Introduction: Sepsis is a major contributor to the global disease burden with estimated 20 million cases per year. With every hour prior to begin of therapy mortality increases by over 7%. Hence, early diagnosis is indispensable for patient survival, though it is hampered by the lack of known pathognomonic symptoms. Sepsis diagnosis is set according to an increase of the sequential organ failure assessment (SOFA) score, reflecting organ dysfunction. Severe symptoms comprising dysfunctional hemostasis are disseminated intravascular coagulation and formation of massive edema due to loss of vascular integrity. Functional hemostasis and maintenance of vascular integrity depends particularly on the platelet surface receptor glycoprotein (GP)VI that dimerizes upon activation and transduces its signals via the associated immunoreceptor tyrosine activation motif (ITAM) of the Fc-gamma chain. Sepsis-associated thrombocytopenia has been extensively investigated and low platelet counts are implemented in the SOFA score.Platelet function during sepsis, however, remains ill-defined with few reports using ADP or thrombin receptor activating peptide (TRAP) and overall conflicting results.

Objectives: We assessed platelet function in patients with sepsis III criteria in a single center study at three times during disease: (I) intensive care unit (ICU) admission day; (II) day five to seven at ICU; (III) day of ICU discharge.

Methods: Fifteen patients were recruited from the ICU of the University Hospital Würzburg, Germany (m:f ratio 9:6; median age 70; mortality: 40%). Platelets in whole blood were stimulated with ADP, TRAP, or the GPVI agonist collagen-related peptide (CRP). Activation was detected flow cytometrically by P-selectin (CD62P) exposure and integrin GPIIb/IIIa activation (PAC-1 binding). Results were correlated with light transmission aggregometry (LTA) and immunoblotting of phospho-Syk and -LAT. Co-incubation experiments were performed with plasma, whole blood or patient-borne bacterial isolates comprising Gram-negative (E. coli, K. pneumoniae) or Gram-positive (E. faecalis) strains.

Results: Platelet function was markedly reduced in all patients assessed by flow cytometry and LTA despite overall unaltered receptor expression. The defect was most prominent after GPVI stimulation with CRP(Mean Geo-MFI ± SD; Control: 9356 ± 3460; Sepsis: 1438 ± 2090; p<.0001). In 14/15 patients GPVI-dysfunction was already present at time (I). In contrast, only 7/15 patients were thrombocytopenic at this early time. We did not consistently find elevated GPVI ectodomain plasma levels, low platelet counts, increased mean platelet volume, or elevated reticulated platelet levels in our cohort at early time points and minor alterations only late during disease progression. We conclude that none of these is a bona fide biomarker, neither for early sepsis diagnosis nor as a prognostic marker.

Sepsis platelets failed to transduce the GPVI signal to induce tyrosine phosphorylation of Syk kinase or LAT. ITIM receptor-based tyrosine phosphatases SHP1 and SHP2 were neither preactivated in resting platelets nor responsive to GPVI stimulation. Of note, platelet aggregation upon GPVI stimulation increased in those patients whose condition ameliorated. We found a strong correlation of restored platelet aggregation by LTA with survival and could distinguish survivors from non-survivors using this marker for stratification (%max aggregation ± SD; survivors: 25 ± 8%; non-survivors: 4 ± 3%; p<0.01). Responsiveness to CRP was unaltered when platelets of healthy donors were co-incubated with patient blood borne bacterial isolates. Platelet reactivity was also unaltered when they were pre-incubated in plasma of sepsis patients. However, pre-incubation in whole blood clearly diminished the response toward CRP,suggesting that the abrogated GPVI signaling is in part mediated by cellular components.

Conclusion: Our results strongly imply that GPVI hypo-responsiveness could serve as a bona fide marker for early sepsis diagnosis and also bears potential to act as a prognostic indicator for therapy monitoring, independent from the source of infection.

Acknowledgement: Financial support was granted by the Deutsche Forschungsgemeinschaft (374031971 - TRR 240).


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.