Influence of Cardiopulmonary Bypass on the Interaction of Recombinant Factor VIIa with Activated Platelets

Abstract:

Recombinant factor VIIa (rFVIIa) interacts preferentially with coated platelets characterized by a high exposure of phosphatidyl serine (PS), FV, FVIII, FIX, and FX binding, and fibrinogen. Cardiopulmonary bypass (CPB) is known to impair platelet function. In this study, the influence of CPB on formation of coated platelets and the interaction of rFVIIa with the platelets were studied. Blood was either exposed to a closed CPB circuit or obtained from patients undergoing CPB-assisted cardiac surgery, and platelets were analyzed by flow cytometry with and without dual agonist stimulation with thrombin and a GPVI collagen receptor agonist known to induce coated platelet formation. Platelets circulated within a closed CPB circuit did not spontaneously form coated platelets. Dual agonists stimulation caused formation of coated platelets at a reduced level compared to pre-CPB level (51 ± 21% vs. 80 ± 17% before CPB, p < .001). The rFVIIa interaction with the coated platelets was not impaired after CPB. Platelets isolated from patients undergoing CPB-assisted cardiac surgery also formed coated platelets only after dual agonist stimulation but to the same level as before surgery (76 ± 8% vs. 83 ± 14% before surgery, p = .17, n = 10). rFVIIa interaction with the coated platelets was not impaired after surgery. No spontaneous rFVIIa-binding platelets were found. The data indicate that CPB exposure in vivo does not compromise the platelet-dependent effects of rFVIIa either by spontaneous formation of coated platelets, thereby limiting the risk of systemic coagulation, or by impairing rFVIIa interaction with the agonist-induced coated platelets, thereby retaining the hemostatic potential of rFVIIa after CPB.

Patients undergoing cardiopulmonary bypass (CPB)-assisted cardiac surgery are at risk for excessive micro-vascular bleeding, which often leads to allogeneic blood transfusion. Excessive bleeding after cardiac surgery is generally related to either a surgical cause or a combination of several CPB-related alterations of the hemostatic system. In addition to hemodilution, there is excessive activation of the hemostatic system, which may be related to the interaction of blood with the extensive non-endothelial CPB surfaces, activation of the extrinsic clotting pathway secondary to surgical trauma, and re-transfusion of pericardial blood (1,2). Not only has CPB surgery been associated with altered hemostasis, but inflammatory responses are also affected (3). It has been shown that CPB triggers multiple events that lead to platelet and leukocyte activation and subsequent adhesion. Activated platelets rapidly externalize P-selectin (CD62P), an adhesive ligand for leukocytes, from platelet alpha granules to the external cell surface, and spontaneous (i.e., non–agonist-induced) exposure of P-selectin has been implicated after exposure of platelets to CPB (4,5). After activation of adhesive molecules on both platelets and leukocytes, leukocyte–platelet conjugates are formed, and this interaction may augment cellular activation by promoting the release of reactive oxygen species, pro-inflammatory cytokines, and thromboxane.

Recombinant FVIIa (rFVIIa; NovoSeven, Novo Nordisk A/S, Bagsvaerd, Denmark) was developed for the treatment of hemophilia patients with inhibitors (6). Pharmacologic levels of rFVIIa are suggested to promote hemostasis at least in part by interaction with activated platelets (7). rFVIIa promotes thrombin generation and formation of a stable fibrin clot at the site of vascular injury (8). rFVIIa has been reported to control or reduce bleeding in several clinical hemorrhagic indications (9–12). Additionally, there are several case reports on the effective use of rFVIIa after cardiac surgery (13–15). One randomized double-blind placebo-controlled pilot study showed that rFVIIa treatment after CPB surgery significantly reduced the need for allogeneic transfusion compared to placebo without causing adverse events (16).

We have previously reported that rFVIIa interacts preferentially with coated platelets (17) who express high levels of negatively charged phospholipids and retain high levels of several adhesive and procoagulant α-granule proteins on their surface as well as increased FV, FVIII, FIX, and FX binding with increased levels of FXa and thrombin generation (18–20). Although the physiologic significance of coated platelets remains to be elucidated, it is speculated that they play an important role in hemostasis (21). In vitro–coated platelets form when activated by the dual agonists thrombin and a GPVI collagen receptor agonist, whereas no coated platelet formation occurs when platelets are activated with single agonists such as either thrombin or a GPVI collagen receptor agonist (18). We have further shown that formation of coated platelets was impaired in hemophilia A–like conditions in a cell-based model of coagulation and that the addition of rFVIIa partially restored coated platelet formation in hemophilia A–like conditions, most likely as a consequence of enhanced thrombin generation (17).

A crucial question is whether formation of coated platelets in an unexpected and uncontrolled way might be responsible for pathologic effects (21). To date, no studies have, to our knowledge, evaluated the effect of CPB on coated platelets. To this end, we conducted a study to determine whether CPB circulation affects coated platelet formation in vitro in blood undergoing high shear stress in a closed CPB circuit and ex vivo in blood samples collected from cardiac patients undergoing CPB-assisted surgery. Analysis of P-selectin expression was included in the study. Furthermore, we studied the interaction of rFVIIa with platelets exposed to CPB.

MATERIALS AND METHODS

This study was in accordance with the Helsinki Second (1983) Declaration and approved by the local ethical committee. Patients and healthy volunteers gave informed consent to participate in the study before any study-related activities started.

In Vitro Platelet Analysis during Closed CPB Circulation

Whole blood from six healthy volunteer donors was collected and stabilized with heparin (10 IU/mL) to avoid clotting of the blood. A heart–lung machine (Sarns 9000; Bloomfield, CT) with semi-occlusive rollerpumps was used. The circuit consisted of sterile uncoated PVC tubings from Terumo (Tokyo, Japan), a Jostra Quadrox (Maquet, Hirrlingen, Germany), and hollow fiber oxygenator with integrated hardshell venous reservoir (VHK 4210). The oxygenator was coated with bioline coating (Jostra, Hirrlingen, Germany), which is heparin based. In the arterial line, there was a 40-μm arterial filter (Pall, East Hill, NY). Normothermic (35.5–36.5°C) conditions were achieved with a Polystan (Værløse, Denmark) heater-cooler.

The circuit was primed with 700 mL lactated Ringer solution, two portions of red blood cells (200 mL each), two portions of fresh-frozen plasma (200 mL each), and 450 mL blood from each individual donor to achieve a normal concentration of coagulation factors and red blood cells, a colloid osmotic pressure between 16 and 18 mmHg, and a platelet count of ∼100 × 10⁹/L in the CPB circuit. After collection, the whole blood was immediately transferred to the CPB circuit. The diluted blood was circulated at a flow rate at 4.5 L/min. Heparin was added to avoid clotting and to mimic similar conditions as in a clinic setup; activated clotting time (ACT) >480 seconds was maintained at all times.

Samples were obtained for analysis (see below) at 0-, 1-, and 2-hour circulations into citrate-containing vacutainers (BD, Franklin Lakes, NJ) after appropriate heparin reversal with protamine sulfate at 1 mg/100 IU of heparin.

Ex Vivo Platelet Analysis of Patients Undergoing CPB Surgery

Blood samples from 10 patients undergoing CPB surgery were collected from central venous catheters into citrate-containing vacutainers (BD) immediately before surgery started and 5 minutes and 1 hour after heparin reversal with protamine sulfate as guided by the ACT. None of the patients were receiving anticoagulant or antithrombotic therapies 5 days before surgery. Platelet count was measured on an automated hematology cell counter (Sysmex SF-3000; Sysmex Filial Denmark, Almind, Denmark). Demographic characteristics and medical history, including amount of blood loss, transfusion requirements, and time on CPB, was recorded for each patient. Blood samples were analyzed as described below.

A heart–lung machine (Sarns 9000) with semi-occlusive rollerpumps was used. The circuit consisted of sterile uncoated PVC tubings from Terumo and a Jostra Quadrox (Maquet) hollow fiber oxygenator with integrated hard-shell venous reservoir(VHK 4210). The oxygenator was coated with bioline coating (Jostra), which is heparin based. In the arterial line, there was a 40-μm arterial filter (Pall). Normothermic (35.5–36.5°C) conditions were achieved with a Polystan heater-cooler.

The system was primed with 1400 mL lactated Ringer solution, 100 mL mannitol, and 10,000 IU bovine heparin. The patients’ blood was heparinized according to department protocol (350 IU/kg). An ACT > 480 seconds was maintained at all times. Pump flow was set according to department protocol at ≥2.4 L/min/m² and adjusted during the procedure if patient parameters required it. After termination of CPB, heparin was neutralized with protamine (1 mg/100 IU of heparin).

Platelet Preparation and Flow Cytometry

Platelet-rich plasma was prepared within 15 minutes of collection of the blood samples by 10-minute centrif ugation at 200g, and platelets were isolated by gel filtration as described by Albeiro et al. (18). The platelet count was adjusted to ∼40 × 10⁹/L with 15 mmol/L HEPES (Sigma-Aldrich, St. Louis, MO), 138 mmol/L NaCl, 2.7 mmol/L KCl, 1 mmol/L MgCl₂, 5 mmol/L CaCl₂, 5.5 mmol/L dextrose, and 1 mg/mL bovine serum albumin (BSA), pH 7.4. Five microliters platelet suspension was transferred to tubes with or without thrombin (final concentration, 5 nmol/L; Roche Diagnostics, Mannheim, Germany) combined with convulxin (final concentration, 0.5 μg/mL; Pentapharm, Basel, Switzerland), rFVIIa (final concentration, 100 nmol/L; NovoSeven; Novo Nordisk, Bagsvaerd, Denmark), and phycoerythrin (PE)-labeled anti-fibrinogen IgG¹⁸ or PE-labeled control IgG (BD Biosciences, Franklin Lakes, NJ). After 10-minute incubation at 37°C with gentle mixing (samples with agonists) or without agitation (non-stimulated samples), the reactions were stopped by adding 4 volumes of 4°C, 2% para-formaldehyde (Bie and Berntsen, Herlev, Denmark) in HBS (20 mmol/L HEPES, pH 7.5, 150 mmol/L NaCl) with 5 nmol/L CaCl₂. Samples were further stained with PE-labeled anti P-selectin (CD62-PE; BD Biosciences) and fluorescein (FITC)-labeled anti-rFVIIa IgG¹⁸ combined with PerCP-labeled anti-CD61 IgG (PerCP-CD61; BD Biosciences). Samples were analyzed by flow cytometry on a FACScan or a FACS Canto flow cytometer (both BD Biosciences) with forward and side scatter light channels and fluorescence channels set on log. Platelets were identified as CD61-positive cells, and data for 5000 platelets were collected. In some experiments with blood circulated in a closed CPB circuit, data were collected for 1000 CD61-positive events because of low platelet count. Single color-activated platelets before exposure to CPB were used for compensation of overlapping fluorescence. The data were analyzed using FlowJo version 7.2.2 software (Tree Star, Ashland, OR). Strongly P-selectin–positive platelets were quantified in the fluorescence channel measuring anti–P-selectin IgG by using a gate (inclusion criteria) excluding 97% of the platelets in agonist-stimulated samples obtained before exposure to CPB and stained with control IgG (see marker in Figure 1A). Intermediate P-selectin–positive platelets were quantified by using a gate excluding 97% of the platelets in non-stimulated samples obtained before exposure to CPB and stained with anti-CD62 IgG (see marker in Figure 1B). Coated platelets were identified as the platelets with high fibrinogen exposure after stimulation with thrombin combined with convulxin. The gate for the coated platelet population was defined in the fluorescence channel measuring anti–fibrinogen IgG and excluded 97% of the platelets in the control IgG-stained agonist-stimulated sample before exposure to CPB (i.e., gate corresponds to the top left and top right quadrants in Figure 2F). The number of platelets within this gate (coated platelets) was quantified. The mean fluorescence of FVIIa binding to the coated platelets was determined as the mean FITC-fluorescence of the coated platelets. The gate for quantification of number of FVIIa-binding platelets excluded 97% of events in the anti-rFVIIa IgG-channel obtained from the stimulated sample before CPB without rFVIIa added (i.e., containing only anti-FVIIa IgG; corresponding to the top right and the bottom right quadrants in Figure 2F). In two patient data sets and in the data sets from the closed CPB circuit, the gate for rFVIIa was set to exclude 95% of events in the channel measuring anti-rFVIIa IgG to allow optimal separation between samples with and without rFVIIa added. The number of platelets positive for anti-FVIIa IgG after addition of rFVIIa was quantified.

Figure 1.

P-selectin exposure on platelets before and after circulation in a closed CPB circuit. A, Platelets were isolated from blood before circulation in a CBP circuit and stimulated with thrombin combined with convulxin (red and blue histograms) or left non-stimulated (black and green histograms) before staining with anti-P-selectin IgG (red and black histograms) or control IgG (green and blue histograms). The marker indicates platelets highly positive for P-selectin after agonist stimulation. B, Spontaneous P-selectin exposure (noted with the marker) was observed after 1-hour circulation in CPB circuit (red histogram). The blue histogram are the sample obtained after 1-hour circulation in CPB circuit stained with control IgG, and the black histogram is the sample before CPB stained with anti-P-selectin IgG. C, Agonist-induced P-selectin exposure was impaired after circulation in a closed CPB circuit. Platelets before CPB were activated with thrombin combined with convulxin (red histogram) or left non-stimulated (black histogram) and stained with anti-P-selectin IgG. Agonist-activated platelets obtained after 1-hour circulation in the CPB circuit were stained with anti-P-selectin IgG (green histogram) or with control IgG (blue histogram). Data are representative for six experiments.

Figure 2.

Formation of coated platelets and rFVIIa binding to platelets before and after circulation in a closed CPB circuit. Platelets were isolated from blood before (A, D, and F) and after (B, C, and E) 1-hour circulation in a closed CBP circuit, left non-stimulated (A–C), or activated with thrombin combined with convulxin (D–F), and stained with anti-fibrinogen IgG and anti-rFVIIa IgG after addition of 100 nmol/L rFVIIa (A–E) or control IgG and anti-rFVIIa-IgG without rFVIIa added (F). A, B, and D–F are samples from the same experiment, whereas C represents a sample from another experiment treated as sample B.

Statistical Analysis

A two-tailed paired t test was used to compare percent positive platelets before and after exposure to CPB.

RESULTS

In Vitro Effect of Closed CPB Circulation on Platelets

Platelets that had not undergone circulation within CPB could be stimulated with agonists (thrombin plus the GPVI collagen receptor agonist convulxin) to expose a high amount of P-selectin on 79.6% ± 21.9% (SD) of the platelets (Figure 1A; Table 1). After 1 hour of closed CPB circulation spontaneous exposure of low level of P-selectin was indicated by a significant increase in binding of anti-P-selectin IgG in the absence of agonists (Figure 1B; Table 1). However, the anti–P-selectin IgG binding was not different from the control IgG binding, indicating that the signal was caused by increased unspecific binding of the antibodies to the platelets after exposure to CPB. The number of platelets positive for P-selectin after agonist stimulation was significantly decreased after exposure to CPB (48.6% ± 26.5%, p < .01; Figure 1C ; Table 1), supporting CPB-induced impaired platelet function. Data obtained for platelets exposed to CPB for a total of 2 hours were not different from those exposed for 1 hour (Table 1). The total platelet count was substantially decreased after 1-hour circulation (12 ± 9 × 10⁹/L vs. 41 ± 15 × 10⁹ /L before CPB, p < .05, n = 6) in the closed CPB and remained low through 2 hours of circulation (16 ± 9 × 10⁹/L at 2 hours, p < .05 compared with before CPB, n = 5). The data reported here reflect only the intact platelets remaining in circulation.

Table 1.

Flow cytometry analysis of blood before and after 1- or 2-hour circulation of blood in a closed CBP circuit.

			After CPB Circuit
	Before CPB Circuit		1 Hour		2 Hours
	Relevant IgG	Control§	Relevant IgG	Control*	Relevant IgG	Control*
P-selectin
Intermediate positive platelets without agonists added (%)	3.0 ± 0.1	3.9 ± 1.9	8.8 ± 3.7†	8.3 ± 6.7	9.1 ± 3.3‡	5.7 ± 3.9
Highly positive platelets before stimulation with agonists (%)	0.3 ± 0.1	1.4 ± 0.8	1.1 ± 1.4	3.1 ± 2.3	0.8 ± 0.7	1.7 ± 0.9
Highly positive platelets after stimulation with agonists (%)	79.6 ± 21.9	3.1 ± 0.1	48.6 ± 26.5‡	3.6 ± 1.0	50.4 ± 26.1‡	3.2 ± 1.8
Fibrinogen (coated platelets)
Positive platelets before stimulation with agonists (%)	0.3 ± 0.3	1.4 ± 0.9	1.4 ± 2.0	3.5 ± 2.6	0.6 ± 0.6	1.9 ± 0.9
Positive platelets after stimulation with agonists (%)	82.0 ± 17.1	2.9 ± 0.1	51.2 ± 21.0§	3.8 ± 1.2	41.1 ± 24.9‡	2.8 ± 1.1
rFVIIa binding to platelets*
Positive platelets before stimulation with agonists (%)	4.9 ± 1.6	2.6 ± 2.6	18.6 ± 6.5‡	4.3 ± 2.2	22.8 ± 9.2‡	4.0 ± 3.2
Positive platelets after stimulation with agonists (%)	66.7 ± 20.5	6.6 ± 3.5	55.8 ± 17.4	8.8 ± 5.1	60.4 ± 20.5	7.3 ± 5.9

Data are mean and SD of n = 6 experiments determining the percentage of platelets within the noted gates (inclusion criteria, see Methods and Figures 1 and 2).

^*Controls for background fluorescence are control IgG for P-selectin and fibrinogen staining and samples with only anti-rFVIIa IgG, i.e., without rFVIIa added, for analysis of rFVIIa-binding platelets.

^†p < .05,

^‡p < .01, and

^§p < .001 compared with sample with same treatment and staining before CPB.

In the presence of thrombin plus convulxin, platelets could be stimulated to form coated platelets after circulation in a closed CPB circuit (51.2% ± 21.0%); however, the amount of coated platelets formed was significantly decreased compared with before CPB (82.0% ± 17.1%, p < .001; Figure 2 ; Table 1). The coated platelets formed after CPB were, however, still able to bind rFVIIa with the same intensity [mean fluorescence intensity (MFI) of anti-rFVIIa IgG binding after adding 100 nmol/L rFVIIa was 145 ± 72 before CPB, 157 ± 37 after 1 hour of CPB, and 175 ± 50 after 2 hours of CPB; no significant differences, n = 6]. No spontaneous formation of coated platelets (i.e., without agonist stimulation) was observed by exposure to CPB (Table 1). In the absence of agonist stimulation, significantly more rFVIIa was found on fibrinogen-negative platelets after exposure to CPB compared with before circulation (18.6% ± 6.5% positive cells after 1 hour of CPB and 22.8% ± 9.2% at 2 hours vs. 4.9% ± 1.6% before CPB, p < .01; Figure 2A–C; Table 1), although this was only a minor fraction of the total number of platelets. When the number of agonist-stimulated platelets positive for rFVIIa binding was quantified (see Materials and Methods), no significant decrease was observed after CPB (Table 1), most likely because of the counteracting contribution of the spontaneous rFVIIa-binding platelets.

Ex Vivo Effect of CPB on Platelets in Patients Undergoing Cardiac Surgery

Patient characteristics are presented in Table 2. Four of the patients underwent standard coronary artery bypass grafting surgery (CABG), one patient underwent single valve replacement (VR), and the remaining five underwent the more complicated double VR, expected to require longer CBP time and thereby having a higher risk of CPB-induced complications. Platelets were isolated from citrate-stabilized blood samples obtained before surgery and 5 minutes and 1 hour after heparin reversal with protamine sulfate as guided by the ACT and analyzed for spontaneous and agonist-induced P-selectin exposure, formation of coated platelets, and rFVIIa binding, as for blood samples exposed to the closed CPB circuit. Because no differences were found between the patients undergoing CABG and the single or double VR procedure, all flow cytometry data were pooled (Table 3). Unlike the results from the closed CPB circuit, P-selectin exposure in ex vivo agonist-activated platelets was not impaired in patients undergoing CPB surgery, nor was formation of coated platelets impaired. Binding of rFVIIa to agonist-activated platelets was not different before and after surgery. Importantly, no rFVIIa binding to non–agonist-stimulated fibrinogen-negative platelets was detected.

Table 2.

Demographic data on patients undergoing CPB-assisted cardiac surgery.

Patient No.	Age/Sex	Surgery	Time on CPB (min)	Platelet Count (×109/L)			Bleeding Intra/Post-operative (mL)	Transfusion
				Before Surgery	5 Minutes after Heparin Reversal	1 Hour after Heparin Reversal
1	80/male	CABG	60	256	200	213	500/300	2 RBC
2	68/male	CABG	84	188	128	166	200/400	No
3	66/male	CABG	65	235	170	198	100/450	No
4	66/male	VR	97	380	200	195	400/250	No
5	60/male	CABG	47	256	158	168	200/200	No
6	45/female	2 × VR	110	330	210	223	1000/900	4 RBC
7	19/male	2 × VR	85	229	160	167	400/750	2 RBC
8	72/female	2 × VR	93	310	213	210	300/500	No
9	68/male	2 × VR	88	179	121	148	1000/300	2 RBC
10	55/male	2 × VR	100	268	142	155	800/350	1 RBC

CABG, cardiac artery bypass grafting; VR, valve replacement; 2 × VR, double VR; RBC, red blood cells.

Table 3.

Flow cytometry analysis of blood samples obtained before and after CPB-assisted cardiac surgery.

	Pre-Surgery		5 Minutes after Heparin Reversal		1 Hour after Heparin Reversal
	Relevant IgG	Control*	Relevant IgG	Control*	Relevant IgG	Control*
P-selectin (n = 9)
Intermediate positive platelets without agonists added (%)	3.1 ± 0.1	2.9 ± 3.7	5.1 ± 4.8	3.9 ± 2.9	3.5 ± 2.6	2.9 ± 2.8
Highly positive platelets before stimulation with agonists (%)	0.3 ± 0.3	0.9 ± 0.7	0.4 ± 0.7	0.4 ± 0.7	0.2 ± 0.3	0.7 ± 0.5
Highly positive platelets after stimulation with agonists (%)	87.4 ± 10.3	3.1 ± 0.1	86.3 ± 11.3	4.4 ± 2.5	90.4 ± 7.4	3.7 ± 2.9
Fibrinogen (coated platelets, (n = 10)
Positive platelets before stimulation with agonists (%)	0.3 ± 0.2	1.1 ± 0.8	0.5 ± 0.6	1.6 ± 1.5	0.5 ± 0.7	0.9 ± 0.4
Positive platelets after stimulation with agonists (%)	83.3 ± 14.2	2.8 ± 0.6	75.8 ± 8.4	4.5 ± 3.1	82.1 ± 6.6	3.5 ± 3.3
rFVIIa binding to platelets (n = 10)*
Positive platelets before stimulation with agonists (%)	7.4 ± 6.5	1.5 ± 0.8	5.6 ± 3.8	2.6 ± 1.7	3.7 ± 2.0	1.4 ± 0.8
Positive platelets after stimulation with agonists (%)	74.8 ± 13.4	3.6 ± 1.3	68.9 ± 14.0	6.1 ± 4.1†	82.8 ± 7.3	5.2 ± 4.4†

Data are mean and SD of percentage of platelets within the noted gates (inclusion criteria, see Methods and Figures 1 and 2).

^*Controls for background fluorescence are control IgG for P-selectin and fibrinogen staining and samples with only anti-rFVIIa IgG, i.e., without rFVIIa added, for analysis of rFVIIa-binding platelets.

^†p < .05 compared with sample with same treatment and staining before CPB.

DISCUSSION

The effect of CPB circulation on platelet membrane proteins including coated platelet formation was analyzed by flow cytometry. Two scenarios were analyzed. Blood samples from normal donors were analyzed in vitro before and after circulation in a closed CPB circuit. Subsequently, ex vivo analysis was performed on blood samples obtained from cardiac patients undergoing CPB-assisted surgery.

In the absence of agonists, we detected a platelet population that bound low levels of anti–P-selectin IgG after exposure to a closed CPB circuit. These results are consistent with other studies showing increased binding of anti–P-selectin IgG after exposure to CPB (4,5). However, it is not clear from these studies whether exposure to CPB also induced increased binding of a control IgG. Our data showing a similar increase in binding of control IgG and anti–P-selectin IgG to platelets after exposure to CPB ruled out spontaneous (i.e., non–agonist-induced) P-selectin expression on platelets exposed to CPB in this study. In contrast, the agonist-induced P-selectin exposure was ∼30% lower after circulation in the closed CPB circuit, suggesting impaired platelet function after CPB. This is in contrast to our and other (22,23) results on platelets isolated from patients undergoing CPB-assisted cardiac surgery, where the percentage of platelets positive for P-selectin after stimulation was not different from the values obtained before surgery.

The use of a dual agonist in stimulation of platelets allowed analysis of the platelets’ capacity to form coated platelets after exposure to CPB. Although maximal P-selectin exposure is obtained after thrombin activation, formation of coated platelets requires stimulation with two strong agonists (i.e., thrombin combined with a collagen GPVI receptor agonist-like convulxin) (18,20). Agonist-induced formation of coated platelets was substantially impaired after circulation in a closed CPB circuit. Because coated platelets are hypothesized to constitute the first layer of activated platelets attached to the injured vessel wall, thereby playing a pivotal role in hemostasis (21), the capacity to form coated platelets may be essential for preventing bleeding after surgery. The normal agonist-induced formation of coated platelets after CPB-assisted cardiac surgery indicated—similar to the data for P-selectin—that the patients’ platelets retain the capacity for normal platelet function. Importantly, coated platelets were not formed spontaneously (i.e., without stimulation with agonists) either after exposure to the high shear stress in the closed CPB circuit or after in vivo exposure to CPB. These results indicate that coated platelet formation cannot be induced by high shear and/or contact with artificial surfaces in the CPB circuit and is regulated by agonists and signals released from within the injured vessel wall. Alternatively, CPB-induced coated platelets may have been trapped in the CPB circuit or in other ways cleared from circulation and therefore not detected in our study.

rFVIIa binding to the agonist-induced coated platelets was not impaired by CPB either after circulation in the closed CPB circuit or in patients undergoing CPB-assisted cardiac surgery. We found some rFVIIa on platelets with low fibrinogen exposure (i.e., not coated platelets) after circulation in the closed CPB circuit. However, in the ex vivo patient samples, no rFVIIa was found on non–agonist-stimulated samples, suggesting that this platelet population was not formed in vivo or was cleared from circulation. rFVIIa was previously found to localize preferentially coated platelets, thereby suggesting restricted thrombin generation at the site of injury where bothcollagen and tissue factor are exposed (17). The binding of rFVIIa to coated platelets was dependent on the Gladomain of rFVIIa and on exposure of negatively charged phospholipids on the surface of the platelets. This suggests that the binding of rFVIIa to non–agonist-stimulated platelets after circulation in the closed CPB circuit may be caused by spontaneous exposure of negatively charged phospholipids after exposure to the high shear stress of the CPB circuit. Pathologic high shear stress (117–388 dyn/cm²) has been shown to trigger apoptotic events in platelets, including exposure of the negatively charged phospholipid phosphatidyl serine (PS)(24), and it is possible that the enhanced shear provided by the closed CPB (∼110 dyn/cm²) may have induced low-level PS exposure and thereby facilitated rFVIIa binding to the platelets. Induction of apoptosis by mitochondrial permeability transition pore (MPTP) activators has been shown to result in expression of markers of coated platelets including PS exposure on the outer platelet leaflet (25). However, in this study, shear-induced formation of coated platelets could not be detected. The difference between PS exposure as a consequence of shear-induced apoptotic events and the formation of coated platelets remains to be fully understood.

Because rFVIIa is a procoagulant protein, there has been concern that use of such a hemostatic agent would induce uncontrolled thrombosis or disseminated intravascular disease under pathologic conditions such as during cardiac surgery (26). Although there is concern that CPB alters platelet function during CPB (1,27), this study confirmed that CPB does not spontaneously induce formation of the highly procoagulant coated platelets where maximal procoagulant function of rFVIIa is expected. The hemostatic potential of the rFVIIa-binding fibrinogen-negative platelets found after in vitro exposure to the closed CPB circuit is not known. However, the lack of this platelet subpopulation after in vivo exposure to CPB suggested limited risk of systemic coagulation with rFVIIa therapy after CPB-assisted cardiac surgery. Additionally, because interaction of rFVIIa with the agonist-induced coated platelets was not impaired by CPB exposure, our data provide evidence to support a hemostatic mechanism by which rFVIIa may safely control bleedings after CPB-assisted cardiac surgery. However, the effective and safe use of rFVIIa in cardiac surgery remains to be confirmed in randomized controlled studies.

This study has important limitations. First, the closed CPB circuit does not necessarily reflect in vivo conditions and, hence, the results obtained should be interpreted with caution. Second, the clinical significance of the ex vivo study is limited by the fact that none of the patients included actually developed a critical bleeding condition where administration of rFVIIa would be considered. We therefore recommend that such a study should be performed.

In summary, our study showed that CPB circulation alone did not spontaneously induce coated platelet formation and that coated platelet formation could only be stimulated by the presence of dual agonists, thrombin and convulxin, both before and after CPB. Furthermore, platelet interaction with rFVIIa was not impaired by CPB-assisted cardiac surgery.