• Z Zhang and H Ni.
• Department of Critical Care Medicine, Jinhua Central Hospital, Zhejiang Province, China. zh_zhang1984@hotmail.com
• Anaesth Intensive Care. 2011 Sep 1;39(5):854-61.

AbstractC-reactive protein is a marker of inflammatory response and has been widely investigated in cardiovascular and infectious diseases, especially to monitor therapeutic success. However, its role as a predictor of clinical outcome in critically ill patients remains uncertain and controversial. The objective of this study was to investigate the predictive value of C-reactive protein in critically ill patients. The databases of PubMed, the Cochrane clinical trial database and EMBASE (from inception to August 2010) were searched. Prospective non-randomised clinical studies comparing C-reactive protein concentrations between survivors and non-survivors were included. Pooled mean difference in C-reactive protein concentrations between survivors and non-survivors was calculated. Heterogeneity was analysed by I2. Sensitivity and subgroup analyses were conducted to explore the heterogeneity. Fourteen studies containing a total of 1969 patients were finally included in our analysis. The weighted mean difference in the C-reactive protein levels between survivors and non-survivors was 9.15 mg/l (95% confidence interval -6.50 to 24.81). The heterogeneity was large with I2 = 92%. Subsequent investigation of the heterogeneity with sensitivity analyses yielded no significant differences. The subgroup analysis showed that the weighted mean difference in early (within 48 hours) C-reactive protein levels between survivors and non-survivors was not significantly different, in contrast to the late (beyond 48 hours) C-reactive protein level. This was significantly greater in non-survivors with a weighted mean difference of 63.80 mg/l (95% confidence interval 35.67 to 91.93). Our systematic review shows that while the early C-reactive protein concentration is not a good predictor of survival in critically ill patients, the late C-reactive protein concentration may help to identify patients who are at risk of death.

12 keywords...

hide…