• Jia-hui Dong, Jie Sun, Rui Chen, Jian Liu, An Zeng, Yun Chen, Qiao Jiang, Ri-cheng Xiong, Fei Xiao, Zhou Yu, and Zhen-hui Guo.
• Department of Critical Care Medicine, General Hospital of Guangzhou Military Command, Guangzhou 510010, Guangdong, China, Corresponding author: Guo Zhen-hui, Email: irwind@163.com.
• Zhonghua Wei Zhong Bing Ji Jiu Yi Xue. 2013 Oct 1;25(10):594-9.

ObjectiveTo assess and compare the performance of acute physiology and chronic health evaluation II/IV (APACHEII/IV) prognostic models in elderly patients with sepsis.MethodsA totally of 82 elderly patients with sepsis were retrospectively assessed in geriatric intensive care unit of General Hospital of Guangzhou Military Command between July 2011 and December 2011. APACHEII/IV scores were recorded within 24 hours after admission. The prognosis accuracy of both scores was assessed by area under the receiver operating characteristic curve (AUC). Based on the best cutoff value corresponding with the highest accuracy, patients were divided into the low and high risk of hospital mortality group. The predictive power of APACHEII/IV in total population and subgroups was compared.ResultsPatients with severe sepsis constituted 57.3% (47/82) of all patients with sepsis, and hospital mortality was 61.0% (50/82). APACHEII/IV scores of the patients were 17.5±6.3 and 55.8±22.3, and mortality rate was 22.5% (18.4/82) and 17.9% (14.6/82) respectively, with significant differences compared with actual mortality (both P<0.01). Both APACHEII/IV scores showed underestimation of hospital mortality in total population [standardized mortality ratio (SMR) with APACHEII=2.71, 95% confidence interval (95%CI) 1.92-3.48 and SMR with APACHEIV=3.33, 95%CI 2.79-4.37]. APACHEII (AUC 0.664±0.066), and APACHEIV presented poor estimation(AUC 0.716±0.056). There was no difference in accuracy in prognosticating hospital death prognosis between the two APACHE models (Z=0.991, P=0.322). Cutoff values of APACHEII/IV were >11 and >59. According to the value, patients were divided into the low and high risk hospital mortality group. There was no significant difference between actual mortality and prognostic mortality in APACHEII low risk group [0-11, 20.0% (3/15) vs. 1.6% (0.2/15), Z=-1.023, P=0.306]. The actual mortality in high risk group with APACHEII (>11) was significantly higher than prognostic mortality [70.1% (47/67) vs. 27.2% (18.2/67), t=6.989, P=0.000]. In the high risk group, APACHEII underestimated mortality (SMR=2.58, 95%CI 2.22-3.51). The actual mortality of the low (0-59) and high (>59) risk group of APACHEIV were higher than prognostic mortality [lower risk group: 44.0% (22/50) vs. 7.5% (3.8/50), Z=-2.235, P=0.025; higher risk group: 87.5% (28/32) vs. 34.1% (10.9/32), Z=-4.712, P=0.000]. Two groups of patients with APACHEIV score, the mortality was underestimated (low risk group: SMR=5.90, 95%CI 5.19-7.07; high risk group: SMR=2.56, 95%CI 2.07-3.24). Mortality rate of the low risk group with APACHEIV score was prone to be underestimated.ConclusionsThe accuracy of APACHEII/IV are not ideal in foretelling mortality rate. Hospital mortality was underestimated with APACHEII in high risk patients, and it was underestimated with APACHEIV both in low and high risk patients, and it is even more prone to be underestimated in low risk group of APACHEIV. More accurate prognostic modality is in need in elderly patients with sepsis.

12 keywords...

hide…