Larissa Heart Failure Risk Score: A Proposed Simple Score for Risk Stratification in Chronic Heart Failure

Heart failure (HF) is a clinical syndrome with high morbidity and mortality.1 Although many prognostic models have been developed in HF, most require the use of calculators limiting their use in daily practice. We have developed a simple score—the Larissa Heart Failure Risk Score (LHFRS)—that risk stratifies patients admitted with acute HF (AHF).2 The purpose of this study was to examine the prognostic value of LHFRS in ambulatory patients with chronic HF (CHF). This is a retrospective study including 454 consecutive ambulatory CHF patients (age ≥ 18 years) followed up for 1 year. Patients with haemoglobin <10 g/dL, acute coronary syndromes (<3 months), sepsis, and those who had received blood transfusions (the year preceding the inclusion in the study), had been hospitalized for AHF the month preceding the clinical evaluation or had a history of malignancy were excluded. The study endpoint was the 1-year all-cause death or hospitalization for HF (whichever occurred first). This study conforms to the principles of the Declaration of Helsinki and was approved by the local institutional review committees. The derivation of LHFRS has been described in detail.2 Variables included are hypertension history, myocardial infarction history, and red cell distribution width (RDW) at admission. Absence of hypertension history, presence of myocardial infarction history, and admission RDW value ≥15% (median value) are assigned 2, 1, and 1 points, respectively. Patients with LHFRS = 0 have the best, whereas those with LHFRS = 4 the worst score. Continuous variables exhibiting normal or non-normal distribution (Kolmogorov–Smirnov test) are presented as mean ± standard deviation and median (interquartile range), respectively. Categorical variables are presented as frequencies and percentages. Differences at baseline between events and non-events were assessed using the Student's t-test, the Wilcoxon rank-sum test, and the Fisher's exact test as appropriate. A visual assessment of LHFRS performance was made using a bar chart, depicting number of events for each score value. Subsequently, the LHFRS was inserted as the only variable into a logistic regression model. A receiver operating characteristic (ROC) curve was then produced and the area under the ROC curve (AUC) estimated. ROC curves were also constructed and compared for patients with HF with preserved (HFpEF) or reduced ejection fraction (HFrEF).3 P-values were two-sided and those <0.05 were considered as statistically significant. The baseline characteristics of the study population are depicted in Table 1. The study composite endpoint occurred in 159 patients (35%). An increasing trend was noted in the number of patients experiencing an event, from lower to higher LHFRS values (Figure 1A). The obtained odds ratio was 2.7 95% confidence interval (CI) 2.2–3.4; P < 0.0001, meaning approximately three times higher odds of an event for every unit increase of the LHFRS value. Furthermore, LHFRS discriminated patients who experienced the study endpoint from those who did not with an AUC of 0.78 (0.73–0.82) (Figure 1B), and showed a similar good performance in patients with HFrEF AUC 0.76 (0.71–0.82), n = 294 and HFpEF AUC 0.77 (0.67–0.87), n = 146. These results indicate that LHFRS may accurately stratify ambulatory CHF patients. Three validated and commonly used scores in CHF are the Seattle HF Risk Model, the MAGGIC Risk Score and the MECKI score.4-7 The first consists of 14 continuous variables and 10 categorical values, the second includes 13 variables, whereas the third encompasses six different variables. The vast majority of scores used in acute HF do not perform well in CHF. Our model consists of only three variables, all with proven prognostic ability both in acute and chronic HF.8-14 This study has several limitations. First, the very small sample size deriving from only two centres; second the retrospective nature of the data collection, although this is more representative of ‘real-world patients’; third, the inclusion of ambulatory patients with CHF in New York Heart Association (NYHA) classes II and III limiting the implementation of the LHFRS in other NYHA classes; fourth, as neurohumoral inhibitors are used both in HF and hypertension, labelling some patients hypertensive based on treatment may occasionally be problematic; fifth, our patient cohort included a Caucasian Greek population and the results may not be applicable to other ethnicities. Regarding the therapeutic implications of the LHFRS, it is reasonable to assume that HF patients with high LHFRS should be treated more aggressively compared to those with low LHFRS. In conclusion, LHFRS is a simple prognostic score with substantial prognostic ability in ambulatory CHF patients. However, LHFRS should be further validated in larger prospective studies. Conflict of interest: none declared.