Reduced Responsiveness of Blood Leukocytes to Lipopolysaccharide Does not Predict Nosocomial Infections in Critically Ill Patients

ABSTRACT Critically ill patients show signs of immune suppression, which is considered to increase vulnerability to nosocomial infections. Whole-blood stimulation is frequently used to test the function of the innate immune system. We here assessed the association between whole-blood leukocyte responsiveness to lipopolysaccharide (LPS) and subsequent occurrence of nosocomial infections in critically ill patients admitted to the intensive care unit (ICU). All consecutive critically ill patients admitted to the ICU between April 2012 and June 2013 with two or more systemic inflammatory response syndrome criteria and an expected length of ICU stay of more than 24 h were enrolled. Age- and sex-matched healthy individuals were included as controls. Blood was drawn the first morning after ICU admission and stimulated ex vivo with 100 ng/mL ultrapure LPS for 3 h. Tumor necrosis factor-&agr;, interleukin-1&bgr; (IL-1&bgr;), and IL-6 were measured in supernatants. Seventy-three critically ill patients were included, of whom 10 developed an ICU-acquired infection. Compared with healthy subjects, whole-blood leukocytes of patients were less responsive to ex vivo stimulation with LPS, as reflected by strongly reduced tumor necrosis factor-&agr;, IL-1&bgr;, and IL-6 levels in culture supernatants. Results were not different between patients who did and those who did not develop an ICU-acquired infection. The extent of reduced LPS responsiveness of blood leukocytes in critically ill patients on the first day after ICU admission does not relate to the subsequent development of ICU-acquired infections. These results argue against the use of whole-blood stimulation as a functional test applied early after ICU admission to predict nosocomial infection.


INTRODUCTION
Nosocomial infection is a common complication in critically ill patients and strongly associated with prolonged stay in the intensive care unit (ICU) and increased morbidity and mortality (1). Critical illness results in a disturbed homeostasis leading to features of both hyperinflammation and immune suppression (2). Although originally immune suppression was considered to follow the initial proinflammatory phase (3,4), a recent study conducted in trauma and burn patients revealed evidence for both activation and impairment of immune pathways within hours of the injury using whole-blood leukocyte transcriptional profiling (5). A hallmark functional feature of immune suppression accompanying critical illness is the reduced capacity of whole blood to produce proinflammatory cytokines on stimulation with bacterial agonists ex vivo (3,4,6).
Immune suppression is considered an important risk factor for secondary infection in ICU patients, and immunostimulatory therapy is a newly proposed treatment strategy in this population (4). A clinical challenge herein is to identify patients who may benefit from such therapies, that is, those at high risk for nosocomial infections. Although multiple ways of immune monitoring have been developed (6), whole-blood stimulation tests provide the potential advantage that they can be done in a rapid and reproducible way in a routine setting (7). In this study, we aimed to determine whether the extent of reduced whole-blood leukocyte responsiveness to lipopolysaccharide (LPS) relates to the subsequent development of ICU-acquired infections in critically ill patients.

Patient population and study design
This study was incorporated in the MARS (Molecular Diagnosis and Risk Stratification of Sepsis) project (clinicaltrials.gov identifier NCT01905033) (8,9). Consecutive patients older than 18 years admitted between April 2012 and June 2013 to the ICU of the Academic Medical Center (Amsterdam, the Netherlands) were screened by trained research physicians. Patients were included when they had at least two systemic inflammatory response syndrome criteria on the day of ICU admission (body temperature e36-C or Q38-C, tachycardia 990/min, tachypnea 920/min or pCO 2 G4.3 kPA, leukocyte count G4 Â 10 9 /L or 912 Â 10 9 /L) (10). Patients transferred from other ICUs, receiving antibiotics for more than 48 h before ICU admission, and/or with an expected length of ICU stay less than 24 h were excluded. Patients were subjected to a single blood draw at 9:00 AM of the first morning after ICU admission. Healthy subjects were matched with regard to age, sex, and timing of blood draw. The medical ethical committee of the Academic Medical Center in Amsterdam gave ethical approval for the conduction of the study (no. NL 34294.018.10), and the study was registered at the Central Committee for Human Research. Written informed consent was obtained from all patients (or legal representative) and healthy controls. Heparin-anticoagulated whole blood was stimulated for 3 h at 37-C in pyrogen-free RPMI 1640 (Life Technologies, Bleiswijk, the Netherlands) with or without 100 ng/mL ultrapure LPS (from Escherichia coli 0111:B4; InvivoGen, Toulouse, France). Tumor necrosis factor-! (TNF-!), interleukin-6 (IL-6), and IL-1" were measured in supernatants using a cytometric bead array assay (BD Biosciences, San Jose, Calif). Lipopolysaccharide-induced cytokine release was calculated by subtraction of cytokine levels in samples incubated without LPS from those measured in samples obtained after incubation with LPS.
An ICU-acquired infection was defined by the systemic therapeutic administration of antibiotics for a suspected new infection of more than 48 h after ICU admission. The presence of infection (either on admission or ICU acquired) was established for every affected organ or site using Center for Disease criteria (11) and International Sepsis Forum consensus definitions (12) as described in detail elsewhere (8). Dedicated research physicians categorized the plausibility of infection based on a post hoc review of all available clinical, radiological, and microbiological evidence (8); patients treated for a suspected infection but with a post hoc infection likelihood of none were not considered infectious.

Statistical analysis
All results are presented as numbers (percentages) for categorical variables, median and interquartile ranges (IQRs) for nonparametric quantitative variables, and mean T SD for parametric quantitative variables. Continuous nonparametric data were analyzed using a Mann-Whitney U test or a Kruskal-Wallis test; categorical data were analyzed using a / 2 or Fisher exact test. Continuous parametric data were analyzed using a Student t test or analysis of variance when appropriate. A value of P G 0.05 was considered statistically significant.

Whole-blood stimulations
Whole-blood leukocytes of critically ill patients, harvested on the first morning after ICU admission, released significantly less TNF-!, IL-1", and IL-6 on stimulation with LPS than blood leukocytes from healthy controls (Fig. 1). Whole blood from patients with a sepsis admission diagnosis released less TNF-! when compared with whole blood from patients with a noninfectious admission diagnosis (P = 0.002), whereas IL-1" and IL-6 release did not differ between these groups (Fig. 2). The whole-blood cytokine production capacity did not differ between patients who subsequently developed an ICUacquired infection and ICU controls (Fig. 1). Similarly, when analyzed separately, whole-blood cytokine production capacity in patients with a sepsis admission diagnosis did not differ between those who did and those who did not develop an ICUacquired infection (data not shown).

DISCUSSION
In accordance with previous investigations (3,4,6), we report a strongly impaired release of TNF-!, IL-1", and IL-6 in LPS-stimulated whole blood obtained from critically ill patients when compared with healthy controls. However, the extent of the reduction in cytokine release capacity did not relate to the subsequent development of ICU-acquired infections. These results argue against the use of whole-blood stimulation as a functional test of innate immunity applied early after ICU admission to predict nosocomial infection.
The incidence of secondary infection in our ICU cohort was 13.7%. Previous studies have reported incidence rates varying between 9% and 37%, largely dependent on the population studied and the definitions used (1). Our study was conducted in a mixed surgical-medical ICU in an academic hospital. We used strict definitions and post hoc classification by dedicated research physicians to diagnose ICU-acquired infections (8).
In theory, functional tests, such as ex vivo stimulation of whole blood, represent the best method to establish the function of the innate immune system because they directly measure the capacity of relevant cells to react to a microbial challenge (6). Whole-blood stimulation is an easy-to-perform test that could be implemented in routine practice (7). In the present study, we intentionally used a short incubation period (3 h) considering that, if deemed clinically relevant, the test should yield results relatively quickly. The cytokines measured on a 3-h LPS stimulation of whole blood likely are mainly produced by monocytes. In accordance, monocytes demonstrated a reduced capacity to release proinflammatory cytokines in response to LPS in a variety of clinical settings, including sepsis and noninfectious systemic inflammatory conditions (3,6). In our cohort, whole-blood cytokine production corrected for absolute monocyte counts (collected in 81% of the patients included) also did not discriminate between patients who did and those who did not develop an ICUacquired infection (data not shown).
The results obtained with LPS-induced whole-blood stimulation were not able to predict the subsequent development of ICU-acquired infections. One might argue that our sample size was too small. However, although median TNF-! levels were slightly lower in patients who developed a secondary infection, based on the variation in TNF-! concentrations in the 73 patients included in the present study, we calculated that a study encompassing more than 800 patients would be required to show a statistically significant difference. Hence, such a test will unlikely be of clinical value in daily practice. We observed a relatively uniformly depressed blood leukocyte Previous studies investigated the value of surrogate markers of suppression of the adaptive immune system to predict secondary infections in ICU patients, especially in those admitted with sepsis. Specifically, reduced expression of human leukocyte antigenYDR and increased expression of programmed cell death (PD)-1, PD-ligand 1, and PD-ligand 2 on blood monocytes, determined by flow cytometry 3 to 5 days after ICU admission, correlated with an enhanced incidence of secondary infections in patients with septic shock (13,14). Measurements were not done earlier after ICU admission. In both earlier investigations, the incidence of nosocomial infections was much higher (24.2% and 29.7%, respectively) (13,14) than observed here (13.7%), suggesting that the populations studied and/or the definitions used for ICU-acquired infection differed. We performed whole-blood stimulation on the first morning after ICU admission, at 9:00 AM, seeking to evaluate a potential early test and avoiding potential circadian variation. Further research is needed to establish whether whole-blood stimulation conducted at later time points after ICU admission can assist in identifying patients at risk for nosocomial infections. Indeed, in a study encompassing 70 critically ill children, among whom 30 with sepsis, a reduced ex vivo LPS-induced TNF-! response in whole blood on day 7 after admission was associated with development of nosocomial infection and death (15).

CONCLUSIONS
The extent of reduced LPS responsiveness of whole-blood leukocytes in critically ill patients determined on the first FIG. 1. Whole-blood leukocyte responsiveness to LPS in critically ill patients. Similarly reduced responsiveness of whole-blood leukocytes to LPS in critically ill patients who do and those who do not develop a secondary infection. Whole blood was drawn from 73 patients (10 of whom subsequently developed a secondary infection) on the first morning (9:00 AM) after admission to the ICU and from 18 age-and sex-matched healthy controls. Blood was stimulated for 3 h with ultrapure LPS (100 ng/mL), and TNF-!, IL-1", and IL-6 were measured in supernatants. Data are expressed as box and whisker diagrams depicting the median and lower quartile, upper quartile, and their respective 1.5 IQR as whiskers (as specified by Tukey). Data beyond the end of the whiskers are outliers and plotted as points. HV, healthy volunteers; ICU-AI, ICU-acquired infection. *P e 0.05, **P e 0.01, ***P e 0.001.

FIG. 2.
Whole-blood leukocytes from sepsis patients show reduced LPS-induced TNF-! production compared with noninfectious patients. Whole blood was drawn from 73 patients (47 of whom had sepsis on admission) on the first morning (9:00 AM) after admission to the ICU and from 18 age-and sexmatched healthy controls. Blood was stimulated for 3 h with ultrapure LPS (100 ng/mL), and TNF-!, IL-1", and IL-6 were measured in supernatants. Data are expressed as box and whisker diagrams depicting the median and lower quartile, upper quartile, and their respective 1.5 IQR as whiskers (as specified by Tukey). Data beyond the end of the whiskers are outliers and plotted as points. HV, healthy volunteers. **P e 0.01, ***P e 0.001. morning after ICU admission does not relate to the subsequent development of ICU-acquired infections.