Objective To develop a tool capable of early and exactly predicting various outcomes in comatose survivors who restore spontaneous circulation after cardiopulmonary resuscitation （CPR） and validate its performance. ...Objective To develop a tool capable of early and exactly predicting various outcomes in comatose survivors who restore spontaneous circulation after cardiopulmonary resuscitation （CPR） and validate its performance. Methods Variables that were both readily available and predictive of outcomes were identified by systematically reviewing published literature on resuscitation. A value was assigned to these variables. We used these variables in combination with APACHE II score to devise a multifactorial prediction score system, which we called PRCSs Prognostication Score （PRCSs-PS）. Outcomes in 115 hospitalized comatose survivors after CPR were retrospectively reviewed using PRCSs-PS. Score of patients with different outcomes was compared. The area under the receiver- operating characteristic （ROC） curve was determined to evaluate performance of this tool to identify patients with a poor outcome （CPC4 and 5） and other outcomes （CPC1, 2, and 3）. Results There were differences of PRCSs-PS score among multiple groups with five different outcomes （CPC 1-5）（F=65.91, P=0.000）. Pairwise groups with different CPC were compared： no significant difference was noted between CPC1 and CPC2 （12.41±6.49 vs 17.38±6.91,P=0.092）, but difference between other pairwise CPC groups was statistically significant （CPC2 vs CPC3：17.38±6.91 vs 24.50±5.80, P=0.041, CPC3 vs CPC4：24.50±5.80 vs 32.29±5.24, P=0.006）. The performance of PRCSs-PS to discriminate patients with a poor outcome from patients with other outcomes went as follows： it had 100% sensitivity, 78.6% specificity, and 178.6 diagnostic index at the score cut-off22.5; it had 77.8% sensitivity, 100% specificity and 176.4 diagnostic index at the score cut-off32.5. Score 23 and 33 were two key cut-offpoints. The area under the ROC curve was 0.968, showing excellent discrimination. Conclusions The final outcomes in post-resuscitation comatose survivors can be accurately predicted using PRCSs-PS Score.展开更多
文摘Objective To develop a tool capable of early and exactly predicting various outcomes in comatose survivors who restore spontaneous circulation after cardiopulmonary resuscitation （CPR） and validate its performance. Methods Variables that were both readily available and predictive of outcomes were identified by systematically reviewing published literature on resuscitation. A value was assigned to these variables. We used these variables in combination with APACHE II score to devise a multifactorial prediction score system, which we called PRCSs Prognostication Score （PRCSs-PS）. Outcomes in 115 hospitalized comatose survivors after CPR were retrospectively reviewed using PRCSs-PS. Score of patients with different outcomes was compared. The area under the receiver- operating characteristic （ROC） curve was determined to evaluate performance of this tool to identify patients with a poor outcome （CPC4 and 5） and other outcomes （CPC1, 2, and 3）. Results There were differences of PRCSs-PS score among multiple groups with five different outcomes （CPC 1-5）（F=65.91, P=0.000）. Pairwise groups with different CPC were compared： no significant difference was noted between CPC1 and CPC2 （12.41±6.49 vs 17.38±6.91,P=0.092）, but difference between other pairwise CPC groups was statistically significant （CPC2 vs CPC3：17.38±6.91 vs 24.50±5.80, P=0.041, CPC3 vs CPC4：24.50±5.80 vs 32.29±5.24, P=0.006）. The performance of PRCSs-PS to discriminate patients with a poor outcome from patients with other outcomes went as follows： it had 100% sensitivity, 78.6% specificity, and 178.6 diagnostic index at the score cut-off22.5; it had 77.8% sensitivity, 100% specificity and 176.4 diagnostic index at the score cut-off32.5. Score 23 and 33 were two key cut-offpoints. The area under the ROC curve was 0.968, showing excellent discrimination. Conclusions The final outcomes in post-resuscitation comatose survivors can be accurately predicted using PRCSs-PS Score.
