Presepsin and Procalcitonin Levels in Sepsis
Presepsin and Procalcitonin Levels in Sepsis
Additional file 3 gives the selected clinical characteristics of patients at the time of study enrollment according to median levels of presepsin assessed on day 1 (1,494 pg/ml). Higher levels of presepsin were significantly associated with worse SOFA scores and reduced diuresis.
Baseline clinical characteristics and circulating biomarker concentrations measured at day 1 in survivors and nonsurvivors at ICU discharge are presented in Table 1. The only significant difference between the two groups was a higher concentration of presepsin in nonsurvivors (P = 0.0015). The main clinical characteristics, risk scores (SOFA score at enrollment and SAPS II score), as well as PCT levels, were not significantly different.
The evolution of presepsin levels over time in survivors was significantly different from that in nonsurvivors (Figure 1) (P value for time–survival interaction = 0.03). At all the three time points, presepsin levels were significantly higher in the decedents than in the survivors (P < 0.005). There was no significant difference in presepsin concentrations measured on day 2 or day 7 compared to that on day 1 in both nonsurvivors and survivors. Conversely, PCT levels fell rapidly and similarly in survivors and nonsurvivors (Figure 1), and concentrations were significantly different between the two groups only on day 7 (P = 0.01). PCT levels measured on day 2 or day 7 were significantly lower than those on day 1 (P < 0.0001 for survivors and P < 0.001 for nonsurvivors). The interaction between 7-day time course and survival was significant for SOFA score (P = 0.008), but not for serum lactate concentration (P = 0.08) or mixed venous oxygen saturation (SvO2) (P = 0.49).
(Enlarge Image)
Figure 1.
Time course of plasma concentrations of presepsin and procalcitonin during ICU stay by survival status. Plasma concentrations of presepsin and procalcitonin 1, 2 and 7 days after enrollment in decedents (n = 50, black) and survivors (n = 50, gray) at ICU discharge. Data are shown as median and interquartile range. Two-way ANOVA for repeated measurements was done on log-transformed biomarker concentrations. **P < 0.005, *P = 0.01 by Mann–Whitney U test.
Plasma concentrations of presepsin were similar between patients with severe sepsis and those with septic shock at the time of study enrollment, both on day 1 (1,571 pg/ml (793 to 2,440) (n = 26) vs. 1,485 pg/ml (960 to 3,501) (n = 74); P = 0.51) and at subsequent times (data not shown). The same was true for SOFA score and PCT (data not shown).
In the 74 patients with septic shock, presepsin was significantly higher in decedents than in survivors, both at baseline (2,590 pg/ml (1,631 to 4,310) vs. 1,170 pg/ml (890 to 1,799); P = 0.0007) (Figure 2) and on the following days (P < 0.001). In contrast, early PCT levels (days 1 and 2) were not different between the two groups with septic shock.
(Enlarge Image)
Figure 2.
Time course of plasma concentrations of presepsin and procalcitonin in patients with septic shock during ICU stay by survival status. Plasma concentrations of presepsin and procalcitonin 1, 2 or 7 days after enrollment in decedents (n = 34, black) and survivors (n = 40, gray) with septic shock at ICU discharge. Data are shown as median and interquartile range. Two-way ANOVA for repeated measurements was done on log-transformed biomarker concentrations. **P < 0.001, *P = 0.007 by Mann–Whitney U test.
Presepsin levels did not differ significantly in relation to the type of infection by either infection sites or blood cultures (bacterial, fungal, mixed or undetermined) or by the type of bacterial infection (purely Gram-negative, purely Gram-positive, mixed or undetermined; data not shown).
Patients were stratified according to the time from fulfillment of inclusion criteria and study enrollment (within 6 hours or between 6 and 24 hours). In patients with early enrollment (within 6 hours), plasma concentrations of presepsin on day 1 were not significantly different between decedents (2,138 pg/ml (1,062 to 3,101); n = 24) and survivors at ICU discharge (1,335 pg/ml (879 to 2,856) (n = 25); P = 0.17). The same was true for PCT (23.2 μg/L (3.3 to 64.1) vs. 10.9 μg/L (2.8 to 30.8); P = 0.20). However, the plasma concentration of presepsin on day 1 was significantly higher in decedents randomized between 6 and 24 hours after the onset of inclusion criteria (2,621 pg/ml (1,223 to 4,959); n = 26) than in the survivors at ICU discharge (1,044 pg/ml (875 to 1,331) (n = 25); P = 0.003), whereas PCT did not differ (16.2 μg/L (3.4 to 40.6) vs. 7.1 μg/L (2.7 to 41.8); P = 0.88).
The prognostic value of presepsin was evaluated at ICU discharge and during follow-up period on days 28 and 90 after study enrollment according to prespecified study endpoints. In univariate Cox proportional hazards models (Table 2), presepsin on day 1 was associated with mortality at ICU discharge (HR = 1.65 (95% CI = 1.22 to 2.24) for each 1 unit increase in log-transformed concentration; P = 0.0012), at 28 days (HR = 1.73 (95% CI = 1.32 to 2.27); P < 0.001) and at 90 days (HR = 1.50 (95% CI = 1.18 to 1.91); P = 0.001). The corresponding HRs and 95% CIs for PCT were 1.01 (95% CI =0.87 to 1.18; P =0.88), 1.07 (95% CI = 0.92 to 1.25; P =0.40), and 1.03 (95% CI =0.91 to 1.18; P =0.63). After adjustment for clinically relevant variables, including SAPS II score, SvO2 level and mean arterial pressure 6 hours after study enrollment, as well as SOFA score and serum lactate level on day 1, presepsin on day 1 remained independently related to outcome at ICU discharge (HR = 1.48 (95% CI = 1.04 to 2.11; P = 0.03) and after 28 days (HR = 1.54 (95% CI = 1.12 to 2.12); P = 0.01) (Table 2), but not after 90 days (HR = 1.29 (95% CI = 0.96 to 1.72); P = 0.09).
Presepsin concentrations, but not PCT concentrations, on days 2 and 7 were independently related to mortality in the ICU, as well as on days 28 and 90 (Table 2). The SOFA score was the only clinical variables independently associated with mortality in the multivariable models including PCT; when presepsin was included in place of PCT in these models, the SOFA score was no longer significantly associated with mortality (data not shown).
The prognostic accuracy of presepsin was evaluated on the basis of ROC curves, yielding AUCs for ICU survival of 0.69, 0.70 and 0.74 on days 1, 2 and 7, respectively (Table 3). The SOFA score had similar accuracy. Corresponding values for PCT were 0.56, 0.55 and 0.64, respectively.
Results
Presepsin Concentration and Baseline Clinical Characteristics
Additional file 3 gives the selected clinical characteristics of patients at the time of study enrollment according to median levels of presepsin assessed on day 1 (1,494 pg/ml). Higher levels of presepsin were significantly associated with worse SOFA scores and reduced diuresis.
Time Course of Presepsin Levels During the Study Period
Baseline clinical characteristics and circulating biomarker concentrations measured at day 1 in survivors and nonsurvivors at ICU discharge are presented in Table 1. The only significant difference between the two groups was a higher concentration of presepsin in nonsurvivors (P = 0.0015). The main clinical characteristics, risk scores (SOFA score at enrollment and SAPS II score), as well as PCT levels, were not significantly different.
The evolution of presepsin levels over time in survivors was significantly different from that in nonsurvivors (Figure 1) (P value for time–survival interaction = 0.03). At all the three time points, presepsin levels were significantly higher in the decedents than in the survivors (P < 0.005). There was no significant difference in presepsin concentrations measured on day 2 or day 7 compared to that on day 1 in both nonsurvivors and survivors. Conversely, PCT levels fell rapidly and similarly in survivors and nonsurvivors (Figure 1), and concentrations were significantly different between the two groups only on day 7 (P = 0.01). PCT levels measured on day 2 or day 7 were significantly lower than those on day 1 (P < 0.0001 for survivors and P < 0.001 for nonsurvivors). The interaction between 7-day time course and survival was significant for SOFA score (P = 0.008), but not for serum lactate concentration (P = 0.08) or mixed venous oxygen saturation (SvO2) (P = 0.49).
(Enlarge Image)
Figure 1.
Time course of plasma concentrations of presepsin and procalcitonin during ICU stay by survival status. Plasma concentrations of presepsin and procalcitonin 1, 2 and 7 days after enrollment in decedents (n = 50, black) and survivors (n = 50, gray) at ICU discharge. Data are shown as median and interquartile range. Two-way ANOVA for repeated measurements was done on log-transformed biomarker concentrations. **P < 0.005, *P = 0.01 by Mann–Whitney U test.
Plasma concentrations of presepsin were similar between patients with severe sepsis and those with septic shock at the time of study enrollment, both on day 1 (1,571 pg/ml (793 to 2,440) (n = 26) vs. 1,485 pg/ml (960 to 3,501) (n = 74); P = 0.51) and at subsequent times (data not shown). The same was true for SOFA score and PCT (data not shown).
In the 74 patients with septic shock, presepsin was significantly higher in decedents than in survivors, both at baseline (2,590 pg/ml (1,631 to 4,310) vs. 1,170 pg/ml (890 to 1,799); P = 0.0007) (Figure 2) and on the following days (P < 0.001). In contrast, early PCT levels (days 1 and 2) were not different between the two groups with septic shock.
(Enlarge Image)
Figure 2.
Time course of plasma concentrations of presepsin and procalcitonin in patients with septic shock during ICU stay by survival status. Plasma concentrations of presepsin and procalcitonin 1, 2 or 7 days after enrollment in decedents (n = 34, black) and survivors (n = 40, gray) with septic shock at ICU discharge. Data are shown as median and interquartile range. Two-way ANOVA for repeated measurements was done on log-transformed biomarker concentrations. **P < 0.001, *P = 0.007 by Mann–Whitney U test.
Presepsin levels did not differ significantly in relation to the type of infection by either infection sites or blood cultures (bacterial, fungal, mixed or undetermined) or by the type of bacterial infection (purely Gram-negative, purely Gram-positive, mixed or undetermined; data not shown).
Patients were stratified according to the time from fulfillment of inclusion criteria and study enrollment (within 6 hours or between 6 and 24 hours). In patients with early enrollment (within 6 hours), plasma concentrations of presepsin on day 1 were not significantly different between decedents (2,138 pg/ml (1,062 to 3,101); n = 24) and survivors at ICU discharge (1,335 pg/ml (879 to 2,856) (n = 25); P = 0.17). The same was true for PCT (23.2 μg/L (3.3 to 64.1) vs. 10.9 μg/L (2.8 to 30.8); P = 0.20). However, the plasma concentration of presepsin on day 1 was significantly higher in decedents randomized between 6 and 24 hours after the onset of inclusion criteria (2,621 pg/ml (1,223 to 4,959); n = 26) than in the survivors at ICU discharge (1,044 pg/ml (875 to 1,331) (n = 25); P = 0.003), whereas PCT did not differ (16.2 μg/L (3.4 to 40.6) vs. 7.1 μg/L (2.7 to 41.8); P = 0.88).
Plasma Presepsin Concentration in Relation to Mortality
The prognostic value of presepsin was evaluated at ICU discharge and during follow-up period on days 28 and 90 after study enrollment according to prespecified study endpoints. In univariate Cox proportional hazards models (Table 2), presepsin on day 1 was associated with mortality at ICU discharge (HR = 1.65 (95% CI = 1.22 to 2.24) for each 1 unit increase in log-transformed concentration; P = 0.0012), at 28 days (HR = 1.73 (95% CI = 1.32 to 2.27); P < 0.001) and at 90 days (HR = 1.50 (95% CI = 1.18 to 1.91); P = 0.001). The corresponding HRs and 95% CIs for PCT were 1.01 (95% CI =0.87 to 1.18; P =0.88), 1.07 (95% CI = 0.92 to 1.25; P =0.40), and 1.03 (95% CI =0.91 to 1.18; P =0.63). After adjustment for clinically relevant variables, including SAPS II score, SvO2 level and mean arterial pressure 6 hours after study enrollment, as well as SOFA score and serum lactate level on day 1, presepsin on day 1 remained independently related to outcome at ICU discharge (HR = 1.48 (95% CI = 1.04 to 2.11; P = 0.03) and after 28 days (HR = 1.54 (95% CI = 1.12 to 2.12); P = 0.01) (Table 2), but not after 90 days (HR = 1.29 (95% CI = 0.96 to 1.72); P = 0.09).
Presepsin concentrations, but not PCT concentrations, on days 2 and 7 were independently related to mortality in the ICU, as well as on days 28 and 90 (Table 2). The SOFA score was the only clinical variables independently associated with mortality in the multivariable models including PCT; when presepsin was included in place of PCT in these models, the SOFA score was no longer significantly associated with mortality (data not shown).
Prognostic Accuracy of Presepsin
The prognostic accuracy of presepsin was evaluated on the basis of ROC curves, yielding AUCs for ICU survival of 0.69, 0.70 and 0.74 on days 1, 2 and 7, respectively (Table 3). The SOFA score had similar accuracy. Corresponding values for PCT were 0.56, 0.55 and 0.64, respectively.
Source...