Citrate versus heparin anticoagulation in chronic haemodialysis patients

Transcription

1 Nephrol Dial Transplant (1993) 8: Original Article Nephrology Dialysis Transplantation Citrate versus heparin anticoagulation in chronic haemodialysis patients M. J. F. M. Janssen 1, P. C. Huijgens 2, A. A. Bouman 3, P. L. Oe 1, A. J. M. Donker 1 and J. van der Meulen 1 'Nephrology Unit, 2 Department of Haematology, 'Laboratory for Clinical Chemistry, Free University Hospital, Amsterdam, The Netherlands Abstract. Anticoagulation with citrate at a rate of 0.68 mm/min in combination with a calcium and magnesium-free dialysate and i.v. supplementation of calcium and magnesium at rates of 0.18 mm/min and 0.08 mm/min respectively, was compared with lowdose heparin. The heparin dose was a loading dose of 2500 IU and a sustaining infusion of IU/h; or a loading dose of 1250 IU and a sustaining infusion of IU/h until 1 h before the end of the dialysis if the patient was taking concomitantly coumarin anticoagulation for a Goretex shunt. Six chronic haemodialysis patients changed from heparin to citrate anticoagulation because they reported bleeding between dialyses. Heparin, after 2 h dialysis, induced a significant 10% prolongation of each patient's wholeblood activated clotting time (WBACT) as compared to the predialysis value; while the WBACT at the dialyser outlet was less than 3% prolonged as compared to the patient's WBACT. However, after 2 h citrate the patient's WBACT was not prolonged but the WBACT at the dialyser outlet was % longer, indicating a better anticoagulation of the extracorporeal system without systemic effects. With heparin the shunt pressure time (SPT), i.e. the time needed to stop bleeding from the puncture sites of the Goretex shunts, was 12 of 28 times 20 min or more. Citrate reduced these episodes by 75%. Thus citrate should be considered for chronic haemodialysis patients who are at risk of bleeding because of the concomitant use of anticoagulants. Other patients who could benefit from citrate are those with premorbid vascular abnormalities such as intestinal arteriovenous malformations, diabetic retinopathy, malignant hypertension or adult polycystic kidney disease. Claims that citrate gave improved biocompatibility, i.e. less leukopenia or thrombocytopenia, were not confirmed. Indications that citrate caused better dialysis efficiency were found, but should be confirmed in a greater number of patients. Key words: heparin; citrate; whole-blood activated Correspondence and offprint requests to: Dr MJFM Janssen, Nephrology Unit, Academic Hospital, Free University, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands clotting time; shunt pressure time; activated partial thromboplastin time Introduction Pinnick et al. advised the use of citrate in combination with a calcium-free dialysate (to prevent neutralization of citrate) and supplementation of calcium after the dialyser instead of heparin for acute renal failure patients with active bleeding [1]. They observed no systemic anticoagulation or other side-effects. However, a comparison between citrate and heparin anticoagulation was not made in these acute patients. During six dialyses in four chronic haemodialysis patients they noticed that the dialysis-induced thrombocytopenia and leukopenia, which were observed earlier with heparin, were less with citrate. Furthermore, during these six dialyses the dialyser clearance of creatinine remained stable with citrate, whereas it decreased with heparin. This decrease was ascribed to a less optimal anticoagulation by heparin, i.e. clotting of the dialyser. We reported that long-term use, i.e. at least 4 weeks, of citrate anticoagulation in chronic haemodialysis patients was hampered by complaints of paraesthesias and muscular cramps, possibly due to the combination of insufficient calcium supplementation and metabolic alkalosis [2]. When we reduced the buffer content of the dialysate and halved the infusion rate of citrate, no metabolic alkalosis occurred. To prevent clotting at such a low dose of citrate the dialysate was made not only free of calcium but also of magnesium. With this regimen, six chronic haemodialysis patients were dialysed for an average of 4 months without any sideeffects [3]. This report compares the effects of low-dose citrate anticoagulation with heparin in six chronic haemodialysis patients who started to bleed between two dialyses. In particular the effects on coagulation tests, leukopenia, thrombocytopenia, and dialyser efficiency are described European Dialysis and Transplant Association-European Renal Association

2 Citrate anticoagulation Subjects and methods Patients Six patients with chronic renal failure, four women and two men, age years (median 63 years), underwent haemodialysis for at least 18 months. Five patients were dialysed twice a week and one patient three times a week. Three patients used coumarin anticoagulants because of a Goretex shunt (Gore Inc., Flagstaff, USA). Two of the three developed subconjunctival bleeding, and the third one faecal blood loss due to diverticulitis. These bleeding episodes occurred when the prothrombin time was within the acceptable range. The three other patients did not use any anticoagulants apart from the heparin during dialysis. One patient had slight retinal bleeding, most probably due to high blood pressure. One patient with adult polycystic disease had recurrent haematuria, and one patient developed recurrent faecal blood loss due to intestinal arteriovenous malformations. During the dialysis that the patient reported the bleeding and during the two consecutive dialyses, data on anticoagulation, leukopenia, thrombocytopenia, and dialyser efficiency were collected. The data on body weight, SPT, clotting phenomena, and blood rests in the dialyser were retrieved from the dialysis charts of the last 4 weeks of heparin. After the patient was changed over to citrate anticoagulation, the same data were collected for 4 weeks. Anticoagulation Haemodialysis was performed using a Gambro AK 10 dialysis monitor with Travenol CF hollow cuprophane fibre dialysers ST 25. The blood flow was 200 ml/min and the dialysate flow 500 ml/min. Heparin anticoagulation was performed as follows: a loading dose of 2500 IU and a sustaining infusion of IU/h until the end of the dialysis when the patient was not taking coumarins. A loading dose of 1250 IU, and a sustaining infusion of IU/h until 1 h before the end of the dialysis was given when the patient was taking coumarins. The dialysate contained 1.50 mm calcium and 0.75 mm magnesium. Citrate anticoagulation was performed by continuous infusion of sterile 510 mm trisodium citrate in the arterial line at a rate of 0.68 mm/min. A calcium- and magnesium-free dialysate containing acetate 29.5 mm after 35-fold dilution was used to prevent neutralization of the citrate effect. To correct for the induced calcium and magnesium loss, a solution containing calcium and magnesium chloride was infused in the venous line at rates of 0.18 mm/min and 0.08 mm/min respectively. Coagulation studies Whole-blood activated clotting time (WBACT) was measured with a Hemochron coagulometer (International Technidyne Corp., New Jersey, USA); the normal range is s. Samples were taken before and after 2 h dialysis from the arterial line before the heparin c.q. citrate infusion (A) and at the dialyser outlet before the calcium and magnesium supplementation (O). Activated partial thromboplastin time (APTT) was measured pre- and postdialysis in samples taken from A. APTT was performed by incubating the sample with PTTa reagens (Boehringer, Mannheim, Germany) for 3 min at 37 C. The normal range, determined daily from pooled plasma, is s Prothrombin time (PT), measured with thromboplastin (Instrumentation Lab., Ascoli, Italy), was expressed as International Normalized Ratio (INR). PT was only tested predialysis in patients using coumarin anticoagulants. The desired INR to prevent thrombus formation in the Goretex graft is Shunt pressure time (SPT), measured by the nurses after each dialysis, was performed by applying gentle manual pressure at the puncture sites of the shunt. The sites were inspected every 5 min until bleeding had stopped. The SPT is routinely registered in all patients and recorded on the dialysis charts. Clotting phenomena such as a rise in venous pressure, fibrin formation in the drip chamber, unusual darkening of the blood, and collapse of the tubing when the blood flow rate was suddenly augmented, were registered on the same charts if encountered. Thrombus formation in the dialyser was estimated visually using an arbitrary three-grade scale (clean, moderate blood clots, full of blood clots) and also recorded on the dialysis charts. Other laboratory tests White blood cells (WBC) and platelets were counted with a Coulter counter in blood samples containing EDTA. The samples were taken at 0 and 20 min after the start of the dialysis. Serum creatinine, urea, calcium, and magnesium were measured by an automatic analyser. The values of ph and pco 2 were measured with a blood gas analyser (Ciba- Corning 228). Plasma bicarbonate (HCO 3 ) was calculated according to the standard method. Serum citrate concentrations were measured according to the Boehringer-Mannheim method [4]. For calculating the dialyser clearance blood samples were taken at the dialyser inlet after the citrate infusion (I) and at O. Dialyser clearance was calculated by using the formula [(Q Cot/Q] x blood flow, where C is solute concentration in plasma. The clearance was measured after 1 and 4 h of dialysis. The flux over the dialyser was calculated with the formula Q o x C o Q, x Q, where Q is the plasma flow rate. Statistics The non-parametric Wilcoxon test for paired data and the Mantel-Haenzel test were used when appropriate. A P-value less than 0.05 was accepted as significant. Results Coagulation studies Predialysis WBACT of the three patients taking coumarins, measured when the three heparin and three citrate dialysis sessions started, ranged from 166 to 200 s (median 175 s). Predialysis WBACT of the three patients not taking coumarins ranged from 119 to 146 s (median 135 s). After 2h dialysis with heparin; WBACT A of all six patients ranged from 171 to 212 s (median 175 s). These values were, as compared to the predialysis values, significantly (Wilcoxon P<0.05) increased by at least 10%. After 2 h dialysis with citrate the range was s (median 149 s), which was a non-significant increase of less than 3% when com-

3 1230 pared to the predialysis value. The WBACT O after 2 h heparin was 6 to 18s (median 15 s) longer than the WBACT A. After 2 h citrate the WBACT O was s (median 98 s) longer than the WBACT A measured at the same time. This was a significant (Wilcoxon P<0.05) difference. Predialysis APTT of all six patients during heparin and citrate anticoagulation ranged from 29 to 40 s (median 39 s) and from 25 to 40 s (median 38 s) respectively. When heparin was used, postdialysis APTT was significantly (Wilcoxon P<0.05) prolonged (range s, median 59 s). When citrate was used no prolongation of postdialysis APTT occurred. Predialysis PT of the three patients using coumarins at the dialysis where they reported the bleeding varied from 1.57 to 2.75 INR. SPT, measured after the 28 dialyses of the three patients taking coumarins, was shorter than (16 times) and equal to or longer than (12 times) 20 min when heparin was used. During the 28 citrate dialyses the SPT of the same patients was shorter than (25 times) and equal to (three times) 20 min. Citrate induced a significant decrease (Mantel-Haenzel P<0.05) in the number of prolonged SPT measurements. During 24 heparin dialyses of the three patients not taking coumarins, SPT was always shorter than 20 min (13 times 5 min only). During 24 citrate dialyses of the same three patients every SPT was again shorter than 20 min (15 times 5 min). When compared to heparin this was a non-significant difference. Clotting phenomena such as an increase in venous pressure, fibrin formation in the drip chamber, unusual darkening of the blood, and collapse of the tubing when the blood flow rate was suddenly augmented did not happen during the two observation periods. The frequency of thrombus formation in the dialyser did not differ between heparin and citrate. Moderate blood clots were noted in the dialyser seven times during heparin and five times during citrate. During the other sessions the dialysers were clean. Other laboratory tests During heparin anticoagulation, WBCs ranged from 3.5 to 14.2 x 10 9 /l (median 6.9 x 10 9 /l) and from 1.4 to 6.8 x 10 9 /l (median 1.6 x 10 9 /l) at 0 and 20 min respectively. During citrate WBCs ranged from 2.9 to 7.9 x 10 9 /l (median 6.5xlO 9 /l) and from 1.0 to 4.6 x 10 9 /l (median 1.7 x 10 9 /l) at 0 and 20 min respectively. Dialysis-induced leukopenia did not differ significantly between heparin and citrate. During heparin anticoagulation the platelets ranged from 148 to 211 x 10 9 /! (median 167 x 10 9 /l) and from 142 to 198 x 10 9 /l (median 170 x 10*/l) at 0 and 20 min respectively. During citrate the platelets ranged from 140 to 268 x 10 9 /l (median 179x 10 9 /l) and from 140 to 268 x lo^/l (median 162xlO 9 /l) at 0 and 20 min respectively. The number of platelets after 20 min did not differ significantly between heparin and citrate. The pre- and postdialysis creatinine, urea, calcium, magnesium, bicarbonate, citrate, and body weight M.J.F.M. Janssenera/. during heparin and citrate anticoagulation are shown in Table 1. The decrease in creatinine during heparin was of the same magnitude as during citrate. The median value of postdialysis creatinine, expressed as a percentage of the predialysis value, was 43% (range 33-50%) during heparin and 43% (range 33-51%) during citrate anticoagulation. The same applied for the decrease in urea during the two anticoagulants. However, this time postdialysis urea, expressed as a percentage of the predialysis value, was slightly but non-significantly greater during heparin (median 33%, range 29-40%) than during citrate (median 30%, range 23-41%). Postdialysis calcium was significantly (Wilcoxon / > <0.05) greater than the predialysis value during heparin, but not during citrate. During the two anticoagulants there was a similar non-significant increase in bicarbonate. Postdialysis citrate concentrations, measured during citrate anticoagulation only, were far less than toxic concentrations, i.e. 2.5 mmol/1. Pre- and postdialysis body weight did not differ significantly between heparin or citrate. The dialyser clearance of creatinine after 1 and 4 h heparin ranged from 128 to 142ml/min (median 134ml/min) and from 109 to 150ml/min (median 125 ml/min) respectively. During citrate the dialyser clearance of creatinine ranged from 119 to 151 ml/min (median 129 ml/min) and from 114 to 149 ml/min (median 139 ml/min) after 1 and 4h respectively. The median values of the creatinine clearance during heparin and citrate anticoagulation did not differ significantly, although during heparin the median value after 4 h was lower, whereas during citrate this value was higher than the respective 1 h values (Figure 1). The dialyser clearance of urea after 1 and 4 h heparin Table l.-pre- and postdialysis creatinine, urea, calcium, magnesium, citrate, and body weight during heparin and citrate anticoagulation Creat. (umol/1) Urea Calcium Magnesium Bicarbonate Citrate Weight (kg) Heparin Pre 1022 ( ) 28.3 ( ) 2.28 ( ) 1.10 ( ) 22.5 ( ) 65.4 ( ) Post 378* ( ) 7.6* ( ) 2.50* ( ) 1.13 ( ) 25.0 ( ) 64.0* ( ) Citrate Pre 891 ( ) 29.3 ( ) 2.35 ( ) 1.16 ( ) 22.8 ( ) 0.1 ( ) 65.6 ( ) Post 381* ( ) 8.9* ( ) 2.28 ( ) 1.13 ( ) 25.0 ( ) 0.5* ( ) 64.2* ( ) The biochemical results are the average of two dialysis sessions, the body weight is the average of 4 weeks dialysis. The results are expressed as median values of the averages with the range between brackets. * P < 0.05 as compared to predialysis value.

4 Citrate anticoagulation heparin citrate 1231 with citrate and the calcium- and magnesium-free dialysate were respectively from to mm/min (median mm/min) and from to mm/min (median mm/min). During heparin the magnesium flux over the dialyser ranged from to mm/min (median mm/min) and from to mm/min (median mm/min) after 1 and 4 h respectively with the 1.5 mm calcium- and 0.75 mm magnesiumcontaining dialysate. During citrate and the calciumand magnesium-free dialysate the ranges were from to mm/min (median mm/min) and from to mm/min (median mm/min) after 1 and 4h. Fig. 1. One- and 4-h dialysis creatinine clearance during hepann and citrate dialysis in six patients. Bold line: median values; pre and post denote 1- and 4-h values respectively. ranged from 163 to 188ml/min (median 176ml/min) and from 131 to 191 ml/min (median 167ml/min) respectively. During citrate the dialyser clearance of urea ranged from 127 to 188 ml/min (median 174 ml/min) and from 177 to 187 ml/min (median 180 ml/min) after 1 and 4 h respectively. These differences between 1 and 4 h dialysis were non-significant, but again during heparin the 4-h value was less and during citrate greater than the respective 1-h value (Figure 2). During heparin the calcium flux over the dialyser ranged from to mm/min (median mm/min) and from to mm/min (median mm/min) after 1 and 4 h respectively with the 1.5 mm calcium- and 0.75 mm magnesiumcontaining dialysate. The ranges of the calcium flux heparin citrate Fig. 2. One- and 4-h dialysis urea clearance during heparin and citrate dialysis in six patients. Bold line: median values; pre and post denote 1- and 4-h values respectively. Discussion By virtue of its chelation of ionized calcium, citrate is a powerful anticoagulant. It could be the ideal regional anticoagulant because its effect is easily reversed when calcium is added. Citrate is dialysable with an extraction coefficient of at least 60%. With our citrate infusion rate of 0.68 mm/min in the arterial line, the citrate load for the body is around 0.30 mm/min. If, due to unforeseen circumstances, citrate bypasses the dialyser, toxic accumulation in the body will not happen within 5 h because it is rapidly metabolized in the tricarboxylic acid energy pathway. We found that citrate at an infusion rate of 1.43 mm/min in the arterial line, i.e. a citrate load of 0,57 mm/min, did not result in citrate concentrations greater than 2.5 mm in patients during a 5 h dialysis [2]. Therefore the citrate infusion rate of 0.68 mm/min is safe. The fixed, low-dose citrate of 0.68 mm/min differed from the dose of Pinnick et al. They increased the dose from 0.50 to 1.00 mm/min to maintain the wholeblood clotting time over 20 min [1]. Our fixed, lowdose citrate achieved acceptable prolongation of the WBACT because the dialysate was not only devoid of calcium but also of magnesium. In a previous study we found that magnesium partially reversed citrate's anticoagulant effect because citrate chelates ionized magnesium as well [3]. With the 0.75 mm magnesiumcontaining dialysate the efflux of magnesium over the dialyser was approximately 0.02 mm/min compared to 0.11 mm/min with the magnesium-free dialysate, i.e. C o magnesium was approximately 0.70 mm less with a magnesium-free dialysate. Because the efflux over the dialyser concerns small molecules, this difference in C o magnesium could only be due to dialysable magnesium. Thus with a magnesium-free dialysate C o citrate could be approximately 0.47 mm (assuming that three magnesium molecules are binding two citrate molecules) less than with a 0.75 mm magnesium-containing dialysate. Since C o citrate was 40% of C, citrate, the latter value could be 1.18 mm less with a magnesium-free dialysate. Assuming Q, 70% of the blood flow, i.e. 140 ml/min, the citrate infusion rate could be 0.17 mm/min less. Thus a citrate infusion rate of 0.68 mm/min in combination with a magnesium-free

5 1232 dialysate corresponds with a citrate infusion rate of 0.85 mm/min in combination with a 0.75 mm magnesium-containing dialysate. The coagulation studies indicated that the combination of citrate at a rate of 0.68 mm/min and a calciumand magnesium-free dialysate achieves adequate regional anticoagulation. All six patients showed no significant prolongation of the WBACT in the arterial blood line after 2 h dialysis with citrate; whereas the WBACT became at least 10% prolonged after 2h dialysis with heparin. Moreover, the postdialysis APTT was significantly prolonged with heparin. As expected, the postdialysis APTT remained stable during citrate anticoagulation. The reduction of the long SPT in the three patients using coumarin anticoagulation because of the Goretex shunt was clinically relevant. The SPT was prolonged 40% of the time, even with the low-dose and limitedtime heparin anticoagulation, necessitating a longer period of nursing care after the dialysis. Citrate anticoagulation reduced these episodes by 75%. This finding is a confirmation of the randomized prospective study of low-dose heparin and citrate anticoagulation, in which it was reported that citrate reduced dialysisassociated bleeding from 50% to 18% [5]. Dialysis-induced thrombocytopenia and leukopenia are both considered to be bioincompatible events. Thrombocytopenia reflects thrombocyte activation, which may be related to increased thrombin formation [6]. Leukopenia is closely related to alternative pathway complement activation [7]. Thrombin formation and complement activation both need ionized calcium. Thus, theoretically, citrate anticoagulation could have a favourable influence on dialysis-induced thrombocytopenia and leukopenia. In our small series we could not find such an influence, which is in contrast with the finding of Pinnick et al. [1]. Whether this discrepancy could be explained by the fact that our fixed dose of citrate was roughly 0.15 mm/min lower than their maximum dose remains to be elucidated. On the other hand, with higher doses of citrate during long-term use in chronic renal failure, the patients experienced paraesthesias around the mouth and in hands and feet [2]. This could mean that citrate has a narrow margin between effective prevention of bioincompatible events and toxicity. Extracorporeal anticoagulation is needed to maintain maximum dialyser clearance throughout a dialysis session. Rough indicators of insufficient anticoagulation, i.e. clotting of the dialyser, are an increase in venous pressure, fibrin formation in the drip chamber, unusual darkening of the blood, and collapse of the tubing when the blood flow rate is suddenly augmented. These events did not happen during the two observation periods. Also after the sessions the dialysers did not show less thrombus formation when citrate was used. However, we could demonstrate that citrate caused a significantly more prolonged WBACT of the extracorporeal system. At 2 h the WBACTs measured at the dialyser outlet were 20-80% longer with citrate M J.F.M. Janssen et al. than with heparin in the three patients on coumarins; in the other three patients this difference was %, again in favour of citrate. These results demonstrate the more efficient extracorporeal anticoagulation with citrate. A more sensitive indicator of extracorporeal clotting is decreased dialyser efficiency, which is ascribed to a loss of effective dialyser membrane surface area. There exists a linear relationship between effective membrane surface area and clearance. To measure dialysis efficiency we collected pre- and postdialysis serum creatinine and urea, and we calculated 1- and 4-h dialyser clearances of creatinine and urea. The median values of postdialysis creatinine, expressed as a percentage of the predialysis values, were the same after heparin and citrate anticoagulation. However, the median value of postdialysis urea, expressed as a percentage of the predialysis value, was slightly but non-significantly lower after citrate anticoagulation. During heparin the median values of the creatinine and urea clearance after 4 h were lower, but non-significantly, than after 1 h, whereas during citrate both the 4-h values were higher, but again non-significantly, than the 1 -h values. Casino et al. recently reported that during the first 3 h of a heparin dialysis the mean urea clearance declined 15 ml/min in 20 patients [8]. Because of the magnitude of that decline, approximately 12 patients have to be studied before statistically significant differences will be reached between heparin and citrate. Thus, in spite of the differences we found, it is still too early to reject the hypothesis that citrate does not give a better dialyser efficiency. Postdialysis serum calcium was significantly greater than the predialysis value when heparin was used (Table 1). This phenomenon was not seen with citrate. The explanation was the difference in calcium flux over the dialyser. When heparin was used with a 1.5 mm calcium-containing dialysate, the median value of the calcium flux was mm/min after 1 h and mm/min after 4 h. The median values of the calcium flux with citrate and the calcium-free dialysate were 0.18 mm/min and mm/min respectively. From these findings one could conclude that the calcium supplementation should be slightly greater than the 0.18 mm/min we used. Postdialysis serum magnesium was not different whether citrate or heparin was used, i.e. the supplementation of magnesium at a rate of 0.08 mm/min during citrate is adequately compensating the effect of the 0.75 mm magnesium in the dialysate used during heparin. A calcium and magnesium-free dialysate allowed us to use a fixed, low-dose citrate anticoagulation in chronic haemodialysis patients. No signs of systemic anticoagulation with this regimen were found. In those patients using coumarins for Goretex shunts, the pressure time required to stop bleeding from the puncture sites were often prolonged when low-dose heparin was used. Citrate anticoagulation significantly reduced the pressure time in these patients. The anticoagulation of the extracorporeal system with citrate, as measured by the whole-blood activated clotting time, was signific-

6 Citrate anticoagulation antly better than with heparin. Therefore this regimen should be considered for the chronic haemodialysis patients who are at risk either of bleeding because of the concomitant use of anticoagulants, or because of premorbid vascular abnormalities such as intestinal arteriovenous malformations, diabetic retinopathy, malignant hypertension, or adult polycystic kidney disease. Claims of improved biocompatibility by citrate were not comfirmed. Indications that citrate causes better dialysis efficiency may be confirmed in studies comprising a greater number of patients. In view of the increasing use of low-molecular-weight heparin for haemodialysis patients, the safety and efficacy of this anticoagulant should also be compared with citrate [9,10]. Finally the use of the calcium- and magnesium-free dialysate would make it possible to manufacture single proportioning dialysate systems that deliver absolute acetatefree dialysate for bicarbonate dialysis. When such a dialysate is combined with citrate anticoagulation, the bicarbonate buffer content should be 25 mm to prevent metabolic alkalosis [3]. References 1. Pinnick RV, Wiegmann TB, Diederich DA. Regional citrate anticoagulation for hemodialysis in the patient at high risk for bleeding. N Engl J Med 1983; 308: Faber LM, de Vries PMJM, Oe PL, Van der Meulen J, Donker AJM. Citrate haemodialysis. Selh J Med 1990; 37: Van der Meulen J, Janssen MFJM, Langendijk PNJ, Bouman AA, Oe PL. Citrate anticoagulation and dialysate with reduced buffer content in chronic hemodialysis. Clin Nephrol 1992; 37: Toffegard Nielsen T. A method for enzymatic determination of citrate in serum and urine. Scand J Clin Lab Invest 1976; 36: Flanigan MJ, Von Brecht J, Freeman RM, Lim VS. Reducing the hemorrhagic complications of hemodialysis: a controlled comparison of low-dose heparin and citrate anticoagulation. Am J Kidney Dis 1987; 9: Sultan Y, London GM, Goldfarb B, Toulon P, Marchais S. Activation of platelets, coagulation and fibrinolysis in patients on long-term haemodialysis: influence of cuprophan and polyacrylonitrile membranes. Nephrol Dial Transplant 1990; 5: Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS. Complement and leukocyte-mediated pulmonary dysfunction in hemodialysis. N Engl J Med 1977; 296: Casino FG, Gaudiano V, Santarsia G el al. Reconciliation between urea kinetics and direct dialysis quantification. Nephrol Dial Transplant 1992; 7: Graue E. LMWH vs unfractioned heparin in chronic HD patients. Nephron 1992; 62: Ljunberg B. et al. Effective anticoagulation by a LMWH in HD with a highly permeable polysulfone membrane. Clin Nephrol 1992; 38 (2): Received for publication Accepted in revised form