Integrated glycaemic and inflammatory risk stratification for postoperative complications in diabetic patients undergoing cardiac surgery and interventions: a retrospective observational analysis
Abstract
Diabetic patients’ post-operative cardiac events are affected by several modifiable risk factors, with most prominence attributed to perioperative glycaemic variability and systemic inflammation. The proposed research seeks to examine the multivariate interactions between glycaemic control, pharmacologic interventions, and inflammatory markers in predicting post-operative complications. A retrospective observational analysis was carried out in 111 diabetic patients who received cardiac surgical or interventional procedures at Velavan Hospital, Tiruchengode, over a period of six months. Information was obtained from patient records such as demographic factors, comorbid conditions, pharmacologic treatment, glycaemic profiles, and inflammatory biomarkers (neutrophil-to-lymphocyte ratio [NLR], C-reactive protein [CRP]). Outcomes measured were atrial fibrillation (AF), acute kidney injury (AKI), ICU stay, readmissions, and mortality. The statistical analyses involved chi-square testing and multivariate logistic regression. Both hyperglycaemic (RBS ≥180 mg/dL) and inflamed (NLR >3.5) patients showed significantly increased incidences of AF (41.2%), AKI (32.4%), longer ICU stays (64.7%), and death (8.8%). Regression analysis also validated both glycaemic instability (OR: 2.74) and increased inflammation (OR: 2.96) as independent predictors of adverse outcomes. Pharmacologic analysis indicated that complication rates were lower in patients on metformin–vildagliptin compared to sulfonylurea or insulin. The additive effect of glycaemic and inflammatory burden is a potent predictor of post-operative complication in diabetic cardiac patients. Merging real-time glucose control with inflammatory monitoring and tailored pharmacologic regimens has the potential to significantly enhance clinical outcomes and diminish hospital burden.