Authors’ reply

We would like to thank Szarpak et al¹ for their interest in and insightful comments on our recently published article on the evolving phenotype of heart failure (HF).² We fully agree with their emphasis on the observed shift in HF characteristics toward HF with preserved (HFpEF) and mildly reduced ejection fraction (HFmrEF). It should be stressed, however, that the precise definition of HFpEF was established in the 2016 European Society of Cardiology (ESC) guidelines, that is, 2 years after the beginning of our study. Furthermore, HF phenotype classification is inherently dynamic: individual patients may be reclassified into a “higher” EF phenotype (HFmrEF or HFpEF) following the implementation of appropriate therapeutic strategies.³ Conversely, reclassification in an opposite direction is also possible.

The authors correctly highlighted the increasing burden of comorbidities in the HF population we analyzed. This trend is consistent with findings from other publications.^4,5 It should also be acknowledged that the rising prevalence of comorbidities may influence the observed decline in the prescription rates of guideline‑recommended pharmacotherapies and modify the prognosis.^5,6

With regard to the authors’ comments on pharmacotherapy, it is important to note that angiotensin receptor‑neprilysin inhibitors (ARNIs) were first included in the ESC guidelines on HF treatment in 2016, whereas a recommendation regarding sodium‑glucose cotransporter 2 inhibitors (SGLT2is) was added in 2021, for HF with reduced EF (HFrEF). As the authors rightly noted, ESC recommendations supporting the use of SGLT2is across the entire HF spectrum were introduced as recently as 2023. Adoption of these therapies in Poland was initially limited, as sacubitril / valsartan has never been reimbursed for HF patients, while SGLT2is became partially reimbursed for HFrEF in May 2022, and for all HF phenotypes in July 2025. We are fully aware that having pharmacotherapy data available only at the time of hospital discharge, without information on the subsequent 5‑year follow‑up, represents a substantial limitation. One may reasonably speculate that the quality of pharmacotherapy has improved during long‑term management. Nonetheless, as demonstrated both by other researchers⁴ and our group in a separate study,⁷ the use of older HF medications has decreased, whereas increases in prescriptions of SGLT2is and ARNIs remain insufficient, indicating suboptimal implementation of guideline‑directed medical therapy.^4,7

Despite the absence of a dedicated HF echocardiographic protocol, we carefully reviewed all available echocardiographic data and clinical information to establish the HF phenotype and to exclude cases with incorrectly coded HF hospitalizations. For these reasons, we believe that the HF phenotypes identified in our study reflect true clinical presentations more closely, as compared with studies relying solely on administrative coding or general medical history.

Regarding the concern raised by the commentary authors, admission tertiles in our study reflected successive enrollment periods rather than unequal follow‑up duration. This approach allowed for a balanced comparison across equally sized groups and was justified by the relatively low number of HF hospitalizations in the early years of the study. Importantly, all tertiles had the same follow‑up duration, limited to 5 years. We also presented Kaplan–Meier survival curves, allowing survival patterns for each HF phenotype to be assessed across tertiles.

Another important observation concerns survival trends across tertiles within each phenotype. In our cohort, survival improved and mortality declined across successive tertiles, driven mainly by patients with HFrEF. Nevertheless, when examining survival curves for HFmrEF and HFpEF, one cannot overlook a potential trend toward reduced mortality, although statistical significance was not reached. This finding is particularly noteworthy given that, at the time, no clear guidelines for life‑saving therapies existed for these phenotypes. We fully agree with the authors’ conclusion that pharmacotherapy constitutes only one of the HF care components and does not capture the complexity of long‑term management. As noted earlier, appropriate diagnosis and treatment of an increasing number of comorbidities, along with improved access to long‑term care, likely played a crucial role in the observed outcomes. Moreover, in another project, we found that survival in patients from a simple educational study conducted 20 years ago did not differ from that observed in a contemporary HF population from the National Health Fund database (unpublished data).

In summary, despite limitations inherent to long‑term observational studies, we believe our findings provide a unique, real‑world insight into the HF population. The documented evolution of HF phenotypes and comorbidity burden offers valuable information that may support future improvements in HF care quality and guide strategic planning. The ongoing implementation of the National Cardiology Network should incorporate the specific needs of patients with HF, providing an opportunity to leverage system‑level improvements to enhance long‑term outcomes in this population.