Reviews/SGLT2 Inhibitors/SGLT2 inhibitors reduce hospitalisation for heart failure in T2D
· SG-CV-03 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
This claim asks whether SGLT2-inhibitor drugs cut the number of times people with type 2 diabetes end up in hospital for heart failure. The evidence here points the same way from several different angles: a randomised comparison against placebo in diabetic patients [3], a randomised trial in diabetic patients hospitalised for worsening heart failure [6], and several large real-world studies tracking patients in everyday care [11][12][14][18][22]. All of them show fewer heart-failure hospitalisations with these drugs, and the size of the benefit is broadly similar across studies and across different drugs in the class. The main caveat the panel noted is that the cleanest heart-failure-hospitalisation result that is specific to diabetes comes either from a small genetic sub-analysis of one trial or from real-world data, rather than from a dedicated placebo-controlled trial of this exact outcome in this exact population. Because the different study types agree so consistently, the panel treated the claim as well supported while flagging that the strongest purpose-built randomised evidence for this precise question is thinner than the drugs' reputation suggests.
Randomised evidence in a type 2 diabetes population shows dapagliflozin lowering hospitalisation for heart failure versus placebo (noncarrier hazard ratio 0.70, 95% CI 0.57-0.86) [3], and sotagliflozin lowering a composite of cardiovascular death and heart-failure hospitalisations and urgent visits (hazard ratio 0.67, 95% CI 0.52-0.85) in diabetic patients with recent worsening heart failure [6]. Large observational analyses converge on the same direction: a heart-failure admission hazard ratio of 0.70 (0.54-0.92) for canagliflozin versus a DPP-4 inhibitor [14], hospitalisation-for-heart-failure incidence rate ratios of 0.29 and 0.17 in a self-controlled case series [22], a cardiovascular-death or first heart-failure-hospitalisation hazard ratio of 0.70 (0.52-0.95) in the Swedish registry [11], a 16-56% risk reduction in a propensity-matched cohort without established cardiovascular disease [12], and an on-treatment heart-failure hazard ratio of 0.54 (0.30-0.99) [18]. Effect sizes cluster in the 0.5-0.7 range across designs and across individual agents. The contested point: The division is not about direction but about the type of evidence carrying the type 2 diabetes-specific, heart-failure-hospitalisation-isolated result. The dedicated placebo-controlled trials in the corpus [7][8] enrol heart-failure-defined populations with and without diabetes rather than a diabetes-restricted cohort, so the cleanest diabetes-specific signal rests on a DECLARE genetic substudy arm [3] and on observational data [11][12][14][18][22].
All 22 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from SGLT2 Inhibitors, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The cleanest HHF reduction in a pure T2D randomised setting comes from a DECLARE genetic substudy noncarrier arm [3], not a headline trial result reported in this corpus; the two dedicated SGLT2i-vs-placebo RCTs [7][8] are HF-defined populations with and without diabetes, not T2D-restricted.
Most T2D-specific HHF evidence [11][12][14][18] is observational and vulnerable to confounding-by-indication and healthy-adherer bias despite propensity matching; SGLT2i initiators were younger with better renal function in [11].
Where it remains uncertain. The purest diabetes-specific heart-failure-hospitalisation signal is observational or a substudy; the dedicated placebo-controlled trials [7][8] are heart-failure populations with and without diabetes, and the randomised diabetic trial [6] uses a composite endpoint.
Aleatory
The carrier arm of [3] rests on 121 variant carriers, giving a wide interval (hazard ratio 0.18, 95% CI 0.04-0.86) and leaving the diabetes-specific randomised estimate sensitive to chance.
The load-bearing papers from a corpus of 22, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Randomised T2D evidence (dapagliflozin vs placebo) of HHF reduction; noncarrier HR 0.70 represents the general T2D population.
“dapagliflozin lowered the risk of HHF more strongly in carriers (hazard ratio 0.18, 95% confidence interval 0.04-0.86) than in noncarriers (hazard ratio 0.70, 95% confidence interval 0.57-0.86; P interaction 0.03)”
Panel auditInternal validity: high · scope of inference: Adults with T2D and increased cardiovascular risk; carrier analysis is exploratory in a small subgroup.Key limitations: carrier subgroup very small (n=121) · genetic analysis post hoc
Randomised T2D trial; SGLT2 inhibition reduces the composite dominated by HF hospitalisations and urgent visits.
“The rate (the number of events per 100 patient-years) of primary end-point events was lower in the sotagliflozin group than in the placebo group (51.0 vs. 76.3; hazard ratio, 0.67; 95% confidence interval [CI], 0.52 to 0.85; P<0.001).”
Panel auditInternal validity: medium · scope of inference: T2D patients recently hospitalised for worsening heart failure; sotagliflozin is a combined SGLT1/2 inhibitor.Key limitations: composite endpoint · trial stopped early · SGLT1/2 inhibitor not pure SGLT2
Real-world T2D+HF registry corroborating reduced HF hospitalisation.
“SGLT2i use was associated with a 30% lower risk of cardiovascular (CV) death/first HF hospitalisation (hazard ratio 0.70, 95% confidence interval 0.52-0.95), which was consistent regardless of ejection fraction, background metformin treatment and kidney function”
Panel auditInternal validity: medium · scope of inference: Contemporary Swedish HF patients with T2DM in routine care.Key limitations: observational · HHF within composite endpoint · small treated group
The primary outcome is a composite of all-cause death, MI, stroke or heart failure over 3 years and HHF is not reported separately, so the direction for HHF specifically cannot be read.
Panel auditInternal validity: medium · scope of inference: Working-age Japanese T2D patients, many at low CVD risk, for a composite cardiovascular outcome.Key limitations: HHF not isolated from composite · observational · young low-risk population
4 of 22 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Convergent randomised and observational T2D evidence all reduce HHF in the same direction, held just above threshold because the purest T2D HHF signal is observational while the randomised components are a substudy and a composite endpoint.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Dedicated diabetes-restricted RCT of heart-failure hospitalisation. The cleanest diabetes-specific heart-failure-hospitalisation evidence is a substudy [3] or observational [11][12][14][18][22]; no corpus trial tests this outcome as a primary endpoint in a diabetes-only population against placebo. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A placebo-controlled randomised trial enrolling only type 2 diabetes patients without prevalent heart failure, with hospitalisation for heart failure as the primary endpoint.
Open question: Does SGLT2 inhibition reduce hospitalisation for heart failure in type 2 diabetes patients who do not already have heart failure, measured as a stand-alone primary endpoint? Why it is still open: Corpus randomised placebo evidence in diabetes is a genetic substudy [3] and a composite endpoint in an established-heart-failure population [6]; the primary-prevention, isolated-outcome question is not directly tested.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 23 Oct 2026 · protocol v0.19