[Predictive value of SOFA score combined AGI grading system in elderly patients with sepsis: a retrospective analysis...

To explore the predictive value of sequential organ failure assessment (SOFA) score combined the acute gastrointestinal injury (AGI) grading system in critical elderly patients with sepsis.

A retrospective analysis was conducted. Elderly patients with sepsis aged > 60 years admitted to medical intensive care unit (MICU) of General Hospital of Guangzhou Military Command from March 2014 to December 2015 and experiencing critical care over 48 hours were enrolled. Age, gender, acute physiology and chronic health evaluation II (APACHE II) score, sequential organ failure assessment (SOFA) score and AGI score at admission (SOFA_initial, AGI_initial), the highest SOFA score and AGI score within the first week (SOFA_max, AGI_max), serum procalcitonin (PCT), C-reactive protein (CRP), albumin (ALB), platelet (PLT), hemoglobin (Hb) and lactate (Lac) levels, length of ICU stay, usage of mechanical ventilation and renal replacement therapy were recorded. The primary end point was 28-day mortality. To extract factors affecting 28-day mortality, the risk factor of death of the senile sepsis patients were analyzed by binary logistic regression analysis (stepwise method). Fitness of the model was assessed by the Hosmer-Lemeshow test and calibration plot (P > 0.05). Receiver operating characteristic (ROC) analysis was performed for APACHE II score, SOFA_initial score, SOFA_max score, AGI_initial score, AGI_max score and SOFA_max and combined AGI_max score.

Ninety-one patients were enrolled, the incidence of AGI in elderly patients with sepsis was 100%; 34 patients died 28 days after the admission, and the 28-day mortality rate was 37.4%. Non-survivors presented a higher APACHE II score, SOFA_initial score, SOFA_max score, AGI_initial score, AGI_max score and longer usage of mechanical ventilation and renal replacement therapy. SOFA_max score [odds ratio (OR) = 1.576] and AGI_max score (OR = 5.695) were associated with 28-day mortality in binary logistic regression analysis (both P < 0.01). The area under the curve (AUC) and 95% confidence interval (95%CI) of SOFA_max score combined AGImax score was significantly higher than that of SOFA_initial score, SOFA_max score, AGI_initial score, AGI_max score and APACHE II score [0.806 (0.710-0.881) vs. 0.723 (0.619-0.812), 0.786 (0.688-0.865), 0.641 (0.533-0.739), 0.633 (0.526-0.881), 0.638 (0.531-0.736), all P < 0.05]. The Youden index (55.37) and positive predict value (5.51) of SOFA_max score combined AGI_max score were the largest. When its cut-off value reached 14, the sensitivity and specificity was 67.65% and 87.72%, respectively. According to score of APACHE II, SOFA_initial, SOFA_max or AGI_max, the higher of each score, the higher mortality rate could be.

The combination of SOFA_max score and AGI_maxscore enable accurate prediction in elderly patients with sepsis.