Procalcitonin (PCT) has been widely investigated for its prognostic value in septic patients. However, studies have produced conflicting results. The purpose of the present meta-analysis is to explore the diagnostic accuracy of a single PCT concentration and PCT non-clearance in predicting all-cause sepsis mortality. We searched PubMed, Embase, Web of Knowledge and the Cochrane Library. Articles written in English were included. A 2 × 2 contingency table was constructed based on all-cause mortality and PCT level or PCT non-clearance in septic patients. Two authors independently evaluated study eligibility and extracted data. The diagnostic value of PCT in predicting prognosis was determined using a bivariate meta-analysis model. We used the Q-test and I
2 index to test heterogeneity. Twenty-three studies with 3,994 patients were included. An elevated PCT level was associated with a higher risk of death. The pooled relative risk (RR) was 2.60 (95% confidence interval (CI), 2.05–3.30) using a random-effects model (I
2 = 63.5%). The overall area under the summary receiver operator characteristic (SROC) curve was 0.77 (95% CI, 0.73–0.80), with a sensitivity and specificity of 0.76 (95% CI, 0.67–0.82) and 0.64 (95% CI, 0.52–0.74), respectively. There was significant evidence of heterogeneity for the PCT testing time (P = 0.020). Initial PCT values were of limited prognostic value in patients with sepsis. PCT non-clearance was a prognostic factor of death in patients with sepsis. The pooled RR was 3.05 (95% CI, 2.35–3.95) using a fixed-effects model (I
2 = 37.9%). The overall area under the SROC curve was 0.79 (95% CI, 0.75–0.83), with a sensitivity and specificity of 0.72 (95% CI, 0.58–0.82) and 0.77 (95% CI, 0.55–0.90), respectively. Elevated PCT concentrations and PCT non-clearance are strongly associated with all-cause mortality in septic patients. Further studies are needed to define the optimal cut-off point and the optimal definition of PCT non-clearance for accurate risk assessment.