Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
https://doi.org/10.1007/s00392-019-01521-y
ORIGINAL PAPER
Abstract
Background Diuretic resistance is a common issue in patients with acute decompensation of advanced chronic heart failure
(ACHF). The aim of this trial was to compare boluses and continuous infusion of furosemide in a selected population of
patients with ACHF and high risk for diuretic resistance.
Methods In this single-centre, double-blind, double-dummy, randomized trial, we enrolled 80 patients admitted for acute
decompensation of ACHF (NYHA IV, EF ≤ 30%) with criteria of high risk for diuretic resistance (SBP ≤ 110 mmHg, wet
score ≥ 12/18, and sodium ≤ 135 mMol/L). Patients were assigned in a 1:1 ratio to receive furosemide by bolus every 12 h
or by continuous infusion. Diuretic treatment and dummy treatment were prepared by a nurse unassigned to patients’ care.
The study treatment was continued for up to 72 h. Coprimary endpoints were total urinary output and freedom from conges-
tion at 72 h.
Results 80 patients were enrolled with 40 patients in each treatment arm. Mean daily furosemide was 216 mg in continu-
ous-infusion arm and 195 mg in the bolus intermittent arm. Freedom from congestion (defined as jugular venous pressure
of < 8 cm, with no orthopnea and with trace peripheral edema or no edema) occurred more in the continuous infusion than
in the bolus arm (48% vs. 25%, p = 0.04), while total urinary output after 72 h was 8612 ± 2984 ml in the bolus arm and
10,020 ± 3032 ml in the continuous arm (p = 0.04). Treatment failure occurred less in the continuous-infusion group (15%
vs. 38%, p = 0.02), while there was no significant difference between groups in the incidence of worsening of renal function.
Conclusion Among patients with acute decompensation of ACHF and high risk of diuretic resistance, continuous infusion
of intravenous furosemide was associated with better decongestion.
DRAIN trial ClinicalTrials.gov number NCT03592836.
* Simone Frea
frea.simone@gmail.com
1
Division of Cardiology, Cardiovascular and Thoracic
Department, Città della Salute e della Scienza University
Hospital of Turin, Corso Bramante 88, 10126 Turin, Italy
13
Vol.:(0123456789)
Clinical Research in Cardiology
Graphic abstract
Keywords Diuretic resistance · Advanced heart failure · Furosemide · Bolus intermittent · Continuous infusion
13
Clinical Research in Cardiology
Inclusion criteria were: age ≥ 18 years, all criteria for dobutamine ≥ 10 γ/kg/min) were required due to shock evolu-
ACHF as defined by Metra et al. [8] EF at admission ≤ 30% tion (development of INTERMACS ≤ 2), vasopressors were
and all the following parameters of high risk for diu- started and patients left the study protocol.
retic resistance (systolic blood pressure ≤ 110 mmHg, Thiazide diuretics were not allowed during the study proto-
serum sodium ≤ 135 mEq/L, and severe congestion (wet col. Other p.o. therapies were administered according to ESC
score ≥ 12/18; according to a modified version—the wet guidelines recommendations and were administered according
score—of a congestion grading score from Gheorghiade [10, 11]. All patients followed a water restriction of 1000 ml/
et al. [9]—see Supplementary Material). Exclusion crite- day and laid in bed up to the end of the protocol.
ria were: reversible causes of acute heart failure (such as The study treatment was continued for up to 72 h. After
acute coronary syndrome, myocarditis, acute pulmonary 72 h, all treatments were open label at the discretion of the
embolism, rhythm disorders, and severe primary organic treating physician. Patients were followed for clinical events
valve disease), cardiogenic shock at admission (defined as (overall and cardiac death, hospitalization for worsening heart
INTERMACS class ≤ 2), and eGFR (MDRD) lower than failure, urgent heart transplantation, and LVAD implantation)
15 ml/min/1.73 m2. until day 90.
Patients who met inclusion/exclusion criteria were randomly At admission, clinical history (age, gender, etiology of heart
assigned in a 1:1 ratio to administration of furosemide either failure, INTERMACS class, and medications intake at home)
by intravenous bolus every 12 h or by continuous intrave- was recorded and physical examination was performed in all
nous infusion. Randomization was carried out by the use patients. Clinical examination included grading of congestion
of sequentially numbered cases prepared before starting the (according to wet score) and assessment of hypoperfusion
study by a computerized sequence. A double-blind, double- (according to the cold-modified model [12]).
dummy design was used. A nurse unassigned to patients’ An assessment of biomarkers, including creatinine, blood
care prepared a syringe pump for continuous infusion and urea nitrogen (BUN), serum neutrophil gelatinase-associated
syringes for boluses. According to the assigned treatment lipocalin (NGAL), serum and urinary electrolytes, hepatic
arm, syringes contained the assigned dose of furosemide or function (GOT/GPT and bilirubin), ammonia [13], high sensi-
a 5% glucose solution placebo. All patients received a bolus tivity Troponin T, and NT-proBNP, was performed at baseline,
of furosemide (40 or 60 mg) in the emergency department at 24, 48, and 72 h.
(length of stay < 90 min). Maximum time from admission
to diuretic treatment according to study protocol was 2 h. Echocardiogram
The dose of furosemide (low dose—120 mg/day—vs.
high dose—240 mg/day) was defined before randomization Two-dimensional transthoracic echocardiography was per-
according to the following criteria: patients with severe sys- formed at admission using Philips i33 (Philips Medical Sys-
temic congestion (wet score ≥ 14/18) and/or on daily out-of- tems, Andover, MA, USA). Left ventricular (LV) chamber
hospital oral furosemide equivalent dose ≥ 125 mg and/or on size and systo-diastolic function (with EF evaluation and E/E’
supplemental metolazone therapy on a long-term basis were ratio) were measured. Aortic and mitral regurgitation were
treated with high doses; otherwise, they were included in assessed using color Doppler. Basal right ventricular (RV)
the low-dose arm. After 48 h, the treating physician on the end-diastolic diameter and tricuspid annular plane systolic
basis of the clinical response had the option to reduce the excursion (TAPSE) were measured. Trans-tricuspid systolic
dose from 240 mg/day to 120 mg/day. gradient (TR gradient) and systolic pulmonary artery pressure
Pharmacological treatment and continuous positive air- (sPAP) were estimated by tricuspid regurgitation peak veloc-
way pressure ventilation, when indicated, were started after ity (TRV) and right atrial pressure. Right atrial pressure was
randomization. All patients underwent a prespecified phar- estimated by inferior vena cava diameter and collapse with
macological protocol. If SBP at admission was > 90 mmHg, sniff and the tricuspid E/e’ ratio and hepatic vein flow analysis
intravenous vasodilator (sodium nitroprusside) was started. in doubtful cases [14]. Finally, RV contraction pressure index
Subsequently, if SBP remained ≥ 90 mmHg without signs was calculated as TAPSE x TR gradient [15].
of hypoperfusion, vasodilator therapy was continued; oth-
erwise, it was discontinued. In patients with hypoperfusion Endpoints
(assessed by Cold Modified 2014), dopamine or dobutamine
was started. Inotropes were started in patients with cold pro- The primary endpoint was freedom from congestion at 72 h
file according to Cold Modified 2014 and titrated according (defined as jugular venous pressure of < 8 cm, with no ortho-
to clinical response. If high doses of inotropes (dopamine or pnea and with trace peripheral edema or no edema).
13
Clinical Research in Cardiology
Secondary endpoints were: total urinary output after assigned to receive continuous intravenous furosemide and
72 h; treatment failure (defined as persistent congestion (wet 40 patients to receive bolus intermittent intravenous admin-
score ≥ 12/18) with the need of increasing diuretic treatment istration of furosemide. There were no crossovers. Seventy-
until renal replacement treatment) at 72 h; worsening renal seven patients completed the 72 h protocol. As shown in
function (defined as an absolute increase in serum creati- the diagram in Fig. 1 according to CONSORT 2010 [19],
nine > 0.3 mg/dl or > 1.5-fold from baseline) at any time three patients left the protocol before 72 h due to evolu-
from randomization to 72 h; diuretic response (defined as ∆ tion to cardiogenic shock, but were included in the analysis.
weight/40 mg furosemide [16, 17]) after 72 h; and worsen- 57 patients met criteria for high-dose furosemide, while 23
ing or persistent heart failure (defined as worsening of HF patients received low-dose furosemide.
symptoms or failure of the patient’s condition to improve Baseline characteristics are shown in Table 1. The patient
with treatment requiring the initiation, or increase in intra- population had several very high-risk features, including
venous inotropic therapy for HF and/or the implementation severe symptoms of ACHF (NYHA class IV for all with a
of mechanical circulatory or ventilator support [18]) up until mean INTERMACS class at admission of 3.7 ± 0.9), crite-
72 h and variations of NT-proBNP after 72 h. Renal replace- ria for high-risk in-hospital mortality for acute heart failure
ment treatment was indicated in patients with persistence [average SBP 98 ± 8.6 mmHg, mean EF 19.3 ± 7.7%, BUN
of congestion (wet score ≥ 12/18) and reduced response to 46.4 ± 20.6, and NT-proBNP 7820 (IQR 4508–14,302)], and
diuretic (absence of weight reduction and/or sodium excre- criteria for high-risk in-hospital diuretic resistance (serum
tion < 30 mmol/l) despite mean arterial pressure ≥ 65 mmHg sodium 129.6 ± 5.3 mmol/L and creatinine 1.8 ± 0.6 mg/dl).
and optimal medical therapy including sequential nephrone The demographic and clinical characteristics were similar
blockade. in both groups as well as medical therapy. Particularly, wet
score was similar (13.9 ± 1.2 in bolus arm vs. 13.6 ± 1.5 in
Statistics infusion arm, p = 0.32), as well as cold profile at admission
(47% in bolus arm vs. 53% in infusion arm).
This trial was designed to assess the superiority of con-
tinuous infusion over bolus infusion. A sample size of 78 Clinical outcomes
patients was required to have a 70% chance of detecting,
as significant at the 10% level, a decrease in the primary The primary endpoint of freedom from congestion after
outcome measure from 48% in the bolus group to 25% in 72 h occurred in 10 patients (25%) in the bolus arm and in
the infusion group. 19 (48%) in the infusion arm (OR 2.71, 95% CI 1.05–7.00;
Continuous variables are reported as mean (SD) and cate- p = 0.04, Fig. 2).
gorical variables as frequencies and percentages. Differences Secondary endpoints are shown in Table 2. Total uri-
in the basal characteristics and endpoints among treatment nary output after 72 h was 8612 ± 2984 ml in the bolus arm
arms were analyzed using Student’s t test for quantitative 10,020 ± 3032 ml in the continuous arm (p = 0.04). Diuretic
data (or Kruskal–Wallis non-parametric test if the variable response was higher in the infusion arm than in the bolus
did not fit the normal distribution) and the χ2 test for cat- arm (− 1 ± 0.7 vs. − 0.6 ± 0.6 kg/40 mg furosemide/72 h;
egorical data. p < 0.01). Treatment failure showed a higher incidence in the
Prespecified subgroup analyses (by renal function, bolus arm (38% vs. 15%, p = 0.02). This result was driven
haemodynamic profile, and diuretic dose) were performed. from the need to increase the dose of diuretic at the end of
The significant threshold was corrected by Bonferroni protocol, while renal replacement treatment was needed in
correction. only one patient in each arm. Change in NT-proBNP at 72 h
All data were analyzed by intention-to-treat. A two- was not significantly different between arms.
sided p value < 0.05 was considered statistically significant; Worsening heart failure after 72 h of treatment occurred
all analyses were performed with SPSS 20.0 (IBM corp., in 20 patients (25%), without any significant difference
Armonk, NY, USA). between the two groups (bolus arm 30% vs. 20% infusion
arm). Worsening renal function was similar between the
bolus arm and the infusion arm (18% vs. 18%) and between
Results the low-dose arm and the high-dose arm. Particularly, no
difference was found in the change in serum creatinine
Baseline characteristics (− 0.04 ± 0.50 vs. − 0.03 ± 0.29 mg/dl for bolus arm and
infusion arm, respectively; p = 0.91—Fig. 3) and in glomeru-
From May 2013 to December 2016, 80 patients were lar filtration rate (+ 4.92 ± 20.42 vs. + 0.86 ± 15.33 ml/min
enrolled. The 90 day follow-up and the study were com- for bolus arm and infusion arm, respectively; p = 0.32) after
pleted on December 2017. 40 patients were randomly 72 h. In addition, changes in serum sodium, BUN, BUN/
13
Clinical Research in Cardiology
Excluded (n=137)
Not meeting inclusion criteria (n=133)
Declined to participate (n=4)
Randomized (n=80)
Allocation
Allocated to bolus intermittent (n=40) Allocated to continuous infusion (n=40)
- Low dose (n=15) - Low dose (n=8)
- High dose (n=25) - High dose (n=32)
Discontinued intervention (evolution to Discontinued intervention (evolution to
cardiogenic shock) (n=2) cardiogenic shock) (n=1)
Follow-Up
Death (n=3) Death (n=3)
LVAD implantation (n=3) LVAD implantation (n=1)
Heart transplantation (n=1) Heart transplantation (n=1)
Lost to follow-up (n=0) Lost to follow-up (n=0)
Analysis
Analysed (n=40) Analysed (n=40)
Excluded from analysis (n=0) Excluded from analysis (n=0)
crea, and NGAL at 72 h were not different between treat- Subgroup analysis
ment arms. At echo-Doppler after 72 h, the infusion arm
showed a lower estimated right atrial pressure (13.5 ± 4.9 A prespecified subgroup analysis testing the primary end-
vs. 16.4 ± 4.5 mmHg, p < 0.01), while other echo-Doppler point was performed, as shown in Table 3. In particular,
variables were not different between treatment arms. furosemide continuous infusion and bolus intermittent were
Within the 90 day follow-up, 6 patients died, 4 patients compared in patients with or without severe renal dysfunc-
underwent LVAD implantation, and two patients underwent tion (eGFR < 30 ml/min), in those with or without hemody-
heart transplantation. In-hospital death occurred in four namic cold profile at admission, and in high vs. low diuretic
(5%) patients. There were no differences in the rate of sin- dose group.
gle events or in the composite of death, LVAD implantation As shown in Table 4, the primary endpoint occurred less
and heart transplantation between continuous-infusion group in the continuous-infusion arm in patients with a warm pro-
and the bolus group (7 vs. 5 patients, respectively, p 0.8). file at admission (OR 0.23, 95% CI 0.06–0.81, p = 0.02).
13
Clinical Research in Cardiology
Dyspnoea Absent Use of one pillow Use of more than one pillow Sleeps in an
armchair or
in a seated
position
Jugular venous pressure (cm) < 8 8–10 11–15 > 16
Hepatomegaly Absent Liver edge Moderate enlargement Massive tender
enlargement
extending to
midline
Oedema None Below knee Over knee Sacral/anasarca
NT-proBNP (pg/ml) < 400 400–1500 1500–3000 > 3000
6 min walk test (m) > 300 200–300 100–200 < 100
Wet score (= sum of all items score): < 6 mild congestion, > 12/18 severe congestion
13
Clinical Research in Cardiology
Table 2 Baseline characteristics
Bolus intermittent (N = 40) Continuous infusion (N = 40) p value
General characteristics
Age (years) 58.7 ± 10.3 63.0 ± 13.1 0.11
Male sex—no. (%) 35 (88) 37 (93) 0.71
Ischemic etiology—no. (%) 22 (55) 25 (63) 0.65
EF (%) 19.2 ± 6.4 19.4 ± 9.0 0.91
ACE-I or ARB—no. (%) 15 (38) 21 (53) 0.26
MRA—no. (%) 32 (80) 30 (75) 0.79
β-blockers—no. (%) 38 (95) 34 (85) 0.26
Thiazides—no. (%) 22 (55) 21 (53) 0.99
Daily furosemide (mg/die) 242.2 ± 176.9 198.6 ± 188.0 0.29
Clinical, laboratory and echo-Doppler characteristics at admission
Systolic blood pressure (mmHg) 97.0 ± 8.6 99.0 ± 8.6 0.30
Heart rate (bpm) 83.0 ± 15.3 81.7 ± 17.5 0.72
INTERMACS profile (n/7) 3.6 ± 0.9 3.7 ± 0.9 0.62
Composite congestion score—no. (%) 13.9 ± 1.2 13.6 ± 1.5 0.32
Cold profile at admission—no. (%) 19 (47) 21 (53) 0.82
Sodium (mmol/L) 130.7 ± 5.4 129.3 ± 5.2 0.23
Creatinine (mg/dl) 1.8 ± 0.6 1.7 ± 0.7 0.49
eGFR (MDRD) (ml/min/1.73 m2) 45.9 ± 16.9 52.1 ± 33.2 0.30
BUN (mg/dl) 45.8 ± 22.5 46.9 ± 30.8 0.86
BUN/crea 56.6 ± 19.0 59.7 ± 22.2 0.50
NGAL (μg/L) 324.7 ± 193.2 213.8 ± 148.6 0.12
Ammonia (μg/dl) 131.8 ± 42.2 118.8 ± 69.7 0.32
Bilirubin (mg/dl) 1.57 ± 0.82 1.65 ± 1.08 0.71
NT-proBNP (pg/ml) 6250 (4303–11,820) 10,101 (4575–19,359) 0.39
EF (%) 19.2 ± 6.4 19.4 ± 9.0 0.91
E/E’ 19.9 ± 7.6 19.7 ± 6.6 0.90
TAPSE (mm) 12.8 ± 2.7 13.4 ± 3.9 0.43
eRAP (mmHg) 18.5 ± 3.5 19.6 ± 2.4 0.23
Use of medication from randomization through 72 h—no. (%)
High-dose furosemide (240 mg die)—no. (%) 32 (80) 25 (63) 0.14
Inotropic agent (DA or DBT ≥ 5 μg/kg/min) 5 (13) 4 (10) 0.7
Vasodilator 34 (85) 36 (90) 0.52
EF left ventricular ejection fraction, ACE-I ACE inhibitors, ARB angiotensin receptor blockers, MRA mineralocorticoid receptor antagonist,
eGFR estimated glomerular filtration rate, BUN blood urea nitrogen, NGAL neutrophil gelatinase-associated lipocalin, TAPSE tricuspid annular
plane systolic excursion, eRAP estimated right atrial pressure, DA dopamine, DBT dobutamine
13
Clinical Research in Cardiology
Table 4 Endpoints
Bolus intermittent (N = 40) Continuous infusion (N = 40) p value OR (95% CI)*
13
Clinical Research in Cardiology
13