MAF-5 score and cT1 for cardiovascular risk: evidence from UK Biobank
J Yang, H Zhou, S J Wang, X Y Zheng, J P WengAbstract
Background
Metabolic dysfunction-associated steatotic liver disease is a recognized yet underappreciated cardiovascular risk factor. The MRI-derived iron-corrected T1 (cT1) provides a quantitative measure of hepatic fibro-inflammation that is associated with adverse cardiometabolic and cardiac structural phenotypes. The MAF-5 score—a low-cost, non-invasive tool integrating waist circumference, body mass index, diabetes status, aspartate aminotransferase, and platelet count—outperform at detecting liver fibrosis in the general population. Little is known about the relationship between MAF-5 and cT1 and their joint association with cardiovascular disease (CVD).
Purposes
The aim of the present analysis was to investigate whether cT1 modulates the association between MAF-5 and incident CVD in the general population.
Methods
Participants from UK Biobank without prior CVD with complete MAF-5 and cT1 data were enrolled. MAF-5 was assessed at the initial visit and cT1 at the imaging visit. The primary outcome was composite incident CVD, including coronary artery disease, atrial fibrillation, heart failure, stroke, and venous thromboembolism after the imaging visit. Landmark-adjusted Cox proportional hazards models were utilized to assess the associations of MAF-5 and cT1 with the primary outcome, with full adjustments for demographics, lifestyle, systolic blood pressure, low-density lipoprotein cholesterol, C-reactive protein, proton density fat fraction, estimated glomerular filtration rate, and prior medication use. To evaluate joint association, participants were stratified into four groups based on median MAF-5 (-1.09) and cT1 (691 ms): Group 1 (low MAF-5/low cT1), Group 2 (high MAF-5/low cT1), Group 3 (low MAF-5/high cT1), and Group 4 (high MAF-5/high cT1). Model performance was assessed using likelihood ratio test. Statistical analyses were performed using R (version 4.3.1).
Results
A total of 27,052 participants (mean age 64 years; 45.3% male) were included. Over a median follow-up of 6.48 years, per 1-SD increase in MAF-5 was associated with a 9% higher risk of incident CVD (HR 1.09, 95% CI 1.02–1.17, P=0.006), while per 1-SD increase in cT1 was associated with an 18% higher risk (HR 1.18, 95% CI 1.10–1.27, P<0.001) (Figure 1). Survival analysis showed significant difference across four groups, with Group 4 exhibiting the highest hazard compared to Group 1 (HR 1.48, 95% CI 1.23–1.78, P<0.001) (Figure 2). No statistically significant interaction was observed between MAF-5 and cT1 (P_interaction = 0.206). Integrating both markers significantly improved model performance compared to either marker alone (P < 0.001).
Conclusion
In the general population, the MAF-5 score demonstrated the independent and additive prognostic value to cT1 for incident CVD. Combination of MAF-5 score and cT1 may enhance the prediction of cardiovascular risk.Landmark Cox model for incident CVDCumulative incidence of incident CVD