American Journal of Respiratory and Critical Care Medicine

Rationale: Sepsis can be complicated by secondary infections. We explored the possibility that patients with sepsis developing a secondary infection while in the intensive care unit (ICU) display sustained inflammatory, vascular, and procoagulant responses.

Objectives: To compare systemic proinflammatory host responses in patients with sepsis who acquire a new infection with those who do not.

Methods: Consecutive patients with sepsis with a length of ICU stay greater than 48 hours were prospectively analyzed for the development of ICU-acquired infections. Twenty host response biomarkers reflective of key pathways implicated in sepsis pathogenesis were measured during the first 4 days after ICU admission and at the day of an ICU-acquired infection or noninfectious complication.

Measurements and Main Results: Of 1,237 admissions for sepsis (1,089 patients), 178 (14.4%) admissions were complicated by ICU-acquired infections (at Day 10 [6–13], median with interquartile range). Patients who developed a secondary infection showed higher disease severity scores and higher mortality up to 1 year than those who did not. Analyses of biomarkers in patients who later went on to develop secondary infections revealed a more dysregulated host response during the first 4 days after admission, as reflected by enhanced inflammation, stronger endothelial cell activation, a more disturbed vascular integrity, and evidence for enhanced coagulation activation. Host response reactions were similar at the time of ICU-acquired infectious or noninfectious complications.

Conclusions: Patients with sepsis who developed an ICU-acquired infection showed a more dysregulated proinflammatory and vascular host response during the first 4 days of ICU admission than those who did not develop a secondary infection.

Scientific Knowledge on the Subject

Recent observational studies have found that patients with sepsis show signs of prolonged immune suppression, which has been postulated to enhance susceptibility toward secondary infections, thereby contributing to late sepsis mortality. Indeed, several investigators have documented a variety of immune defects in patients with sepsis, such as hyporesponsiveness and a profound loss of innate and adaptive immune cells. However, a systemic hyperinflammatory reaction is not captured by the assays used to study immune suppression in previous investigations.

What This Study Adds to the Field

Patients with sepsis who went on to develop an intensive care unit (ICU)-acquired infection demonstrated a more dysregulated host response during the first 4 days after admission, as reflected by enhanced inflammation, stronger endothelial cell activation, a more disturbed vascular integrity, and evidence for enhanced coagulation activation. This enhanced hyperinflammation was sustained up to the day of ICU-acquired infection development, and no differences were found between the host response during ICU-acquired infection and noninfectious ICU-acquired complications (acute kidney injury or acute respiratory distress syndrome). Although this study does not contradict earlier investigations reporting immune suppression in patients with sepsis, it indicates that patients with sepsis who develop secondary infections while in the ICU also demonstrate hyperinflammation.

Sepsis is characterized by an injurious host response to an infection, and a leading cause of hospitalization, morbidity, and mortality (1, 2). Most research seeking to obtain insight into mechanisms contributing to sepsis mortality focused on early lethality, presumably caused by an overzealous activation of the innate immune system in response to acute infection (3, 4). However, most deaths in sepsis occur more than 1 week after admission to the intensive care unit (ICU) (58). This relatively late sepsis mortality has received much attention in recent years, and it has been suggested that immune suppression and, as consequence thereof, ICU-acquired infections are key causative denominators herein (3, 6, 911). Indeed, a variety of immune defects have been documented in ICU patients with sepsis, most notably impaired responsiveness of immune cells to bacterial antigens and a profound loss of T and B cells because of apoptotic cell death (3, 10, 11).

We recently reported on the incidence, risk factors, and attributable mortality of ICU-acquired infections in patients admitted to the ICU with sepsis (12). In a prospectively enrolled cohort consisting of 1,719 consecutive sepsis admissions, ICU-acquired infections occurred in 13.5% of cases, bearing a population-attributable mortality fraction of 10.9% by Day 60 (12). Although earlier studies on susceptibility to ICU-acquired infections focused on immune suppression (3, 4, 911, 13), we here explored the possibility that the more dysregulated host response in patients with sepsis who acquire an infection while in the ICU is also reflected by a systemic hyperinflammatory reaction that is not captured by the assays used to delineate immune suppression. As such, the primary objective of this study was to compare systemic proinflammatory host responses in patients with sepsis who during their ICU stay acquire a new infection with those who do not. For this we performed a substudy in the previously described cohort (12), and report the levels of 20 host response biomarkers reflective of key pathways implicated in sepsis pathogenesis, measured during the first 4 days after ICU admission and at onset of ICU-acquired infection.

Study Design and Population

This was an explorative substudy of a previously reported cohort used to determine the incidence and attributable mortality of ICU-acquired infections in critically ill patients with sepsis (12). The study was conducted as part of the MARS (Molecular Diagnosis and Risk Stratification of Sepsis) project, a prospective observational cohort study in the mixed ICUs of two tertiary teaching hospitals in the Netherlands (12, 14, 15). Consecutive patients older than 18 years of age admitted to the two ICUs were included via an opt-out method approved by the medical ethical committees of the participating hospitals, the Academic Medical Center in Amsterdam and the University Medical Center in Utrecht. Both ICUs used protocolized care, including selective decontamination of the digestive tract (12, 16).

For every admitted patient the plausibility of an infection was assessed daily using a four-point scale (ascending from none, possible, probable, to definite) (14). Sepsis was defined as the presence of infection diagnosed within 24 hours after ICU admission with a probable or definite likelihood, accompanied by at least one additional parameter as described in the 2001 International Sepsis Definitions Conference (17). ICU-acquired infection was defined as any new-onset infection (with likelihood possible, probable, or definite) starting greater than 48 hours after ICU admittance for which the clinical team started a new antibiotic regimen. Organ failures, shock, and comorbidities were defined as described in the online supplement; acute kidney injury (AKI) and acute respiratory distress syndrome (ARDS) (18, 19) were deemed ICU-acquired when starting greater than 48 hours after ICU admittance.

For the current study consecutive patients with sepsis admitted to the ICU from January 2011 until July 2013 with a length of ICU stay greater than 48 hours were analyzed. Patients with an infection diagnosed greater than or equal to 24 hours but less than or equal to 48 hours after ICU admission were excluded because their infection could not with certainty be deemed the reason for admission or ICU-acquired. Readmissions, defined as any second admission within the 2.5-year study period, were analyzed as new unique admissions. For patients who were readmitted to the ICU demographic and long-term follow-up data (≥30 d) are shown for the first ICU admission only.

Sampling and Assays

Daily (on admission and at 6 a.m. thereafter) ethylenediaminetetraacetic acid anticoagulated plasma harvested from blood obtained for regular patient care was stored within 4 hours after blood draw at −80°C. For assays, see the online supplement.

Statistical Analysis

Biomarker measurements were analyzed using all unique admissions, and readmitted patients were not excluded. Biomarkers were transformed to their 10 log scale for plotting purposes. Biomarker distribution over time was analyzed using a general mixed model analysis in which a linear regression model was fitted on logarithmically transformed biomarker data using the different data time points (i.e., admission, Day 2, and Day 4). Different mixed models were fitted taking the group (ICU-acquired infection vs. no ICU-acquired infection), time, and their interaction as fixed effects and patient-specific intercept and slope of time as random effects. The model with the best fit was regarded most appropriate. The overall P value reported in the figures and the tables is derived from the fixed-effect model in which group + time was used in addition to the random effects model, unless otherwise specified. The rate of change in biomarker levels over time was analyzed using the mixed effects model using the regression coefficient of time alone or in combination with the interaction between group and time when significant. In addition, this regression coefficient was transformed into percentage change over time. Biomarker distribution at a single time point was compared using a nonparametric Mann-Whitney U test. Multiple-comparison–adjusted (Benjamini-Hochberg) P value less than 0.05 defined significance of plasma biomarkers. For more details, see the online supplement.

Propensity Score Matching

Considering that the release of host response biomarkers in sepsis often is proportional to disease severity (20), propensity score matching was used in patients with biomarkers to account for disease severity on ICU admission and other baseline differences between patients who did and those who did not develop an ICU-acquired infection. A logistic regression implemented in the R library MatchIt version 2.4–21 (http://gking.harvard.edu/matchit) (21) was used, including variables associated with disease severity and baseline variables that were different between groups. The propensity score included Acute Physiology and Chronic Health Evaluation (APACHE)-IV score, Sequential Organ Failure Assessment (SOFA) score, source of infection, and shock, all on ICU admission. Patients developing an ICU-acquired infection were matched 1:3 to patients without the development of ICU-acquired infection, using nearest matching with a caliper of 0.20 SD of the normally distributed propensity score. If less than three control subjects could be matched, fewer matches were allowed, making optimal use of the control subjects. The individual time points (i.e., admission, Day 2, and Day 4) were analyzed separately taking clustering of matching into account by including match-pair identifiers in a mixed model analysis. Both the mixed model analysis and the Mann-Whitney U test showed similar results and for consistency the Mann-Whitney U test is reported.

Patients

We studied 1,237 admissions for sepsis with a length of ICU stay greater than 48 hours (1,089 patients) (Figure 1). Of these, 178 admissions (14.4%), concerning 156 patients, were complicated by one or more ICU-acquired infections, involving a total of 262 ICU-acquired infections. Patients developing an ICU-acquired infection were more often admitted for primary bacteremia and less often for neurologic and soft tissue infection (Table 1). Patients with sepsis who developed a secondary infection while in the ICU were more severely ill on admission than those who did not, as reflected by higher APACHE IV and SOFA scores, and a higher proportion of shock (Table 1).

Table 1. Baseline Characteristics and Outcome of Patients Admitted with Sepsis Stratified according to Development of ICU-acquired Infection or Not

 ICU-acquired InfectionNo ICU-acquired InfectionP Value
Patients156933 
Demographics   
 Age, yr, mean (SD)60 (13.9)60 (14.9)0.96
 Male sex, n (%)100 (64.1)542 (58.1)0.17
Chronic comorbidity   
 Any comorbidity, n (%)118 (75.6)682 (73.1)0.53
 Immunocompromised state, n (%)44 (28.2)219 (23.5)0.22
 Charlson comorbidity index, median (IQR)4 (3–6)4 (2–6)0.93
Admissions1781,059 
Source of sepsis admission diagnosis, n (%)   
 Pulmonary tract73 (41.0)431 (40.7)>0.99
 Abdomen40 (22.5)206 (19.5)0.34
 Bloodstream infection11 (6.2)24 (2.3)0.01
 Catheter-related bloodstream infection4 (2.2)18 (1.7)0.76
 Neurologic3 (1.7)61 (5.8)0.02
 Urinary tract9 (5.1)72 (6.8)0.41
 Soft tissue infection2 (1.1)51 (4.8)0.02
 Other*36 (20.2)196 (18.5)0.60
Admission type, n (%)   
 Medical138 (77.5)801 (75.6)0.65
 Readmission22 (12.4)126 (11.9)0.91
Severity of disease   
 APACHE IV score, mean (SD)92 (27.5)82 (27.9)<0.0001
 SOFA score, median (IQR)8 (6–11)7 (5–9)<0.0001
 Shock, n (%)82 (46.1)371 (35.0)<0.01
Corticosteroid treatment in the first 4 d after ICU admission, n (%)   
 Any hydrocortisone use118 (66.3)605 (57.1)0.03
 Hydrocortisone >200 mg/d100 (56.2)463 (43.7)<0.01
 SDD use123 (69.1)705 (66.6)0.54
Outcome   
 Length of ICU stay, d, median (IQR)24 (15–34)6 (4–10)<0.0001
 Length of hospital stay, d, median (IQR)37 (22–66)20 (10–39)<0.0001
 Mortality, n (%)   
  ICU§69 (38.8)174 (16.4)0.001
  Hospital||81 (51.9)267 (28.6)<0.0001
  30 d||49 (31.4)237 (25.4)0.13
  60 d||73 (46.8)281 (30.1)<0.0001
  90 d||82 (52.6)308 (33.0)<0.0001
  1 yr||91 (58.3)406 (43.4)<0.0001

Definition of abbreviations: APACHE IV = Acute Physiology and Chronic Health Evaluation IV; ICU = intensive care unit; IQR = interquartile range; SDD = selective decontamination of the digestive tract; SOFA = Sequential Organ Failure Assessment.

*Other infections include lung abscess, sinusitis, pharyngitis, tracheobronchitis, endocarditis, mediastinitis, myocarditis, postoperative wound infection, bone and joint infection, oral infection, eye infection, and viral infections.

Use of hydrocortisone or its equivalent (hydrocortisone dose = 4 × prednisolone dose, 5 × methylprednisolone dose, 25 × dexamethasone dose).

Patients not on SDD received selective oropharyngeal decontamination.

§ICU mortality was calculated using all ICU admissions for sepsis.

||Follow-up data were calculated using the first ICU admission for sepsis for each patient during the study period; readmissions were not included in this analysis.

Twenty-six patients were lost to 1-year follow-up (3.8% in patients with sepsis developing an ICU-acquired infection and 2.1% in patients with sepsis with no ICU-acquired infection; P = 0.25).

ICU-acquired Infections and Outcome

The first ICU-acquired infections occurred at a median of Day 10 (interquartile range, 6–13). The most common ICU-acquired infections were catheter-related bloodstream infections (n = 73; 27.9%), pneumonia (n = 64; 24.4%), and abdominal infection (n = 42; 16.0%) (Table 2). The most common causative pathogens were gram-positive bacteria (n = 118; 45.0%), followed by gram-negative bacteria (n = 74; 28.2%), fungi (n = 24; 9.2%), and viruses (n = 27; 10.3%) (Table 2).

Table 2. Characteristics of ICU-acquired Infections

Number and timing of infections 
 Admissions associated with an ICU-acquired infection, n (%)178 (14.4)
 ICU-acquired infections262
 Admissions associated with multiple ICU-acquired infections, n (%)61 (34.3)
 Day of first ICU-acquired infection, median (IQR)10 (6–13)
Source of infection, n (%) 
 Pulmonary64 (24.4)
  Hospital-acquired pneumonia18 (6.9)
  Ventilator-associated pneumonia46 (17.6)
 Cardiovascular85 (32.4)
  Bacteremia12 (4.6)
  Catheter-related bloodstream infection73 (27.9)
 Abdomen42 (16.0)
  Abdominal infection41 (15.6)
  Gastrointestinal infection1 (0.4)
 Neurologic3 (1.1)
  Primary meningitis1 (0.4)
  Secondary meningitis2 (0.8)
 Soft tissue infection11 (4.2)
 Urinary tract3 (1.1)
 Other*54 (20.6)
Causative pathogen, n (%) 
 Gram-positive bacteria118 (45.0)
 Gram-negative bacteria74 (28.2)
 Fungi24 (9.2)
 Viral (including reactivation)27 (10.3)
 Other8 (3.1)
 Unknown64 (24.4)

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile range.

*Other infections include lung abscess, sinusitis, pharyngitis, tracheobronchitis, endocarditis, mediastinitis, myocarditis, postoperative wound infection, bone and joint infection, oral infection, eye infection, and viral infections.

Percentages depict the portion of ICU-acquired infections (total n = 262) caused by the pathogen group indicated. In total 251 pathogens were assigned to 262 ICU-acquired infections; in 51 (19.5%) of all ICU-acquired infections more than one pathogen was assigned as causative.

The median ICU length of stay was longer in patients who acquired a secondary infection than in those who did not (24 [15–24] vs. 6 [4–10] d, respectively; P < 0.001) (Table 1). ICU mortality was higher in patients developing a secondary infection than in patients who did not (38.8% vs. 16.4%; P < 0.001); the mortality difference between groups remained until 1 year after ICU admission (Table 1).

Host Response Biomarkers in Patients with Sepsis Who Did and Those Who Did Not Develop an ICU-acquired Infection

In a subset of patients (n = 1,010; 81.6%), biomarkers indicative of the host response during sepsis were measured. Patients with sepsis displayed a profound systemic inflammatory reaction (elevated plasma levels of interleukin (IL)-6, IL-8, IL-10, and matrix metalloproteinase-8) (Figure 2; see Figure E1 in the online supplement), activation of the vascular endothelium (elevated plasma concentrations of soluble E-selectin, soluble intercellular adhesion molecule-1 [ICAM-1], and fractalkine), increased loss of vascular integrity (increased levels of angiopoietin-2 and reduced levels of angiopoietin-1) (Figure 3), and a net procoagulant state (elevated plasma levels of D dimer, reduced levels of the anticoagulants antithrombin and protein C, and prolonged activated partial thromboplastin time [aPTT] and prothrombin time) (Figure 4). Most of these characteristic sepsis responses were exaggerated in patients who developed an ICU-acquired infection relative to those who did not, significantly so for IL-6, IL-8, IL-10, soluble ICAM-1, fractalkine, angiopoietin-2, the angiopoietin 2:1 ratio, and aPTT (all P < 0.01). Platelet counts were significantly lower in patients who developed an ICU-acquired infection (P < 0.001 vs. those who did not). This more disturbed host response remained after exclusion of readmissions (see Table E1). Plasma levels of tumor necrosis factor-α, IL-1β, IL-13, granulocyte–macrophage colony–stimulating factor, and IFN-γ were undetectable in most patients and not different between groups (data not shown). The rate of biomarker change in the first 4 days was comparable between patients developing ICU-acquired infections and patients that did not except for fractalkine and platelets (higher in the former group) (see Table E2).

Considering that patients who went on to acquire a secondary infection while on the ICU had higher baseline APACHE IV and SOFA scores than those who did not, and considering that the levels of host response biomarkers in sepsis often are proportional to disease severity (20), we matched patients who did and those who did not develop an ICU-acquired infection for disease severity on ICU admission. A total of 133 admissions complicated by ICU-acquired infection were matched to 322 admissions without ICU-acquired infection with comparable disease severity and source of infection on ICU admission (Table 3). In this matched cohort, many sepsis host response biomarkers remained more aberrant in patients who developed an ICU-acquired infection (Table 4), significantly so for IL-6, IL-8, and IL-10. This outcome was consistent in sensitivity analyses including immunocompromised state in the matching procedure or including immunocompromised state and corticosteroid treatment during the first 4 days of ICU stay in the matching procedure (see Tables E3–E6).

Table 3. Baseline Characteristics of Patients Admitted with Sepsis Who Did and Those Who Did Not Develop an ICU-acquired Infection Propensity Matched for APACHE IV Score, SOFA Score, Source of Infection, and Shock on ICU Admission

 ICU-acquired InfectionNo ICU-acquired InfectionP Value
Patients123301 
Demographics   
 Age, yr, mean (SD)60.9 (14.0)61.1 (14.0)0.93
 Sex, male, n (%)74 (60.2)166 (55.1)0.38
Chronic comorbidity   
 Any comorbidity, n (%)90 (73.2)237 (78.7)0.25
 Immunocompromised state, n (%)33 (26.8)74 (24.6)0.70
 Charlson comorbidity index, median (IQR)4 (3–6)4 (3–6)0.21
Admissions133322 
Source of infection, n (%)   
 Pulmonary tract56 (42.1)145 (45.0)0.61
 Abdomen33 (24.8)77 (23.9)0.90
 Cardiovascular5 (4.0)13 (4.0)>0.99
 Neurologic3 (2.3)9 (2.8)0.78
 Urinary tract7 (5.3)20 (6.2)0.83
 Skin sepsis2 (1.5)5 (1.6)>0.99
 Other*27 (20.3)53 (16.5)0.33
Admission type, n (%)   
 Medical102 (76.7)246 (76.4)>0.99
 Readmission10 (7.5)21 (6.5)0.70
Severity of disease   
 APACHE IV score, mean (SD)89 (26.7)85 (26.0)0.10
 SOFA score, median (IQR)8 (6–10)8 (6–10)0.33
 Shock, n (%)55 (41.4)128 (39.8)0.76
Corticosteroid treatment in the first 4 d after ICU admission, n (%)   
 Any hydrocortisone use89 (66.9)189 (58.7)0.12
 Hydrocortisone >200 mg/d73 (54.9)147 (45.6)0.08
 SDD use91 (68.4)225 (69.9)0.82
Outcome   
 Length of ICU stay, d, median (IQR)24 (15–35)7 (4–10)<0.0001
 Length of hospital stay, d, median (IQR)35 (22–65)20 (10–43)<0.0001
 Mortality, n (%)   
  ICU§54 (39.4)71 (21.6)<0.001
  Hospital||70 (53.8)100 (33.0)<0.001
  30 d||44 (33.8)86 (28.4)0.30
  60 d||64 (49.2)102 (33.7)<0.01
  90 d||70 (53.8)107 (35.3)<0.001
  1 yr||80 (61.5)148 (48.8)<0.01

Definition of abbreviations: APACHE IV = Acute Physiology and Chronic Health Evaluation IV; ICU = intensive care unit; IQR = interquartile range; SDD = selective decontamination of the digestive tract; SOFA = Sequential Organ Failure Assessment.

*Other infections include lung abscess, sinusitis, pharyngitis, tracheobronchitis, endocarditis, mediastinitis, myocarditis, postoperative wound infection, bone and joint infection, oral infection, eye infection, and viral infections.

Use of hydrocortisone or its equivalent (hydrocortisone dose = 4 × prednisolone dose, 5 × methylprednisolone dose, 25 × dexamethasone dose).

Patients not on SDD received selective oropharyngeal decontamination.

§ICU mortality was calculated using all ICU admissions for sepsis.

||Follow-up data were calculated using the first ICU admission for sepsis for each patient during the study period; readmissions were not included in this analysis.

Twelve patients were lost to 1-year follow-up (4.6% in patients with sepsis developing an ICU-acquired infection and 2.0% in patients with sepsis with no ICU-acquired infection; P = 0.19).

Table 4. Plasma Biomarkers in Patients Admitted with Sepsis Who Did and Did Not Develop an ICU-acquired Infection Propensity Matched for APACHE IV Score, SOFA Score, Source of Infection, and Shock on ICU Admission

 AdmissionDay 2Day 4Overall P Value
ICU-acquired Infection (n = 133)No ICU-acquired Infection (n = 322)ICU-acquired Infection (n = 127)No ICU-acquired Infection (n = 308)ICU-acquired Infection (n = 126)No ICU-acquired Infection (n = 234)
Inflammation       
 IL-6, pg/ml337.25 (63.8–1940.83)146.77 (41.78–770.55)113.38 (35.2–485.08)*69.82 (22.38–233.34)68.92 (25.02–200.65)33.39 (15.24–92.77)<0.0001
 IL-8, pg/ml203.8 (71.89–1003.74)133.17 (62.27–401.55)149.92 (55.7–542.58)79.10 (33.34–190.68)119.61 (53.99–276.66)62.80 (23.45–131.59)<0.0001
 IL-10, pg/ml19.64 (5.98–96.05)11.80 (4.46–31.33)13.65 (4.1–46.19)6.56 (2.46–15.43)10.40 (4.34–21.04)3.78 (2.29–8.97)<0.0001
 MMP-8, ng/ml2.93 (1.22–10.45)3.07 (0.96–9.62)3.12 (0.93–11.26)2.23 (0.7–6.39)2.16 (0.95–5.74)*1.32 (0.45–3.55)0.73
Endothelial cell activation       
 sE-selectin, ng/ml8.95 (4.76–23.12)9.33 (4.14–23.62)8.75 (4.66–18.74)9.25 (4.58–19.45)6.49 (4.05–17.88)7.83 (4.18–15.47)>0.99
 sICAM-1, ng/ml221.00 (128.12–333.49)188.46 (106.75–337.02)249.84 (157.06–400.87)218.94 (136.4–376.7)264.69 (181.95–379.76)237.18 (144.76–355.63)>0.99
 Fractalkine, pg/ml36.81 (19.87–82.73)26.96 (17.9–58.7)41.01 (22.57–87.25)25.03 (15.88–55.58)50.51 (24.07–123.79)24.07 (16.55–59.12)0.01
 ANG-1, ng/ml1.92 (0.8–6.85)1.78 (0.71–5.09)1.49 (0.67–3.57)1.36 (0.66–3.72)1.16 (0.54–3.01)1.62 (0.62–4.33)>0.99
 ANG-2, ng/ml9.01 (4.73–17.07)7.41 (3.75–15.23)11.59 (6.39–21.58)8.25 (3.73–17.68)9.54 (5.16–16.15)§6.03 (3.08–11.03)0.09
 ANG-2:ANG-13.96 (1.08–13.29)3.23 (0.93–16.64)7.67 (2.21–21.53)5.44 (1.34–23.97)7.69 (2.02–24.22)*3.57 (0.89–12.03)0.99
Coagulation activation       
 D dimer, μg/ml10.12 (3.63–17.8)9.20 (4.56–17.09)10.39 (4.44–19)9.25 (3.99–17.51)11.47 (5.84–17.63)9.66 (4.96–16.41)>0.99
 PT, s15.55 (12.78–18.52)15.15 (12.78–18.4)15 (12.6–17.2)14.90 (12.4–17.12)14.70 (12.5–16.3)13.70 (11.8–15.38)>0.99
 aPTT, s42 (32.5–54.5)38 (31–51)45.50 (35–52)40.00 (31.25–50)40.00 (32–51)43.00 (29–54)>0.99
 Protein C, ng/ml119.99 (95.45–163.1)122.14 (94.46–159.19)123.95 (88.92–157.85)125.96 (94.64–163.81)132.44 (98.17–171.16)138.23 (98.63–189.29)>0.99
 Antithrombin, ng/ml807.28 (490.91–1198.3)757.55 (527.75–1154.03)702.62 (417.08–1027.7)739.21 (457.62–1078.2)849.68 (500.09–1325.9)895.16 (595.66–1373.3)>0.99
 Platelets, 109/L209 (125–292)228 (133–318)147 (76–259)179 (103–287)140 (55–268)*198 (103–312)>0.99

Definition of abbreviations: ANG = angiopoietin; APACHE IV = Acute Physiology and Chronic Health Evaluation IV; aPTT = activated partial thromboplastin time; ICAM = intercellular adhesion molecule; ICU = intensive care unit; MMP = matrix metalloproteinase; PT = prothrombin time; sE-selectin = soluble E-selectin; sICAM-1 = soluble ICAM-1; SOFA = Sequential Organ Failure Assessment.

Patients who did and did not develop an ICU-acquired infection were matched for APACHE-IV score, SOFA score, source of infection, and shock (all on ICU admission). Data are presented as median (interquartile range). Overall P values represent the fixed-effect model in which group + time was used in the random effects model.

*P < 0.05, multiple-comparison adjusted (Benjamini-Hochberg).

P < 0.001, multiple-comparison adjusted (Benjamini-Hochberg).

For biomarkers (i.e., ANG-1, ANG-1:ANG-2, and D dimer) in which no fixed-effect model could be fitted, the model with merely random intercept of time was used.

§P < 0.01, multiple-comparison adjusted (Benjamini-Hochberg).

Host Response Biomarkers at the Time of ICU-acquired Infection

To obtain insight into the host response at the time of ICU-acquired infection, we first compared biomarkers measured in samples obtained within 24 hours after the diagnosis of an ICU-acquired infection (n = 104) or a noninfectious ICU-acquired complication (i.e., AKI, n = 71; ARDS, n = 34) (Table 5). Most host response parameters were not different between groups (Figure 5).

Table 5. Baseline Characteristics and Outcome of Patients Developing ICU-acquired Infection, Acute Kidney Injury, and Acute Respiratory Distress Syndrome

 ICU-acquired InfectionAcute Kidney InjuryAcute Respiratory Distress SyndromeP Value
Patients1027034 
Demographics    
 Age, yr, mean (SD)59.5 (14.0)62.6 (13.3)59.6 (14.1)0.30
 Sex male, n (%)63 (61.8)44 (62.9)23 (67.6)0.85
Chronic comorbidity    
 Any comorbidity, n (%)82 (80.4)55 (78.6)27 (79.4)0.98
 Charlson comorbidity index, median (IQR)4 (3–5)4 (3–6)5 (3–7)0.19
Admissions1047134 
 Time of event, d, median (IQR)10 (7–15)4 (2–7)3 (2–4)<0.0001
Severity of disease during event    
 SOFA score, median (IQR)7 (5–10)7 (5–9)7 (5–9)0.45
 Shock, n (%)26 (25.0)25 (35.2)8 (23.5)0.28
Outcome    
 Length of ICU stay, d, median (IQR)24 (14–33)11 (7–19)11 (6–16)<0.0001
 Length of hospital stay, d, median (IQR)35 (21–64)20 (11–50)22 (12–41)<0.001
 Mortality, n (%)    
  ICU*43 (41.3)23 (32.4)8 (23.5)0.15
  Hospital58 (56.9)36 (51.4)12 (35.3)0.08
  30 d35 (34.3)29 (41.4)10 (29.4)0.48
  60 d51 (50.0)33 (47.1)12 (35.3)0.36
  90 d59 (57.8)36 (51.4)12 (35.3)0.05
  1 yr66 (64.7)42 (60.0)16 (47.1)0.13

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile range; SOFA = Sequential Organ Failure Assessment.

*ICU mortality was calculated using all ICU admissions for sepsis.

Follow-up data were calculated using the first ICU admission for sepsis for each patient during the study period; readmissions were not included in this analysis.

Five patients were lost to 1-year follow-up (3.9% in patients with sepsis developing an ICU-acquired infection and 2.9% in patients with sepsis developing an ICU-acquired acute respiratory distress syndrome).

In a final analysis, biomarker distribution at the last standardized sampling moment (i.e., Day 4) was compared with biomarker distribution at the time of ICU-acquired infection in all patients from whom paired samples were available (n = 84), revealing no differences (Table 6).

Table 6. Host Response Biomarkers at Day 4 after Admission and at the Time of an ICU-acquired Infection

 Day 4 (n = 84)Time of ICU-acquired Infection (n = 84)P Value*
Inflammation   
 IL-6, pg/ml68.92 (24.45–194.38)57.72 (23.3–133.07)0.44
 IL-8, pg/ml132.28 (54.99–246.13)93.71 (41.73–224.73)0.18
 IL-10, pg/ml10.67 (4.56–26.02)7.00 (2.89–16.05)0.07
 MMP-8, ng/ml2.39 (1.04–5.79)2.17 (0.82–4.72)0.31
Endothelial cell activation   
 sE-selectin, ng/ml6.29 (3.82–12.95)5.58 (2.91–12.71)0.23
 sICAM-1, ng/ml238.05 (168.21–359.23)241.76 (140.37–344.83)0.60
 Fractalkine, pg/ml55.07 (25.98–151.59)38.50 (20.92–80.33)0.11
 ANG-1, ng/ml1.18 (0.57–3.11)1.18 (0.51–4.48)0.77
 ANG-2, ng/ml9.14 (5.54–16.19)6.85 (4.3–13.73)0.11
 ANG-2:ANG-18.15 (2.56–28.06)8.27 (1.66–21.85)0.43
Coagulation activation   
 D dimer, μg/ml12.47 (7.11–21.09)12.09 (5.58–23.2)0.64
 PT, s14.80 (12.3–16.1)14.60 (12.55–16)0.93
 aPTT, s41 (31–47)38 (30–53)0.96
 Protein C, ng/ml138.11 (90.38–185.55)148.67 (111.57–187.56)0.30
 Antithrombin, ng/ml883.85 (513.08–1358.3)949.65 (558.14–1415.43)0.40
 Platelets, 109/L111 (49.5–243.5)168 (93–323)0.02

Definition of abbreviations: ANG = angiopoietin; aPTT = activated partial thromboplastin time; ICAM = intercellular adhesion molecule; ICU = intensive care unit; MMP = matrix metalloproteinase; PT = prothrombin time; sE-selectin = soluble E-selectin; sICAM-1 = soluble ICAM-1.

Data are presented as median (interquartile range).

*Multiple-comparison–adjusted (Benjamini-Hochberg) P value for all >0.99, except for platelets P = 0.44.

Sepsis is associated with prolonged immune suppression and it has been suggested that immune suppression renders patients with sepsis susceptible to secondary infections (3, 6, 911, 13, 22). We here examined the possibility that patients with sepsis who, during their ICU stay acquire a secondary infection, besides immune-suppressive features, also display more profound “hyperinflammatory” responses when compared with those who do not develop a secondary infection. For this we measured 20 host response biomarkers reflective of typical proinflammatory sepsis responses, including cytokine release and activation of the vascular endothelium and the coagulation system, in a large cohort of patients with sepsis during the first 4 days after ICU admission and at the time of an ICU-acquired complication (infectious or noninfectious). Our main findings are (1) patients with sepsis who went on to develop an ICU-acquired infection demonstrated enhanced cytokine release and stronger endothelial cell and coagulation activation than those who did not develop a secondary infection, (2) hyperinflammation was sustained up to the day of occurrence of the secondary infection, and (3) the hyperinflammatory host response detected at the time of an ICU-acquired infection was not different from that measured in patients with a noninfectious ICU-acquired complication (i.e., AKI or ARDS).

Our current data do not contradict previous investigations reporting immune suppression in patients with sepsis (3, 6, 911, 13, 2227). These studies focused on mononuclear cells, particularly their responsiveness to bacterial products, antigen presentation capacity, and features of apoptosis, and in some an association was demonstrated between the extent of immune suppression in patients with sepsis and the subsequent development of a secondary infection (2327). Our measurements reveal proinflammatory responses generated at least in part by host mediator systems not captured in the studies on immune suppression cited previously (3, 6, 911, 13, 2227), especially with regard to activation of the endothelium. In line with our finding, one earlier study, performed in 98 patients with septic shock, showed elevated plasma midregional-proadrenomedulin levels in those who developed a secondary infection (28).

We argue that patients with sepsis who develop secondary infections while in the ICU demonstrate concurrent immune suppression and hyperinflammation, and both to a larger extent than patients with sepsis who do not develop an ICU-acquired infection. This overall more disturbed host response is in accordance with our previous finding that patients with sepsis who develop a secondary infection are more severely ill than those who do not (12), which we confirmed in the present subgroup analysis. Likewise, earlier studies in trauma patients have reported a strong association between injury severity and an increased susceptibility to nosocomial infection (2931). Biomarker concentrations showed a large overlap between patient groups, which precludes firm conclusions on the clinical relevance of the differences detected, yet confirms the heterogeneity of the sepsis population and the accompanying host response. The relative hyperinflammation detected in patients with sepsis who went on to develop a secondary infection partially remained detectable after correction for disease severity. In accordance, in a recent investigation in patients with trauma matched for injury characteristics and severity, multiple proinflammatory mediators were elevated within the first 24 hours after trauma in patients who subsequently developed a nosocomial infection (31).

Whole-genome expression profiles of blood leukocytes harvested from patients with sepsis (12) and trauma (32) showed sustained and concurrent activation of multiple proinflammatory, antiinflammatory, and immune-suppressive pathways. In the trauma literature these findings have led to the concept of the so-called persistent inflammation, immunosuppression and catabolism syndrome (33). The present results indicate that sepsis can also lead to persistent inflammation, immunosuppression, and catabolism syndrome, further suggesting that the host response to sepsis and severe noninfectious injury is not fundamentally different (34). As such, we argue that patients who remain critically ill for prolonged periods of time enter a state of sustained hyperinflammation and immune suppression irrespective of the inciting event (sepsis or noninfectious injury), which together with invasive procedures and devices, such as mechanical ventilation and intravenous catheters (12, 35), render patients more susceptible to ICU-acquired complications. Although this observational study does not prove a causal link between enhanced inflammatory responses during the first 4 days of ICU stay and subsequent development of ICU-acquired infections, we consider hyperinflammation and disturbed barrier integrity part of a syndrome that has been named a “failure of homeostasis” (34), resulting in dysfunction of immune and other cells, at least in part caused by mitochondrial damage and impaired cellular oxygen use, which together with a lengthy requirement of invasive care, are main drivers of the occurrence of both noninfectious and infectious complications on the ICU.

Of note, the previously reported whole-blood leukocyte genome transcriptome changes were not different between patients who did and those who did not develop an ICU-acquired infection in this cohort (12), which at least in part can be explained by the fact that gene expression analyses of blood leukocytes only provide insight in immune pathways regulated at mRNA level in circulating cells, whereas the protein biomarkers reported here mostly are derived from extravascular cells. Furthermore, although the blood genomic response was measured at a single time point within 24 hours after ICU admission, plasma protein biomarkers were measured at multiple time points.

Biomarker analyses at the time of an ICU-acquired infection versus a noninfectious ICU-acquired complication (AKI or ARDS) showed comparable host response reactions. In paired analyses, no differences were found between Day 4 and the day of the ICU-acquired infection (Table 4) or AKI/ARDS (data not shown). These results indicate that the dysregulation of key host mediator systems as measured here is sustained and not different in patients with either one of these major ICU-acquired complications.

Our study has strengths and limitations. We provide information on a large, well-defined, prospectively collected cohort with extensive information on ICU-acquired complications. Although we measured 20 host response biomarkers reflecting activation of key pathways implicated in sepsis pathogenesis, we did not perform functional and/or flow cytometry measurements that would have provided information on the extent of immune suppression. Hence, we cannot examine potential correlations between hyperinflammatory, procoagulant, and immune-suppressive responses in individual patients. In addition, most measurements were confined to the first 4 days after ICU admission; however, paired analyses of biomarker levels at Day 4 and the day of an ICU-acquired infection did not show differences with the single exception of platelet counts. In this study, statistical methods are used to adjust for differences between groups; however, we cannot exclude the effect of unmeasured covariables residually confounding our outcome. In addition, in some cases persistent infections may be difficult to distinguish from new-onset infections, especially in abdominal sepsis. We can therefore not exclude the possibility that occasionally an ongoing infection was deemed ICU acquired.

Conclusions

Patients with sepsis developing secondary infections during ICU stay showed a more dysregulated proinflammatory and vascular host response in the first 4 days of ICU admission than patients with sepsis who did not develop an ICU-acquired infection.

The authors acknowledge all members of the MARS consortium for the participation in data collection and especially acknowledge Friso M. de Beer, M.D., Lieuwe D. J. Bos, M.D., Ph.D., Gerie J. Glas, M.D., and Roosmarijn T. M. van Hooijdonk, M.D., Ph.D. (Department of Intensive Care, Academic Medical Center, University of Amsterdam); Michaëla A. M. Huson, M.D., Ph.D. (Center for Experimental and Molecular Medicine, Academic Medical Center, University of Amsterdam); David S. Y. Ong, M.D., Ph.D. (Department of Intensive Care Medicine and Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands); and Laura R. A. Schouten, M.D., Marleen Straat, M.D., Esther Witteveen, M.D., and Luuk Wieske, M.D., Ph.D. (Department of Intensive Care, Academic Medical Center, University of Amsterdam).

1. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med 2013;369:840851.
2. Cohen J, Vincent JL, Adhikari NK, Machado FR, Angus DC, Calandra T, Jaton K, Giulieri S, Delaloye J, Opal S, et al. Sepsis: a roadmap for future research. Lancet Infect Dis 2015;15:581614.
3. Hotchkiss RS, Monneret G, Payen D. Immunosuppression in sepsis: a novel understanding of the disorder and a new therapeutic approach. Lancet Infect Dis 2013;13:260268.
4. van der Poll T, Opal SM. Host-pathogen interactions in sepsis. Lancet Infect Dis 2008;8:3243.
5. Abraham E, Reinhart K, Opal S, Demeyer I, Doig C, Rodriguez AL, Beale R, Svoboda P, Laterre PF, Simon S, et al.; OPTIMIST Trial Study Group. Efficacy and safety of tifacogin (recombinant tissue factor pathway inhibitor) in severe sepsis: a randomized controlled trial. JAMA 2003;290:238247.
6. Otto GP, Sossdorf M, Claus RA, Rödel J, Menge K, Reinhart K, Bauer M, Riedemann NC. The late phase of sepsis is characterized by an increased microbiological burden and death rate. Crit Care 2011;15:R183.
7. Ranieri VM, Thompson BT, Barie PS, Dhainaut JF, Douglas IS, Finfer S, Gårdlund B, Marshall JC, Rhodes A, Artigas A, et al.; PROWESS-SHOCK Study Group. Drotrecogin alfa (activated) in adults with septic shock. N Engl J Med 2012;366:20552064.
8. Opal SM, Laterre PF, Francois B, LaRosa SP, Angus DC, Mira JP, Wittebole X, Dugernier T, Perrotin D, Tidswell M, et al.; ACCESS Study Group. Effect of eritoran, an antagonist of MD2-TLR4, on mortality in patients with severe sepsis: the ACCESS randomized trial. JAMA 2013;309:11541162.
9. Torgersen C, Moser P, Luckner G, Mayr V, Jochberger S, Hasibeder WR, Dünser MW. Macroscopic postmortem findings in 235 surgical intensive care patients with sepsis. Anesth Analg 2009;108:18411847.
10. Boomer JS, To K, Chang KC, Takasu O, Osborne DF, Walton AH, Bricker TL, Jarman SD II, Kreisel D, Krupnick AS, et al. Immunosuppression in patients who die of sepsis and multiple organ failure. JAMA 2011;306:25942605.
11. Hotchkiss RS, Monneret G, Payen D. Sepsis-induced immunosuppression: from cellular dysfunctions to immunotherapy. Nat Rev Immunol 2013;13:862874.
12. van Vught LA, Klein Klouwenberg PM, Spitoni C, Scicluna BP, Wiewel MA, Horn J, Schultz MJ, Nürnberg P, Bonten MJ, Cremer OL, et al.; MARS Consortium. Incidence, risk factors, and attributable mortality of secondary infections in the intensive care unit after admission for sepsis. JAMA 2016;315:14691479.
13. Leentjens J, Kox M, van der Hoeven JG, Netea MG, Pickkers P. Immunotherapy for the adjunctive treatment of sepsis: from immunosuppression to immunostimulation. Time for a paradigm change? Am J Respir Crit Care Med 2013;187:12871293.
14. Klein Klouwenberg PM, Ong DS, Bos LD, de Beer FM, van Hooijdonk RT, Huson MA, Straat M, van Vught LA, Wieske L, Horn J, et al. Interobserver agreement of Centers for Disease Control and Prevention criteria for classifying infections in critically ill patients. Crit Care Med 2013;41:23732378.
15. Scicluna BP, Klein Klouwenberg PM, van Vught LA, Wiewel MA, Ong DS, Zwinderman AH, Franitza M, Toliat MR, Nürnberg P, Hoogendijk AJ, et al. A molecular biomarker to diagnose community-acquired pneumonia on intensive care unit admission. Am J Respir Crit Care Med 2015;192:826835.
16. Oostdijk EA, Kesecioglu J, Schultz MJ, Visser CE, de Jonge E, van Essen EH, Bernards AT, Purmer I, Brimicombe R, Bergmans D, et al. Effects of decontamination of the oropharynx and intestinal tract on antibiotic resistance in ICUs: a randomized clinical trial. JAMA 2014;312:14291437.
17. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL, Ramsay G; SCCM/ESICM/ACCP/ATS/SIS. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 2003;31:12501256.
18. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P; Acute Dialysis Quality Initiative workgroup. Acute renal failure: definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204R212.
19. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M, Legall JR, Morris A, Spragg R. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818824.
20. Pierrakos C, Vincent JL. Sepsis biomarkers: a review. Crit Care 2010;14:R15.
21. Ho DE, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw 2011;42:128.
22. Hotchkiss RS, Swanson PE, Freeman BD, Tinsley KW, Cobb JP, Matuschak GM, Buchman TG, Karl IE. Apoptotic cell death in patients with sepsis, shock, and multiple organ dysfunction. Crit Care Med 1999;27:12301251.
23. Drewry AM, Samra N, Skrupky LP, Fuller BM, Compton SM, Hotchkiss RS. Persistent lymphopenia after diagnosis of sepsis predicts mortality. Shock 2014;42:383391.
24. Grimaldi D, Louis S, Pène F, Sirgo G, Rousseau C, Claessens YE, Vimeux L, Cariou A, Mira JP, Hosmalin A, et al. Profound and persistent decrease of circulating dendritic cells is associated with ICU-acquired infection in patients with septic shock. Intensive Care Med 2011;37:14381446.
25. Guignant C, Lepape A, Huang X, Kherouf H, Denis L, Poitevin F, Malcus C, Chéron A, Allaouchiche B, Gueyffier F, et al. Programmed death-1 levels correlate with increased mortality, nosocomial infection and immune dysfunctions in septic shock patients. Crit Care 2011;15:R99.
26. Landelle C, Lepape A, Voirin N, Tognet E, Venet F, Bohé J, Vanhems P, Monneret G. Low monocyte human leukocyte antigen-DR is independently associated with nosocomial infections after septic shock. Intensive Care Med 2010;36:18591866.
27. Shubin NJ, Monaghan SF, Heffernan DS, Chung CS, Ayala A. B and T lymphocyte attenuator expression on CD4+ T-cells associates with sepsis and subsequent infections in ICU patients. Crit Care 2013;17:R276.
28. Guignant C, Voirin N, Venet F, Lepape A, Monneret G. Persistent high level of circulating midregional-proadrenomedullin and increased risk of nosocomial infections after septic shock. J Trauma Acute Care Surg 2012;72:293296.
29. Jamulitrat S, Narong MN, Thongpiyapoom S. Trauma severity scoring systems as predictors of nosocomial infection. Infect Control Hosp Epidemiol 2002;23:268273.
30. El-Masri MM, Joshi M, Hebden J, Korniewicz DM. Use of the injury severity score to predict nosocomial bloodstream infections among critically ill trauma patients. AACN Clin Issues 2002;13:367372.
31. Namas RA, Vodovotz Y, Almahmoud K, Abdul-Malak O, Zaaqoq A, Namas R, Mi Q, Barclay D, Zuckerbraun B, Peitzman AB, et al. Temporal patterns of circulating inflammation biomarker networks differentiate susceptibility to nosocomial infection following blunt trauma in humans. Ann Surg 2016;263:191198.
32. Xiao W, Mindrinos MN, Seok J, Cuschieri J, Cuenca AG, Gao H, Hayden DL, Hennessy L, Moore EE, Minei JP, et al.; Inflammation and Host Response to Injury Large-Scale Collaborative Research Program. A genomic storm in critically injured humans. J Exp Med 2011;208:25812590.
33. Vanzant EL, Lopez CM, Ozrazgat-Baslanti T, Ungaro R, Davis R, Cuenca AG, Gentile LF, Nacionales DC, Cuenca AL, Bihorac A, et al. Persistent inflammation, immunosuppression, and catabolism syndrome after severe blunt trauma. J Trauma Acute Care Surg 2014;76:2129, discussion 29–30.
34. Deutschman CS, Tracey KJ. Sepsis: current dogma and new perspectives. Immunity 2014;40:463475.
35. Vincent J-L. Nosocomial infections in adult intensive-care units. Lancet 2003;361:20682077.
Correspondence and requests for reprints should be addressed to Lonneke Alette van Vught, M.D., Center for Experimental and Molecular Medicine, Academic Medical Center, Meibergdreef 9 1105 AZ Amsterdam, the Netherlands. E-mail:

Supported by the Center for Translational Molecular Medicine (http://www.ctmm.nl), project MARS (grant 04I-201).

Author Contributions: L.A.v.V. and T.v.d.P. contributed to the design of the study. L.A.v.V., M.A.W., A.J.H., B.P.S., J.F.F., J.H., P.M.C.K.K., O.L.C., M.J.S., M.M.J.B., and T.v.d.P. acquired the data. A.H.Z. provided statistical guidance and R.L. guided biomarker measurements. L.A.v.V. had full access to all the data in the study and takes full responsibility for the integrity of the data and the accuracy of the data analysis. L.A.v.V. and T.v.d.P. were involved in the interpretation of the data. L.A.v.V. and T.v.d.P. drafted the manuscript, and all authors revised it critically for important intellectual content. All authors gave final approval of this version and declare that they have no conflicts of interest.

This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org

Originally Published in Press as DOI: 10.1164/rccm.201606-1225OC on January 20, 2017

Author disclosures are available with the text of this article at www.atsjournals.org.

Comments Post a Comment




New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine
196
4

Click to see any corrections or updates and to confirm this is the authentic version of record