Heart Failure Diagnosis and Prognosis in the Elderly: The Proof of the Pudding is in the Eating

This editorial refers to ‘Clinical evaluation of geriatric outpatients with suspected heart failure: value of symptoms, signs, and additional tests’ by I. Oudejans et al., and ‘Predictors of clinical outcome in elderly patients with heart failure’ by L. Manzano et al., published in this issue on pages 518–527 and 528–536. The prevalence of heart failure (HF) increases with age, with the highest rates occurring in those aged >80 years.1 Older patients also have higher hospitalization rates, longer length of hospital stay, and are more often discharged to long-term care facilities.2 Moreover, the pathophysiology of HF in the elderly is increasingly complex, and may include abnormalities of diastolic function and arterial stiffness, diminished responsiveness to beta-adrenergic stimulation, impaired endothelial function, as well as the detrimental effects associated with multiple co-morbidities.3 Heart failure with preserved ejection fraction, seems to be the predominant clinical phenotype in elderly patients with signs and symptoms of HF.4 The treatment of HF in the elderly (especially in those with preserved ejection fraction) remains empirical due to the fact that these patients are usually excluded from randomized clinical trials. However, recent data suggest that the use of life-saving medications in the elderly can reduce new cardiac ischaemic events and re-hospitalizations and can improve outcome.5–7 In a sub-analysis of the Study of the Effects of Nebivolol Intervention on Outcomes and Rehospitalization in Seniors with Heart Failure (SENIORS) it was found that treatment with a β-blocker reduced the incidence of cardiac ischaemic events in patients with HF of ischaemic aetiology.6 In the CHARM (Candesartan in Heart Failure—Assessment of Mortality and Morbidity) trial, 7599 HF patients were randomized to treatment with candesartan or placebo; this included 3169 patients aged >70 years. Mean follow-up was 37.7 months. The proportional hazards model was used to estimate the effect of treatment on efficacy and safety within the following five age groups: <50 years (8% of total cohort), 50–59 years (19%), 60–69 years (31%), 70–79 years (33%), and ≥80 years (9%). The risk of cardiovascular death or HF hospitalization (primary outcome) increased from 24% in the lowest age group to 46% in the highest age group (and mortality from 13 to 42%). The relative reduction in risk of the primary outcome with treatment was similar irrespective of age. Thus, the absolute benefit was greater in HF patients of more advanced age (3.8 patients avoided an adverse outcome per 100 patients treated in the lowest age group compared with 6.8 in the highest).7 Thus, the early recognition and treatment of HF, especially in the elderly, can lead to a reduction in long-term mortality and/or morbidity. According to the European Society of Cardiology,8 the diagnosis of HF requires the presence of typical signs and symptoms and objective evidence of a structural or functional abnormality of the heart at rest (Figure 1). A common problem in everyday clinical practice is the delayed detection and diagnosis of HF in the elderly; this is often because in the majority of cases there is an overlap between signs and symptoms of HF and the patient's various other co-morbidities. On the other hand, this overlap could also lead to over-diagnosis and unnecessary drug therapy. Moreover, limited access to echocardiography and the fact that age and co-morbidities (such as renal dysfunction and atrial fibrillation) may affect plasma levels of natriuretic peptides, further limits the diagnostic ability of the available tools and the adequate evaluation of elderly patients suspected of HF.9 Thus, standardized diagnostic criteria would be useful for the diagnosis of incipient HF, especially in primary care where gradual onset and a long period of mild symptoms can mask the diagnosis. The most common scores used for the diagnosis, definition, and classification of HF are presented in Table 1. However, most of these scores have only been used in clinical trials and epidemiological or population studies, they have not been evaluated extensively in a clinical setting, especially in the elderly. In this issue of the European Journal of Heart Failure Oudejans et al.15 describe how they screened a population of geriatric outpatients with suspected HF and confirmed the diagnosis in half of the patients. Using simple demographic and clinical characteristics, the authors created a non-invasive and easily applied rule that adequately predicted the existence of HF in this complicated patient population. Given the fact that classical signs of HF are not present in ~30–40% of HF patients, this new score that includes male gender, increase in age per 10 years, loss of appetite, absence of wheezing, low body mass index, and the classical symptom of nocturnal dyspnoea may be useful in clinical practice for the evaluation of geriatric patients. Moreover, the addition of the biomarker N-terminal-pro B-type natriuretic peptide (NT-proBNP) seemed to substantially improve the diagnostic accuracy of the proposed clinical model. It should be noted that breathlessness was the main reason for referral in only 34.5% of patients in the study cohort, while functional impairment (39.8%), cognitive impairment (30.6%), and mobility disorders (28.6%) or their combination was the main reason for referral in the other patients. Moreover, the average number of co-morbidities per patient was six. A potential diagnostic algorithm based on these clinical observations is shown in Figure 2. This is an arbitrary interpretation of the data presented in the Oudejans study, to show how difficult is to implement these variables in everyday clinical practice. For most cardiologists and general practitioners HF is related to typical signs and symptoms; although the concept introduced by Oudejans et al., suggesting that the presentation of HF in the elderly is different, has been discussed for many years, it is a new concept for many. Moreover, since they are not always available in primary care, the exact role of the electrocardiogram (ECG) and chest X-ray for the diagnosis of HF in the elderly is not clear from the present study. However, it is clear that the studied population only included patients from a regional database, and thus this clinical model now needs to be tested prospectively in larger patient cohorts in order to confirm its reproducibility independently of geographic and regional characteristics. There is strong clinical evidence to suggest that advanced age is a powerful prognostic indicator in patients with acute or chronic HF.16–18 However, it is very difficult to predict outcome and provide risk stratification within an elderly population with HF, because most data come from clinical trials in which the elderly are under-represented. In one of the few available surveys, EuroHeart Failure Survey II, in addition to age; prior myocardial infarction, creatinine level, and low plasma sodium were all independently associated with 1 year mortality in patients hospitalized for HF.19 A recent report from Komajda et al.20 evaluating data from 4128 elderly patients in the I-PRESERVE (Irbesartan in Heart Failure with Preserved Ejection Fraction) study, showed using multivariable Cox regression models that log NT-proBNP, age, diabetes mellitus, and left ventricular ejection fraction were the strongest independent factors for all-cause mortality. Other independent factors associated with poor outcome included quality of life, a history of chronic obstructive lung disease, log neutrophil count, heart rate, and estimated glomerular filtration rate. This sub-analysis of I-PRESERVE shows that in a large sample of elderly patients with HF and preserved ejection fraction, simple clinical, demographic, and biological parameters can identify sub-groups at very high risk of events. Extending this concept, in this issue of the European Journal of Heart Failure Manzano et al.21 propose a new risk model for elderly HF outpatients using information obtained from routine clinical assessments and biochemical measurements. Using data from the SENIORS trial and a sophisticated statistical model, the authors found that in addition to conventional prognostic markers [age, New York Heart Association (NYHA) class, body mass index, poor exercise capacity, history of atherosclerotic disease and prior myocardial infarction, diabetes, renal function, and ECG abnormalities], the two laboratory markers, left atrial dimension and plasma uric acid level, also provided useful prognostic information in elderly HF patients. Recently, the MeRGE collaboration22 that includes prospective data from 18 HF studies, also demonstrated the prognostic importance of left atrial size in patients with systolic dysfunction. In multivariate analysis, left atrial area was significantly associated with prognosis independently of age, NYHA class, left ventricular ejection fraction, and restrictive filling pattern. In addition, the findings of Manzano et al. regarding the prognostic role of plasma uric acid in elderly HF patients are in keeping with the proposed risk factors in the Seattle HF risk model23 as well as previous studies in the general HF population.24–26 It has been suggested that a high uric acid level may only predict poor outcome when it is a marker of increased xanthine oxidase activity, but not when hyperuricaemia is the result of impaired renal excretion of uric acid.26 Nevertheless, plasma uric acid as a marker of oxidative stress, catabolic processes, and/or severity of the syndrome, seems to preserve its prognostic importance in the elderly. Consequently, the two studies published in this issue of the European Journal of Heart Failure may provide new tools for the diagnosis and risk stratification of elderly HF patients based on history, clinical examination, and simple biochemical and/or echocardiographic parameters. This new knowledge may help clinicians to optimize the management of elderly patients with HF and lead to improved patient prognosis. However, these new parameters will have varying usefulness in different populations. Prospective trials and registries are needed in order to examine the reproducibility and validity of the suggested parameters in routine clinical practice. Because as we all know … the proof of the pudding is in the eating.