To rule out age- and gender-dependent influences, female and male volunteers were grouped in three age clusters: 20 to 30, 31 to 40, and 41 to 50 years, with 6 volunteers per gender and age cluster, resulting in a total of 36 volunteers.

To ensure that recruited volunteers were healthy, a carefully structured interview was conducted and various variables (for example, blood pressure, pulse, temperature, and peripheral oxygen saturation) were determined before platelets were isolated and stimulated in vitro. Volunteers with a recent history of fever, surgery, or intake of drugs possibly influencing platelet function (for example, aspirin and clopidrogel) were excluded.

Following placement of an ICP probe, patients with sTBI were treated in the intensive care unit (ICU) according to a standardized protocol. Routine treatment and decision making were not influenced by the present investigations, and the obtained data were not integrated in the current treatment concept. Continuously infused midazolam (Dormicum^® and fentanyl (Sintenyl^® were tapered according to ICP values. Volume and norepinephrine administration were adjusted to maintain CPP values above 70 mm Hg. Patients did not receive heparin or low-molecular-weight heparin. All flush systems were maintained without heparin.

Platelet activation was measured in platelet-rich plasma (PRP) using monoclonal antibodies and three-color flow cytomtery. Within 30 minutes of blood withdrawal, samples were centrifugated at 5,000 rounds per minute for 15 minutes. Thereafter, 5 μL of PRP was added to a 12 × 75-mm tube containing 15 μL of each of the following fluorescent-labelled monoclonal antibodies: CD61-fluorescein isothiocyanate and CD62P-phycoerythrin. CD62P (P-selectin) is an antigen present on the surface of activated platelets 10. Anti-CD61 recognizes the platelet glycoprotein receptor, GPIIIa, which is found on all resting and activated platelets and which is used to identify platelets.

After 20 minutes of incubation with monoclonal antibodies in the dark at room temperature, 1 mL of 1% paraformaldehyde was added to each tube for fixation of platelets. Mouse immunoglobulin G 1 (fluorescein isothiocyanate) and phycoerythrin were used as isotype controls. Antibodies and isotype controls were purchased from Becton Dickinson Immunocytometry Systems (San Jose, CA, USA). All samples were analyzed within 90 minutes on a FACSscan flow cytometer (Becton Dickinson, Mountain View, CA, USA) using Cell Quest^® software (Becton Dickinson Immunocytometry Systems). Flow cytometer performance used to analyze microparticles was verified employing 1-μm calibration beads (Bangs Laboratories, Inc., Fishers, IN, USA).

A total of 5,000 CD61-positive events were collected with all light scatter and fluorescence parameters in a logarithmic mode. Platelets were gated on the basis of light scatter and CD61 expression. Activated platelets were defined as the percentage of CD61-positive events expressing the activated confirmation of P-selectin (CD62P). Platelet-derived microparticles were also measured and identified as CD61-positive events in a gate obtained using uniform microspheres of 7.4 μm in diameter (Bangs Laboratories, Inc.).

Double samples of isolated peripheral venous, jugular venous, and arterial platelets were incubated for 20 minutes with different norepinephrine concentrations (Noradrenaline Sintetica 0.1%; Sintetica S.A., Mendrisio, Switzerland) ranging from 10 nM to 100 μM. The same norepinephrine as employed in the routine treatment in our ICU was used for the in vitro stimulation. Thrombin receptor-activating peptide (TRAP) (Becton Dickinson Immunocytometry Systems), known to maximally activate platelets, served as a positive control. Upon stimulation, changes in expression of P-selectin-positive platelets and changes in the number of CD61-positive platelet-derived microparticles were assessed to reveal the degree of platelet activation. All samples were analyzed within 90 minutes after blood withdrawal.

Differential blood counts were analyzed in the ISO-IEC 17025 accredited university hospital laboratory at the University Hospital Zuerich.

sP-selectin was measured in plasma using a DuoSet^® ELISA [enzyme-linked immunosorbent assay] Development System (R&D Systems, Inc., Minneapolis, MN, USA) in accordance with the instructions of the manufacturer.

Continuously recorded ICP, CPP, temperature, and SjvO₂ were assessed in 1-hour intervals. Drug dosage was also determined in 1-hour intervals. A daily median was calculated using these 24 values. Daily administration of hydroxyethyl starch (HES) (Voluven^® was recorded. AJVDs in glucose and lactate were assessed in 4- to 6-hour intervals, allowing us to calculate a daily median. AJVDs in platelets, leukocytes, and sP-selectin were measured once daily.

Jugular venous values were substracted from arterial values, thus yielding the calculated AJVDs. Positive AJVDs reflect cerebral retention or uptake as the arterial levels exceed the jugular venous concentration. Negative AJVD values reveal sustained release or decreased uptake/binding within the cerebral compartment as jugular venous levels exceed arterial concentrations.

Results are presented as median or mean ± standard error of the mean, where applicable. Differences between groups, time points, and norepinephrine concentrations were rated significant at a probability level of less than 0.05 using analysis of variance on ranks with post hoc multiple pairwise comparisons. Statistical analysis was performed using SigmaStat^® 3.5 (SPSS Inc. Headquarters, Chicago, Illinois, USA). Figures were created with SigmaPlot^® 10.0 (SPSS Inc. Headquarters, Chicago, Illinois, USA).

Demographic data of the investigated critically ill patients suffering from sTBI are presented in Table 2. Changes in absolute blood platelet and leukocyte counts, AJVDs of platelets, leukocytes, glucose, and lactate as well as mean arterial blood pressure, ICP, CPP, SjvO₂, temperature, and average drug dosage are presented in Table 3. Data were pooled for the first and second week. During the second week, absolute platelet and leukocyte counts were significantly increased. Whereas platelets remained within normal limits, leukocytes surpassed the upper limit of normal values. Whereas ICP was significantly increased, CPP, SjvO₂, and temperature were significantly decreased during the second week compared with the first week. These changes, however, remained within clinically acceptable limits. Administered drug dosages were similar for norepinephrine, midazolam, and fentanyl during the first and second week. In a total of 751 SjvO₂, CPP, and ICP values which were recorded at the same time as jugular venous blood gas analysis only 0.4% SjvO₂ were less than 50%, 0.1% of CPP values were less than 60 mm Hg, and 17% of ICP was greater than 20 mm Hg. In eight of the 11 patients, pneumonia was diagnosed on (a median of) 8.5 days after trauma (range 3 to 13 days). In 1 patient (#3), bacteremia with coagulase-negative Staphylococcus aureus was diagnosed. In 1 multiply injured patient (#8), pulmonary embolism was diagnosed clinically and verified radiologically on day 12 after trauma after the patient was mobilized. A deep venous thrombosis was not found. A vena cava filter was inserted and removed after 14 days. Thereafter, the patient had an uneventful recovery.

Arterial and jugular venous catheters remain in place until these catheters can or need to be removed. Over time, local thrombus formation at the tip of the catheter is possible. New daily insertions of catheters to avoid any local thrombus formation, however, are not feasible under clinical conditions due to hemodynamic instability, generalized edema formation related to capillary leakage, and a limited number of accessible vessels. Local thrombus formation at the tip of the catheters activates platelets, possibly resulting in false-positive results. As a standardized procedure to reduce the risk of possible thrombus-related confounding influences, 2 mL of blood was withdrawn and discarded before the actual blood sample was taken. Nevertheless, local activation might have occurred, possibly explaining the reduced number of P-selectin-expressing platelets during the second week. In addition to local catheter-related effects, the underlying tissue damage might have contributed to platelet activation with subsequent P-selectin shedding and sustained sP-selectin concentrations. Due to the fact that the post-traumatic significantly increased sP-selectin levels exceeded normal values by several fold, any additional shedding might remain obscured. In addition, isolation procedures can activate cells. As to our own preliminary experiments, the chosen isolation procedure is associated with an activation of less than 2%.

As shown by Scherer and Spangenberg 11, Jacoby and colleagues 12, and Nekludov and colleagues 1314, plasmatic coagulation, platelet count, and platelet function are significantly and reversibly altered during the early phase following sTBI. In this context, activation of the coagulation cascade which occurs within the first hours after trauma within the injured brain 1113 as reflected by an elevated transcranial gradient precedes systemic hypercoagulability which is followed by fibrinolytic activity. These alterations, in turn, could explain the observed decrease in platelet count and fibrinogen level and subsequent increase in thrombin-antithrombin III complex, prothrombin fragment F1+2, and D-dimer concentrations 11. Following TBI, platelets were significantly activated in the face of depressed function as reflected by prolonged collagen/epinephrine closure times during the first 3 post-traumatic days 12. In addition, prolonged disturbance in platelet function was significantly sustained in non-surviving patients, which underlines the pathophysiologic importance of disturbed coagulation. In conjunction with a prolonged bleeding time, platelets showed a decreased responsiveness to arachidonic acid as determined by thromboelastography 14. As shown by the present study, functional depression in isolated platelets is expanded to 7 days following sTBI and reflects prolonged functional disturbance in thrombocytic coagulation. Clinically, however, there were no signs of coagulation disorder. Following the initial functional depression, platelet function was significantly increased in the second week following sTBI, which coincided with sustained cerebral retention of platelets and signs of disturbed cerebral perfusion. Thus, these changes clearly unmask temporally differentiated changes in platelet function which are of pathophysiologic importance.

Under physiologic conditions, quantitative and qualitative features of platelets are tightly controlled by various mediators within the bone marrow, blood, and along the endothelial cells 15. Following injury, excessive loss and consumption of platelets exceeding production and release from bone marrow result in a significant decrease in circulating platelets, reaching its nadir by the second post-traumatic day. Subsequent significant increase reflects upregulated compensatory production within the bone marrow aimed at normalizing the amount of circulating platelets. In this context, thrombopoietin is of crucial importance 16. Thrombopoietin also contributes to enhanced platelet activation under clinical conditions 17. Newly produced and freshly released platelets might be activated more easily than senescent platelets. This, in turn, might explain the preserved and exaggerated in vitro norepinephrine-mediated stimulation during the second week as observed in the present study. The preserved functionality in platelets despite decreased baseline P-selectin expression as found in the second week is in line with results from Michelson and colleagues 18, who showed that circulating platelets remain active for at least 24 hours following shedding of surface P-selectin. In this context, we suggest that reduced P-selectin-positive platelets in the face of signs of cerebral worsening reflect functional disturbance of the isolated platelets, assuming that platelets contribute to pathophysiologic cascades within the injured brain in these patients. While P-selectin expression determines size and stability of platelet aggregates 19, reduced surface P-selectin expression does not imply functional impairment 18. Shedding of P-selectin reflects previous platelet activation and could result in facilitated release of various toxic mediators 2021 which have been shown to induce and promote tissue damage. This warrants further investigations.

Activation of α₂ adrenergic receptors by norepinephrine routinely infused to elevate CPP following sTBI enhanced platelet aggregability concentration dependently and increased platelet secretion of beta-thromboglobulin during high-dose infusion 22. In addition, norepinephrine stimulated the expression of surface P-selectin and intracellular prothrombotic microparticles. Stimulation of different surface receptors results in a stereotypic amplified activation of intracellular G-protein-mediated cascades involving the Rho/Rho-kinase pathway, phospholipase C, and protein kinase C, which are essential for conformational changes in platelet shape as well as aggregation and degranulation 23.

Despite the tedious analysis and difficult interpretation of concentrations of blood norepinephrine (due to its short half-life and fast response to changes in infusion parameters), Johnston and colleagues 24 determined the pharmacokinetic profile of norepinephrine in eight patients suffering from sTBI. Based on their findings, plasma norepinephrine levels significantly correlated with the rate of norepinephrine infusion during steady-state conditions of the norepinephrine infusion period. The average norepinephrine dose infused in the presently investigated patients ranged from 0.1 ± 0.07 to 0.16 ± 0.11 μg/kg per minute. Assuming a similar norepinephrine distribution volume and clearance in our patients, we are to expect plasma levels of between 22.98 ± 16.98 and 37.08 ± 20.15 nM/L according to the results published by Johnston and colleagues 24.

Based on the assumptions that norepinephrine exhibits minimal regional and temporal fluctuations during steady-state conditions and that in vitro concentrations are equally potent as those in vivo, it appears as if extremely high norepinpehrine doses were required to activate isolated platelets. The lowest norepinephrine concentration associated with a significant effect in the presently isolated platelets was 500 nM, which exceeded the extrapolated blood levels of 25 nM by 20-fold. Thus, it remains unclear to what extent the observed effects are also valid under in vivo conditions.

The fact that isolated platelets exhibited a temporally differentiated response to the same norepinephrine concentration in the first versus second week coinciding with a preserved and even increased TRAP-mediated platelet activation suggests altered susceptibility of platelet receptors. In this context, functional adaptation of platelet α₂ adrenergic receptors in terms of receptor downregulation or upregulation might be of pharmacologic and pathophysiologic importance. Clinical as well as experimental studies have shown that elevated catecholamine concentrations are associated with a reduction in expression and affinity of α₂ adrenergic receptors 25262728. This also resulted in a decreased platelet aggregation response to epinephrine 29. Intracellular adaptive processes in conjunction with regained sensitization of previously desensitized α₂ adrenergic receptors might lead to the observed sustained in vitro stimulation during the second week during continuous norepinephrine stimulation following the depressed stimulation during the first week. This could also account for the stimulatory effect at a lower norepinephrine concentration compared with healthy controls (500 nM versus 10 μM).

Sedative agents (for example, midazolam) might have contributed to the decreased expression of platelet surface P-selectin as shown by Tsai and colleagues 30 and Gries and colleagues 31. The inhibitory mechanism of midazolam is best explained by concentration-dependent blocking of platelet aggregation, inhibition of phosphoinositide breakdown and intracellular Ca⁺² mobilization, increased formation of cyclic AMP, inhibition of increases in intracellular pH, and attenuated protein kinase C activation 32. Adaptive intracellular processes upon initial midazolam-induced functional depression might have contributed to the sustained norepinephrine-mediated stimulation of platelets isolated during the second week despite the administration of amounts comparable to those in the first week.

Whether inflammation-induced cytokine release might have contributed to the sustained in vitro stimulation of isolated platelets appears doubtful since interleukin (IL)-6 levels were not significantly increased during the second week in the presently investigated patients despite significant leukocytosis. This is in line with findings reported by Leytin and colleagues 33 showing that the pro-inflammatory cytokines IL-1β, IL-6, and IL-8 did not stimulate platelets and failed to promote thrombin-mediated platelet activation. Other mechanisms related to bacterial infections, however, have been shown to activate platelets, a circumstance that was not reflected by an increase in leukocytes 34. In those 8 patients with pneumonia and the single patient with bacteremia, there was no significant difference in baseline P-selectin expression and susceptibility to norepinephrine-mediated stimulation of isolated platelets compared with the remaining 5 patients. An inflammation-induced influence, however, needs to be specifically addressed in a larger study population.

In clinical routine, colloids (for example, HES) are combined with cristalloids to maintain adequate organ perfusion and to reduce catecholamine dose by inducing normovolemia. As reported by Chen and colleagues 35, HES 130/0.4 (Voluven^®, which is routinely used in our ICU, induced transient reduction in platelet-mediated coagulation reflected by decreased platelet membrane glycoprotein and P-selectin expression in patients undergoing elective minor surgery.

Under in vitro conditions, HES 130/0.4 did not influence the expression of various membrane proteins on platelets isolated from healthy volunteers 36. Thus, decreased baseline P-selectin expression observed in the second week does not appear to be induced by HES since patients required significantly less HES 130/0.4 compared with the first week. In fact, baseline P-selectin and microparticle expression were comparable to healthy volunteers during the first week despite a significantly larger amount of HES 130/0.4 administered per day compared with the single administration of HES 130/0.4 during minor surgery as studied by Chen and colleagues 35.

Following TBI, impaired pericontusional microcirculation shows a dynamic temporal and heterogeneous regional profile with impaired as well as increased cerebral perfusion 3738. Impaired perfusion is related to vasoconstriction and endovascular occlusion due to microthrombosis evolving within the first 24 hours and promoting edema formation. Under experimental conditions, thrombotic occlusion is followed by spontaneous resolution during the second post-traumatic day as evidenced by histology, intravital microscopy, and laser Doppler flowmetry 7893940.

Sustained platelet adhesion and activation are functionally interwoven with activated leukocytes, thereby facilitating thrombus formation as well as attraction and tissue penetration of various leukocyte subpopulations 6. This, in principle, enables and promotes tissue repair. Upon excessive stimulation, however, platelet-induced attraction and activation of leukocytes can aggravate underlying tissue injury in conjunction with evolving microthrombosis formation, thereby promoting perpetuating autodestructive cascades.

Whether the increased platelet count in conjunction with leukocytosis, sustained norepinephrine-mediated platelet activation, and increased retention of platelets within the brain (positive arterio-jugular venous platelet difference) contributed to the signs of cerebral deterioration as reflected by elevated ICP, decreased SjvO₂, and sustained lactate release during the second week remains unclear.

Based on findings obtained in other neurodegenerative diseases, activated platelets could be of increasing pathophysiologic importance also following clinical TBI. As reported by Mathew and colleagues 41, transcerebral activation of platelets occurred following the release of aortic crossclamp in patients subjected to cardiac surgery and was associated with neurocognitive worsening. Altered platelet function resulting in impaired uptake and sustained release of glutamate might also promote cerebral injury as discussed for cerebral ischema 42, migraine 43, and epilepsy 44.

The finding of norepinephrine-mediated increased platelet activation during the second week with a significantly attenuated effect during the first week does not automatically imply functional disturbance of platelets resulting in additional hemorrhage or contusion growth. Further analysis, however, is required to determine norepinephrine-induced release of platelet-derived toxic mediators despite nearly unchanged expression of P-selectin in the early phase following sTBI.