Serum Levels of TNF Receptor Ligands Are Dysregulated in Sepsis and Predict Mortality in Critically Ill Patients

Abstract

Introduction: TNF superfamily members, including TNF-related weak inducer of apoptosis (TWEAK) and Glucocorticoid-Induced TNFR-Related Protein Ligand (GITRL) have been described as serum based biomarkers for inflammatory and immune mediated diseases. However, up to now the role of TWEAK and GITRL has not been analyzed in critical illness and sepsis.

Methods: GITRL and TWEAK serum concentrations were measured in 121 critically ill patients (84 fulfilled with septic disease), in comparison to 50 healthy controls. Results were correlated with clinical data.

Results: Serum levels of TWEAK and GITRL were strongly decreased in critically ill patients compared with healthy controls. Concentrations of TWEAK (but not GITRL) were further decreased in patients with sepsis and correlated with routinely used markers of inflammation and bacterial infection such as C-reactive protein, procalcitonin and Interleukin-6. Notably, we failed to detect a correlation to other TNFR ligands such as TNF or APRIL. Finally, TWEAK levels of the upper quartile of the cohort were prognostic for mortality during ICU treatment.

Conclusion: TWEAK and GITRL levels were lower in intensive care unit medical patients. Levels of TWEAK were further decreased in septic patients, and alterations in TWEAK concentrations were linked to an unfavorable outcome. Together with recently published results on other TNFR ligands, these data indicate specific functions of the different TNFR ligands in septic diseases.

Fig 1. Serum TWEAK and GITRL concentrations of critically ill patients at ICU admission.

(a,b) TWEAK and GITRL serum levels at admission to the ICU were significantly (P < 0.001, U-test) lower in critically ill patients (n = 121) compared to healthy controls (n = 50). (c) TWEAK concentrations at admission to the ICU were independent on the severity of disease. (d) Serum TWEAK concentrations were similar in critically ill patients with or without obesity. (e) Serum TWEAK concentrations were similar in critically ill patients with or without diabetes mellitus typ 2. *** p< 0.001

Fig 2. Serum TWEAK concentrations are decreased in sepsis.

(a) Critically ill patients with sepsis (n = 84) displayed significantly lower TWEAK serum concentrations (n = 37, U-test) compared to patients without sepsis. (b) TWEAK serum concentrations were not different in patients with different etiologies of critical illness and are highest in patients with cardiac diseases. (c) ROC curve analyses comparing the diagnostic power in predicting sepsis of TWEAK with well-established laboratory markers: C-reactive protein (CRP), procalcitonin, and white blood cell count (leukocytes).

Fig 3. Prediction of ICU mortality by TWEAK serum concentrations.

(a) Patients that succumbed to death during ICU treatment displayed higher TWEAK serum levels compared to survivors. (b) Kaplan-Meier survival curves of ICU patients showed that patients with low TWEAK levels an increased short-term survival at the ICU. (c) Kaplan-Meier survival curves of ICU patients revealed that patients with TWEAK levels < 209 ng/ml had an increased ICU survival. (d) ROC curve analyses comparing the prognostic value of TWEAK at admission for ICU survival with the patients´ age and routinely used laboratory parameters.

Fig 4. Prediction of long-term survival by analysis of TWEAK serum levels.

(a) TWEAK levels were unaltered in critically ill patients that survived in the long term follow up compared to patients that succumbed to death (U-test).

Fig 5. Time course of TWEAK concentrations in critically ill patients.

(a) TWEAK serum concentrations are unchanged in critically ill patients at admission to the ICU and after three days of ICU treatment (U-test). (b, c) TWEAK serum concentrations were determined at admission to the ICU and after three days of ICU treatment. The differences in TWEAK levels between these time-points were not predictive for ICU or long-term survival.

Fig 6. TNF and TWEAK levels for the diagnosis of sepsis.

(a) Patients that fulfilled sepsis criteria had higher TNF levels when compared with critically ill patients without sepsis. (b) ROC analyses comparing the diagnostic power of TWEAK and TNF for sepsis.