Patient blood management knowledge and practice among clinicians from seven European university hospitals: a multicentre survey.

BACKGROUND AND OBJECTIVES: The aim of this survey was to evaluate the knowledge about Patient Blood Management (PBM) principles and practices amongst clinicians working in seven European hospitals participating in a European Blood Alliance (EBA) project. MATERIALS AND METHODS: A web-based questionnaire was sent to 4952 clinicians working in medical, surgery and anaesthesiology disciplines. The responses were analysed, and the overall results as well as a comparison between hospitals are presented. RESULTS: A total of 788 responses (16%) were obtained. About 24% of respondents were not aware of a correlation between preoperative anaemia (POA) and perioperative morbidity and mortality. For 22%, treatment of POA was unlikely to favourably influence morbidity and mortality even before surgery with expected blood loss. More than half of clinicians did not routinely treat POA. 29%, when asked which is the best way to treat deficiency anaemia preoperatively, answered that they did not have sufficient knowledge and 5% chose to 'do nothing'. Amongst those who treated POA, 38% proposed red cell transfusion prior to surgery as treatment. Restrictive haemoglobin triggers for red blood cell transfusion, single unit policy and reduction of number and volumes of blood samples for diagnostic purposes were only marginally implemented. CONCLUSION: Overall, the responses indicated poor knowledge about PBM. Processes to diagnose and treat POA were not generally and homogeneously implemented. This survey should provide further impetus to implement programmes to improve knowledge and practice of PBM.

Patient Blood Management in Europe: surveys on top indications for red blood cell use and Patient Blood Management organization and activities in seven European university hospitals.

BACKGROUND AND OBJECTIVES: Patient Blood Management (PBM) in Europe is a working group of the European Blood Alliance with the initial objective to identify the starting position of the participating hospitals regarding PBM for benchmarking purposes, and to derive good practices in PBM from the experience and expertise in the participating teams with the further aim of implementing and strengthening these practices in the participating hospitals. METHODS: We conducted two surveys in seven university hospitals in Europe: Survey on top indications for red blood cell use regarding usage of red blood cells during 1 week and Survey on PBM organization and activities. RESULTS: A total of 3320 units of red blood cells were transfused in 1 week at the seven hospitals. Overall, 61% of red cell units were transfused to medical patients and 36% to surgical patients, although there was much variation between hospitals. The organization and activities of PBM in the seven hospitals were variable, but there was a common focus on optimizing the treatment of bleeding patients, monitoring the use of blood components and treatment of preoperative anaemia. CONCLUSION: Although the seven hospitals provide a similar range of clinical services, there was variation in transfusion rates between them. Further, there was variable implementation of PBM activities and monitoring of transfusion practice. These findings provide a baseline to develop joint action plans to further implement and strengthen PBM across a number of hospitals in Europe.

Peri-partum reference ranges for ROTEM(R) thromboelastometry.

BACKGROUND: Post-partum haemorrhage (PPH) causes rapidly developing deficiencies in clotting factors and contributes to substantial maternal morbidity and mortality. Rotational thromboelastometry (ROTEM(®)) is increasingly used as a point of care coagulation monitoring device in patients with massive haemorrhage; however, there are limited data on reference ranges in the peri-partum period. These are required due to the haemostatic changes in pregnancy. METHODS: In a Dutch multi-centre trial, 161 subjects were included; blood samples were obtained during labour (T1) and within 1 h of delivery (T2). Reference ranges of ROTEM(®) INTEM, EXTEM, FIBTEM, and APTEM were set and correlation with laboratory results was investigated using the guidelines of the International Federation of Clinical Chemistry. RESULTS: Reference ranges were obtained for clotting time (CT), clot formation time (CFT), α-angle, clot firmness at 10 and 20 min (A10, A20), maximum clot firmness (MCF), and maximum lysis (ML). These were comparable from centre to centre, and between T1 and T2. Reference ranges T1: EXTEM: CT 31-63 s, CFT 41-120 s, and MCF 42-78 mm. INTEM: CT 109-225 s, CFT 40-103, and MCF 63-78 mm. FIBTEM: CT 31-79 s and MCF 13-45 mm. APTEM: CT 33-62 s, CFT 42-118, and MCF 61-79 mm. CONCLUSIONS: Reference values for ROTEM(®) parameters are reported. The previously published correlation between FIBTEM parameters and plasma fibrinogen levels by the Clauss method is confirmed. Further research is needed to define threshold values for haemostatic therapy in the course of PPH. Clinical trial registration NTR 2515 (http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=2515).

Perioperative dilutional coagulopathy treated with fresh frozen plasma and fibrinogen concentrate: a prospective randomized intervention trial.

BACKGROUND AND OBJECTIVES: Treatment of dilutional coagulopathy by transfusing fresh frozen plasma (FFP) remains sub-optimal. We hypothesized that partial replacement of transfused FFP by fibrinogen concentrate results in improved coagulant activity and haemostasis. This was tested in a controlled clinical intervention trial with patients experiencing massive bleeding during major surgery. METHODS: Patients undergoing major elective surgery were treated according to current protocols. When transfusion with FFP was required, patients were randomized as follows: group A received 4 units FFP and group B received 2 units FFP plus 2 g fibrinogen concentrate. Blood samples were taken before and after the intervention. Analysts were blinded to the treatment type. RESULTS: Group A (B) consisted of 21 (22) patients, in 16 (17) of whom bleeding stopped after intervention. Plasma fibrinogen increased significantly more in group B (0·57 g/l) than in group A (0·05 g/l). However, levels of prothrombin and factors VIII, IX and X increased more in group A than in group B. Rotational thromboelastometry (ROTEM) of whole blood and plasma revealed improved fibrin clot formation in group B but not in group A. Thrombin generation [calibrated automated thrombogram (CAT)] in plasma increased more in group A. Principal parameters determining whole-blood thromboelastometry were the fibrinogen level and platelet count. In vitro addition of fibrinogen and prothrombin complex concentrate to pre-intervention samples restored both ROTEM and CAT parameters. CONCLUSIONS: Partial replacement of transfused FFP by fibrinogen increases fibrin clot formation at the expense of less improved thrombin generation. Coagulation factors other than fibrinogen alone are required for full restoration of haemostasis.

Thrombotic and infectious complications of central venous catheters in patients with hematological malignancies.

Central venous catheters (CVCs) have considerably improved the management of patients with hematological malignancies, by facilitating chemotherapy, supportive therapy and blood sampling. Complications of insertion of CVCs include mechanical (arterial puncture, pneumothorax), thrombotic and infectious complications. CVC-related thrombosis and infections are frequently occurring complications and may cause significant morbidity in patients with hematological malignancies. CVC-related thrombosis and infections are related and can therefore not be seen as separate entities. The incidence of symptomatic CVC-related thrombosis had been reported to vary between 1.2 and 13.0% of patients with hematological malignancy. The incidence of CVC-related bloodstream infections varies between 0.0 and 20.8%. There is need for a specific approach regarding diagnosis and treatment of CVC-related thrombosis and infection with specific attention to the preservation of the catheter. Since data on CVC-related infections and thrombosis in hematological patients have been obtained mainly from retrospective studies of small sample size, prospective, randomized studies of prophylactic measures concerning CVC-related thrombosis and infection are warranted.

[Coagulation disorders in patients with trauma to the skull and brain: a frequent and potentially fatal combination]. / Stollingsstoornissen bij patiënten met schedel-hersentrauma: een frequente en potentieel fatale combinatie.

Abnormalities in blood coagulation are relatively common after traumatic brain injury (TBI) and play an important role in the morbidity and mortality after head injuries. Exposure oftissue factor, which is abundantly present in brain tissue, is the initiator of the coagulation cascade and plays an important role in the pathogenesis of coagulopathy after TBI. Coagulopathy after TBI is actually a manifestation of the disseminated intravascular coagulation (DIC) syndrome. The interplay between hypothermia, acidosis and progressive coagulopathy, referred to as the 'lethal triad', results in high mortality. This necessitates damage control in the treatment of such TBI patients. Repeated laboratory evaluation of the coagulation parameters in TBI patients is indicated, even if the initial values are normal. The DIC score, a combination of frequently used coagulation parameters, is not only a measure of the coagulopathy but can also predict the outcome and prognosis following TBI. Primary and secondary prevention of coagulopathy together with timely and effective intervention are the most important elements in the treatment of coagulation disorders. Nevertheless, the risk of death remains high.

Recombinant factor VIIa enhances platelet adhesion and activation under flow conditions at normal and reduced platelet count.

BACKGROUND: Recombinant factor VIIa (rFVIIa), which was developed for treatment of inhibitor-complicated hemophilia, appears suitable as prohemostatic agent in other clinical disorders including patients with thrombocytopenia. It is generally accepted that rFVIIa functions by enhancement of thrombin generation at the site of injury. It is, however, unknown if and how this affects platelet adhesion and aggregation. OBJECTIVES: To determine the effect of rFVIIa-mediated thrombin generation on platelet adhesion and aggregation under flow conditions at normal and reduced platelet counts. METHODS: Washed platelets and red cells were combined to obtain plasma-free blood with different platelet counts. The reconstituted blood was perfused over a collagen- or fibrinogen-coated surface in the absence or presence of a thrombin generating system consisting of purified coagulation factors rFVIIa, factor (F)X and prothrombin. RESULTS: Addition of coagulation factors rFVIIa, FX and prothrombin to washed platelets and red cells enhanced platelet adhesion and aggregation to collagen and adhesion and spreading to fibrinogen at normal platelet count and at platelet numbers as low as 10 000 microL(-1). rFVIIa-mediated thrombin generation enhanced the activation state of platelets as measured by intracellular calcium fluxes, and enhanced the exposure of procoagulant phospholipids as measured by annexin A5 binding. CONCLUSIONS: Taken together, increased platelet adhesion and aggregation by rFVIIa-mediated thrombin formation may explain the therapeutic effects of rFVIIa in thrombocytopenic conditions and in patients with a normal platelet count by (i) enhancement of primary hemostasis and (ii) enhancement of procoagulant surface leading to elevated fibrin formation.

Elevated levels of FVIII:C within families are associated with an increased risk for venous and arterial thrombosis.

Elevated levels of coagulation factor VIII:C (FVIII:C) are associated with an increased risk for venous and arterial thromboembolism. Whether relatives of patients with elevated levels of FVIII:C are also at increased risk for thrombotic disease is unknown. The objective was to determine the annual incidences of both venous and arterial thrombotic events in first-degree relatives of patients with elevated levels of FVIII:C and venous thromboembolism (VTE) or premature atherosclerosis. A retrospective study with 584 first-degree relatives of 177 patients with elevated levels of FVIII:C was performed. The level of FVIII:C was determined and relatives with elevated and normal levels of FVIII:C were compared. Of the participants, 40% had elevated levels of FVIII:C. The annual incidence of a first episode of VTE was 0.34% and 0.13% in relatives with elevated levels of FVIII:C and those with normal levels, respectively [OR 3.7 (95% CI 1.9-7.5)]. The absolute annual incidence in the youngest age group with elevated levels of FVIII:C was 0.16% (0.05-0.37) and gradually increased to 0.99% (0.40-2.04) in those older than 60 years of age, although the odds ratios were not statistically significant. The annual incidences of a first arterial thrombotic event were 0.29% and 0.14% in relatives with and without elevated levels of FVIII:C, respectively [OR 3.1 (1.4-6.6)]. In particular the risks for a first myocardial infarction [OR 4.3 (1.0-18.1); P =0.046] and a first peripheral arterial thrombosis [OR 8.6 (1.6-47.6)] were increased. Within families of patients with elevated levels of FVIII:C and VTE or premature atherosclerosis, 40% of their first-degree relatives has elevated levels of FVIII:C as well, and they are at increased risk for both VTE and arterial thrombosis as compared with their relatives with normal levels.

Cardiac tamponade from misplaced central venous line in pericardiophrenic vein.

A patient in whom a left internal jugular vein catheter had first migrated into the left pericardiophrenic vein, and subsequently had perforated into the pericardium leading to a cardiac tamponade is described. Although this malposition has rarely been reported, it does not seem to be so infrequent, as three other similar misplacements have occurred in our institution. This malposition can be prevented by a high degree of suspicion, preferential use of the right internal jugular vein for catheterization, routine use of a J-tipped guidewire, limiting the depth of insertion of the guidewire during cannulation, routine roentgenographic control of radiopaque catheters, and (slow) injection of a small volume of radiopaque dye through the central venous catheter.

The prognostic value of the soluble urokinase-type plasminogen activator receptor (s-uPAR) in plasma of breast cancer patients with and without metastatic disease.

Elevated levels of soluble uPAR (s-uPAR) and other fibrinolytic parameters functionally related to the urokinase-type plasminogen activator system might indicate the presence of cancer cells. In 25 breast cancer patients with metastases s-uPAR was significantly increased compared with 25 patients without metastases and with 25 healthy controls: 420 pg mL-1 vs. 145 pg mL-1 (P = 0.005) and 190 pg mL-1 (P = 0.003). Plasmin-alpha2-antiplasmin (PAP) complexes and d-dimers were significantly increased in breast cancer patients with metastases compared with patients without metastases and with healthy controls. The levels of plasminogen activator inhibitor (PAI)-1 activity, uPA antigen and factor (F)XIIa did not significantly differ between the patient groups and healthy controls. PAP complexes (529 microg L-1 vs. 420 microg L-1; P = 0.03), d-dimers (278.5 ng mL-1 vs. 79.0 ng mL-1; P = 0.005) and FXIIa (1.64 ng mL-1 vs. 1.19 ng mL-1; P = 0.01) were significantly higher in patients with metastases not surviving compared with patients with metastases surviving the 3-year follow-up period. Plasma s-uPAR levels in the patients with metastases did not discriminate between patients surviving and patients not surviving after 3-year follow-up. No significant differences in s-uPAR or any of the other parameters were found in the five patients developing metastases during follow-up. A single value of s-uPAR is of limited value in the follow-up of breast cancer patients with and without metastatic disease and does not predict survival or future metastases.

Risk of venous thromboembolism in carriers of factor V Leiden with a concomitant inherited thrombophilic defect: a retrospective analysis.

Factor V Leiden is the most common genetic defect associated with venous thromboembolism. Its clinical expression is limited and shows a wide intrafamilial and interfamilial variation, which might be explained by the influence of other genetic risk factors. We retrospectively studied 226 patients with factor V Leiden and documented venous thromboembolism (probands) and 400 first-degree carrier relatives to assess the contribution of concomitant genetic risk factors to the occurrence of venous thromboembolism. The prothrombin G20210A mutation was found in 8.3%, homozygosity of factor V Leiden in 7.2%, and inherited deficiencies of antithrombin, protein C or protein S in 4.7% of symptomatic carriers (probands and relatives), as compared with 6.0, 3.4 and 0.9% of asymptomatic carriers, respectively. The total follow-up time in relatives was 11 049 years. Prevalences of venous thromboembolism were 10.8% in single heterozygous factor V Leiden carrier relatives, 16.0% in double-heterozygotes for factor V Leiden and the prothrombin mutation, 36.8% in homozygotes for factor V Leiden, and 40.0% in double-heterozygotes for factor V Leiden and an inherited deficiency of protein C or protein S. Annual incidences in these groups were 0.39, 0.57, 1.41, and 4.76%, respectively. Multivariate analysis showed a small, non-significant additional effect of the prothrombin mutation on the risk of venous thromboembolism in heterozygotes for factor V Leiden [adjusted hazard ratio, 1.3; 95% confidence interval (CI), 0.5-3.8]. This effect was more pronounced for homozygosity of factor V Leiden (adjusted hazard ratio, 3.9; 95% CI, 1.7-9.0) and inherited protein C or protein S deficiencies (adjusted hazard ratio, 17.5; 95% CI, 3.8-81.2). Our data provide evidence of clustering of the evaluated genetic thrombophilic defects in symptomatic factor V Leiden carriers and support the assumption that the clinical expression of factor V Leiden depends on clustering in a part of carriers.

Impaired thrombin generation and fibrin clot formation in patients with dilutional coagulopathy during major surgery.

Patients subjected to haemodilution during surgery are at increased risk of bleeding. We hypothesised that, in the acquired dilutional coagulopathy, insufficient haemostasis is due to either insufficient thrombin generation or insufficient fibrin clot formation. In tissue factor-activated plasmas from patients with coagulation deficiency, we measured time curves of thrombin generation and fibrin clot formation (thromboelastography). Investigated were in study A: 10 patients treated with vitamin K antagonist and five healthy subjects; in study B: 30 patients undergoing cardiopulmonary bypass (CPB) surgery and infused with on average 2,000 ml crystalloids and colloids (no major bleeding); in study C: 58 patients undergoing major general surgery, and transfused with >5,000 ml crystalloids, colloids and red cell concentrates, who experienced major bleeding and were post-transfused with fresh frozen plasma. The treatment with vitamin K antagonist led to a progressive reduction in thrombin generation but not fibrin clot formation. In CPB patients, plasma factor levels post-surgery were 53-60% of normal. This was accompanied by moderate reduction in both haemostatic processes. In plasmas from patients undergoing major surgery, factor levels were 38-41% of normal, and these levels increased after plasma transfusion. Taking preset thresholds for normal thrombin generation and fibrin clot formation, at least one of these processes was low in 88-93% of the patients with (persistent) bleeding, but only in 40-53% of the patients without bleeding. In conclusion, the ability of thrombin generation and fibrin clot formation is independently reduced in acquired dilutional coagulopathy, while minimal levels of both are required for adequate haemostasis.

Myelodysplasia presenting with pulmonary manifestations associated with neutrophilic dermatosis.

We present the case of a patient who suffered from severe pneumonia. Extensive diagnostic procedures revealed negative cultures and a severe neutrophilia in the bronchoalveolar lavage fluid suggestive of Sweet's syndrome. The persistent toxic leukocyte differentiation (not induced by infection) suggested an underlying hematologic disorder. One week later, she developed cutaneous lesions indicative of neutrophilic dermatosis. Bone marrow investigation showed refractory anemia with excess of blasts (RAEB). Both the pulmonary and cutaneous infiltrates improved after treatment with chemotherapy.

Influence of aspirin on human megakaryocyte prostaglandin synthesis.

To evaluate whether currently popular aspirin regimens have an effect on the prostaglandin synthesis in human megakaryocytes we measured thromboxane B2 (TXB2) synthesis in response to thrombin stimulation in human megakaryocytes ex vivo. Human megakaryocytes were purified by Counterflow Centrifugal Elutriation from bone marrow punctures, taken from volunteers before and 2 hours after ingestion of one dose of 500 mg (n = 4), 80 mg (n = 4) or 40 mg (n = 2) aspirin. Subsequently, megakaryocytes were purified before and 12 h after ingestion of 80 mg (n = 3) aspirin twice daily for 1 week and 12 h after 500 mg (n = 3) aspirin. On average, 140 +/- 102 x 10(3) (mean +/- 1 SD) megakaryocytes were recovered. We found that aspirin inhibits megakaryocyte cyclooxygenase in a dose-dependent manner. Two hours after 500 mg of aspirin, TXB2 synthesis in megakaryocytes was inhibited by 96.8 +/- 2%, whereas one dose of 80 and 40 mg aspirin showed an inhibition of 79.4 +/- 13.7% and 80 +/- 6.2% respectively. However, the inhibition of TXB2 synthesis seems not to be long-lasting since, 12 h after the ingestion of aspirin, an increase of megakaryocyte TXB2 production could be observed which reached significance after the 500 mg aspirin dosage (p < 0.048). We conclude that human megakaryocyte cyclooxygenase is sensitive to aspirin inhibition and that low doses of aspirin (40 and 80 mg) enter the systemic circulation and are able to inhibit megakaryocyte cyclooxygenase, but this inhibition is incomplete and megakaryocyte cyclooxygenase seems to recover within 12 h after ingestion of aspirin.

Circulating human megakaryocytes in cardiac diseases.

Both bone marrow and circulating megakaryocytes may produce platelets. Changes in number and properties of the circulating megakaryocytes in cardiac diseases may provide information about thrombopoiesis in these disease states. Circulating megakaryocytes were obtained by retrograde aspiration through wedged pulmonary artery catheters from patients without cardiac abnormalities, patients with cardiac diseases and patients with disorders known to affect the megakaryocyte-platelet axis or marrow-blood barrier. Their number, ploidy distribution and expression of the beta 1-integrin adhesion antigens were estimated. We found evidence for an increased number of circulating megakaryocytes during acute myocardial infarction and in patients with triple vessel disease. A significantly higher proportion of circulating megakaryocytes obtained within 60 h after acute myocardial infarction, expressed CDw49b (alpha-chain of VLA2), CDw49e (alpha-chain of VLA5), and CDw49f (alpha-chain of VLA6) compared to megakaryocytes of patients without cardiac diseases. In myocardial infarction the increase in number of circulating megakaryocytes together with the higher proportion that expressed the beta 1-integrin adhesion antigens favour the hypothesis that these megakaryocytes are acutely released from the bone marrow. This could reflect either a selective response to an increased demand for circulating platelets or be the result of an a selective mechanism resulting from damage to the marrow-blood barrier following the acute infarction or thrombolytic therapy administered.

[Corneal opacity]. / Hornhauttrübung.

A case with bilateral corneal opacification in the form of a half-ring is described. Routine laboratory tests revealed no abnormalities, and serum lipoprotein composition was normal. The authors suspect that this was a case of primary lipoidal keratopathy.

Tissue factor activity in human monocytes is regulated by plasma: implications for the high and low responder phenomenon.

The 'high and low responder phenomenon' of monocyte tissue factor (MTF) activity has been attributed to effects on monocytes by granulocytes, platelets and lipopolysaccharide (LPS). To study the possible contribution of plasma to the high and low responder phenomenon, we measured the MTF activity in isolated cryopreserved human monocytes from two donors (monocytes A and monocytes B) after incubation in a plasma environment depleted of granulocytes, platelets and LPS. In buffer only, MTF activity was 643 and 679 fM (fM = final concentration of tissue factor), in normal pooled plasma, it was 1478 and 1615 fM (P = 0.001), respectively, in monocytes A and in monocytes B. Incubation with individual plasma samples from healthy controls (n = 43) gave a median MTF of 1355 fM (range 1044-1976 fM) and 1329 fM (range 858-1951 fM) respectively. A plasma consistently induced a higher or lower level of MTF activity in both monocytes: r = 0.82 (P < 0.00001). Coumarin use did not influence the high and low responder phenomenon. In the absence of granulocytes, platelets and LPS, plasma determines the high and low responder phenomenon. This phenomenon is not influenced by coumarin treatment.

Purification of human megakaryocytes using a glycoprotein Ib dependent agglutination technique.

In order to study normal and abnormal human megakaryocytopoiesis we modified a glycoprotein Ib dependent megakaryocyte purification technique originally described in guinea pigs. After Percoll density centrifugation, 2 ml bone marrow aspirates were subjected to 3 steps of bovine plasma induced cell clumping, which should selectively agglutinate glycoprotein Ib positive cells. In 13 experiments the resulting samples contained 28-102 x 10(3) megakaryocytes (median 70) with a median purity of 79%. Megakaryocyte recovery ranged from 30-86% (median 51%). The DNA profile showed median values of 10% for 8 N, 44% for 16 N and 43% for 32 N megakaryocytes. The described method is relatively simple, inexpensive and gives results comparable with more elaborate techniques.

Factor XIII Val34Leu and the risk of venous thromboembolism in factor V Leiden carriers.

A mutation in factor XIII (Val34Leu) was reported to protect against venous thromboembolism. We evaluated the effect of Val34Leu on thrombotic risk in 352 factor V Leiden carriers who were first-degree relatives of 132 thrombotic propositi carrying factor V Leiden. The total observation period was 2,594 years in 92 Val34Leu carriers and 7,444 years in 260 non-carriers. The annual incidence of a first episode of venous thromboembolism was 0.31% in Val34Leu carriers and 0.44% in non-carriers [relative risk (RR) for venous thromboembolism: 0.7, 95% CI 0.3-1.5]. Age-specific RR for venous thromboembolism were (for Val34Leu carriers and non-carriers respectively): 1.0 (95% CI 0.3-3.2) in the age group of 15-30 years, 0.4 (95%, CI 0.05-3.0) in the age group of 30-45 years, 0.6 (95% CI 0.1-2.9) in the group aged 45-60 years and 0.5 (95% CI 0.06-4.5) in relatives older than 60 years. In conclusion, the impact of FXIII Val34Leu on the venous thromboembolic risk is modest, suggesting that screening for this mutation in factor V Leiden carriers is not justified.

Combining counterflow centrifugal elutriation and glycoprotein Ib-dependent purification of human megakaryocytes: efficacy and selectivity.

To estimate the efficacy of human megakaryocyte purification techniques, mixtures of known numbers of megakaryocytes with a known ploidy range and of bone marrow or peripheral blood mononuclear cells were made. These artificial bone marrow samples were submitted to either a counterflow centrifugal elutriation or Percoll density separation followed by the glycoprotein Ib-dependent agglutination procedure. Also crude bone marrow samples were submitted to counterflow centrifugal elutriation directly followed by a glycoprotein Ib-dependent agglutination. The counterflow centrifugal elutriation resulted in a mean megakaryocyte recovery of 81% (mean 81% +/- 2.3). Purification by glycoprotein Ib-dependent agglutination after either a Percoll density separation or counterflow centrifugal elutriation resulted in a recovery of 61% (mean 61% +/- 15%) and 81% (mean 81% +/- 6) respectively. Purity of the resulting material was 87% (mean 87% +/- 11) and 83% (mean 83% +/- 5) respectively. The various isolation procedures did not affect the ploidy distribution of megakaryocytes greater than or equal to 8N. Counterflow centrifugal elutriation was preferred as the preparing step before glycoprotein Ib-dependent agglutination because of the lower variability in recovery and purity of megakaryocyte populations. When large numbers of rather pure and mature megakaryocytes are required, counterflow centrifugal elutriation followed by the glycoprotein Ib-dependent agglutination is a relatively simple method to purify human megakaryocytes without an appreciable loss in ploidy class greater than or equal to 8N.