Clinical Features and Management of Intracranial Hemorrhage in Patients Undergoing Maintenance Dialysis Therapy

Transcription

1 Neurol Med Chir (Tokyo) 44, , 2004 Clinical Features and Management of Intracranial Hemorrhage in Patients Undergoing Maintenance Dialysis Therapy Mamoru MURAKAMI, Tomoyuki HAMASAKI, SatoshiKIMURA, Daisuke MARUYAMA, andkiyohitokakita Department of Neurosurgery, Kyoto First Red Cross Hospital, Kyoto Abstract The management and outcome were retrospectively investigated in patients with chronic renal failure receiving maintenance blood purification who suffered intracranial hemorrhage. Patients with intracerebral hemorrhage (ICH, n = 36) or subarachnoid hemorrhage (SAH, n = 5) were evaluated. Both groups were initially managed using continuous hemofiltration (HF) after admission, except for two patients with SAH receiving maintenance peritoneal dialysis. Patients with ICH were managed with HF three times a week after computed tomography showed decreased peripheral edema. Nafamostat mesilate was used as the anticoagulant for both continuous HF and HF. Hemodialysis (HD) three times a week was initiated after confirming the absence of neurological deterioration using HF. Craniotomy was not performed in any patient with ICH, but if necessary, the hematoma was aspirated using burr-hole surgery. Angiography was performed on the day of admission in patients with SAH. Delayed neck-clipping surgery was performed after continuous HF for 2 weeks with lumbar cerebrospinal fluid drainage. In patients with ICH, continuous HF was continued for 2 9 days after admission (mean 5.2 ± 2.2 days), followed by 2 9 courses of HF (mean 4.7 ± 2.1 courses). HD was initiated 9 26 days after admission (mean 15.5 ± 4.6 days). Favorable outcomes were achieved by 13 of the 36 patients with ICH and two of the five patients with SAH, whereas 22 patients with ICH and three patients with SAH died. Death occurred in 12 of 16 patients with ICH and diabetic nephropathy. In contrast, 10 of 20 non-diabetic patients with ICH had favorable outcomes. Ten of the 16 patients with initial GCS Ã8 and six of the 20 with GCS Æ9 were diabetic. Therefore, there were significant differences between diabetic and non-diabetic patients (p = 0.05). Poor outcomes in diabetic patients with ICH are caused by primary brain damage, reflected in the initial disturbance of consciousness. Key words: cerebral hemorrhage, subarachnoid hemorrhage, hemodialysis, continuous hemofiltration, diabetic nephropathy Introduction Stroke is the third most common cause of death among Japanese patients undergoing dialysis therapy, following heart failure and infections. 8) Several studies of dialysis patients have shown a higher annual incidence of stroke ( per 1000 personyears) compared with the general population, and an incidence of intracerebral hemorrhage (ICH) 2- to 3-fold higher than the incidence of infarction. 6,7,11) Hemodialysis (HD) is used in more than 90% of patients with end-stage renal disease. The number of dialysis patients in Japan has been increasing at a rate of 13,000 per year, with the total number exceeding 219,000 in ) Therefore, neurosurgeons can expect to see more dialysis patients presenting with such neurological conditions in the future. Management of intracranial hemorrhage in HD patients is frequently complicated by factors such as: hypotension during dialysis; active bleeding and coagulopathy that may be exacerbated by the systemic anticoagulant therapy used in HD; and dialysis disequilibrium syndrome (DDS), which is attributed to edema of the central nervous system caused by a rapid osmolar shift, often resulting in increased peripheral edema and neurological de- Received March 3, 2003; Accepted January 16, 2004 Author's present address: M. Murakami, M.D., Department of Neurosurgery, Kyoto National Hospital, Kyoto, Japan. 225

2 226 M. Murakami et al. terioration. 19) Continuous hemofiltration (HF) is recommended to prevent such complications, but the duration of continuous HF and the timing of restart of conventional HD remain unclear. The present study retrospectively analyzed 36 patients with ICH and five with subarachnoid hemorrhage (SAH) managed with continuous HF followed by HF and conventional HD, to evaluate the outcomes of these patients, and the methods and durations of blood purification. Materials and Methods I. Patient population Forty-one patients receiving maintenance HD or peritoneal dialysis (PD) for end-stage renal disease were admitted to our institute with ICH (36 patients) and SAH (5 patients) between November 1997 and December The present study excluded patients who initially received HD after admission, and those with traumatic lesions. II. Clinical standard management Computed tomography (CT) was performed as soon as possible after hemodynamic stabilization. Thevolumeofthehematomawasdeterminedas: A B C/2,whereA,B,andCarethemaximum orthogonal diameters of the hematoma. Patients with ICH in score 3 or 4 of the Glasgow Coma Scale (GCS) due to cerebral herniation were not given further treatment with blood purification or surgery. Since evacuation of hematoma in dialysis patients is rarely beneficial, aggressive craniotomy was not performed. Stereotaxic aspiration of hematoma was performed if CT revealed ICH À35 ml at the basal ganglia or thalamus, or subcortical hematoma À50 60 ml, with GCS score Ã10. Patients with lobar hemorrhage Àapprox. 100 ml, or aged over 75 years old were contraindicated for surgery. Ventricular drainage was utilized if hydrocephalus was present. Patients with SAH in grade V according to the World Federation of Neurological Surgeons scale were treated conservatively without blood purification or surgery. After angiography on admission, delayed neck-clipping surgery was performed in patients with favorable grades about 14 days after admission. Lumbar cerebrospinal fluid drainage tubes were placed to control intracranial pressure (ICP) and prevent vasospasm. III. Blood purification after admission The principle of HD is the diffusion of low molecular weight solutes across a semipermeable membrane, and fluid removal occurs via ultrafiltration. HD effectively removes solutes and fluid, but the complications of hypotension during dialysis and DDS, which often results in increased cerebral edema, prevent the use of HD in the acute phase of stroke. Therefore, continuous HF was used initially as a slow, continuous renal replacement therapy. 13) Continuous HF was performed with a specific hemofilter (Hemofeel ; Toray, Tokyo) using nafamostat mesilate (20 30 mg/hr) as anticoagulant. Replacement solution (Sublood -B; Fuso, Tokyo) was delivered at l/20 24 hrs. Continuous HF was performed until CT showed cerebral edema had decreased in patients with ICH, and until neck clipping surgery in patients with SAH. The treatment was then changed to HF three times a week. Patients with small ICH º15 ml without mass effects, particularly subcortical hematoma, underwent HF initially instead of continuous HF. HF is less effective at removing low molecular weight solutes than HD, but less frequently causes DDS. Patients were managed with HF using a dialyzer (UF-210F ; Nipro, Tokyo) with administration of mg/5 6 hrs nafamostat mesilate. Supplement solution (HF-Solita ;Shimizu- Takeda, Tokyo) was delivered over 5 6 hours ( ml/hr). If signs and symptoms of uremia still developed following use of HF only, use of hemodiafiltration (HDF) was considered, but was required in only one patient (Case 36). HD three times a week using a dialyzer (FB ;Nipro)andlow molecular weight heparin as anticoagulant 16) was not instituted in patients with either ICH or SAH until HF had been confirmed to cause no neurological deterioration. Treatment algorithms are provided in Fig. 1. Patient outcomes were evaluated at least 3 months after admission according to the Glasgow Outcome Scale (GOS). 9) IV. Representative case Case 24: A 51-year-old man presented with a history of chronic renal failure resulting from chronic glomerulonephritis first indicated in He had received neither antihypertensive nor antiplatelet agents. He had been admitted comatose to another hospital. CT demonstrated a 5 3 6cm ICH within the right putamen (Fig. 2A). He was transferred to our institute with an initial GCS score 7 (E2, V1, M4). Continuous HF was started on the following day (Day 1). His consciousness gradually improved to GCS score 10 (E3, V3, M4) from Day 4. Since CT revealed increased peripheral edema, aspiration of the hematoma was performed on Day 8. HF was started from Day 10, and was performed eight times due to persistent peripheral edema on

3 Intracranial Hemorrhage in Dialysis Patients 227 Fig. 1 Treatment algorithms for dialysis patients with intracerebral hemorrhage (ICH) (A) or subarachnoid hemorrhage (SAH) (B). CSF: cerebrospinal fluid, CT: computed tomography, GCS: Glasgow Coma Scale, HD: hemodialysis, HF: hemofiltration, WFNS: World Federation of Neurological Surgeons. CT (Fig. 2B). Following that, conventional HD was safely performed from Day 26. He was transferred to the previous dialysis hospital after recovery to GCS score 14 (E4, V4, M6) on Day 41. Results Fig. 2 Case 24. Computed tomography scans on admission (A), revealing intracerebral hemorrhage measuring 5 3 6cm within the right putamen, and 10 days after admission (B), demonstrating persistent peripheral edema despite surgery for aspiration of the hematoma. I. Outcome Tables 1 and 2 show the clinical characteristics of the patients with ICH and SAH. Thirteen of the 36 patients with ICH achieved favorable outcomes (GOS of good recovery or moderate disability), but 22 of 36 patients died (overall mortality 61.1%). Sixteen of these deaths were directly attributable to ICH, with six deaths due to other causes. Two patients (Cases 18 and 33) died of ventricular fibrillation on Days 0 and 15, respectively. One patient (Case 10) was well manageduntilday17(gcsscore14),butdiedof sepsis on Day 75 after rectal bleeding. Two patients (Cases 1 and 29) died of respiratory failure during

4 228 M. Murakami et al. Table 1 A: Characteristics (A) and blood purification (B) of 36 patients with intracerebral hematoma (ICH) Case No. Age Sex Diagnosis Volume of hematoma (ml) Ventricular perforation GCS on admission (E,V,M) Surgery (day of surgery) GOS Cause of death 1 64 F rt putaminal h (4,4,5) - D respiratory failure 2 56 F rt thalamic h (4,5,6) VPshunt MD - (Day 18) 3 65 M rt thalamic h (4,2,6) - D respiratory failure 4 61 M rt thalamic h (1, 1, 2) - D ICH 5 67 M rt lobar h (2,2,6) - MD M lt thalamic h (1, 1, 2) - D ICH 7 43 M lt putaminal h (4,3,5) - MD F lt putaminal h (2,3,6) - MD F lt lobar h (1,1, 2) - D ICH F rt thalamic h (4,5,6) - D sepsis M rt thalamic h (1, 1, 2) - D ICH F rt putaminal h (4,5,6) stereo MD - (Day 3) M lt putaminal h. 6-15(4,5,6) - GR F lt cerebellar h (1, 2, 4) VD (Day 0) D ICH F lt thalamic h (2, 2, 4) VD (Day 0) D ICH M lt lobar h (2, 3, 4) - D ICH F lt frontal (1, 1, 1) - D ICH M rt putaminal h (3,3,6) stereo D arrhythmia (Day 1) M rt lobar h (1, 1, 1) - D ICH M rt temporal (1, 1, 2) - D ICH F lt frontal (1, 1, 1) - D ICH M rt thalamic h (4,5,6) - MD M rt lobar h (1, 1, 2) - D ICH M rt putaminal h (2, 1, 4) stereo MD - (Day 8) F rt lobar h (1, 1, 2) - D ICH M pontine h (1, 1, 2) - D ICH F rt thalamic h (4,4,6) - MD F lt putaminal h. 9-15(4,5,6) - MD F lt thalamic h (1, 1, 2) - D respiratory failure M rt parietal (3,4,6) - D ICH M lt cerebellar h (4,5,6) - MD M lt thalamic h (1,1, 2) - D ICH M rt temporal 40-11(4,2,5) - D arrhythmia F rt putaminal h. 5-15(4,5,6) - MD M lt occipital (3,2,5) - SD M rt putaminal h (3,2,5) - MD - Mean±SD 61.3± ±55.2 D: dead, GCS: Glasgow Coma Scale, GOS: Glasgow Outcome Scale, GR: good recovery, h.: hemorrhage, MD: moderate disability, mean±sd: mean±standard deviation, SD: severe disability, stereo: stereotaxic aspiration of hematoma, VD: external ventricular drainage, VP: ventriculoperitoneal. HF. One patient (Case 3) died of respiratory failure probably attributable to DDS during the fourth HD on Day 9. He had undergone continuous HF for 7 days without neurological deterioration prior to HD initiated on Day 9. This was the only patient not to receive HF before HD. Subsequently, HD was never started until good management with HF had been achieved, even if continuous HF could maintain

5 Intracranial Hemorrhage in Dialysis Patients 229 B: Case No. Initial blood purification* (time after admission, hours) Period of CHF (days) Courses of HF (HDF)* Time from admission to start of HD* (days) Presence of DM Duration of previous dialysis** (years) Medication with antihypertensive agents History of ischemic heart disease** Medication with antiplatelet agents** 1 CHF (9) HF (12) CHF (38) / / 5 CHF (7) / - 7 CHF (24) CHF (20) / / 10 HF (3) HF (40) HF (13) CHF (14) CHF (6) CHF (6) / - / / 18 CHF (4) / CHF (2) CHF (18) CHF (1) CHF (14) CHF (2) CHF (14) CHF (16) CHF (2) CHF (14) CHF (15) 7 3 (HDF) Mean±SD (12.8±10.5) 5.2± ± ± ±5.7 CHF: continuous hemofiltration, DM: diabetus mellitus, HD: hemodialysis, HDF: hemodiafiltration, HF: hemofiltration, mean±sd: mean ± standard deviation, *-: not performed, **/: not confirmed. stable neurological symptoms. Three of the five patients with SAH died. Two patients (Cases 3 and 5) died from severe SAH (neurological grade V), whereas one patient (Case 1) underwent neck clipping on Day 14, but died on Day 15 of cerebral infarction probably caused by vasospasm. II. Duration of blood purification Maintenance dialysis before admission utilized HD in all 36 ICH patients and three of the five SAH patients. The other two SAH patients had been managed with PD. Thirteen of the 36 ICH patients were contraindicated for blood purification. Nineteen patients with ICH underwent continuous HF for 5.2 ± 2.2 days (range 2 9 days). After HF was performed (mean 4.7 ± 2.1 courses, range 2 9 courses), HD was started 9 26 days after admission (mean 15.5 ± 4.6 days). Four patients with ICH of º15 ml were initially treated with HF (Cases 2, 10, 12, and 13), all of whom were well managed during HF. One patient (Case 10) died of sepsis, whereas the other three achieved favorable outcomes.

6 230 M. Murakami et al. Table 2 A: Characteristics (A) and blood purification (B) of five patients with subarachnoid hemorrhage (SAH) Case No. Age Sex H&K grade WFNS grade Fisher's group Location of aneurysm Operation (operative day) GOS 1 67 F III IV 4 AComA neck clipping (Day 14) D 2 35 M II II 3 unknown SAH - GR 3 76 M V V D 4 73 F III IV 3 lt IC-PComA neck clipping (Day 28) MD 5 61 F V V D B: Case No. Initial blood purification (time after admission, hours)* Period of CHF* (days) Presence of DM Previous dialysis method (duration, years) 1 CAPD (9) - - PD (5) 2 CAPD (4) - - PD (6) HD (6) 4 CHF (3) 21 - HD (7) 5 CHF (34) - - HD (11) AComA: anterior communicating artery, CAPD: continuous ambulatory peritoneal dialysis, CHF: continuous hemofiltration, D: dead, DM: diabetes mellitus, GOS: Glasgow Outcome Scale, GR: good recovery, H&K: Hunt and Kosnik, HD: hemodialysis, IC-PComA: internal carotid artery-posterior communicating artery, MD: moderate disability, PD: peritoneal dialysis, WFNS: World Federation of Neurological Surgeons, *-: not performed. III. Factors influencing the outcome of patients with ICH Various factors for ICH patients were identified retrospectively as possible discriminators of favorableandunfavorableoutcomes.the36patientswith ICH were divided into those with or without diabetes mellitus (DM) to evaluate the effect on outcome. Three of the 16 diabetic patients and 10 of the 20 non-diabetic patients had favorable outcomes, whereas 12 diabetic patients and 10 non-diabetic patients died. There were significant differences with regard to both outcomes (x 2 test, p = 0.05). Twenty patients had initial GCS scores Æ9, whereas 16 had Ã8. Six patients with GCS scores Æ9 and10 patients with Ã8 had DM, representing a significant difference (p = 0.05). All four patients aged over 75 years (Cases 10, 15, 20, and 35) had poor outcomes, whereas 13 of the 32 patients aged Ã75 years experienced favorable outcomes. However, no significant difference was observed (p = 0.11). Medication with antihypertensive and antiplatelet agents, and history of ischemic heart disease showed no significance (p = 0.12, p = 0.41, and p = 0.90, respectively). Patients undergoing maintenance HD for À10 years (n = 9) or Ã10 years (n = 25) displayed no significant differences in outcomes (p = 0.11). Discussion The present study found the overall mortality for patients with ICH was 61.1%, confirming previous reports of 53 79% mortalities in dialysis patients. 7,10,11,17) Whether ICH occurs more frequently in dialysis patients with DM compared to those without DM remains unclear. 8) This study found that DM was associated with poor outcome, and that poor outcomes in diabetic patients were not caused by complications after admission but by primary damage leading to poor GCS. We suggest that DM is not associated with increased onset of ICH, 8,17) but instead influences the severity of the event. Diabetic nephropathy has been increasing, and was the most common cause of dialysis induction in ) Management of dialysis patients with severe ICH derived from DM will represent a serious problem in the future. Many blood purification methods are available for dialysis patients. HD has the following disadvantages: promotion of cerebral edema and increased ICP (so-called DDS); anticoagulants given during HD may exacerbate hemorrhaging and risk cerebral herniation, depending on the bleeding site; and control of blood pressure during HD is difficult. In contrast, PD offers the following advantages: reduced burden on the cardiovascular system; no anticoagulants are required, so risk of further

7 Intracranial Hemorrhage in Dialysis Patients 231 hemorrhaging is reduced; and continuous dialysis is associated with lower risk of exacerbating cerebral edema. PD has been recommended for post-onset management of patients with ICH. 2 4,18) PD can maintain and lower ICP compared to HD, 12) but also has some disadvantages: the technique is less efficient and less useful in highly catabolic patients; fluid removal by only PD may be insufficient; and satisfactory removal of fluid and solutes is obtained at least 2 or 3 days after placement of the abdominal catheter. PD should be continued after admission for patients treated with maintenance PD, but continuous HF is recommended in intensive care for patients with ICH and SAH treated with maintenance HD. 1 3) This study indicated that HF is essential as an intermediate between continuous HF and HD. Favorable outcome was achieved in 13 of 19 patients treated using HF. One of the six patients with poor outcomes (Case 35) was severely disabled. None of the causes of death, respiratory failure (Cases 1 and 29), sepsis (Case 10), primary brain damage (Case 14), and arrhythmia (Case 18), were directly associated with HF. We encountered a patient who developed DDS after the early induction of HD (Case 3). Persistent peripheral edema was often found in dialysis patients, so HF should be performed before HD to prevent DDS. HF removes fewer solutes than HD, but unsatisfactory removal of solutes occurred in only one patient (Case 36), who later required HDF. Surgical indications for the removal of ICH remain controversial, particularly for dialysis patients. In our series, no craniotomies were performed for aggressive evacuation, and aspiration of hematoma was avoided wherever possible. Freehand aspiration for large periventricular hemorrhage is beneficial, 5) but platelet dysfunction due to uremiaandcoagulopathybysystemicanticoagulation can result in difficulties maintaining hemostasis and increased risk of rebleeding after surgery. Of the three patients who underwent stereotaxic aspiration of hematoma, two had favorable outcomes and one died of arrhythmia. In our patients with favorable outcomes, a maximum volume of ICH in the basal ganglia or thalamus was 45 ml (Case 24), whereas one patient (Case 18) with ICH of 48 ml died of arrhythmia after aspiration of hematoma. We propose that a volume of ICH of Ã45 50 ml in the basal ganglia or thalamus represents the appropriate indication for aspiration of hematoma, whereas a volume of À60 ml indicates a poor prognosis irrespective of the surgical procedure. None of the six patients with subcortical hemorrhage underwent surgery. Three patients (Cases 17, 20, and 21) died of primary brain damage, one (Case 33) died of arrhythmia on the day of admission, and another (Case 35), who underwent no surgery due to age, was severely disabled. The other patient (Case 30) with ICH of 60 ml in the parietal lobe was managed with continuous HF from 14 hours after admission, but suddenly developed respiratory failureonday6anddied.hewastheonlypatient who displayed deterioration during continuous HF, so retrospectively, aspiration surgery should have been performed. We had too small numbers of these cases, but ICH of À60 ml in the subcortical hematoma cannot be managed without surgery. The optimal timing of neck-clipping surgery for SAH is not well understood. Early surgery performed less than hours after the onset of SAH offers some advantages for prevention of rebleeding and vasospasm, and has also been recommended for dialysis patients with SAH. 4,15) We performed early surgery before 1997, but 10 of our 11 patients died (Kakita et al., unpublished data), forcing a change to delayed surgery. In this study, one patient was well managed, but the outcomes of other patients obviously indicate the need for further evaluation of appropriate protocols. Intra-aneurysmal embolization may result in good outcomes in dialysis patients with unruptured cerebral aneurysms, 14) so endovascular surgery may be suitable. Acknowledgments WewishtothankDrs.N.Iwamoto,S.Yamasaki,T. Nakanouchi,M.Koyama,N.Sato,T.Ueda,andK. Yamamoto of the Department of Urology at Kyoto First Red Cross Hospital for support in the management of blood purification, and Drs. K. Yoda, A. Saito, and M. Hirata of the Department of Anesthesia, Kyoto First Red Cross Hospital for leading the intensive care unit. References 1) Caruso DM, Vishteh AG, Greene KA, Matthews MR, Carrion CA: Continuous hemodialysis for the management of acute renal failure in the presence of cerebellar hemorrhage. Case report. J Neurosurg 89: , ) Gondo G, Fujitsu K, Kuwabara T, Mochimatsu Y, Ishiwata Y, Oda H, Takagi N, Yamashita T, Fujino H, Kim I, Nakajima F: [Comparison of five modes of dialysis in neurosurgical patients with renal failure]. Neurol Med Chir (Tokyo) 29: , 1989 (Jpn, with Eng abstract) 3) Gondo G, Yamanaka Y, Fujii S, Yamamoto I, Yakagi N, Suzuki N, Sugiyama M, Hosoda H: [Treatment strategy of cerebral apoplectic patients with renal

8 232 M. Murakami et al. failure]. Surgery for Cerebral Stroke 28: , 2000 (Jpn, with Eng abstract) 4) Hirano K, Ishii R, Suzuki Y, Kikuoka M, Hirano H, Ohsawa G, Ohtsuka R, Itoh Y: [Neurosurgical management of dialyzed patients]. Neurol Med Chir (Tokyo) 31: , 1991 (Jpn, with Eng abstract) 5) Imaizumi S, Onuma T, Mino M, Kameyama M, Motohashi O: Free hand aspiration for large periventricular hemorrhage: case report. Surg Neurol 55: , ) Iseki K, Fukiyama K: Clinical demographics and long-term prognosis after stroke in patients on chronic haemodialysis. The Okinawa Dialysis Study (OKIDS) Group. Nephrol Dial Transplant 15: , ) Iseki K, Kinjo K, Kimura Y, Osawa A, Fukiyama K: Evidence for high risk of cerebral hemorrhage in chronic dialysis patients. Kidney Int 44: , ) Japanese Society for Dialysis Therapy: [An Overview of Regular Dialysis Treatment in Japan]. Tokyo, Japanese Society for Dialysis Therapy, 2001, pp (Jpn) 9) Jennett B, Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1: , ) Kawahata N: [Brain hemorrhage associated with maintenance hemodialysis. CT analysis of 19 cases]. No Socchu 16: 79 86, 1994 (Jpn, with Eng abstract) 11) Kawamura M, Fujimoto S, Hisanaga S, Yamamoto Y, Eto T: Incidence, outcome, and risk factors of cerebrovascular events in patients undergoing maintenance hemodialysis. Am J Kidney Dis 31: , ) Krane NK: Intracranial pressure measurement in a patient undergoing hemodialysis and peritoneal dialysis. Am J Kidney Dis 13: , ) Macias WL, Mueller BA, Scarim SK, Robinson M, Rudy DW: Continuous venovenous hemofiltration: An alternative to continuous arteriovenous hemofiltration and hemodiafiltration in acute renal failure. Am J Kidney Dis 18: , ) Nakajima T, Ezura M, Takahashi A, Yoshimoto T: [A case of unruptured cerebral aneurysm treated by intra-aneurysmal embolization during outpatient hemodialysis for chronic renal failure]. No Shinkei Geka 27: , 1999 (Jpn, with Eng abstract) 15) Sasaki H, Gotoh S, Takamura H, Ozaki N, Takahashi H, Wakita K, Mita N, Miyazaki S, Hohkin K, Isu T: [Early operation for ruptured aneurysm in the case of a chronic hemodialysis patient. Report of three cases]. Surgery for Cerebral Stroke 19: , 1991 (Jpn, with Eng abstract) 16) Schrader J, Valentin R, Tonnis HJ, Hildebrand U, Stibbe W, Armstrong VW, Kandt M, Kostering H, Quellhorst E: Low molecular weight heparin in hemodialysis and hemofiltration patients. Kidney Int 28: , ) Shimojo S, Ebisawa T, Hasegawa T, Toyohara K, Sakai O: [Intracranial hemorrhage in chronic renal failure]. No Socchu 19: 66 70, 1997 (Jpn, with Eng abstract) 18) Yorioka N, Oda H, Ogawa T, Taniguchi Y, Kushihata S, Takemasa A, Usui K, Shigemoto K, Harada S, Yamakido M: Continuous ambulatory peritoneal dialysis is superior to hemodialysis in chronic dialysis patients with cerebral hemorrhage. Nephron 67: , ) Yoshida S, Tajika T, Yamasaki N, Tanikawa T, Kitamura K, Kubo K, Lyden PD: Dialysis disequilibrium syndrome in neurosurgical patients. Neurosurgery 20: , 1987 Address reprint requests to: M. Murakami, M.D., Department of Neurosurgery, Kyoto National Hospital, 1 1 Mukaihata cho, Fushimi ku, Kyoto , Japan. Mamoru.Murakami@ma8.seikyou.ne.jp Commentary on this paper appears on the next page.

9 Intracranial Hemorrhage in Dialysis Patients 233 Commentary The authors have retrospectively analyzed the outcome of 41 patients with chronic renal failure who suffered a spontaneous intracranial hemorrhage. Thirty-six of these patients experienced an intracerebral hemorrhage and five a subarachnoid hemorrhage. The review emphasizes the difficulty in treating these challenging patients and the poor outcomes associated with their management. In analyzing their experience, the authors propose an algorithm utilizing continuous hemofiltration to avoid the complications associated with hemodialysis in this group of patients. This algorithm appears to be a very reasonable way to manage these patients from a medical standpoint while preparing them for surgical intervention, if necessary. Another recent adjunct to the management of these patients is the development of activated factor VII, which rapidly and reliably reverses coagulopathies in patients with liver and renaldisease,allowingmoreaggressivesurgicalintervention if indicated. The authors are to be congratulated on their analysis and recommendations in the management of these critically ill patients. Daniel L. BARROW, M.D. Department of Neurosurgery The Emory Clinic Atlanta, Georgia, U.S.A. This scrupulous retrospective study on patients with chronic renal failure and intracranial hemorrhage is a detailed report on the experiences from 41 patients. Obviously the number of patient requiring hemodialysis increases with time and patients with additional intracranial hematomas are seen with increasing frequency. Several observations from this report are of great practicable value: Diabetes was associated with a more severe brain damage as reflected by an initial disturbance of consciousness and was found to be a predictor of a less favorable outcome. Altogether these patients have a mortality of 61%. 45 to 50 ml of blood in basal ganglia can well be treated by evacuation, but larger volumes appear to have a less favorable outcome irrespective of surgery. Hemofiltration proved to be essential as an intermediate between continuous hemofiltration and hemodialysis. The courses of event after subarachnoid hemorrhage are of particular interest. It is noteworthy to realize that early surgery of aneurysms had been associated with a mortality of 10 out of 11 patients and therefore the authors resorted to delayed surgery for aneurysms. This report is obviously a good source of data for anybody confronted with intracranial hemorrhage with concomitant renal failure. Raimund FIRSCHING, M.D., L.R.C.P., M.R.C.S. Klinik f äur Neurochirurgie Otto-von-Guericke-Universit äat Magdeburg, Germany