Presepsin as a predictor of severe sepsis in urinary tract infection


 Background Recently, presepsin is reported to be a biomarker for early diagnosis of sepsis and evaluation of prognosis in septic patients, but there are few reports about urinary-tract infections. The objective of this study is to evaluate whether presepsin is a recent marker for detecting severe sepsis, and whether it can predict the therapeutic course in UTI when compared with procalcitonin (PCT) and C-reactive protein (CRP), already used markers.Methods From April 2014 to December 2016, a total of 50 patients, who were admitted into Gunma university hospital with urinary-tract infections, were enrolled in this study. Vital signs, presepsin, PCT, CRP, white blood cell (WBC), causative diseases of urinary-tract infections and other data were evaluated at the enrollment, third and fifth days. The patients were divided into two groups; with (n=11) or without (n=39) septic shock at the enrollment day, and with (n=7) or without (n=43) sepsis at the fifth day, respectively. Presepsin was evaluated for systemic inflammatory response syndrome (SIRS) or septic shock. Results Concerning the enrollment day, there was no significant difference of presepsin between SIRS and non-SIRS groups (p=0.276). The median presepsin (pg/mL) was significantly higher in the septic shock group (p<0.001). Multivariate logistic regression analysis showed presepsin (≧ 500 pg/ml) was an independent risk factor associated with septic shock (p=0.007). ROC curve for diagnosing septic shock indicated an area under the curve (AUC) at 0.881 for presepsin (vs. 0.690, 0.583 and 0.527 for PCT, CRP and WBC, respectively). Concerning the 5th day after admission, the median presepsin of the enrollment day was significantly higher in SIRS groups than non-SIRS groups (p=0.006). On the other hand, PCT (≥ 2 ng/ml) of the enrollment day was an independent risk factor associated with SIRS. ROC curve for diagnosing sepsis at the fifth day indicated an AUC at 0.837 for PCT (vs. 0.817, 0.811 and 0.802 for presepsin, CRP and WBC, respectively).Conclusions This study shows that presepsin may be a good marker for diagnosis of severe patients who need vasopressor therapy at the data of admission, and PCT may be a good marker for predicting hard-to-treat cases in UTI.


Introduction
Urinary tract infections (UTI) are common, and sometimes progresses to sepsis or septic shock, which can be lethal. Mortality from severe sepsis and septic shock is reported to be between 20-50% [1,2], and the ratio of UTI in all-cause of severe sepsis and septic shock is 9-31% [2]. Therefore, the diagnosis and severity of sepsis or septic shock are important at the beginning of treatment in UTI.
Clinically, C-reactive protein (CRP) and procalcitonin (PCT) are used as markers of disease severity in UTI.
However, both CRP and PCT have some limitations. The response time after bacterial infections is delayed (CRP; 6hr, PCT; 2-3hr), the trigger for production is not living bacteria (CRP; cytokine, PCT; endotoxin and cytokine), and the serum half-time is long (CRP; 4-6hr, PCT; 20-24hr) [3,4]. Therefore, a new biomarker of bacterial infection, which re ects the clinical condition at the time of measurement is required.
Presepsin is a 13KDa fragment of the N-terminal of soluble CD14 [5]. Granulated leukocyte phagocytose both bacteria and CD14 and expel presepsin into the blood after enzymatic digestion of bacteria within two hours [5,6]. Recently, presepsin has been reported to have a high sensitivity for detecting sepsis and to be a biomarker for early diagnosis of sepsis [7]. Moreover, elevated presepsin levels on day 1 can evaluate the prognosis of septic patients in intensive-care unit [8,9]. However, there are few reports on the use of presepsin in UTI.
The objective of this study is to evaluate whether presepsin is a useful marker for detecting sepsis or severe sepsis, and whether it can predict the therapeutic courses in UTI compared with other markers, such as PCT and CRP.

Patients
We performed a prospective observational study. From April 2014 to December 2016, a total of 57 patients, who were admitted into Gunma university hospital with UTI, were enrolled in this study. Seven patients were excluded from this study due to data unavailability. UTI was diagnosed by a urologist based on urinary sediment (≥ 5 leucocytes/high power eld) and symptoms (fever and/or micturition pain and/or ank pain). Patient age, sex, medical history, oral medicine, blood pressure, body temperature, heart rate, respiratory rate, urine and blood culture results, surgical procedure for UTI after admission, causative diseases of UTI, presepsin, PCT, CRP, white blood cell (WBC), Aspartate transaminase (ALT), Alanine transaminase (AST), γ-glutamyl transpeptidase (γGPT) and Creatinine (Cr) were collected. Vital signs and blood data were evaluated at enrollment, and on the 3rd and 5th day after admission. This study was approved by the institutional review board of Gunma University Hospital (No.1650).

Assessment of systemic in ammatory response syndrome (SIRS) and septic shock
The diagnosis of SIRS and septic shock were made according to the criteria set by the American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM) [10].

Outcomes
Three outcome variables: SIRS on the enrollment day, septic shock on the enrollment day, and SIRS on the 5th day after admission were evaluated by dividing the participants into two groups for each of these variables; with (n=39) or without (n=11) SIRS on the enrollment day, with (n=11) or without (n=39) septic shock on the enrollment day, and with (n=7) or without (n=43) SIRS on the 5th day after admission, respectively.

Statistical Analysis
Mann-Whitney U-test was used for the continuous variables (age, CRP, presepsin, PCT, ALT, AST, γGPT and Cr). We estimated the independence by the chi-squared test or Fisher`s exact test for categorical variables (sex, placement of urinary catheter, urological cancer, urinary calculi, diabetes mellitus and internal use of steroid). Independent predictors were evaluated using logistic regression analysis. The predictive accuracy of presepsin, PCT, CRP and WBC for septic shock or SIRS was evaluated by the area under the curve (AUC) of a receiver operating characteristics (ROC) analysis. The Youden's index (sensitivity + speci city -1) was used to calculate optimal cutoff values of presepsin. P values ≤ 0.05 were considered evidence of a signi cant difference. SPSS Statistics Ver. 25 (IBM Corp. IL, USA) was used for statistical analysis.

Prediction of SIRS on the enrollment day
The overall median baseline presepsin, PCT, and CRP levels were 483 pg/mL, 0.87 ng/mL, and 10.45 mg/L, respectively, with no signi cant difference between SIRS and non-SIRS groups (supplemental Table   1)." Prediction of Septic shock on the enrollment day The median presepsin level (pg/mL) was signi cantly higher in the septic shock group (1380 vs. 399, p<0.001). The PCT and CRP levels were not signi cantly different between septic shock and non-septic shock groups. Other blood investigations that were signi cantly higher in the septic group included: AST (p=0.003), ALT (p=0.049), γGTP (p=0.002 and Cr (p=0.02), respectively (Table 1a). Logistic regression analysis to evaluate factors associated with septic shock on enrollment day is shown in Table 1b. Factors associated with septic shock on univariate analysis included presepsin (≥ 500 pg/mL), PCT (≥ 2 ng/ml), AST (≥ 34 U/L) and γGTP (≥ 47 U/L). Only presepsin level retained signi cance after controlling for confounders in multivariate logistic regression analysis (Table 1b). ROC curve for diagnosing septic shock indicated AUC of 0.881 for presepsin, which is larger than other markers (Fig.1). The Cutoff level of presepsin with the optimum diagnostic e ciency by the ROC curves were 492 ng/ml, which was broadly similar to clinical cut off values (500 pg/mL).  Table 2b. Factors associated with SIRS on the 5th day after admission in univariate analysis included day1/presepsin (≥ 500 pg/mL), day1/PCT (≥ 2 ng/ml) and urinary calculi. Only day1/PCT level retained signi cance after controlling for confounders in multivariate logistic regression analysis (Table 2b). ROC curve in patients with de nitive prediction of SIRS on the 5th day after admission indicated an AUC at 0.837 for PCT, which is larger than other markers (Fig.1). Concerning presepsin, the cutoff level with the optimum diagnostic e ciency by the ROC curves were 492 ng/ml, which was broadly similar to clinical cut off values (500 pg/mL).

Discussion
In this study, elevated presepsin on admission was an independent risk factor for septic shock, while elevated PCT on admission was an independently associated with SIRS on the 5th day after admission for patients admitted with UTI. Presepsin and PCT have the different origin [3][4][5]. These results support our hypothesis that presepsin is a useful marker for detecting severe sepsis from UTI. On the other hand, PCT can predict therapeutic courses in UTI better than presepsin.
Presepsin is one of the biomarkers which increases after bacterial infections [11]. The levels are increased in acute pyelonephritis patients with bacteremia [12] and elevation of presepsin levels before treatment might predict the development of sepsis in patients with obstructive acute pyelonephritis [13].
In this study, we evaluated the patients who needed hospitalization due to not only pyelonephritis but also prostatitis, and elevation of presepsin levels on the enrollment day was a predictor of septic shock Therefore, it seems that presepsin is useful for detecting severe urosepsis which need vasopressor therapy.
Sever sepsis and septic shock are fatal, with mortality rates of 28.3%-41.1% [14]. Clinically, it is very useful to predict septic shock at the time of admission. It is, therefore, essential that biomarkers increase immediately after infections and have a high sensitivity for sepsis. Presepsin levels have been shown to increase within 2 hours or later together with blood bacterial counts, and peak at 3 hours [6]. On the other hand, elevated PCT and CRP levels are detected within 3-6 hours and 6 hours, and peak at 6-8 hours and 36-50 hours, respectively [15,16]. It has also been reported that patients with severe sepsis have signi cantly higher presepsin levels than those with sepsis, local infection or SIRS [17]. In this study, elevation of presepsin levels on the enrollment day was a predictor of septic shock, but not of SIRS on the day of enrollment. These results suggest that presepsin may be used to identify patients at increased risk of more severe infections at early stages.
One important characteristics of biomarkers is the prognostic value. Presepsin levels on day 1 were reported to be correlated with 60-day in-hospital mortality in patients with sepsis, severe sepsis or septic shock [18], a longer intensive-care unit stay, and a lower degree of resolution of the primary infection [8,9]. Mortality is one of the prognostic variables. Since there were no mortalities in this current study, we used presence or absence of SIRS on day 5 for prognosis. In this study, the levels of presepsin on day1 was signi cantly higher in SIRS on day 5 groups than non-SIRS on day 5 groups, but the levels of PCT on day1 was only picked up as a predictor of SIRS on day5 by multivariate logistic regression analysis. Concerning PCT, the AUC for PCT to predict 30-day mortality in febrile UTI was reported to be 0.71 (95% CI: 0.56-0.85) [19]. There are few reports about the comparison of presepsin with PCT for evaluating the prognostic value in UTI. Further researches which include a lot of more sever patients of urosepsis than this study are needed to study the availability of each biomarkers to predict treatment outcomes.
This study had some limitations. First, since 2016, sepsis has been de ned using the Sequential (Sepsis-Related) Organ Failure Assessment (SOFA) [20]. This study began in 2014, and we did not check conscious level which is a requirement for SOFA scoring. Therefore, instead of SOFA, the SIRS criteria [10] was used for de ning sepsis and septic shock. Second, we used a single-center design with a small sample size. Third, we did not consider renal function in setting the reference values of presepsin and PCT. The levels of presepsin and PCT have been reported to be affected by renal function [21,22]. It would be ideal to adjust the reference values depending on renal function. However, these adjustments are yet to be clari ed.

Conclusions
This study showed that, in UTI, presepsin may be a good marker for diagnosis of severe patients who need vasopressor therapy at the data of admission, and PCT may be a good marker for predicting hardto-treat cases.