Controlled study of heparin versus epoprostenol sodium (prostacyclin) as the sole anticoagulant for chronic hemodialysis.

We performed a controlled parallel study comparing the effects of heparin sodium to epoprostenol sodium (prostacyclin sodium, Flolan) during hemodialysis in 30 dialysis patients. Mean fiber bundle volume loss and dialyzer function were similar with both anticoagulation methods. Intradialytic symptoms occurred in 100% of the epoprostenol dialyses and 88% of the heparin dialyses, but only 10/325 epoprostenol and 3/374 heparin dialysis were discontinued prematurely because of symptoms. Long-term hemodialysis with epoprostenol is safe and effective. Epoprostenol may be a suitable alternative to heparin in some dialysis settings.     

Prostacyclin substitution for heparin in long-term hemodialysis

We studied prostacyclin as a substitute for heparin in 12 patients who underwent maintenance hemodialysis. All subjects underwent initial hemodialysis with prostacyclin as the sole anticoagulant; 10 of the 12 were restudied during heparin hemodialysis. Few adverse reactions occurred during prostacyclin hemodialysis in the 10 patients in whom dialysis was performed against a bicarbonate-containing dialysate; however, significant hypotension developed in two subjects when an acetate bath was used. Platelet aggregation progressively decreased during prostacyclin hemodialysis (p less than 0.02), but not during heparin hemodialysis, and returned toward control values after hemodialysis. Platelet thromboxane release decreased during both prostacyclin and heparin hemodialysis. Intradialytic percent decrements in serum urea nitrogen and creatinine were greater during prostacyclin than heparin administration (42 +/- 2.9 percent versus 36 +/- 2.6 percent [p less than 0.05] and 33 +/- 2.6 percent versus 29 +/- 2.1 percent [0.05 less than p less than 0.1], respectively). The plasma concentrations of 6-keto-prostaglandin-F1 alpha, a prostacyclin metabolite, reached peak levels by 120 minutes of hemodialysis and declined biexponentially toward predialysis concentrations during 120 minutes after hemodialysis, thereby providing an index of cumulative prostacyclin dosage. We conclude that prostacyclin is not only a safe alternative to heparin anticoagulation during hemodialysis, but that prostacyclin might also increase the efficiency of hemodialysis.     

Mechanisms of Recovery From Neutropenia Induced by Hemodialysis

Profound, transient neutropenia is observed early in hemodialysis. The presentstudies were performed during 11 twin-coil dialyses in eight patients withchronic renal failure to determine if therapid recovery reflects remobilization ofsequestered neutrophils and/or if marrow reserves are utilized. In five studies,in vitro cell tagging with diisopropylfluorophosphate-³²P (DF ³²P) and reinfusion were performed 4-6 hr beforeroutine hemodialysis. In all 11 studies amarked decrease of the absolute neutrophil count occurred within 15 min ofstarting blood return from the coil. Innine studies carried on over a sufficientperiod of time, a rebound of the neutrophil count to levels significantly higherthan the highest control counts occurredby 1-3 hr after starting dialysis. Inthe radioactive studies during the rebound phase, most or all of the expectednumber of labeled neutrophils returnedto the circulation, but simultaneously inall cases a large number of extra unlabeled neutrophils appeared in circulation. Since there are two sources foradditional circulating neutrophils—themarginated pool and the marrow reserves—and since the marginated poolunder the study conditions would beexpected to have the same specific radioactivity as the circulating pool, the dataare consistent with the hypothesis thatthere was an immediate marrow responseto neutropenia.

Submitted on March 24, 1971 Revised on May 25, 1971 Accepted on May 27, 1971     

Toxicity of Plasma from Hemodialysis Patients Treated with Heparin or Prostacyclin

Abstract Hemodialysis was performed in 6 uremic patients with either a bolus dose of 5 000 IU heparin or prostacyclin at a constant infusion rate of 5 ng/kg/min. Clinical data, plasma triglycerides (TG), free fatty acids (FFA), platelet aggregation, white blood cell count and plasma toxicity were measured prior to and during both procedures. No serious sideeffects were recorded. Heparin induced a fall in plasma TG, a rise in FFA and increased plasma toxicity. Prostacyclin infusion had no effect on these parameters. During both tests a marked drop in white blood cell count was found 15 min after the start of hemodialysis. During heparin dialysis no clotting was observed in the extracorporeal circuit. During prostacyclin dialysis clotting was observed at the venous line in three cases     

Recombinant human granulocyte colony-stimulating factor therapy for patients with neutropenia and/or neutrophil dysfunction secondary to glycogen storage disease type 1b

The purpose of this study was to evaluate the efficacy and toxicity of recombinant human granulocyte colony-stimulating factor (rhG-CSF) therapy in patients with neutropenia and/or neutrophil dysfunction secondary to glycogen storage disease (GSD) type 1b. Thirteen patients with neutropenia and/or neutrophil dysfunction secondary to GSD type 1b were treated with rhG-CSF. The effects of therapy on neutrophil numbers and in vitro neutrophil function and on bone marrow cellularity and morphology were studied. The clinical status of the patients and the occurrence of adverse events associated with rhG-CSF use were monitored. Use of rhG-CSF therapy was associated with a significant increase in circulating neutrophil numbers (P <. 01) and an improvement in neutrophil function as assessed in vitro. In addition, rhG-CSF therapy produced a significant increase in marrow cellularity and an increase in myeloid:erythroid (M:E) ratio, indicating stimulation of granulopoeisis. No adverse effects on marrow function were noted; in particular, no myelodysplasia or marrow exhaustion was seen. Use of rhG-CSF therapy was associated with objective and subjective improvements in infection-related morbidity. The therapy was well tolerated, although all patients developed splenomegaly, and 5 patients developed mild hypersplenism that did not require any specific treatment. rhG-CSF therapy is efficacious in the management of neutropenia and neutrophil dysfunction associated with GSD type 1b. Patients on this therapy need to be monitored for hypersplenism. Continued follow-up will be necessary to confirm long-term safety; however, no significant short-term toxicity was noted.     

Role of the adhesion molecule CD99 in platelet–neutrophil interactions

The interaction of platelets with neutrophils plays an important role in inflammation and thrombosis and is coordinated by multiple adhesive interactions. The adhesion molecule CD99 is a key mediator of neutrophil migration across the endothelium but whether it is involved in platelet–neutrophil adhesive interactions has not previously been addressed. We found that platelet CD99 is predominantly localized on the cell surface and is not shed following platelet activation. Blocking of either platelet or neutrophil CD99 significantly diminished neutrophil migration across surface‐adherent activated platelets in a quantitatively equivalent manner. In contrast, the blocking of CD99 affected neither neutrophil adhesion to surface‐adherent activated platelets nor formation of circulating platelet–neutrophil conjugates. Thus, homophilic CD99 interaction mediates neutrophil transplatelet migration but is not involved or is redundant in neutrophil adhesion to surface‐adherent or circulating platelets.     

Neutropenia during combination therapy of interferon alfa and ribavirin for chronic hepatitis C

Interferon therapy of hepatitis C causes a decrease in neutrophil counts, and neutropenia is a common reason for dose adjustment or early discontinuation. However, it is unclear whether neutropenia caused by interferon is associated with an increased rate of infection. In this study, we assessed factors associated and clinical consequences of neutropenia before and during interferon therapy of chronic hepatitis C. A total of 119 patients with chronic hepatitis C treated with the combination of interferon alfa and ribavirin were analyzed. In these studies, neutropenia was not used as an exclusion or dose modification criterion. In multivariate analysis, only black race was associated with baseline neutropenia. During treatment, neutrophil counts decreased by an average of 34%. Among 3 blacks with baseline neutropenia without cirrhosis or splenomegaly, there was little or no decrease in neutrophil counts (despite typical decreases in platelet and lymphocyte counts). Documented or suspected bacterial infections developed in 22 patients (18%), but in no patient with neutropenia. United States population estimates suggest that 76,000 blacks with hepatitis C have neutrophil counts below 1,500 cells/microL and might be denied therapy if this exclusion criterion was generally applied. In conclusion, neutropenia is frequent during treatment of hepatitis C with interferon and ribavirin, but it is not usually associated with infection. Constitutional neutropenia, which is common among blacks, should not exclude patients from therapy with interferon as these patients usually have minimal further decreases in neutrophil counts on therapy and are not excessively prone to bacterial infections.     

Experimental diabetes mellitus inhibits prostacyclin synthesis by the rat penis: pathological implications

Summary In view of the marked increase in blood flow into the penis during erection and the association of diabetes mellitus with impotence, we used the diabetic rat model to investigate the possibility that: (a) the penis may produce prostacyclin; and (b) prostacyclin secretion may be decreased in diabetes. Rats given a high dose of streptozotocin (120 mg/kg body weight) developed acute ketotic diabetes and were killed after 48 h. Animals given a low dose of streptozotocin (65 mg/ kg body weight) developed non-ketonuric diabetes and were killed after 7 or 62 days. Aortic rings and penile tissue discs were incubated in buffer, which was assayed for 6-oxo-pros-taglandin F₁, the stable and spontaneous breakdown product of prostacyclin. Penile tissue from control, ketotic and non-ketonuric (7 days) animals released similar quantities of prostacyclin, whereas that from long-term non-ketonuric animals (62 days) produced significantly less prostacyclin. Production of this prostanoid by the aortic rings paralleled these changes. We conclude that: (a) penile tissue releases prostacyclin in quantities comparable to those of the aorta; (b) long-term diabetes leads to diminished prostacyclin release by penile and aortic tissue: the former may contribute to the pathogenesis of diabetic impotence; and (c) since short-term ketotic diabetes does not inhibit aortic or penile prostacyclin release, duration of diabetes rather than its severity is responsible for diminished prostacyclin release.     

Outbreak of pyrogenic reactions at a dialysis center. Association with infusion of heparinized saline solution

Twenty-three pyrogenic reactions occurred in 16 patients undergoing hemodialysis at a private dialysis center in the south central United States between November 23 and December 2, 1978. No deaths were attributed to reactions; however, 10 patients were hospitalized for observation after experiencing a reaction. Cultures of all blood specimens obtained from the patients gave negative results. Chills (75 percent), nausea and/or vomiting (30 percent), and fever (90 percent) were the most common signs and symptoms, with mean times of onset after starting dialysis of 1.1, 1.6, and 3.6 hours, respectively. An epidemiologic and laboratory investigation documented that reactions occurred only in patients who had anticoagulation with a dilute solution of heparin. Analyses of heparinized saline solution used during the outbreak revealed a bacterial count of 7.4 X 10(5)/ml and a bacterial endotoxin level of 1,300 ng/ml. Acinetobacter calcoaceticus var. Iwoffi was isolated from the solution. Diluted heparin solution was prepared at the dialysis center by adding commercially supplied sodium heparin to 0.9 percent sodium chloride infusion fluid. Bacteria and endotoxin were not detected in vials of stock heparin and bags of unopened 0.9 percent sodium chloride infusion fluid. We concluded that contamination of the solution occurred at the dialysis center. After changes in the preparation and use of heparin were instituted on December 4, 1978, no pyrogenic reactions occurred in more than 400 subsequent dialyses.     

The effect of unfractionated and low-molecular-weight heparin on the release of prostacyclin from the arterial wall

Heparin is widely used as an antithrombotic agent, but one reported complication is thrombocytopenia associated with platelet aggregation. The mechanism is not fully clear but heparin interference in the prostaglandin production has been proposed. To investigate if heparin interacts with the production of prostacyclin from the vessel wall, and if low-molecular-weight heparin differs from unfractionated heparin in this respect, excised rabbit aortas were studied in a perfusion model. The vessels were perfused ex vivo for 5 x 15 min and in the last period arachidonic acid was added. Unfragmented heparin (500 IU/kg body weight) or low-molecular-weight heparin (LMWH) (500 antifactor Xa units/kg body weight) were given either 15 min before harvesting the vessels or added directly to the perfusate. The stable degradation product for prostacyclin, 6-keto-PGF1 alpha was not altered by addition of these agents. It is concluded that heparin and LMWH per se do not interact with the prostacyclin system in normal rabbit aortas in the doses studied.     

Delayed-onset neutropenia associated with rituximab therapy

The characteristics of severe neutropenia with a delayed onset following administration of rituximab have been evaluated in 53 consecutively treated patients. All but one patient received rituximab for the treatment of non-Hodgkin's lymphoma. Eight episodes of grade 4 neutropenia were detected between 1 and 5 months after rituximab, when administered alone on five occasions, and on three occasions in combination with chemotherapy, where neutrophil counts had recovered prior to the development of neutropenia. In three episodes, the patients presented with sepsis. Development of neutropenia did not correlate with either the presence of detectable disease or the administration of further treatment. Neutropenia was associated with selective depletion of neutrophil precursors in all but one episode, where it was associated with generalized bone marrow hypoplasia. All episodes developed after a period of either normal or mildly depressed neutrophil counts following treatment with rituximab, and persisted for between several days and several months, before undergoing spontaneous recovery in four instances, and after administration of filgrastim in the remainder. Episodes of neutropenia were associated with disordered immune status manifested by lymphopenia and hypogammaglobulinaemia, raising the possibility that either disturbance of the balance of lymphocyte subsets or an immune dyscrasia induced by rituximab resulted in the development of this type of neutropenia.     

Effects of density and of dehydration of sickle cells on their adhesion to cultured endothelial cells

Abnormal adhesion of sickle cells to vascular endothelium may be a factor in the initiation of painful vaso-occlusive crisis. The sickle cell population contains an unusually large number of less dense reticulocytes that are known to be more adhesive than mature red cells, but there is contradictory evidence regarding the adhesiveness of dense sickle cells. We used a flow-based assay of adhesion to cultured human umbilical vein endothelial cells to test the properties of density fractions of sickle cells, prepared either by density gradient or by centrifugation of packed cells. We also examined the effects of incubating sickle cells with or without cyclical deoxygenation on their adhesion. After fractionation on a Percoll-Isopaque gradient, the less dense 10% (reticulocyte-rich) cells and the most dense 10% cells adhered in greater number than the remainder (by about twofold). However, after centrifugation of packed cells, the less dense 10% were again more adhesive than the “middle” cells, but the most dense were not. Exposing sickle cells to constituents of the gradient had no consistent effect on adhesion, while centrifugal packing induced a degree of hemolysis, and tended to reduce adhesiveness of the dense fraction previously obtained from a gradient. Incubation in air at 37°C for 15 hr reduced the number of reticulocytes and the adhesiveness of less dense sickle cells compared to those held at 4°C. On the other hand, incubation at 37°C for 15 hr with cyclical deoxygenation caused formation of dense cells and increased adhesiveness compared to incubation without cyclical deoxygenation. We conclude that young, less dense sickle cells are unusually adhesive, but that this adhesiveness is reduced during maturation. However, repeated sickling in vivo causes formation of an abnormally dense subpopulation of cells which either redevelop an increased tendency to adhere to endothelial cells or preserve their initial adhesiveness. Both adhesive cell populations may be implicated in promoting vascular obstruction. © 1996 Wiley-Liss, Inc.     

Use of oral mucosal neutrophil counts to detect the onset and resolution of profound neutropenia following high-dose myelosuppressive chemotherapy

Severe neutropenia following cytotoxic, anti-cancer chemotherapy is well-known to be associated with an increased risk of infections that may be life-threatening, particularly if not treated immediately. Consequently, serial measurements of neutrophil counts in peripheral blood are done routinely following the administration of high-dose myelosuppressive chemotherapy in order to monitor the onset, severity, and duration of iatrogenic neutropenia. We have studied a non-invasive method of quantifying neutrophils recoverable from the oral mucosa, a normal tissue site of neutrophil turnover, as an alternative approach for monitoring severe, chemotherapy-induced neutropenia. This method is based on the quantification of fluorochrome-stained neutrophils present in timed mouthwash specimens. Blood neutrophil (ANC) and mucosal neutrophil counts (MNC) were measured repeatedly in 23 patients who had been treated with dose-intensive chemotherapy for a variety of indications. All 23 patients developed profound neutropenia (ANC < 100/mm3), and 19 developed neutropenic fever (>101 degrees F) during the 2 weeks following treatment. Nadirs of neutropenia defined by MNC were significantly less prolonged than those defined by the ANC. Furthermore, the onset and resolution of neutropenic fever coincided more precisely with nadirs of neutropenia defined by the MNC than with those defined by the ANC. Our findings indicate that oral mucosal neutrophil counts predict the timing of clinical events associated with neutropenia (e.g., the onset and resolution of fever) with significantly greater accuracy than blood neutrophil counts.     

Adult-onset cyclic neutropenia responsive to cyclosporine therapy in a patient with ankylosing spondylitis

A 45-year-old female with a long history of HLA-B27-positive ankylosing spondylitis and ulcerative colitis developed cyclic neutropenia. She was hospitalized for high fever during each of three consecutive episodes of absolute neutropenia. On the third hospitalization, granulocyte-colony-stimulating factor (G-CSF), 5 μg/kg/day, was given by subcutaneous injection and resulted in an increase of absolute neutrophil count from 0 to 2.2 × 10⁹/liter and an associated decrease of platelet count and hemoglobin as well as severe bone and joint pain predominantly in the middle and lower back and purulent diarrhea. The back pain necessitated discontinuation of the drug. Oral cyclosporine therapy was begun, and although the neutrophil count continued to oscillate, both the peaks and the nadirs were higher than previously, and symptoms of neutropenia subsided. We conclude that cyclosporine can be an effective treatment for cyclic neutropenia associated with autoimmunity since G-CSF may cause exacerbations of autoimmune disorders.     

Effects of prostacyclin infusion on platelets and hemodynamics in coronary bypass surgery

Twenty patients undergoing aorto-coronary bypass were randomly assigned to a prostacyclin treatment group or control group. Eight patients received 2 mg/kg heparin and 10 ng/kg/min prostacyclin before cardiopulmonary bypass (CPB) and 50 ng/kg/min during CPB. Twelve patients, serving as controls, received 3 mg/kg heparin. Heparinization resulted in a slight but significant increase of plasma beta-thromboglobulin in the control group but not in the prostacyclin group, and of plasma platelet factor 4 (PF-4) in both groups. After 90 minutes of CPB, beta-thromboglobulin was 408 (SD 128) ng/ml in the control group and 111 (SD 50) ng/ml (p less than 0.001) in the prostacyclin group. Platelet count, corrected for hemodilution, was 92% (SD 10) of the pre-CPB value after 10 minutes of CPB and 89% (SD 7) one hour after CPB in the control group, as compared to 113% (SD 10) and 145% (SD 18), respectively, in the prostacyclin group. Prostacyclin infusion before CPB reduced systemic vascular resistance to half of that of the control group, lowered mean arterial blood pressure, and increased cardiac index by 60% to 80%. An infusion of prostacyclin before CPB does not add to the already excellent platelet protective effect of 50 ng/kg/min prostacyclin during CPB, but may be used for vasodilation.     

In vitro analysis of pulmonary inflammation using rat lung organ cultures.

We have analyzed leukocyte to lung adhesive interactions and neutrophil-mediated lung injury using a rat lung organ culture system. Rat lung slices were maintained in tissue culture on gelatin sponges floating at the gas-liquid interface. Maintenance of three-dimensional alveolar structure, critical to the viability of lung tissue, was achieved by instilling 0.5% agarose (in 37 degrees C RPMI 1640) into the lungs during tissue explanation. Quantitative neutrophil to lung adhesive interactions were examined using an adaptation of the Woodruff-Stamper frozen section binding assay. Pretreatment of organ cultures with recombinant human tumor necrosis factor (rhTNF) resulted in a protein synthesis-dependent three- to fourfold increase in adhesiveness for neutrophils. Time course and mononuclear leukocyte binding experiments revealed that TNF-induced rat lung adhesiveness peaks at 4 h and is largely neutrophil-specific. Agonist-induced activation of neutrophils in the presence of [3H]leucine-labeled organ cultures resulted in lung injury as assessed by radioisotope release. These observations are consistent with endothelial cell culture data that indicate that TNF-induced endothelium exhibits a protein synthesis-dependent increase in adhesiveness for neutrophils. These data validate rat lung organ cultures as a model system that can be used to assess mechanisms of neutrophil adhesion and leukocyte-mediated tissue injury.     

Blood Platelet Economy during Moderate and Intensive Heparin Therapy

A study was made of the effects of heparin on platelet survival. The samesubjects were studied during control periods and during heparin treatment attwo levels of dosage. In most subjects, platelet survival was considerably prolonged and platelet turnover correspondingly diminished when 8000 units ofheparin were given 8-hourly, but not with smaller dosage. The evidence in noway implies that such a dosage is optimal—several patients showed little orno response to it. Comparable results were obtained with various in vitromeasures of blood coagulability. There was some evidence suggesting thatplatelet adhesiveness was paradoxically increased when the effect of heparinwas wearing off.

These findings raise questions of clinical importance. Moreover, they provideyet further evidence that the survival of the blood platelet is considerably influenced by factors in its external environment.

Submitted on October 20, 1962 Accepted on March 15, 1963     

Does heparin modify protease-antiprotease balance in acute experimental pancreatitis in rats

A rat model of taurocholate-induced acute pancreatitis has been employed to investigate the effect of heparin on the protease-antiprotease balance. Heparin was applied intraperitoneally at a dose of 6 mg/kg body weight during 24 hrs. At 24 and 48 hours of acute pancreatitis, heparin evidently diminished the consumption of trypsinogen in pancreatic tissue and decreased trypsin generation. The use of heparin prevented the consumption of alpha 1 anti-chymotrypsin, alpha 1-anti-trypsin and AT-III in pancreatic tissue, whereas in plasma the concentration of the mentioned inhibitors was restored or even increased. Heparin does not affect evidently lowered alpha 2-macroglobulin concentration, either in pancreatic tissue or in plasma. We conclude that heparin applied in acute pancreatitis markedly moderates the dysfunction of protease-antiprotease balance both in plasma and in pancreatic tissue.     

Continuation of deferiprone therapy in patients with mild neutropenia may not lead to a more severe drop in neutrophil count

Approximately 6% of patients with thalassemia receiving deferiprone develop neutropenia. Present practice is to monitor absolute neutrophil count (ANC) weekly and to interrupt treatment at the first sign of neutropenia, lest continuation lead to progressive neutrophil reduction. In a 6‐month study evaluating the safety and efficacy of a liquid form of deferiprone in 100 children, ANC was initially checked weekly for all patients. For individuals experiencing mild neutropenia, deferiprone was continued but monitoring was increased to daily until resolution. Therapy was to be suspended only if the episode was prolonged or if it worsened. Four patients experienced single episodes of mild neutropenia, and two others each experienced two episodes. All eight episodes resolved within 4–7 d despite continued therapy. (One patient later developed agranulocytosis and had treatment terminated.) This study showed that not all cases of mild neutropenia during deferiprone therapy develop into agranulocytosis, and suggests that many may not be caused by deferiprone. Transient declines in ANC to levels defined as neutropenic are common even in healthy individuals, particularly children; and it could be that the frequent monitoring of ANC mandated during deferiprone therapy may reveal cases of transient neutropenia that would otherwise have gone undetected and resolved on their own without clinical consequences. In patients with thalassemia, several factors increase the probability of a transient fall in ANC. These findings raise the question of whether deferiprone should be routinely stopped in cases of mild neutropenia, provided that such patients have their ANC monitored more frequently during the neutropenic episode.     

Antithrombin III supplementation reduces heparin requirement and platelet loss during hemodialysis of patients with fulminant hepatic failure

Previous studies have shown that antithrombin III levels are low in fulminant hepatic failure, and heparin kinetics are abnormal, making control of heparinization difficult during hemodialysis of these patients who are at risk of bleeding. In this study, we have performed a controlled, randomized trial of antithrombin III supplementation on heparin activity, occurrence of bleeding and the platelet count and activation during hemodialysis in 24 patients with fulminant hepatic failure. The treated group of 12 patients was given 3,000 units of antithrombin III before hemodialysis. Antithrombin III supplementation was shown to normalize antithrombin III levels during hemodialysis (prelevels: 0.22 +/- 0.03 U/ml S.E.; at 1 hr 0.99 +/- 0.06 U/ml; p less than 0.001; control prelevels: 0.24 +/- 0.03 U/ml; at 1 hr 0.23 +/- 0.04 U/ml). Total heparin usage was significantly decreased by antithrombin III supplementation (median 5,200 U; range = 2,000 to 13,000) as compared with the control group (median 10,200 U; range = 5,000 to 16,500; p less than 0.005). Blood heparin level (antifactor Xa activity) after the initial bolus was significantly greater in the antithrombin III-supplemented subjects (0.40 +/- 0.07 U/ml compared with 0.22 +/- 0.05 U/ml in the control group; p less than 0.05). The significant reduction in platelet count observed in the control patients (18% +/- 6% at 1 hr; p less than 0.05) did not occur in antithrombin III patients (6% +/- 4% at 1 hr), which was reflected by a lower release of the platelet-specific protein beta-thromboglobulin. Two of 12 patients in both groups showed minor bleeding around vascular access sites during the first hemodialysis.(ABSTRACT TRUNCATED AT 250 WORDS)