Cardiac fibrosis is a key pathological substrate that drives diastolic dysfunction, arrhythmogenesis, and heart failure progression across a spectrum of cardiometabolic disorders. Sodium-glucose cotransporter 2 (SGLT2) inhibitors, initially developed for glucose lowering, have demonstrated pleiotropic effects on myocardial structure, notably attenuating fibrotic remodeling. Experimental models of diabetes, hypertension, ischemia, and cardiotoxicity consistently show that SGLT2 inhibitors mitigate interstitial and perivascular fibrosis through modulation of oxidative stress, mitochondrial function, autophagy, and canonical profibrotic signaling cascades, including TGF-β/Smad, STAT3, and mTOR. These actions are largely preserved in non-diabetic settings and appear to extend beyond hemodynamic or glycemic benefits. Clinical data, including cardiac magnetic resonance-based assessments, support the notion of diffuse fibrosis regression, particularly in heart failure with preserved ejection fraction and diabetic cardiomyopathy. Moreover, reductions in serum collagen biomarkers and improvements in myocardial energetics further substantiate their antifibrotic capacity. Nonetheless, fibrosis-specific endpoints remain underrepresented in major cardiovascular outcome trials, and histological validation in human tissue is lacking. Integrating artificial intelligence-driven fibrosis quantification, spatial transcriptomics, and high-resolution imaging may refine phenotyping and enable precision antifibrotic therapy. Whether fibrosis regression translates into durable clinical benefit remains an open question. This review comprehensively synthesizes the mechanistic, translational, and clinical evidence supporting the role of SGLT2 inhibitors as modulators of cardiac fibrosis across diverse cardiovascular disease states.
Background: Orthotopic Cardiac Transplantation (OCTx) improves survival in advanced heart failure. Currently, a tool in United Kingdom from NHS Blood and Transplant (NHSBT) helps predict likelihood of OCTx from waitlist. However, it does not use predictive variables such as age, or Human Leukocyte Antibody (HLA%). We aimed to develop OCTx predictive models incorporating known prognostic variables at 3-, 6-, 9- and 12-months.
Methods: All patients who were urgent-listed for OCTx at Harefield Hospital between 2014 and 2018 (n = 125) were analysed. Variables included age, gender, blood group (BG), midline sternotomy, ventricular assist device (VAD), body mass index (BMI) and HLA%. Multivariable logistic regression models were constructed following internal validation per timepoint. A separate validation dataset was collected using 52 patients transplanted between 2019 and 2023, to compare model effectiveness against the current NHSBT tool.
Results: At 3-months, variables included were age, gender, sternotomy, BG O and HLA%=0, with model area under curve (AUC) of 0.74 (0.66-0.83 95 % confidence interval [CI]). 6-month model included variables age, gender, BG O, sternotomy, BMI and HLA%=0, model AUC of 0.80 (0.72-0.89 95 % CI). 9-month model used age, BG O, VAD, BMI and HLA%=0, giving an AUC of 0.80 (0.71-0.89 95 % CI). The final 12-month model included midline sternotomy, BMI and HLA%=0 and HLA%=1-24, with AUC 0.78 (0.68-0.88 95 % CI). Our predictive models recorded an 85 % win-ratio compared to the NHSBT tool.
Conclusion: We were able to develop models to predict urgent OCTx, with greater accuracy than the currently available tool. Multicentre external validation would help enable its wider implementation.

