“Change is the only constant; stability is a necessary illusion”

–Heraclitus

Heart failure (HF) is classically defined as a chronic disease characterized by periods of clinical stability of variable duration, interrupted by occasional episodes of worsening despite continued therapy. This concept has led to the widespread adoption of the ambiguous and misleading term “stable heart failure” in medical records.1 During an outpatient visit, symptoms are often assessed using the New York Heart Association functional classification, which is highly subjective and lacks precision.2 Nevertheless, a thorough evaluation of HF should include a comprehensive assessment of activities of daily living, physical examination findings, diuretic needs, natriuretic peptide levels, and multimodality imaging. When data obtained from these assessment tools are integrated into HF risk calculators, the sense of stability often dissipates, giving way to the perception of imminent risk.3

In recent years, health-related quality of life (QoL) has emerged as a key outcome reflecting patients’ priorities and needs. Patients with HF often have worse QoL than the general population and even than patients with other chronic diseases.4 Among disease-specific questionnaires, the Minnesota Living with Heart Failure Questionnaire and the Kansas City Cardiomyopathy Questionnaire (KCCQ) have been extensively validated and widely used.5 In a recent paper published in Revista Española de Cardiología, Kim et al.6 performed a post-hoc exploratory analysis of the Prospective Comparison of Angiotensin Receptor Neprilysin Inhibitor with Angiotensin Receptor Blocker Global Outcomes in HF With Preserved Ejection Fraction (PARAGON-HF) trial.7 The authors categorized the KCCQ Clinical Summary Score into 5 health states, including death, using 25-point ranges, and evaluated dynamic changes in QoL at prespecified protocol visits using a multistate Markov model.

One of the methodological aspects that warrants careful consideration is the definition of health states based on symmetrical KCCQ Clinical Summary Score cutoff points. Although the use of 25-point intervals has been widely adopted in previous studies and facilitates comparability with existing literature, this categorization remains largely pragmatic. Current guidelines do not provide specific recommendations for classifying KCCQ scores into discrete health states, and there is limited empirical evidence formally validating that these intervals correspond to clinically or perceptually equivalent distinctions in health status.

Unlike other health-related QoL instruments, such as the EuroQol Five-Dimension (EQ-5D), in which health states are derived from structured evaluation frameworks incorporating population preferences (eg, time trade-off or standard gamble techniques), the health state classification applied to the KCCQ does not stem from a formal modeling or utility-based validation process.8 Consequently, although the symmetrical structure of these cutoffs offers analytical simplicity and consistency with prior research, it does not necessarily ensure clinical equivalence or homogeneous perceived differences between adjacent categories. Therefore, results related to patient assignment among KCCQ-based health states should be interpreted with appropriate caution, acknowledging this inherent methodological limitation.

The Markov model is an innovative machine learning-based statistical method that assumes that variations in QoL depend solely on the current health state. It allows estimation of the probabilities of transitioning between adjacent KCCQ states and the average duration a patient remains in a specific QoL status (also known as sojourn time).9 By applying the Markov model to the 4707 participants in the PARAGON-HF trial with available baseline KCCQ scores, the authors analyzed: a) the mean duration of sojourn time within the different KCCQ states; b) the association between QoL and mortality; and c) the impact of sacubitril/valsartan on transitions between health states.

Approximately 90% of patients had KCCQ completed questionnaires at the end of follow-up (36 months). Most individuals (86.5%) had an excellent or good baseline health state (≥ 50 of 100 points on the KCCQ). However, the QoL trajectory showed a steady decline of approximately 4 points per year, with a clinically relevant decrease in the proportion of excellent and good states and a corresponding increase in the proportion of fair and poor strata.

Interestingly, QoL was highly dynamic over time. A total of 22% of patients in the excellent stratum transitioned to worse health states during follow-up. Variation was more pronounced among patients in the lower KCCQ categories: 42.6%, 49.3%, and 60.1% of patients in the good, fair and poor strata, respectively, shifted to other health states. Consistent with these findings, the sojourn time was twice as long in the excellent stratum compared with the other health states (1.57 vs 0.70-0.80 years). One noteworthy and surprising observation was that many of the individuals with lower baseline KCCQ scores transitioned to better health states over time: 55.4% of transitions within the good stratum were from good to excellent; 76.5% within the fair stratum were from fair to good, and 87.5% within the poor stratum were from poor to fair.

It is therefore important to emphasize that HF is a progressive condition with a natural tendency toward continued worsening over time.10 Nevertheless, individual health trajectories may be bidirectional and highly dynamic, which highlights the importance of clinical vigilance and the timely, comprehensive implementation of disease-modifying therapies.11

Moreover, HF with preserved ejection fraction has become one of the major contemporary epidemics and poses a huge economic and administrative burden on health care systems.12 Understanding sojourn times within each QoL stratum may help optimize resource allocation: greater focus and effort should be directed toward patients with lower KCCQ scores, whereas follow-up intervals may be safely extended in individuals with baseline and persistently excellent QoL over time.

The authors reported an inverse and exponential relationship between QoL and mortality rates (annual mortality was 5 times higher in the poor state compared with the excellent state: 15.6% vs 3.1%).6 Furthermore, as QoL worsened, the proportion of nonsudden cardiac deaths and noncardiovascular deaths progressively increased. These observations are consistent with previous studies, confirming that QoL is a strong and independent predictor of all-cause mortality across the entire spectrum of HF, in all geographic regions, and in both reduced and preserved ejection fraction.13

Last but not least, there was a weak suggestion that sacubitril/valsartan may improve clinical outcomes among patients in the poor health state. However, this stratum included a tiny number of individuals (68 of 4707), no other transitions were affected, and sojourn times within each health state did not differ between the sacubitril/valsartan and valsartan groups. Therefore, these results should be interpreted with caution.

A previous subanalysis of the PARAGON-HF trial showed that sacubitril/valsartan had a borderline benefit of 1 point in KCCQ Clinical Summary Score at 8 months compared with valsartan. This modest improvement was supported by greater odds (odds ratio,=1.31; 95% confidence interval, 1.06-1.61) of patients reporting a clinically meaningful improvement (≥ 5-point increase) in KCCQ scores.14 HF with preserved ejection fraction is a complex and multifaceted condition in which meaningful improvements in QoL are hard to achieve. Indeed, therapies that have demonstrated substantial reductions in hard outcomes but do not significantly affect body weight, such as sodium-glucose cotransporter 2 inhibitors and mineralocorticoid receptor antagonists, are associated with only modest increases of 1 to 2 points in KCCQ scores.15,16

Kim et al.6 should be congratulated for this compelling study, which aimed to thoroughly characterize health state trajectories in HF with preserved ejection fraction. The results emphasize the value of evaluating QoL not only as isolated time-point measurements, but also as a dynamic continuum throughout the natural course of HF. Worsening health states over time, and even failure to achieve the excellent stratum, should prompt a comprehensive and timely medical evaluation.

We believe that the implementation and integration of telemedicine into routine clinical practice could facilitate the systematic collection of QoL data through standardized questionnaires. This valuable information may be of great help in detecting early signs of decompensation, optimizing pharmacological and lifestyle interventions, assessing treatment response, and tailoring follow-up according to patient risk.