Futility in cardiology: the need for a change in perspectives

Abstract

In the complexity of heart failure (HF) syndrome, correct staging and prognostic stratification is a great issue to solve in order to tailor therapies on patient’s needs. We read with great interest the article by Straw et al.1 addressing the prognostic value of the Surprise Question ‘Would you be surprised if this patient were to die within the next year?’ among 129 patients hospitalized with acute HF. The Surprise Question was able to .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . identify nearly all patients who were within the last year of life and those unlikely to die (sensitivity 85%, negative predictive value 88%). Further, the ‘not surprised’ response from cardiologists was significantly associated with reduced survival, even when adjusted for important clinical covariates. Hence, the Surprise Question has been proposed as a useful tool to assist in prognostic stratification of patients with HF, identifying those who may benefit from palliative care referral.2 Although this simple tool appears to be intriguingly and quick, it is misleading and inadequate when applied to the high complexity of HF and, moreover, its futility is corroborated by the low specificity (59%) and positive predictive value (52%) that led to overestimation of mortality and severity of HF. The limitation of current prognostic tools is low specificity that is mainly driven by the lack of understanding of the myocardial abnormalities underlying HF. Prognostic discrimination should imply mechanistic insights, grouping patients based on critical pathophysiological abnormalities. Moreover, to improve prognostic stratification, a comprehensive ‘staging’ of multi-organ damage should be pursued, going beyond the cardiocentric vision based on symptom by New York Heart Association (NYHA) class and evaluation of ejection fraction.3 The results reported by Straw et al.1 support our considerations. Variables significantly associated with mortality were poor renal function, anaemia, hypoalbuminaemia, but neither NYHA class nor ejection fraction predicted mortality. We proposed a new staging system for HF, named HLM (A-B),4 in analogy to TNM used in oncology: ‘H’ for heart damage, analogous with ‘T’ of tumour; lung involvement (L), since lungs are functionally and anatomically the lymph node stations of the heart, in analogy with ‘L’ of TNM; malfunction (M) of peripheral organs such as the kidney, liver, brain and hematopoietic system, in analogy with ‘M’ of TNM. Each parameter is allocated in four levels of severity.4 Such holistic approach may provide a global assessment, allowing for better definition of pathophysiology and prognosis of HF,5 resulting as simple, all-encompassing and non-simplistic.