Clin Exp Nephrol (2002) 6:140–146
© Japanese Society of Nephrology 2002
ORIGINAL ARTICLE Tsuneyoshi Oh · Osamu Sakayori · Chisako Kamano Yuichi Komaba · Yasuhiko Iino · Yasuo Katayama
Optimal hematocrit based on regional cerebral blood flow in hemodialysis patients with diabetic nephropathy
Received: May 30, 2001 / Accepted: May 31, 2002
Abstract Background. The optimal hematocrit (Hctopt) in hemodialysis (HD) patients has yet to be determined based on the etiology and complications of their endstage renal disease (ESRD). To investigate this problem, we compared regional cerebral oxygen supply (rCOS) in diabetic (DM group) and non-diabetic HD patients (non-DM group) with data from subjects without renal disease or DM (control group). Methods. Regional cerebral blood flow (rCBF) was measured with single-photon emission computed tomography (SPECT) by the N-isopropyl-p-[123I]-iodoamphetamine (123I-IMP)-autoradiographic (ARG) method, and both the O2 content (O2CT) of arterial blood and hematocrit (Hct) were evaluated. Using the regression lines of rCBF vs Hct and O2CT vs Hct, we established a convex curve between rCOS and Hct. The peak of the curve indicates the maximum rCOS (rCOSmax) and Hctopt for rCOSmax. Results. The rCBF in both the DM and non-DM groups was lower than that of the control group at the same Hct level, and the DM group had the lowest values. The rCOSmax values in the DM and non-DM groups were nearly equal, but both were lower than in controls. The Hctopt in the DM group was lower than that in the non-DM group by 6.3% ⫾ 3.3%. Conclusions. Although the difference in Hctopt values in the DM and non-DM groups was 6.3%, the rCOSmax values in both groups were nearly equal. This suggests that differences in the Hctopt may depend on complications or causes of ESRD. The optimal Hct in the DM group was 22.6% ⫾ 1.9%, and that for the non-DM group was 29.0% ⫾ 1.8%. Key words Optimal hematocrit (Hctopt) · Hemodialysis (HD) · Diabetes mellitus (DM) · Single-photon emis-
T. Oh (*) · O. Sakayori · C. Kamano · Y. Komaba · Y. Iino · Y. Katayama Second Department of Internal Medicine, Nippon Medical School, 1-1-5 Sendagi, Bunkyo-ku, Tokyo 113-8603, Japan Tel. ⫹81-3-3822-2131 (ext. 6496); Fax ⫹81-3-3822-4865 e-mail:
[email protected] sion computed tomography (SPECT) · Regional cerebral blood flow (rCBF) · Regional cerebral oxygen supply (rCOS)
Introduction The treatment of anemia due to chronic renal failure (CRF) is important for improving the level of energy and quality of life (QOL) of hemodialysis (HD) patients. A low hematocrit exacerbates the hypoxic conditions associated with heart diseases such as chronic heart failure or left ventricular hypertrophy. Hypoxia also obstructs cognitive function and brain metabolism, which consumes about 20% of whole body oxygen supply.1–6 Some studies of the optimal hematocrit (Hctopt) have been reported.3,7–9 The Hctopt proposed by the Working Group of the Japanese Ministry of Welfare10 is 30%, while the guidelines of the National Kidney Foundation (NKF)11 propose 33% to 36%. Recombinant human erythropoietin (rHuEPO) is remarkably successful in treating anemia due to CRF, but correcting the Hct to near normal levels (42%) using rHuEPO may induce ischemia of the brain or heart, or occlusion of vascular access sites.1,4,12–15 Cerebrovascular ischemia is one of the most common causes of death in HD patients,16 and this suggests the importance of determining the Hctopt of HD patients. Most patients undergoing HD have complications, such as hypertension and hyperlipidemia, which induce progression of arteriosclerosis, and the severity of arteriosclerosis is linked to the duration of HD.17–19 Therefore, the Hctopt for each patient may also depend on the cause of the endstage renal disease (ESRD) and the dialysis history. The percentage of patients who started HD as a result of diabetic nephropathy increased from 15.6% in 1983 to 35.7% in 1998, and the percentage of all HD patients who were diabetic increased from 7.4% in 1983 to 24.0% in 1998. We compared regional cerebral blood flow (rCBF) and regional cerebral oxygen supply (rCOS) in HD patients with and without DM and compared these values with those in healthy controls to determine the Hctopt in HD patients.
141
All patients were HD patients with no history of cerebrovascular ischemia and negative magnetic resonance image (MRI) scanning; 11 patients had diabetic nephropathy (DM group) and 18 did not have DM (non-DM group). As a control group, 56 subjects without renal disease, DM, or cerebrovascular ischemia were selected (Table 1). There were no significant differences in age or duration of HD therapy between the DM and non-DM groups, and the prevalence of hypertension and hyperlipidemia did not differ. The objective and methods were explained to all subjects who were involved in this study, and informed consent was obtained from each patient before they underwent single-photon emission computed tomography (SPECT).
sampled from the opposite brachial artery 10 min after the start of injection of 123I-IMP and measured using a scintillation counter. At the same time, the oxygen content of the arterial blood (O2CT; ml·O2/dl) and Hct (%) were measured (Table 2). Scanning was initiated 25 min after the injection of 123I-IMP, and scanning time was 30 min. Images reflecting CBF were created by the SPECT ARG method with 68 circular regions of interest (ROI) placed in the CBF images (Fig. 1).21,22 Mean rCBF values in cerebral cortex, white matter, striatum, thalamus, brain stem, and cerebellum were obtained in each subject (Table 3). We established regression lines for rCBF vs Hct in these regions, and similar regression lines for O2CT vs Hct based on the measurements given above. Because rCOS (ml·O2/100 g per min) is expressed as the product of rCBF and O2CT (rCOS ⫽ rCBF ⫻ O2CT/100),3 we established a convex curve for rCOS vs Hct; the peak indicates the maximum rCOS (rCOSmax) and its corresponding Hctopt (Fig. 2).
Methods
Statistical analysis
We measured rCBF (ml/100 g per min), using SPECT with the N-isopropyl-p-[123I]-iodoamphetamine (123I-IMP)20 autoradiographic (ARG) method. HD patients were tested on the next day of HD. All subjects underwent SPECT in a resting state; the ears were not plugged, but the eyes were closed. The SPECT scanner was a HeadtomeSET080 (Shimadzu, Kyoto, Japan). Radioactivity was
Data values are presented as means ⫾ SD. Differences between groups were analyzed using a one-way analysis of variance (ANOVA). Differences with a P value of ⬍0.05 were considered significant. The differences in the prevalence of hypertension and hyperlipidemia between the DM and non-DM groups were analyzed with Fisher’s exact probability test.
Subjects and methods Patients and controls
Table 1. Patient demographics and clinical characteristics Group
Control group DM group Non-DM group
n
56 11 18
Age (years)
Duration of dialysis (months)
60.9 ⫾ 12.5 62.4 ⫾ 10.3 67.1 ⫾ 9.5
– 16.1 ⫾ 25.9 17.4 ⫾ 27.2
Complications
HbA1c
Hypertension
Hyperlipidemia
–
– 1 2
7 12
– 7.0 ⫾ 1.7 –
The differences in average age between each group were not significant, and differences in the average duration of dialysis and the prevalence of hypertension and hyperlipidemia between the diabetes mellitus (DM) and non-DM groups were not significant HbA1c, hemoglobin Alc
Table 2. Blood cell counts and arterial blood gas analysis Group
RBC (⫻104/µl)
Hb (g/dl)
Control group DM group Non-DM group
436.0 ⫾ 31.6 285.7 ⫾ 38.5* 304.0 ⫾ 48.4*
12.7 ⫾ 1.6 8.8 ⫾ 1.3* 9.0 ⫾ 1.0*
Group
pH
PCO2 (mmHg)
Control group DM group Non-DM group
7.41 ⫾ 0.03 7.39 ⫾ 0.06 7.40 ⫾ 0.04
41.5 ⫾ 5.6 37.9 ⫾ 3.8 38.1 ⫾ 3.4
Hct (%)
MCH (pg)
MCV (fl)
MCHC (pg/fl)
29.1 ⫾ 1.7 29.2 ⫾ 1.3* 30.4 ⫾ 1.6*
85.2 ⫾ 4.0 96.0 ⫾ 6.1* 97.9 ⫾ 5.2*
34.2 ⫾ 0.5 30.5 ⫾ 1.5* 31.2 ⫾ 2.4*
PO2 (mmHg)
HCO⫺3 (mmol/l)
O2SAT (%)
O2CT (ml·O2/dl)
106.2 ⫾ 17.3 105.6 ⫾ 19.5 111.0 ⫾ 18.1
25.5 ⫾ 1.6 22.5 ⫾ 3.0 23.3 ⫾ 2.9
97.6 ⫾ 1.1 97.8 ⫾ 0.7 98.0 ⫾ 0.8
17.6 ⫾ 2.2 11.3 ⫾ 1.6 12.5 ⫾ 2.0
37.2 ⫾ 4.3 28.6 ⫾ 3.6* 29.2 ⫾ 5.1*
Hb, Hemoglobin; Hct, Hematocrit; MCH, mean red blood cell hemoglobin; MCV, mean red blood cell volume; MCHC, mean red blood cell hemoglobin concentration; O2SAT, O2 saturation of arterial blood; O2CT, O2 content of arterial blood * P ⬍ 0.0001 vs control group
142 Table 3. Regional cerebral blood flow (rCBF) and regional cerebral oxygen supply (rCOS) in regions of the brain rCBF (ml/100 g per min)
Cerebral cortex White matter Striatum Thalamus Brain stem Cerebellum
Control
DM
46.2 ⫾ 29.9 ⫾ 48.6 ⫾ 52.4 ⫾ 44.4 ⫾ 47.6 ⫾
44.4 29.9 48.5 49.8 45.6 48.5
5.0 2.9 5.9 6.5 4.9 5.7
rCOS (ml·O2/100 g per min)
⫾ 9.2 ⫾ 6.0 ⫾ 11.5 ⫾ 11.7 ⫾ 11.4 ⫾ 11.4
Non-DM
Control
DM
Non-DM
47.8 ⫾ 28.8 ⫾ 52.1 ⫾ 53.9 ⫾ 46.7 ⫾ 49.5 ⫾
8.1 ⫾ 5.2 ⫾ 8.5 ⫾ 9.2 ⫾ 7.7 ⫾ 8.3 ⫾
5.2 ⫾ 0.8* 3.5 ⫾ 0.6* 5.6 ⫾ 0.9* 5.8 ⫾ 1.0* 5.3 ⫾ 0.9* 5.6 ⫾ 0.9*
5.7 3.4 6.2 6.4 5.6 6.0
11.0 5.6 12.7 12.4 10.0 9.6
0.8 0.6 0.9 1.2 0.8 0.8
⫾ 0.8* ⫾ 0.5* ⫾ 1.0* ⫾ 1.1* ⫾ 0.9* ⫾ 0.9*
* P ⬍ 0.01 vs control group
Fig. 1. Cerebral blood flow (CBF) images obtained from six slices of brain parallel to the orbitomeatal plane (OM) and region of interest (ROI) in each image
Results Relationship between rCBF and Hct In all regions of cerebral cortex, white matter, striatum, thalamus, brain stem, and cerebellum, there was a negative correlation between rCBF and Hct. As Hct increased, rCBF fell. This was most remarkable in the DM group. At a higher Hct level (Hct ⬎30%), the rCBF values in the DM and non-DM groups were lower than in controls (Fig. 3).
Fig. 2. The relationship between regional cerebral oxygen supply (rCOS) and hematocrit (Hct). Because rCOS can be expressed as the product of regional cerebral blood flow (rCBF) and Hct, rCOS vs Hct can be shown as a convex curve with the equations of the regression lines of rCBF vs Hct and O2 content of arterial blood (O2CT) vs Hct. The peak of the curve can be considered as the point of maximum rCOS (rCOSmax) and optimal hematocrit (Hctopt)
Differences in mean red blood cell hemoglobin (MCH), mean red blood cell volume (MCV), and mean red blood cell hemoglobin concentration (MCHC) between controls and HD patient groups There was no significant difference in MCH (pg) between groups. The MCV (fl) in HD patients was significantly greater than in controls, and MCHC (pg/fl) in HD patients was significantly lower than in controls (Fig. 5). For HD patients, there were no significant differences in any of these values between the DM and non-DM groups (Table 2).
Relationship between O2CT and Hct There was a positive correlation between Hct and O2CT. At the same Hct level, the O2CT values in the DM and nonDM groups were lower than in controls. Furthermore, the regression lines in the DM and non-DM groups were similar (Fig. 4).
Relationship between rCOS and Hct The correlation between Hct and rCOS was expressed as a convex curve; the peak of each curve was the point of Hctopt and rCOSmax in the cerebral cortex, white matter, striatum,
143 Fig. 3. Negative correlation between rCBF and Hct. In each region of the cerebral cortex, white matter, striatum, thalamus, brain stem, and cerebellum, the slope of the regression line is smallest in the diabetes mellitus (DM) group (closed circles, continuous line) suggesting that a change in the Hct in the DM group produces a greater change in the rCBF than it does in the other two groups. Crosses, dotted line, Control group; open circles, dashed line, non-DM group
thalamus, brain stem, and cerebellum (Fig. 6). The Hctopt of these regions differed between each group. There were significant differences in rCOSmax between the DM and control groups, and between the non-DM and control groups. There was no significant difference between the DM and non-DM groups (Table 4).
Discussion Many patients with chronic renal failure suffer from anemia. Because of this decline in hemoglobin and oxygen supply, peripheral tissues are exposed to long-term hypoxic conditions. The development and clinical use of rHuEPO has made a great change in the treatment of anemia in patients with chronic renal failure and HD patients. The target hematocrit (Hct) is 30% at most facilities in Japan. In recent studies, a sustained Hct of more than 35% improved the quality of life (QOL) significantly, and certain studies
reported that a higher Hct normalized the weight of left ventricular muscle.7,12 Some studies have reported the advantages of setting a higher Hct for HD, but these studies may not apply to all HD patients. The optimal Hct in HD patients with different causes of ESRD and complications has not been investigated. Although rHuEPO corrects anemia, it may also result in adverse effects due to an elevation in Hct. During 4 h of HD treatment, the increase in Hct and the decrease in body fluids will increase the viscosity of blood. These factors increase peripheral vascular resistance, and the resultant occlusive complications may outweigh the advantages of treatment with rHuEPO. DM has been shown to worsen arteriosclerosis,23,24 and elevations in Hct may heighten this adverse risk in diabetic HD patients because of compromised organ perfusion compared with non-diabetic HD patients. As evidence for this, rCBF in the DM group was lower than that in the other groups at the same Hct (Fig. 3). The rCBF in whole brain, excluding white matter, is about 50 ml/100 g per min, and in white matter it is about
144
Fig. 4. The relationship of O2CT vs Hct. The regression lines of the hemodialysis (HD) patient groups (DM group and non-DM group) are lower than those for controls, and the DM and non-DM regression lines are similar. Symbols, As in Fig. 3
30 ml/100 g per min, as reported previously5,25,26 and in this study (Table 3). As a result of a decline in oxygen supply due to anemia, rCBF is augmented. The rCBF declines to normal when oxygen supply recovers with the reversal of anemia.25 However, even if the Hct improves to over 30%, rCBF and maximum oxygen intake in HD patients are lower than normal. The O2CT in HD patients is lower than in normals at the same Hct, i.e., oxygen transport ability is poorer in HD patients. Referring to the relationship between Hct and O2CT in this study (Fig. 4), O2CT in both the DM and the non-DM groups was lower than in controls; the regression lines of these two groups overlapped each other. This suggests that a reduction in oxygen transport ability in HD patients is caused by factors other than DM. We found that the MCV of HD patients was greater than in controls, and that the MCHC of HD patients was variable but significantly lower than in controls (Fig. 5). Anemia in HD patients may be due not only to a deficiency of erythropoietin but also to loss of blood, hemolysis caused by extracorporeal circulation, suppression of hematopoiesis, and disorders of nutrition or metabolism. Because the hemoglobin concentration is closely related to oxygen transport, it is possible that the lower O2CT in HD patients was caused by the lower hemoglobin concentration in HD patients at the same Hct. Improving the Hct will increase oxygen transport ability, but it will also increase the viscosity of blood and reduce blood flow to organs. These factors promote the risk of occlusive complications.13,15 The purpose of improving Hct and blood flow is to secure the oxygen supply to tissues and organs; the optimal Hct should achieve the maximum oxygen supply. In this study, the rCOSmax values in the DM and non-DM groups were lower than in the control group. In studies with positron emission tomography (PET), it has
Fig. 5. Red cell indices in HD patients and controls. MCH, mean red blood cell hemoglobin; MCV, mean red blood cell volume; MCHC, mean red blood cell hemoglobin concentration; ns, not significant
145 Fig. 6. The relationship of rCOS vs Hct. Hctopt in the DM group is 21.1% to 25.9% (22.6% ⫾ 1.9%); in the non-DM group, 27.2% to 31.8% (28.9% ⫾ 1.8%); and in controls, 43.2% to 54.3% (48.1% ⫾ 4.0%). The rCOS in the DM group is 3.56 to 6.31 ml·O2/100 g per min; in the non-DM group, 3.55 to 6.79 ml·O2/100 g per min; and in controls, 5.93 to 11.43 ml·O2/100 g per min. Dotted lines, Control group; dashed lines, non-DM group; continuous lines, DM group
Table 4. Optimal hematocrit (Hctopt) and maximum rCOS (rCOSmax) in regions of the brain Hctopt (%)
Cerebral cortex White matter Striatum Thalamus Brain stem Cerebellum Mean ⫾ SD
rCOSmax (ml·O2/100 g per min)
Control
,
DM* ***
Non-DM*
Control
DM**
Non-DM**
47.3 54.3 43.2 49.7 48.3 44.5 47.9 ⫾ 4.0
23.8 25.9 21.6 22.3 21.1 21.1 22.6 ⫾ 1.9
27.2 28.1 27.4 28.7 30.5 31.8 29.0 ⫾ 1.8
8.6 5.9 10.2 11.4 9.6 8.7
5.4 3.6 6.2 6.3 5.9 6.3
5.9 3.5 6.4 6.8 5.7 6.1
* P ⬍ 0.0001 vs control group; ** P ⬍ 0.001 vs control group; *** P ⬍ 0.01 vs non-DM group
been reported that rCOS in HD patients was lower than in controls.26,27 A study of rCOS and rCBF of the cerebral cortex in HD patients and patients without renal disease, using SPECT, has already been performed,28 but the data were not classified by complications or causes of ESRD. When we classified HD patients into DM and non-DM
groups, it became clear that there was a difference in optimal Hct between patients whether they were diabetic or not. There was a significant difference in Hctopt between the DM and non-DM groups, but there was no difference in rCOSmax. Our data suggest that, although rCBF in the DM group was lower than in the non-DM group, the rCOSmax in
146
patients with DM can be kept at the same level at that in non-DM patients, as long as the optimal Hct is maintained. Because the ultrafiltration volume and weight gain differ in each patient, the change in Hct during the time of HD treatment varies widely. We have found that some HD patients have Hct changes of more than 5% after HD. On the basis of studies of cerebral infarction patients29,30 and experiments in dogs,31 it has been reported that there is a wide range of Hct for maintaining maximum oxygen supply. It is thought to be desirable to use the average Hct before and after HD as the target value. In comparison with other studies of the optimal Hct,12,15,26,27 we found a lower optimal Hct in our DM group. This study aimed only at improving cerebral blood flow and oxygen supply. It did not address QOL, patient motility, or oxygen supply to other organs. However, it is clear that there are differences in the optimal Hct based on the complications and causes of ESRD; the ideal treatment strategy is to set optimal Hct targets for each patient relative to the cause of ESRD and any complications.
Conclusions We found a significant difference in optimal hematocrit for the maximum cerebral oxygen supply between diabetic and non-diabetic hemodialysis patients. The optimal hematocrit of diabetic hemodialysis patients was 22.6% ⫾ 1.9% and that for non-diabetic hemodialysis patients was 29.0% ⫾ 1.8%. Although the optimal hematocrit we obtained in this study is only relevant to brain circulation, our results suggest that the target hematocrit in hemodialysis patients must be determined relative to specific complications or duration of hemodialysis.
References 1. Besarab A, Bolton WK, Browne JK, Egrie JC, Nissenson AR, Okamoto DM, et al. The effects of normal as compared with low hematocrit values in patients with cardiac disease who are receiving hemodialysis and epoetin. N Engl J Med 1998;339:584–90. 2. Brown WS, Marsh JT, Wolcott D, Takushi R, Carr CR, Higa J, et al. Cognitive function, mood and P3 latency: effects of the amelioration of anemia in dialysis patients. Neuropsychologia 1991;29:35– 45. 3. Hirakata H, Kanai H, Fukuda K, Tsuruya K, Ishida I, Kubo M, et al. Optimal hematocrit for the maximum oxygen delivery to the brain with recombinant human erythropoietin in hemodialysis patients. Clin Nephrol 2000;53:354–61. 4. Metry G, Wikström B, Valind S, Sandhagen B, Linde T, Beshara S, et al. Effect of normalization of hematocrit on brain circulation and metabolism in hemodialysis patients. J Am Soc Nephrol 1999;10:854–63. 5. Lasen NA. Normal value of cerebral blood flow in younger adults is 50 ml/100 g/min. J Cereb Blood Flow Metab 1985;5:347–9. 6. Temple RM, Langan SJ, Deary IJ, Winney RJ. Recombinant erythropoietin improves cognitive function in chronic hemodialysis patients. Nephrol Dial Transplant 1992;7:240–5. 7. Eschbach JW, Glenny R, Robertson T. Normalizing the hematocrit (HCT) in hemodialysis patients (HDP) with EPO improves quality of life (Q/L) and is safe (abstract). J Am Soc Nephrol 1993;4:425.
8. Hint H. The pharmacology of dextran and the physiological background for the clinical use of Rheomacrodex and Macrodex. Acta Anaesthesiol Belg 1968;2:119–38. 9. Paganini EP. In search of an optimal hematocrit level in dialysis patients: rehabilitation and quality-of-life implications. Am J Kidney Dis 1994;24:S10–6. 10. The Working Group of the Japanese Ministry of Welfare. The guideline for the treatment for anemia due to chronic renal failure with recombinant human erythropoietin (in Japanese). Tokyo: Japanese Ministry of Welfare; 1989 (revised in 1991). 11. National Kidney Foundation (NKF) Anemia Work Group. NKFDOQI clinical practice guidelines for the treatment of anemia of chronic renal failure. Am J Kidney Dis 1997;30:S192–240. 12. 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 1987;316:73–8. 13. Onoyama K, Kumagai H, Miishima T, Tsuruda H, Tomooka S, Motomura K. Incidence of strokes and its prognosis in patients on maintenance hemodialysis. Jpn Heart J 1986;27:685–91. 14. Standage BA, Schuman ES, Ackerman D, Gross GF, Ragsdale JW. Does the use of erythropoietin in hemodialysis patients increase dialysis graft thrombosis rates? Am J Surg 1993;165:650–4. 15. Valderrabano F. Adverse drug effects of recombinant erythropoietin (r-HuEPO) (abstract). Kidney Int 1989;35:265. 16. Statistical Survey Group of the Japanese Society for Dialysis Therapy. An overview of dialysis treatment in Japan (as of Dec. 31, 1998). J Jpn Soc Dial Ther 2000;33:1–27. 17. Rostand SG, Gretes JC, Kirk KA. Ischemic heart disease in patients with uremia undergoing maintenance hemo-dialysis. Kidney Int 1979;16:600–11. 18. Lindner A, Charra B, Sherrard DJ, Scribner BH. Acceleration of arteriosclerosis in prolonged maintenance hemodialysis. N Engl J Med 1974;290:697–701. 19. Castro L, Hofling B, Hassler R. Progression of coronary and valvular heart disease in patients on dialysis. Trans Am Soc Artif Intern Organs 1985;31:647–50. 20. Iida H, Itoh H, Nakazawa M, Hatazawa J, Nishimura H, Onishi Y, et al. Quantitative mapping of regional cerebral blood flow using Iodo-123-IMP and SPECT. J Nucl Med 1994;35:2019–30. 21. Komaba Y, Kitamura S, Terashi A. Effect of prostaglandin E1 on cerebral blood flow in patients with chronic cerebral infarction. Intern Med 1998;37:841–6. 22. Komaba Y, Osono E, Kitamura S, Katayama Y. Crossed cerebellocerebral diaschisis in patients with cerebellar stroke. Acta Neurol Scand 2000;101:8–12. 23. Hara S, Arizono K, Ubara Y. Risk factors for renal retardation of renal function in IDDM and NIDDM with nephropathy. J Diabetes Complications 1991;5:131–3. 24. Yamada N. Increased risk factors for coronary artery disease in Japanese subjects with hyperinsulinemia or glucose intolerance. Diabetes Care 1994;17:107–14. 25. Kee DB, Wood JH. Rheology of the cerebral circulation. Neurosurgery 1984;15:125–31. 26. Hirakata H, Yao H, Osato S, Ibayashi S, Onoyama K, Otsuka M. CBF and oxygen metabolism in hemodialysis patients: effects of anemia correction with recombinant human EPO. Am J Physiol 1992;262:F737–43. 27. Herold S, Brozovic M, Gibbs J, Lammertsma AA, Leenders KL, Carr D. Measurements of regional cerebral blood flow, blood volume and oxygen metabolism in patients with sickle cell disease using positron emission tomography. Stroke 1986;17:692–8. 28. Ishiwata A, Sakayori O, Kitamura S, Tsuganesawa T, Terashi A. Is brain circulation maintained sufficiently in chronic renal failure patients under hemodialysis treatment? J Cereb Blood Flow Metab 1997;17(suppl 1):S465. 29. Fan FC, Chen RYZ, Schuessler GB, Chien S. Effects of hematocrit variations on regional hemodynamics and oxygen transport in the dog. Am J Physiol 1980;238:H545–52. 30. Kusunoki M, Kimura K, Nakamura M, Isaka Y, Yoneda S, Abe H. Effects of hematocrit variations on cerebral blood flow and oxygen transport in ischemic cerebrovascular disease. J Cereb Blood Flow Metab 1981;1:413–7. 31. Jan KM, Chien S. Effect of hematocrit variation on coronary hemodynamics and oxygen utilization. Am J Physiol 1997;233: H106–13.