Not All Albumin Ratios Are Equal: A Comparative Analysis of Four Biomarkers in Predicting In-Stent Restenosis After ST-Elevation Myocardial Infarction

Abstract

This study aimed to comparatively evaluate the predictive value of four albumin-based biomarker ratios — C-reactive protein-to-albumin ratio (CAR), platelet-to-albumin ratio (PAR), calcium-to-albumin ratio (CALR), and uric acid-to-albumin ratio (UAR) — for in-stent restenosis (ISR) in patients undergoing primary percutaneous coronary intervention (PCI) for ST-elevation myocardial infarction (STEMI). A total of 608 STEMI patients who underwent primary PCI between January 2020 and January 2025 were retrospectively analyzed; the median follow-up was 59.3 months (IQR: 52.5–67.0), and ISR developed in 63 patients (10.4%). The association between the four ratios and restenosis was assessed using univariable and multivariable logistic and Cox regression analyses, receiver operating characteristic (ROC) curve analysis, and restricted cubic spline (RCS) modeling. All four biomarkers were significantly higher in the ISR (+) group and showed significant discriminative ability on ROC analysis (UAR AUC=0.741, CALR AUC=0.710, CAR AUC=0.632, PAR AUC=0.614). All four ratios were independent predictors in multivariable logistic regression; however, in multivariable Cox regression, CAR lost statistical significance (HR=1.28, p=0.071), whereas UAR (HR=2.93, p<0.001), CALR (HR=2.97, p<0.001), and PAR (HR=1.16, p=0.010) remained independent predictors. RCS analysis revealed nonlinear threshold-type relationships for CAR, CALR, and UAR, whereas PAR demonstrated a linear dose-response relationship. In conclusion, CALR and UAR emerged as strong and consistent independent predictors of ISR risk in both cross-sectional and long-term analyses, whereas the predictive value of CAR was largely confined to the cross-sectional assessment