Frailty in TOPCAT: A Deep Dive into the Deficit Index Approach for Defining Frailty

This article refers to 'The frailty syndrome and outcomes in the TOPCAT trial' by N.A. Sanders et al., published in this issue on pages 1570–1577. Heart failure with preserved ejection fraction (HFpEF) disproportionately affects older adults, making HFpEF a common disease treated by geriatric cardiologists. Incorporating frailty measures into our understanding of older adults with HFpEF is recommended,1 but little data currently exist to guide clinicians. While frailty can be conceptualized as the cumulation of physiologic deficits over time that results in vulnerability to stressors,2 HFpEF can be considered a disease borne out of the accumulation of both age and key co-morbid illnesses such as obesity, hypertension, diabetes, and atrial fibrillation.3 It is fitting, therefore, that in the current issue of the Journal, Sanders and colleagues use the deficit index approach to describe frailty in patients enrolled in the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist (TOPCAT) trial.4 The deficit index is based on the theory that it is the number, rather than the nature of physiologic deficits that increase a patient's vulnerability to stress, resulting in poor outcomes such as hospitalization or death.5 The deficit index is calculated as the number of deficits a patient has (e.g. co-morbid illnesses, abnormal lab values, inability to complete activities of daily living, etc.) divided by the total number of deficits assessed. The result is a fraction between 0 and 1, with 0 indicating no frailty and 1 indicating maximal frailty. The deficit index has the convenient feature of being moulded to whatever data are available; if the main principles of constructing a deficit index are adhered to,6 then deficit indices can be compared as 'apples to apples' among data sets. The deficit index is designed to measure the total illness burden a person has, and thus helps predict when the entire biologic system might become overloaded and fail, leading to death. Therefore, construction of a deficit index must follow several rules of thumb: the deficits assessed should not saturate with age, should be across multiple organ systems, should be readily available in routine clinical care, and should number between 30 and 40.6 TOPCAT randomized patients with HFpEF to either spironolactone or placebo, with the primary outcome being cardiovascular death, hospitalization for heart failure (HF) or aborted sudden cardiac death.7 Due to irregularities in the patients enrolled in Russia and Georgia,8 Sanders and colleagues appropriately included only patients enrolled in the Americas in their analysis (n = 1767).4 The authors chose to include individual questions from the Kansas City Cardiomyopathy Questionnaire (KCCQ) as part of their deficit index. The KCCQ is a well validated, patient-reported HF-specific health status questionnaire9 that is scored from 0–100, with higher scores indicating better health status. The individual questions of the KCCQ comprised 15/39 components of the authors' deficit index. The KCCQ is already known to predict hospitalization and mortality outcomes in patients with HFpEF.10 The remainder of the deficits reflected primarily the cardiovascular system (ejection fraction, blood pressure, pulse pressure, history of coronary artery disease, myocardial infarction, atrial fibrillation, hypertension, revascularization, dyslipidaemia and stroke) with fewer components of the endocrine (blood sugar, history of diabetes, and body mass index), renal (creatinine, estimated glomerular filtration rate, sodium and potassium), haematologic (haemoglobin and white blood cell count), and respiratory (history of chronic obstructive pulmonary disease) systems. Therefore, the mandate to use a wide variety of organ systems6 was somewhat met in these authors' deficit index. However, the significant reliance on cardiovascular and HF deficits makes their deficit index more of a 'cardiovascular deficit index' than a global frailty deficit index. Therefore, it is unsurprising to find that the authors' deficit index, which includes already known powerful prognostic indicators among patients with cardiovascular disease, predicts the TOPCAT primary outcome. It is also unsurprising that the highest deficit index class was the youngest group; again this finding is likely driven by heavy reliance on the KCCQ, which is inversely associated with age.11 What is of interest in this study was the lack of interaction between deficit index class and treatment assignment in determining the primary outcome. As expected, the absolute risk of the TOPCAT primary outcome increased with deficit index class. In the absence of a treatment interaction with spironolactone, the number needed to treat in the highest deficit index class is smaller than that of the lower classes. In other words, treating patients in the highest deficit index class with spironolactone may be an efficient intervention in this patient population. This finding goes against the traditional wisdom of a 'less is more approach' in frail, older adults. However, given the data included in this particular deficit index, the highest deficit index class really represents younger patients with symptomatic HFpEF and high burden of cardiovascular disease. The findings by Sanders and colleagues confirm what many of us suspect clinically: frailty in the setting of HFpEF should garner as much of our clinical and research attention as frailty in the setting of other cardiovascular diseases. However, additional studies examining which treatments might still be safe and effective despite the presence of frailty—as measured by both disease-specific and global frailty metrics—remain sorely needed in the study of older adults with HFpEF. Conflict of interest: none declared.