HIGHLIGHTS
- •
Prompt initiation of diuretic therapy was not correlated with faster decongestion in acute heart failure.
- •
The timing of decongestion but not diuretics was associated with better biomarker trajectories.
- •
Residual congestion rather than the timing of decongestion or diuretics predicted prognosis.
Prompt treatment may mitigate the adverse effects of congestion in the early phase of heart failure (HF) hospitalization, which may lead to improved outcomes. We analyzed 814 acute HF patients for the relationships between time to first intravenous loop diuretics, changes in biomarkers of congestion and multiorgan dysfunction, and 1-year composite end point of death or HF hospitalization. B-type natriuretic peptide (BNP), high sensitivity cardiac troponin I (hscTnI), urine and serum neutrophil gelatinase–associated lipocalin, and galectin 3 were measured at hospital admission, hospital day 1, 2, 3 and discharge. Time to diuretics was not correlated with the timing of decongestion defined as BNP decrease ≥ 30% compared with admission. Earlier BNP decreases but not time to diuretics were associated with earlier and greater decreases in hscTnI and urine neutrophil gelatinase–associated lipocalin, and lower incidence of the composite end point. After adjustment for confounders, only no BNP decrease at discharge was significantly associated with mortality but not the composite end point (p = 0.006 and p = 0.062, respectively). In conclusion, earlier time to decongestion but not the time to diuretics was associated with better biomarker trajectories. Residual congestion at discharge rather than the timing of decongestion predicted a worse prognosis.
In patients with acute heart failure (AHF), prompt initiation of decongestive therapy may help mitigate the adverse effects of congestion and multi-organ injury in the early phase of hospitalization and result in improved outcomes. Several studies have examined the relationship between early initiation of diuretics and/or vasoactive agents and clinical outcomes. However, a pathophysiologic link between early treatment, decongestion, organ damage and clinical outcomes has not been fully assessed, especially from the viewpoint of biomarker trajectories. B-type natriuretic peptide (BNP) is a well-established biomarker of congestion and the timing of BNP decrease can serve as an objective surrogate for time to decongestion. In this sub-analysis of the Acute Kidney Injury Neutrophil gelatinase–associated lipocalin (NGAL) Evaluation of Symptomatic heart failure Study (AKINESIS), we investigated whether (1) early diuretic therapy results in early decongestion, defined by a decrease in BNP, (2) early diuretic therapy versus decongestion correlates with favorable biomarker changes, and (3) early diuretic therapy versus decongestion is associated with better clinical outcomes.
Methods
We retrospectively analyzed patients enrolled in the AKINESIS. AKINESIS was a prospective, international, multicenter cohort study of AHF patients, which enrolled 927 AHF patients at 16 sites in the United States and Europe from January 2011 through September 2013. The study methods have been previously described. Briefly, subjects with symptoms and signs of AHF and having received or planned administration of intravenous diuretic therapy were enrolled. In the current analysis, we excluded 70 patients discharged from the emergency department, 28 patients without BNP data on admission, and 15 patients lacking serial BNP measurements, leaving 814 patients for this analysis.
Time to diuretics was defined as the time from presentation to the first intravenous loop diuretic administration. Time to BNP decrease was defined as the day when BNP decreased by ≥ 30% compared with admission and the last BNP value was also ≥ 30% lower than admission. Serum and urine specimens for biomarker assessment were tested at 5 time-points if the participant remained hospitalized for the duration of collection times: (1) day of enrolment within 2-hour of first diuretic administration; (2) hospital day 1; (3) hospital day 2; (4) hospital day 3, and (5) the day of discharge or anticipated discharge. Biomarkers analyzed, in addition to BNP, were high sensitivity cardiac troponin I (hscTnI), urine and serum NGAL (uNGAL and sNGAL) and galectin 3 (Gal3). Details of biomarker assessment were reported in elsewhere. uNGAL was indexed to urine creatinine to account for urine tonicity. Levels of serum creatinine were measured each day during hospitalization.
The primary outcome was a 1-year composite of death or HF hospitalization. In-hospital, 1-year mortality and 1-year HF hospitalization were analyzed as secondary outcomes.
Patient characteristics were evaluated stratifying time to diuretics and BNP decrease with analysis of variance, Kruskal–Wallis or chi-square test as appropriate. Time to diuretics was compared by the time to BNP decrease both as a continuous variable and dichotomized as before or after 60 minutes. Loop diuretic equivalents were converted with 1 mg bumetanide = 20 mg torsemide = 40 mg furosemide for intravenous doses and 1 mg bumetanide = 20 mg torsemide = 80 mg furosemide for oral doses. , Relative changes of biomarkers from admission to each sample collection were compared stratifying quartiles of time to diuretics and time to BNP decrease with Kruskal–Wallis test. Kaplan-Meier, log-rank and Cox regression analysis were performed for 1-year outcomes. In multivariable Cox regression analysis, confounding factors were selected based on prior studies including age, African American race, history of chronic obstructive pulmonary disease , edema, systolic blood pressure , heart rate, sodium, hemoglobin, blood urea nitrogen (BUN) and hscTnI. To investigate factors associated with BNP ≥ 30% decrease at day 1, and with BNP < 30% decrease at discharge, univariable and multivariable logistic regression analyses were performed. Factors with p-value < 0.05 in univariable analysis were included in the multivariable model. All biomarkers were log-2 transformed in Cox and logistic regression analyses. All statistical analyses were performed using R version 3.6.3 for Windows.
Results
The mean age of the 814 patients was 69 ± 14 years, 63% were male, 47% had a history of coronary artery disease (CAD), 44% had diabetes, and 26% reported a history of chronic kidney disease. Median creatinine was 1.20 mg/dl (interquartile ranges [IQR] 0.94 to 1.61 mg/dl) and median BNP was 569 ng/l (IQR 233 to 1108 ng/l). Median dose of intravenous furosemide equivalents within the first 3 days of hospitalization was 60 mg/day (IQR 40 to 100 mg/day). BNP decrease at day 1, 2, 3 and discharge were observed in 166 (20%), 148 (18%), 55 (6.7%), and 52 patients (6.4%), respectively. Patients with earlier BNP decrease were younger, more likely to be non-white race, less frequently had a history of chronic kidney disease and had jugular vein distention on admission ( Table 1 ). They had higher systolic blood pressure, and lower levels of creatinine and Gal3 on admission. The prevalence of CAD and interventions were not different between the groups.
| Variables | Day 1 | Day 2 | Day3 | At discharge | No Decrease | |
|---|---|---|---|---|---|---|
| n = 166 | n = 148 | n = 55 | n = 52 | n = 393 | p-value | |
| Age (years), mean±SD | 65±14 | 66±14 | 68±12 | 71±14 | 71±14 | <0.001 |
| Non-white | 73 (44%) | 65 (44%) | 22 (40%) | 10 (19%) | 124 (32%) | 0.001 |
| Men | 103 (62%) | 91 (62%) | 37 (67%) | 32 (62%) | 247 (63%) | 0.958 |
| Coronary artery disease | 70 (42%) | 68 (46%) | 30 (55%) | 25 (48%) | 193 (49%) | 0.471 |
| Myocardial infarction | 44 (27%) | 38 (26%) | 18 (33%) | 11 (21%) | 111 (28%) | 0.693 |
| Percutaneous coronary intervention | 32 (19%) | 28 (19%) | 11 (20%) | 16 (31%) | 93 (24%) | 0.330 |
| Coronary artery bypass grafting | 23 (14%) | 23 (16%) | 14 (26%) | 8 (15%) | 73 (19%) | 0.305 |
| Hypertension | 135 (81%) | 124 (84%) | 46 (84%) | 37 (71%) | 310 (79%) | 0.306 |
| Hyperlipidemia | 84 (51%) | 84 (57%) | 29 (53%) | 26 (50%) | 197 (50%) | 0.724 |
| Diabetes mellitus | 71 (43%) | 66 (45%) | 24 (44%) | 27 (52%) | 173 (44%) | 0.843 |
| Chronic obstructive pulmonary disease | 45 (27%) | 46 (31%) | 11 (20%) | 5 (10%) | 102 (26%) | 0.034 |
| Chronic kidney disease | 27 (16%) | 40 (27%) | 9 (16%) | 20 (39%) | 117 (30%) | 0.001 |
| Smoker | 31 (19%) | 24 (16%) | 12 (22%) | 7 (14%) | 53 (14%) | 0.359 |
| Angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker | 52 (31%) | 43 (29%) | 23 (42%) | 16 (31%) | 122 (31%) | 0.528 |
| Beta-blockers | 123 (74%) | 108 (73%) | 39 (71%) | 42 (81%) | 270 (69%) | 0.352 |
| Diuretic agents | 118 (71%) | 103 (70%) | 40 (73%) | 39 (75%) | 276 (70%) | 0.950 |
| Loop diuretic dose within the first 3 days of hospitalization (mg/day), median [IQR] | 60 [33, 80] | 63 [40, 99] | 67 [40, 103] | 62 [40, 133] | 60 [40, 107] | 0.081 |
| Edema | 117 (71%) | 110 (74%) | 47 (86%) | 35 (67%) | 305 (78%) | 0.087 |
| Jugular vein distension | 24 (15%) | 40 (27%) | 15 (27%) | 21 (40%) | 118 (30%) | 0.001 |
| Rales | 66 (40%) | 61 (41%) | 27 (49%) | 30 (58%) | 177 (45%) | 0.176 |
| Systolic blood pressure (mm Hg), mean±SD | 148±30 | 145±30 | 141±28 | 136±28 | 135±29 | <0.001 |
| Heart rate (beats/min), mean±SD | 89±21 | 88±22 | 93±26 | 84±19 | 87±24 | 0.185 |
| Sodium (mEq/l), median [IQR] | 139 [137, 141] | 139 [138, 142] | 140 [137, 142] | 140 [137, 142] | 139 [136, 141] | 0.001 |
| Hemoglobin (g/dl), mean±SD | 12.0±2.5 | 11.6±2.2 | 11.8±2.3 | 11.4±2.0 | 11.5±2.7 | 0.188 |
| Blood urea nitrogen (mg/dl), median [IQR] | 21.0 [16.0, 29.0] | 23.0 [15.0, 32.5] | 24.0 [16.0, 31.5] | 33.0 [18.9, 58.3] | 26.0 [18.3, 42.0] | <0.001 |
| Creatinine (mg/dl), median [IQR] | 1.10 [0.83, 1.36] | 1.21 [1.00, 1.64] | 1.17 [1.00, 1.50] | 1.22 [1.05, 1.63] | 1.25 [0.96, 1.70] | 0.001 |
| B-type natriuretic peptide (ng/l), median [IQR] | 550 [268, 1065] | 769 [313, 1260] | 713 [252, 1168] | 688 [371, 1268] | 456 [195, 1027] | 0.003 |
| High-sensitivity cardiac troponin I (ng/l), median [IQR] | 22.2 [11.8, 51.3] | 24.1 [12.6, 62.1] | 32.1 [17.1, 98.8] | 27.5 [15.3, 56.0] | 25.5 [12.5, 55.5] | 0.229 |
| Urine neutrophil gelatinase-associated lipocalin (ug/g), median [IQR] | 23.2 [13.6, 57.8] | 31.7 [12.9, 71.0] | 26.8 [13.8, 54.2] | 42.7 [16.3, 104.2] | 25.7 [12.8, 70.0] | 0.340 |
| Serum neutrophil gelatinase-associated lipocalin (ng/ml), median [IQR] | 121.4 [74.3, 196.4] | 140.5 [77.4, 249.0] | 119.7 [87.8, 227.8] | 196.5 [101.3, 359.6] | 146.7 [87.2, 247.5] | 0.008 |
| Galectin 3 (ng/ml), median [IQR] | 22.8 [18.3, 30.7] | 24.2 [19.2, 31.2] | 23.3 [18.7, 29.9] | 29.0 [22.3, 43.6] | 27.9 [20.4, 41.2] | <0.001 |
| IQR, Interquartile range; SD, standard deviation | ||||||
The median time to first intravenous diuretics was 2.95 [IQR 1.75 to 5.78] hours. Patients in the earliest quartile of time to diuretic therapy (< 1.75 hours) were the oldest, had highest percentage of white race, and had the highest BUN on admission (Supplemental Table 1). Those with the latest quartile of time to diuretic therapy were most likely to be male and have rales. The dose of diuretic was not associated with the timing of the first dose. Although a numerical stepwise increase in time to diuretics was observed with BNP decrease at day 1 to discharge, no BNP decrease at discharge was also associated with shorter time to diuretics ( Figure 1A ). Receiving diuretics within 1 hour of the presentation was not correlated with the time to BNP decrease either (12% in patients with BNP decrease at day1, 11% in day 2, 8% in day 3, 13% in BNP decrease at discharge and 13% in no BNP decrease, p = 0.822). Urine output and changes in body weight were not correlated with time to diuretics ( Figure 1B ). Although there was some evidence of those with an earlier BNP decrease having an earlier and greater body weight decrease, this was not statistically significant ( Figure 1C ).
No relationship was observed regarding time to diuretics and change in biomarkers ( Figure 2A ). Patients with later time to diuretics had higher relative ratios of hscTnI, uNGAL, and sNGAL at discharge, though these findings were not statistically significant. In contrast, earlier BNP decrease was associated with earlier and greater decreases in hscTnI and uNGAL ( Figure 2B ). Creatinine increased more in earlier BNP decrease groups, and a similar finding was observed with sNGAL and Gal3.
In-hospital mortality was 3.3% (27 patients). Patients in earlier time to diuretic therapy had a numerically higher in-hospital mortality, but this was not statistically significance ( Figure 3A ). Those without BNP decrease during hospitalization had the highest in-hospital mortality, but the timing of BNP decrease at day 1, 2 and 3 was not correlated with mortality Figure 3B . The 1-year composite end point was observed in 276 patients (34%), with 147 patients (18%) dying and 163 patients (20%) hospitalized because of HF. The earliest time to diuretics was associated with the highest risk of the composite end point and mortality in univariate analysis. These relationships were not statistically significant after adjustment for confounders. ( Figure 4A and Table 2 ). There was a stepwise increase in risk for the composite end point and mortality in patients with later timing of decrease or no decrease of BNP compared with BNP decrease at day 1 ( Figure 4B and Table 3 ). After adjustment for confounders, only no BNP decrease at discharge remained statistically significant for mortality but not for the composite end point.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
