CHA₂DS₂-VASc Score for Pre-Procedural Prediction of Contrast-Associated Acute Kidney Injury After Primary Percutaneous Coronary Intervention for ST-Elevation Myocardial Infarction: A Head-to-Head Comparison with the Mehran Score

Technical Summary: CHA₂DS₂-VASc vs. Mehran Score for CA-AKI Prediction in STEMI

Problem Statement
Contrast-associated acute kidney injury (CA-AKI), formerly known as contrast-induced nephropathy (CIN), is a significant complication of primary percutaneous coronary intervention (PPCI) for ST-elevation myocardial infarction (STEMI), occurring in 6–25% of cases and independently increasing morbidity and mortality. Effective risk stratification is critical for guiding pre-procedural strategies such as contrast volume minimization and fluid loading. However, established tools like the Mehran score require post-procedural data (e.g., contrast volume, post-procedural creatinine) that are unavailable at the moment of decision-making before angiography. In resource-limited settings, particularly in South Asia where patients often present without prior biochemical workup, there is a lack of validated pre-procedural risk prediction tools that rely solely on clinical history.

Methodology
This study employed a single-centre, prospective, observational cohort design conducted at Shahid Gangalal National Heart Centre in Kathmandu, Nepal, between February and September 2024.

• Population: 300 consecutive adult patients (aged 18–80) with STEMI undergoing successful PPCI.
• Exclusion Criteria: Cardiogenic shock, chronic kidney disease (CKD) stage ≥3b (eGFR <45 mL/min/1.73 m²), and baseline creatinine >3.0 mg/dL.
• Definitions: CA-AKI was defined as an absolute increase in serum creatinine of ≥0.5 mg/dL or a relative increase of ≥25% at 48–72 hours post-procedure.
• Intervention/Measurement:
  • The CHA₂DS₂-VASc score was calculated pre-angiography using only clinical history and physical examination. Notably, the "Vascular disease" component was scored as 1 for all patients (due to established CAD) and excluded from variable-level regression to avoid collinearity, though it remained part of the aggregate score.
  • The Mehran score was calculated post-procedurally using all eight standard variables, including contrast volume.
• Statistical Analysis: Group differences were assessed using Student's t-test, Mann–Whitney U test, and Fisher's exact test for all 2×2 categorical comparisons. Univariable logistic regression was used for predictors (multivariable modeling was avoided due to the low event count of 32). Discriminatory performance was evaluated via Receiver Operating Characteristic (ROC) curve analysis, with AUC comparison performed using DeLong's method.

Key Results

• Incidence: CA-AKI occurred in 32 of 300 patients (10.7%).
• Score Comparison: The mean CHA₂DS₂-VASc score was significantly higher in the CA-AKI group (3.24 ± 1.45) compared to the non-CA-AKI group (1.48 ± 1.21; p < 0.001). Similarly, the Mehran score was higher in the CA-AKI group (8.1 ± 2.3 vs. 4.7 ± 1.9; p < 0.001).
• Diagnostic Performance:
  • CHA₂DS₂-VASc: Achieved an Area Under the Curve (AUC) of 0.734 (95% CI 0.664–0.804). At a cut-off of ≥3, it demonstrated 87.5% sensitivity and a Negative Predictive Value (NPV) of 97.3%, with a specificity of 53.7%.
  • Mehran Score: Achieved an AUC of 0.752 (95% CI 0.685–0.819). At a cut-off of ≥6, it showed 71.9% sensitivity and 69.8% specificity.
• Statistical Significance: The difference in AUC between the two scores was not statistically significant (p = 0.345).
• Regression: Univariable analysis identified age ≥65, diabetes, congestive heart failure, prior stroke/TIA, Killip class III, CHA₂DS₂-VASc ≥3, and Mehran ≥6 as significant predictors of CA-AKI.

Key Contributions

1. Pre-Procedural Utility: The study demonstrates that the CHA₂DS₂-VASc score, derived entirely from clinical history, offers predictive accuracy for CA-AKI comparable to the Mehran score, which requires post-procedural data.
2. Resource-Limited Applicability: The findings validate a tool that can be calculated in under a minute at the bedside without laboratory data, addressing a critical gap in settings where biochemical workups are incomplete prior to catheterization.
3. Methodological Rigor: The analysis utilized Fisher's exact test for categorical comparisons and DeLong's method for AUC comparison, while appropriately limiting statistical modeling to univariable analysis given the event count (n=32), avoiding the instability of multivariable models.
4. Geographic Extension: This represents the first such analysis from Nepal, extending previous findings from Egyptian and Turkish populations to a South Asian context.

Significance and Claims
The paper claims that the CHA₂DS₂-VASc score is a clinically usable alternative to the Mehran score for pre-procedural risk stratification in STEMI patients. Its high NPV (97.3% at a cut-off of ≥3) allows clinicians to reliably "rule out" high-risk patients before the procedure, streamlining monitoring for low-risk individuals. Conversely, a score ≥3 prompts practical, low-burden interventions—such as minimizing contrast volume, considering iso-osmolar agents, and deferring non-culprit interventions—without delaying revascularization.

The authors emphasize that while the Mehran score remains valuable for post-procedural prognostication, the CHA₂DS₂-VASc score fills the specific need for pre-procedural decision-making in resource-constrained environments. The study concludes that the score's performance is comparable to Mehran's, but its availability at the moment of STEMI diagnosis makes it uniquely suited for guiding contrast strategy before the catheterization laboratory is activated.

Limitations Noted by Authors
The authors modestly acknowledge that the study is single-centre, limiting generalizability. The exclusion of cardiogenic shock and advanced CKD means findings cannot be extrapolated to the highest-risk populations. The small number of CA-AKI events (32) precluded multivariable adjustment, and the optimal cut-off was derived without bootstrap validation, suggesting sensitivity and specificity estimates may be optimistic and require external confirmation. Long-term renal outcomes were not captured.