This registry study assessed the real-world implementation of SGLT2 inhibitors (SGLT2i) in patients with heart failure (HF) across the spectrum of left ventricular ejection fraction (EF). The primary objective was to describe uptake of SGLT2i and identify patient and care factors independently associated with their use in routine clinical practice using data from the Swedish Heart Failure Registry (SwedeHF).
Patients enrolled in SwedeHF from 1 September 2022 to 16 December 2024 were included. The analytic cohort comprised 24,578 patients with HF. Median age was 75 years (interquartile range 65–81) and 33% were female. By EF category, 43% had heart failure with reduced ejection fraction (HFrEF), 31% had heart failure with mildly reduced ejection fraction (HFmrEF), and 26% had heart failure with preserved ejection fraction (HFpEF).
During the observation period SGLT2i prescriptions increased substantially across all EF groups. In patients with HFrEF, use rose from 78% at the beginning of the period to 89% by December 2024. In HFmrEF, SGLT2i use increased from 54% to 80% over the same interval. In HFpEF, use rose from 41% to 80% by the end of follow-up. The authors note that SGLT2i implementation was rapid in clinical practice, reaching approximately 90% in HFrEF and approaching a plateau close to 80% in HFmrEF and HFpEF.
Multivariable logistic regression was used to identify characteristics independently associated with SGLT2i use in the overall cohort and stratified by EF. Factors independently associated with higher likelihood of SGLT2i prescription included:
Type 2 diabetes (T2DM) was independently associated with SGLT2i use across EF categories; the association was stronger in HFpEF than in the other EF strata.
Renin–angiotensin–aldosterone system (RAAS) inhibitor use was associated with SGLT2i prescription across the EF spectrum, with stronger associations observed in HFrEF and HFmrEF. Mineralocorticoid receptor antagonists (MRAs) also showed associations with SGLT2i use, with patterns stronger in HFrEF and HFmrEF compared with HFpEF.
Follow-up in specialist care (as opposed to primary care) was associated with higher SGLT2i use in HFrEF and HFmrEF. Concomitant cardiovascular therapies, notably beta-blockers, RAAS inhibitors, and MRAs, were linked to greater likelihood of SGLT2i prescription, suggesting clustering of guideline-directed therapies in certain patient groups and care settings.
In this large Swedish registry cohort, SGLT2i were rapidly implemented in routine care for patients with heart failure across the EF spectrum between September 2022 and December 2024. By the end of the study period, use reached about 89–90% in HFrEF and approximately 80% in HFmrEF and HFpEF. Despite rapid overall uptake, the findings highlight remaining gaps in implementation: lower use among women, patients with HFpEF without T2DM, those followed in primary care, and patients with lower socioeconomic status. The observed associations with specialist follow-up and concurrent use of RAAS inhibitors, MRAs, and beta-blockers point to care delivery and treatment-pattern factors that influence SGLT2i adoption.
Details beyond those summarized in the abstract (for example, full model covariates, absolute effect sizes or odds ratios, temporality of prescriptions, reasons for nonuse, and adverse events) were not reported in the abstract provided here. The source text should be consulted for the full methods, detailed results, and any additional limitations discussed by the authors.
Rapid real-world adoption of SGLT2i occurred across EF categories in Sweden, with near-universal use in HFrEF and high uptake in HFmrEF and HFpEF by late 2024.
Persistent underuse was identified in specific subgroups (women, HFpEF without T2DM, patients in primary care, and those with lower socioeconomic status), indicating targets for quality-improvement initiatives.
Specialist follow-up and concurrent use of other heart-failure therapies were associated with higher SGLT2i use, underlining the role of specialty care and integrated treatment regimens in implementation.
For full numeric results, detailed regression outputs, and discussion of limitations and implications, refer to the complete article available via the cited DOI.