Early drop in systolic blood pressure is associated with poor diuretic response and prognosis in patients with acute heart failure 

Abstract

Aims

Although an excessive drop in systolic blood pressure (SBP) during acute heart failure (AHF) negatively impacts prognosis, the association between changes in SBP and the diuretic response (DR) is unclear. We aimed to clarify the association between an early drop in SBP and DR/prognosis in patients with AHF.

Methods and results

This was a sub-analysis of the REALITY-AHF study, which registered patients with AHF admitted through emergency departments (EDs). An early SBP drop was defined as the difference between baseline SBP and the lowest value during the first 48 h of hospitalization. DR was defined as the urine output achieved per 40 mg of intravenous furosemide administered. SBP was measured on admission, at 90 min, and 6, 24, and 48 h after admission. Patients were divided into four groups according to their median SBP drop and DR: greater SBP drop/poor DR (n = 322), smaller SBP drop/poor DR (n = 409), greater SBP drop/good DR (n = 419), and smaller SBP drop/good DR (n = 314). The study included 1,464 patients. A greater SBP drop/poor DR was associated with higher baseline SBP and vasodilator use. Multivariable linear regression analysis showed that a greater drop in SBP was associated with poorer DR following adjustment for potential covariates. Cox proportional hazards analysis demonstrated that a greater SBP drop/poor DR was independently associated with 1-year mortality. Both SBP and DR changes were independently associated with prognosis.

Conclusion

An early drop in SBP during the first 48 h of hospitalization was associated with poor DR and 1-year mortality in patients with AHF.

Clinical Trial Registration

URL: http://www.umin.ac.jp/ctr/Unique identifier: UMIN000014105

Graphical Abstract

Open in new tabDownload slide

Introduction

Although acute heart failure (AHF) is a critical condition requiring prompt treatment,^1,2 recent studies have suggested that an excessive early drop in systolic blood pressure (SBP), which can occasionally be attributed to treatment with vasodilators, is associated with potential adverse events. ^3–5 It has been proposed that this association could be attributable to the subsequent impairment of renal function, since several studies have demonstrated an association between an SBP drop and an increase in creatinine levels in patients with AHF. ^3,4 However, our recent study based on the RELAX-AHF-2 cohort, which included patients whose SBP remained relatively unchanged, demonstrated that the relationship between an early SBP drop and poor prognosis was not significantly associated with renal dysfunction, defined as an increase in creatinine levels. ⁵ Therefore, the mechanism underlying the association between an early drop in SBP and poor prognosis is yet to be clarified.

Poor diuretic response (DR) has also been reported to be associated with a poor prognosis in patients with AHF.^6–8 Although low baseline blood pressure has been shown to be associated with poor DR in previous studies, whether the changes in SBP are associated with DR independent of the baseline SBP has yet to be clarified. Considering that the autoregulation of renal blood flow in response to changes in SBP is blunted in patients with heart failure,⁹ the SBP drop and subsequent reduction in renal blood flow can lead to a poor DR, which could be the mechanism responsible for the association between an early SBP drop and poor prognosis. In this study, we aimed to address two clinical questions: (i) Does an SBP drop impact DR? and (ii) If so, is the SBP drop associated with a poor prognosis, independent of its association with the DR?

Methods

Study design and patient population

This was a post hoc analysis of the REALITY-AHF (Registry Focused on Very Early Presentation and Treatment in Emergency Department of Acute Heart Failure) registry. The detailed study protocol and the main results have been reported previously.¹⁰ Briefly, the REALITY-AHF study was a prospective, multicentre, observational cohort study that included 1 682 consecutive hospitalized patients with AHF who were admitted to emergency departments (EDs). The study patients were enrolled from August 2014 to December 2015 in 20 hospitals; 9 were university hospitals and 11 were non-university teaching hospitals in Japan. AHF was diagnosed by an attending physician at each site based on the Framingham criteria. The exclusion criteria were as follows: (i) treatment with an intravenous drug before ED admission; (ii) a history of heart transplantation; (iii) chronic peritoneal dialysis or haemodialysis; (iv) acute myocarditis; and (v) acute coronary syndrome requiring emergent or urgent revascularization. Patients with missing data on B-type natriuretic peptide (BNP) or N-terminal pro-BNP (NT-proBNP) levels were excluded, as were patients with a BNP level < 100 pg/mL or an NT-proBNP level < 300 pg/mL at baseline. All patients were registered with an ED, and baseline data, including physical examination results, echocardiography findings, and laboratory data, were obtained from the ED.

The REALITY-AHF registry was conducted in accordance with the Declaration of Helsinki and the Japanese Ethical Guidelines for Medical and Health Research Involving Human Subjects. All study patients were notified of their participation in the study and were free to opt out of the study at any time. The study protocol was approved by the ethics committee of each participating hospital. Study information, including the aims, inclusion and exclusion criteria, and names of the participating hospitals, were reported in the publicly available University Hospital Information Network (UMIN-CTR, unique identifier: UMIN000014105) before the first patient was enrolled.

Definitions of early SBP drop, DR, and worsening renal function

Blood pressure measurements were obtained at baseline (upon ED arrival) and at 1.5, 6, 24, and 48 h after the patients’ arrival in the ED. An early SBP drop was defined as the difference between the baseline value and the lowest value recorded during the first 48 h of the study period. For instance, if the baseline SBP was 160 mmHg and the lowest SBP recorded during the 48 h period was 130 mmHg, the early SBP drop was 30 mmHg (not –30 mmHg).

DR was defined as the urine output achieved per 40 mg of intravenous furosemide (or equivalent doses) administered during the first 48 h. The converted oral furosemide dose was calculated as half the dose of intravenous furosemide.^11,12 The doses of the oral loop diuretics that were considered equivalent to 40 mg of intravenous furosemide were 10 mg for torsemide and 60 mg for azosemide. Worsening renal function (WRF) was defined as an increase in the serum creatinine concentration of ≥ 0.3 mg/dL during the first 48 h.¹³

To elucidate the effects of patient characteristics based on the early SBP drop and DR, patients were divided into the following four groups: greater SBP drop/poor DR (n = 322), smaller SBP drop/poor DR (n = 409), greater SBP drop/good DR (n = 419), and smaller SBP drop/good DR (n = 314). These classifications were based on the median values of the SBP drop and DR.

Outcomes

The primary endpoint was 1-year all-cause mortality, which was determined from the day of admission. Follow-up data were prospectively obtained for all patients with medical records and was confirmed by follow-up contact. For patients who were followed-up in different institutions from where they were registered, the follow-up data were collected from telephone interviews by the medical records departments of other medical facilities caring for the patients or from information provided by the family members.

Statistical analysis

Data are presented as the mean and standard deviation for the normally distributed variables and as the median and interquartile range (IQR) for non-normally distributed variables; these data were compared using either the Student’s t-test or Mann–Whitney U test. Categorical variables are expressed as numbers and percentages and were compared between groups using chi-square or Fisher’s exact tests.

Univariate and multivariate linear regression analyses were conducted to assess the association between the DR and SBP drop. Age, baseline SBP, prescription loop diuretics at baseline, and haemoglobin, blood urea nitrogen, creatinine, and alanine aminotransferase levels were included in the multivariable analysis, since these variables have already been identified as being independently associated with DR measured at 48 h in this cohort.¹⁴ To compare the survival rates between the groups, adjusted survival curves were constructed based on the Cox proportional hazards model. Variables that were already known to be prognostic factors in patients with heart failure were included in the adjusted multivariable Cox regression model; these included age, sex, history of heart failure, New York Heart Association functional class, atrial fibrillation, diabetes, coronary artery disease, left ventricular ejection fraction, SBP, heart rate, haemoglobin, serum sodium, and creatinine concentrations, the log-transformed BNP measurement at the time of admission, and prescription of loop diuretics, angiotensin-converting-enzyme inhibitor/angiotensin II receptor blockers, or beta-blockers. To clarify the impact of the early SBP drop and that of DR on mortality, the adjusted hazard ratio (HR) for 1-year all-cause mortality was evaluated using a linear regression model and a non-linear model with restricted cubic splines of 3, 4, and 5 knots for both the early SBP drop and DR. Similarly, the impact of an early SBP drop on DR was evaluated using restricted cubic spline analysis after adjustment for covariates.¹⁴ The goodness-of-fit was compared between models using an analysis of variance test, and the model with the best fit was chosen.

All data were analysed using R version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org). Statistical significance was defined as a two-sided P-value < 0.05.

Results

Of the 1 682 patients in the REALITY-AHF cohort, patients with missing SBP (n = 26) or DR (n = 216) data were excluded (some of these overlapped). In total, 1 464 patients (mean age, 77.8 ± 12.3 years; 55.5% male) were finally included in the analysis. The baseline SBP was 144 mmHg (IQR, 123–171 mmHg), and the median SBP drop and DR were 38 mmHg (IQR, 18–62 mmHg) and 2 365 mL/40 mg (IQR, 1 425–4 052 mL/40 mg) of intravenous furosemide administered, respectively. The baseline characteristics of the patients are described in Table 1. Overall, the patients in the SBP drop/poor DR group were older and had higher baseline SBP and serum creatinine levels. A total of 396 patients (23.6%) experienced acute onset (within the 6 h prior to index hospitalisation), 330 (83.3%) of whom had a baseline SBP ≥ 125 mmHg.

WRF occurred more often in patients in the greater SBP drop/poor DR group than in the other three groups (P < 0.001) (see Supplementary material online, Figure S1). As for the association between the DR and SBP drop, univariate linear regression analysis revealed that a greater SBP drop was positively associated with the DR (β, 0.09; t-value, 3.56; P < 0.001). However, multivariable linear regression analysis showed that a greater drop in SBP was associated with poorer DR following the adjustment for potential covariates, including baseline SBP (β = –0.13; t = –2.33; P =0.020) (Table 2).

The adjusted survival analysis demonstrated that the greater SBP drop/poor DR group had the highest risk of 1-year mortality compared with that of the smaller SBP drop/good DR group [HR: 3.40; 95% confidence interval (CI): 1.95–5.95, P < 0.001] (Figure 1). Having a smaller SBP drop/poor DR or a greater SBP drop/good DR was also independently associated with an increased risk of 1-year mortality following adjustment for the covariates (smaller SBP drop/poor DR group: HR: 2.18; 95% CI: 1.38–3.46, P < 0.001; greater SBP drop/good DR group: HR, 1.90; 95% CI, 1.06–3.40, P =0.030) (Table 3). Moreover, when the greater SBP drop/poor DR group was used as a reference, having a smaller SBP drop/poor DR was associated with a better prognosis (HR: 0.62; 95% CI: 0.42–0.93, P = 0.019). In addition, the multivariable Cox regression model, which included early SBP drop and DR together as continuous variables and other adjustment variables, demonstrated that both an early SBP drop (HR: 1.10 per 10 mmHg increase, 95% CI: 1.00–1.20, P = 0.038) and DR (HR: 0.99/100 mL/48 h/40 mg increase in intravenous furosemide administered, 95% CI: 0.98–0.99, P < 0.001) were associated with prognosis, independent of each other. The impact of the early SBP drop and DR on mortality is shown in Figure 2. After constructing the linear regression model and non-linear models with restricted cubic splines of 3, 4, and 5 knots, we found that the model with a restricted cubic spline of three knots was the best fit for 1-year mortality, and a greater SBP drop and poorer DR were both associated with a poor prognosis. The association between an early SBP drop and the DR is shown in Figure 3. The greater the early SBP drop, the poorer the DR.

Discussion

In the analysis of the REALITY-AHF registry data of patients with AHF, we found that a greater early drop in SBP was independently associated with a poorer DR. Moreover, early drops in SBP and DR were significantly associated with a poor prognosis, independent of each other, and patients with a greater SBP drop and poor DR had the highest incidence of 1-year mortality.

The prognostic impact of the SBP drop in patients with AHF

Although several studies have previously investigated the relationship between an SBP drop and prognosis in patients with AHF, the results of these studies were not completely consistent. A post hoc study of patients from the ESCAPE trial showed that an SBP drop from admission to discharge was not associated with 180-day mortality.¹⁵ Conversely, in the post hoc analysis of the Value of Endothelin Receptor Inhibition with Tezosentan in Acute heart failure Study (VERITAS), which was a randomized and placebo-controlled study evaluating the efficacy of tezosentan for patients with AHF whose SBP remained relatively unchanged, a greater SBP drop during the first 24 h was associated with increased 180-day mortality.³ Moreover, a post hoc analysis of the Relaxin for the Treatment of Acute Heart Failure 2 study (RELAX-AHF-2), which assessed the prognostic value of serelaxin for patients with AHF who had an SBP ≥ 125 mmHg, demonstrated that an early drop in SBP during the first 48 h was independently associated with worsening heart failure at 5 days and cardiovascular death at 180 days.⁵ Considering that the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial focused neither on acute phase treatment nor on patients with AHF, the results of the previous studies can be considered to be in line with our own findings. However, it should be noted that these studies were all randomized controlled trials, which mainly assessed the use of vasodilators and, therefore, needed to exclude patients with relatively low blood pressure. For instance, VERITAS excluded patients with an SBP ≤ 100 mmHg who were not treated with intravenous vasodilators and those with SBP ≤ 120 mmHg who were treated with vasodilators.¹⁶ The RELAX-AHF-2 study excluded patients with an SBP < 125 mmHg at screening. The results of these studies are only applicable to the group of patients with AHF whose baseline SBP values were relatively preserved. Considering that low blood pressure is strongly associated with poor outcomes in patients with AHF, the generalizability of the findings concerning the association between the SBP drop and adverse outcomes is limited. Thus, our study is the first to demonstrate an association between an early drop in SBP during the first 48 h and mortality, independent of baseline SBP in real-world patients with AHF. Moreover, REALITY-AHF is a registry focusing on the very acute phase of patients with AHF, and data on blood pressure were obtained from the time of ED arrival. Using the same database, Kitai et al. had reported an association between an early drop in SBP during the first 6 h and 1-year all-cause mortality in patients treated with intravenous vasodilators.¹¹ However, our study investigated the mechanism of DR without taking into account the use of vasodilators during the 48-h period. Therefore, the present study is the first to reconfirm this association using registry data to expand its generalizability to the daily clinical setting.

SBP drop, DR, and phenotypes of heart failure

It should be noted that the magnitude of the SBP drop and the DR are associated with the phenotype of AHF. In our study, the group exhibiting a greater SBP drop/poor DR had a higher baseline SBP and the highest proportions of patients with preserved left ventricular ejection fraction (≥50%) and pulmonary oedema, suggesting that the changes seen in this group might have been associated with the phenotype of hypertensive heart failure.¹⁷ Vasodilators have been shown to be used more frequently for patients with this phenotype,¹⁸ and, in general, there is a limited association between this phenotype and systemic congestion.^18,19 Thus, a greater SBP drop in those with a poor DR may have been influenced by the phenotype of hypertensive heart failure. However, physical examinations revealed no significant between-group differences in the presence of jugular venous distention or peripheral oedema in the present study, suggesting that the group exhibiting a greater SBP drop/poor DR was not the only one with an especially low level of systemic congestion. Additionally, previous studies have shown that the phenotype of hypertensive heart failure was not associated with a poorer prognosis compared with that of other phenotypes.^17,18 Therefore, we believe that the combination of an SBP drop and poor DR was not driven solely by AHF phenotypes.

The association between the SBP drop and DR, and prognosis

One of the novel findings of the present study was that an early drop in SBP was significantly and negatively associated with DR, even after adjustment for other factors, including baseline SBP. Although the pathophysiological mechanisms underlying this phenomenon have yet to be elucidated, there are some possible explanations for this association. For instance, as the renal autoregulation function is blunted in patients with heart failure, a blood pressure drop can directly reduce renal blood flow, resulting in a reduction in the glomerular filtration rate.⁹ Moreover, a reduction in the renal flow elevates levels of blood urea nitrogen and decreases the delivery of loop diuretics to the active sites.^12,20 Considering that all these factors have been shown to be associated with a poor response to diuretics,²¹ the association between an SBP drop and the subsequent reduction in renal flow and diminished DR may be biologically plausible. Indeed, in the present study, the combination of a greater SBP drop and poor DR was significantly associated with WRF.

However, our study also showed that an early SBP drop and DR were both independently associated with a poor prognosis, suggesting that mechanisms other than the impairment of DR could be involved in the association between the excessive SBP drop and poor prognosis. Indeed, the unfavourable impact of low blood pressure on prognosis has been reported in patients other than those with heart failure. For instance, several studies have reported that blood pressure that is too low is associated with reduced coronary blood flow, subclinical myocardial damage, and increased cardiovascular events.²² A previous study involving 10 001 patients with coronary artery disease showed that low blood pressure resulted in an increased risk of future cardiovascular events.²³ Another study that included 22 672 patients with stable coronary artery disease undergoing treatment for hypertension suggested that a decrease in blood pressure beyond a certain degree was related to an increased risk of cardiovascular death, all-cause mortality, myocardial infarction, and hospital admission for heart failure.²⁴ Although these studies did not focus on patients with heart failure, these factors may also support the significant association between an excessive SBP drop and poor outcomes in patients with heart failure. Nevertheless, further randomized controlled trials are required to confirm whether an excessive SBP drop leads to a poorer prognosis among patients with AHF.

Limitations

This study had several limitations. First, since this analysis was conducted only on Japanese cohorts, there is still room for consideration of other populations. Therefore, the generalizability of our outcomes may be limited since the characteristics of the patients or the medical care system may differ between countries. Second, in the present study, a number of patients were excluded due to acute coronary syndrome, acute myocarditis, or haemodialysis, even though such diseases are the main causes of heart failure in the real world. Finally, since there were numerous differences in baseline characteristics across the groups, all of the possible confounders could not be excluded.

Conclusions

An early drop in the SBP during the treatment of AHF was associated with the DR and all-cause mortality, independent of baseline SBP in a real-world setting. The results suggest that excessive SBP drops should be prevented, irrespective of the SBP in patients with AHF, although further studies are needed to determine the optimal means of managing blood pressure in the acute phase of AHF.

Supplementary material

Supplementary material is available at European Heart Journal : Acute Cardiovascular Care online.

Funding

The REALITY-AHF study was funded by The Cardiovascular Research Fund, Japan. This research was partially supported by AMED under Grant Number JP 22ek0109543. This work was also partially supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant number 22K16152).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes