Heart failure affects an estimated 56–64 million people worldwide and remains one of the few cardiovascular syndromes whose prevalence continues to rise, driven by population ageing, improved survival after myocardial infarction and the growing burden of hypertension and diabetes [1]. Acute decompensated heart failure (ADHF) — the abrupt or gradual worsening of symptoms and signs that compels hospital admission — is the single most common reason for hospitalisation in adults older than 65 years in high-income countries and accounts for a disproportionate share of cardiovascular health expenditure [2]. Hospitalisation is not simply an episode of care but a prognostic watershed: patients who survive to discharge carry a post-discharge mortality several times higher than that of stable ambulatory heart failure, and the index admission itself remains lethal for a substantial minority [3].

Contemporary guidelines from the European Society of Cardiology and the American College of Cardiology/American Heart Association recommend early risk assessment in every patient hospitalised with acute heart failure, so that monitoring intensity, the timing of vasoactive therapy and referral for advanced therapies can be matched to the probability of adverse outcome [4,5]. The evidence base supporting that recommendation comes largely from a small number of large registries. The Acute Decompensated Heart Failure National Registry (ADHERE) identified admission blood urea nitrogen, systolic blood pressure and serum creatinine as the dominant determinants of in-hospital death in a classification and regression tree analysis of more than 65,000 hospitalisations, with mortality varying from 2.1% to 21.9% across the resulting strata [10]. The Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) extended this to a multivariable model in which age, heart rate, systolic blood pressure, serum sodium, creatinine and the presence of left ventricular systolic dysfunction independently predicted death [11]. The Get With The Guidelines–Heart Failure risk score and the EFFECT model have since provided further validated instruments [12,13].

These models were derived in populations that differ substantially from those seen in much of the world. Patients in the ADHERE and OPTIMIZE-HF cohorts had a mean age above 72 years, approximately half had preserved ejection fraction, and nearly all had access to specialist cardiology care from the moment of presentation [10,11]. By contrast, the International Congestive Heart Failure (INTER-CHF) study demonstrated that heart failure mortality varies more than threefold across world regions even after adjustment for patient characteristics, with the highest rates in Africa and South Asia [7]. Registries from South Asia describe patients a decade or more younger, with a higher prevalence of rheumatic and dilated cardiomyopathic aetiology, later presentation, lower rates of guideline-directed medical therapy before admission, and in-hospital mortality that is several times the North American figure [8]. Systematic reviews of heart failure care in low- and middle-income countries confirm both the excess mortality and the scarcity of prospectively collected outcome data [9]. European registry data, in turn, show that in-hospital mortality is strongly determined by clinical phenotype at presentation, ranging from under 3% in patients admitted with warm–wet congestion to more than 25% in cardiogenic shock [6].

Whether the predictors established in Western registries perform comparably in such populations, and whether variables reflecting the early in-hospital course add prognostic information to admission data alone, has not been adequately addressed. We therefore conducted a prospective observational study of consecutive patients admitted with ADHF to a tertiary care teaching hospital, with the objective of identifying independent clinical, biochemical and in-hospital course predictors of all-cause in-hospital mortality, and of assembling them into a bedside risk stratification tool suitable for use where echocardiography and biomarker testing may not be immediately available.

Study design and setting

This was a prospective, single-centre, observational cohort study conducted in the Department of Medicine and Cardiology of a tertiary care teaching hospital over 24 months, from January 2024 to December 2025. The hospital serves as a referral centre for a predominantly semi-urban and rural catchment population and admits patients both directly through the emergency department and by transfer from peripheral facilities. The study is reported in accordance with the STROBE statement for observational research [17].

Participants

Consecutive patients aged 18 years or older admitted with a primary diagnosis of acute decompensated heart failure were screened for eligibility. ADHF was defined as the new onset or acute worsening of symptoms and signs of heart failure requiring urgent hospitalisation and intravenous diuretic, vasodilator or inotropic therapy, diagnosed using the modified Framingham criteria (two major, or one major and two minor criteria) [14], and supported in all cases by transthoracic echocardiography and by measurement of N-terminal pro–B-type natriuretic peptide (NT-proBNP), interpreted against the age-stratified rule-in thresholds recommended by current guidelines [4].

Patients were excluded if the dominant clinical problem was an acute coronary syndrome requiring immediate revascularisation, if dyspnoea was attributable primarily to a non-cardiac cause such as pneumonia, exacerbation of chronic obstructive pulmonary disease or end-stage renal disease on maintenance dialysis, if they had isolated right heart failure due to primary pulmonary disease or chronic thromboembolic pulmonary hypertension, if they were pregnant, if they had an established terminal non-cardiac illness with an anticipated survival of less than three months, or if informed consent was declined. Patients transferred out of the hospital before an in-hospital outcome could be ascertained were also excluded.

Data collection and definitions

All data were collected by the investigators on a pre-tested structured proforma at the bedside, within 24 hours of admission, and were verified against the hospital record before entry into the study database. Recorded variables comprised demographic details; the duration of known heart failure and whether the presentation was de novo or a decompensation of chronic heart failure; comorbidities including hypertension, diabetes mellitus, chronic kidney disease, coronary artery disease, atrial fibrillation, chronic obstructive pulmonary disease and prior stroke; the presumed aetiology of heart failure; and the identified precipitant of decompensation.

Vital signs were taken as the first set recorded after arrival. Hypotension was defined as an admission systolic blood pressure below 100 mmHg, and cardiogenic shock as systolic blood pressure below 90 mmHg for more than 30 minutes with clinical and biochemical evidence of hypoperfusion. Functional status was graded by the New York Heart Association classification at presentation. Transthoracic echocardiography was performed within 48 hours by a cardiologist blinded to the laboratory data; left ventricular ejection fraction (LVEF) was measured by the modified Simpson biplane method, and heart failure was categorised as reduced (LVEF < 40%), mildly reduced (40–49%) or preserved (≥ 50%) ejection fraction.

Venous blood was drawn at admission for complete blood count, serum electrolytes, renal and liver function, serum albumin, random blood glucose, high-sensitivity cardiac troponin I and NT-proBNP. Hyponatraemia was defined as a serum sodium below 135 mmol/L and anaemia by World Health Organization haemoglobin thresholds. The estimated glomerular filtration rate was calculated with the CKD-EPI equation [16]. Worsening renal function was defined, in keeping with KDIGO criteria for acute kidney injury, as a rise in serum creatinine of at least 0.3 mg/dL from the admission value at any point during hospitalisation [15]. High-sensitivity troponin was considered elevated above the sex-specific 99th percentile upper reference limit of the assay. In-hospital management, including diuretic strategy, vasodilator and inotropic use, non-invasive and invasive ventilation and renal replacement therapy, followed prevailing guideline recommendations and was at the discretion of the treating team [4,5,19].

Outcome

The primary outcome was all-cause mortality occurring between admission and discharge. Patients were followed daily until discharge or death. Length of stay and the use of organ support were recorded as secondary descriptive outcomes.

Sample size

Assuming an in-hospital mortality of approximately 12% in this setting, an absolute precision of 3.5% and 95% confidence, a minimum of 332 patients was required. To permit a multivariable model containing up to ten candidate predictors while retaining at least ten outcome events per variable, and allowing for incomplete data, a target of 420 patients was set.

Statistical analysis

Analyses were performed in SPSS version 27.0 (IBM Corp., Armonk, NY). Continuous variables are presented as mean ± standard deviation when normally distributed and as median with interquartile range otherwise, with normality assessed by the Shapiro–Wilk test; categorical variables are presented as counts and percentages. Survivors and non-survivors were compared using the independent-samples t test or Mann–Whitney U test for continuous variables and the chi-square or Fisher exact test for categorical variables.

Variables associated with in-hospital death at p < 0.10 on univariable logistic regression, together with variables of established clinical importance, were entered into a multivariable logistic regression model using backward stepwise elimination. Collinearity was assessed by variance inflation factors, and no retained variable had a factor above 2.5. Results are reported as adjusted odds ratios with 95% confidence intervals. Model discrimination was assessed by the area under the receiver operating characteristic curve and calibration by the Hosmer–Lemeshow goodness-of-fit test. A simple additive risk score was constructed by counting the number of independent predictors present in each patient. Two-sided p values below 0.05 were considered statistically significant. Missing data were below 3% for all variables analysed and were handled by complete-case analysis.

Ethics

The study protocol was approved by the Institutional Ethics Committee before enrolment began, and written informed consent was obtained from every participant or, where the patient was unable to consent, from a legally authorised representative. The study was conducted in accordance with the Declaration of Helsinki [18]. No identifying information is reported.

Cohort characteristics

Of 512 patients screened during the study period, 92 were excluded — 31 with a dominant acute coronary syndrome requiring immediate revascularisation, 26 in whom dyspnoea was ultimately attributed to a non-cardiac cause, 14 on maintenance haemodialysis, 11 with isolated right heart failure of pulmonary origin, 6 transferred out before an outcome could be ascertained and 4 who declined consent. The final cohort comprised 420 patients.

Mean age was 63.2 ± 13.0 years and 264 patients (62.9%) were men. Heart failure was a de novo presentation in 164 patients (39.0%). Ischaemic heart disease was the commonest aetiology (49.5%), followed by hypertensive (17.4%), valvular (13.6%) and dilated cardiomyopathic (13.6%) disease. Mean LVEF was 33.6 ± 12.5%, and 269 patients (64.0%) had heart failure with reduced ejection fraction. Infection was the most frequently identified precipitant of decompensation (26.4%), followed by non-adherence to medication or dietary salt restriction (20.7%).

Fifty-five patients died before discharge, an in-hospital mortality of 13.1% (95% CI 10.2–16.7). Median length of stay was 6 days (IQR 4–9) overall and was shorter among those who died, at 4 days (IQR 2–7), reflecting a concentration of deaths in the first 72 hours: 32 of 55 deaths (58.2%) occurred within three days of admission.

Baseline characteristics by outcome

Non-survivors were older, more often diabetic and more often had pre-existing chronic kidney disease, and presented in a more advanced functional class with markedly lower systolic blood pressure (Table 1). Sex, the proportion with de novo heart failure and the distribution of aetiology did not differ significantly between the groups.

Table 1. Baseline demographic and clinical characteristics by in-hospital outcome (n = 420).

SD, standard deviation; COPD, chronic obstructive pulmonary disease; BP, blood pressure; NYHA, New York Heart Association; LVEF, left ventricular ejection fraction.

Laboratory findings and in-hospital course

The biochemical separation between survivors and non-survivors was wider than the clinical one (Table 2). Mean blood urea nitrogen in those who died was 46.7 ± 21.3 mg/dL against 28.4 ± 14.6 mg/dL in survivors, and more than half of non-survivors met the ADHERE threshold of 43 mg/dL. Hyponatraemia was twice as common among those who died, and median NT-proBNP was more than twofold higher. Worsening renal function developed in 56.4% of non-survivors compared with 19.7% of survivors, and 63.6% of those who died required intravenous inotropic or vasopressor support at some point during admission.

Table 2. Admission laboratory values and in-hospital course by outcome (n = 420).

IQR, interquartile range; eGFR, estimated glomerular filtration rate; hs, high-sensitivity; NT-proBNP, N-terminal pro–B-type natriuretic peptide; WHO, World Health Organization.

Univariable analysis

Eighteen candidate variables were associated with in-hospital death at p < 0.10 on univariable logistic regression (Table 3). The strongest unadjusted associations were with markers of haemodynamic failure and its consequences — invasive mechanical ventilation, inotrope requirement, cardiogenic shock and worsening renal function — followed by azotaemia and hyponatraemia. Ejection fraction below 40% and anaemia reached the pre-specified screening threshold but were not significant at the conventional level.

Table 3. Univariable logistic regression for in-hospital mortality.

OR, odds ratio; CI, confidence interval. Variables with p < 0.10 are shown.

Multivariable analysis

Cardiogenic shock at admission and invasive mechanical ventilation were excluded a priori from the multivariable model: both are better understood as manifestations of the terminal event than as antecedent predictors, and in this cohort 19 of the 21 intubations occurred within 24 hours of death. The remaining variables were entered into backward stepwise logistic regression. Six remained independently associated with in-hospital death (Figure 1). Variables that did not survive elimination included age ≥ 65 years (aOR 1.98, 95% CI 0.94–4.17; p = 0.072), infection as precipitant (1.87, 0.89–3.93; p = 0.098), serum albumin < 3.4 g/dL (1.61, 0.76–3.42; p = 0.21), chronic kidney disease (1.52, 0.70–3.30; p = 0.29) and LVEF < 40% (1.34, 0.62–2.90; p = 0.46).

The final model discriminated well, with an area under the receiver operating characteristic curve of 0.86 (95% CI 0.81–0.91), and showed acceptable calibration (Hosmer–Lemeshow χ² = 6.33, 8 df, p = 0.61). The Nagelkerke R² was 0.41.

Figure 1. Adjusted odds ratios (95% confidence intervals) for in-hospital death from the final multivariable logistic regression model, 420 patients and 55 deaths. The dashed line marks an odds ratio of 1.

All six retained variables carried lower confidence bounds above unity, and the two renal variables — admission azotaemia and in-hospital worsening of renal function — entered the model independently of one another, indicating that a rising creatinine during treatment adds prognostic information to the admission value.

Additive risk stratification

Counting the six independent predictors present in each patient produced a monotonic gradient of risk (Figure 2). Patients with none of the six — 40.5% of the cohort — had a mortality of 1.8%, while the 5.0% of patients with four or more had a mortality of 61.9%, a relative risk of 35.1. The simple count discriminated almost as well as the full regression model (AUC 0.83, 95% CI 0.77–0.89 versus 0.86, 95% CI 0.81–0.91; p for difference 0.21).

Figure 2. In-hospital mortality by the number of the six independent predictors present. Deaths / patients are shown beneath each bar; the dashed line marks the whole-cohort mortality of 13.1%.

In this prospective cohort of 420 consecutive patients admitted with acute decompensated heart failure, one in eight died before discharge. Six variables — admission hypotension, azotaemia, hyponatraemia, elevated cardiac troponin, worsening renal function during admission and the need for inotropic support — were independently associated with death, and a simple count of these features separated patients whose risk of dying differed more than thirty-fold. Four of the six are available within an hour of presentation from a blood pressure cuff and a basic biochemistry panel.

Mortality in context

The in-hospital mortality of 13.1% observed here is three to four times that reported by the large North American registries, where in-hospital death occurred in 3.8–4.2% of admissions [10,11]. It is closer to, though still above, the 6.7% recorded in the ESC Heart Failure Long-Term Registry [6] and the 10.7% of the ALARM-HF global survey, which deliberately enrolled a more severely ill population [36]. Two observations help explain the difference. First, our patients were a decade younger than those in ADHERE and OPTIMIZE-HF yet presented in a far more advanced state: nearly half were in NYHA class IV and almost one in twenty was in cardiogenic shock at the door. Second, the INTER-CHF study has shown that heart failure mortality is substantially higher in South Asia and Africa than in Western populations even after adjusting for age, ejection fraction and comorbidity, implicating late presentation, limited access to guideline-directed therapy and the absence of organised post-discharge pathways rather than differences in disease biology alone [7]. The Trivandrum Heart Failure Registry, the best-characterised Indian cohort, reported a mean age of 61 years and in-hospital mortality of 8.5%, figures that bracket our own [8]. Our higher rate is plausibly explained by the referral nature of the centre, which receives transfers from peripheral facilities after initial therapy has failed.

Haemodynamic predictors

Admission systolic blood pressure was the single most powerful bedside variable in our model, as it has been in every major analysis of this syndrome. In the OPTIMIZE-HF registry, in-hospital mortality fell progressively from 7.2% in patients presenting with systolic pressure below 120 mmHg to 1.8% in those above 161 mmHg [20]. Low systolic pressure identifies the patient in whom cardiac output, rather than congestion alone, has become the limiting problem, and it carries the additional practical consequence of precluding the vasodilator and up-titrated neurohormonal therapy that benefits the hypertensive decompensator [19]. The strength of the association with inotrope requirement — the highest adjusted odds ratio in our model — is consistent with the ADHERE analysis of vasoactive therapy, in which patients treated with inotropes had substantially higher in-hospital mortality than those receiving vasodilators, a difference that persisted after risk adjustment and propensity matching [35]. We interpret inotrope use in our cohort primarily as a marker of low-output physiology rather than as evidence of harm from the drugs themselves; the design cannot separate the two.

Renal and electrolyte predictors

The most striking feature of our model is that two renal variables entered independently. Blood urea nitrogen at or above 43 mg/dL — the identical threshold identified by recursive partitioning in ADHERE [10] — remained significant after adjustment, and worsening renal function during admission added further risk. Urea is not simply a marker of glomerular filtration; it is reabsorbed under the influence of vasopressin and sympathetic activation, so an elevated level reflects the neurohormonal intensity of the decompensation as much as the kidney itself [21]. Worsening renal function during treatment for acute heart failure has been associated with higher mortality in a meta-analysis of more than one million patients, in which both baseline renal impairment and in-hospital deterioration carried independent prognostic weight [22], and in the OPTIME-CHF cohort, where change in renal function during hospitalisation predicted post-discharge survival independently of the admission value [25]. Our finding that both variables survive mutual adjustment supports the view that the cardiorenal interaction in acute heart failure is dynamic rather than a fixed baseline characteristic [24]. It should be noted, however, that not all creatinine rises are equivalent: a modest increase accompanying effective decongestion carries a different prognosis from one occurring alongside persistent congestion and hypoperfusion [23].

Hyponatraemia was present in nearly a third of the cohort and in more than half of those who died. The OPTIMIZE-HF analysis of admission sodium found a 19.5% increase in in-hospital mortality for every 3 mmol/L decrement below normal, independent of the other variables in the model [26], and an individual patient data meta-analysis across the MAGGIC collaboration confirmed that the relationship holds in both reduced and preserved ejection fraction [27]. Hyponatraemia in this setting is a surrogate for non-osmotic vasopressin release driven by perceived arterial underfilling, which places it on the same pathophysiological axis as hypotension and azotaemia.

Myocardial injury and natriuretic peptides

Elevated high-sensitivity troponin remained independently predictive after adjustment. In the ADHERE analysis of more than 67,000 patients with measurable troponin, a positive result was associated with lower systolic pressure, lower ejection fraction and an in-hospital mortality of 8.0% versus 2.7% in troponin-negative patients, with an adjusted odds ratio of 2.55 — almost exactly the estimate we obtained [28]. Troponin release in heart failure without epicardial occlusion reflects subendocardial ischaemia from raised wall stress, neurohormonal toxicity and microvascular dysfunction, and it marks ongoing myocyte loss rather than a discrete infarction [29]. NT-proBNP above 10,000 pg/mL was strongly predictive on univariable analysis but did not survive multivariable adjustment, which is consistent with the ICON pooled analysis showing that the prognostic signal from natriuretic peptides is partly shared with renal function and clinical severity [30]. For centres where natriuretic peptide assays are unavailable or costly, this is a reassuring result: the model retains its discrimination without them.

Precipitants

Infection was the commonest identified precipitant and was significantly more frequent among non-survivors, but it did not retain independence after adjustment. The intercontinental GREAT registry similarly found that infection as a trigger was associated with the worst 90-day outcome among all precipitants, while acute coronary syndrome and uncontrolled hypertension carried intermediate and favourable prognoses respectively [32]. Analyses from Get With The Guidelines reached comparable conclusions across the ejection fraction spectrum [31]. Our result suggests that the adverse effect of infection operates largely through the haemodynamic and renal variables already in the model rather than through an independent pathway — but the study was not powered to resolve this, and the point estimate of 1.87 is clinically non-trivial.

Variables that did not predict

Neither ejection fraction nor aetiology independently predicted in-hospital death. This accords with registry data from Japan and Europe showing that short-term outcome in acute heart failure is determined far more by presenting haemodynamics and end-organ function than by the chronic substrate [6,37]. Hypoalbuminaemia and anaemia, both strongly associated with death on univariable testing and both established predictors of longer-term mortality in chronic heart failure [33,34], likewise lost significance after adjustment, suggesting they operate as markers of chronic illness burden whose influence is expressed over months rather than days.

Clinical implications

The practical value of these findings lies in their availability. A patient who arrives with a systolic pressure below 100 mmHg, a urea above 43 mg/dL and a sodium below 135 mmol/L can be identified as high risk before echocardiography has been performed or a natriuretic peptide result has returned. In our cohort such patients had a mortality above 40%, which in most health systems would justify admission to a high-dependency area, early arterial and central venous access, consideration of invasive monitoring, and an explicit conversation with the family about the severity of the illness. Conversely, the 40% of patients with none of the six features had a mortality below 2% and could reasonably be managed on a general ward — a judgement of equal practical importance where critical care beds are scarce. The gradient demonstrated in Figure 2 closely parallels that of the ADHERE risk tree [10] and the Get With The Guidelines score [12], suggesting that the underlying physiology travels well across populations even when the absolute event rates do not.

Strengths and limitations

The principal strengths of this study are its prospective design with consecutive enrolment, complete in-hospital follow-up, the use of standardised definitions applied at the bedside, and the inclusion of in-hospital course variables alongside admission data.

Several limitations temper the conclusions. First, this was a single-centre study at a tertiary referral hospital, and the case mix is likely to be more severe than that of a community hospital; the absolute mortality and the prevalence of each predictor should not be generalised without local validation. Second, the model was derived and tested in the same cohort; despite acceptable calibration and good discrimination, optimism bias is likely and external validation in an independent population is required. Third, two of the six predictors — worsening renal function and inotrope requirement — are not available at admission and are themselves influenced by treatment decisions, so reverse causation cannot be excluded and the model should be understood as a dynamic rather than a purely admission-based instrument. We excluded mechanical ventilation and admission cardiogenic shock from the model for this reason, but the boundary between a predictor and a manifestation of dying is not a sharp one. Fourth, with 55 events the model is at the limit of what the data can support at ten events per variable, and the confidence intervals around the adjusted odds ratios are correspondingly wide. Fifth, we did not record pre-admission guideline-directed medical therapy in sufficient detail to assess its contribution, nor did we collect post-discharge outcomes, so the clinical importance of these predictors beyond the index admission is unknown. Finally, causes of death were assigned clinically without autopsy confirmation.

Among patients hospitalised with acute decompensated heart failure at a tertiary centre, in-hospital mortality was 13.1%, substantially higher than that reported by North American and European registries in an appreciably younger population. Six variables were independently associated with death: admission systolic blood pressure below 100 mmHg, blood urea nitrogen at or above 43 mg/dL, serum sodium below 135 mmol/L, elevated high-sensitivity cardiac troponin, worsening renal function during admission, and the need for intravenous inotropic support. Ejection fraction, aetiology, anaemia and hypoalbuminaemia, though associated with death on unadjusted analysis, carried no independent weight over the short horizon of the index admission.

The practical message is that short-term prognosis in this syndrome is governed by perfusion and end-organ function rather than by the chronic cardiac substrate, and that the variables carrying this information are inexpensive and almost universally available. A simple count of the six predictors stratified mortality from 1.8% to 61.9% and performed nearly as well as the full regression model, making it suitable for bedside use in settings where neither natriuretic peptide assays nor immediate echocardiography can be relied upon. External validation in independent cohorts, and prospective testing of whether early escalation of care in high-scoring patients alters outcome, are the necessary next steps.

Declarations

Ethics approval and consent to participate. Approved by the Institutional Ethics Committee; written informed consent was obtained from all participants.

Consent for publication. Not applicable.

Availability of data. The dataset analysed is available from the corresponding author on reasonable request.

Competing interests. The authors declare no competing interests.

Funding. This study received no specific grant from any funding agency.

1. Savarese G, Becher PM, Lund LH, et al. Global burden of heart failure: a comprehensive and updated review of
2. Cardiovasc Res. 2023;118(17):3272–3287. doi:10.1093/cvr/cvac013
3. Ambrosy AP, Fonarow GC, Butler J, et al. The global health and economic burden of hospitalizations for heart failure: lessons learned from hospitalized heart failure registries. J Am Coll Cardiol. 2014;63(12):1123–1133. doi:10.1016/j.jacc.2013.11.053
4. Arrigo M, Jessup M, Mullens W, et al. Acute heart failure. Nat Rev Dis Primers. 2020;6(1):16. doi:10.1038/s41572-020-0151-7
5. McDonagh TA, Metra M, Adamo M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42(36):3599–3726. doi:10.1093/eurheartj/ehab368
6. Heidenreich PA, Bozkurt B, Aguilar D, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145(18):e895–e1032. doi:10.1161/CIR.0000000000001063
7. Chioncel O, Mebazaa A, Harjola VP, et al. Clinical phenotypes and outcome of patients hospitalized for acute heart failure: the ESC Heart Failure Long-Term Registry. Eur J Heart Fail. 2017;19(10):1242–1254. doi:10.1002/ejhf.890
8. Dokainish H, Teo K, Zhu J, et al. Global mortality variations in patients with heart failure: results from the International Congestive Heart Failure (INTER-CHF) prospective cohort study. Lancet Glob Health. 2017;5(7):e665–e672. doi:10.1016/S2214-109X(17)30196-1
9. Harikrishnan S, Sanjay G, Anees T, et al. Clinical presentation, management, in-hospital and 90-day outcomes of heart failure patients in Trivandrum, Kerala, India: the Trivandrum Heart Failure Registry. Eur J Heart Fail. 2015;17(8):794–800. doi:10.1002/ejhf.283
10. Callender T, Woodward M, Roth G, et al. Heart failure care in low- and middle-income countries: a systematic review and meta-analysis. PLoS Med. 2014;11(8):e1001699. doi:10.1371/journal.pmed.1001699
11. Fonarow GC, Adams KF Jr, Abraham WT, et al. Risk stratification for in-hospital mortality in acutely decompensated heart failure: classification and regression tree analysis. JAMA. 2005;293(5):572–580. doi:10.1001/jama.293.5.572
12. Abraham WT, Fonarow GC, Albert NM, et al. Predictors of in-hospital mortality in patients hospitalized for heart failure: insights from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF). J Am Coll Cardiol. 2008;52(5):347–356. doi:10.1016/j.jacc.2008.04.028
13. Peterson PN, Rumsfeld JS, Liang L, et al. A validated risk score for in-hospital mortality in patients with heart failure from the American Heart Association Get With the Guidelines program. Circ Cardiovasc Qual Outcomes. 2010;3(1):25–32. doi:10.1161/CIRCOUTCOMES.109.854877
14. Lee DS, Austin PC, Rouleau JL, et al. Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model. JAMA. 2003;290(19):2581–2587. doi:10.1001/jama.290.19.2581
15. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971;285(26):1441–1446. doi:10.1056/NEJM197112232852601
16. Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179–c184. doi:10.1159/000339789
17. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. doi:10.7326/0003-4819-150-9-200905050-00006
18. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453–1457. doi:10.1016/S0140-6736(07)61602-X
19. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191–2194. doi:10.1001/jama.2013.281053
20. Mebazaa A, Yilmaz MB, Levy P, et al. Recommendations on pre-hospital and early hospital management of acute heart failure: a consensus paper from the Heart Failure Association of the European Society of Cardiology, the European Society of Emergency Medicine and the Society of Academic Emergency Medicine. Eur J Heart Fail. 2015;17(6):544–558. doi:10.1002/ejhf.289
21. Gheorghiade M, Abraham WT, Albert NM, et al. Systolic blood pressure at admission, clinical characteristics, and outcomes in patients hospitalized with acute heart failure. JAMA. 2006;296(18):2217–2226. doi:10.1001/jama.296.18.2217
22. Harjola VP, Mullens W, Banaszewski M, et al. Organ dysfunction, injury and failure in acute heart failure: from pathophysiology to diagnosis and management. Eur J Heart Fail. 2017;19(7):821–836. doi:10.1002/ejhf.872
23. Damman K, Valente MAE, Voors AA, et al. Renal impairment, worsening renal function, and outcome in patients with heart failure: an updated meta-analysis. Eur Heart J. 2014;35(7):455–469. doi:10.1093/eurheartj/eht386
24. Metra M, Nodari S, Parrinello G, et al. Worsening renal function in patients hospitalised for acute heart failure: clinical implications and prognostic significance. Eur J Heart Fail. 2008;10(2):188–195. doi:10.1016/j.ejheart.2008.01.011
25. Ronco C, Haapio M, House AA, Anavekar N, Bellomo R. Cardiorenal syndrome. J Am Coll Cardiol. 2008;52(19):1527–1539. doi:10.1016/j.jacc.2008.07.051
26. Klein L, Massie BM, Leimberger JD, et al. Admission or changes in renal function during hospitalization for worsening heart failure predict postdischarge survival: results from OPTIME-CHF. Circ Heart Fail. 2008;1(1):25–33. doi:10.1161/CIRCHEARTFAILURE.107.746933
27. Gheorghiade M, Abraham WT, Albert NM, et al. Relationship between admission serum sodium concentration and clinical outcomes in patients hospitalized for heart failure: an analysis from the OPTIMIZE-HF registry. Eur Heart J. 2007;28(8):980–988. doi:10.1093/eurheartj/ehl542
28. Rusinaru D, Tribouilloy C, Berry C, et al. Relationship of serum sodium concentration to mortality in a wide spectrum of heart failure patients with preserved and with reduced ejection fraction: an individual patient data meta-analysis (MAGGIC). Eur J Heart Fail. 2012;14(10):1139–1146. doi:10.1093/eurjhf/hfs099
29. Peacock WF 4th, De Marco T, Fonarow GC, et al. Cardiac troponin and outcome in acute heart failure. N Engl J Med. 2008;358(20):2117–2126. doi:10.1056/NEJMoa0706824
30. Kociol RD, Pang PS, Gheorghiade M, et al. Troponin elevation in heart failure: prevalence, mechanisms, and clinical implications. J Am Coll Cardiol. 2010;56(14):1071–1078. doi:10.1016/j.jacc.2010.06.016
31. Januzzi JL Jr, van Kimmenade R, Lainchbury J, et al. NT-proBNP testing for diagnosis and short-term prognosis in acute destabilized heart failure: an international pooled analysis of 1256 patients. Eur Heart J. 2006;27(3):330–337. doi:10.1093/eurheartj/ehi631
32. Kapoor JR, Kapoor R, Ju C, et al. Precipitating clinical factors, heart failure characterization, and outcomes in patients hospitalized with heart failure with reduced, borderline, and preserved ejection fraction. JACC Heart Fail. 2016;4(6):464–472. doi:10.1016/j.jchf.2016.02.017
33. Arrigo M, Gayat E, Parenica J, et al. Precipitating factors and 90-day outcome of acute heart failure: a report from the intercontinental GREAT registry. Eur J Heart Fail. 2017;19(2):201–208. doi:10.1002/ejhf.682
34. Horwich TB, Kalantar-Zadeh K, MacLellan RW, Fonarow GC. Albumin levels predict survival in patients with systolic heart failure. Am Heart J. 2008;155(5):883–889. doi:10.1016/j.ahj.2007.11.043
35. Groenveld HF, Januzzi JL, Damman K, et al. Anemia and mortality in heart failure patients: a systematic review and meta-analysis. J Am Coll Cardiol. 2008;52(10):818–827. doi:10.1016/j.jacc.2008.04.061
36. Abraham WT, Adams KF, Fonarow GC, et al. In-hospital mortality in patients with acute decompensated heart failure requiring intravenous vasoactive medications: an analysis from the Acute Decompensated Heart Failure National Registry (ADHERE). J Am Coll Cardiol. 2005;46(1):57–64. doi:10.1016/j.jacc.2005.03.051
37. Follath F, Yilmaz MB, Delgado JF, et al. Clinical presentation, management and outcomes in the Acute Heart Failure Global Survey of Standard Treatment (ALARM-HF). Intensive Care Med. 2011;37(4):619–626. doi:10.1007/s00134-010-2113-0
38. Sato N, Kajimoto K, Keida T, et al. Clinical features and outcome in hospitalized heart failure in Japan (from the ATTEND Registry). Circ J. 2013;77(4):944–951. doi:10.1253/circj.cj-13-0187