Accuracy of procalcitonin for diagnosis of sepsis in adults: a Meta-analysis

Abstract

To assess the clinical value of procalcitonin (PCT ) in the diagnosis of sepsis in adults. An extensive search for related literature from the Wanfang data, CNKI, VIP, Medline/PubMed, Embase/OvidSP and the Cochrane Library up to December 2014 was performed. The articles, including prospective observational studies or randomized controlled trials, regarding PCT for the diagnosing of sepsis were enrolled. Only patients older than 18 years were included. Patients with sepsis, severe sepsis, or septic shock served as the experimental group, and those with a systemic inflammatory response syndrome (SIRS) of non-infectious origin as control group. The language of literature included was English or Chinese. The quality of the studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. Heterogeneity, pooled diagnostic odds ratio (DOR), pooled sensitivity, pooled specificity, pooled positive likelihood ratio, pooled negative likelihood ratio, the area under the summary receiver operating characteristic curve (SROC) and subgroup analysis were analyzed with the software of Metadisc 1.4. A total of 6 385 published reports were collected, and among them 24 met the inclusion criteria, including a total. of 3 107 patients. The studies showed substantial heterogeneity (I2 = 69.4%), and random effect model was used for Meta analysis, showing that the pooled DOR was 10.37 [95% confidence interval (95%CI) = 7.10-15.17]. No evidence of a threshold effect was found (Spearman correlation coefficient = 0.27, calculated by logarithm of sensitivity and logarithm of 1-specificity, P = 0.20). The DOR values of pooled and each study were not distributed along the same line in forest plots, and Cochran-Q = 78.33, P = 0.000 0, showing that there was heterogeneity in result from non threshold effect. Except for partial heterogeneity caused by non threshold effect, the result of Meta regression analysis including PCT detection method, categories of disease, research location and so on showed P values were all higher than 0.05. Thus, the heterogeneity could not be explained by Meta regression analysis. The pooled sensitivity was 74% (95% CI = 72%-76%), the pooled specificity was 70% (95% CI = 67%-72%), the pooled positive likelihood ratio was 2.79 (95% CI = 2.31-3.38), the pooled negative likelihood ratio was 0.34 (95% CI = 0.28-0.41), and the pooled AUC was 0.83 (95% CI = 0.79-0.87). AUC in medical patients was 0.80 (95% CI = 0.75-0.85), which was higher than that in surgical patients [0.71 (95% CI = 0.65-0.81)]. Our results indicate a moderate degree of value of PCT for diagnosis of sepsis in adult patients. The diagnostic accuracy in medical patients is higher than that in surgical patients. PCT is a good auxiliary biomarker for diagnosis of sepsis.