Plasma presepsin for mortality prediction in patients with sepsis-associated acute kidney injury requiring continuous kidney replacement therapy
Article information
Abstract
Background
The reliability of presepsin as a biomarker of sepsis may be reduced in patients with acute kidney injury (AKI) requiring continuous kidney replacement therapy (CKRT). This study analyzed the utility of plasma presepsin values in predicting mortality in patients with AKI requiring CKRT, particularly those with sepsis-associated AKI.
Methods
This single-center retrospective study included 57 patients who underwent CKRT, with plasma presepsin measurements, from April 2022 to March 2023; 35 had sepsis-associated AKI. The predictive values of plasma presepsin, as well as Acute Physiology and Chronic Health Evaluation II (APACHE II) and Sequential Organ Failure Assessment (SOFA) scores, for 28-day mortality were analyzed using receiver operating characteristic curves. Multivariate Cox regression analysis was performed to identify risk factors for 28-day mortality in the sepsis-associated AKI subgroup.
Results
Overall, plasma presepsin showed a lower area under the curve value (0.636; 95% confidence interval [CI], 0.491–0.781) than the APACHE II (0.663; 95% CI, 0.521–0.804) and SOFA (0.731; 95% CI, 0.599–0.863) scores did. However, in sepsis-associated AKI, the area under the curve increased to 0.799 (95% CI, 0.653–0.946), which was higher than that of the APACHE II (0.638; 95% CI, 0.450–0.826) and SOFA (0.697; 95% CI, 0.519–0.875) scores. In the multivariate Cox regression analysis, a high presepsin level was an independent risk factor for 28-day mortality in sepsis-associated AKI (hazard ratio, 3.437; p = 0.03).
Conclusion
Presepsin is a potential prognostic marker in patients with sepsis-associated AKI requiring CKRT.
Introduction
Presepsin is a soluble cluster of differentiation (CD) subtype that results from the cleavage of CD14. CD14 is a lipopolysaccharide receptor, mainly expressed in monocytes and macrophages. It is degraded inside the cell during phagocytosis during infection and is released into the blood as a soluble CD14 subtype [1–3]. Presepsin is a novel marker that is highly specific for infections and has been extensively studied. Presepsin is characterized by a faster rise in blood levels during infection compared to other biomarkers, such as procalcitonin (PCT) and C-reactive protein (CRP), and its short half-life is expected to be useful for determination of prognosis [4–6]. Recent clinical studies have reported that presepsin is a good indicator for distinguishing between noninfectious organ failure and sepsis, and its utility in clinical practice is expected to increase in the future [7].
Presepsin has a low molecular weight (13 kDa) and is eliminated by the kidneys. Therefore, in patients with reduced kidney function, the cutoff value of presepsin varies and the reliability of the test may decrease [8,9]. In patients with acute kidney injury (AKI), especially those requiring continuous kidney replacement therapy (CKRT), there is a greater risk of a reduction in the reliability of presepsin because kidney function fluctuates very dynamically. Furthermore, there is an additional risk that presepsin may be removed by CKRT machines, further reducing its reliability [10,11]. Therefore, presepsin has not been studied as a diagnostic or prognostic marker in patients with AKI receiving CKRT.
Presepsin is highly specific for infection. Although it is affected by kidney function, we hypothesized that plasma presepsin levels measured before CKRT initiation may be useful as a prognostic marker, especially in patients with sepsis-associated AKI (SA-AKI) undergoing CKRT. Therefore, we analyzed the predictive value of the plasma presepsin level as a prognostic marker for mortality in patients with AKI who underwent CKRT.
Methods
Study design and population
This single-center, retrospective study included critically ill patients who underwent CKRT for AKI at Konyang University Hospital, College of Medicine, Konyang University Hospital between April 2022 and March 2023, in addition to plasma presepsin measurements. The exclusion criteria were age of <18 years, already being on maintenance renal replacement therapy (RRT) for end-stage kidney disease (ESKD), and not having a plasma presepsin measurement within the 24 hours prior to CKRT initiation (measurement more than 24 hours prior to CKRT initiation, measurement after CKRT initiation). AKI was diagnosed according to the 2012 KDIGO (Kidney Disease: Improving Global Outcomes) AKI guidelines, and CKRT was initiated in patients with clinical hemodynamic instability with volume overload, electrolyte imbalance, and metabolic acidosis that could not be controlled with conservative treatment [12]. The protocol for CKRT initiation is a continuous veno-venous hemodiafiltration mode with an effluent dose of 30 mL/kg/hr. Half of the total effluent dose is dialysate, the other half are pre- and post-replacement doses in a 2:1 ratio. The blood flow rate is 150 mL/min. The anticoagulation regimen is heparin, nafamostat, or no anticoagulation, depending on the patient’s status and the physician’s consideration of the bleeding risk.
A total of 74 patients were recruited; 57 patients were finally analyzed after excluding three patients who were already undergoing RRT for ESKD and 14 patients whose plasma presepsin measurement was not performed within the 24 hours prior to CKRT initiation. The cause of AKI was assessed to establish an SA-AKI subgroup (Fig. 1). This study was approved by the Institutional Review Board (IRB) of Konyang University Hospital, College of Medicine, Konyang University (No. KYUH 2023-09-009). The need to obtain informed patient consent was waived by the IRB because the study was retrospective.
Study design and population.
Of the 74 patients who underwent CKRT with plasma presepsin measurement at Konyang University Hospital between April 2022 and March 2023, 17 who met the exclusion criteria were excluded. Finally, the data of 57 patients were analyzed; among them, 35 patients who met both the sepsis and AKI criteria were classified into the SA-AKI subgroup.
AKI, acute kidney injury; CKRT, continuous kidney replacement therapy; ESKD, end-stage kidney disease; RRT, renal replacement therapy; SA-AKI, sepsis-associated acute kidney injury.
Classification of the sepsis-associated acute kidney injury subgroup
The cause of AKI in each patient was identified through a rigorous review of the electronic medical records. Clinical infection was defined as cases with eminent infection (e.g., pneumonia, urinary tract infection) with or without the identification of specific pathogens, or cases with any organ infection by identified pathogens. Sepsis was defined as the presence of two or more of the following three items from the quick Sequential Organ Failure Assessment (SOFA) criteria, along with a suspected or documented infection according to the Sepsis-3 definition: 1) low blood pressure (systolic blood pressure, ≤100 mmHg), 2) high respiratory rate (≥22 rates per minutes), or 3) altered mental status (Glasgow coma scale [GCS] score, <14) [13]. Patients who met both the sepsis and AKI criteria were classified into the SA-AKI subgroup.
Data collection and primary outcome
The patients’ baseline demographic and clinical data were collected, including age, sex, cause of AKI, and comorbidities. Laboratory test results at the time of CKRT initiation were collected, including complete blood cell counts; coagulation; arterial blood gas; and electrolyte, bilirubin, creatinine, arterial lactate, CRP, and PCT levels. The mean arterial pressure, heart rate, respiratory rate, body temperature, partial pressure of oxygen in the arterial blood/fraction of inspired oxygen (PaO2/FiO2) ratio, GCS score, mechanical ventilator status, and inotrope administration status were recorded at the time of CKRT initiation. Based on the collected data, Acute Physiology and Chronic Health Evaluation II (APACHE II) [14] and SOFA [15] scores were calculated.
The primary outcome was the 28-day mortality, which was defined as death within 28 days of CKRT initiation.
Measurement of plasma presepsin
The plasma presepsin levels were measured using PATHFAST Presepsin (LSI Science Corp.). Blood samples were collected in ethylenediaminetetraacetic acid tubes and stored at room temperature. Plasma presepsin levels were measured within 4 hours of collection, according to the manufacturer’s instructions. The assay range of plasma presepsin is 20 to 20,000 pg/mL and the coefficient of variation is 4% to 5% [16].
Statistical analysis
All continuous variables are expressed as mean ± standard deviation or median (interquartile range). The Student t test was used to compare variables showing a normal distribution, and the Mann-Whitney U test was used for variables showing a non-normal distribution. Categorical variables were compared using the chi-square test or Fisher exact test, as appropriate, and expressed as proportions. The predictive values of plasma presepsin, and APACHE II and SOFA scores, for 28-day mortality were analyzed using receiver operating characteristic (ROC) curve analysis in the overall cohort and the SA-AKI subgroup. The cutoff values were determined to simultaneously represent the highest sensitivity and specificity [17]. The Kaplan-Meier survival curve analysis was performed to compare patient survival between the high- and low-presepsin groups in the SA-AKI subgroup according to the cutoff value of plasma presepsin, which was compared using the log-rank test. Multivariate Cox proportional hazard regression analysis was performed to identify risk factors for 28-day mortality in the SA-AKI subgroup. Patients who were transferred to another medical institution within the 28 days after CKRT initiation were censored. A p-value of <0.05 was considered statistically significant. All analyses were performed using IBM SPSS version 24 (IBM Corp.).
Results
Comparison of baseline characteristics between the survivor and non-survivor groups in the overall cohort and the sepsis-associated acute kidney injury subgroup
Twenty-eight days after CKRT initiation, 25 survivors and 32 non-survivors were included in the overall cohort. Presepsin measurement was conducted an average of 3.5 hours before CKRT initiation, and there was no significant difference between the survivor and non-survivor groups (3.4 hours in the survivor group vs. 3.6 hours in the non-survivor group, p = 0.86). Table 1 shows the baseline characteristics of the survivor and non-survivor groups. The mean arterial pressure and GCS score were significantly lower in the non-survivor group than in the survivor group, and the total bilirubin, arterial lactate levels, and inotrope administration rates were significantly higher in the non-survivor group than in the survivor group. Among the comorbidities, the prevalence of hypertension and chronic kidney disease was significantly lower in the non-survivor group than in the survivor group, and there was no difference in the Charlson Comorbidity Index (CCI) between the two groups. There were no differences between the two groups regarding other demographics, laboratory findings, or causes of AKI requiring CKRT.
Comparison of baseline characteristics between the survivor and non-survivor groups in the overall cohort
Table 2 shows the baseline characteristics of survivors and non-survivors in the SA-AKI subgroup. Presepsin measurement was conducted an average of 3.7 hours before CKRT initiation, and there was no significant difference between the survivor and non-survivor groups (3.8 hours in the survivor group vs. 3.6 hours in the non-survivor group, p = 0.70). Similar to the overall cohort results, the mean arterial pressure was significantly lower in the non-survivor group than in the survivor group, and the respiratory rate, total bilirubin, and arterial lactate levels were significantly higher in the non-survivor group than in the survivor group. The prevalence of hypertension was significantly lower in the non-survivor group than in the survivor group, and there was no difference in the CCI between the two groups, similar to the overall cohort results. There were no significant differences in other demographics or laboratory findings between the two groups, other than serum creatinine levels.
Comparison of baseline characteristics between the survivor and non-survivor groups in the SA-AKI subgroup
Comparison of plasma presepsin levels, and APACHE II and SOFA scores, between the survivor and non-survivor groups in the overall cohort and the sepsis-associated acute kidney injury subgroup
Fig. 2 shows a comparison of the plasma presepsin levels, and the APACHE II and SOFA scores, between the survivor and non-survivor groups in the overall cohort. Plasma presepsin levels were not significantly different between the two groups (1,536 pg/mL in survivor group vs. 2,021 pg/mL in non-survivor group, p = 0.08) (Fig. 2A). In contrast, the APACHE II score (27.0 in survivor group vs. 31.0 in non-survivor group, p = 0.02) (Fig. 2B) and SOFA score (10.0 in survivor group vs. 13.5 in non-survivor group, p = 0.001) (Fig. 2C) were significantly higher in the non-survivor group than in the survivor group.
Comparison of plasma presepsin levels, and APACHE II and SOFA scores, between the survivor and non-survivor groups in the overall cohort.
Comparison of median (A) plasma presepsin levels, (B) APACHE II scores, and (C) SOFA scores between the survivor and non-survivor groups in the overall cohort. The median plasma presepsin level was 1,536 pg/mL in the survivor group and 2,021 pg/mL in the non-survivor group (p = 0.08). The median APACHE II score was 27.0 in the survivor group and 31.0 in the non-survivor group (p = 0.02). The median SOFA score was 10.0 in the survivor group and 13.5 in the non-survivor group (p = 0.001).
APACHE II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sequential Organ Failure Assessment.
Fig. 3 shows the plasma presepsin levels, and the APACHE II and SOFA scores, of the survivor and non-survivor groups in the SA-AKI subgroup. In contrast to the overall cohort results, the plasma presepsin level was significantly higher in the non-survivor group than in the survivor group (1,509 pg/mL in survivor group vs. 3,049 pg/mL in non-survivor group, p = 0.002) (Fig. 3A). The APACHE II score tended to be higher in the non-survivor group (28.0 in survivor group vs. 31.0 in non-survivor group, p = 0.12) (Fig. 3B), although this result was not statistically significant. The SOFA score was significantly higher in the non-survivor group than in the survivor group (12.5 in survivor group vs. 14.0 in non-survivor group, p = 0.02) (Fig. 3C).
Comparison of plasma presepsin levels, and APACHE II and SOFA scores, between the survivor and non-survivor groups in the SA-AKI subgroup.
Comparison of median (A) plasma presepsin levels, (B) APACHE II scores, and (C) SOFA scores between survivors and non-survivors in the SA-AKI subgroup. The median plasma presepsin level was 1,509 pg/mL in the survivor group and 3,049 pg/mL in the non-survivor group (p = 0.002). The median APACHE II score was 28.0 in the survivor group and 31.0 in the non-survivor group (p = 0.12). The median SOFA score was 12.5 in the survivor group and 14.0 in the non-survivor group (p = 0.02).
APACHE II, Acute Physiology and Chronic Health Evaluation II; SA-AKI, sepsis-associated acute kidney injury; SOFA, Sequential Organ Failure Assessment.
Predictive values of plasma presepsin, and APACHE II and SOFA scores, for 28-day mortality in the overall cohort and the sepsis-associated acute kidney injury subgroup
In the overall cohort, the SOFA score had the highest area under the ROC curve (AuROC) value (0.731; 95% confidence interval [CI], 0.599–0.863), followed by the APACHE II score (0.663; 95% CI, 0.521–0.804). Plasma presepsin showed the lowest predictive value, with an AuROC value of 0.636 (95% CI, 0.491–0.781) (Fig. 4A). However, in the ROC curve analysis performed in the SA-AKI subgroup, the AuROC value of plasma presepsin was 0.799 (95% CI, 0.653–0.946), which was a significantly better predictive value than the APACHE II and SOFA scores; the AuROC value of the APACHE II score = 0.638 (95% CI, 0.450–0.826) and the AuROC value of the SOFA score = 0.697 (95% CI, 0.519–0.875) (Fig. 4B).
Predictive values of plasma presepsin levels, and APACHE II and SOFA scores, for 28-day mortality in the overall cohort and SA-AKI subgroup.
(A) the receiver operating characteristic (ROC) curve analysis in the overall cohort. Plasma presepsin showed the lowest area under the ROC (AuROC) value, compared with the APACHE II and SOFA scores. (B) The ROC curve analysis in the SA-AKI subgroup. Plasma presepsin showed the highest AuROC value, compared with the APACHE II and SOFA scores.
APACHE II, Acute Physiology and Chronic Health Evaluation II; CI, confidence interval; SA-AKI, sepsis-associated acute kidney injury; SOFA, Sequential Organ Failure Assessment.
In the SA-AKI subgroup, patients were classified into the high- and low-presepsin groups based on a plasma presepsin cutoff level of 1,951 pg/mL. The Kaplan-Meier survival curve analysis showed that the survival rate of the high-presepsin group was prominently decreased (log-rank p < 0.001) (Fig. 5).
The Kaplan-Meier survival curve analysis according to plasma presepsin levels in the SA-AKI subgroup.
The SA-AKI subgroup was further classified into high- and low-presepsin groups based on the cutoff plasma presepsin level of 1,951 pg/mL observed through receiver operating characteristic curve analysis. The high-presepsin group showed a prominent decrease in the survival rate compared to the low-presepsin group. After 28 days of continuous kidney replacement therapy, the survival rates of the high- and low-presepsin groups were 15.8% and 68.8%, respectively.
SA-AKI, sepsis-associated acute kidney injury.
Multivariate Cox regression analysis for 28-day mortality in the sepsis-associated acute kidney injury subgroup
Table 3 shows the results of the multivariate Cox proportional hazard regression analysis performed to identify the risk factors affecting 28-day mortality in the SA-AKI subgroup. The analysis included known significant factors (age, PaO2/FiO2 ratio, and the CCI), baseline characteristics that showed statistically significant differences between the two groups (mean arterial pressure, respiratory rate, serum creatinine level, total bilirubin level, presence of hypertension, and presence of chronic kidney disease), high arterial lactate (lactate level ≥3.10 mmol/L), high CRP (CRP level ≥17.45 mg/dL), high PCT (PCT level ≥7.36 ng/mL), and high presepsin (plasma presepsin level ≥1,951 pg/mL) (for cutoff levels of arterial lactate, PCT, and CRP) (Supplementary Fig. 1, available online). The univariate hazard ratio (HR) of a high presepsin level was 4.723 (95% CI, 1.698–13.136); even after adjusting for several confounding factors, it was an independent risk factor with a multivariate HR of 3.437 (95% CI, 1.126–10.491).
Multivariate Cox regression analysis for 28-day mortality in the SA-AKI subgroup
Discussion
The plasma presepsin level was not a useful marker for predicting 28-day mortality in the overall study cohort undergoing CKRT for AKI. However, in the SA-AKI subgroup, plasma presepsin levels were superior to the APACHE II and SOFA scores in predicting 28-day mortality. Additionally, there was a significant difference in patient survival between the high- and low-presepsin groups according to the plasma presepsin cutoff value. Furthermore, in the multivariate Cox regression analysis, a high presepsin level was observed to be an independent risk factor for 28-day mortality.
Regarding the baseline characteristics of the overall cohort, non-survivors showed significant differences in mean arterial pressure, GCS score, total bilirubin, arterial lactate, and inotrope administration proportions compared with survivors. These are components of the APACHE II and SOFA scores, which are well-known scoring systems for the evaluation of critically ill patients. The results of this study are consistent with those of previous studies [18–20]. The prevalence of chronic kidney disease was higher in the survivor group than in the non-survivor group. This may be because the rate of CKRT initiation for renal problems was higher in the survivor group than in the non-survivor group (16.0%, survivor group vs. 6.3%, non-survivor group). The prevalence of hypertension was higher in the survivor group than in the non-survivor group; however, it is difficult to provide a clear explanation for this. The number of patients analyzed in this study was small, and since this was a retrospective study, the possibility of bias during the data collection process cannot be ruled out. The baseline characteristics of the SA-AKI subgroup also showed significant differences in the components of the APACHE II and SOFA scores (mean arterial pressure, respiratory rate, total bilirubin, and arterial lactate), similar to those of the overall cohort.
The APACHE II and SOFA scores were observed to have relatively good predictive values for 28-day mortality in both the overall cohort and the SA-AKI subgroup, confirming that they are reliable indicators for evaluating critically ill patients [21–23]. In contrast, plasma presepsin showed poor predictive power for 28-day mortality in the overall cohort; however, in the SA-AKI subgroup, plasma presepsin was a better predictor of 28-day mortality than the APACHE II and SOFA scores. In other words, because plasma presepsin is affected by kidney function, its usefulness may be reduced in cases where kidney function changes dynamically, such as in AKI patients requiring CKRT. In contrast, presepsin is highly specific to infection, rises quickly during infection, and has a short half-life; therefore, in the case of SA-AKI, plasma presepsin measured before CKRT initiation can better reflect the severity of infection [9,24,25]. This study showed that plasma presepsin could be a useful prognostic indicator in patients with SA-AKI undergoing CKRT.
In the SA-AKI subgroup, the Kaplan-Meier survival analysis, performed in the high- and low-presepsin groups according to the plasma presepsin cutoff level of 1,951 pg/mL, showed a clear difference in the survival rate between the two groups. The 28-day survival rates in the high-presepsin and low-presepsin groups were 15.8% and 68.8%, respectively. Furthermore, in the multivariate Cox regression analysis, which included several confounding factors reflecting patient prognosis, plasma presepsin was observed to be an independent risk factor for predicting 28-day mortality in the SA-AKI subgroup. Additionally, in the multivariate Cox regression analysis, performed using the APACHE II and SOFA scores, plasma presepsin remained an independent risk factor (Supplementary Table 1, available online). The results of this study are expected to help determine patient prognosis by measuring plasma presepsin levels before CKRT initiation in patients with SA-AKI in clinical settings. Furthermore, screening patients with high plasma presepsin levels and providing them with more meticulous management can improve patient prognosis.
Arterial lactate, CRP, and PCT levels are widely used laboratory indicators; their correlation with plasma presepsin levels was analyzed. The correlations between plasma presepsin and CRP levels, and between plasma presepsin and PCT levels, showed a moderate correlation (Spearman’s ρ = 0.482 and Spearman’s ρ = 0.546, respectively). However, plasma presepsin and arterial lactate levels showed no correlation (Spearman’s ρ = 0.033) (Supplementary Fig. 2, available online). In particular, arterial lactate showed good AuROC values in both the overall cohort and the SA-AKI subgroup (Supplementary Fig. 1, available online). This appears to be because arterial lactate is not infection-specific and increases in all situations with tissue hypoxia. On the other hand, plasma presepsin showed a notable increase in AuROC value in the SA-AKI subgroup, showing a predictive value comparable to that of arterial lactate, which was superior to CRP or PCT (Supplementary Fig. 1, available online). This result is consistent with those of previous studies that analyzed the prognostic value of plasma presepsin, CRP, and PCT levels in critically ill patients [26,27]. In this study, we observed a novel finding that the 28-day mortality predictive power of plasma presepsin was superior to that of CRP or PCT, even in patients who underwent CKRT for SA-AKI.
This study had several limitations. First, it was a single-center retrospective study and the number of patients analyzed was relatively small. In this study, the highest AuROC value of plasma presepsin was observed in the SA-AKI subgroup, but the Delong test results with the AuROC values of APACHE II and SOFA scores did not show statistical significance (Delong test p = 0.19 for plasma presepsin vs. APACHE II score, Delong test p = 0.30 for plasma presepsin vs. SOFA score). These results may be due to the small sample size of this study. Second, the Simplified Acute Physiology Score 3 (SAPS 3) was not used as an indicator to evaluate critically ill patients in this study. In a previous study, SAPS 3 was reported to be superior to the APACHE II and SOFA scores [28]. However, the SAPS-3 score is calculated using results at the time of intensive care unit (ICU) admission. Because we had to analyze patient data at the time of CKRT, not at ICU admission, we used APACHE II and SOFA scores. Third, the APACHE II score is already a well-known, reliable indicator for evaluating critically ill patients, but in this study, in the SA-AKI subgroup, although the APACHE II score tended to be higher in the non-survivor group, there was no statistical significance (p = 0.12). We are not able to rule out the possibility that this was due to the significantly lower serum creatinine level and the prevalence of hypertension in the non-survivor group. These results imply that several confounding factors may have intervened due to the small sample size of this study. However, despite these limitations, this study is considered valuable in that it identified the potential utility of plasma presepsin in patients requiring CKRT for SA-AKI for the first time.
In conclusion, plasma presepsin levels had the highest predictive power for 28-day mortality compared to the well-known APACHE II and SOFA scores in the SA-AKI subgroup. It was also found to be an independent risk factor after adjusting for other risk factors. Additionally, plasma presepsin was the most useful biomarker of 28-day mortality, compared to PCT and CRP, in the SA-AKI subgroup. Therefore, the plasma presepsin level has the potential to be a strong prognostic indicator in patients with SA-AKI receiving CKRT. The usefulness of plasma presepsin levels in individuals undergoing CKRT should be studied in a larger cohort.
Supplementary Materials
Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.23.301).
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This study was supported by the Young Investigator Research Grant from the Korean Nephrology Research Foundation (2021-08-004) and the National Research Foundation of Korea (NRF) grant, funded by the Korean government (MSIT) (No. RS-2022-00166592). This manuscript is a researcher-led paper with no specific participation by funders.
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization, Formal analysis: GBL, YP
Data curation, Investigation: GBL, JWL, SHY, WMH, SRY, DHK, YP
Funding acquisition: YP
Writing–original draft: GBL, YP
Writing–review & editing: All authors
All authors read and approved the final manuscript.