In July 2023, heart failure guideline recommendations were expanded to include the class of SGLT2 inhibitors. This study assessed whether the guideline expansion was followed by differential changes in primary care prescribing of three marketed monotherapy SGLT2 inhibitors in England: dapagliflozin, empagliflozin, and canagliflozin. The primary objective was to compare postintervention trends in utilisation, using national aggregated prescribing data.
The analysis used national monthly primary care prescribing data for England covering January 2019 through December 2025. Prescribing was analysed at the drug level using aggregate metrics; no patient-level data or indication-specific information were available in the dataset.
The primary outcome was monthly utilisation measured in defined daily doses (DDDs). Secondary outcomes included monthly prescription items, the market share of total SGLT2 inhibitor DDDs attributable to each drug, and net ingredient cost (NIC) per DDD.
A controlled interrupted time series (ITS) design was implemented using segmented regression. Models used Newey-West standard errors to account for potential autocorrelation and included month fixed effects to control for seasonality. Canagliflozin served as the comparator in the primary controlled ITS contrasts. Aggregate drug-level data limited analyses to population-level inferences rather than individual prescribing indications.
Overall utilisation of the three monotherapy SGLT2 inhibitors increased across the study period, but prescribing trajectories diverged markedly following July 2023. Market share of total SGLT2 DDDs shifted substantially: dapagliflozin rose from 33.2% to 55.9%, empagliflozin declined to 41.0%, and canagliflozin declined to 3.1%.
The controlled ITS model did not identify an immediate step change coinciding with July 2023. However, there was a significant differential increase in the postintervention slope for dapagliflozin relative to canagliflozin (β = 0.042, 95% CI 0.035 to 0.049). The source did not report the exact p-value associated with this estimate. Empagliflozin showed a smaller differential effect compared with canagliflozin.
A substantial reduction in dapagliflozin NIC was observed during the period studied, with NIC falling from £1.38 to £0.35 per DDD. The NIC change temporally coincided with acceleration in dapagliflozin prescribing.
The temporal pattern—marked acceleration in dapagliflozin prescribing after July 2023 against continued stable trends for canagliflozin—is consistent with an association between the guideline expansion and differential drug uptake. However, the aggregate ecological design prevents direct attribution of observed changes to prescribing for heart failure specifically; alternative indications such as type 2 diabetes or chronic kidney disease could also contribute to observed utilisation shifts.
The concurrent and large reduction in dapagliflozin NIC presents a plausible competing explanation. In exploratory analyses adjusting for NIC, the dapagliflozin estimate was attenuated by roughly one-third, but because NIC may be endogenous to prescribing and market forces, that attenuation cannot be interpreted as evidence of causal mediation. The more modest price change for empagliflozin and its smaller differential effect are noted as hypothesis-generating observations that require methods capable of disentangling guideline from price influences.
Key limitations reported by the authors include:
Aggregate (ecological) data: prescribing was analysed at the drug level without patient-level indications, so the analysis cannot distinguish whether increased utilisation was driven by heart failure or other conditions such as diabetes or chronic kidney disease.
Potential endogeneity of NIC: price and cost dynamics may both influence and be influenced by prescribing, limiting causal interpretation of adjustment analyses.
Incomplete statistical reporting in the source: the p-value reported for the dapagliflozin slope change was not provided in the source text excerpt.
Design limits on causal inference: the ITS with drug-level controls provides temporal associations but cannot definitively attribute changes to guideline publication given concurrent market and policy changes.
These findings indicate that class-wide guideline updates do not necessarily produce uniform adoption across individual drugs within the class. The pronounced shift toward dapagliflozin following July 2023 highlights the need for ongoing, drug-specific monitoring of guideline uptake in real-world prescribing data. Policy and formulary decisions should consider both clinical guideline recommendations and concurrent economic factors such as net ingredient cost, since price changes may materially influence adoption patterns.
Further research using patient-level linkage and designs that can separate indication-specific uptake from pricing or supply effects is needed to determine whether the observed changes reflect heart failure treatment adoption versus other clinical or market drivers.