Summary: In this post-hoc sub-analysis of the open-label PROCEED trial, 57 adults with type 2 diabetes and chronic kidney disease were randomised to ipragliflozin or non-SGLT2-inhibitor standard therapy. Over 24 weeks the primary endpoint, change in the E/e’ ratio, did not differ between groups (group difference -0.82, 95% CI -2.44 to 0.81; P=0.317). Only exploratory subgroups with baseline LVEF >=60% or BMI >=25 kg/m2 showed a significant reduction.
PICO Summary
| Element | Detail |
|---|---|
| Population | 57 adults with type 2 diabetes and chronic kidney disease and paired echocardiography at baseline and 24 weeks; post-hoc sub-analysis of the multicentre, open-label, randomised PROCEED trial (Japan). |
| Intervention | Ipragliflozin, an SGLT2 inhibitor, added to standard care for 24 weeks (n=28). |
| Comparison | Non-SGLT2-inhibitor standard therapy, open-label control (n=29). |
| Outcome | Primary endpoint (change in E/e’ ratio) showed no between-group difference: -0.82 (95% CI -2.44 to 0.81; P=0.317). Exploratory subgroups: LVEF >=60% favoured ipragliflozin (-1.42 [-2.76 to -0.08]; P=0.038; interaction P=0.048); BMI >=25 kg/m2 favoured ipragliflozin (-1.95 [-3.56 to -0.34]; P=0.020; interaction P=0.016). No effect in LVEF <60% (n=7) or BMI <25 (n=9) subgroups. No effect modification by NT-proBNP, eGFR, or UACR. |
Ipragliflozin and LV diastolic function in diabetic CKD (PROCEED)
RCT sub-analysis · T2D + CKD · 24 weeks
The primary endpoint was null: ipragliflozin did not improve diastolic function (E/e') overall. Only exploratory LVEF >=60% and BMI >=25 subgroups favoured ipragliflozin, and these tiny, open-label strata are hypothesis-generating only.
Expert Commentary
The honest reading of this paper is a negative one: the pre-specified primary endpoint was not met, and ipragliflozin did not improve left ventricular diastolic function, measured by the E/e’ ratio, beyond standard therapy across the whole cohort. What remains are subgroup signals in patients with preserved ejection fraction or higher body mass index, and these must be treated as hypothesis-generating rather than confirmatory. The arms are tiny, with as few as seven and nine patients in the contrasting strata, so the interaction tests are fragile and vulnerable to chance, multiplicity, and unmeasured confounding. The parent PROCEED trial was open-label, which is a meaningful limitation here, because an unblinded design can bias an operator-dependent echocardiographic index even when measurement is intended to be objective. So, can I use this with my patients? Not yet. Nothing here changes prescribing, and it would be wrong to offer ipragliflozin specifically to improve diastolic function on the strength of a null trial with favourable post-hoc strata. The signal is biologically plausible and aligns with the broader SGLT2-inhibitor heart-failure story, but plausibility is not proof. What is needed is a prospectively powered, blinded trial that enrols patients with preserved ejection fraction and elevated body mass index and treats diastolic function as the primary outcome. Until that exists, these findings belong in the discussion section, not the clinic.
References
Teragawa H, Tanaka A, Takahashi K, Oshita C, Uchimura Y, Kamei N, et al. Impact of baseline left ventricular ejection fraction and body mass index on the effect of 24-week ipragliflozin treatment on left ventricular diastolic function in patients with type 2 diabetes and chronic kidney disease: insights from the PROCEED trial. Cardiovasc Diabetol. 2025;24(1):190. doi:10.1186/s12933-025-02745-1