应用简化急性生理评分Ⅱ预测危重神经疾病患者死亡概率

Predicting possibility of mortality in critically ill patients with neurological diseases by using Simplified Acute Physiology Score Ⅱ

目的 评估简化急性生理评分（SAPS）Ⅱ预测危重神经疾病患者死亡概率的可行性和可靠性.方法 纳入2005年1月至2007年12月神经内科重症监护病房（N-ICU）患者653例,记录入院后0、24、48、72 h的SAPSⅡ分值,并计算医院死亡概率（PHM）.疾病病种分为脑出血、脑梗死、中枢神经系统感染、神经肌肉疾病和其他神经疾病5类.经两样本均数比较的t检验比较生存组与死亡组SAPSⅡ分值,经单因素回归分析判断不同时间点SAPSⅡ分值与预后的关系;经校准度（Hosmer-Lemeshow拟合优度检验）和分辨度（ROC曲线下面积）分析判断SAPSⅡ性能;经多因素Logistic回归分析判断神经疾病种类与预后的关系.结果 不同时间点死亡组SAPSⅡ分值（分别为46.53±12.22、47.28±13.84、48.58±14.18和49.06±14.61）明显高于生存组分值（分别为34.70±11.78、30.28±12.24、29.79±12.36和29.69±12.96,t=11.12、14.02、14.43和13.49,均P＜0.01）;不同时间点SAPSⅡ分值与预后均相关（OR值分别为1.080、1.100、1.109和1.100,P＜0.01）;不同时间点SAPSⅡ分值Hosmer-Lemeshow拟合优度检验结果为校准度均较高,分别x2=5.305、7.557、6.369和8.540（P＞0.05）;48和72 h分值预测病死率与实际病死率的总一致率分别为82.6%和83.4%,ROC曲线下面积分别为0.825和0.847;多因素分析疾病病种与预后无关.结论 SAPSⅡ适用于N-ICU危重症神经疾病患者的病情评估和死亡概率预测,48和72 h评估最为准确,且不受疾病病种影响.

Objective To evaluate feasibility and reliability of using Simplified Acute Physiology Score Ⅱ （SAPS Ⅱ）in predicting mortality in critically ill patients with neurological diseases.Methods All 653 patients hospitalized in neurological intensive care unit （N-ICU） from Jan 2005 to Dec 2007 were retrospectively studied.SAPSⅡ scores were scaled upon admission at 24,48 and 72 h,and possibility of hospital mortality （PHM） was calculated based on SAPS Ⅱ score.Neurological diseases diagnosis made initially at time of hospitalization was classified into five categories：intracranial hemorrhage,cerebral infarction,neurologic infection,neuromuscular diseases and other neurologic diseases.At each of four time points,the SAPSⅡscores were compared between the survival group and death group,and the relationship of SAPS Ⅱ score and prognosis were analyzed.The calibration of the SAPS Ⅱ were accessed with the Hosmer-Lemeshow goodness-of-fit chi-squared statistic and the discrimination with area under the receiver operating characteristic curve （AUROC）.Multivariate logistic regression was used to analyze the relationship between disease category and prognosis.Results SAPS Ⅱ scores in the death group （46.53±12.22,47.28±13.84,48.58±14.18 and 49.06±14.61）at each time point were significantly higher than those in the survival group （34.70±11.78,30.28±12.24,29.79±12.36 and 29.69±12.96;t=11.12,14.02,14.43 and 13.49 at 0,24,48 and 72 h,respectively,P〈0.01）.Furthermore,univariate logistic regression analyses demonstrated that SAPS Ⅱ score was correlated with prognosis （OR=1.080,1.100,1.109,1.100 at 0,24,48 and 72 h,respectively,P〈0.01）.The scores at 48 and 72 h were more accurate in predicting mortality.SAPS Ⅱ had good calibration at each time points （x2=5.305,7.557,6.369 and 8.540,P〉0.05）,however,the consistency of expected mortality with observed mortality was satisfactory only at 48 and 72 h（82.6%,83.4%）,so was the discrimination （ AUROC=0.825,0.847 respectively）.There was no correlation between disease categories and outcome.Conclusion SAPS Ⅱ scoring,best evaluated at 48 and 72 h after hospitalization,can be used as a reliable predictor of probability of mortality in patients hospitalized in N-ICU and prediction can be applied in these patients with all different neurology diseases.