- Email: [email protected]
Prevalence of preterminal pulmonary thromboembolism among patients on maintenance hemodialysis treatment before and after introduction of recombinant erythropoietin
Prevalence of Preterminal Pulmonary Thromboembolism Among Patients on Maintenance Hemodialysis Treatment Before and After Introduction of Recombinant Erythropoietin Martin Wiesholzer, MD, Melitta Kitzwo¨gerer, MD, Ferdinand Harm, MD, Gabriele Barbieri, MD, Anna-Christine Hauser, MD, Andreas Pribasnig, MD, Hans Bankl, MD, and Peter Balcke, MD ● The prevalence of pulmonary thromboembolism at autopsy was assessed in a retrospective study of a cohort of 185 patients undergoing maintenance hemodialysis treatment who died in the last decade. The overall frequency of thromboembolism was 12.43% in the dialysis population, which statistically was significantly less than in a control group of 8,051 nondialysis patients (21.77%; P ⴝ 0.0023). Moreover, pulmonary thromboembolism was less frequently fatal or contributing to death in the dialysis group than in the control group (P ⴝ 0.039). The prevalence of pulmonary thromboembolism in the dialysis group remained statistically unchanged over the 10-year period and was independent of a steady increase in the percentage of patients receiving recombinant erythropoietin therapy and the average hematocrit values. The occurrence of preterminal pulmonary thromboembolism was associated with a shorter period since onset of hemodialysis treatment and with infection as cause of death (P ⴝ 0.031; P ⴝ 0.029, respectively). No statistically significant influence of the type of basic renal disease, type of dialysis anticoagulation, or dialysis access could be found. Our data suggest that, at least in the preterminal stage, the introduction of recombinant erythropoietin within the last decade had no substantial influence on the prevalence of pulmonary thromboembolism. 娀 1999 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis; pulmonary thromboembolism; erythropoietin.
P
ULMONARY EMBOLISM is a common finding at autopsy. In a general hospital, moderate to severe thromboembolic findings can be detected at autopsy in up to approximately 25%, having been fatal or contributing to death in up to approximately 37% of these cases.1,2 Congenital or acquired blood coagulation disorders leading to thrombophilia, along with prolonged immobilization, may contribute to this high figure.3-7 Conversely, in the setting of abnormal hemostasis, such as uremic platelet dysfunction, thromboembolic complications would be expected to be decreased.8-12 Fatal pulmonary embolic events seem to be rare in the dialysis population according to various registries. However, no relevant data are available on the occurrence of pulmonary thromboembolism in the modern recombinant erythropoietin therapy era.13,14 In a retrospective study, we assessed the preva-
From the Department of Internal Medicine and Ludwig Boltzmann Institute of Nephrology; and the Institute of Clinical Pathology, St Poelten, Austria. Received May 27, 1998; accepted in revised form October 16, 1998. Address reprint requests to Peter Balcke, MD, First Medical Department and Ludwig Boltzmann Institute of Nephrology, General Hospital A-3100, St Poelten, Austria. E-mail: [email protected]
娀 1999 by the National Kidney Foundation, Inc. 0272-6386/99/3304-0011$3.00/0 702
lence of pulmonary embolism at autopsy in a cohort of central dialysis unit patients within the last decade, during which time recombinant erythropoietin has been introduced, and tried to point out factors that could have contributed to its occurrence. METHODS
Study Patients In a 10-year period from 1987 to 1996, 185 chronic hemodialysis patients without a history of peritoneal dialysis or routine anticoagulation therapy for more than 3 months during the dialysis period died at the hospital, all of whom were autopsied. The mean age at death was 69.5 ⫾ 10.59 (standard deviation) years. Dialysis was performed three times weekly for 4 hours per session. Anticoagulation was performed with unfractionated heparin (Heparin Novo; NovoNordisk, Bagsvaerd, Denmark) or low-molecular-weight heparin (Deltaparin, Fragmin; Kabi Pharmacia, Uppsala, Sweden). The mean period since the start of hemodialysis treatment was 142.60 ⫾ 150.2 weeks. The types of basic renal disease were chronic glomerulonephritis in 42 patients, diabetic nephropathy in 42 patients, renal vascular disease in 39 patients, chronic pyelonephritis in 16 patients, druginduced nephropathy in 23 patients, hereditary nephropathy in 6 patients, and uncertain cause in 17 patients. The number of patients receiving chronic hemodialysis during the period of observation was 360. The mean annual death rate was 22.30% ⫾ 1.22%. In the same 10-year period, 8,660 patients died of nonrenal causes at the general hospital, which has divisions of internal medicine, surgery, gynecology, opthalmology, pediatrics, and neurology. Of these, 8,051 patients (92.96%) were autopsied. The mean age was 72.3 ⫾ 2.5 years.
American Journal of Kidney Diseases, Vol 33, No 4 (April), 1999: pp 702-708
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
Autopsies were performed by pathologists supervised by the same leading staff members. Pulmonary embolism was considered fatal when at least one of these criteria was found: (1) massive thrombi occluding the main pulmonary arteries, (2) multiple thrombi occluding more than 80% of the peripheral area of the pulmonary arterial bed, (3) pulmonary emboli in the absence of any other cause of death, or (4) pulmonary infarction affecting more than 80% of the pulmonary circulation. Pulmonary embolism was considered to contribute to death when, in a patient with severe or critical disease, pulmonary embolism was suspected as a relevant cofactor for death. Because it is difficult in practice to distinguish between fatal or contributing pulmonary embolisms in multimorbid patients, both are considered an equivalent occurrence in the study.
Statistical Analyses Data are expressed as mean ⫾ standard deviation. For comparisons of groups, the Mann-Whitney U test or the chi-squared test were performed. For correlation testings, Spearman’s rank correlation analysis was used. The level of statistical significance was defined as P less than 0.05.
RESULTS
From 1987 to 1996, pulmonary thromboembolism was detected in 12.43% (23 of 185) of the autopsies of chronic hemodialysis patients. This percentage is statistically significantly less than that found at 8,051 autopsies of nondialysis patients (21.77%; P ⫽ 0.0023). In the cases with pulmonary thromboembolism, embolism was estimated to have been fatal or to have contributed to death in 71.7% of the nonrenal cases and 52.17% of the hemodialysis patients (P ⫽ 0.039; Fig 1). Thus, pulmonary thromboembolism was fatal or contributing to death in 6.5% of the hemodialysis patients. In none of these patients had pulmonary thromboembolism been diagnosed premortally. The annual thromboembolism prevalence in the dialysis population remained statistically unchanged during the last 10-year observation period. It was 23.1% in 1987, 7.1% in 1988, 19.0% in 1989, 6.3% in 1990, 8.0% in 1991, 9.5% in 1992, 13.6% in 1993, 20.0% in 1994, 11.1% in 1995, and 6.7% in 1996. During this 10-year period, recombinant erythropoietin was administered to an increasing number and percentage of patients beginning in 1988. In 1996, 90% of the patients were receiving erythropoietin therapy. Simultaneously, the mean hematocrit values of all patients increased statistically significantly within these years from 25.5% ⫾ 5.0% in 1985
703
Fig 1. Percent of pulmonary thromboembolism at autopsy in 185 chronic hemodialysis patients (DP) and 8,051 nonrenal patients (controls). The prevalence of pulmonary thromboembolism (fatal or contributing to death and incidental) was statistically significantly less in the dialysis group than controls (P ⴝ 0.0023). Pulmonary thromboembolism was more frequently fatal or contributing to death in the controls than in the dialysis group (P ⴝ 0.039). (N), pulmonary thromboembolism fatal or contributing to death; (䊐), incidental pulmonary thromboembolism.
to 31.6% ⫾ 8.1% in 1996 (P ⫽ 0.0054). No statistically significant correlation between the prevalence rate of thromboembolism and the percentage of patients receiving recombinant erythropoietin or the hematocrit value could be noted (P ⫽ 0.624 and P ⫽ 0.273, respectively; Fig 2). Mean platelet values showed no statistically significant variation over these years. Anticoagulation during hemodialyses was increasingly performed with low-molecular-weight heparin (in 37.88% of the patients in 1996), whereas the percentage of anticoagulation performed with unfractionated heparin decreased. The prescription of salicylates increased continuously to 28.57% in 1996. Other potentially relevant parameters are listed in Table 1. In a further analysis, data obtained less than 2 weeks before death for chronic hemodialysis patients with and without thromboembolism were compared. Pulmonary thromboembolism was associated with a statistically significantly greater number of deaths from infection and a shorter period since onset of hemodialysis treatment. No statistically significant difference between groups could be found regarding age, sex, basic renal disease, current hematocrit values, platelet count, blood urea nitrogen levels, type of dialysis anticoagulant, administration of salicylates, days of
704
WIESHOLZER ET AL
Fig 2. Percentage of patients with pulmonary thromboembolism at autopsy (bars) and percent hemodialysis unit patients receiving recombinant erythropoietin therapy (line) from 1987 to 1996. The parameters have no statistically significant relationship (P ⴝ 0.624).
immobilization, prevalence of diabetes, and type of dialysis access (Table 2). DISCUSSION
Our study shows that pulmonary thromboembolism is a frequent occurrence in preterminal chronic hemodialysis patients. For the assessment of thromboembolism frequency, autopsy findings were chosen because conventional clinical methods of detection are far less reliable.2,15 Although pulmonary emboli could be detected in a substantial percentage of hemodialysis patients, the frequency was statistically significantly less and the fatality rate from thromboembolism was less than those in nondialysis patients autopsied during the observation period at the same general hospital and by the identical leading medical staff. In our opinion, the discrepancy between both groups could be caused by various factors. Thromboembolic risk may be greater in nondialysis patients as a consequence of different basic or concomitant diseases or therapies triggering thromboembolism, such as malignancy or major surgery.16 Conversely, thromboembolism risk might be reduced in the dialysis population by the abnormal hemostasis encountered in chronic hemodialysis patients, which results principally from perturbed platelet adhesion and aggregation.8-12 In addition to the abnormal hemostasis, anticoagulation with unfractionated or lowmolecular-weight heparin during dialysis procedures could also contribute to this discrepancy.
The prevalence of pulmonary thromboembolism seems to be stable during the last decade. This is amazing because so many modifications in hemodialysis therapy have occurred within the last 10 years; most importantly, therapy with recombinant erythropoietin.17 Erythropoietin therapy leads to a reduction of the prolonged bleeding time through the increase in erythrocyte counts, thereby modifying rheological factors, as well as through an improved platelet function by increasing the number of glycoprotein IIb-IIIa molecules on the platelet membrane.18,19 As a further potentially thrombosis-promoting factor by recombinant erythropoietin, an enhanced release of plasminogen activator inhibitor-1 by endothelial cells was described recently.20 However, although a similar influence on hemostasis can be assumed in our patients treated with recombinant erythropoietin, no increase in the prevalence of pulmonary thromboembolism could be noted despite a steady increase in the percentage of patients receiving erythropoietin therapy and the increment in the average hematocrit value of the patients within the last decade. Two mechanisms could be responsible. First, the improvement of anemia leading to target hematocrit values in our patients of between 30% and 35% is not sufficient to correct the abnormal hemostasis in uremia relevantly to the thrombosis risk. Second, anticoagulation performed with unfractionated or low-molecular-weight heparin during the thrice-weekly dialysis procedure has a
Pulmonary embolism at autopsy Patients with erythropoietin therapy Hematocrit (%) BUN (mg/dL) Platelets/µL Unfractionated heparin (%) Low-molecular-weight heparin (%) Patients with dialysis catheters (%) Patients with diabetic nephropathy (%) Patients with salicylates (%) Age (yr)
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
3/13 (23.1)
1/14 (7.1)
4/21 (19)
1/16 (6.3)
2/25 (8)
2/21 (9.5)
3/22 (13.6)
4/20 (20)
2/18 (11.1)
1/15 (6.7)
P 0.62
0 (0)
4 (4)
14 (14)
29 (23)
27 (26)
41 (40)
55 (53)
60 (61)
53 (62)
63 (90)
0.0027
25.5 ⫾ 5 79.9 ⫾ 18.1 241,329 ⫾ 70,093 100
26.6 ⫾ 4.9 81.9 ⫾ 21.2 252,444 ⫾ 81,746 100
26.1 ⫾ 4.4 80.3 ⫾ 23.1 231,896 ⫾ 84,166 100
27.7 ⫾ 4.7 85.7 ⫾ 25 218,707 ⫾ 59,828 100
27.4 ⫾ 3.7 75.5 ⫾ 18.6 241,985 ⫾ 79,938 91.87
28.7 ⫾ 3.9 79.5 ⫾ 25.2 246,250 ⫾ 89,963 37.88
28.6 ⫾ 3.5 81.7 ⫾ 22 250,500 ⫾ 69,843 38.16
28 ⫾ 2.5 83 ⫾ 20.9 228,366 ⫾ 65,557 25.33
30.3 ⫾ 2.9 75.1 ⫾ 18.2 224,500 ⫾ 74,295 31.43
31.6 ⫾ 8.1 77 ⫾ 22.1 236,027 ⫾ 69,450 62.12
0.0054 0.283 0.413
0
0
5.32
7.69
10.68
21.15
1.23 62.9 ⫾ 14.5
23.4
0
0
0.014
8.33
62.12
61.84
74.67
68.57
37.88
0.014
7.81
11.76
9.71
8.65
10.2
15.29
20
0.023
18.45
17.97
9.8
20.39
21.15
21.43
22.35
31.43
0.274
2.35
1.15
4.04
9.72
6.8
9.62
12.24
21.18
28.57
0.0054
58.5 ⫾ 15.8
59.5 ⫾ 16.9
60.6 ⫾ 16.1
60.5 ⫾ 15.4
60.9 ⫾ 15.3
60.8 ⫾ 16
62.2 ⫾ 16.4
63 ⫾ 15.4
62.9 ⫾ 14.5
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
Table 1. Changes in the Prevalence of Pulmonary Embolism at Autopsy and in Various Parameters Potentially Relevant to Pulmonary Thromboembolism in All Chronic Hemodialysis Unit Patients From 1987 to 1996
0.0865
NOTE. Values expressed as number (%) or mean ⫾ standard deviation. Abbreviation: BUN, blood urea nitrogen.
705
706
WIESHOLZER ET AL Table 2. Premortal Parameters of Patients With and Without Pulmonary Thromboembolism at Autopsy Patients With PE
Patients Age (yr) Sex Men Women Basic renal disease Glomerulonephritis Pyelonephritis Drug-associated nephropathy Hereditary nephropathy Renal vascular disease Diabetes Other Cause of death Fatal pulmonary thromboembolism Cardiac Vascular Infection Liver Miscellaneous Other Erythropoietin therapy Erythropoietin dose/wk (units) Hematocrit value (%) Platelets/µL BUN (mg/dL) Unfractionated heparin Low-molecular-weight heparin Salicylates Days of immobilization Diabetics Type of dialysis access Arteriovenous fistula Permcath catheter Femoral vein cannula Subclavian or jugular vein cannula GoreTex shunt Arteriovenous graft Weeks since start of hemodialysis
Without PE
P
23 72.4 ⫾ 7.7
162 69.1 ⫾ 10.9
9 (39.1) 14 (60.9)
98 (60.5) 64 (39.5)
7 (30.4) 2 (8.7) 1 (4.3) 1 (4.3) 3 (13) 7 (30.4) 2 (8.7)
35 (21.6) 14 (8.6) 22 (13.6) 5 (3.1) 36 (22.2) 35 (21.6) 15 (9.3)
0.344 0.993 0.209 0.749 0.313 0.344 0.93
8 (34.8) 3 (13.0) 2 (7.8) 9 (39.1) 0 (0) 1 (4.3) 0 (0) 8 (34.8) 10,000 ⫾ 4,751 27.3 ⫾ 4.5 232,496 ⫾ 124,336 80.54 ⫾ 29.3 18 (78.3) 5 (21.7) 3 (13) 19 ⫾ 22.7 8 (34.8)
0 (0) 85 (52.5) 23 (14.2) 31 (19.1) 1 (0.6) 19 (11.7) 3 (1.9) 55 (34) 10,581 ⫾ 6,559 27.1 ⫾ 4.9 244,188 ⫾ 103,047 73.89 ⫾ 21.5 132 (81.5) 30 (18.5) 26 (16) 13.1 ⫾ 21.8 61 (37.7)
0.0004 0.47 0.029 0.705 0.286 0.51 0.985 0.83 0.73 0.42 0.44 0.92 0.9 0.71 0.11 0.757
13 (56.5) 4 (17.4) 2 (8.7) 1 (4.3) 1 (4.3) 2 (8.7) 103.3 ⫾ 107.2
78 (48.2) 40 (24.7) 7 (4.3) 20 (12.3) 14 (8.6) 3 (1.9) 148.21 ⫾ 155
0.23 0.052
0.452 0.441 0.361 0.257 0.48 0.058 0.031
NOTE. Values expressed as number (%) or mean ⫾ standard deviation. P refers to numbers. Abbreviations: PE, pulmonary thromboembolism; BUN, blood urea nitrogen.
substantial protective effect on thromboembolism by mitigating any potential thrombosispromoting activity of recombinant erythropoietin. In the last years, a protective effect of anticoagulants could theoretically have become more relevant by the more frequent use of lowmolecular-weight heparin instead of unfractionated heparin for hemodialysis and salicylates prescribed for various causes. Low-molecularweight heparin has a longer lasting anticoagulant
activity than unfractionated heparin and could thus have had a more protective influence,21,22 and salicylates administered for cardiovascular or cerebrovascular therapy reasons have a prolonged efficiency in renal failure.23 However, our statistical analyses of parameters at the premortal period do not suggest this explanation, that a potentially greater thromboembolic risk by recombinant erythropoietin therapy could have been compensated by such modifica-
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
tions, because the thromboembolism and nonthromboembolism groups did not differ statistically significantly regarding the type of heparin used during the dialysis session or the prevalence of salicylate use. The nondifferent prevalence of erythropoietin therapy and hematocrit values in the premortal period underline additionally that erythropoietin therapy does not have a major role for thromboembolism. We believe that abnormal hemostasis in the dialysis population is such a prevailing factor; thus, the effect of erythropoietin therapy is not sufficient to increase the thromboembolic risk substantially. The analyses of premortal parameters did not show statistically significant influences of the prevalence of diabetes, blood urea nitrogen levels and platelet counts, days of immobilization, and type of dialysis access. However, thromboembolism was associated with a greater prevalence of death from infection and fewer weeks of dialysis; the latter finding most likely caused by more severe morbidity. Moreover, the thromboembolic risk may be influenced by prethrombotic disorders that were not assessed in our study, such as the inherited resistance to activated protein C as a consequence of factor V Leiden mutation; congenital deficiencies of antithrombin III, protein C, protein S, or plasminogen; and hyperhomocysteinemia, which as genetic disorders can be assumed to have remained unchanged within the last decade, as can be considered for the prevalence of antiphospholipid antibodies.3-7 We believe that the factor V Leiden mutation deserves the most interest in this context because its prevalence in the general population ranges from 2% to 15%, and its impact for thromboembolism is markedly increased when hyperhomocysteinemia coexists, as in the setting of chronic hemodialysis treatment, in which a slight increase in blood homocysteine levels has been described for most patients.24-27 Consequently, these factors could have an important role for thromboembolism in various conditions, including the state of severe infection. The pulmonary embolisms found at autopsy were not diagnosed antemortem in one single case, showing the difficulty of detecting pulmonary emboli clinically in principle. This is consistent with various studies that showed antemor-
707
tem diagnoses of pulmonary embolism were not made in up to 70% of autopsy-proven events in general.1,2 Although pulmonary thromboembolism was fatal or contributing to death in only a small percentage of patients, nonfatal events may also be relevant regarding such consecutive complications as pneumonia, for which as a secondary event, venous thromboembolism could be an underestimated cause in this population. However, regarding the preterminal prevalence of pulmonary thromboembolism, our data suggest that the recombinant erythropoietin therapy introduced within the last decade did not result in a substantial increase in the rate of thromboembolic events in the hemodialysis population. REFERENCES 1. Lindblad B, Sternby NH, Bergqvist D: Incidence of venous thromboembolism verified by necropsy over 30 years. BMJ 302:709-711, 1991 2. Stein PD, Henry JW: Prevalence of acute pulmonary embolism among patients in a general hospital and at autopsy. Chest 108:978-981, 1995 3. Diuguid DL: Oral anticoagulant therapy for venous thromboembolism. N Engl J Med 336:433-434, 1997 4. Ginsberg JS: Drug therapy: Management of venous thromboembolism. N Engl J Med 335:1816-1828, 1997 5. Heijboer H, Brandjes DPM, Bu¨ller HR, Sturk A, ten Cate JW: Deficiencies of coagulation-inhibiting and fibrinolytic proteins in outpatients with deep-vein thrombosis. N Engl J Med 323:1512-1516, 1990 6. Dahlback B: Inherited thrombophilia: Resistance to activated protein C as a pathogenetic factor of venous thromboembolism. Blood 85:607-614, 1995 7. Ginsberg JS, Wells PS, Brill-Edwards P, Donovan D, Moffatt K, Johnston M, Stevens P, Hirsh J: Antiphospholipid antibodies and venous thromboembolism. Blood 86:36853691, 1995 8. Eberst ME, Berkowitz LR: Hemostasis in renal disease: Pathophysiology and management. Am J Med 96:168178, 1994 9. Rabelink TJ, Zwaginga JJ, Koomans HA, Sixma JJ: Thrombosis and hemostasis in renal disease. Kidney Int 46:287-293, 1994 10. Gawaz MP, Dobos G, Spath M: Impaired function of platelet membrane glycoprotein IIb-IIIa in end-stage renal disease. J Am Soc Nephrol 5:36-46, 1994 11. Heistinger M, Stockenhuber F, Schneider B, Pabinger I, Brenner B, Wagner B, Balcke P, Lechner K, Kyrle PA: Effect of conjugated estrogens on platelet function and prostacyclin generation in chronic renal failure. Kidney Int 38:1181-1187, 1990 12. Macconi D, Vigano G, Bisogno G, Galbusera M, Orisio S, Remuzzi G, Livio M: Defective platelet aggregation in response to platelet-activating factor in uremia asso-
708
ciated with low platelet thromboxane A2 generation. Am J Kidney Dis 19:318-325, 1992 13. Bloembergen WE, Port FK, Mauger EA, Wolfe RA: Causes of death in dialysis patients: Racial and gender differences. J Am Soc Nephrol 5:1231-1242, 1994 14. Mailloux LU, Bellucci AG, Wilkes BM, Napolitano B, Mossey RT, Lesser M, Bluestone PA: Mortality in dialysis patients: Analysis of the causes of death. Am J Kidney Dis 18:326-335, 1991 15. The PIOPED Investigators: Value of the ventilation/ perfusion scan in acute pulmonary embolism diagnosis. JAMA 363:2753-2759, 1990 16. Giuntini C, Di Ricco G, Marini C, Melillo E, Palla A: Pulmonary embolism: Epidemiology. Chest 107:3S-9S, 1995 (suppl) 17. Eschbach JW, Egrie JC, Downing MR, Browne JK, Adamson JW: Correction of the anemia of end-stage renal disease with recombinant human erythropoietin. Results of a combined phase I and II clinical trial. N Engl J Med 316:73-78, 1987 18. Cases A, Escolar G, Reverter JC, Ordinas A, LopezPedret J, Revert L, Castillo R: Recombinant human erythropoietin treatment improves platelet function in uremic patients. Kidney Int 42:668-672, 1992 19. Livio M, Gotti E, Marchesi D: Uremic bleeding: Role of anemia and beneficial effect of red cell transfusion. Lancet 2:1013-1017, 1982 20. Nagai T, Akizawa T, Kohjiro S, Koiwa F, Nabeshima K, Niikura K, Kino K, Kanonori E, Kinugasa E, Ideura T: RHuEPO enhances the production of plasminogen activator
WIESHOLZER ET AL
inhibitor-1 in cultured endothelial cells. Kidney Int 50:102107, 1996 21. Lockner D, Bratt G, To¨rnebohm E, Aberg W: Pharmacokinetics of intravenously and subcutaneously administered Fragmin in healthy volunteers. Hemostasis 16:8-10, 1986 22. Follea G, Laville M, Pozet N, Dechavanne M: Pharmacokinetic studies of standard heparin and low-molecularweight heparin in patients with chronic renal failure. Haemostasis 16:147-151, 1986 23. Williams ME, Weinblatt M, Rosa RM, Griffin VL, Goldlust B, Shang SF, Harrison LI, Brown RS: Salicylate kinetics in patients with chronic renal failure undergoing hemodialysis. Clin Pharmacol Ther 39:420-428, 1986 24. Beauchamp NJ, Daly ME, Hampton KK, Cooper PC, Preston FE, Peake IR: High prevalence of a mutation in the factor V gene within the UK population: Relationship to activated protein C resistance and familial thrombosis. Br J Haematol 88:219-222, 1994 25. Dahlback B, Hillarp A, Rosen S, Zoller B: Resistance to activated protein C, the FV:Q506 allele, and venous thrombosis. Ann Hematol 72:166-176, 1996 26. Mandel H, Brenner B, Berant M, Rosenberg N, Lanir N, Jakobs C, Fowler B, Seligsohn U: Coexistence of hereditary homocystinuria and factor V Leiden–Effect on thrombosis. N Engl J Med 334:763-768, 1996 27. Bachmann J, Tepel M, Raidt H, Riezler R, Graefe U, Langer K, Zidek W: Hyperhomocysteinemia and the risk for vascular disease in hemodialysis patients. J Am Soc Nephrol 6:121-125, 1995
P
ULMONARY EMBOLISM is a common finding at autopsy. In a general hospital, moderate to severe thromboembolic findings can be detected at autopsy in up to approximately 25%, having been fatal or contributing to death in up to approximately 37% of these cases.1,2 Congenital or acquired blood coagulation disorders leading to thrombophilia, along with prolonged immobilization, may contribute to this high figure.3-7 Conversely, in the setting of abnormal hemostasis, such as uremic platelet dysfunction, thromboembolic complications would be expected to be decreased.8-12 Fatal pulmonary embolic events seem to be rare in the dialysis population according to various registries. However, no relevant data are available on the occurrence of pulmonary thromboembolism in the modern recombinant erythropoietin therapy era.13,14 In a retrospective study, we assessed the preva-
From the Department of Internal Medicine and Ludwig Boltzmann Institute of Nephrology; and the Institute of Clinical Pathology, St Poelten, Austria. Received May 27, 1998; accepted in revised form October 16, 1998. Address reprint requests to Peter Balcke, MD, First Medical Department and Ludwig Boltzmann Institute of Nephrology, General Hospital A-3100, St Poelten, Austria. E-mail: [email protected]
娀 1999 by the National Kidney Foundation, Inc. 0272-6386/99/3304-0011$3.00/0 702
lence of pulmonary embolism at autopsy in a cohort of central dialysis unit patients within the last decade, during which time recombinant erythropoietin has been introduced, and tried to point out factors that could have contributed to its occurrence. METHODS
Study Patients In a 10-year period from 1987 to 1996, 185 chronic hemodialysis patients without a history of peritoneal dialysis or routine anticoagulation therapy for more than 3 months during the dialysis period died at the hospital, all of whom were autopsied. The mean age at death was 69.5 ⫾ 10.59 (standard deviation) years. Dialysis was performed three times weekly for 4 hours per session. Anticoagulation was performed with unfractionated heparin (Heparin Novo; NovoNordisk, Bagsvaerd, Denmark) or low-molecular-weight heparin (Deltaparin, Fragmin; Kabi Pharmacia, Uppsala, Sweden). The mean period since the start of hemodialysis treatment was 142.60 ⫾ 150.2 weeks. The types of basic renal disease were chronic glomerulonephritis in 42 patients, diabetic nephropathy in 42 patients, renal vascular disease in 39 patients, chronic pyelonephritis in 16 patients, druginduced nephropathy in 23 patients, hereditary nephropathy in 6 patients, and uncertain cause in 17 patients. The number of patients receiving chronic hemodialysis during the period of observation was 360. The mean annual death rate was 22.30% ⫾ 1.22%. In the same 10-year period, 8,660 patients died of nonrenal causes at the general hospital, which has divisions of internal medicine, surgery, gynecology, opthalmology, pediatrics, and neurology. Of these, 8,051 patients (92.96%) were autopsied. The mean age was 72.3 ⫾ 2.5 years.
American Journal of Kidney Diseases, Vol 33, No 4 (April), 1999: pp 702-708
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
Autopsies were performed by pathologists supervised by the same leading staff members. Pulmonary embolism was considered fatal when at least one of these criteria was found: (1) massive thrombi occluding the main pulmonary arteries, (2) multiple thrombi occluding more than 80% of the peripheral area of the pulmonary arterial bed, (3) pulmonary emboli in the absence of any other cause of death, or (4) pulmonary infarction affecting more than 80% of the pulmonary circulation. Pulmonary embolism was considered to contribute to death when, in a patient with severe or critical disease, pulmonary embolism was suspected as a relevant cofactor for death. Because it is difficult in practice to distinguish between fatal or contributing pulmonary embolisms in multimorbid patients, both are considered an equivalent occurrence in the study.
Statistical Analyses Data are expressed as mean ⫾ standard deviation. For comparisons of groups, the Mann-Whitney U test or the chi-squared test were performed. For correlation testings, Spearman’s rank correlation analysis was used. The level of statistical significance was defined as P less than 0.05.
RESULTS
From 1987 to 1996, pulmonary thromboembolism was detected in 12.43% (23 of 185) of the autopsies of chronic hemodialysis patients. This percentage is statistically significantly less than that found at 8,051 autopsies of nondialysis patients (21.77%; P ⫽ 0.0023). In the cases with pulmonary thromboembolism, embolism was estimated to have been fatal or to have contributed to death in 71.7% of the nonrenal cases and 52.17% of the hemodialysis patients (P ⫽ 0.039; Fig 1). Thus, pulmonary thromboembolism was fatal or contributing to death in 6.5% of the hemodialysis patients. In none of these patients had pulmonary thromboembolism been diagnosed premortally. The annual thromboembolism prevalence in the dialysis population remained statistically unchanged during the last 10-year observation period. It was 23.1% in 1987, 7.1% in 1988, 19.0% in 1989, 6.3% in 1990, 8.0% in 1991, 9.5% in 1992, 13.6% in 1993, 20.0% in 1994, 11.1% in 1995, and 6.7% in 1996. During this 10-year period, recombinant erythropoietin was administered to an increasing number and percentage of patients beginning in 1988. In 1996, 90% of the patients were receiving erythropoietin therapy. Simultaneously, the mean hematocrit values of all patients increased statistically significantly within these years from 25.5% ⫾ 5.0% in 1985
703
Fig 1. Percent of pulmonary thromboembolism at autopsy in 185 chronic hemodialysis patients (DP) and 8,051 nonrenal patients (controls). The prevalence of pulmonary thromboembolism (fatal or contributing to death and incidental) was statistically significantly less in the dialysis group than controls (P ⴝ 0.0023). Pulmonary thromboembolism was more frequently fatal or contributing to death in the controls than in the dialysis group (P ⴝ 0.039). (N), pulmonary thromboembolism fatal or contributing to death; (䊐), incidental pulmonary thromboembolism.
to 31.6% ⫾ 8.1% in 1996 (P ⫽ 0.0054). No statistically significant correlation between the prevalence rate of thromboembolism and the percentage of patients receiving recombinant erythropoietin or the hematocrit value could be noted (P ⫽ 0.624 and P ⫽ 0.273, respectively; Fig 2). Mean platelet values showed no statistically significant variation over these years. Anticoagulation during hemodialyses was increasingly performed with low-molecular-weight heparin (in 37.88% of the patients in 1996), whereas the percentage of anticoagulation performed with unfractionated heparin decreased. The prescription of salicylates increased continuously to 28.57% in 1996. Other potentially relevant parameters are listed in Table 1. In a further analysis, data obtained less than 2 weeks before death for chronic hemodialysis patients with and without thromboembolism were compared. Pulmonary thromboembolism was associated with a statistically significantly greater number of deaths from infection and a shorter period since onset of hemodialysis treatment. No statistically significant difference between groups could be found regarding age, sex, basic renal disease, current hematocrit values, platelet count, blood urea nitrogen levels, type of dialysis anticoagulant, administration of salicylates, days of
704
WIESHOLZER ET AL
Fig 2. Percentage of patients with pulmonary thromboembolism at autopsy (bars) and percent hemodialysis unit patients receiving recombinant erythropoietin therapy (line) from 1987 to 1996. The parameters have no statistically significant relationship (P ⴝ 0.624).
immobilization, prevalence of diabetes, and type of dialysis access (Table 2). DISCUSSION
Our study shows that pulmonary thromboembolism is a frequent occurrence in preterminal chronic hemodialysis patients. For the assessment of thromboembolism frequency, autopsy findings were chosen because conventional clinical methods of detection are far less reliable.2,15 Although pulmonary emboli could be detected in a substantial percentage of hemodialysis patients, the frequency was statistically significantly less and the fatality rate from thromboembolism was less than those in nondialysis patients autopsied during the observation period at the same general hospital and by the identical leading medical staff. In our opinion, the discrepancy between both groups could be caused by various factors. Thromboembolic risk may be greater in nondialysis patients as a consequence of different basic or concomitant diseases or therapies triggering thromboembolism, such as malignancy or major surgery.16 Conversely, thromboembolism risk might be reduced in the dialysis population by the abnormal hemostasis encountered in chronic hemodialysis patients, which results principally from perturbed platelet adhesion and aggregation.8-12 In addition to the abnormal hemostasis, anticoagulation with unfractionated or lowmolecular-weight heparin during dialysis procedures could also contribute to this discrepancy.
The prevalence of pulmonary thromboembolism seems to be stable during the last decade. This is amazing because so many modifications in hemodialysis therapy have occurred within the last 10 years; most importantly, therapy with recombinant erythropoietin.17 Erythropoietin therapy leads to a reduction of the prolonged bleeding time through the increase in erythrocyte counts, thereby modifying rheological factors, as well as through an improved platelet function by increasing the number of glycoprotein IIb-IIIa molecules on the platelet membrane.18,19 As a further potentially thrombosis-promoting factor by recombinant erythropoietin, an enhanced release of plasminogen activator inhibitor-1 by endothelial cells was described recently.20 However, although a similar influence on hemostasis can be assumed in our patients treated with recombinant erythropoietin, no increase in the prevalence of pulmonary thromboembolism could be noted despite a steady increase in the percentage of patients receiving erythropoietin therapy and the increment in the average hematocrit value of the patients within the last decade. Two mechanisms could be responsible. First, the improvement of anemia leading to target hematocrit values in our patients of between 30% and 35% is not sufficient to correct the abnormal hemostasis in uremia relevantly to the thrombosis risk. Second, anticoagulation performed with unfractionated or low-molecular-weight heparin during the thrice-weekly dialysis procedure has a
Pulmonary embolism at autopsy Patients with erythropoietin therapy Hematocrit (%) BUN (mg/dL) Platelets/µL Unfractionated heparin (%) Low-molecular-weight heparin (%) Patients with dialysis catheters (%) Patients with diabetic nephropathy (%) Patients with salicylates (%) Age (yr)
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
3/13 (23.1)
1/14 (7.1)
4/21 (19)
1/16 (6.3)
2/25 (8)
2/21 (9.5)
3/22 (13.6)
4/20 (20)
2/18 (11.1)
1/15 (6.7)
P 0.62
0 (0)
4 (4)
14 (14)
29 (23)
27 (26)
41 (40)
55 (53)
60 (61)
53 (62)
63 (90)
0.0027
25.5 ⫾ 5 79.9 ⫾ 18.1 241,329 ⫾ 70,093 100
26.6 ⫾ 4.9 81.9 ⫾ 21.2 252,444 ⫾ 81,746 100
26.1 ⫾ 4.4 80.3 ⫾ 23.1 231,896 ⫾ 84,166 100
27.7 ⫾ 4.7 85.7 ⫾ 25 218,707 ⫾ 59,828 100
27.4 ⫾ 3.7 75.5 ⫾ 18.6 241,985 ⫾ 79,938 91.87
28.7 ⫾ 3.9 79.5 ⫾ 25.2 246,250 ⫾ 89,963 37.88
28.6 ⫾ 3.5 81.7 ⫾ 22 250,500 ⫾ 69,843 38.16
28 ⫾ 2.5 83 ⫾ 20.9 228,366 ⫾ 65,557 25.33
30.3 ⫾ 2.9 75.1 ⫾ 18.2 224,500 ⫾ 74,295 31.43
31.6 ⫾ 8.1 77 ⫾ 22.1 236,027 ⫾ 69,450 62.12
0.0054 0.283 0.413
0
0
5.32
7.69
10.68
21.15
1.23 62.9 ⫾ 14.5
23.4
0
0
0.014
8.33
62.12
61.84
74.67
68.57
37.88
0.014
7.81
11.76
9.71
8.65
10.2
15.29
20
0.023
18.45
17.97
9.8
20.39
21.15
21.43
22.35
31.43
0.274
2.35
1.15
4.04
9.72
6.8
9.62
12.24
21.18
28.57
0.0054
58.5 ⫾ 15.8
59.5 ⫾ 16.9
60.6 ⫾ 16.1
60.5 ⫾ 15.4
60.9 ⫾ 15.3
60.8 ⫾ 16
62.2 ⫾ 16.4
63 ⫾ 15.4
62.9 ⫾ 14.5
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
Table 1. Changes in the Prevalence of Pulmonary Embolism at Autopsy and in Various Parameters Potentially Relevant to Pulmonary Thromboembolism in All Chronic Hemodialysis Unit Patients From 1987 to 1996
0.0865
NOTE. Values expressed as number (%) or mean ⫾ standard deviation. Abbreviation: BUN, blood urea nitrogen.
705
706
WIESHOLZER ET AL Table 2. Premortal Parameters of Patients With and Without Pulmonary Thromboembolism at Autopsy Patients With PE
Patients Age (yr) Sex Men Women Basic renal disease Glomerulonephritis Pyelonephritis Drug-associated nephropathy Hereditary nephropathy Renal vascular disease Diabetes Other Cause of death Fatal pulmonary thromboembolism Cardiac Vascular Infection Liver Miscellaneous Other Erythropoietin therapy Erythropoietin dose/wk (units) Hematocrit value (%) Platelets/µL BUN (mg/dL) Unfractionated heparin Low-molecular-weight heparin Salicylates Days of immobilization Diabetics Type of dialysis access Arteriovenous fistula Permcath catheter Femoral vein cannula Subclavian or jugular vein cannula GoreTex shunt Arteriovenous graft Weeks since start of hemodialysis
Without PE
P
23 72.4 ⫾ 7.7
162 69.1 ⫾ 10.9
9 (39.1) 14 (60.9)
98 (60.5) 64 (39.5)
7 (30.4) 2 (8.7) 1 (4.3) 1 (4.3) 3 (13) 7 (30.4) 2 (8.7)
35 (21.6) 14 (8.6) 22 (13.6) 5 (3.1) 36 (22.2) 35 (21.6) 15 (9.3)
0.344 0.993 0.209 0.749 0.313 0.344 0.93
8 (34.8) 3 (13.0) 2 (7.8) 9 (39.1) 0 (0) 1 (4.3) 0 (0) 8 (34.8) 10,000 ⫾ 4,751 27.3 ⫾ 4.5 232,496 ⫾ 124,336 80.54 ⫾ 29.3 18 (78.3) 5 (21.7) 3 (13) 19 ⫾ 22.7 8 (34.8)
0 (0) 85 (52.5) 23 (14.2) 31 (19.1) 1 (0.6) 19 (11.7) 3 (1.9) 55 (34) 10,581 ⫾ 6,559 27.1 ⫾ 4.9 244,188 ⫾ 103,047 73.89 ⫾ 21.5 132 (81.5) 30 (18.5) 26 (16) 13.1 ⫾ 21.8 61 (37.7)
0.0004 0.47 0.029 0.705 0.286 0.51 0.985 0.83 0.73 0.42 0.44 0.92 0.9 0.71 0.11 0.757
13 (56.5) 4 (17.4) 2 (8.7) 1 (4.3) 1 (4.3) 2 (8.7) 103.3 ⫾ 107.2
78 (48.2) 40 (24.7) 7 (4.3) 20 (12.3) 14 (8.6) 3 (1.9) 148.21 ⫾ 155
0.23 0.052
0.452 0.441 0.361 0.257 0.48 0.058 0.031
NOTE. Values expressed as number (%) or mean ⫾ standard deviation. P refers to numbers. Abbreviations: PE, pulmonary thromboembolism; BUN, blood urea nitrogen.
substantial protective effect on thromboembolism by mitigating any potential thrombosispromoting activity of recombinant erythropoietin. In the last years, a protective effect of anticoagulants could theoretically have become more relevant by the more frequent use of lowmolecular-weight heparin instead of unfractionated heparin for hemodialysis and salicylates prescribed for various causes. Low-molecularweight heparin has a longer lasting anticoagulant
activity than unfractionated heparin and could thus have had a more protective influence,21,22 and salicylates administered for cardiovascular or cerebrovascular therapy reasons have a prolonged efficiency in renal failure.23 However, our statistical analyses of parameters at the premortal period do not suggest this explanation, that a potentially greater thromboembolic risk by recombinant erythropoietin therapy could have been compensated by such modifica-
PULMONARY EMBOLISM IN HEMODIALYSIS PATIENTS
tions, because the thromboembolism and nonthromboembolism groups did not differ statistically significantly regarding the type of heparin used during the dialysis session or the prevalence of salicylate use. The nondifferent prevalence of erythropoietin therapy and hematocrit values in the premortal period underline additionally that erythropoietin therapy does not have a major role for thromboembolism. We believe that abnormal hemostasis in the dialysis population is such a prevailing factor; thus, the effect of erythropoietin therapy is not sufficient to increase the thromboembolic risk substantially. The analyses of premortal parameters did not show statistically significant influences of the prevalence of diabetes, blood urea nitrogen levels and platelet counts, days of immobilization, and type of dialysis access. However, thromboembolism was associated with a greater prevalence of death from infection and fewer weeks of dialysis; the latter finding most likely caused by more severe morbidity. Moreover, the thromboembolic risk may be influenced by prethrombotic disorders that were not assessed in our study, such as the inherited resistance to activated protein C as a consequence of factor V Leiden mutation; congenital deficiencies of antithrombin III, protein C, protein S, or plasminogen; and hyperhomocysteinemia, which as genetic disorders can be assumed to have remained unchanged within the last decade, as can be considered for the prevalence of antiphospholipid antibodies.3-7 We believe that the factor V Leiden mutation deserves the most interest in this context because its prevalence in the general population ranges from 2% to 15%, and its impact for thromboembolism is markedly increased when hyperhomocysteinemia coexists, as in the setting of chronic hemodialysis treatment, in which a slight increase in blood homocysteine levels has been described for most patients.24-27 Consequently, these factors could have an important role for thromboembolism in various conditions, including the state of severe infection. The pulmonary embolisms found at autopsy were not diagnosed antemortem in one single case, showing the difficulty of detecting pulmonary emboli clinically in principle. This is consistent with various studies that showed antemor-
707
tem diagnoses of pulmonary embolism were not made in up to 70% of autopsy-proven events in general.1,2 Although pulmonary thromboembolism was fatal or contributing to death in only a small percentage of patients, nonfatal events may also be relevant regarding such consecutive complications as pneumonia, for which as a secondary event, venous thromboembolism could be an underestimated cause in this population. However, regarding the preterminal prevalence of pulmonary thromboembolism, our data suggest that the recombinant erythropoietin therapy introduced within the last decade did not result in a substantial increase in the rate of thromboembolic events in the hemodialysis population. REFERENCES 1. Lindblad B, Sternby NH, Bergqvist D: Incidence of venous thromboembolism verified by necropsy over 30 years. BMJ 302:709-711, 1991 2. Stein PD, Henry JW: Prevalence of acute pulmonary embolism among patients in a general hospital and at autopsy. Chest 108:978-981, 1995 3. Diuguid DL: Oral anticoagulant therapy for venous thromboembolism. N Engl J Med 336:433-434, 1997 4. Ginsberg JS: Drug therapy: Management of venous thromboembolism. N Engl J Med 335:1816-1828, 1997 5. Heijboer H, Brandjes DPM, Bu¨ller HR, Sturk A, ten Cate JW: Deficiencies of coagulation-inhibiting and fibrinolytic proteins in outpatients with deep-vein thrombosis. N Engl J Med 323:1512-1516, 1990 6. Dahlback B: Inherited thrombophilia: Resistance to activated protein C as a pathogenetic factor of venous thromboembolism. Blood 85:607-614, 1995 7. Ginsberg JS, Wells PS, Brill-Edwards P, Donovan D, Moffatt K, Johnston M, Stevens P, Hirsh J: Antiphospholipid antibodies and venous thromboembolism. Blood 86:36853691, 1995 8. Eberst ME, Berkowitz LR: Hemostasis in renal disease: Pathophysiology and management. Am J Med 96:168178, 1994 9. Rabelink TJ, Zwaginga JJ, Koomans HA, Sixma JJ: Thrombosis and hemostasis in renal disease. Kidney Int 46:287-293, 1994 10. Gawaz MP, Dobos G, Spath M: Impaired function of platelet membrane glycoprotein IIb-IIIa in end-stage renal disease. J Am Soc Nephrol 5:36-46, 1994 11. Heistinger M, Stockenhuber F, Schneider B, Pabinger I, Brenner B, Wagner B, Balcke P, Lechner K, Kyrle PA: Effect of conjugated estrogens on platelet function and prostacyclin generation in chronic renal failure. Kidney Int 38:1181-1187, 1990 12. Macconi D, Vigano G, Bisogno G, Galbusera M, Orisio S, Remuzzi G, Livio M: Defective platelet aggregation in response to platelet-activating factor in uremia asso-
708
ciated with low platelet thromboxane A2 generation. Am J Kidney Dis 19:318-325, 1992 13. Bloembergen WE, Port FK, Mauger EA, Wolfe RA: Causes of death in dialysis patients: Racial and gender differences. J Am Soc Nephrol 5:1231-1242, 1994 14. Mailloux LU, Bellucci AG, Wilkes BM, Napolitano B, Mossey RT, Lesser M, Bluestone PA: Mortality in dialysis patients: Analysis of the causes of death. Am J Kidney Dis 18:326-335, 1991 15. The PIOPED Investigators: Value of the ventilation/ perfusion scan in acute pulmonary embolism diagnosis. JAMA 363:2753-2759, 1990 16. Giuntini C, Di Ricco G, Marini C, Melillo E, Palla A: Pulmonary embolism: Epidemiology. Chest 107:3S-9S, 1995 (suppl) 17. Eschbach JW, Egrie JC, Downing MR, Browne JK, Adamson JW: Correction of the anemia of end-stage renal disease with recombinant human erythropoietin. Results of a combined phase I and II clinical trial. N Engl J Med 316:73-78, 1987 18. Cases A, Escolar G, Reverter JC, Ordinas A, LopezPedret J, Revert L, Castillo R: Recombinant human erythropoietin treatment improves platelet function in uremic patients. Kidney Int 42:668-672, 1992 19. Livio M, Gotti E, Marchesi D: Uremic bleeding: Role of anemia and beneficial effect of red cell transfusion. Lancet 2:1013-1017, 1982 20. Nagai T, Akizawa T, Kohjiro S, Koiwa F, Nabeshima K, Niikura K, Kino K, Kanonori E, Kinugasa E, Ideura T: RHuEPO enhances the production of plasminogen activator
WIESHOLZER ET AL
inhibitor-1 in cultured endothelial cells. Kidney Int 50:102107, 1996 21. Lockner D, Bratt G, To¨rnebohm E, Aberg W: Pharmacokinetics of intravenously and subcutaneously administered Fragmin in healthy volunteers. Hemostasis 16:8-10, 1986 22. Follea G, Laville M, Pozet N, Dechavanne M: Pharmacokinetic studies of standard heparin and low-molecularweight heparin in patients with chronic renal failure. Haemostasis 16:147-151, 1986 23. Williams ME, Weinblatt M, Rosa RM, Griffin VL, Goldlust B, Shang SF, Harrison LI, Brown RS: Salicylate kinetics in patients with chronic renal failure undergoing hemodialysis. Clin Pharmacol Ther 39:420-428, 1986 24. Beauchamp NJ, Daly ME, Hampton KK, Cooper PC, Preston FE, Peake IR: High prevalence of a mutation in the factor V gene within the UK population: Relationship to activated protein C resistance and familial thrombosis. Br J Haematol 88:219-222, 1994 25. Dahlback B, Hillarp A, Rosen S, Zoller B: Resistance to activated protein C, the FV:Q506 allele, and venous thrombosis. Ann Hematol 72:166-176, 1996 26. Mandel H, Brenner B, Berant M, Rosenberg N, Lanir N, Jakobs C, Fowler B, Seligsohn U: Coexistence of hereditary homocystinuria and factor V Leiden–Effect on thrombosis. N Engl J Med 334:763-768, 1996 27. Bachmann J, Tepel M, Raidt H, Riezler R, Graefe U, Langer K, Zidek W: Hyperhomocysteinemia and the risk for vascular disease in hemodialysis patients. J Am Soc Nephrol 6:121-125, 1995