Mortality prediction at admission to intensive care: A comparison of microalbuminuria with acute physiology scores after 24 hours

OBJECTIVETo compare low level albumin excretion (microalbuminuria), a marker of systemic capillary permeability, with mortality, Acute Physiologic And Chronic Health Evaluation (APACHE II) score, the Simplified Acute Physiologic (SAP II) score, and their derived mortality probabilities in patients admitted to a general intensive care unit. DESIGNProspective observational study. SETTINGA 14-bed intensive care unit in a university teaching hospital. PATIENTSA total of 140 consecutive patients (59 surgical, 48 medical, 22 trauma, and 11 burns). INTERVENTIONSUrine collection within 15 mins of intensive care unit admission for assessment of microalbuminuria. MEASUREMENTS AND MAIN RESULTSMicroalbuminuria, expressed as the albumin-creatinine ratio (ACRnormal, 5.9 mg/mmol gave a sensitivity for death of 100%, specificity of 59%, positive predictive value of 25%, and negative predictive value of 100%. Mortality probability receiver operator characteristic curve areas for ACR, APACHE II, and SAP II were 0.843 (p < .0001), 0.793 (p = .0004), and 0.770 (p = .0017), respectively. ACR was associated with intensive care unit stay (p = .0021) and highest serum C-reactive protein (p = .0002), serum creatinine (p < .0001), and bilirubin (p = .0009). For 48 medical patients, of whom 34 survived, admission ACRs for survivors and nonsurvivors were 8.3 (5.7–10.8) and 10.7 (4.1–48.2) mg/mmol, respectively (p = .32). SAP II, but not APACHE II, score was significantly higher for nonsurvivors. CONCLUSIONSFor surgical, trauma, and burn patients, but not medical patients, microalbuminuria within 15 mins of intensive care unit admission predicted death as well as APACHE II and SAP II scores calculated after 24 hrs, and it shows promise as a predictor of outcome.