Bone Marrow Transplantation, (1999) 23, 991–996 1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt A prospective study of G-CSF effects on hemostasis in allogeneic blood stem cell donors R LeBlanc1, J Roy1, C Demers1,LVu2 and G Cantin1 1St Sacrement Hospital, Laval University, Quebec City; and 2St Sacrement Hospital, Quebec City, Quebec, Canada Summary: The last few years have witnessed an important increase in the use of peripheral blood stem cells (PBSC) for allogeneic Granulocyte colony-stimulating factor (G-CSF) is used in transplantation. Granulocyte colony-stimulating factor (G- healthy donors of peripheral blood stem cells (PBSC) for CSF) is currently the most widely used cytokine for stem allogeneic transplantation. However, some data have cell mobilization. Exposing a large number of otherwise recently suggested that G-CSF may induce a hypercoag- healthy individuals to G-CSF raises concerns about poten- ulable state, prompting us to study prospectively 22 tial adverse effects.1 Although usually well tolerated, G- PBSC donors before and after G-CSF 5 ␮g/kg twice CSF administration commonly induces side-effects such as daily. We sought evidence for changes in the following bone pain, headaches, fatigue and nausea. Anxiety, noncar- parameters: platelet count, von Willebrand factor anti- diac chest pain, myalgia, insomnia, fever, night sweats, skin gen (vWF:Ag) and activity (vWF activity), ␤-thrombo- rash, anorexia, dizziness, weight gain, local reactions at the globulin (␤-TG), platelet factor 4 (PF-4), platelet acti- injection site and vomiting have been infrequently reported. vation markers (GMP-140 and PAC-1), activated partial G-CSF has been extensively used for the treatment of neu- thromboplastin time (aPTT), prothrombin time (PT), tropenia following chemotherapy, but short- and long-term coagulant factor VIII (FVIII:C), thrombin–antithrombin side-effects of G-CSF in normal individuals are not well complex (TAT), prothrombin fragment 1+2(F1+2), defined. thrombomodulin (TM) and tissue plasminogen activator In a few small studies, G-CSF was found to induce plate- antigen (tPA:Ag) prior to G-CSF and immediately before let activation. Shimoda and others have shown that G-CSF leukapheresis. ADP-induced platelet aggregation studies receptors are present on the surface of platelets.2,3 These were also performed. G-CSF administration produced receptors appear to be functional in vitro2 and in vivo4,5 in only mild discomfort. We found a significant increase in healthy volunteers. However, the consequences on hemo- vWF:Ag (from 0.99 ؎ 0.32 U/ml to 1.83 ؎ 0.69 U/ml; P stasis of these findings are conflicting. Other investigators Ͻ 0.001), in vWF activity (from 1.04 ؎ 0.34 U/ml to have also observed an increase in platelet aggregation, in U/ml; P Ͻ 0.001) and in FVIII:C (from platelet activation, in von Willebrand factor and in 0.50 ؎ 1.78 U/ml to 1.73 ؎ 0.57 U/ml; P Ͻ 0.001) after thrombin–antithrombin complex after G-CSF, but these 0.37 ؎ 1.12 G-CSF. Of note, four donors with low baseline vWF had results could not be confirmed.2,4–10 a two- to three-fold increase after receiving G-CSF. G- To date, two cases of thrombosis occurring in normal CSF had no impact on the platelet count, ␤-TG, PF-4, donors have been reported.11 The first one was a healthy GMP-140 or PAC-1. The final% of platelet aggregation 54-year-old woman who developed a cerebrovascular acci- decreased from 73 ؎ 22% to 37 ؎ 26% after G-CSF (P dent 2 days after completing an uneventful stem cell Ͻ 0.001). We found a significant decrease in aPTT after donation. The second one was a 64-year-old man with a -G-CSF (29.9 ؎ 3.1 s to 28.3 ؎ 3.3 s; P = 0.004), but the history of coronary artery disease who experienced a myo PT was unaffected. In addition, we also observed a sig- cardial infarction shortly after PBSC collection. nificant increase in TAT, F1+2 and TM, but not in These clinical and biological observations raise concerns tPA:Ag. Our data suggest that G-CSF may possibly about the possibility that G-CSF might in fact induce a hy- induce a hypercoagulable state by increasing levels of percoagulable state in healthy donors and consequently, FVIII:C and thrombin generation. In contrast to this increase the risk of thrombosis. In order to better define the information, we found reduced platelet aggregation after effects of G-CSF on hemostasis, we undertook a prospec- G-CSF administration. The clinical implications of these tive study in healthy PBSC donors focusing on short-term findings remain unclear and larger studies are defi- adverse effects and on hemostatic changes immediately nitely required. after G-CSF administration. Keywords: granulocyte colony-stimulating factor; plate- let activation; thrombosis; healthy peripheral blood stem cell donors Materials and methods Study design Correspondence: Dr R LeBlanc, CHA, Pavillon Saint-Sacrement, 1050, chemin Sainte-Foy, Que´bec (Qc), Canada G1S 4L8 This is a cohort study, conducted at St Sacrement Hospital, Received 27 July 1998; accepted 22 December 1998 Laval University, Quebec City, Canada. It was approved Effect of G-CSF on hemostasis in allogeneic blood stem cell donors R LeBlanc et al 992 by the institutional review board and written consent was coagulant factor VIII (FVIII:C), von Willebrand factor anti- obtained from all healthy donors prior to participation. gen (vWF:Ag) and activity (vWF activity), thrombin– antithrombin complex (TAT), prothrombin fragment 1+2 + Patients (F1 2), thrombomodulin (TM) and tissue plasminogen acti- vator antigen (tPA:Ag) assays were performed on plasma Between July 1996 and April 1998, all consecutive healthy obtained from anticoagulated blood with citrate (Becton allogeneic PBSC donors were considered for enrollment in Dickinson, 0.105 mol/l) and centrifuged at 2500 g for 15 the study. All donors studied were used as their own con- min at 4°C. Plasma samples were analyzed immediately or trol. Donors were excluded if they had an abnormal platelet stored in aliquots at Ϫ80°C until tested. For assaying, count, bleeding time, activated partial thromboplastin time samples were thawed for 15 min in a 37°C water-bath. The (aPTT) or prothrombin time (PT) prior to G-CSF. Other Monovette blood collection system (Sarstedt, NC, USA) exclusion criteria were a history of unstable angina, myo- containing sodium citrate (0.105 mol/l) was used for plate- cardial infarction, ischemic or thrombotic stroke in the past let aggregation. Platelet-rich plasma was obtained by centri- 3 months, active neoplasia, active inflammatory disease, fugation at 250 g for 6 min at 27°C and was adjusted with pregnancy or a contra-indication to blood donation. platelet-poor plasma (2500 g for 15 min at 27°C) to a plate- let count of 200 × 109/l. Intervention Assays: Ivy bleeding time was performed with the Simp- All donors were evaluated initially by a physician and a late IIR device (Organon Teknika, Durham, NC, USA) with nurse with a complete medical history and physical examin- longitudinal incisions. Platelet counts were measured on a ation. At their first visit, they were instructed on how to Coulter STKS (Miami, FL, USA). Platelet activation mark- administer their medication and on the potential side- ers (GMP-140 and PAC-1) were measured by flow cytome- effects. Before the first dose of G-CSF, blood samples were try with argon laser generating laser light at 488 nm drawn for laboratory evaluation. All donors received G- (Coulter Epics XL-MCL) using 20 ␮l of anti-P-selectin (Pe, CSF (Neupogen, Amgen, Mississauga, Canada) at 5 ␮g/kg B-D) fluorescing at 575 nm and 20 ␮l anti-gpIIb-IIIa subcutaneously twice daily for nine consecutive doses. (FITC, B-D) fluorescing at 525 nm, respectively. Platelet After the ninth dose on day 5, blood samples were drawn aggregation was performed on the PACKS-4 platelet aggre- for the same assays immediately before undergoing leu- gometer (Helena Laboratories, Beaumont, TX, USA) using kapheresis. A personal diary was provided to all donors in 3 ␮mol/l ADP (Boehringer Mannheim, Indianapolis, IN, order to record prospectively any side-effects experienced USA). PT and aPTT were carried out on the Behring during the period of G-CSF administration. Side-effects Fibrintimer A (BFA, Behringwerke, Marburg, Germany) were graded by one investigator after discussion with the using Thromborel-S (Behring) and Automated APTT donor on a scale of 1 to 4; grade 1 were mild side-effects (Organon Teknika) respectively. FVIII:C was measured by with no intervention needed; grade 2 were moderate side- one stage clotting assay on the BFA instrument using effects needing some intervention (acetaminophen to immunoadsorbed FVIII deficient plasma (Dade, Miami, FL, relieve bone pain or fever) but not interfering with drug USA). vWF:Ag, ␤-TG, PF-4, TM and tPA:Ag were perfor- administration; grade 3 were severe side-effects interfering med by enzyme-linked immunosorbent assays (ELISA) with normal daily activity or with the need to discontinue using Asserachrom kits (Diagnostica Stago, Asnie`res-sur- medication; grade 4 were life-threatening complications Seine, France). vWF activity was determined using IMUB- related to the medication. In order to attribute symptoms IND vWF ELISA kit (American Diagnostica, Greenwich, only to G-CSF administration, all donors were instructed CT, USA). Enzygnost ELISA kits (Behringwerke) were to start recording symptoms in their diary 5 days prior to used for TAT and F1+2 determinations. ELISA assays were G-CSF administration. all carried out on CODA Automated EIA Analyzer (Bio- Rad, Hercules, CA, USA). Normal pooled plasma (NPP) Laboratory testing was used for calibration with a defined value of 1 U/ml.