- Email: [email protected]
Incidence, outcome, and risk factors of cerebrovascular events in patients undergoing maintenance hemodialysis
Incidence, Outcome, and Risk Factors of Cerebrovascular Events in Patients Undergoing Maintenance Hemodialysis Mitsunobu Kawamura, MD, Shouichi Fijimoto, MD, Shuichi Hisanaga, MD, Yoshitaka Yamamoto, MD, and Tanenao Eto, MD ● We retrospectively investigated the incidence and prognosis of and risk factors for cerebrovascular events in 1,064 patients with chronic uremia who received maintenance hemodialysis (HD) for more than 3 months during 24 years in our dialysis units in Miyazaki, Japan. Cerebrovascular events developed in 98 patients (9.2%). The confirmed incidences of cerebral hemorrhage (CH) and infarction were 8.7 and 3.7 per 1,000 patient-years, respectively. Of the 56 patients with CH, 40 (71.4%) died within 3 months of the onset of CH. Ganglio-thalamic lesion was observed in 32 (80.0%) of 40 patients with CH confirmed by a brain computed tomography. The incidence of polycystic kidney disease was higher in the CH group than in the overall HD population (12.5% v 3.9%, P F 0.01). Of the 13 patients with diabetes mellitus and nephrosclerosis, nine (69.2%) developed CH within 36 months of the initiation of HD; 11 (78.6%) of 14 patients with chronic glomerulonephritis developed CH after 36 months. CH developed in six patients (15.0%) within 6 hours of a previous HD session. We compared laboratory values, the supine blood pressure, and electrocardiographic (ECG) findings in 35 patients with CH and a control group (66 patients) matched in age, sex, basal renal disease, age at the initiation of HD, and the duration of HD. Data were obtained before and after HD 3 to 4 months before the first attack of CH. The systolic and diastolic blood pressure (SBP, DBP) before and after HD were significantly higher in the CH group than in the control group (pre-HD SBP: 171 ⴞ 22.5 v 154 ⴞ 19.3 mm Hg, P F 0.001; pre-HD DBP: 89 ⴞ 13.6 v 81 ⴞ 9.6 mm Hg, P F 0.001). The incidence of left ventricular hypertrophy was higher, and the Kt/V was significantly lower (1.23 ⴞ 0.26 v 1.38 ⴞ 0.34, P F 0.05) in the CH group than in the control group. However, there were no significant differences in the serum levels of albumin and cholesterol or the total dose of heparin during HD sessions between groups. In conclusion, the incidence of CH was high, and its prognosis was poor, in patients undergoing maintenance HD. Reversible risk factors include hypertension and possibly the amount of HD prescribed, but not anticoagulation with heparin. r 1998 by the National Kidney Foundation, Inc. INDEX WORDS: Cerebral hemorrhage, cerebrovascular disease, hemodialysis patients, hypertension.
C
EREBROVASCULAR DISEASES and cardiovascular diseases are important causes of death in patients receiving maintenance hemodialysis (HD).1-5 The number of patients receiving HD has been increasing, and HD is being initiated increasingly in older patients in Japan, where HD is used in more than 90% of patients with end-stage renal disease.6 Comorbid conditions, such as hypertension and hyperlipidemia, which are risk factors for cerebrovascular disease, are commonly observed in HD patients. There is a higher prevalence of cerebral hemorrhage (CH) than cerebral infarction (CI) in HD patients,7,8 unlike in the general population. However, only a few studies have examined the characteristics of HD patients who develop cerebrovascular events (CVE), and no studies have compared the characteristics of CH patients with those of matched controls on maintenance HD. In Miyazaki in the Kyushu area of southwestern Japan, HD patients who develop CVE are usually transferred to one of only a few regional hospitals, where a brain computed tomography (CT) is performed. Because the movement of HD patients is less frequent in this area, we are able to survey the follow-up data of CVE pa-
tients on HD with relative ease and completely. We retrospectively investigated the clinical profiles of HD patients who developed CH in 17 dialysis centers in Miyazaki, Japan. PATIENTS AND METHODS
Entry Criteria and Definitions Of 1,064 patients who received maintenance HD (4 or 5 hours, 3 times weekly) for more than 3 months at 17 dialysis centers in Miyazaki, Japan, between January 1971 and December 1994, 98 patients (9.2%) developed a first attack of CVE after beginning HD. CVE was diagnosed based on CT findings or clinical signs, such as the sudden onset of disturbances in consciousness, speech, visual, motor, or sensory functions. Patients with a history of cerebrovascular disease, including transient ischemic attack, and accidental brain CT findings compatible with cerebrovascular disease were excluded. Of the 98 patients,
From the First Department of Internal Medicine, Miyazaki Medical College, Miyazaki, Japan; and the Yamamoto Clinic, Miyazaki, Japan. Received March 17, 1997; accepted December 12, 1997. Address reprint requests to Tanenao Eto, MD, First Department of Internal Medicine, Miyazaki Medical College, 5200 Kihara, Kiyotake, Miyazaki, Japan 889-16. E-mail: [email protected]
r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3106-0013$3.00/0
American Journal of Kidney Diseases, Vol 31, No 6 (June), 1998: pp 991-996
991
992
KAWAMURA ET AL
13 patients were excluded because they did not undergo CT at the onset of CVE. We first investigated the demographic and clinical characteristics of these 85 patients in terms of the CVE subtype. We further analyzed clinical features of the 40 patients with CH for whom clinical data, including medical records, laboratory data, and the brain CT films at the onset of disease, were available. Finally, to identify factors that contributed to the development of CH, the characteristics of 35 of these patients with CH were compared with characteristics of a control group consisting of 66 patients from the original group of 1,064 HD patients who had no clinical evidence of CVE and who were matched in age, sex, basal renal disease, age at the initiation of HD, and the duration of HD.
Data Collection Medical and clinical records were collected by one investigator. Laboratory data, chest x-ray films, electrocardiograms (ECGs), and supine blood pressure were obtained before and after HD 3 to 4 months before the first attack of CVE. The cardiothoracic ratio (CTR) and the voltage of SV1 plus RV5 in the precordial lead were determined from x-ray and ECGs obtained before HD. The Kt/V was calculated as follows: Kt/V ⫽ ⫺1n (R ⫺ 0.03 ⫺ 0.75 ⫻ UF/W), where R is the ratio of the postdialysis to the predialysis (post/pre) level of urea nitrogen, UF is the volume removed by ultrafiltration in liters, and W is the postdialysis weight in kilograms. The size of hematomas in patients with CH was determined from CTs obtained at the onset as follows: the maximum transverse diameter ⫻ the maximal anteroposterior diameter ⫻ the maximal superoinferior diameter ⫻ 1⁄2.9
Statistical Analysis Statistical analyses were performed using the Macintosh StatView program (Abacus Concepts Inc., Berkeley, CA, 1992). Differences between groups were evaluated by analysis of variance, an unpaired t-test, the chi-square test, and by determining correlation coefficients for parametric data. Results are expressed as the mean ⫾ standard deviation (SD). A P value ⬍ 0.05 was accepted as indicating statistical significance.
RESULTS
Clinical Background CVE was diagnosed in 85 (8.0%) of 1,064 patients (Table 1). They included 56 patients with CH, 24 with CI, and five with subarachnoid hemorrhage. The definite incidence of CVE was 13.2 per 1,000 patient-years. The incidences of CH, CI, and subarachnoid hemorrhage were 8.7, 3.7, and 0.8 per 1,000 patient-years, respectively. Male sex predominated in the overall HD patients and in the CVE group. Patients in the CH group were significantly younger at initiation of HD and at the last observation compared with patients in the CI group. The mean age was greater in the CI group than in the total population. Although the duration of HD did not differ among groups, the time from the initiation of HD
Table 1. Clinical Characteristics of HD Patients CVE
Number Sex, male Age at start of HD (yr) Duration of HD (mo) Age at last observation (yr) Basal renal disease Chronic glomerulonephritis Diabetes mellitus Nephrosclerosis Polycystic kidney disease Others Number of deaths
Total
Total
CH
CI
SAH
1,064 634 (59.6%) 53.1 ⫾ 15.5 72.4 ⫾ 60.0 59.1 ⫾ 13.9
85 52 (61.2%) 53.8 ⫾ 13.6 70.6 ⫾ 59.6 59.7 ⫾ 11.7
56 36 (64.3%) 50.3 ⫾ 13.1* 79.7 ⫾ 66.6 56.8 ⫾ 11.3*
24 14 (58.3%) 61.5 ⫾ 12.6† 56.0 ⫾ 40.3 66.1 ⫾ 11.3‡
5 2 (40.0%) 57.4 ⫾ 7.6 39.9 ⫾ 27.7 60.7 ⫾ 6.1
576 (54.1%) 187 (17.6%) 108 (10.2%) 42 (3.9%) 151 (14.2%) 297 (27.9%)
35 (41.2%) 19 (22.4%) 9 (10.6%) 8 (9.4%)§ 14 (16.4%) 54 (63.5%)¶
25 (44.6%) 10 (17.9%) 4 (7.1%) 7 (12.5%)\ 10 (17.9%)§ 40 (71.4%)¶
6 (25.0%)\ 9 (37.5%)§ 5 (20.8%) 0 (0%) 4 (16.7%) 12 (50.0%)
4 (80.0%) 0 (0%) 0 (0%) 1 (20.0%) 0 (0%) 2 (40.0%)
Abbreviations: CVE, cerebrovascular events; CH, cerebral hemorrhage; CI, cerebral infarction; SAH, subarachnoid hemorrhage; HD, hemodialysis. *P ⬍ 0.01 v CI by ANOVA. †P ⬍ 0.01 v all patients by t-test. ‡P ⬍ 0.05 v all patients by t-test. §P ⬍ 0.05 v all patients by chi-square test. \P ⬍ 0.01 v all patients by chi-square test. ¶P ⬍ 0.001 v all patients by chi-square test.
CEREBROVASCULAR EVENTS IN HD PATIENTS
to the first attack of CVE was shorter in the CI group than in the CH group (39.4 ⫾ 34.4 v 63.5 ⫾ 57.5 months, P ⬍ 0.05). The incidence of polycystic kidney disease was significantly higher in the CH group and the total CVE group than in all maintenance HD patients. The incidence of chronic glomerulonephritis was higher in the CH group than in the CI group, and the incidences of diabetes mellitus and nephrosclerosis were higher in the CI group, but the differences were not significant. There were significantly more deaths in the CH group and the total CVE group than in the total HD group. The cause of death in the total HD group was CVE in 57 patients (19.2%), infection in 40 patients (13.5%), congestive heart failure (CHF) in 38 patients (12.8%), malignancy in 33 patients (11.1%), myocardial infarction in 13 patients (4.4%), and other causes in 116 patients (39.0%). CVE was the cause of death in 35 (87.5%) of 40 patients in the CH group. The causes of death in the CI group were CVE in two patients (16.7%), infection in three patients (25.0%), and CHF in two patients (16.7%). Comparison of CH and CI Groups Patients in the CH group were significantly younger at the onset of CVE than patients in the CI group (Table 2). The survival period after the onset of CVE was shorter in the CH group than in the CI group; the mortality rate of CVE tended to be higher in the CH group (71.4%) than in the Table 2. Characteristics of Cerebral Hemorrhage (CH) and Cerebral Infarction (CI) Groups
Age at onset of CVE (yr) Survival after the onset of CVE (mo) Pre-HD SBP (mm Hg) Pre-HD DBP (mm Hg) Post-HD SBP (mm Hg) Post-HD DBP (mm Hg) Albumin (g/dL) High-density lipoprotein cholesterol (mg/dL)
CH (56)
CI (24)
55.9 ⫾ 11.4*
64.7 ⫾ 11.4
6.4 ⫾ 12.4* 172 ⫾ 21.6† 90 ⫾ 13.4† 158 ⫾ 24.0* 86 ⫾ 11.3* 4.08 ⫾ 0.40†
17.9 ⫾ 16.4 160 ⫾ 24.5 77 ⫾ 8.3 142 ⫾ 19.9 73 ⫾ 9.3 3.80 ⫾ 0.45
40.3 ⫾ 12.1†
33.0 ⫾ 9.5
Abbreviations: HD, hemodialysis; CVE, cerebrovascular events; SBP, systolic blood pressure; DBP, diastolic blood pressure. *P ⬍ 0.01 v CI (t-test). †P ⬍ 0.05 v CI (t-test).
993
CI group (50.0%). The systolic blood pressure (SBP) and diastolic blood pressure (DBP) before and after HD were significantly higher in the CH group than in the CI group. Actually, 89% (50 of 56) and 67% (16 of 24) of patients were prescribed antihypertensive drugs in the CH and the CI groups, respectively. Serum levels of albumin and high-density lipoprotein cholesterol were significantly lower in the CI group, but there were no significant differences in the CTR, the ultrafiltration volume, the amount of heparin administered, or other laboratory parameters, such as levels of total cholesterol, triglycerides, and production of calcium and phosphorus (data not shown). Clinical Features in the CH Group In the group of 40 CH patients for whom all clinical data were available, males accounted for 26 patients (65.0%) and for 17 (65.4%) of 26 patients who died within 3 months of the onset of CH. Hemorrhagic lesions occurred in the ganglionic-thalamus in 32 (80.0%) of 40 patients; one patient experienced a lobar hemorrhage, and none experienced cerebellar hemorrhage. To examine the differences in the clinical characteristics by the dialysis periods at the onset of CVE, we classified CH patients according to the time from the initiation of HD to the first attack of CVE, as follows: ⱕ36 months (group A, n ⫽ 17) and ⬎36 months (group B, n ⫽ 23). Patients in group A were significantly older than patients in group B (57.7 ⫾ 8.4 v 45.6 ⫾ 12.6 years, P ⬍ 0.01), and the ECG voltage of the left ventricle (Sv1 ⫹ Rv5 mV) was higher in group A than in group B (4.70 ⫾ 1.51 v 3.58 ⫾ 1.37 mV, P ⬍ 0.05). There were no differences between groups in SBP, DBP, and Kt/V. All patients with nephrosclerosis and half of those with diabetes mellitus belonged to group A, whereas most of those with chronic glomerulonephritis were in group B. There were no significant differences between CH patients who survived and those who died in clinical characteristics, such as the SBP/DBP (167 ⫾ 19.3/86 ⫾ 13.7 mm Hg v 175 ⫾ 22.6/ 92 ⫾ 13.0 mm Hg), the dose of standard heparin (unfractionated) (4,414 ⫾ 1,311 v 4,379 ⫾ 1,005 unit/HD), onset time of CH after a previous HD session (16.0 ⫾ 14.2 v 25.0 ⫾ 13.9 hours), the serum level of albumin (4.25 ⫾ 0.37 v
994
4.00 ⫾ 0.40 g/dL), and the Kt/V (1.32 ⫾ 0.36 v 1.15 ⫾ 0.21). Hematomas were lager in patients who died (170.6 ⫾ 119.3 cm3 ) than in survivors (26.7 ⫾ 32.8 cm3, P ⬍ 0.001), and the size of the hematoma was correlated with the SBP at the start of HD (r ⫽ 0.504, P ⬍ 0.05) (Fig 1). Patients without intraventricular hemorrhage had a better survival rate. We studied the effects of anticoagulation therapy during HD on the occurrence of CH. CH occurred more than 6 hours after the end of an HD session in 34 (85%) of 40 patients and 6 hours or less after HD in six patients. In the early-onset group, three of six patients had a CVE during an HD session. There were no significant differences in the dose of heparin, the size of the hematoma, or the blood pressure (BP) between groups. Survival was prolonged (12.2 ⫾ 15.3 v 7.0 ⫾ 11.4 months, P ⬍ 0.01), and the age at the onset of CVE was younger (46.2 ⫾ 10.7 v 57.8 ⫾ 10.2 years, P ⬍ 0.05) in the early-onset group than in the late-onset group. Differences in Clinical Parameters Between CH and Control Groups There were no significant differences in patient characteristics between the CH (n ⫽ 35) and control (n ⫽ 66) groups. Histories of cerebrovascular diseases, and of coronary or peripheral vascular diseases before beginning HD, were documented in zero, nine, and five patients in the CH group, and five, three, and two patients in the control group. The number of survivors at the last observation was significantly lower in the
KAWAMURA ET AL Table 3. Clinical, Laboratory, and ECG Data in CH and Control Groups CH
Control
Pre-HD SBP (mm Hg) 171 ⫾ 22.5 154 ⫾ 19.3* Pre-HD DBP (mm Hg) 89 ⫾ 13.6 81 ⫾ 9.6* Post-HD SBP (mm Hg) 157 ⫾ 25.5 140 ⫾ 24.8* Post-HD DBP (mm Hg) 86 ⫾ 11.4 76 ⫾ 10.7* CTR (%) 51.9 ⫾ 6.0 50.8 ⫾ 6.2 ECG, Sv1 ⫹ Rv5 (mV) 4.12 ⫾ 1.55 3.19 ⫾ 1.29† Ultrafiltration (L/HD) 2.41 ⫾ 1.16 2.84 ⫾ 1.01 Kt/V 1.23 ⫾ 0.26 1.38 ⫾ 0.34‡ Heparin (unit/HD) 4,290 ⫾ 1,036 4,090 ⫾ 1,453 Total cholesterol (mg/dL) 163 ⫾ 29.3 164 ⫾ 36.2 Triglycerides (mg/dL) 135 ⫾ 72.3 123 ⫾ 63.0 High-density lipoprotein cholesterol (mg/dL) 40.6 ⫾ 12.5 42.0 ⫾ 14.0 Albumin (g/dL) 4.08 ⫾ 0.40 4.10 ⫾ 0.41 Ca ⫻ Pi 53.9 ⫾ 16.6 56.6 ⫾ 17.1 Abbreviations: CH, cerebral hemorrhage; HD, hemodialysis; SBP, systolic blood pressure; DBP, diastolic blood pressure; CTR, cardiothoracic ratio; ECG, electrocardiogram; Ca ⫻ Pi, production of calcium and phosphorus. *P ⬍ 0.001 (t-test). †P ⬍ 0.01 (t-test). ‡P ⬍ 0.05 (t-test).
CH group (21 of 26 patients died of CVE) compared with the control group. The SBP and DBP before and after HD were significantly higher in the CH group than in the control group (Table 3). The BP values just before CVE were significantly higher than those obtained 3 to 4 months before CVE (pre-HD SBP/DBP, 180 ⫾ 25.8/94 ⫾ 15.7 mm Hg v 171 ⫾ 22.5/89 ⫾ 13.6 mm Hg, P ⬍ 0.01). Conversely, the means of long-term BP values before CVE were not different from those of 3 to 4 months before CVE (pre-HD SBP/DBP, 168 ⫾ 22.0/88 ⫾ 13.1 mm Hg v 171 ⫾ 22.5/ 89 ⫾ 13.6 mm Hg, NS). The ECG voltage was also significantly higher in the CH group. There were no differences between groups in the ultrafiltration volume and the total dose of heparin per HD session. The Kt/V was lower in the CH group than in the control group. DISCUSSION
Fig 1. Correlation between SBP and size of hemotoma in CH patients.
Cerebrovascular disease was the most common cause of deaths among the 1,064 HD patients in the current study, accounting for 19.2% of all deaths. The CVE-related mortality rate in chronic dialysis patients is 12.7% in Japan6 and 4.8% in the United States.10 By the causes of
CEREBROVASCULAR EVENTS IN HD PATIENTS
differences in human species or nutritional enviroments, and so forth, as well as general population, the CVE-related mortality rate in chronic dialysis patients may be higher in Japan than in the United States. The higher frequency of HD as the treatment modalities for the patients with end-stage renal disease also may contribute to the higher mortality rate due to CVE in Japan. The incidence of confirmed stroke events during maintenance HD per 1,000 patient-years in the current study was 8.7 in the CH group, 3.7 in the CI group, and 0.8 in the subarachnoid hemorrhage group. The incidences of CVE and, especially, of CH were higher in the current study than in the general population, in which the reported incidences of stroke and CH are 0.95 and 0.1 to 2.2 per 1,000 patient-years.12,13 Approximately 75% of strokes in the general population are caused by CI.14 The current results confirm findings of previous studies,3,4 which also found a higher prevalence of CH among dialysis patients. These findings may be related to the high prevalence of hypertension in the dialysis patients, which is established as the risk factor for the occurrence of CVE in the general population. However, asymptomatic CI is frequently detected by CT and magnetic resonance imaging,15 and decreased cerebral blood flow has been detected by positron emission tomography in HD patients.16 Furthermore, in this study, we examined the definite cases with a first attack of CVD after beginning HD (patients with a history of cerebrovascular disease were excluded). Therefore, our results seemed to be underestimate the true incidence of cerebrovascular disease, especially CI, during maintenance HD treatment. More than two thirds of the CH group died within 3 months of the onset of CVE in the current study. The higher stroke-related death rate in HD patients may have been related to the large size of hemorrhagic lesions and the high incidence (73%) of the rupture into the ventricles. The size of hematomas was positively correlated with the SBP at the start of HD, as it is in the general population.17 Anticoagulation therapy with heparin during HD can lead to a bleeding tendency and may promote enlargement of cerebral lesions. Onoyama et al18 previously reported that the frequency of lobar hemorrhage in HD patients was three times that of the general population because of heparin-induced
995
bleeding in HD patients. However, lobar hemorrhage occurred in only one patient in the current study, and most hemorrhages occurred more than 6 hours after the end of the previous HD session. There was no difference in the dose of heparin, the size of hematomas, or the frequency of intraventricular hemorrhage between patients with early and late onset of CH after a previous HD session. These findings suggest that factors other than anticoagulation therapy may have contributed to the high mortality rates in the CH group. Patients who developed CH early after initiation of HD were older and had a higher ECG voltage in the left ventricle than patients who developed CH at a later stage. Diabetic nephropathy and nephrosclerosis were the most common renal diseases in patients with early-onset CH. These findings suggest that long-standing medical conditions before the initiation of renal replacement therapy may be more important contributors to stroke than uremia itself. However, 11 of 23 patients who developed CH more than 36 months after the start of HD had chronic glomerulonephritis; these patients were significantly younger than patients with early-onset CH. It is still controversial whether uremia or HD itself contributes to the progression of atherosclerosis.1,5,19 The current findings do not rule out the possibility that uremia or HD may contribute to an increased risk of CH and stroke-related death in patients receiving long-term HD. The SBP and DBP before and after HD were significantly higher, and ECG evidence of left ventricular hypertrophy was more frequent in the CH group than in the control group in the current study. Elevated BP and left ventricular hypertrophy are also risk factors for CH in the general population.11-13 Hemorrhage was most common in the putamen and the thalamus, which are sites vulnerable to hypertensive cerebral hemorrhage in general.20-22 These findings suggest that poor control of hypertension may accelerate the progression of preexisting atherosclerotic vascular changes to angionecrosis of the cerebral arteries, resulting in CH in HD patients, as well as in the general population. Decreased serum levels of albumin and cholesterol are risk factors for survival in the HD patients,8,23-25 but these levels were not decreased in the current study. However, the Kt/V, which is correlated with the survival rate,26-28 was significantly lower in the
996
KAWAMURA ET AL
CH group. The current findings indicate that the prescribed amount of HD may be one of the important factors related to the incidence of stroke and the survival rate in HD patients. In conclusion, patients undergoing maintenance HD had a high incidence of CH, which was associated with a poor prognosis. We could not find out the effect of the anticoagulation with heparin on the accident of CH. One of the most important and reversible risk factors related to the occurrence of CH was hypertension. The amount of HD prescription also may be important. REFERENCES 1. Lindner A, Charra B, Sherrard DJ, Scribner BH: Accelerated atherosclerosis in prolonged maintenance hemodialysis. N Engl J Med 290:697-701, 1974 2. Lazarus JM, Lowrie EG, Hampers CL, Merrill JP: Cardiovascular disease in uremic patients on hemodialysis. Kidney Int 7:167-175, 1975 (suppl 2) 3. Degoulet P, Legraim M, Re´ach I, Aime´ F, Devrie`s C, Rojas P, Jacobs C: Mortality risk factors in patients treated by chronic hemodialysis. Nephron 31:103-110, 1982 4. Iseki K, Kawazoe N, Osawa A, Fukiyama K: Survival analysis of dialysis patients in Okinawa, Japan. Kidney Int 43:404-409, 1993 5. Said G: Neurological aspects of dialysis patients, in Jacobs C (ed): Replacement of Renal Function by Dialysis, chap 50. Dordrecht, the Netherlands, Kluwer Academic Publishers, 1996, pp 1243-1259 6. Japanese Dialysis Organization: Annual Report on Dialysis in Japan. 1994 (in Japanese) 7. Onoyama K, Kumagai H, Miishima T, Tsuruda H, Tomooka S, Motomura K, Fujishima M: Incidence of strokes and its prognosis in patients on maintenance hemodialysis. Jpn Heart J 27:686-691, 1986 8. Iseki K, Kinjo K, Kimura Y, Osawa A, Fukiyama K: Evidence for high risk of cerebral hemorrhage in chronic dialysis patients. Kidney Int 44:1086-1090, 1993 9. Helweg-Larsen S, Sommer W, Strange P, Lester J, Boysen G: Prognosis for patients treated conservatively for spontaneous intracerebral hematomas. Stroke 15:10451048, 1984 10. United States Renal Data System: USRDS 1994 annual data report. Am J Kidney Dis 24:88-95, 1994 (suppl 2) 11. Qizilbach N, Lewingtorr S, Duffy S, Peto R: Cholesterol, diastolic blood pressure, and stroke: 13,000 strokes in 450,000 people in 45 prospective cohorts. Lancet 346:16471653, 1995 12. Håheim LL, Holme I, Hijermann I, Leren P: Risk factor of stroke incidence and mortality: A 12-year followup of the Oslo Study. Stroke 24:1484-1489, 1993
13. Omae T, Ueda K: Hypertension and cerebrovascular disease: The Japanese experience. J Hypertens 6:343-349, 1988 14. Kaplan NM: Clinical Hypertension (ed 6). Baltimore, MD, Williams & Wilkins, 1994, pp 109-144 15. Kameyama M, Sakaguchi K, Nishinaka K, Tsujimura T, Udaka F, Fushimi H: Cerebrovascular disease in patients under chronic hemodialysis. J Jpn Soc Dial Ther 27:13631371, 1994 (in Japanese) 16. Hirakata H, Yano H, Osato S, Ibayashi S, Onoyama K, Otsuka M, Ichiya Y, Kuwabara Y, Masuda Y, Fujishima M: CBF and oxygen metabolism in hemodialysis patients: Effects of anemia correction with recombinant human EPO. Am J Physiol 262:737-743, 1992 17. Strandgaard S, Paulson OB: Hypertension and human cerebrovascular disease, in Swales JD (ed): Textbook of Hypertension, chap 36. London, England, Blackwell, 1994, pp 690-697 18. Onoyama K, Ibayashi S, Nanishi F, Okuda S, Oh Y, Hirakata H, Nishimura Y, Fujishima M: Cerebral hemorrhage in patients on maintenance hemodialysis. CT analysis of 25 cases. Eur Neurol 26:171-175, 1987 19. Lundin AP, Friedman EA: Vascular consequences of maintenance hemodialysis: An unproven case. Nephron 21: 177-180, 1978 20. Kase CS, Williams JP, Wyatt DA, Mohr JP: Lobar intracerebral hematomas: Clinical and CT analysis of 22 cases. Neurology 32:1146-1151, 1982 21. Wiggins WS, Moody DM, Toole JF, Laster DW, Ball MR: Clinical and computerized tomographic study of hypertensive intracerebral hemorrhage. Arch Neurol 35:832-833, 1978 22. Weisberg LA: Computerized tomography in intracranial hemorrhage. Arch Neurol 36:422-426, 1979 23. Zusman CJ, Cohen JJ, Harrington JT, Madias NE: Morbidity and mortality in dialysis patients. Kidney Int 46:1728-1737, 1994 24. Avram MM, Mittman N, Bonomini L, Chattopadhyay J, Fein P: Markers for survival in dialysis: A seven-year prospective study. Am J Kidney Dis 26:209-219, 1995 25. Iseki K, Kawazoe N, Fukiyama K: Serum albumin is a strong predictor of death in chronic dialysis patients. Kidney Int 44:115-119, 1993 26. Owen WF, Lew NL, Liu Y, Lazarus JM: The urea reduction ratio and serum albumin concentration as predictor of mortality in patients undergoing hemodialysis. N Engl J Med 329:1001-1006, 1993 27. Held PJ, Port FK, Wolfe RA, Stannard DC, Carroll CE, Daugirdas JT, Bloembergen WE, Greer JW, Hakim RM: The dose of hemodialysis and patient mortality. Kidney Int 50:550-556, 1996 28. Bloembergen WE, Stannard DC, Port FK, Wolfe RA, Pugh JA, Jones CA, Greer JW, Golper TA, Held PJ: Relationship of dose of hemodialysis and cause-specific mortality. Kidney Int 50:557-565, 1996
C
EREBROVASCULAR DISEASES and cardiovascular diseases are important causes of death in patients receiving maintenance hemodialysis (HD).1-5 The number of patients receiving HD has been increasing, and HD is being initiated increasingly in older patients in Japan, where HD is used in more than 90% of patients with end-stage renal disease.6 Comorbid conditions, such as hypertension and hyperlipidemia, which are risk factors for cerebrovascular disease, are commonly observed in HD patients. There is a higher prevalence of cerebral hemorrhage (CH) than cerebral infarction (CI) in HD patients,7,8 unlike in the general population. However, only a few studies have examined the characteristics of HD patients who develop cerebrovascular events (CVE), and no studies have compared the characteristics of CH patients with those of matched controls on maintenance HD. In Miyazaki in the Kyushu area of southwestern Japan, HD patients who develop CVE are usually transferred to one of only a few regional hospitals, where a brain computed tomography (CT) is performed. Because the movement of HD patients is less frequent in this area, we are able to survey the follow-up data of CVE pa-
tients on HD with relative ease and completely. We retrospectively investigated the clinical profiles of HD patients who developed CH in 17 dialysis centers in Miyazaki, Japan. PATIENTS AND METHODS
Entry Criteria and Definitions Of 1,064 patients who received maintenance HD (4 or 5 hours, 3 times weekly) for more than 3 months at 17 dialysis centers in Miyazaki, Japan, between January 1971 and December 1994, 98 patients (9.2%) developed a first attack of CVE after beginning HD. CVE was diagnosed based on CT findings or clinical signs, such as the sudden onset of disturbances in consciousness, speech, visual, motor, or sensory functions. Patients with a history of cerebrovascular disease, including transient ischemic attack, and accidental brain CT findings compatible with cerebrovascular disease were excluded. Of the 98 patients,
From the First Department of Internal Medicine, Miyazaki Medical College, Miyazaki, Japan; and the Yamamoto Clinic, Miyazaki, Japan. Received March 17, 1997; accepted December 12, 1997. Address reprint requests to Tanenao Eto, MD, First Department of Internal Medicine, Miyazaki Medical College, 5200 Kihara, Kiyotake, Miyazaki, Japan 889-16. E-mail: [email protected]
r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3106-0013$3.00/0
American Journal of Kidney Diseases, Vol 31, No 6 (June), 1998: pp 991-996
991
992
KAWAMURA ET AL
13 patients were excluded because they did not undergo CT at the onset of CVE. We first investigated the demographic and clinical characteristics of these 85 patients in terms of the CVE subtype. We further analyzed clinical features of the 40 patients with CH for whom clinical data, including medical records, laboratory data, and the brain CT films at the onset of disease, were available. Finally, to identify factors that contributed to the development of CH, the characteristics of 35 of these patients with CH were compared with characteristics of a control group consisting of 66 patients from the original group of 1,064 HD patients who had no clinical evidence of CVE and who were matched in age, sex, basal renal disease, age at the initiation of HD, and the duration of HD.
Data Collection Medical and clinical records were collected by one investigator. Laboratory data, chest x-ray films, electrocardiograms (ECGs), and supine blood pressure were obtained before and after HD 3 to 4 months before the first attack of CVE. The cardiothoracic ratio (CTR) and the voltage of SV1 plus RV5 in the precordial lead were determined from x-ray and ECGs obtained before HD. The Kt/V was calculated as follows: Kt/V ⫽ ⫺1n (R ⫺ 0.03 ⫺ 0.75 ⫻ UF/W), where R is the ratio of the postdialysis to the predialysis (post/pre) level of urea nitrogen, UF is the volume removed by ultrafiltration in liters, and W is the postdialysis weight in kilograms. The size of hematomas in patients with CH was determined from CTs obtained at the onset as follows: the maximum transverse diameter ⫻ the maximal anteroposterior diameter ⫻ the maximal superoinferior diameter ⫻ 1⁄2.9
Statistical Analysis Statistical analyses were performed using the Macintosh StatView program (Abacus Concepts Inc., Berkeley, CA, 1992). Differences between groups were evaluated by analysis of variance, an unpaired t-test, the chi-square test, and by determining correlation coefficients for parametric data. Results are expressed as the mean ⫾ standard deviation (SD). A P value ⬍ 0.05 was accepted as indicating statistical significance.
RESULTS
Clinical Background CVE was diagnosed in 85 (8.0%) of 1,064 patients (Table 1). They included 56 patients with CH, 24 with CI, and five with subarachnoid hemorrhage. The definite incidence of CVE was 13.2 per 1,000 patient-years. The incidences of CH, CI, and subarachnoid hemorrhage were 8.7, 3.7, and 0.8 per 1,000 patient-years, respectively. Male sex predominated in the overall HD patients and in the CVE group. Patients in the CH group were significantly younger at initiation of HD and at the last observation compared with patients in the CI group. The mean age was greater in the CI group than in the total population. Although the duration of HD did not differ among groups, the time from the initiation of HD
Table 1. Clinical Characteristics of HD Patients CVE
Number Sex, male Age at start of HD (yr) Duration of HD (mo) Age at last observation (yr) Basal renal disease Chronic glomerulonephritis Diabetes mellitus Nephrosclerosis Polycystic kidney disease Others Number of deaths
Total
Total
CH
CI
SAH
1,064 634 (59.6%) 53.1 ⫾ 15.5 72.4 ⫾ 60.0 59.1 ⫾ 13.9
85 52 (61.2%) 53.8 ⫾ 13.6 70.6 ⫾ 59.6 59.7 ⫾ 11.7
56 36 (64.3%) 50.3 ⫾ 13.1* 79.7 ⫾ 66.6 56.8 ⫾ 11.3*
24 14 (58.3%) 61.5 ⫾ 12.6† 56.0 ⫾ 40.3 66.1 ⫾ 11.3‡
5 2 (40.0%) 57.4 ⫾ 7.6 39.9 ⫾ 27.7 60.7 ⫾ 6.1
576 (54.1%) 187 (17.6%) 108 (10.2%) 42 (3.9%) 151 (14.2%) 297 (27.9%)
35 (41.2%) 19 (22.4%) 9 (10.6%) 8 (9.4%)§ 14 (16.4%) 54 (63.5%)¶
25 (44.6%) 10 (17.9%) 4 (7.1%) 7 (12.5%)\ 10 (17.9%)§ 40 (71.4%)¶
6 (25.0%)\ 9 (37.5%)§ 5 (20.8%) 0 (0%) 4 (16.7%) 12 (50.0%)
4 (80.0%) 0 (0%) 0 (0%) 1 (20.0%) 0 (0%) 2 (40.0%)
Abbreviations: CVE, cerebrovascular events; CH, cerebral hemorrhage; CI, cerebral infarction; SAH, subarachnoid hemorrhage; HD, hemodialysis. *P ⬍ 0.01 v CI by ANOVA. †P ⬍ 0.01 v all patients by t-test. ‡P ⬍ 0.05 v all patients by t-test. §P ⬍ 0.05 v all patients by chi-square test. \P ⬍ 0.01 v all patients by chi-square test. ¶P ⬍ 0.001 v all patients by chi-square test.
CEREBROVASCULAR EVENTS IN HD PATIENTS
to the first attack of CVE was shorter in the CI group than in the CH group (39.4 ⫾ 34.4 v 63.5 ⫾ 57.5 months, P ⬍ 0.05). The incidence of polycystic kidney disease was significantly higher in the CH group and the total CVE group than in all maintenance HD patients. The incidence of chronic glomerulonephritis was higher in the CH group than in the CI group, and the incidences of diabetes mellitus and nephrosclerosis were higher in the CI group, but the differences were not significant. There were significantly more deaths in the CH group and the total CVE group than in the total HD group. The cause of death in the total HD group was CVE in 57 patients (19.2%), infection in 40 patients (13.5%), congestive heart failure (CHF) in 38 patients (12.8%), malignancy in 33 patients (11.1%), myocardial infarction in 13 patients (4.4%), and other causes in 116 patients (39.0%). CVE was the cause of death in 35 (87.5%) of 40 patients in the CH group. The causes of death in the CI group were CVE in two patients (16.7%), infection in three patients (25.0%), and CHF in two patients (16.7%). Comparison of CH and CI Groups Patients in the CH group were significantly younger at the onset of CVE than patients in the CI group (Table 2). The survival period after the onset of CVE was shorter in the CH group than in the CI group; the mortality rate of CVE tended to be higher in the CH group (71.4%) than in the Table 2. Characteristics of Cerebral Hemorrhage (CH) and Cerebral Infarction (CI) Groups
Age at onset of CVE (yr) Survival after the onset of CVE (mo) Pre-HD SBP (mm Hg) Pre-HD DBP (mm Hg) Post-HD SBP (mm Hg) Post-HD DBP (mm Hg) Albumin (g/dL) High-density lipoprotein cholesterol (mg/dL)
CH (56)
CI (24)
55.9 ⫾ 11.4*
64.7 ⫾ 11.4
6.4 ⫾ 12.4* 172 ⫾ 21.6† 90 ⫾ 13.4† 158 ⫾ 24.0* 86 ⫾ 11.3* 4.08 ⫾ 0.40†
17.9 ⫾ 16.4 160 ⫾ 24.5 77 ⫾ 8.3 142 ⫾ 19.9 73 ⫾ 9.3 3.80 ⫾ 0.45
40.3 ⫾ 12.1†
33.0 ⫾ 9.5
Abbreviations: HD, hemodialysis; CVE, cerebrovascular events; SBP, systolic blood pressure; DBP, diastolic blood pressure. *P ⬍ 0.01 v CI (t-test). †P ⬍ 0.05 v CI (t-test).
993
CI group (50.0%). The systolic blood pressure (SBP) and diastolic blood pressure (DBP) before and after HD were significantly higher in the CH group than in the CI group. Actually, 89% (50 of 56) and 67% (16 of 24) of patients were prescribed antihypertensive drugs in the CH and the CI groups, respectively. Serum levels of albumin and high-density lipoprotein cholesterol were significantly lower in the CI group, but there were no significant differences in the CTR, the ultrafiltration volume, the amount of heparin administered, or other laboratory parameters, such as levels of total cholesterol, triglycerides, and production of calcium and phosphorus (data not shown). Clinical Features in the CH Group In the group of 40 CH patients for whom all clinical data were available, males accounted for 26 patients (65.0%) and for 17 (65.4%) of 26 patients who died within 3 months of the onset of CH. Hemorrhagic lesions occurred in the ganglionic-thalamus in 32 (80.0%) of 40 patients; one patient experienced a lobar hemorrhage, and none experienced cerebellar hemorrhage. To examine the differences in the clinical characteristics by the dialysis periods at the onset of CVE, we classified CH patients according to the time from the initiation of HD to the first attack of CVE, as follows: ⱕ36 months (group A, n ⫽ 17) and ⬎36 months (group B, n ⫽ 23). Patients in group A were significantly older than patients in group B (57.7 ⫾ 8.4 v 45.6 ⫾ 12.6 years, P ⬍ 0.01), and the ECG voltage of the left ventricle (Sv1 ⫹ Rv5 mV) was higher in group A than in group B (4.70 ⫾ 1.51 v 3.58 ⫾ 1.37 mV, P ⬍ 0.05). There were no differences between groups in SBP, DBP, and Kt/V. All patients with nephrosclerosis and half of those with diabetes mellitus belonged to group A, whereas most of those with chronic glomerulonephritis were in group B. There were no significant differences between CH patients who survived and those who died in clinical characteristics, such as the SBP/DBP (167 ⫾ 19.3/86 ⫾ 13.7 mm Hg v 175 ⫾ 22.6/ 92 ⫾ 13.0 mm Hg), the dose of standard heparin (unfractionated) (4,414 ⫾ 1,311 v 4,379 ⫾ 1,005 unit/HD), onset time of CH after a previous HD session (16.0 ⫾ 14.2 v 25.0 ⫾ 13.9 hours), the serum level of albumin (4.25 ⫾ 0.37 v
994
4.00 ⫾ 0.40 g/dL), and the Kt/V (1.32 ⫾ 0.36 v 1.15 ⫾ 0.21). Hematomas were lager in patients who died (170.6 ⫾ 119.3 cm3 ) than in survivors (26.7 ⫾ 32.8 cm3, P ⬍ 0.001), and the size of the hematoma was correlated with the SBP at the start of HD (r ⫽ 0.504, P ⬍ 0.05) (Fig 1). Patients without intraventricular hemorrhage had a better survival rate. We studied the effects of anticoagulation therapy during HD on the occurrence of CH. CH occurred more than 6 hours after the end of an HD session in 34 (85%) of 40 patients and 6 hours or less after HD in six patients. In the early-onset group, three of six patients had a CVE during an HD session. There were no significant differences in the dose of heparin, the size of the hematoma, or the blood pressure (BP) between groups. Survival was prolonged (12.2 ⫾ 15.3 v 7.0 ⫾ 11.4 months, P ⬍ 0.01), and the age at the onset of CVE was younger (46.2 ⫾ 10.7 v 57.8 ⫾ 10.2 years, P ⬍ 0.05) in the early-onset group than in the late-onset group. Differences in Clinical Parameters Between CH and Control Groups There were no significant differences in patient characteristics between the CH (n ⫽ 35) and control (n ⫽ 66) groups. Histories of cerebrovascular diseases, and of coronary or peripheral vascular diseases before beginning HD, were documented in zero, nine, and five patients in the CH group, and five, three, and two patients in the control group. The number of survivors at the last observation was significantly lower in the
KAWAMURA ET AL Table 3. Clinical, Laboratory, and ECG Data in CH and Control Groups CH
Control
Pre-HD SBP (mm Hg) 171 ⫾ 22.5 154 ⫾ 19.3* Pre-HD DBP (mm Hg) 89 ⫾ 13.6 81 ⫾ 9.6* Post-HD SBP (mm Hg) 157 ⫾ 25.5 140 ⫾ 24.8* Post-HD DBP (mm Hg) 86 ⫾ 11.4 76 ⫾ 10.7* CTR (%) 51.9 ⫾ 6.0 50.8 ⫾ 6.2 ECG, Sv1 ⫹ Rv5 (mV) 4.12 ⫾ 1.55 3.19 ⫾ 1.29† Ultrafiltration (L/HD) 2.41 ⫾ 1.16 2.84 ⫾ 1.01 Kt/V 1.23 ⫾ 0.26 1.38 ⫾ 0.34‡ Heparin (unit/HD) 4,290 ⫾ 1,036 4,090 ⫾ 1,453 Total cholesterol (mg/dL) 163 ⫾ 29.3 164 ⫾ 36.2 Triglycerides (mg/dL) 135 ⫾ 72.3 123 ⫾ 63.0 High-density lipoprotein cholesterol (mg/dL) 40.6 ⫾ 12.5 42.0 ⫾ 14.0 Albumin (g/dL) 4.08 ⫾ 0.40 4.10 ⫾ 0.41 Ca ⫻ Pi 53.9 ⫾ 16.6 56.6 ⫾ 17.1 Abbreviations: CH, cerebral hemorrhage; HD, hemodialysis; SBP, systolic blood pressure; DBP, diastolic blood pressure; CTR, cardiothoracic ratio; ECG, electrocardiogram; Ca ⫻ Pi, production of calcium and phosphorus. *P ⬍ 0.001 (t-test). †P ⬍ 0.01 (t-test). ‡P ⬍ 0.05 (t-test).
CH group (21 of 26 patients died of CVE) compared with the control group. The SBP and DBP before and after HD were significantly higher in the CH group than in the control group (Table 3). The BP values just before CVE were significantly higher than those obtained 3 to 4 months before CVE (pre-HD SBP/DBP, 180 ⫾ 25.8/94 ⫾ 15.7 mm Hg v 171 ⫾ 22.5/89 ⫾ 13.6 mm Hg, P ⬍ 0.01). Conversely, the means of long-term BP values before CVE were not different from those of 3 to 4 months before CVE (pre-HD SBP/DBP, 168 ⫾ 22.0/88 ⫾ 13.1 mm Hg v 171 ⫾ 22.5/ 89 ⫾ 13.6 mm Hg, NS). The ECG voltage was also significantly higher in the CH group. There were no differences between groups in the ultrafiltration volume and the total dose of heparin per HD session. The Kt/V was lower in the CH group than in the control group. DISCUSSION
Fig 1. Correlation between SBP and size of hemotoma in CH patients.
Cerebrovascular disease was the most common cause of deaths among the 1,064 HD patients in the current study, accounting for 19.2% of all deaths. The CVE-related mortality rate in chronic dialysis patients is 12.7% in Japan6 and 4.8% in the United States.10 By the causes of
CEREBROVASCULAR EVENTS IN HD PATIENTS
differences in human species or nutritional enviroments, and so forth, as well as general population, the CVE-related mortality rate in chronic dialysis patients may be higher in Japan than in the United States. The higher frequency of HD as the treatment modalities for the patients with end-stage renal disease also may contribute to the higher mortality rate due to CVE in Japan. The incidence of confirmed stroke events during maintenance HD per 1,000 patient-years in the current study was 8.7 in the CH group, 3.7 in the CI group, and 0.8 in the subarachnoid hemorrhage group. The incidences of CVE and, especially, of CH were higher in the current study than in the general population, in which the reported incidences of stroke and CH are 0.95 and 0.1 to 2.2 per 1,000 patient-years.12,13 Approximately 75% of strokes in the general population are caused by CI.14 The current results confirm findings of previous studies,3,4 which also found a higher prevalence of CH among dialysis patients. These findings may be related to the high prevalence of hypertension in the dialysis patients, which is established as the risk factor for the occurrence of CVE in the general population. However, asymptomatic CI is frequently detected by CT and magnetic resonance imaging,15 and decreased cerebral blood flow has been detected by positron emission tomography in HD patients.16 Furthermore, in this study, we examined the definite cases with a first attack of CVD after beginning HD (patients with a history of cerebrovascular disease were excluded). Therefore, our results seemed to be underestimate the true incidence of cerebrovascular disease, especially CI, during maintenance HD treatment. More than two thirds of the CH group died within 3 months of the onset of CVE in the current study. The higher stroke-related death rate in HD patients may have been related to the large size of hemorrhagic lesions and the high incidence (73%) of the rupture into the ventricles. The size of hematomas was positively correlated with the SBP at the start of HD, as it is in the general population.17 Anticoagulation therapy with heparin during HD can lead to a bleeding tendency and may promote enlargement of cerebral lesions. Onoyama et al18 previously reported that the frequency of lobar hemorrhage in HD patients was three times that of the general population because of heparin-induced
995
bleeding in HD patients. However, lobar hemorrhage occurred in only one patient in the current study, and most hemorrhages occurred more than 6 hours after the end of the previous HD session. There was no difference in the dose of heparin, the size of hematomas, or the frequency of intraventricular hemorrhage between patients with early and late onset of CH after a previous HD session. These findings suggest that factors other than anticoagulation therapy may have contributed to the high mortality rates in the CH group. Patients who developed CH early after initiation of HD were older and had a higher ECG voltage in the left ventricle than patients who developed CH at a later stage. Diabetic nephropathy and nephrosclerosis were the most common renal diseases in patients with early-onset CH. These findings suggest that long-standing medical conditions before the initiation of renal replacement therapy may be more important contributors to stroke than uremia itself. However, 11 of 23 patients who developed CH more than 36 months after the start of HD had chronic glomerulonephritis; these patients were significantly younger than patients with early-onset CH. It is still controversial whether uremia or HD itself contributes to the progression of atherosclerosis.1,5,19 The current findings do not rule out the possibility that uremia or HD may contribute to an increased risk of CH and stroke-related death in patients receiving long-term HD. The SBP and DBP before and after HD were significantly higher, and ECG evidence of left ventricular hypertrophy was more frequent in the CH group than in the control group in the current study. Elevated BP and left ventricular hypertrophy are also risk factors for CH in the general population.11-13 Hemorrhage was most common in the putamen and the thalamus, which are sites vulnerable to hypertensive cerebral hemorrhage in general.20-22 These findings suggest that poor control of hypertension may accelerate the progression of preexisting atherosclerotic vascular changes to angionecrosis of the cerebral arteries, resulting in CH in HD patients, as well as in the general population. Decreased serum levels of albumin and cholesterol are risk factors for survival in the HD patients,8,23-25 but these levels were not decreased in the current study. However, the Kt/V, which is correlated with the survival rate,26-28 was significantly lower in the
996
KAWAMURA ET AL
CH group. The current findings indicate that the prescribed amount of HD may be one of the important factors related to the incidence of stroke and the survival rate in HD patients. In conclusion, patients undergoing maintenance HD had a high incidence of CH, which was associated with a poor prognosis. We could not find out the effect of the anticoagulation with heparin on the accident of CH. One of the most important and reversible risk factors related to the occurrence of CH was hypertension. The amount of HD prescription also may be important. REFERENCES 1. Lindner A, Charra B, Sherrard DJ, Scribner BH: Accelerated atherosclerosis in prolonged maintenance hemodialysis. N Engl J Med 290:697-701, 1974 2. Lazarus JM, Lowrie EG, Hampers CL, Merrill JP: Cardiovascular disease in uremic patients on hemodialysis. Kidney Int 7:167-175, 1975 (suppl 2) 3. Degoulet P, Legraim M, Re´ach I, Aime´ F, Devrie`s C, Rojas P, Jacobs C: Mortality risk factors in patients treated by chronic hemodialysis. Nephron 31:103-110, 1982 4. Iseki K, Kawazoe N, Osawa A, Fukiyama K: Survival analysis of dialysis patients in Okinawa, Japan. Kidney Int 43:404-409, 1993 5. Said G: Neurological aspects of dialysis patients, in Jacobs C (ed): Replacement of Renal Function by Dialysis, chap 50. Dordrecht, the Netherlands, Kluwer Academic Publishers, 1996, pp 1243-1259 6. Japanese Dialysis Organization: Annual Report on Dialysis in Japan. 1994 (in Japanese) 7. Onoyama K, Kumagai H, Miishima T, Tsuruda H, Tomooka S, Motomura K, Fujishima M: Incidence of strokes and its prognosis in patients on maintenance hemodialysis. Jpn Heart J 27:686-691, 1986 8. Iseki K, Kinjo K, Kimura Y, Osawa A, Fukiyama K: Evidence for high risk of cerebral hemorrhage in chronic dialysis patients. Kidney Int 44:1086-1090, 1993 9. Helweg-Larsen S, Sommer W, Strange P, Lester J, Boysen G: Prognosis for patients treated conservatively for spontaneous intracerebral hematomas. Stroke 15:10451048, 1984 10. United States Renal Data System: USRDS 1994 annual data report. Am J Kidney Dis 24:88-95, 1994 (suppl 2) 11. Qizilbach N, Lewingtorr S, Duffy S, Peto R: Cholesterol, diastolic blood pressure, and stroke: 13,000 strokes in 450,000 people in 45 prospective cohorts. Lancet 346:16471653, 1995 12. Håheim LL, Holme I, Hijermann I, Leren P: Risk factor of stroke incidence and mortality: A 12-year followup of the Oslo Study. Stroke 24:1484-1489, 1993
13. Omae T, Ueda K: Hypertension and cerebrovascular disease: The Japanese experience. J Hypertens 6:343-349, 1988 14. Kaplan NM: Clinical Hypertension (ed 6). Baltimore, MD, Williams & Wilkins, 1994, pp 109-144 15. Kameyama M, Sakaguchi K, Nishinaka K, Tsujimura T, Udaka F, Fushimi H: Cerebrovascular disease in patients under chronic hemodialysis. J Jpn Soc Dial Ther 27:13631371, 1994 (in Japanese) 16. Hirakata H, Yano H, Osato S, Ibayashi S, Onoyama K, Otsuka M, Ichiya Y, Kuwabara Y, Masuda Y, Fujishima M: CBF and oxygen metabolism in hemodialysis patients: Effects of anemia correction with recombinant human EPO. Am J Physiol 262:737-743, 1992 17. Strandgaard S, Paulson OB: Hypertension and human cerebrovascular disease, in Swales JD (ed): Textbook of Hypertension, chap 36. London, England, Blackwell, 1994, pp 690-697 18. Onoyama K, Ibayashi S, Nanishi F, Okuda S, Oh Y, Hirakata H, Nishimura Y, Fujishima M: Cerebral hemorrhage in patients on maintenance hemodialysis. CT analysis of 25 cases. Eur Neurol 26:171-175, 1987 19. Lundin AP, Friedman EA: Vascular consequences of maintenance hemodialysis: An unproven case. Nephron 21: 177-180, 1978 20. Kase CS, Williams JP, Wyatt DA, Mohr JP: Lobar intracerebral hematomas: Clinical and CT analysis of 22 cases. Neurology 32:1146-1151, 1982 21. Wiggins WS, Moody DM, Toole JF, Laster DW, Ball MR: Clinical and computerized tomographic study of hypertensive intracerebral hemorrhage. Arch Neurol 35:832-833, 1978 22. Weisberg LA: Computerized tomography in intracranial hemorrhage. Arch Neurol 36:422-426, 1979 23. Zusman CJ, Cohen JJ, Harrington JT, Madias NE: Morbidity and mortality in dialysis patients. Kidney Int 46:1728-1737, 1994 24. Avram MM, Mittman N, Bonomini L, Chattopadhyay J, Fein P: Markers for survival in dialysis: A seven-year prospective study. Am J Kidney Dis 26:209-219, 1995 25. Iseki K, Kawazoe N, Fukiyama K: Serum albumin is a strong predictor of death in chronic dialysis patients. Kidney Int 44:115-119, 1993 26. Owen WF, Lew NL, Liu Y, Lazarus JM: The urea reduction ratio and serum albumin concentration as predictor of mortality in patients undergoing hemodialysis. N Engl J Med 329:1001-1006, 1993 27. Held PJ, Port FK, Wolfe RA, Stannard DC, Carroll CE, Daugirdas JT, Bloembergen WE, Greer JW, Hakim RM: The dose of hemodialysis and patient mortality. Kidney Int 50:550-556, 1996 28. Bloembergen WE, Stannard DC, Port FK, Wolfe RA, Pugh JA, Jones CA, Greer JW, Golper TA, Held PJ: Relationship of dose of hemodialysis and cause-specific mortality. Kidney Int 50:557-565, 1996