
ejection fraction, diabetes, estimated glomerular 0.001) were significantly higher in the acetazolamide
filtration rate, serum sodium, and logarithmically vs placebo arm of the ADVOR trial, respectively.
transformed N-terminal pro–B-type natriuretic pep- Allocation to acetazolamide was associated with
tide (NT-proBNP). Restricted cubic splines according increased natriuresis according to all metrics tested.
to the best fit method were used to evaluate break- Detailed information on UNa, FENa, total natriuresis,
points in the relationship between natriuresis and and urine output according to treatment allocation is
clinical outcomes. P values and 95% CIs presented in presented in Table 1. Patients with a greater
2016 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

T A B L E 2 Baseline Characteristics of the Study Group

Overall UNa <50 mmol/L UNa 50-79 mmol/L UNa 80-99 mmol/L UNa $100 mmol/L
(N ¼ 462) (n ¼ 36) (n ¼ 88) (n ¼ 142) (n ¼ 196) P Value

Allocation to acetazolamide 231 (50) 12 (33) 33 (38) 61 (43) 125 (64) <0.001
Age, y 78
9 74
10 80
8 78
9 78
9 0.003
Women 164 (36) 13 (36) 39 (44) 47 (33) 65 (33) 0.282
Heart rate, beats/min 78
18 80
18 78
19 78
17 78
18 0.927
Systolic blood pressure, mm Hg 126
21 119
18 120
18 127
22 130
20 <0.001
Diastolic blood pressure, mm Hg 72
13 69
11 68
10 73
13 74
14 0.002
NYHA functional class 0.541
II 62 (13) 4 (11) 12 (14) 20 (14) 26 (13)
III 267 (58) 16 (44) 53 (60) 81 (57) 117 (60)
IV 133 (29) 16 (44) 23 (26) 41 (29) 53 (27)
Congestion score (0-10) (n ¼ 461) 4.38
1.68 4.58
1.78 4.59
1.70 4.23
1.60 4.35
1.72 0.370
Ischemic heart failure 208 (45) 16 (44) 43 (49) 67 (47) 82 (42) 0.659
Ejection fraction, % 43
15 39
17 45
16 43
15 43
15 0.275
Duration of heart failure, y 3 (1-8) 4 (1-10) 4 (1-8) 3 (1-7) 2 (1-7) 0.420
Hypertension 345 (75) 27 (75) 71 (81) 102 (72) 145 (74) 0.505
Atrial fibrillation 338 (73) 25 (69) 68 (77) 101 (71) 144 (74) 0.724
Diabetes 219 (47) 19 (53) 42 (48) 65 (46) 93 (47) 0.903
COPD 90 (20) 8 (22) 20 (23) 32 (23) 30 (15) 0.286
Laboratory measurements
Hemoglobin, g/dL (n ¼ 461) 11.9
2.0 12.4
2.4 11.9
1.9 11.8
1.9 11.9
1.9 0.480
Sodium, mmol/L 139
4 136
7 139
4 139
4 141
3 <0.001
eGFR, mL/min/1.73 m2 (n ¼ 435)a 48
19 44
19 46
21 46
19 50
19 0.121
NT-proBNP, ng/L (n ¼ 458) 6,028 (2,873-10,618) 7,524 (3,429-12,347) 7,188 (2,698-13,952) 5,472 (3,280-10,118) 5,476 (2,560-9,487) 0.118
Medication use
ACE inhibitor/ARB/ARNi 240 (52) 17 (47) 49 (56) 71 (50) 103 (53) 0.787
b-blocker 368 (80) 27 (75) 76 (86) 110 (78) 155 (79) 0.336
MRA 198 (43) 21 (58) 37 (42) 60 (42) 80 (41) 0.273
Digoxin 34 (7) 2 (6) 3 (3) 10 (7) 19 (10) 0.287
Loop diuretic dose (mg furosemide equivalents) 60 (40-100) 100 (80-180) 80 (40-200) 60 (40-100) 45 (40-100) <0.001
Cardiac device
Cardiac resynchronization therapy 59 (13) 4 (11) 15 (17) 15 (11) 25 (13) 0.541
Defibrillator 74 (16) 9 (25) 16 (18) 17 (12) 32 (16) 0.240

Values are n (%), mean
SD, or median (IQR). aResult in stable circumstances 70 days (23-211 days) before randomization.
ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin receptor blocker; ARNi ¼ angiotensin-neprilysin inhibitor; COPD ¼ chronic obstructive pulmonary disease; eGFR ¼ estimated glomerular filtration
rate according to the Chronic Kidney Disease Collaboration formula; MRA ¼ mineralocorticoid receptor blocker; NT-proBNP ¼ N-terminal pro–B-type natriuretic peptide; NYHA ¼ New York Heart Association;
UNa ¼ urine sodium concentration.

natriuretic response had higher blood pressure and compared with an adjusted group average) and a
higher serum sodium levels, and they used lower 115 mmol greater total natriuresis (corresponding to a
doses of loop diuretic agents as maintenance therapy 32% increase compared with an adjusted group
(Table 2). Supplemental Table 3 provides the natri- average) after multivariable adjustment (Table 3).
uresis metrics and urine output according to the UNa Male sex, higher systolic blood pressure, better
strata. Pearson’s r for the correlation between urine glomerular filtration rate, and higher serum sodium
output and total natriuresis was 0.903 (P < 0.001) levels were other independent predictors, although
(Supplemental Figure 4). Pearson’s r for the correla- they had a less pronounced impact. The dose of loop
tion between urine output and UNa was 0.422 diuretic agents was inversely and independently
(P <0.001) (Supplemental Figure 5). associated with weighted UNa, but it did not influ-
DETERMINANTS OF NATRIURESIS. Allocation to ence total natriuresis.
acetazolamide was the strongest independent pre- NATRIURESIS AND CLINICAL DECONGESTION. Pa-
dictor of natriuresis within 48 hours in the ADVOR tients with a higher weighted UNa had faster relief
group; it was associated with a 16 mmol/L increase in from congestion, as assessed by the sequential
weighted UNa (corresponding to a 19% increase congestion score (Supplemental Figure 6). Despite a
JACC VOL. 81, NO. 20, 2023 Verbrugge et al 2017
MAY 23, 2023:2013–2024 Acetazolamide in Heart Failure

similar congestion score at randomization, patients

T A B L E 3 Predictors of Natriuresis
with higher vs lower UNa had a significantly lower
congestion score (indicating less fluid overload) as Univariable Multivariable
Unstandardized b Unstandardized b
early as the next morning, with this difference (95% CI) (95% CI) P Value
increasing during the following 2 days that consti- Predictors of UNa, mmol/L
tuted the randomized treatment period of the trial. Treatment with acetazolamide 12 (8 to 17) 16 (12 to 20) <0.001
The interaction term between UNa and treatment Age (per 10 y) 0 (2 to 3) 0 (2 to 3) 0.779

allocation to acetazolamide or placebo had a signifi- Female 3 (8 to 2) 5 (9 to 1) 0.019
Systolic blood pressure 3 (1 to 4) 2 (1 to 3) 0.002
cant impact on the evolution of the congestion score (per 10 mm Hg)
from randomization to discharge (P ¼ 0.007). Signif- Congestion score (per D1) 1 (2 to 1) 0 (1 to 2) 0.589
icant differences in congestion scores were observed eGFR (per 10 mL/min/1.73 m2) 2 (1 to 3) 2 (1 to 3) <0.001
among the UNa <50 mmol/L, 50 to 79 mmol/L, and 80 Serum sodium (per 1 mmol/L) 2 (2 to 3) 2 (2 to 3) <0.001
to 99 mmol/L strata, whereas the $100 mmol/L group Log(NT-proBNP) 6 (12 to 0) 3 (8 to 3) 0.340
Log(loop diuretic dose) 20 (28 to 13) 13 (21 to 6) <0.001
had similar findings compared with the 80 to
Predictors of total natriuresis, mmol
99 mmol/L group, except at discharge, where UNa
Treatment with acetazolamide 101 (59 to 142) 115 (74 to 156) <0.001
was associated with better decongestion. Similarly,
Age (per 10 y) 27 (51 to -3) 11 (35 to 12) 0.349
the interaction between total natriuresis and treat- Female 40 (85 to 4) 47 (90 to 4) 0.034
ment allocation to acetazolamide or placebo had a Systolic blood pressure 22 (12 to 33) 21 (11 to 31) <0.001
significant impact on the evolution of the congestion (per 10 mm Hg)
Congestion score (per D1) 0 (13 to 13) 6 (7 to 18) 0.382
score from randomization to discharge (P ¼ 0.028).
eGFR (per 10 mL/min/1.73 m2) 19 (8 to 30) 19 (8 to 30) 0.001
Successful decongestion, according to the definition
Serum sodium (per 1 mmol/L) 8 (3 to 12) 8 (3 to 13) 0.001
of the primary ADVOR endpoint, within 72 hours and
Log(NT-proBNP) 76 (130 to 22) 49 (103 to 5) 0.074
without the need for escalation of diuretic therapy, Log(loop diuretic dose) 28 (102 to 46) 18 (54 to 90) 0.620
increased from 8% in the UNa <50 mmol/L group to
45% in the UNa $100 mmol/L group (Figure 1A). Values are unstandardized b (95% CI).
Abbreviations as in Table 2.
Successful decongestion at discharge increased from
44% in the UNa <50 mmol/L group to 73% in the
UNa $100 mmol/L group (Figure 1B). Similar patterns
were observed for higher total natriuresis, yet with HF readmission (HR: 0.92; 95% CI: 0.85-0.99).
less discrimination between the highest strata Restricted cubic spline analysis (Figure 3) showed
(Figures 1C and 1D). For each of these endpoints, the clear breakpoints around UNa w50 mmol/L, below
interaction between natriuresis and the treatment which the risk increased exponentially and around
effect of acetazolamide was significant (P < 0.036 UNa w100 mmol/L, above which the risk decreased
for all). exponentially. Kaplan-Meier curves according to
UNa strata are presented in Figure 4A. Total natri-
NATRIURESIS AND LENGTH OF HOSPITAL STAY. uresis as a continuous variable after adjustments
Patients who survived to hospital discharge with a was not associated with a reduced risk of death or
higher-weighted UNa had a shorter length of stay for HF readmission (HR: 0.96; 95% CI: 0.92-1.01 per
their index admission, ranging from a median of 50 mmol increase in total natriuresis), although its
12.5 days in the UNa <50 mmol/L group to 8 days in categorical analysis among quartiles was statistically
the UNa $100 mmol/L group (Figure 2A). Patients who significant (P ¼ 0.029) (Figure 4B).
survived to hospital discharge with higher total
NATRIURETIC RESPONSE TO ACETAZOLAMIDE AND
natriuresis had a shorter length of stay as well,
CLINICAL OUTCOME. Overall, 186 of 231 or 80.5% of
ranging from a median of 10.5 days in the lowest
patients in the acetazolamide arm of the ADVOR trial
quartile to 8 days in the highest quartile (Figure 2B).
had a good or excellent natriuretic response with a
DEATH OR READMISSION FOR HEART FAILURE. UNa $80 mmol/L. In contrast, in the placebo arm, only
Over 92 days (IQR: 77-99 days) of follow-up, 60 152 of 231 or 65.8% had a UNa $80 mmol/L, which was
patients died (13.0%) and 85 were readmitted for HF significantly lower (P < 0.001). Their baseline charac-
(18.4%). After adjustments for age, sex, systolic teristics are described in Supplemental Table 4. Pa-
blood pressure, ischemic origin of HF, ejection tients with a UNa $80 mmol/L had a similar clinical
fraction, diabetes, estimated glomerular filtration outcome irrespective of their treatment allocation
rate, serum sodium, and logNT-proBNP, every to acetazolamide vs placebo, yet patients allocated
10 mmol/L increase in UNa was associated with a to acetazolamide with a remaining poor natriuretic
lower risk for the combined endpoint of death or response (UNa <80 mmol/L) had significantly worse
2018 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

F I G U R E 1 Percentage of Patients With Successful Decongestion According to Natriuretic Response

A Successful Decongestion Within 3 Days B Successful Decongestion at Discharge

(Primary ADVOR Endpoint)

100% 100%

80% 80%

60% 60%

40% 40%

20% 20%

0% 0%
<50 50-79 80-99 ≥100 <50 50-79 80-99 ≥100
Weighted Urine Sodium Concentration (mmol/L) Weighted Urine Sodium Concentration (mmol/L)

C Successful Decongestion Within 3 Days D Successful Decongestion at Discharge

(Primary ADVOR Endpoint)

100% 100%

80% 80%

60% 60%

40% 40%

20% 20%

0% 0%
14-258 259-402 405-537 542-1,667 14-258 259-402 405-537 542-1,667
Total Natriuresis (mmol) Total Natriuresis (mmol)

Successful decongestion was defined as the absence of any clinical sign of volume overload other than trace edema. (A and C) Within 72 hours and without the need for
escalation of diuretic therapy as defined by the primary endpoint of the ADVOR (Acetazolamide in Decompensated Heart Failure with Volume Overload) trial. (B and D)
At discharge (with deaths counted as unsuccessful decongestion) irrespective of the need for diuretic therapy escalation.

outcome (Central Illustration). When treatment decongestion and clinical outcomes. Key findings are
allocation to acetazolamide vs placebo and UNa as follows: 1) allocation to acetazolamide in the
response $80 mmol/L vs <80 mmol/L were simulta- ADVOR trial was the strongest independent predictor
neously introduced in the adjusted Cox regression of a greater natriuretic response; 2) a stronger natri-
model, only UNa response was significant (P ¼ 0.044), uretic response was associated with faster and more
and treatment allocation was not (P ¼ 0.248). successful decongestion, translating into a shorter
length of stay; 3) the interaction between natriuretic
DISCUSSION response and treatment allocation to acetazolamide
or placebo was statistically significant for all decon-
This analysis from the ADVOR trial offers a unique gestion endpoints; and 4) after multivariable adjust-
perspective on the impact of acetazolamide on ments, every 10 mmol/L increase in UNa was
natriuresis in ADHF and its relationship with associated with an 8% lower risk on the combined
JACC VOL. 81, NO. 20, 2023 Verbrugge et al 2019
MAY 23, 2023:2013–2024 Acetazolamide in Heart Failure

endpoint of death or HF readmissions. These data

F I G U R E 2 Length of Hospital Stay According to Natriuretic Response
further support the upfront use of acetazolamide
together with loop diuretic agents to achieve decon-
gestion in ADHF. Furthermore, they suggest that UNa
A
35 P < 0.001
may be an attractive measure of effective deconges-
tion for future studies.
30

Number of Days in Hospital

Signs and symptoms of fluid overload remain the
most common reasons for hospital admission in HF, 25
irrespective of its cause.11 Inadequate treatment of
congestion is a major driver of adverse outcomes, 20
including death and/or early readmissions.4,5 Impor-
tantly, extracellular volume may be predominantly 15
apparent in the form of fluid (ie, edema, ascites,
pleural effusions), yet it is actually governed by so- 10
dium homeostasis.12 Diuretic treatment in ADHF
produces hypotonic urine (UNa lower than serum 5
sodium concentration), resulting in a more efficient
removal of water compared with salt.10 This discon- 0
<50 50-79 80-99 ≥100
nection becomes even more pronounced after pro-
longed treatment with intravenous diuretic agents Weighted Urine Sodium Concentration (mmol/L)
because natriuresis in such cases drops faster than B
urine output. 8 This effect probably explains why 35
current decongestive strategies perform poorly to P < 0.001
achieve a durable reduction in plasma volume.13 30
Number of Days in Hospital

Proximal diuretic agents such as acetazolamide (but

also sodium-glucose cotransporter-2 inhibitors) have 25
the potential to increase the natriuretic response to
loop diuretic agents because they bring more sodium
20
to Henle’s loop and act as a brake on the neurohu-
15
moral system. 14-17
The results of this prespecified analysis from the
10
ADVOR trial clearly indicate a strong natriuretic effect
of acetazolamide, when provided on top of high-dose 5
loop diuretic agents in ADHF. This effect was
observed for all metrics of natriuresis, including UNa, 0
FENa, and total natriuresis. Although there was a 14-258 259-402 405-537 542-1,667
similarly strong effect on urine output after the initial Total Natriuresis (mmol)
dose of diuretic agents on randomization, this effect
with acetazolamide was no longer significant for Length of the index admission in days according to (A) weighted urine sodium
subsequent diuretic administrations, whereas the concentration strata and (B) quartiles of total natriuresis in patients who survived
natriuretic effect remained clear and highly signifi- to hospital discharge. No corrections for multiple testing were applied.
cant. In other words, treatment with acetazolamide
kept the UNa high, whereas UNa rapidly decreased in
patients who were treated with loop diuretic agents
only, presumably because of more pronounced was the end of the randomized treatment period.7
15,18
neurohumoral activation in this group. This Another important consequence is that these results
finding may explain why the effect of acetazolamide suggest that it is important to use acetazolamide
on decongestion in the main trial continued to grow initially rather than as a rescue therapy to break
stronger after the randomized treatment period diuretic resistance to avoid excessive neurohumoral
ended, up toward discharge. Indeed, the effect of activation and stimulation of salt-preserving systems
acetazolamide on successful decongestion in the in the body. 7 The fact that UNa in the second urine
ADVOR trial was greater at discharge (number needed collection in the acetazolamide group (from the first
to treat of 6) when compared with the third day after to the second morning after randomization) was even
randomization (number needed to treat of 8.5), which higher than UNa immediately after the first dose of
2020 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

F I G U R E 3 Relationship Between Urine Sodium Concentration and Clinical Outcome

10.00

2.00
HR: 1.00
HR

0.50

0.10
20 40 60 80 100 120 140
Weighted UNa

Restricted cubic spline analysis with best fit method to assess the relationship between weighted urine sodium concentration (UNa) and the
HR for death or heart failure readmission during the 90-day follow-up period.

diuretic agents in the loop diuretic–only group further effect of acetazolamide was predominantly driven
corroborates this finding. by an effect on natriuresis.
Urine output was stronger related to total natri- Although allocation to acetazolamide was not
uresis when compared with UNa. Indeed, cumulative associated with a significant effect on the clinical
urine output remains the major determinant of total endpoint of all-cause mortality or HF readmission
sodium excretion. Although total natriuresis may be during the 90-day follow-up period (with the trial
interpreted as a metric for the amount of extracellular underpowered to assess this endpoint), UNa did show
volume and interstitial sodium buffer that is durably a strong and independent association with this
removed, UNa is more a marker of the strength of the outcome in this analysis. Even though patients with
diuretic effect for a given intrinsic renal sodium higher UNa could have represented a group with less
avidity.10,19 Both metrics are obviously linked to some advanced disease and an inherently better outcome,
extent, but the total amount of sodium removed acetazolamide treatment increased the chance of
likely depends on the absolute sodium overload pre- achieving a good natriuretic response, even in very
sent at baseline as well, which may vary among in- old (average 78 years) and sick (NT-proBNP levels
dividual patients. In this analysis, UNa seemed a >5,000 ng/L) patients. Unfortunately, there also
better discriminator with respect to decongestion seems to be a group of patients who, despite treat-
endpoints and its correlation with clinical outcome, ment with acetazolamide, still have a poor natriuretic
although results with total natriuresis were generally response. One may speculate that this group needs an
congruent. even more aggressive approach to achieve decon-
Greater natriuretic response to diuretic agents gestion (ie, higher doses of loop diuretic agents or the
was associated with faster and more complete relief addition of thiazide-like diuretic agents). Alterna-
of congestion, as determined according to a pre- tively, this group could represent a phenotype where
specified congestion score in the ADVOR trial. The the outcome is less driven by extracellular volume
interaction between the treatment effect of acet- overload per se, potentially because the underlying
azolamide and natriuresis was highly significant, disease is too advanced and prognosis is poor
and after adjusting for this interaction, the effect of anyway.
acetazolamide itself was no longer significant. On the basis of the results of this analysis, one may
Findings were very robust when evaluating the hypothesize that UNa could become a valid treatment
evolution of the congestion score over time, the target in ADHF with clear volume overload. Indeed,
proportion of patients who reached successful although dehydration in such patients seemingly re-
decongestion after the randomized treatment period sults in fewer clinical signs of fluid overload, a net
(ie, the primary endpoint of the ADVOR trial), or the negative fluid balance, and effective weight loss,
patients with complete decongestion on discharge. durable decongestion probably needs desalinization,
These findings strongly support the concept that the which has shown a stronger correlation with clinical
JACC VOL. 81, NO. 20, 2023 Verbrugge et al 2021
MAY 23, 2023:2013–2024 Acetazolamide in Heart Failure

F I G U R E 4 Death or Heart Failure Readmission According to Natriuretic Response

A Death or Heart Failure Readmission

50%

40% P = 0.009

30%

20%

10%

0%

0 15 30 45 60 75 90
Days
UNa <50 mmol/L 36 34 29 26 24 22 17
UNa 50-79 mmol/L 88 78 70 67 62 58 49
UNa 80-99 mmol/L 142 135 128 123 119 113 96
UNa ≥100 mmol/L 196 190 183 177 167 159 135

B Death or Heart Failure Readmission

50%

40%

P = 0.029
30%

20%

10%

0%

0 15 30 45 60 75 90
Days
Total Natriuresis
Q1 (14-258 mmol) 115 105 94 90 85 79 67
Q2 (259-402 mmol) 116 113 105 98 91 87 76
Q3 (405-537 mmol) 115 107 101 99 92 88 77
Q4 (542-1,667 mmol) 116 112 109 106 104 98 77

Cumulative incidence of death or heart failure readmission according to (A) weighted urine sodium concentration (UNa) strata and
(B) quartiles (Q) of total natriuresis. No corrections for multiple testing were applied.
2022 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

C E NT R AL IL L U STR AT IO N Effect of Acetazolamide on Natriuresis, Decongestion, and Clinical Outcome

Verbrugge FH, et al. J Am Coll Cardiol. 2023;81(20):2013–2024.

Acetazolamide increased both the urine sodium concentration (UNa) and total natriuresis during the first 48 hours in the ADVOR (Acetazolamide in Decompensated
Heart Failure with Volume Overload) trial. This resulted in more patients (81%) with a strong natriuretic response (urine sodium concentration $80 mmol/L) when
compared with the control group, who received loop diuretic agents and placebo (66%). A strong natriuretic response was associated with faster decongestion, a
shorter length of stay, and a lower risk of death or heart failure readmission. OD ¼ once daily.
JACC VOL. 81, NO. 20, 2023 Verbrugge et al 2023
MAY 23, 2023:2013–2024 Acetazolamide in Heart Failure

outcomes and is supported by this analysis of the should investigate whether a decongestive strategy
ADVOR trial as well. 20,21 Although maybe counterin- titrated on the basis of UNa further improves
tuitive at first sight, patients with ADHF may need decongestion success and, potentially, clinical
enough fluid intake to be able to excrete all the excess outcomes.
salt from their body. One should consider that even ACKNOWLEDGMENTS The authors express their
with an excellent natriuretic response to diuretic gratitude to Katrien Tartaglia and her team from the
agents with UNa w100 mmol/L (or 6 g salt/L urine Clinical Trial Unit Ziekenhuis Oost-Limburg Genk,
output), a urine output of w1.5 L is needed to actually who played a pivotal role in completing the ADVOR
remove 1 L of isotonic fluid (with UNa 154 mmol/L). trial.
With a less pronounced natriuretic response, even
more urine output needs to be generated as the FUNDING SUPPORT AND AUTHOR DISCLOSURES
relative proportion of free water excretion increases
This work was supported by the Belgian Health Care Knowledge
(ie, dehydration). A UNa-driven approach with liberal
Center under the KCE Trials Program (KCE 17001). Dr Damman has
fluid intake, serial UNa measurements, and continued received speaker fees from AstraZeneca, Abbott, and Boehringer
intravenous diuretic therapy with upfront acetazol- Ingelheim. Dr Ruschitzka has not received personal payments by
amide therapy and low-threshold full nephron pharmaceutical companies or device manufacturers in the last 3 years
(remuneration for the time spent in activities, such as participation as
blockade is currently being studied in the randomized
steering committee member of clinical trials and member of the Pfizer
DECONGEST (Diuretic Treatment in Acute Research Award selection committee in Switzerland, was made
Heart Failure With Volume Overload Guided by Serial directly to the University of Zurich); the Department of Cardiology,
Spot Urine Sodium Assessment; NCT05411991) trial. University Hospital of Zurich/University of Zurich has received
research, educational, and/or travel grants from Abbott, Amgen,
STUDY LIMITATIONS. This is a post hoc analysis from AstraZeneca, Bayer, Berlin Heart, B Braun, Biosense Webster, Bio-
a randomized clinical trial. Therefore, the results are sensors Europe, Biotronik, Bristol Myers Squibb, Boehringer Ingel-

exploratory and provide a better understanding and heim, Boston Scientific, Bracco, Cardinal Health Switzerland,
Corteria, Daiichi-Sankyo, Diatools, Edwards Lifesciences, Guidant
context of the results of the main trial, but they
Europe NV (Boston Scientific), Hamilton Health Sciences, Kaneka
should be considered hypothesis-generating. Limita- Corporation, Kantar, Labormedizinisches Zentrum, Medtronic, MSD,
tions of the main trial, including a study group that Mundipharma Medical Company, Novartis, Novo Nordisk, Orion,
Pfizer, Quintiles Switzerland Sarl, Roche Diagnostics, Sahajanand IN,
was exclusively recruited in Belgium and did not use
Sanofi, Sarstedt, Servier, SIS Medical, SSS International Clinical
sodium-glucose cotransporter-2 inhibitors, may Research, Terumo Deutschland, Trama Solutions, V-Wave, Vascular
somewhat limit the external validity of the findings. Medical, Vifor, Wissens Plus, and ZOLL (grants that have not affected
Conversely, baseline characteristics in the ADVOR his personal remuneration). All other others have reported that they
have no relationships relevant to the contents of this paper to
trial were very similar compared with those of other
22
disclose.
important trials in ADHF. Because of specific design
features of the ADVOR trial, UNa was measured from
ADDRESS FOR CORRESPONDENCE: Dr Frederik
a timed urine collection rather than a spot sample
Verbrugge, Centre for Cardiovascular Diseases,
obtained 30 to 120 minutes after loop diuretic
University Hospital Brussels, Laarbeeklaan 101,
administration, as has been recommended by expert
9
1090 Jette, Belgium. E-mail: frederik.verbrugge@
consensus. However, because the strongest diuretic
uzbrussel.be. Twitter: @FH_Verbrugge.
effect takes place within 2 hours of loop diuretic
administration, with effects weaning off after 6 to 8
hours, UNa in such timely spot samples corresponds PERSPECTIVES
well to UNa in timed collections over the whole
period of diuretic therapy efficacy.23 COMPETENCY IN PATIENT CARE AND PROCEDURAL
SKILLS: In patients with volume overload secondary to ADHF,
CONCLUSIONS
the addition of acetazolamide to high-dose loop diuretic therapy
results in substantial natriuresis, shorter hospital stays, fewer
Acetazolamide leads to faster and more complete
readmissions, and longer survival.
decongestion in patients with ADHF and clear clin-
ical signs of volume overload, which is mediated by
TRANSLATIONAL OUTLOOK: In future clinical trials
its effect on natriuresis. UNa may be an attractive
involving patients with acute HF, UNa should be evaluated as a
measure of effective decongestion for future ADHF
simple, objective measure of decongestion associated with clin-
trials because it is simple to assess, objective,
ical outcomes.
strongly related to decongestion, and associated
with clinical outcome. Further randomized studies
2024 Verbrugge et al JACC VOL. 81, NO. 20, 2023

Acetazolamide in Heart Failure MAY 23, 2023:2013–2024

REFERENCES

1. McDonagh TA, Metra M, Adamo M, et al. 2021 reduced ejection fraction. Circ Heart Fail. 2014;7: tive heart failure at high risk for diuretic resis-
ESC guidelines for the diagnosis and treatment of 766–772. tance. Eur J Heart Fail. 2019;21:1415–1422.
acute and chronic heart failure. Eur Heart J.
9. Mullens W, Damman K, Harjola VP, et al. The 17. Griffin M, Rao VS, Ivey-Miranda J, et al. Empa-
2021;42:3599–3726.
use of diuretics in heart failure with congestion – a gliflozin in heart failure: diuretic and cardiorenal
2. Heidenreich PA, Bozkurt B, Aguilar D, et al. position statement from the Heart Failure Associ- effects. Circulation. 2020;142:1028–1039.
2022 AHA/ACC/HFSA guideline for the manage- ation of the European Society of Cardiology. Eur J 18. Verbrugge FH, Tang WH, Mullens W. Renin-
ment of heart failure: a report of the American Heart Fail. 2019;21:137–155. Angiotensin-aldosterone system activation during
College of Cardiology/American Heart Association
10. Verbrugge FH. Utility of urine biomarkers and decongestion in acute heart failure: friend or foe. ?
Joint Committee on Clinical Practice Guidelines.
electrolytes for the management of heart failure. J Am Coll Cardiol HF. 2015;3:108–111.
J Am Coll Cardiol. 2022;79(17):e263–e421. https://
Curr Heart Fail Rep. 2019;16:240–249. 19. Nijst P, Verbrugge FH, Grieten L, et al. The
doi.org/10.1016/j.jacc.2021.12.012
11. Chioncel O, Mebazaa A, Harjola VP, et al. pathophysiological role of interstitial sodium in
3. Felker GM, Lee KL, Bull DA, et al. Diuretic
Clinical phenotypes and outcome of patients heart failure. J Am Coll Cardiol. 2015;65:378–388.
strategies in patients with acute decompensated
hospitalized for acute heart failure: the ESC 20. Verbrugge FH. Decongestion: more than
heart failure. N Engl J Med. 2011;364:797–
Heart Failure Long-Term Registry. Eur J Heart Fail. meets the eye. Eur J Heart Fail. 2016;18:192–194.
805.
2017;19:1242–1254.
4. Metra M, Davison B, Bettari L, et al. Is wors- 21. Hodson DZ, Griffin M, Mahoney D, et al.
ening renal function an ominous prognostic sign in 12. Mullens W, Verbrugge FH, Nijst P, Tang WHW. Natriuretic response is highly variable and associ-
patients with acute heart failure? The role of Renal sodium avidity in heart failure: from path- ated with 6-month survival: insights from the
congestion and its interaction with renal function. ophysiology to treatment strategies. Eur Heart J. ROSE-AHF trial. ? J Am Coll Cardiol HF. 2019;7:
Circ Heart Fail. 2012;5:54–62. 2017;38:1872–1882. 383–391.

5. Rubio-Gracia J, Demissei BG, Ter Maaten JM, 13. Miller WL, Mullan BP. Understanding the het- 22. Mullens W, Dauw J, Martens P, et al. Acet-
et al. Prevalence, predictors and clinical outcome erogeneity in volume overload and fluid distribution azolamide in Decompensated Heart Failure with
of residual congestion in acute decompensated in decompensated heart failure is key to optimal Volume Overload trial (ADVOR): baseline charac-
heart failure. Int J Cardiol. 2018;258:185–191. volume management: role for blood volume quan- teristics. Eur J Heart Fail. 2022;24:1601–1610.
titation. J Am Coll Cardiol HF. 2014;2:298–305.
6. Mullens W, Verbrugge FH, Nijst P, et al. Ratio- 23. Collins SP, Jenkins CA, Baughman A, et al.
nale and design of the ADVOR (Acetazolamide in 14. Lambers Heerspink HJ, de Zeeuw D, Wie L, Early urine electrolyte patterns in patients with
Decompensated Heart Failure with Volume Over- Leslie B, List J. Dapagliflozin a glucose-regulating acute heart failure. ESC Heart Fail. 2019;6:80–88.
load) trial. Eur J Heart Fail. 2018;20:1591–1600. drug with diuretic properties in subjects with type
2 diabetes. Diabetes Obes Metab. 2013;15:853–862.
7. Mullens W, Dauw J, Martens P, et al. Acet- KEY WORDS acetazolamide, cardiorenal
azolamide in Acute Decompensated Heart Failure 15. Verbrugge FH, Dupont M, Steels P, et al. The
syndrome, diuretic agents, heart failure,
with Volume Overload. N Engl J Med. 2022;387: kidney in congestive heart failure: ’are natriuresis,
natriuresis, prognosis
1185–1195. sodium, and diuretics really the good, the bad and
the ugly?’. Eur J Heart Fail. 2014;16:133–142.
8. Verbrugge FH, Nijst P, Dupont M, Penders J,
Tang WH, Mullens W. Urinary composition during 16. Verbrugge FH, Martens P, Ameloot K, et al. A PPE NDI X For supplemental figures and ta-
decongestive treatment in heart failure with Acetazolamide to increase natriuresis in conges- bles, please see the online version of this paper.