Summary
Rapidly assessing outcome in patients with acute decompensated heart failure is important but prognostic factors may differ from those used routinely for stable chronic heart failure. Multiple plasma biomarkers, besides the classic natriuretic peptides, have recently emerged as potential prognosticators. Furthermore, prognostic scores that combine clinical and biochemical data may also be useful. However, compared with the scores used in chronic heart failure, scores for acute decompensated heart failure have not been validated. This article reviews potential biomarkers, with a special focus on biochemical biomarkers, and possible prognostic scores that could be used by the clinician when assessing outcome in patients with acute heart failure.
Résumé
L’évaluation précoce du pronostic chez les patients atteints d’insuffisance cardiaque aiguë décompensée est importante mais les facteurs pronostiques peuvent différer de ceux utilisés couramment pour une insuffisance cardiaque stable chronique. De nouveaux biomarqueurs plasmatiques sont apparus avoir une valeur pronostique ces dernières années, à côté des classiques peptides natriurétiques. En outre, des scores de risque, combinant les données cliniques et biochimiques, peuvent également être utilisés. Cependant, à la différence de ceux utilisés dans l’insuffisance cardiaque chronique, les scores de risque ont été moins validés dans l’insuffisance cardiaque aiguë décompensée. Cet article passe en revue les biomarqueurs pronostiques potentiels, avec un accent particulier sur les marqueurs biochimiques, et les scores de risque qui pourraient être utilisés plus souvent par le clinicien en charge de patients atteints d’insuffisance cardiaque aiguë pour prédire rapidement le pronostic.
Background
Patients with cardiovascular disease generally have a relatively poor prognosis. It is therefore critical for doctors to have simple markers to predict short- and medium-term outcome in order to optimize therapy and management. Such a prognostic evaluation has been successfully developed and applied in cardiology, particularly for coronary artery disease and myocardial infarction. Similar attempts have been made in chronic heart failure (CHF) but CHF is a much more heterogeneous syndrome than acute coronary syndromes. Indeed, the spectrum of CHF has changed over time, with a much higher prevalence of heart failure (HF) with preserved ejection fraction versus HF with systolic dysfunction.
Besides CHF, one can no longer ignore the increasing incidence of acute HF (AHF), which is a major cause of death and hospital readmission. AHF is either at the origin of CHF (de novo AHF) or results from the decompensation of CHF (acute decompensated HF [ADHF]). AHF is an even more heterogeneous condition as it is an acute syndrome that often affects the elderly, in whom co-morbidities are frequent. Consequently, AHF is a highly complex syndrome that has rarely been the focus of attention, until recently, as illustrated by the limited number of therapeutic trials conducted in this setting (SURVIVE, ASCEND, PROTECT, RELAX, ATOMIC-HF). However, it is becoming crucial to conduct studies in AHF, as its prognosis is quite poor, with 30-day and 6-month mortality rates of around 10% and 30%, respectively, not including the high rate of hospital readmission. Patient management for this syndrome is often invasive and sometimes involves relatively aggressive therapeutic procedures, such as circulatory support or cardiac transplantation. It is therefore critical to validate prognostic markers at the patient’s admission, during the first days of hospitalization or at discharge, in order to select the best therapeutic option.
This review discusses the emerging clinical, physiological and biological markers that seem promising, with a view to developing a score that will improve AHF patient management.
Methodological challenges
Before detailing and prioritizing the prognostic markers, it is important to consider a number of methodological challenges.
For which “acute heart failure” setting are we considering prognostic markers?
AHF or ADHF can take multiple forms. De novo AHF, the first manifestation of the disease and subacute decompensation of known CHF are very different conditions in terms of aetiology and underlying associated co-morbidities. In both cases, severity can be either mild (e.g. hypertensive pulmonary oedema in patients with preserved left ventricular ejection fraction) or severe (cardiogenic shock, that is often excluded from the definition of AHF and for which mortality rate is high). It is therefore important to specify the type of AHF/ADHF under consideration, when trying to develop a scoring method.
Which doctor would find such a score useful?
One of the specificities of AHF is that it can be managed by cardiologists, intensivists, emergency physicians or internists in an intensive care unit (ICU), cardiology department or internal medicine ward. The patient is directed to one department or another, depending on his or her profile and the severity of the AHF or associated co-morbidities. For instance, patients with co-morbidities are more likely to be managed in emergency and internal medicine departments. Furthermore, the therapeutic approaches are often different: in ICUs there is extensive use of positive inotropes, ventilation, intra-aortic balloon pumps and ultrafiltration, as opposed to less invasive oral or intravenous treatment in cardiology or internal medicine wards. A more aggressive therapy may also affect patient prognosis (e.g. iatrogenic effects). Therefore, patient diversity and the different departmental therapeutic approaches make it difficult to produce a comprehensive score that can be applied in any circumstance, as prognostic markers may vary from one situation to another.
When is the best time to evaluate prognosis?
One might imagine that the sooner the prognosis is established, the better (i.e. at patient admission, including in the emergency department). However, not all the variables that are necessary to calculate a prognostic score may be available in the first hours of management, such as echocardiography and some biomarkers. In addition, the response to treatment, which is a powerful prognostic marker, especially in the early days, is only available a few hours after initiating treatment. Several studies have also shown that the prognostic scores measured at patient discharge, including those involving biomarkers, have a better prognostic value than those determined at admission. Some studies have also highlighted the prognostic importance of the variations in these markers between admission and discharge. Nevertheless, from a practical standpoint, it is probably true that the sooner the prognosis can be assessed, the better it is for the patient because they can then be directed towards the most appropriate medical strategy (e.g. from immediate discharge after intravenous diuretics to immediate transfer to the ICU for invasive ventilation and/or cardiac assistance).
Which endpoints should be considered?
The most relevant and undebatable criterion is mortality, especially given the high death rate for this disease. Another criterion that is equally relevant in terms of severity and public health cost is the number of hospital readmissions. The predictive criteria for death and for rehospitalization may not, as intuitively thought, be the same. Importantly, it may also be relevant to take into account the cause of hospital readmission or death in such a prognostic score. Indeed, in elderly patients or those with HF with preserved ejection fraction, readmission for co-morbidities is as frequent as for cardiovascular aetiologies. Furthermore, among readmission for cardiovascular reasons, one can distinguish those that are associated with HF and those that are not (e.g. stroke, pulmonary embolism, myocardial infarction, renal failure). Finally, among those hospitalized for HF, it is often difficult to distinguish those cases caused by HF or associated with HF, mixed causes (infection plus cardiac decompensation) being particularly frequent. Likewise, the evaluation of the cause of death should follow similar criteria.
Prognostic criteria: for which follow-up duration?
Prognostic criteria may vary depending on whether one is interested in in-hospital, 1-month, 6-month or even 1-year mortality . Although in-hospital or 1-month mortality may at first seem most relevant in acute pathologies, it is clear that we cannot overlook outcome at 6 months. Many therapies, such as positive inotropic amines, can improve short-term prognosis without beneficial effect and even with deleterious effect, in the long term . In contrast, it is difficult to consider a 6-month or 1-year prognosis defined at admission as being final, as numerous factors — particularly medical interventions — may affect outcome at the medium or long term: for instance, in the recently published RELAX-AHF study, 48-hour treatment with serelaxin seemed to affect 6-month outcome . Moreover, we have seen in the SURVIVE study that one of the most powerful prognostic criteria at 6 months was whether beta-blocker therapy was present at discharge . These studies suggest that discharge may be the best time to evaluate mid- and long-term mortality. However, between hospitalization discharge and the sixth or twelfth month after discharge, many procedures can also be performed: coronary angiography, revascularization, resynchronization, implantable cardioverter defibrillator implantation, rehabilitation or educational programmes. All of these may interfere with the prognosis at 1 year, disrupting prognostic marker values measured at either admission or discharge.
How can the relevance of a prognostic factor be assessed?
Until recently, univariate or multivariable models were used to determine the best single or multiple prognostic criteria. However, statistical superiority does not necessarily mean greater clinical relevance, and new statistical methods are now available (such as the New Reclassification Index) which allow for the clinical relevance of prognostic variables .
Prognostic criteria
Thus far, all of these challenges have greatly hindered the possibility of obtaining a tool for risk stratification in clinical practice. Nevertheless, several studies — such as the ADHERE registry , EFICA , SURVIVE (Mebazaa et al., unpublished data), ALARM , IN-HF and OFICA — have been carried out to identify prognostic criteria ( Table 1 ). These studies, performed in various populations from various countries, and others based on smaller populations, have identified different individual prognostic criteria, which can be grouped under various headings ( Table 2 ).
| Study reference | Time period | Data source | Number of patients ( n ) | Mortality risk factors | Prediction |
|---|---|---|---|---|---|
| EFFECT | 1997–2001 | Registry | 4031 | Age; higher respiratory rate; low SBP; increased BUN; hyponatraemia; cerebrovascular disease; dementia; COPD; cirrhosis; cancer; low haemoglobin | 30-day mortality 1-year mortality |
| OPTIME-HF | 1997–1999 | Clinical trial | 949 | Age; NYHA functional class; SBP; BUN; sodium | 60-day mortality; death or rehospitalization at 60 days |
| ADHERE | 2001–2003 | Registry | 32 229 | BUN; creatinine; SBP; age; heart rate | In-hospital mortality |
| OPTIMIZE-HF | 2003–2004 | Registry | 48 612 | Creatinine; sodium; age; heart rate; liver disease; previous CVA/TIA; peripheral vascular disease; race; left ventricular systolic dysfunction; COPD; SBP; previous HF hospitalization | In-hospital mortality |
| GWTG-HF | 2005–2007 | Registry | 26 837 | Age; COPD; heart rate; SBP; sodium; BUN | In-hospital mortality |
| EFICA | 2011 | Registry | 599 | Shock; renal dysfunction; ischaemia; liver dysfunction; previous ADHF episode; co-morbidity; SBP; pulmonary oedema | 1-month and 12-month mortality |
| IN-HF | 2009 | Registry | 1855 | SBP; age; somnolence/confusion; sodium; creatinine; shock; pulmonary oedema | In-hospital mortality |
| OFICA | 2009 | Registry | 1658 | Age; SV arrhythmia; SBP; creatinine; median; natriuretic peptides | In-hospital mortality |
| PROTECT | 2012 | Clinical trial | BUN; respiratory rate; SBP; heart rate; albumin; cholesterol; diabetes; previous HF hospitalization | ||
| MOCA | 2013 | Registry | 5306 | Age; sex; SBP and DBP; eGFR; sodium; haemoglobin; heart rate; NT-proBNP; CRP; MR-proADM; sST2 | 1-month and 12-month mortality |
| Demographics | Age; male; black; poor social situation |
| Medical history | Previous HF hospitalization; diabetes; ischaemic aetiology; vascular disease; co-morbidities |
| Symptoms and signs | Congestion; low output; low BP/shock; high heart rate; high BR; low SvO 2 ; low body mass index |
| Laboratory | High creatinine; low sodium; low haemoglobin; high urea/BUN; low albumin; high lactates; high natriuretic peptides; high proADM; high sST2; high CRP; high troponins |
Patient characteristics: age, sex and weight
Age is a powerful prognostic factor in all studies, with prognosis getting poorer with age. The role of sex is less clear, although women often have a better prognosis than men. In terms of body weight and, more specifically, body mass index, contrary to what has frequently been observed in the general population and in primary prevention, a low body mass index carries a worse prognosis than a high body mass index, as the most cachectic patients seem to have the poorest prognosis .
Heart failure
The duration and severity of the disease, de novo HF versus decompensation of known CHF, cardiogenic shock, signs of congestion assessed by any method (rales, oedema, body weight, etc.) and blood pressure usually emerge as prognosticators in multivariable analyses, for instance, low blood pressure always predicts grid outcome while hypertension generally confers relatively good prognostic value, contrary to what is observed in the normal population. Additional variables can easily be determined from doppler echocardiography (left ventricular ejection fraction, E/E’ ratio, pulmonary arterial function and right ventricular function), although these variables are not always available at the time of initial patient evaluation.
Co-morbidities
Chronic renal failure, diabetes, chronic obstructive pulmonary disease (COPD) and vascular disease are co-morbidities that are frequently associated with HF, especially in the elderly, and have their own negative impact on outcome.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
