Outside of organ transplantation, a diagnosis of advanced heart failure is a terminal one. It is often difficult to define exactly when a patient is entering end-stage heart failure, although many models and scoring systems exist to aid with the decision-making process.
Once the diagnosis is made or suspected, recognition of factors suggesting poor prognosis is essential in order to initiate conversations about long-term expectations for treatment. That may be to establish candidacy for advanced therapy (internal cardiac defibrillator [ICD], ventricular assist device [VAD] or transplantation) or to opt for palliation (Chaudhry & Stewart, 2016). While determining a patient’s prognosis is complex, it is essential for selection of advanced heart failure therapy. However, it is important to recognise that it is not necessary for referral to an advanced heart failure centre and so should not delay the referral process (Crespo-Leiro et al., 2018).
Identifiable factors which indicate the presence of advanced heart failure (Chaudhry & Stewart, 2016):
These more discriminating factors can be added to and filed into four main categories.
These can be determined via clinical examination and history. Physical examination will reveal, in severe advanced heart failure, the presence of a raised jugular venous pressure and a third heart sound resulting in gallop rhythm (Drazner et al., 2001), and cardiac cachexia. Prior to the development and widespread use of ACE inhibitors, 1-year mortality rates for NYHA class II/III were 15%, while class IV had up to 64% mortality rate (Chaudhry & Stewart, 2016). Hospitalisation is another key metric; Setoguchi et al. (2007) demonstrated a persistently increasing impact with recurrent admissions (Figure 4), and this may be a useful surrogate measure of frailty.
Bendopnea is a recently described symptom of advanced heart failure consisting of shortness of breath when bending forwards. In ambulatory patients, the presence of bendopnea was associated with heart failure admission at three months (HR 2.5, p<0.04), while in an older, hospitalised patient population, it was associated with short-term mortality (6 months) and reduced QoL (Baeza-Trinidad & Mosquera-Lozano, 2017; Thibodeau et al., 2017).
Historically considered to be less important functionally than the left heart, it has been established that right heart maximum ejection fraction is a discriminative indicator of prognosis. Theoretical involvement of the right heart as a mediator of metabolic derangement and playing a role in the development of cardiac cachexia has been suggested (Chaudhry & Stewart, 2016). Use of the Pulmonary Artery Pulsatility Index (PAPi) has been shown to be a powerful predictor of right ventricular (RV) failure in patients with acute inferior myocardial infarction and those receiving a LVAD. It has also now been shown to be a marker of RV dysfunction and an independent predictor of death or hospitalisation at 6 months (HR 0.91; 95% CI, 0.84–0.99, P=0.022) in patients with advanced heart failure (Kochav et al., 2018).
Biomarkers too are gaining increasing relevance; independent predictors of a worse prognosis include:
B-type natriuretic peptides (BNP) are secreted by the heart chambers in response to wall stress and are potent predictors of mortality. Doust et al. (2005) carried out a systematic review, finding that for a 100 pg/mL rise in BNP, there was a corresponding 35% rise in mortality rate.
Late stage heart failure patients are often unable to tolerate medications that are arresting the disease process, due to symptomatic hypotension and renal dysfunction.
Rather than focusing on mortality, it is often helpful to identify factors which may be associated with a poor quality of life (QoL). Allen et al. (2011) sought to identify factors that were associated with either a 6-month mortality or a persistently unfavourable quality of life (Table 3). They were able to identify many factors which would be detectable on admission, and constructed a scoring system.
Table 3. Scoring system developed by Allen et al. (2011) to identify characteristics associated with 6-month mortality or unfavourable quality of life.
Scores in the population studied were graphed against mortality and the probability of a persistently unfavourable QoL and showed an almost linear increase with rising score (Figure 5).
Prognostic scales have also been developed to assess risk in patients with advanced heart failure, particularly those waiting for heart transplant. The Heart Failure Survival Score (HFSS) and the Seattle Heart Failure Model (SHFM) are used in everyday clinical practice and the International Society for Heart and Lung Transplantation (ISHLT) guidelines recommend using both scales to assess the prognosis of ambulatory patients with advanced heart failure qualifying for heart transplant (Mehra et al., 2016). The HFSS scale has previously shown good prognostic strength for assessing patient outcomes, however, the majority of the studies showing this were conducted in a past era of heart failure therapy and therefore the scale's prognostic power in patients treated with current medical therapy is unknown (Ponikowski et al., 2016). Unlike HFSS, the SHFM has been updated to include inotropes, intra-aortic balloon pumping, ventilation, ultrafiltration and new ventricular assist devices (VADs) (Yancy et al., 2013). In addition to these two commonly used scales, there are a number of adapted prognostic scales such as the modified Model for End-Stage Liver Disease (modMELD) and MELD excluding INR (MELD-XI). New prognostic scales are also being developed including the Index for Mortality Prediction After Cardiac Transplantation (IMPACT) and the RADIAL score (right atrial pressure ≥10 mm Hg, recipient age ≥60 years, diabetes mellitus, inotrope dependence, donor age ≥30 years, length of ischemic time ≥240 minutes) (Szczurek et al., 2018).
The haematological parameters determined on admission for advanced heart failure have also been assessed to see if they are associated with patient mortality over time. During a three-year follow-up of 785 patients with advanced heart failure, type 2 diabetes (HR 1.46, 95% CI 1.15–1.86, p=0.002), elevated red blood cell distribution width (HR 1.05, 95% CI 1.04–1.07, p<0.0001) and a low relative lymphocyte count (HR 0.942, 95% CI 0.928–0.956, p<0.0001) were independently associated with mortality. Importantly, these measures are widely available and could represent simple markers for the prediction of three-year mortality (Szygula-Jurkiewicz et al., 2018).
With the potential ability to identify patients in the end stages of their disease process, it becomes possible to initiate discussions about the aims of medical intervention and management. Too frequently, end-stage heart failure patients die in acute hospital settings where the aim of treatment would never be curative. Adequate provision of palliative care and home support can enable patients to stay out of hospital. Positive ionotropic support and diuretic therapy can be more beneficial than analgesic agents, while morphine has a powerful action as a vasodilator and tranquiliser that can reduce the work of breathing (Crespo-Liero & Paniagua Martin, 2004).
Dev et al. (2012) carried out a survey to review patient preferences as they were nearing end of life. An interesting aspect of their findings was the variability of preferences. They sought to establish preference regarding location and circumstance of death – while the majority wished for an ‘unaware death’, others expressed a preference for an ‘aware death’. This conflicted with the fact that most patients did not wish to have their ICDs turned off – even if they were terminally ill. Again, the majority expressed a belief that death was preferable to spending their last days in a coma or requiring a feeding tube (Dev et al., 2012). What is apparent from their findings is the necessity for clinicians to discuss these issues ahead of time, and to come to a shared decision with patients regarding their wishes.
Frailty can be defined as a biological syndrome of decreased homeostatic reserve; it engenders an increased susceptibility to stressors and poor outcomes. A recent study by Madan and colleagues (2016) explored whether frailty was of prognostic value in advanced heart failure. They found that frailty testing with the Fried Frailty Index (assessing weight loss, exhaustion, weakness, slow gait and reduced physical activity) was able to accurately identify those at most risk of death or hospitalisation (Table 4) (Madan et al., 2016).
Table 4. Fried Frailty Index – score of 3 or more indicates frailty (Fried et al., 2001).
The prognostic value of a modified Fried Frailty criteria on mortality was also assessed in combination with peak VO2 and BNP levels in 201 patients with advanced heart failure. Over a median follow-up of 17.5 months, frailty was associated with a twofold increase in risk of death (HR 2.01, 95% CI 1.42–2.84, p<0.0001) although this was no longer significant when adjusted for BNP or peak VO2. However, frail patients with a peak VO2 <12 mL/kg/min were observed to have an increased risk of mortality (HR 1.72, p=0.006) versus those with a higher peak VO2 (≥12 mL/kg/min) (Moayedi et al., 2018).
Cardiopulmonary exercise testing (CPET) can provide important insights into a patient’s cardiovascular reserve and prognosis. While CPET can be a complicated task, requiring skilled staff and thorough interpretation of outputs, it can support the identification of patients who may be viable candidates for heart transplantation or long-term mechanical circulatory support. In fact, the HFA-ECS guidelines recommend CPET should be a part of the work-up for elective patients with advanced heart failure who are being considered for these treatments. For patients not being considered for these treatments, the 60-minute walking test can provide an alternative insight into patients’ functional impairment and help assess frailty (Crespo-Leiro et al., 2018). While CPET can be used to help identify patients suitable for advanced heart failure therapy, a patient’s age should not disqualify a patient. A study of 168 patients referred for therapy assessed the activation (engagement and ability to self-manage) of patients and observed that activation was consistent between age groups (<65 years vs. ≥65 years) (Carey et al., 2018).