Neuropathological findings after bone marrow transplantation - Nature

Bone Marrow Transplantation (2000) 25, 301–307  2000 Macmillan Publishers Ltd All rights reserved 0268–3369/00 $15.00 www.nature.com/bmt

Neuropathological findings after bone marrow transplantation: an autopsy study of 180 cases LF Bleggi-Torres1, BC de Medeiros1, B Werner1, J Zanis Neto2, G Loddo1, R Pasquini2 and CR de Medeiros2 1

Department of Medical Pathology, and 2Bone Marrow Transplantation Unit, Hospital de Clıńicas, UFPR, Curitiba, Brazil

Summary: We prospectively evaluated the neuropathological complications of 180 patients who underwent autopsy studies following bone marrow transplantation (BMT) (177 allogeneic, three autologous). The most frequent underlying disorders included severe aplastic anemia (n = 55), chronic myelogenous leukemia (n = 53), acute myelogenous leukemia (n = 24) and Fanconi anemia (n = 16). There were 114 males and 66 females. Neuropathological findings were detected in 90.55% of the patients. The most frequent findings were subarachnoid hemorrhages (SAH) (n = 57), intraparenchymal hemorrhages (IHP) (n = 49), fungal infections (n = 16), Wernicke’s encephalopathy (n = 10), microglial nodular encephalopathy (n = 10) and neurotoxoplasmosis (n = 8). In only 17 patients was the brain within normal limits. Survival time after BMT averaged 5.4 months and the majority of patients died in the first 3 months post BMT (n = 105). Central nervous system (CNS) pathology was the main cause of death in 17% of the patients (n = 31), with a predominance of IHP in this particular group. Furthermore, the survival time of these patients who died of CNS causes (96.3 days) was almost half of the survival time of those who died of extra-cerebral causes (177.8 days) (P = 0.0162). IHP (70.96 vs 27.22%) (P ⬍ 0.001), fungal infections (25.8 vs 8.88%) (P ⬍ 0.001) and toxoplasmosis (9.67 vs 4.44%) (P ⬍ 0.001) were significantly more frequent in the group of patients who died due to CNS causes than in the control group. The findings of this work provide a possible guide to the possible causes of neurological syndromes following BMT. Bone Marrow Transplantation (2000) 25, 301–307. Keywords: central nervous system; bone marrow transplantation; cerebrovascular disorders; infections

Bone marrow transplantation (BMT) is used to treat several malignant diseases, severe aplastic anemia, immunodeficiencies and metabolic disorders.1,2 It allows use of potentially lethal doses of chemotherapy and/or irradiation to eradicate systemic malignancy or poorly functioning conCorrespondence: Dr LF Bleggi-Torres, Seçaõ de Microscopia Eletroˆnica e Neuropatologia, Serviço de Anatomia Patolo´gica, Hospital de Clıńicas/UFPR, Rua General Carneiro, 181-80060-900-Curitiba, Pr-Brazil Received 9 June 1999; accepted 23 September 1999

stituents of the bone marrow, replacing these with donor bone marrow that will provide normally functioning cells and enzymes.1–3 All these diseases and disorders share common problems in the post-BMT period, including the potential for: (1) toxicity from irradiation therapy and/or chemotherapy; (2) infection caused by immunosuppression and mucosal injury; (3) thrombocytopenia; (4) graft-versus-host disease (GVHD); and (5) disease relapse.1–3 Few communications in the literature have addressed the main findings in the central nervous system (CNS) and these have found a conflicting incidence of neurological2–4 (11–70%) and neuropathological2–5 (45–74%) complications following BMT. The CNS represented less than 10% of the causes of death in these patients.2–5 We present a CNS autopsy study of 180 adult and pediatric patients who underwent BMT. This represents the largest experience presented to date in the literature. Patients and methods From July 1987 to June 1998, 845 patients at our institution underwent either allogeneic or autologous bone marrow transplantation. A total of 371 died and 196 patients underwent autopsy studies. The CNS of 180 patients were studied and their medical records were reviewed. Complete autopsy examinations were performed in most patients and their brains were removed and fixed for at least 3 weeks in 30% buffered formaldehyde solution. They were macroscopically analyzed through several 1 cm thick coronal sections and tissues samples from at least 12 different areas (frontal, temporal, occipital and parietal lobes, insula, thalamus, hypothalamus, mammilary bodies, midbrain, pons, medulla oblongata, cerebellar hemisphere and vermis) were obtained in every case. Areas in the CNS, other than the above mentioned, containing any macroscopic lesions were also sampled. Tissue sections were then prepared by routine neupathological techniques and additional special stains were carried out on histological sections when necessary. The presence of diffuse microglial hyperplasia (DMH), microglial nodular encephalopathy (MNE), Toxoplasma gondii and cytomegalovirus (CMV) infection were confirmed by immunohistochemistry6 using monoclonal antibody (moAb) anti-HLA/DR-CR3/43, moAb anti-CMV (both from Zymed Laboratories, San Francisco, CA, USA) and polyclonal antibody (poAB) anti-Toxoplasma gondii (Biogenesis, Sandown, USA). In a few cases of CNS infections electron microscopy was used to identify etiological

Neuropathological complications after BMT LF Bleggi-Torres et al

302

Table 1

were the cause of death in 31 patients (17%) (Table 4). The majority of patients had more than one neuropathological finding (Table 3).

Underlying disease for which BMT was performed

Disease

No. of patients

Severe aplastic anemia Chronic myelogenous leukemia Acute myelogenous leukemia Fanconi anemia Acute lymphocytic leukemia Myelodysplasia Lymphoma Other causes

Incidence

55 53 24 16 9 8 3 12

The number of BMTs increased annually between 1987 and 1998. A total of 845 procedures were performed during this period. There were 371 deaths and 180 of these cases had their brains studied by autopsy. Therefore, CNS abnormalities could be found in one of every 5.2 transplant recipients, in one of every 2.3 deaths that occurred in this period and in one of every 1.1 brains studied by autopsy. Furthermore, cerebral causes of death were found in one of every 27.2 marrow recipients, in one of every 11.9 deaths that occurred in this period and in one of every 5.8 brains studied by autopsy. However, since an autopsy and brain study were not performed in every single patient that died, it is possible that this incidence may have been underestimated.

agents. Statistical analyses were conducted by analysis of variance, coefficient of determination, chi-square test and Student’s t-test. Results Of the 180 autopsies performed, there were 114 males (63%) and 66 females (37%); the average age was 23.7 years (range, 1–48). Underlying diseases are listed in Table 1; preparatory regimens for BMT depended on the underlying disorder (Table 2). There were 177 allogeneic and three autologous BMTs. Most patients died during the first 3 months (105; 58.33%) following BMT and the average survival time was 163 days (range, 0–1264). One hundred and sixty-three patients (90.55%) were found to have abnormalities within the CNS (Table 3) and these complications Table 2

Cerebrovascular disorders These were the most common identifiable lesions in this study. They could be classified as: subarachnoid hemorrhages (SAH) (31.67%), intraparenchymal hemorrhages (IPH) (27.22%), intraventricular hemorrhages (8.33%) and subdural hemorrhages (6.66%). Thirty-two patients (17.77%) had more than one type of hemorrhage. The lesions ranged in size from petechiae and focal subarachnoid bleeds (Figure 1a) to large, massive, necrotic

BMT preparatory regimens according to the underlying disorder

BU-CY CY-TBI BU-CY-TBI CY BU-CY-VP CY-THYMO

SAA (n = 55)

CML (n = 53)

AML (n = 24)

FA (n = 16)

ALL (n = 9)

MDS (n = 8)

Lymphoma (n = 3)

Others (n = 12)

27 — — 28 — —

52 1 — — — —

19 5 — — — —

— 2 — 11 — 3

— 6 — — 3 —

5 — — — 3 —

— 3 — — — —

9 — — — 2 1

BU = busulphan; CY = cyclophosphamide; TBI = total body irradiation; VP = etoposide; Thymo = anti-thymocytic globulin; SAA = severe aplastic anemia; CML = chronic myelogenous leukemia; AML = acute myelogenous leukemia; FA = Fanconi anemia; ALL = acute lymphoblastic leukemia; MDS = myelodysplasia. Table 3

Summary of neuropathological findings in BMT patients

Normal SAH IPH FI NT MNE WE Others

SAA (n = 55)

CML (n = 53)

AML (n = 24)

FA (n = 16)

ALL (n = 9)

MDS (n = 8)

Lymphoma (n = 3)

Others (n = 12)

3 22 19 8 4 6 1 6

4 12 14 3 2 3 4 5

1 6 5 1 1 0 2 3

0 9 4 2 1 1 0 7

1 3 1 0 0 0 2 4

2 2 2 0 0 0 0 2

2 0 0 0 0 0 1 0

4 3 4 1 0 0 0 2

SAH = subarachnoid hemorrhages; IPH = intraparenchymal hemorrhages; FI = fungal infections; NT = neurotoxoplasmosis; MNE = microglial nodular encephalopathy; WE = Wernicke’s encephalopathy; SAA = severe aplastic anemia; CML = chronic myelogenous leukemia; AML = acute myelogenous leukemia; FA = Fanconi anemia; ALL = acute lymphocytic leukemia; MDS = myelodysplasia. Bone Marrow Transplantation


303

Table 4 Neuropathological findings of the patients who had the CNS as cause of death Neuropathological findings IPH SAH FI NT MNE WE

No. of cases (n = 31) 22 14 8 3 2 1

SAH = subarachnoid hemorrhages; IPH = intraparenchymal hemorrhages; FI = fungal infections; NT = neurotoxoplasmosis; MNE = microglial nodular encephalopathy; WE = Wernicke’s encephalopathy.

Figure 2 Case No. 118: female, acute myeloid leukemia. Histological sections of post-BMT toxoplasmosis showing two pseudo-cysts (arrowheads) in a Virchow-Robin perivascular space with minimal inflammatory reaction (HE ×400).

severe, massive IPH that invaded the subarachnoid space). About half of the cerebrovascular lesions (13.33%) were old, gliotic and organized, while the remaining were recent. Bland cerebral infarcts were detected in 11 patients (6.11%). These were various sizes and ages. None of the cases were associated with nonbacterial thrombotic endocarditis (NTBE), and six cases were associated with fungal infections of the brain parenchyma. CNS infections

Figure 1 (a) Case No. 89: female, 10 years old, Fanconi anemia. Presence of a large collection of blood in subarachnoid space of right frontal lobe (䊉). (b) Case No. 8: female, 19 years old, severe aplastic anemia. Presence of massive intraparenchymal hemorrhage (considered as cause of death), with compression of the ventricle and distortion of the left laterial ventricle (**), and edema due to disruption of the blood brain barrier (arrowheads).

masses distorting adjacent neural structures. Nineteen patients (10.55%) suffered hemorrhagic lesions that were incompatible with life (Figure 1b). Despite being a common finding in both groups, SAH was not considered to be the main cause of death in any patient (simply an extension of a

CNS infections were documented in 27 patients (15%) and constituted the second most common group of lesions. Fungal abscesses were observed in 16 patients (eight cases of Aspergillus sp., five cases of Candida sp., one case of Fusarium sp. and two others), toxoplasmosis (Figure 2) patients (three of each were confirmed by immunohistochemistry), while bacterial infections were seen in three cases. The Aspergillus (Figure 3a) lesions showed typical necrotizing parenchymal lesions with vascular invasion (Figure 3b) and secondary hemorrhages, occasionally associated with meningitis. Candida abscesses demonstrated a much less prominent hemorrhagic component. The Toxoplasmosis lesions were usually single brain areas of softening, characterized histologically by an avascular central zone containing necrotic material. Surrounding this area was a hyperemic region containing numerous free parasites, occasional pseudo-cysts and mononuclear inflammatory cells. One patient had a very unusual hemorrhagic presentation of toxoplasma encephalitis.7 The bacterial lesions manifested as inflammation of the cerebrum and cerebellum with or without brain abscesses and/or purulent meningeal infiltrates. Isolated pathogens included Streptococcus epidermidis (two cases) and Staphylococcus aureus (one case). Metabolic encephalopathy Ten cases (5.55%) of Wernicke’s encephalopathy (WE) were identified. The main aspects of these cases have Bone Marrow Transplantation


304

Figure 3 Case No. 4: male, 28 years old, chronic myelogenous leukemia. (a) Coronal section of cerebritis demonstrating large, granular lesion with central hemorrhage (arrowheads). Edema produced compression of the adjacente parenchyma (★). (b) Histological sections of post-BMT Aspergillus sp. infection showing vascular thrombosis due to fungal vascular invasion. (HE ×400).

already been discussed in a previous publication8. All patients experienced nonspecific clinical symptoms, until physical symptoms of severe metabolic acidosis were detected between days +13 and +37 post BMT (median 22). Preceding symptoms of metabolic acidosis was the appearance of a raspberry tongue. Neurological findings included coma in six patients, confusion in five patients, papilledema in four, nystagmus in three and blindness in two patients. Neuropathological examination showed extensive periventricular foci of hemorrhage affecting particularly the thalamus, midbrain, pons and medula oblongata. No lesions were detected in the mammilary bodies (MBs). Histological findings were extravasation of red blood cells into the perivascular spaces, associated with foci of reactive neovascularization with swollen endothelium and fibrinoid necrosis of the vascular wall. Minimal neuronal loss and gliosis were also evident. Bone Marrow Transplantation

Figure 4 Case No. 72: female, 20 years old, chronic myelogenous leukemia. Histological section of brain tissue showing microglial hyperplasia. In (a) there are activated microglial cells (arrowheads) and oligodendrocytes (**) (HE ×400). In (b) immunohistochemistry with MoAB antiHLA/Dr-CR3/43 showing activated microglial cells (arrowheads) (DAB ×400).

Diffuse microglial hyperplasia (DMH) and microglial nodular encephalopathy (MNE) Due to the difficulties in identifying microglial cells and the reactive nature of these findings immunohistochemistry with moAb anti-HLA/DR-CR3/43 was used to confirm DMH and MNE.8 In 49 cases (27.22%) there was DMH (Figure 4a and b) amongst which there were 10 cases with numerous microglial nodules throughout brain parenchyma consistent with MNE (Figure 5). Of these 10 patients, CMV nuclear positivity was detected in four and Toxoplasma gondii positivity in three cases. Three cases were positive for both antigens. Both forms of toxoplasma infection (bradyzoites and tachyzoites) were noted in necrotic areas of the brain tissue.


Figure 5 Case No. 28: male, 25 years old, severe aplastic anemia. Nodular encephalopathy with microglial nodules in detail (**) demonstrating numerous rod cells (HE ×400).

Other CNS findings A total of 45 distinct neuropathological findings were detected in the 180 cases studied. Secondary CNS lesions included hypoxic/ischemic changes (56 patients), cerebral edema (49 patients), and granular ependimitis (23 cases), possible viral infection, cysticercosis, colloid cyst among others. CNS pathology as cause of death Thirty-one patients (17%) were found to have CNS problems as the cause of death. This group was studied separately and compared to the rest of the 149 patients (83%) whose cause of death was considered to be extra-cerebral (control group). There were 18 males (58%) and 13 females (42%). There were no significant statistical differences between the two groups according to sex, age and underlying disease. The survival time was almost half in the group ‘CNS as cause of death’ (96.29 vs 177.8 days) (P = 0.0162). The main neuropathological findings in this group were cerebrovascular disorders, fungal infections and neurotoxoplasmosis (Table 4). IHP (70.96 vs 27.22%) (P ⬍ 0.001), fungal infections (25.8 vs 8.88%) (P ⬍ 0.001) and toxoplasmosis (9.67 vs 4.44%) (P ⬍ 0.001) were significantly more frequent in this group than in the control group. Discussion BMT is a procedure which has significantly contributed to the definitive therapy of many patients afflicted with lifethreatening hematological malignancies, bone marrow aplasias and certain genetic and metabolic disorders.1–5 However, broad-ranging side-effects with negative consequences on the CNS are frequently encountered. Preparatory regimens are markedly cytotoxic, immunosuppressive, and capable of direct damage to many tissues

including the CNS.9,10 The full extent of GVHD upon the CNS, as well as the effects of post-transplant immunosuppressive agents used to prevent it, have not yet been fully established.9–11 Delayed engraftment leads to prolonged low platelet and leukocyte counts, with risk of hemorrhage and opportunistic infections. Nutritional requirements during periods of intense regenerative cell proliferation and other metabolic stresses must be met in order to avoid injury to susceptible tissues. Underlying malignant disease may not be completely eradicated by conditioning treatment and may recur despite engraftment of the new marrow. We prospectively studied 180 brains of patients who died following BMT (largest experience in the literature). It was found that neuropathological abnormalities occurred in 90.55% of patients and brain pathology was the main cause of death in 17% of our cases. Cerebrovascular complications were the most prevalent finding in our series occurring in 58.69% of the patients. This prevalence is the highest reported in the literature to date (58.69% vs 6.4%3 vs 27%5). A variety of circumstances may have contributed to the large number of hemorrhagic complications seen in our patients. Thrombocytopenia, secondary to chemotherapy or radiationinduced, or secondary to a decrease in megakaryocytes; and autoimmune or drug-induced mechanisms of platelet destruction or dysfunction, have always been considered major risk factors for CNS bleeding post BMT.4,5,12,13 Furthermore, infections and coagulopathies, with secondary liver dysfunction or disseminated intravascular coagulation, have also been correlated with bleeding disorders.3,5 Another contributory factor is the high prevalence of Fanconi anemia among our patients, known for its associated vascular and endothelial fragility, especially when exposed to alkylating agents;14 however, further studies are needed to confirm these observations. Cerebral infarcts were as prevalent as in previously reported studies (6.11% vs 8.77%,2 vs 3.8%,3 vs 12.84%5), even though there were no cases of NBTE in our series. Infections are widely recognized as the most common cause of death among BMT recipients.15 In the CNS, infections have frequently been reported as one of the most common neuropathological findings.2–5,16 BMT recipients are severely immunocompromised because of both the underlying diseases and the BMT regimen. After BMT, GVHD may impair immune recovery and is often itself treated with immunosuppressive agents. During the initial period of granulocytopenia which lasts for about 1 month after BMT, gram-negative sepsis, fungal infections and herpesvirus infections are common. There is a prolonged period of deranged cell-mediated and humoral immunity which may last up to a year, when interstitial pneumonitis, CMV infections and protozoan infections become more frequent.15,17 In our series, CNS infection was the second most commonly seen neuropathological finding after BMT. Etiologic agents were fungi (Aspergillus eight, Candida five, Fusarium one, others two), T. gondii eight and bacterial infections three cases. Fungal infections and toxoplasmosis were significantly more prevalent in the group that died of CNS causes than in the control group. WE was detected in 10 patients (5.5%) of our series. WE is characterized by a classical triad of altered mental status,

305

Bone Marrow Transplantation


306

ataxia and ophthalmoplegia,18 but the clinical course is usually oligosymptomatic.19 Mental status changes are the most frequent findings in these patients (82%), followed by ocular findings (29%), and ataxia (23%).7 Severe refractory metabolic acidosis was found in all of our patients preceding the symptoms of WE. Furthermore, our patients also presented with a ‘raspberry tongue’, which may be an important clinical finding in the early detection of iatrogenic WE.18 Histologically, all had focal periventricular lesions of the gray matter in the medulla oblongata, which is a very characteristic finding of WE.20 The hypothalamus was the second most affected area. The mammilary bodies (MBs) were not affected in any of our patients, supporting the hypothesis21 that the MBs were the last sites to be affected by thiamine deficiency, and suggesting that acute states of thiamine depletion can present without any histological finding of MBs lesions. MNE is characterized by the presence of microglial nodules and macrophage cells in the CNS,22 usually associated with viral (HIV) encephalitis.23 Very few reports have described the presence of MNE, associated with CMV infection3 or varicella encephalitis5 after BMT. In our patients, MNE was noted in 10 of 49 patients with diffuse rod cell hyperplasia. This finding was confirmed by immunohistochemistry to demonstrate activated microglial cells expressing the HLA-DR receptor,8 since these cells present this surface antigen after immune stimulation.24 These 10 positive cases were further studied to possible etiological agents, such as CMV and T. gondii. Four cases were positive for CMV infection, three showed positivity for T. gondii and three were positive for both antigens. The other three cases were not conclusive due to technical difficulties. This is a striking new finding whose clinical significance and diagnostic utility are still to be determined in BMT patients. In previous reports concerning neurologic and neuropathological complications following BMT, the cases where CNS pathology is the main cause of death did not exceed 10% (10%,2 6%3,5). Seventeen percent of the patients in our series had CNS problems as the main cause of death. The most commonly seen neuropathological findings in this group were IPH (70.96%), fungal infections (25.8%) and toxoplasmosis (9.67%). These findings were significantly more prevalent in this group than in the control group (P ⬍ 0.001), supporting the evidence these findings may cause massive and fatal destruction of the brain parenchyma in the post-BMT period. Furthermore, this group had a survival time of almost half that of the control group (96.29 vs 177.8 days) (P = 0.0162). Overall, we have demonstrated an increased prevalence of neuropathological complications and an increased rate of fatal CNS findings after BMT than has been previously reported. IHP, fungal infections and toxoplasmosis were demonstrated to be major CNS risk factors with high mortality rates in BMT patients. Furthermore, the great variability of neuropathological abnormalities seen after BMT might reflect the heterogeneity of possible etiological factors. Nevertheless, an awareness of the main types of neuropathological complications seen after BMT is important in order to make an accurate diagnosis and instigate appro-


priate management complications.

of

these

potentially

dangerous

Acknowledgements We are indebted to Miss Heliz Regina Neves and Marco Bittencourt, MD for data analysis help.

References 1 Armitage JO. Bone marrow transplantation. New Engl J Med 1994; 330: 827–838. 2 Wiznitzer M, Packer RJ, August CS, Burkey ED. Neurological complications of bone marrow transplantation in children. Ann Neurol 1984; 16: 569–576. 3 Patchell RA, White CL III, Clark AW et al. Neurologic complications of bone marrow transplantation. Neurology 1985; 35: 300–306. 4 Graus F, Saiz A, Sierra J et al. Neurologic complications of autologous and allogeneic bone marrow transplantation in patients with leukemia: a comparative study. Neurology 1996; 46: 1004–1009. 5 Mohrmann RL, Mah V, Vinters HV. Neuropathological findings after bone marrow transplantation: an autopsy study. Hum Pathol 1990; 21: 630–639. 6 Colvin RB, Bhan AK, McCluskey RT. Diagnostic Immunopathology, 2nd edn. Raven Press: New York, 1995. 7 Bleggi-Torres LF, Medeiros BC, Werner B et al. Unusual presentation of cerebral toxoplasmosis after BMT. Bone Marrow Transplant (accepted for publication). 8 Graeber MB, Bise K, Mehraein P. CR3/43, a marker to human activated microglia: application to diagnostic neuropathology. Neuropathol Appl Neurobiol 1994; 20: 406–408. 9 Sullivan KM, Storb R, Shulman HM et al. Immediate and delayed neurotoxicity after mechlorethamine preparation for bone marrow transplantation. Ann Intern Med 1982; 97: 182–189. 10 Zimmer WE, Hourihane JM, Wang HZ, Schriber JR. The effect of human leukocyte antigen disparity on cyclosporine neurotoxicity after allogeneic bone marrow transplantation. Am J Neuroradiol 1998; 19: 601–610. 11 Antonini G, Ceschin V, Morino S et al. Early neurologic complications following allogeneic bone marrow transplant for leukemia: a prospective study. Neurology 1998; 50: 1441– 1445. 12 Mehta J, Powles R, Singhal S et al. Early identification of patients at risk of death due to infections, haemorrhage, or graft failure after allogeneic bone marrow transplantation on the basis of leukocyte counts. Bone Marrow Transplant 1997; 19: 349–355. 13 Openshaw H, Slatjin NE. Neurological complications of bone marrow transplantation. In: Forman SJ, Blume KG, Thomas ED (eds). Bone Marrow Transplantation. Blackwell Scientific Pub: Boston, 1994, pp 482–496. 14 Zanis Neto J, Ribeiro RC, Medeiros CR et al. Bone marrow transplantation for patients with Fanconi anemia: a study of 24 cases from a single institution. Bone Marrow Transplant 1995; 15: 293–298. 15 Bombi JA, CArdesa A, Llebaria C et al. Main autopsy findings in bone marrow transplant patients. Arch Pathol Lab Med 1987; 111: 125–129. 16 Bleggi-Torres LF, Medeiros BC, Neto JZ et al. Disseminated fusarium sp infection affecting the brain of a child after bone


17

18

19 20

marrow transplantation. Bone Marrow Transplant 1996; 18: 1013–1015. Proctor SJ, Jackson GH. Intra-cerebral aspergillosis following allogeneic marrow transplantation responding to treatment with non-liposomal amphotericin B and rifampicin. Br J Haematol 1997; 98: 1048–1053. Bleggi-Torres LF, Medeiros BC, Ogasawara VSA et al. Iatrogenic Wernicke’s encephalopathy in allogenic bone marrow transplantation: a study of eight cases. Bone Marrow Transplant 1997; 20: 391–395. Harper C. The incidence of Wernicke’s encephalopathy in Australia – a neuropathological study of 131 cases. J Neurol Neurosurg Psychiatry 1983; 49: 341–345. Peixoto MAL, Santos EC, Pittella JEH. Coma and death in

21

22 23 24

unrecognized Wernicke’s encephalopathy. Arq Neuropsiquiat 1992; 50: 329–333. Zhang SX, Weilersbacher GS, Henderson SW et al. Excitotoxic cytopathology, progression and reversibility of thiamine deficiency-induced diencephalic lesions. J Neuropathol Exp Neurol 1995; 54: 255–267. Budka H. Neuropathology of human immunodeficiency virus infection. Brain Pathol 1991; 1: 163–175. Lantos PL, McLaughlin JE, Scholtz CL et al. Neuropathology of the brain in HIV infection. Lancet 1989; 11: 309–311. Lassman H, Schmied M, Vass K et al. Bone marrow derived elements and resident microglia in brain inflammation. Glia 1993; 7: 19–24.

307
