Development of a risk score for acute kidney injury in infective endocarditis
J Gouveia Fiuza, G R M Ferreira, M Duarte Almeida, F Rodrigues Dos Santos, O Kungel, L Afonso Santos, J Gil, A CostaAbstract
Introduction
Acute kidney injury (AKI) affects a significant proportion of patients with infective endocarditis (IE) and is associated with increased mortality. Early identification of high-risk patients could enable targeted preventive strategies.
Purpose
To determine the prevalence of AKI in IE, identify baseline predictors, and develop a simple risk stratification score.
Methods
Retrospective single-centre cohort of 64 consecutive patients with blood culture-positive IE over a 5-year period. AKI was defined by KDIGO criteria. Baseline variables included demographics, comorbidities, risk scores and laboratory parameters. Multivariable logistic regression identified independent predictors. A simplified risk score was derived.
Results
Mean age was 69±12 years, 67% male. AKI occurred in 42 patients (65.6%). In-hospital mortality (IHM) was 25%. Patients with AKI were significantly older (72±12 vs. 62±9 years, p<0.001), had higher qSOFA (1.1±1.0 vs. 0.1±0.4, p<0.001), SOFA (5.1±3.5 vs. 2.8±2.1, p=0.007) and SHARPEN scores (11.4±2.0 vs. 8.4±1.9, p<0.001), and more often had prior heart failure (40.5% vs. 13.6%, p=0.045). Baseline creatinine, echocardiographic parameters, white blood cell and platelet counts, liver enzymes and vegetation size were similar between groups. In multivariable analysis, independent predictors were age (OR 1.11 per year, p=0.005) and qSOFA (OR 8.51 per point, p=0.008). Prior heart failure was not independently associated with AKI (OR 3.04, p=0.205). The AKI-IE Score incorporated age ≥70 years (1 point), prior heart failure (1 point; retained for clinical relevance and consistency with its univariable association) and qSOFA ≥1 (2 points). In this cohort, observed scores ranged from 0 to 3 points. The score demonstrated excellent discrimination (area under the curve 0.854, p<0.001) and stratified patients into three risk groups: low risk (score 0, 12% AKI incidence, n=17), high risk (score 1, 81% incidence, n=21; and score 2, 82% incidence, n=17) and very high risk (score 3, 100% incidence, n=9). For identifying patients requiring intensified prevention (score ≥2), specificity was 86% and positive predictive value 88%, with sensitivity 55%. IHM was higher in patients with AKI (33.3% vs. 9.1%, OR 5.0, p=0.038).
Conclusion
AKI affects two-thirds of IE patients and confers a 5-fold increase in IHM. The AKI-IE Score uses three readily available baseline variables to stratify patients into three distinct risk groups. This simple bedside tool may support rational allocation of preventive measures and intensive monitoring to high-risk patients.ROC Curve AKI-IEFor image description, please refer to the figure legend and surrounding text.