We report 78 cases of toxoplasmosis diagnosed between 1987 and 1992, f
rom an autopsy study of 205 patients infected by the human immunodefic
iency virus (HIV). Of the 78 patients 22 were females (28 %) and 56 ma
les (72 %). Risk factors were as follows: intravenous drug addiction (
44 cases, 56 %), homosexuality or bisexuality (18 cases; 36 %) and mul
tiple blood transfusions (6 cases). Cerebral toxoplasmosis (CT) was di
agnosed in 73 cases (93 %) and was characterized by abcesses (59 cases
), diffuse encephalitic lesions (8 cases), isolated cysts without infl
ammation (3 cases) and hemispheric involvement with ventricular hemorr
hage (3 cases). Cerebral involvement were isolated (55178 cases; 70 %)
or associated with multivisceral diffusion (181 78 cases; 23 %). Isol
ated extracerebral localization was observed in 5 cases. The most freq
uent extracerebral sites were: cardiac (21 cases), pulmonary (14 cases
) and pancreatic (7 cases). Immunohistochemical study with anti Toxopl
asma gondii antibodies allowed to a diagnosis of extracerebral localiz
ation in 8 cases. Ultrastructural features of Toxoplasma gondii were s
tudied on post mortem myocardial samples (2 cases) open lung biopsy (1
case) and bladder biopsy specimen (1 case). Antemortem diagnosis of c
erebral toxoplasmosis was performed on CTscan in 591 73 (80 %). Antemo
rtem diagnosis of extracerebral toxoplasmosis was performed or suspect
ed in 8123 cases (34 %): by isolation of trophozoites in bronchoalveol
ar lavage (2 cases), on an open lung biopsy (1 case) and on a bladder
biopsy specimen (1 case), and by clinical and echocardiographic data (
4 cases). Antitoxoplasmic serology allowed to the diagnosis of toxopla
smosis in 12/78 cases (15 %) by showing high levels of IgG in the seru
m. The incidence of extracerebral toxoplasmosis is particularly import
ant in our autopsy study (11 %). The high prevalence of toxoplasmosis
in France, the increase of the mean duration of the disease from diagn
osis of AIDS to death since the introduction of azidothymidine and the
high number of post mortem specimen performed, might explain the grea
t frequency of extracerebral involvement by Toxoplasma gondii observed
in this study.