Abstract
BACKGROUND: Geriatric nutritional risk index (GNRI) is a simple and useful nutritional marker for predictor of adverse outcomes in patients undergoing a variety of conditions. This study explored the relationship between GNRI and adverse outcomes of Pyogenic Liver Abscess (PLA) patients and assessed GNRI predictive value. METHODS: This was one retrospective study involving 240 PLA patients. According to one GNRI cutoff value of 90, the patients were divided into two groups. Besides, demographic, laboratory, adverse outcomes were compared between the two groups. Multivariate logistic regression analyses and receiver-operating characteristic (ROC) curve analysis were performed. RESULTS: Compared with high GNRI patients, those with low GNRI had a higher risk of mortality (13.4% vs. 2.0%, P = 0.002), metastatic infection (17.7% vs. 8.2%, P = 0.050), acute hepatic failure (6.8% vs. 1.0%, P = 0.036), acute respiratory failure (7.4% vs. 1.0%, P = 0.024), upper gastrointestinal (UGI) bleeding (11.9% vs. 2.1%, P = 0.006) and empyema (20.1% vs. 10.2%, P = 0.047). Multivariate logistic regression analysis demonstrated GNRI (< 90) as one independent factor in death prediction (odds ratio (OR) = 5.36, 95% of confidence interval (CI) = 1.17-24.48), and adverse outcomes (OR = 2.04, 95% CI = 1.05-3.98). GNRI had the largest area under receiver operating characteristic (ROC) curve than albumin, BMI, platelet, prothrombin time and hemoglobin in death prediction (area under ROC curves (AUC) = 0.771, cutoff value = 79.45, P < 0.01) and all adverse outcomes (AUC = 0.656, cutoff value = 87.43, P < 0.01). CONCLUSIONS: Lower levels of GNRI are an independent risk factor for poor PLA prognosis. Physicians should consider GNRI for PLA outcomes and consider more careful resuscitation and timely and appropriate treatment, especially in those with GNRI< 87.43.