Blood urea nitrogen (BUN) independently predicts mortality in critically ill patients admitted to ICU: A multicenter study

Abstract

Background and purpose: The Microcirculatory Shock Occurrence in Acutely Ill Patients (micro-SOAP) study investigated associations of microcirculation and mortality. Risk stratification in critically ill patients is of utmost interest. Established score such as APACHE2 (Acute Physiology And Chronic Health Evaluation 2) are relatively complex and might therefore be of limited use. Blood urea nitrogen (BUN) was described to be associated with mortality in various diseases. We therefore aimed (i) to evaluate BUN for prediction of mortality in a cohort of critically ill patients and (ii) to investigate associations of BUN with microcirculation.

Methods: 412 patients were included in our post-hoc analysis of the prospective multicenter microSOAP study. Assesment of the sublingual microcirculation (Sidestream Dark Field (SDF) imaging) and collection of laboratory values were performed on the same day in this point prevalence study. Evaluation of associations with mortality was done by logistic regression analysis. An optimal BUN cut-off was calculated by means of the Youden Index.

Results: Median BUN was 9.0 mmol/L. BUN was associated with in-hospital-mortality in a logistic regression analysis (HR 1.03; 95% CI 1.01-1.05; p < 0.001). Per quartile (BUN 0-5.4 mmol/L, 5.4-9.0 mmol/L, 9.0-15.9 mmol/L and above 15.9 mmol/L) in-hospital mortality increased by as much as 51% (HR 1.51; 95% CI 1.23-1.85; p < 0.001). ROC analysis was done (AUC 0.63 95% CI 0.58-0.67) and the statistically optimal cut-off calculated by means of the Youden Index: 9.7 mmol/L. This cut-off was associated with a significant 3-fold increase in mortality (HR 2.97 95% CI 1.88-4.70; p < 0.001) and remained robustly associated with adverse outcome after correction for APACHE2 (HR 2.71 95% CI 1.61-4.59; p < 0.001), renal function as expressed by creatinine (HR 2.63 95% CI 1.59-4.33; p = 0.001), as well in an integrative model (MAP<60 mmHg, tachycardia (heart rate >90/min), lactate above 1.5 mmol/L, age above 80 years; HR 2.43 95% CI 1.50-3.92; p < 0.001). Parameters of microvascular perfusion were associated neither with BUN nor mortality.

Conclusions: BUN is associated with hospital mortality and a combination of BUN and clinical signs might constitute a powerful but easy-to-use tool for risk stratification in critically ill patients and help improve their outcome. BUN was not associated with parameters of microcirculation which were not associated with mortality.

Keywords: BUN; Critically ill; ICU; microcirculation; risk score; risk stratification.