Opportunistic CNS infection after bone marrow transplantation - Nature

Bone Marrow Transplantation, (1999) 23, 1167–1176  1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt

Opportunistic CNS infection after bone marrow transplantation M Maschke1, U Dietrich2, M Prumbaum3, O Kastrup1, B Turowski4, UW Schaefer3 and HC Diener1 1 4

Department of Neurology, 2Department of Neuroradiology, 3Clinic for Bone Marrow Transplantation, University of Essen; and Department of Neuroradiology, University of Frankfurt, Germany

Summary: We retrospectively identified opportunistic CNS infections in 655 patients who had undergone allogeneic, syngeneic or autologous BMT or PBSCT between 1990 and 1997. Twenty-seven patients (4%) developed CNS infections. All CNS infections occurred in allogeneic BMT or PBSCT patients. The most common CNS infections were toxoplasma encephalitis (74%) and cerebral aspergillosis (18%). Furthermore, we identified one patient with candida encephalitis and one patient with viral encephalitis. Overall mortality of patients with opportunistic CNS infection was 67%. There were two different groups of toxoplasma encephalitis with a different appearance on MR imaging. The first group showed edema, but no gadolinium enhancement, whereas the second group exhibited typical MRI appearances with the exception of frequent hemorrhagic transformation. The first group had a significant shorter latency between BMT and onset of CNS infection (mean 45 days vs 180 days, P = 0.02), a significant higher daily dose of corticosteroids as treatment for graft-versus-host disease (GVHD) (P = 0.01), more severe GVHD and a higher mortality (71% vs 36%). This study shows that the most common CNS infections in our patient population are toxoplasma encephalitis and cerebral aspergillosis, that there are two distinct subgroups of toxoplasma encephalitis and that CNS infections occur after allogeneic BMT only. Keywords: CNS infection; cerebral toxoplasmosis; cerebral aspergillosis

Both allogeneic and autologous BMT are associated with several neurologic complications secondary to the underlying disease and prolonged myelosuppression.1–4 The use of immunosuppressive drugs such as corticosteroids and cyclosporine in order to avoid rejection of the allograft and to control graft-versus-host disease (GVHD) is another major risk factor for neurologic complications especially CNS infections. In previous studies the most common complications were cerebral hemorrhage, metabolic encephalopathy and CNS infection.1–5 Neurologic complications were more frequent in patients when AML was the underlying disease than in other leukemia patients.4 Correspondence: Dr M Maschke, Department of Neurology, University Clinic of Essen, Hufelandstr. 55, 45122 Essen, Germany Received 12 August 1998; accepted 17 January 1999

Infections are a major cause of morbidity and mortality and represent the most significant barrier to both immediate and long-term survival after BMT.6,7 CNS infections occurred mostly in patients who had undergone allogeneic BMT and are reported to occur in 3–8% of patients after allogeneic, syngeneic or autologous BMT.1–4 CMV encephalitis and cerebral aspergillosis are the most common CNS infections, whereas cerebral toxoplasmosis seems to be a rare condition.2,4 Fungi were isolated from brain abscesses in up to 90% of cases. Aspergillus species were the most prevalent fungi and Candida species were second in frequency.8 Previous studies have preferentially focused on general neurological complications secondary to BMT and reviewed predominantly autopsy files.2–4 In the present study, we retrospectively analyzed clinical data and re-evaluated MR imaging of patients with CNS infections after BMT to obtain cause, frequency, time of onset and clinical signs of these infections. Patients and methods From January 1990 to October 1997, 655 patients underwent allogeneic, syngeneic or autologous BMT or peripheral blood stem cell transplantation (PBSCT) at the University Clinic of Essen (basic characteristics see Table 1). Seventy-four patients received autologous BMT and/or autologous PBSCT (11%), whereas 571 patients were treated with allogeneic BMT and/or allogeneic PBSCT (87%). Syngeneic BMT and/or syngeneic PBSCT were performed in 10 patients (1%). The mean age of patients was 38.9 years (15–59 years). Underlying diseases were CML (n = 273, 42%), AML or ALL (together n = 238, 36%) and miscellaneous diseases such as plasmocytoma or myelodysplastic syndrome (n = 144, 22%). Clinical data and complications after BMT were routinely recorded in a database in the BMT department. Additional information was obtained from chart records of hospitalized patients and follow-up records in our outpatient clinic in the BMT department and in the neurological department. Data of magnet resonance (MR) imaging were kept in a database and images were saved as hard copies and on optical disks. We reviewed the records and MR imaging of all patients diagnosed as having CNS infections between BMT and death or the last follow-up visit. The diagnosis of CNS infection was confirmed by two or more of the following conditions: (1) autopsy or biopsy studies; and/or (2) typical neuroradiological images; (3) neuroradiological and clinical response to specific unidirectional antimicrobial chemotherapy; and/or (4) positive serology (Table 2). The last

Opportunistic CNS infection after BMT M Maschke et al

1168

Table 1 Basic characteristics of patients who had undergone BMT or PBSCT

Total Age years (range) Sex female male Type of BMT/PBSCT Autologous BMT/PBSCT Allogeneic BMT/PBSCT Syngeneic BMT/PBSCT BMT alone PBSCT alone BMT + PBSCT Underlying disease CML ALL + AML miscellaneous GVHD grade I grade II grade III + IV Mortality in the first year overall mortality

Table 2

n

%

655 39 (15–59)

100

276 379

42 58

74 571 10 532 122 1

11 87 1 81 19 0.2

273 238 144 387 194 86 107

42 36 22 59 30 13 16

235 291

36 44

Diagnostic criteria for CNS infection

and/or and

1 2 3

and/or

4

Autopsy or biopsy studies were available Typical neuroradiological images on MR Neuroradiological and clinical response to specific unidirectional antimicrobial chemotherapy Positive serology

condition only applied to cases of toxoplasma encephalitis, given the fact that all other CNS infections were confirmed by autopsy. ‘Positive serology’ was determined by: (1) positive IgM and/or positive IgG serum antibodies against Toxoplasma gondii; (2) negative Aspergillus and Candida antigen; and (3) negative CMV pp65 antigen. Patients with an uncertain diagnosis or a diagnosis of another secondary neurologic complication, as for example, metabolic encephalopathies or subdural hematoma due to BMT were excluded from the analysis. GVHD was a frequent accompanying disease in our patients and was diagnosed in 387 patients (59%). We graded the severity of this condition according to the involvement of skin, liver and gastrointestinal system by using the clinical classification described by Thomas et al.9 Grade I GVHD was found in 194 patients (29.6% of all patients), grade II in 86 patients (13.1%) and grades III and IV in 107 patients (16.3%). All patients received variable doses of immunosuppressive drugs. To allow comparisons between different patient groups we calculated the mean corticosteroid dose with the following formula which accounts for the stronger glucocorticoid effect of methylprednisolone in comparison to prednisone: prednisone (mg) + [methylprednisolone (mg) ⫻ 1.25]/n = mean corticosteroid dose/day. Descriptive statistics of the main characteristics (age, underlying disease, type of infection, latency between onset

of CNS infection and BMT, mean corticosteroid amount, severity of GVHD, leukocyte count) of the patients were performed. Continuous variables were compared by independent t-test using SPSS. Equal variances were not assumed in the t-test. Values below or above two standard deviations were excluded from comparison by means of the t-test. Results We identified 27 patients (4%) with opportunistic CNS infection from 655 patients who had undergone BMT or PBSCT. Basic characteristics of patients with identified CNS infection (females = 11, males = 15) are shown in Table 3. Mean age was 38.2 years (⫾10.5 years). All CNS infections occurred in patients undergoing allogeneic BMT (n = 25) or allogeneic PBSCT (n = 2). Underlying diseases were CML (n = 15; 56%), ALL (n = 6; 22%), AML (n = 4; 15%), myelodysplastic syndrome (n = 1; 4%) and plasmocytoma (n = 1; 4%). In 12 patients autopsy or biopsy studies were available (44.4% of patients included). In our population, toxoplasma encephalitis was the most common opportunistic CNS infection. Twenty patients had toxoplasma encephalitis (74%), five patients aspergillus encephalitis (18%), one patient candida encephalitis (4%) and one patient viral encephalitis (4%). Two patients showed the criteria of GVHD grade 0, four patients had grade I, seven patients had grade II, seven patients had grade III and seven patients had grade IV. The leukocyte count at time of CNS infection was obtained from the chart records (unavailable in two patients). The mean leukocyte count of the remaining 25 patients was 4520 ⫾ 3086/␮l. Treatment for GVHD included variable doses of corticosteroids, cyclosporine, antithymocyte globulin and monoclonal antibodies. Overall mortality of patients with opportunistic CNS infections was 67% from time of BMT or PBSCT until October 1997 (n = 18/27). Toxoplasma encephalitis Toxoplasma encephalitis was diagnosed in 20 patients (four diagnoses confirmed by autopsy or biopsy studies, 16 diagnoses from neuroradiological and clinical criteria as described above). Cerebral toxoplasmosis occurred from day 14 to day 689 after BMT/PBSCT (median 72.5, mean 123 ⫾ 150). Clinical signs were focal and generalized seizures (n = 9), neuropsychological and cognitive disturbances (n = 7), headache (n = 5), focal cerebral signs (hemiparesis, hemianopia, ataxia, dysarthria) (n = 5), and impairment of consciousness (n = 4) (Table 4). Serological findings of these patients are shown in Table 5. For the MRI study MR images from 18 of 20 patients were available. Diagnoses of the remaining two patients were confirmed by autopsy studies. The overall mortality in the group of patients with toxoplasma encephalitis (not necessarily due to toxoplasma encephalitis) was 55% (11/20 patients). MR imaging of patients with toxoplasma encephalitis revealed two differing main features. There were patients with multiple areas of signal hyperintensity on T2-weighted images (T2WI) in supra- and infra-tentorial locations (Figure 1a).


Table 3 No.

1169

Basic characteristics of patients with CNS infection GVHD (grade)

Leukocytes per ␮l/ neutrophils in %

Age Sex

ToI

ToT

UD

I-BMT (d)

1

51

f

toxo

BMT

AML

14

S II

1000/50

2

38

m

toxo

PBSCT

ALL

33

S, GI III

2000/50

3

52

f

toxo

BMT

ALL

34

S, GI IV

9700/92

4

29

m

toxo

BMT

CML

34

S, L, GI IV

3500/?

5

49

m

toxo

BMT

55

0

6

35

m

toxo

BMT

IgG plasmocytoma ALL

55

S, L, GI IV

3900/95

7

45

m

toxo

BMT

CML

62

S III

8600/60

8

39

m

toxo

BMT

CML

64

S III

?

9 10

31 29

m f

toxo toxo

BMT BMT

CML MDS

65 70

SI S, L, GI IV

3100/50 2200/97

11

28

f

toxo

BMT

CML

75

S, GI II

4000/95

12

17

f

toxo

BMT

ALL

78

S II

2000/?

13

27

m

toxo

BMT

CML

85

S II

2000/70

14

31

m

toxo

BMT

CML

94

SI

3400/65

15

54

f

toxo

BMT

CML

116

SI

8200/85

16 17

50 30

f f

toxo toxo

BMT BMT

CML CML

144 187

S, L II S II

3200/60 9100/80

18 19

48 42

m f

toxo toxo

BMT BMT

CML CML

250 266

S, GI III S, L III

9200/84 5600/70

20 21 22 23 24 25

37 51 33 20 40 33

m f f m m m

toxo asperg asperg asperg asperg asperg

PBSCT BMT BMT BMT BMT BMT

CML ALL, Ph− AML AML CML AML

689 39 109 110 168 290

S, L II S, GI, L IV S, GI, L IV S, GI, L IV SI S III

7500/66 3900/60 9400/87 2000/75 1100/50 4400/95

26 27

38 55

m m

candida viral

BMT BMT

ALL CML

24 88

0 S, L III

4300/45 3100/94

?

Immunosuppress drugs PR 150 CsA 200 SDH 500 CsA SDH 240 CsA SDH 2 g CsA ATG Pr 100 CsA SDH 500 CsA OKT PR 200 CsA PR 200 CsA OKT PR 80 SDH 300 OKT PR 180 CsA PR CsA PR 30 CsA PR 125 CsA PR 62.5 CsA SDH 100 SDH 30 CsA PR 45 CsA PR 200 CsA SDH 2 g SDH 1 g CsA SDH 1g CsA PR 25 PR 45 CsA CsA SDH 1 g CsA

ToI = type of infection; toxo = toxoplasma encephalitis; asperg = cerebral aspergillosis; candida = candida encephalitis; viral = viral encephalitis; ToT = type of transplantation; UD = underlying disease; I-BMT = interval between BMT and CNS infection in days; GVHD = graft-versus-host disease; S = skin; L = liver; GI = gastrointestinal; PR = prednisone in mg, CsA = cyclosporine A; SDH = methylprednisolone in mg or gram (g); OKT = monoclonal antibodies; ATG = antithymocyte globulin; ? = information at time of infection not available.

These lesions showed minimal space-occupying effects. On T1-weighted images (T1WI) they were hypointense or not visible at all and demonstrated no contrast enhancement (Figure 1b). Another group of patients showed typical MRI appearance of toxoplasma encephalitis with multiple lesions on T1WI and T2WI and ring-enhancement with perifocal edema (Figure 2). In half of these patients lesions were hyperintense on non-contrast T1WI indicating hemorrhage (Figure 3). A comparison of the clinical characteristics of these two groups is shown in Table 6.

Most common locations of the MRI lesions were subcortical white matter, basal ganglia and cerebellum. The first type of MRI appearance as described above was found in seven patients. The infection in these patients manifested itself from day 14 to day 75 (mean 45 ⫾ 22 days). Underlying diseases were AML (n = 1), ALL (n = 2), CML (n = 2), plasmocytoma (n = 1) and MDS (n = 1). Leukocyte counts were 1000–9700/␮l (mean 3733 ⫾ 3116/␮l) at the time of infection. GVHD was grade 0 in one patient (14%), grade I in one patient (14%), grade II in one patient


1170

(14%), grade III in one patient (14%) and grade IV in three patients (43%). Immunosuppressive therapy included prednisone (100–180 mg/day, n = 3), methylprednisolone (240–2000 mg/day, n = 4), cyclosporine (n = 6) and monoclonal antibodies (n = 1). Mean corticosteroid dose was 604 ⫾ 854 mg/day. Treatment for toxoplasma encephalitis was begun as soon as the diagnosis was clinically suspected using a standard therapy regimen with pyrimethamine (100 mg/day) and sulfadiazine (4–8 g/day). Five patients (71%) showed clinical and neuroradiological improvement after initiation of treatment. Two of the five patients who had shown improvement with specific therapy died later of heart failure and pneumonia, unrelated to the toxoplasma encephalitis. Overall mortality of these seven patients was 71% (5/7 patients). The second MRI appearance with typical MRI appearance of toxoplasma encephalitis was found in 11 patients. The interval between BMT and onset of cerebral toxoplasmosis for these patients was between 62 and 689 days (mean 180 ⫾ 184.4 days). Underlying diseases were CML (n = 10) and AML (n = 1). The mean leukocyte count at the time of infection was 5370 ⫾ 2983/␮l (2000–9200/␮l). Immunosuppressive therapy included prednisone (5– Table 4 litis

Clinical symptoms of 20 patients with toxoplasma encepha-

Clinical symptoms

No. of patients (%)

Seizures Neuropsychological signs Focal signs Headaches Coma

Table 5

9 7 5 5 4

(45) (35) (25) (25) (20)

200 mg/day, n = 9), methylprednisolone (30–100 mg/day, n = 2), cyclosporine (n = 8) and monoclonal antibodies (n = 2). The mean corticosteroid amount was 89 ⫾ 70 mg/day. In this group GVHD was classified as grade I in three patients (28%), grade II in four patients (36%) and grade III in four patients (36%). Overall mortality of these 11 patients was 36% (4/11 patients).

Cerebral aspergillosis The diagnosis of aspergillus encephalitis was made in five patients. All diagnoses were confirmed by autopsy studies. Time of onset of cerebral aspergillosis was at minimum day 39 and at maximum day 290 after BMT/PBSCT (median 110, mean 143 ⫾ 94). Neurological symptoms and signs were neuropsychological disturbances (n = 4), headache (n = 3), hemianopia (n = 3), focal and generalized seizures (n = 2), coma (n = 2) and hemiparesis (n = 1). The mean leukocyte count was 4160 ⫾ 3225/␮l. Grade I of GVHD was found in one patient (20%), grade III in one patient (20%) and grade IV in three patients (60%). Patients received immunosuppressive therapy with methylprednisolone (1–2 g/day, n = 3), prednisone (25–45 mg/day, n = 2) and cyclosporine (n = 3). MR images on T2WI revealed large lesions with an intermediate signal and with surrounding edema. In all patients ring-enhancement could be demonstrated (Figure 4). In three of five patients we found a significant mass effect and in two patients lesions showed hemorrhagic transformation. Lesions were localized cortically, subcortically, in the basal ganglia and cerebellum. All patients died due to central nervous complications of cerebral aspergillosis.

Results of serological tests of 20 patients with toxoplasma encephalitis

No.

I-BMT (d)

Aspergillus antigen

Candida antigen

Candida Cryptococcus antibody IHA antigen increase

CMV Toxoplasma IgG Toxoplasma IgG pp65 AGG IHA

1 2 3 4 5 6 7 8 9 10 11 12 13 14

14 33 34 34 55 55 62 64 65 70 75 78 85 94

— — — — — — — — — — — — — —

— — — — — — — — — — — — — —

no no no no no no no no no no no no no no

— — — — — — — — — — — — — —

— — — — — — — — — — — — — —

15 16 17 18 19 20

116 144 187 250 266 689

— — — — — —

— — — — — —

no no no no no no

— — — — — —

— — — — — —

Toxoplas. IgG EIA

1:64 ⇒ 1:540 1:128 ⇒ 1:256 NA pos. pos. NA pos. pos. NA NA NA pos. 1:16 ⇒ 1:32.000 1:32 ⇒ 1:4096 NA NA NA pos. NA NA ⬎500 IE/ml NA NA pos. 1:540 1:32 NA 1:64 ⇒ 1:540 1:128 ⇒ 1:256 NA NA NA pos. 1:64 ⇒ 1:16.384 NA NA NA NA pos. NA NA 1:4000 IE/ml 1:180 1:64 NA 1:180 ⇒ 1:540 NA NA 1:1620 1:256 NA NA NA pos. pos. pos. NA 1:4000 pos. NA

Toxoplas. IgM

Toxoplas. IgG donor

Biopsy/Autopsy

neg. neg. neg. neg. pos. neg. neg. neg. neg. neg. pos. pos. neg. pos.

pos. pos. pos. pos. pos. NA pos. pos. pos. pos. NA pos. pos. pos.

NA NA NA NA NA yes, pos. NA NA yes, pos. NA NA NA NA NA

neg. pos. pos. pos. neg. neg.

pos. NA pos. NA NA pos.

yes, pos. NA NA NA NA yes, pos.

Candida antigen = immunohemagglutination; Aspergillus antigen = latex agglutination; Cryptococcus antigen = enzyme-linked immunoassay; Toxoplasma: AGG = immunosorbent agglutination; IHA = immunohemagglutination, EIA = enzyme-linked immunoassay; NA = not available; pos = positive; neg = negative; I-BMT ⫽ interval between BMT and CNS infection in days.


1171

Figure 1 Cerebral toxoplasmosis without contrast enhancement. Axial T2-weighted images (T2WI: 2000/80/1) show right occipital, left occipital and left frontal lesions (arrows) with high signal intensity, note additional bilateral subdural effusions (a). Corresponding postcontrast T1-weighted images (T1WI: 570/15/1) reveal diffuse edema but no enhancing lesion (b). Toxoplasmosis was confirmed by clinical and radiological improvement following 1 month of therapy (c, d).

Candida encephalitis We identified one patient with candida encephalitis. The diagnosis was confirmed by autopsy. The interval between BMT and the diagnosis of candida encephalitis was 24 days. The patient received 150 mg cyclosporine at the time of infection and had a leukocyte count of 4300/␮l. GVHD was not apparent. MR images showed multiple hypointense lesions on T1WI, ring-enhancement after gadolinium application and intermediate-signal lesions within surrounding

high-signal on T2WI. The lesions were localized to the cerebellum and the basal ganglia. The patient died 19 days after onset of symptoms of respiratory failure. Viral encephalitis One patient died of brain stem hematoma probably due to a viral encephalitis, as revealed by autopsy. The underlying disease was CML. The interval between BMT and infection


1172

Figure 2 Cerebral toxoplasmosis with contrast enhancement. T2WI (a) and T1WI (b) reveal a large right temporal lesion with rim enhancement and multiple small lesions in the left basal ganglia with nodular and rim enhancement.

was 88 days. The patient had a leukocyte count of 3100/␮l at the time of infection. GVHD was grade III and the patient received immunosuppressive therapy with 1000 mg methylprednisolone per day and cyclosporine. Clinical signs were nystagmus, diplopia and impairment of consciousness. MR imaging revealed ill-defined hippocampal, cerebellar and brain stem hyperintense lesions on T2WI with hemorrhagic transformation on T1WI and intense gadolinium enhancement. Discussion CNS infections and metabolic encephalopathies are the most frequent neurologic complications following allogeneic BMT.1–5,10 The present study focused on the frequency and cause of opportunistic CNS infections in a large population of patients who had undergone BMT or PBSCT. A frequency of 4% of CNS infections after BMT or PBSCT in our population is comparable to that in former studies. Graus et al4 reported an occurrence of 3% of CNS infections in allogeneic BMT patients, whereas Patchell et al2 found 8% of CNS infections in an autopsy study. We found CNS infections only in patients who had undergone allogeneic BMT or PBSCT. This is in accordance with the study of Graus et al,4 who compared neurologic complications after allogeneic and autologous BMT and identified no CNS infections in patients after autologous BMT. In the present study the overall mortality of patients with opportunistic CNS infection was 67%, which is comparable to the high mortality found in former studies.4 In our series the most common infections were toxoplasma encephalitis and cerebral aspergillosis. This differs from the results in previous post-mortem and clinical studies,1–4,8,10 which reported CMV encephalitis and aspergillosis to be the most frequent cerebral infections.

One explanation for the fact that we did not identify patients with CMV encephalitis is the improvement in early diagnosis and management in the last years. In our hospital each patient was examined twice weekly for active CMV infection as indicated by markers of replication (pp65-antigen, polymerase chain reaction of virus-DNA) from day 20 to month 2 after BMT. Thereafter, pp65-antigen was sought monthly. In the event of active replication the patient received specific antiviral treatment. Clinical studies have shown that the use of CMV-negative blood products in transplantation if donor and patient lack CMV antibodies at the time of BMT, as well as prophylactic administration of acyclovir or ganciclovir to patients at high risk serve to reduce the incidence of systemic and of cerebral CMV infection.11–13 Toxoplasma encephalitis occurs frequently in patients with acquired immunodeficiency syndrome (AIDS).14–16 Prior to AIDS, fulminate necrotizing encephalitis due to toxoplasma infection was observed in patients with significant immunodeficiency such as collagen vascular disease, underlying malignancy, organ transplantation, and patients maintained on steroids, chemotherapy or radiation therapy.17–19 It seemed to be a rare condition in BMT patients. Winston et al6 (UCLA transplant group) reported one case only out of 60 patients after BMT. Patchell et al2 described another patient with presumed toxoplasma encephalitis with a non-diagnostic brain biopsy. Graus et al4 also identified only one patient with toxoplasma encephalitis from 425 patients who had undergone BMT. Another large study was performed by Slavin et al20 who found 12 patients with toxoplasmosis of different organs out of 3294 patients after allogeneic BMT. Two of these patients had toxoplasma encephalitis confirmed by brain biopsy and responded to therapy with pyrimethamine and sulfadiazine, but died later from leukemia relapse. There are a few more single case reports, demonstrating cerebral toxoplasmosis in patients


1173

Figure 3 Cerebral toxoplasmosis with hemorrhage. Non-enhanced axial T1WI demonstrates a diffuse left temporal lesion (large arrow) and multiple small lesions (small arrows) with high signal (a). Postcontrast T1WI shows slight enhancement adjacent to the larger lesion (arrow, b). Hemorrhage was confirmed by computed tomography (c).

after BMT.21,22 However, in an autopsy study performed from 1972 to 1993 at our university, Schwechheimer et al23 found six patients with toxoplasma encephalitis out of nine with severe neuropathological problems after BMT. One reason for the different frequency of toxoplasma encephalitis in comparison to studies which were performed in the USA1–3,6,20 might be the high endemicity of toxoplasmosis reflected by seropositivity of the general population of 70– 96% in Germany and much of Europe24 in contrast to the USA of 3–68%.24,25 Another explanation for the common finding of toxoplasma encephalitis in our study may be the improvement in neuroradiological diagnosis techniques in

comparison to former studies without MRI in some of their patients.20 The neuroradiological findings of cerebral toxoplasmosis on computer tomographic images (CT) typically consists of multiple areas of iso- or hypodensity with ring or nodular enhancement on postcontrast CT.26,27 On T2weighted images, lesions are usually hyperintense. On T1weighted images lesions show iso- or hypointensity. After application of gadolinium MRI reveals ring or nodular enhancement, which is clearly distinguishable from surrounding edema.26,28,29 The MRI appearance in the second group of patients with toxoplasma encephalitis in our population was similar to the characteristic findings in MRI as


1174

Table 6

Comparison of the two different groups of toxoplasma encephalitis defined as differences in MR appearance Group 1

MR characteristics T2 images T1 images Gd-enhancement No. of patients Mortality (%) I-BMT (mean in days) Underlying disease AML ALL CML miscellaneous Mean leukocyte count/␮l Immunosuppressive therapy methylprednisolone prednisone cyclosporine ATG, OKT mean corticosteroids GVHD grades 0–II grades III–IV

Group 2

multiple lesions with edema hyperintense signal lesions no signal no 7 71% 45 ± 22

multiple lesions with edema, partial hemorrhagic transformation and ringenhancement hyperintense signal lesions hyperintense signal lesions yes, ring-enhancement 11 36% 180 ± 184

P value

0.02*

1 2 2 2 (plasmocytoma, MDS) 3733 ± 3116

— 1 10 — 5370 ± 2982

0.31

4 (240–2000 mg) 3 (100–180 mg) 6 1 604 ± 854 mg

2 (30–100 mg) 9 (5–200 mg) 8 2 89 ± 70 mg

0.01*

3 4

7 4

MR = magnet resonance; Gd = gadolinium; I-BMT = interval between BMT and CNS infection in days; ATG = antithymocyte globulin, OKT = monoclonal antibodies, mean corticosteroids = prednisone (mg) + [methyprednisolone (mg) ⫻ 1.25]/n; * = significant.

Figure 4 Cerebral aspergillosis. Multiple lesions in the left cerebellar hemisphere with dark rings on T2WI (a) and moderate rim enhancement on postcontrast T1WI (b) suspicious of fungal infection.

described above with the following exception: half of the patients in the second group had hyperintense lesions on MRI on non-enhanced T1-weighted images, suspicious of hemorrhagic transformation in these lesions. This could be verified on corresponding CT scans in one patient. Even though hemorrhage in toxoplasma lesions is uncommon, it has been reported.20,29 Most neuroradiological toxoplasma

studies have been carried out in patients with AIDS.26–28 In view of the fact that in that situation toxoplasma encephalitis seemed rare in patients after BMT little was known about the typical MRI appearance in these patients. One important difference from patients with AIDS is the usually significant reduction in white cells, red cells and platelets after BMT. The reduction in platelet numbers may lead to


more frequent hemorrhage in toxoplasma lesions in patients after BMT in comparison to the findings in patients with AIDS. In contrast to group two patients with toxoplasma encephalitis the MRI appearance in group one patients was dissimilar to the characteristic MRI features in cerebral toxoplasmosis usually described. In this group MRI imaging revealed only hyperintense lesions in T2-weighted images, localized to subcortical white matter, basal ganglia and cerebellum. In all seven patients in this group we failed to identify MRI features characteristic of CMV encephalitis, eg contrast-enhancing lesions in the subependymal and cortical regions, periventricular rarefaction or progressive ventriculomegaly.30,31 In addition, the pp65-antigen, PCR of herpes-simplex- and varicella-zoster-virus DNA, candida antigen and aspergillus antigen were negative at the time of infection and later on. All patients were positive for IgG antibodies against toxoplasmosis prior to transplant or at the time of infection, so that most of the reported cases of encephalitis were attributed to reactivation, even though toxoplasmosis could have been acquired from donor marrow.32 No serologic test seemed particularly useful in the diagnosis of toxoplasma infection after BMT in former studies20 or in the present study. Besides MRI appearances there were other differences between groups one and two with toxoplasma encephalitis (shown in Table 6). The most important difference was the lower leukocyte count in group one patients. A previous study also reported lack of contrast enhancement in brain parenchymal toxoplasma infection in patients with low leukocyte counts after BMT.33 They explained their findings as a reduced inflammatory response due to neutropenia. This might be also an explanation for the lack of contrast-enhancement in the group one patients in the present study. In summary, severe GVHD and use of high doses of immunosuppressive drugs were risk factors for developing toxoplasma encephalitis in the first 3 months after BMT. The strong association between toxoplasmosis and GVHD and its treatment has also been observed by others.20,34–37 Early toxoplasma encephalitis was associated with a higher mortality than infection occurring later in the course of illness. Aspergillus and candida encephalitis showed similar clinical features and MRI appearances, as described in former studies.38–40 In immunocompromised patients aspergillus encephalitis occurs as part of a disseminated infection and invasive aspergillosis is an important cause of death from infection,41 as confirmed in the present study. We found a mortality of 100% in accordance with former studies.4,38–41 In conclusion, the present study showed toxoplasma encephalitis as the major cause of opportunistic CNS infection after allogeneic BMT. There were two types of toxoplasma encephalitis, which differed in time of onset after BMT, clinical and neuroradiological features and mortality. Risk factors for the occurrence of the early onset cerebral toxoplasmosis were severe GVHD and use of high-dose immunosuppression. Because of the nebulous value of serologic tests20 and uncertain diagnostic value of PCR42 in detection of toxoplasma infection, MR imaging should be performed as early as possible after development of neuro-

logical signs and specific treatment should be started as soon as toxoplasma encephalitis is suspected.

References 1 Davis DG, Patchell RA. Neurologic complications of bone marrow transplantation. Neurol Clin 1988; 6: 377–387. 2 Patchell RA, White CL, Clark AW et al. Neurological complications of bone marrow transplantation. Neurology 1985; 35: 300–306. 3 Patchell RA. Neurological complications of organ transplantation. Ann Neurol 1994; 36: 688–703. 4 Graus F, Saiz A, Sierra J et al. Neurologic complications of autologous and allogeneic bone marrow transplantation in patient with leukemia. Neurology 1996; 46: 1004–1009. 5 Antonini G, Ceschin V, Morino S et al. Early neurologic complications following allogeneic bone marrow transplant for leukemia. Neurology 1998; 50: 1441–1445. 6 Winston DJ, Gale RP, Meyer DV, Lowell SY. Infectious complications of human bone marrow transplantation. Medicine (Baltimore) 1979; 58: 1–31. 7 Elfenbein GJ, Saral R. Infectious disease during immune recovery after bone marrow transplantation. In: Elfenbein GJ, Saral R (eds). Infection and the Compromised Host. Williams and Wilkins: Baltimore, 1981, pp 157–196. 8 Hagensee ME, Bauwens JE, Kjos B, Bowden RA. Brain abscess following marrow transplantation: experience at the Fred Hutchinson Cancer Research Center, 1984–1992. Clin Infect Dis 1994; 19: 402–408. 9 Thomas ED, Storb R, Clift RA et al. Bone marrow transplantation. New Engl J Med 1975; 292: 832–843, 895–902. 10 Mohrmann RL, Mah V, Vinters HV. Neuropathologic findings after bone marrow transplantation: an autopsy study. Hum Pathol 1990; 21: 630–639. 11 Bowden RA, Sayers M, Flournoy N et al. Cytomegalovirus immune globulin and seronegative blood products to prevent primary cytomegalovirus infection after marrow transplantation. New Engl J Med 1986; 314: 1006–1010. 12 Meyers JD, Reed EC, Shepp DH et al. Acyclovir for prevention of cytomegalovirus infection and disease after allogeneic marrow transplantation. New Engl J Med 1988; 318: 70–75. 13 Goodrich JM, Mori M, Gleaves CA et al. Early treatment with ganciclovir to prevent cytomegalovirus infection after allogeneic bone marrow transplantation. New Engl J Med 1991; 325: 1601–1607. 14 Navia BA, Petito CK, Gold JW et al. Cerebral toxoplasmosis complicating the acquired immune deficiency syndrome: clinical and neuropathological findings in 27 patients. Ann Neurol 1986; 19: 224–238. 15 Holliman RE. Toxoplasma and the acquired immune deficiency syndrome. J Infect 1988; 16: 121–128. 16 Harrison MJG, McArthur JC. Opportunistic infections – parasites. In: Harrison MJG, McArthur JC (eds). AIDS and Neurology. Curchill Livingstone: New York, 1995, pp 171–182. 17 Carey RM, Kimball AC, Armstrong D et al. Toxoplasmosis. Clinical experiences in a cancer hospital. Am J Med 1973; 54: 30–38. 18 Remington JS. Toxoplasmosis in the adult. Bull NY Acad Med 1978; 50: 211–227. 19 Vietzke WM, Geldermann AH, Grimley PM et al. Toxoplasmosis complicating malignancy. Cancer 1968; 21: 816–827. 20 Slavin MA, Meyers JD, Remington JS et al. Toxoplasma gondii infection in marrow transplant recipients: a 20 year experience. Bone Marrow Transplant 1994; 13: 549–557.

1175


1176

21 Cancelas JA, Lopez-Jimenez J, Gonzalez C et al. Cerebral toxoplasmosis with atypical presentation in a bone marrow transplant patient. Med Clin Barc 1994; 103: 383–384. 22 Seong DC, Przepiorka D, Bruner JM et al. Leptomeningeal toxoplasmosis after allogeneic marrow transplantation. Case report and review of the literature. Am J Clin Oncol 1993; 16: 105–108. 23 Schwechheimer K, Hashemian A. Neuropathologic findings after organ transplantation. An autopsy study. Gen Diagn Pathol 1995; 141: 35–39. 24 Dannemann B, McCutchan JA, Israelski D et al. Treatment of toxoplasmic encephalitis in patients with AIDS: a randomized trial comparing pyrimethamine plus clindamycin to pyrimethamine plus sulfadiazine. Ann Intern Med 1992; 116: 33–43. 25 Remington JS, Desmonts G. Toxoplasmosis. In: Remington JS, Klein JO (eds). Infectious Diseases of the Fetus and Newborn Infants. WB Saunders Company: Philadelphia, 1990, pp 89–195. 26 Levy RM, Rosenbloom S, Perrett LV. Neuroradiologic findings in AIDS: a review of 200 cases. AJNR 1986; 7: 833–839. 27 Post MJD, Kursunoglu SJ, Hensley GT et al. Cranial CT in acquired immunodeficiency syndrome: spectrum of diseases and optimal contrast enhancement technique. Am J Neuroradiol 1985; 6: 743–754. 28 Post MJD, Sheldon JJ, Hensley GT et al. Central nervous system disease in acquired immunodeficiency syndrome: prospective correlation using CT, MR and pathologic studies. Radiology 1986; 158: 141–148. 29 Whiteman MLH, Bowen BC, Post MJD et al. Intracranial infection. In: Scott W (ed). Magnetic Resonance Imaging of the Brain and Spine. Lippincott-Raven: Philadelphia, 1996, pp 708–772. 30 Post MJ, Hensley GT, Moskowitz LB et al. Cytomegalic inclusion virus encephalitis in patients with AIDS: CT, clinical, and pathologic correlation. Am J Roentgenol 1986; 146: 1229–1234.

31 Kalayjian RC, Cohen ML, Bonomo RA et al. Cytomegalovirus ventriculoencephalitis in AIDS. Medicine 1993; 72: 67–77. 32 Jurges E, Young Y, Eltumi M et al. Transmission of toxoplasmosis by bone marrow transplant associated with Campath-1G. Bone Marrow Transplant 1992; 9: 65–66. 33 Yuh WT, Nguyen HD, Gao F et al. Brain parenchymal infection in bone marrow transplantation patients: CT and MR findings. Am J Roentgenol 1994; 162: 425–430. 34 Beelen DW, Mahmoud HK, Mlynek ML et al. Toxoplasmosis after bone marrow transplantation. Haematol Blood Transfus 1987; 30: 574–578. 35 Ackermann ZVI, Or R, Maayan S. Cerebral toxoplasmosis complicating bone marrow transplantation. Israel J Med Sci 1986; 22: 582–586. 36 Hirsch R, Burke BA, Kersey JH. Toxoplasmosis in bone marrow transplant recipients. J Pediatr 1984; 105: 426–428. 37 Soulier-Lauper M, Zulian G, Pizzolato G et al. Disseminated toxoplasmosis in a severely immunodeficient patient: demonstration of cysts in bone marrow smears. Am J Hematol 1991; 38: 324–326. 38 Miaux Y, Ribaud P, Williams M et al. MR of cerebral aspergillosis in patients who have had bone marrow transplantation. Am J Neuroradiol 1995; 16: 555–562. 39 Beal MF, O’Carroll CP, Kleinman GM et al. Aspergillosis of the nervous system. Neurology 1982; 32: 473–479. 40 Walsh TJ, Hier DB, Caplan LR. Aspergillosis of the central nervous system: clinicopathological analysis of 17 patients. Ann Neurol 1985; 18: 574–582. 41 Cohen J. Clinical manifestations and management of aspergillosis in the compromised patient. In: Warnock DW, Richardson MD (eds). Fungal Infection in the Compromised Patient. John Wiley and Sons: Chichester, 1991, pp 117–151. 42 Johnson JD, Butcher PD, Savva D et al. Application of the polymerase chain reaction to the diagnosis of human toxoplasmosis. J Infect 1993; 26: 147–158.