Autopsy Findings in the Acquired Immunodeficiency Syndrome (AIDS). A Report of 395 Cases from the South of France

PATHOLOGY RESEARCH AND PRAGICE

© Urban & Fischer Verlag http://www.urba nfischer.deJjournalslprp

Autopsy Findings in the Acquired Immunodeficiency Syndrome (AIDS). A Report of 395 Cases from the South of France Paul Hofman, Marie Christine Saint-Paul, veronique Battaglione, Jean-Francois Michiels and Robert l.oubiere Department of Pathology, H6pital Louis Pasteur, University of Nice, Nice, France

Summary

Introduction

Necropsy findings in 395 adult patients with the acquired immunodeficiency syndrome (AIDS) who died in Nice, France , between March 1983 and May 1996 were compared retrospectively with antemortem diagnoses, risk factors and number of positive T CD4 lymphocytes at the time of death. Special emphasis on bacterial infections was made in this study. Lesions observed from 1983 through 1989 and from 1990 through 1996 were compared. We assessed the role of organ lesions in the immediate cause of death. The organ system distribution of major opportunistic infections and neoplasms was similar throughout the years of the study. The most common diagnostic disease entities in all organ sites were cytomegalovirus infection, toxoplasmosis and candidia sis. Toxoplasmosis was more common in the intravenous drug abuser group. Bacterial infections were frequent and contributed to the mortality and morbidity of all risk factor groups . Kaposi' sarcoma continued to occur more frequently in the homosexual population. Cytomegalovirus infection remained one of the most common causes of death from 1983 to 1996. Mortality from fungal and bacterial infections, and mycobacteriosis increased in frequenc y during the course of this study whereas deaths from pneumocystosis declined. The death rate from malignant lymphoma and carcinom a increased after 1989. The clinical cause of death concurred with the pathological cause in 55% of the cases. Lung was the most frequent organ involved followed by the central nervous system, the gastrointestinal tract and the heart.

Despite advanced technolog y in diagnostic medicine over the years, autopsy remains an established tool for providing additional information about disease s and their effects. Given this role, autopsy has played an important role in increasing our knowledge about the human immunodeficiency virus [22 , 36, 38]. This retrospective study was performed to correlate the diagnoses made antemortem to those made postmortem, particularly for the opportunistic infections, malignant lymphomas and Kaposi' sarcoma, in an AIDS population of the South of France, and to determine the causes of death. Lesions observed from 1983 through 1989 and from 1990 through 1996 were compared. The present study was performed from a geographic area where intravenou s drug abuse is the most common risk factor related to AIDS. This series contained a high percentage of women in comparison with previously reported necropsy series [1 , 10, 12, 15, 16, 19, 36, 38]. The incidence of specific infections and malignancies correlate s with the risk factors for HIV infection and the number of CD4+ lymphocytes. We determined whether there have been evolutionary changes in the causes of death, survival and organ system distribution of opportuni stic infections and neoplasms seen with AIDS at autopsy. Finally the toxicity of some therapeutics was evaluated using microscopic examination.

Key words: HIV - AIDS - Autopsy - Opportunistic infection - Neoplasia - Carcinoma - Kaposi' sarcoma Pathol. Res. Pract. 195: 209-217 (1999)

Address for correspondence: Dr. Paul Hofman, Department of Pathology, Hopital Pasteur, 30 avenue de la voie romaine, F - 06002 Nice cedex I, France. Tel.: ++33-492/03 77 65; Fax: ++33-492/038210; E-mail: [email protected] 0344-0338/99/195/4-209 $12 .00/0

210 . P. Hofman et al.

Patients and Methods

Results

Consecutive systematic necropsies of 395 patients who died of AIDS were performed between March 1983 and May 1996 at the Nice University Medical Center. Adults who met standard criteria for the Centers for Disease Control and Prevention (CDC), Atlanta, GA, for diagnosis of AIDS were included in this study [8, 9]. Pediatric AIDS cases were excluded. One hundred and ninety four cases from 1983 through 1989 were compared with 201 cases from 1990 through 1996, separated by the institution of antiretrovira1 therapy and Pneumocystis carinii (PCP) prophylaxis in the latter period. Risk factors for HIV infection were enumerated. The main medical treatments administered were noted: azidothymidine (AZT) and/or dideoxynucleosides didanosine (ddl) and zalcitabine (ddC), antitoxop1asmosisand antipneumocystosis drugs. The duration of the disease after development of full blown AIDS was determined, as well as the number of positive T CD4lymphocytes (CD4+) at the time of death. The immediate cause of death was defined by comparing autopsy and clinical data. All autopsies were complete examination using the standard Rokitansky technique with removal and examination of all visceral organs and the central nervous system. A total of 58 spinal cord specimens were studied. All organs were sampled systematically; gross lesions were also sampled. Hematein eosin safran (HES), gram, giemsa, periodic acid Schiff (PAS), Ziehl-Nielsen and Gomori Grocott stains were routinely used for microorganism detection. Peroxydase immunostaining with the avidine-biotin complex (Vectastain kit) and Toxoplasma gondii (ICN Biomedicals Costa Mesa, USA), Cytomegalovirus and Pneumocystic carinii (Biosoft, San Diego, CA) antibodies were performed when necrotic lesions were present. Toxoplasma encephalitis, cytomegalic colitis and pneumocystis pneumonia served as positive control. Statistical analyses were performed using k2 tests with the significance of correlations evaluated using t tests [28].

Patient characteristics There were 268 males (68%) and 127 females (32%) in which AIDS was present at the time of death. All patients were caucasian. There were 194 cases from 1983 through 1989 and 201 cases from 1990 through 1996. Risk factors for HIV infection included intravenous drug use (257 patients; 65%), homosexuality (80 patients; 20%), heterosexual contact (39 patients; 10%) and blood transfusion (19 patients; 5%). Risk factors before and after 1989 are shown in Fig. 1. The mean age at death was 37 years (range 19-82). The mean number of CD4+ lymphocytes at the time of death was 37/mm3 for the period from 1983 to 1989 and 84/mm3 from the period from 1990 to 1996 (p < 0.05). The duration of the disease after development of full blown AIDS for HIV was 22 months for the period from 1983 to 1989 and 49 months for the period from 1990 to 1996 (p < 0.05). Infections The main pathogens observed at the postmortem study from 1983 to 1989 and from 1990 to 1996 are summarized in Fig. 2 and correlations between the diagnosis made antemortem to those made postmortem are shown in Fig. 3. The different locations of the main pathogens seen at necropsy are summarized in Table 1.

Opportunistic infections: One hundred and fifty one (38%) of the 395 subjects had demonstrable CMV inclusions. 48/151 (31%) cases had single organ involvement (29 of the adrenal gland,

A

B

(194 patients)

(201 patients)

7%

i::::;; ••~

3%

5%

I ••••••••

II • • • • • • • ,

!:i:ii:i~

liiiiir

•••• w .... ' I•••••·

47%

.... ' --- - ----!:::~ I... I'"

!~

Fig. 1. Risk factors noted from 1983 to 1989 (A) and from 1990 to 1996 (B).

f±II

Blood transfusion

D

Heterosexuality

13

Homosexuality

•

Intravenous Drug Abuse

Autopsy in AIDS . 211

Number of cases

•

1983-1989

•

Antemortem

o

1990-1996

o

Post Mortem Diagnosis

Number of cases

150

Diagnosis

250

130 20 0

110

90

150

70 100

50 30

50 10

> ::IE

U

c,

VI

VI

......

u

VI

o

'iii

e,

0

E VI

.!! a. o >C o

u

... s ..

. VI

VI

0

u >. ::IE

Pathogens

a

>

VI

u

0

::IE

u

m

VI

E

VI

"iii 0 ;:

..

c,

VI

.!!

J:I 0

u

a.

u

....

"-

u

'"

0

>.

J:I 0

>C

....

::IE

. ~E .... VI

VI

..... m':

..

Vl

E

>o!

0

s:

.. u

:0 :0

a.

o ..

. c

E

>.

..J

U

u >. ::IE

0

Fig . 2. Main pathogens ob served at the postmortem study from 1983 to 1989 and from 1990 to 1996.

Fig. 3. Correlations between the diagnosis made antemortem and those made po stmortem.

18 of the lung, 1 of the parotid) (Fig. 4). Immunohi stochemical study was useful to confirm the diagnosis in 3 adrenals. Only 33% of the patients had antemortem diagnosis of CMV infection . In 8 patients with antemortem CMV infection, inclusions were not observed on the postmortem samples. One hundred and one (26%) cases of toxoplasmosis were observed. In 1993 we reported 78 of these cases with 73 cerebral and 23 extracerebral involvement [9]. Since, 23 additional cerebral toxoplasmosis have been diagnosed and among

these cases an association with 6 pulmonary toxoplasmosis and I cardiac toxoplasmosis . Forty six percent of the patients had antemortem diagnosis of toxoplasmosis. Pneumocystis carinii was present in 15% of the patients at the time of autopsy. While in 11 cases the P. carinii was not identified antemortem, 51 subjects with P. carinii pneumoniae (PCP) had been treated either with trimethoprim-sulfamethoxazole, pentamidine or both. In 12 cases, immunohistochemical study alone allowed the diagnosis showing a few clusters of organism

Table I . Different locations of the main pathogens seen at necropsy CMV CNS Lung Heart GJ. Liver Pancreas U.G. Bone marrow Skin Lymph nodes Spleen Adrenal Thyroid

33 99 4 35 28 8 17 3 4 3 128 15

Toxoplasmo sis 96 20 22 4 2 7 5 2 2 3 2

PCP

62

Mycobacteriosis

Candidiasis

Cryptosporidiosis

Aspergillosis

8 21 3 19 18 5 4 12

2 8 5 204

21

4 8 4

46 20 2 I

2

5

2

10

2 5 2 7 3 3 2 3 3 2

Leishmaniasis

Bacteria 8

5

112

I

8

7 41

4

4 5

8

8

15

8 8

6

8

4

CNS: Central nervous system; G.1.: Gastrointestinal Tract; U.G. : Urogenital Tract; PCP: Pneumocystis carinii pneumonia.

4

212 . P. Hofman et al.

in alveolae. In 81 additional patients pneumocystosis had been diagnosed and treated antemortem. In 9 of these cases foci of calcification were observed in the

Fig. 4. Numerous cytomegalic inclusions in the parotid (HES x 200). Fig. 5. Cryptococci involving the bone marrow (HES x 320). Fig. 6. Leishmaniasis of the adrenal gland (arrow) (HES x 320).

lung, corresponding to pneumocystosis sequellae. At the postmortem study, acid-fast bacilli were identified in 15% of the cases. Fifty one of these cases were cultured premortem. Twelve cases were positive for Mycobacterium avium-intracellulare (MAl). Acid-fast bacilli were identified at necropsy in only 8 cases. In 33 cases, MAl infection was disseminated, being identified premortem in bone marrow (24 cases), urine (16 cases), blood (16 cases), bronchial brushing and sputum (12 cases). In 5 cases MAl was found in the lymph node alone. Tuberculosis infection was disseminated in all cases and acid-fast bacilli were identified predominantly in lymph node (14 cases) and lung (9 cases). Candidiasis was the most common mycose observed in this series. Two-hundred and forty five patients had antemortem diagnosis of candidiasis. Disseminate infection was found in 12 cases. Twenty one cryptococcal infections observed at necropsy (Fig. 5) were diagnosed antemortem and treated with amphotericin B without success. Disseminated aspergillosis was noted in 11 cases, but antemortem diagnosis was performed in only 5 cases. One patient presented a disseminated histoplasmosis with lung, liver, GI tract and bone marrow locations. At autopsy, 16 patients had evidence of disseminated leishmaniasis infection (Fig. 6). Culture from all these patients yielded Leishmania infantum. Antemortem diagnosis was performed in all cases. Eighteen patients had mucocutaneous Herpes simplex or Herpes zoster infections. Six also had herpes pneumonitis. Oesophageal inclusions were seen in 12 cases. Cryptosporidia which were seen at autopsy in only 14 cases were cultured antemortem in 78 patients. One case of invasive isosporiasis was diagnosed with small bowel, lymph node, spleen and liver involvement. We have reported this case previously [15J. Thirty two patients (8%) had progressive multifocal leukoencephalopathy. In 9 of these, demyelinisation was associated with extensive tissue destruction and cyst formation. HIV-associated encephalopathy was noted in 93 patients (23%) and showed 2 characteristic morphological patterns: progressive diffuse leukoencephalopathy (42 cases) and multifocal giant encephalitis (21 cases). Bacterial infections: Bacterial infections, diagnosed antemortem, occurred during the course of AIDS in 208 (53%) of the 395 cases in our autopsy series. Bacterial infection was diagnosed postmortem in 199 cases. The full assessment of all the infectious complications of AIDS provided by the analysis of our autopsy cases showed that such infections are more common than parasitic, viral, fungal or mycobacterial infection. Among the 208 patients who had bacterial infections diagnosed antemortem or postmortem, 102 (49%) had 2 or more bacterial infections. Eighty-eight patients had bacteriemic infections, including Staphylococcus aureus bacteriema

Autopsy in AIDS . 213

(45 cases), Escherichia coli infections (14 cases), catheter-related coagulase-negative staphylococcal infections (12 cases), and bacteriemic infections each with Pseudomonas aeruginosa and Yersinia pseudotuberculosis. Fifty two percent of the patients had one or more lower respiratory tract infections, including 26% due to S. aureus, 9% due to Sreptococcus pneumoniae, 3% due to Hemophilus influenzae and 3% involving various gram-negative bacilli. Eighty-five patients had urinary tract infections including 47 due to enterococcus and 19 due to P. aeruginosa. S. aureus infections which were diagnosed postmortem in 62 of our autopsies of patients with AIDS and bacterial infection due to other species were identified in only 14 additional cases for a total of 74 with bacterial infections at the time of death. Bacterial infection at the time of death was diagnosed on the basis of a pure culture in 55 of the 76 cases with bacterial infections diagnosed postmortem. All 62 of the S. aureus infections diagnosed postmortem involved the lungs and in 12 cases there were also hematogenous abscesses in the kidneys (12 cases), liver (6 cases), spleen (6 cases), heart (3 cases), brain (3 cases), adrenal (2 cases), thyroid (1 case), hypodermis (1 case), prostate (1 case) and testicle (1 case). Thirtyone of the 62 staphylococcal infections diagnosed postmortem were diagnosed for the first time at autopsy. Malignancies

The different malignancies observed from 1983 to 1989 and from 1990 to 1996 at the postmortem study are summarized in Fig. 7. Thirty nine (10%) of 395 subjects (38 males and 1 females) had a diagnosis of skin Kaposi' sarcoma (KS) during the course oftheir illness. At the time of autopsy, 22 (71 %) of these had a disseminated disease while 17 only had skin or mucosal lesions. A lymphoma was observed postmortem in 58 Number of cases

cases (14%). Antemortem diagnosis was performed in 34 cases. Forty three of the patients developed a nonHodgkin lymphoma of the following varieties: 27 were primary large-cell lymphoma of the CNS, 10 were a disseminated Burkitt's type lymphoma, and 7 were disseminated large cell B lymphoma. Eleven patients had a Hodgkin disease. Eighteen patients had carcinoma of various varieties (10 malpighien bronchial carcinoma, 6 lung adenocarcinoma, 1 hepatocarcinoma associated with a pancreas neurocarcinoma, 1 thyroid adenocarcinoma). Antemortem diagnosis was performed in all cases. Two patients presented a glioblastoma. Lymphocytic inflammation and drug toxicity-related lesions

Isolated lymphocytic infiltration without pathogen associated were seen in 71 patients (18%). These infiltration were mostly associated with parenchymal cell necrosis or ischemia. In these cases, no microorganism was observed after immunohistochemical study using anti-CMV and anti-Toxoplasma antibodies. In 14 cases, lymphocytic inflammation was disseminated and predominantly found in the heart and pericardium (13 cases), adrenal glands (8 cases), pancreas (8 cases) and brain (6 cases). Forty-one cases had single organ involvement (17 ofthe adrenal gland, 11 ofthe heart, 4 of the pancreas, 3 of the testicles, 3 of the prostate, 3 of the liver). In 22 cases, CMV (14 cases), toxoplasmosis (6 cases) or herpetic (2 cases) infections had been at distance in other organs. Iatrogen lesions were determined in 21 cases based on clinical information and lesions seen at necropsy. In 12 patients, diffuse necrotizing pancreatitis were observed after ddI (6 cases), intravenous pentamidine (4 cases) or ddC (2 cases) treatments. Lymphocytic myocarditis were noted in 11 patients treated by adriamycine (7 cases) and alpha interferon (4 cases).

50

•

1983-1989

o

Immediate cause of death

45

1990-1996

Clinical and pathological criteria were used to assess the role of organ involvement and tissue lesions in the terminal course of the patients. Eighty-five patients could not be assigned a cause of death based on the established clinical criteria. Respiratory insufficiency was the cause of death in 199 (50%) patients. Postmortem examination of the lungs of these 199 patients showed that CMV, bacteria, pneumocystosis and KS were the most frequent primary process. For 81 (20%) patients, the primary cause of death was neurologic disease. Thirty nine patients had necrotizing encephalitis due to T. gondii. Twenty-three patients died due to cerebral herniation. Ten patients died due to brainstem involvement. Four patients with severe neurologic impairment died due to aspiration pneumonia. In 5 patients CMV inclu-

40

3S

30

25

20 1S

10

o to

to

E o s:

E

E

~

Co

>.

..J

o

<:

..... ... til

Malignancies

s:

o

to U

Fig. 7. Different malignancies observed at the postmortem study from 1983 to 1989 and from 1990 to 1996.

214 . P. Hofman et al.

sions associated with inflammation were also found in areas of the brain with toxoplasma involvement. Twenty patients with CNS lymphoma became comatose and died without pulmonary associated disease. Four patients died due to gradual unexplained neurologic deterioration. At autopsy 3 had severe progressive multifocal leukoencephalopathy and 1 had a giant cell encephalitis. Fifty eight patients died because of primary cardiac dysfunction. Forty patients showed a myocarditis. Twelve patients with cardiac dysrythmias had ventricular dilatation diagnosed at autopsy. Fifteen patients died due to hypotension that was not caused by identifiable respiratory, neurologic or cardiac failure. Ten patients with severe diarrhea and extreme cachexia elected to have noninterventional care terminally. In these cases hypotension was probably caused by volume depletion due mainly to diarrhea. Seven patients had acute pancreatitis and one patient had massive gastrointestinal bleeding due to diffuse ulcerative KS of the bowel.

Discussion The initial fifteen years of the AIDS epidemic have witnessed impressive advances in techniques to diagnose and treat the underlying retroviral infection and the associated opportunistic infections and tumors. Clinical reports and autopsy studies have described many of the complications which occur in this patient population [1, 2, 3, 10, 11, 12, 15, 16, 19, 22, 26, 31, 34J. However, the clinical manifestations of AIDS have changed with widespread usage of pharmacologic therapies, particularly with prophylaxis for PCP and with antiretroviral therapy [25, 33J. It's obvious now that the incidence of PCP has declined since 1988 and prophylaxis for PCP appears to have shifted the clinical manifestations of AIDS to other illnesses. Then the diagnosis of AIDS appears to be occurring later in the natural history of HIV infection. The present study focused on 395 patients who had both well documented clinical courses and autopsy examinations. Our population was caucasian with a high incidence of intravenous drug abusers (65%) and of women (32%) reflecting the inner-city location of our Hospital and the ethnicity of our patients. Our findings show the increasing importance of intravenous drug abuse transmission of HIV in our area, whereas homosexual and heterosexual transmission decreased. In our study 39 of the 127 women reported heterosexual contact as their only risk factor. The lungs were the most common site of AIDS-diagnostic diseases, similar to previous studies [1, 2, 3, 11, 15, 29 J. However, the antemortem and postmortem diagnosis of PCP was limited to 36% in our group compared to 52-100% in some other studies [1, 3, 11, 15, 29, 31]. This result could be explained by employing invasive procedures as sputum induction or broncho-

scopy for the diagnosis PCP and combined pentamidin and sulfamethoxazole treatment administered to our patients. In our series the most commonly missed infection was CMV which was diagnosed antemortem in only 33% of the subjects with CMV inclusions observed postmortem. In our series, immunohistochemical studies for both CMV and PCP infections were helpful in a very few cases, respectively in 8 and 3 cases. Toxoplasmosis is the most common cause of intracerebral mass lesions in patients with AIDS and the most commonly identified opportunistic infection of the CNS [17, 38J. We have previously described this high incidence of both cerebral and extracerebral toxoplasmosis occurring in our patients [17, 18]. This high incidence could be partially explained by the high prevalence of toxoplasmosis in France. Only 46% of cerebral toxoplasmosis have been diagnosed premortem, probably because MRI has not been performed in all cases. Since 1990 we observed a decrease of toxoplasmosis among our population, maybe due to primary prophylaxis by sulfamethoxazole-trimethoprim. In our experience, MAl infection was documented in less than 16%. However, since 1990, we observed at the postmortem study an increase in the frequency of MAl and widely disseminated disease was frequently found even in treated patients. Some previous reports indicate disseminated Candida albicans is common and significant [19, 36J, whereas others found no cases at all or only found superficial infections in tracheal or gastric ulcers and only one case of lung involvement in a series of 36 patients [38J. The differences between studies may be due to the duration of the disease prior to death since, in comparison with studies where survival is mentioned, our patients appear to have a longer survival after 1990. In our series, aspergillosis diagnosis increased after 1989, maybe due to neutropenic therapy administrated to our patients. In recent years, numerous studies have described isolated cases of visceral leishmaniasis in HIV-infected patients living or travelling to Mediterranean countries. Most cases have been in Spain and also in the South of France because these countries are endemic areas for visceral leishmaniasis [25 J. Sixteen disseminated leishmaniasis were found among our HIV population. In one patient, visceral leishmaniasis occurred previous to HIV infection, suggesting that the disease in some HIV-seropositive patients may result from the reactivation of latent leishmanial infection in the presence of HIV-induced immunosuppression. However, the increasing occurrence of visceralleishmantiasis in our HIV-infected patients favour the hypothesis that this disease is a common opportunistic infection in HIV-infected patients in our city. HIV-associated brain pathology has been previously well documented [7, 13, 14, 21, 23, 24, 39]. Progressive multifocale leukoencephalopathy (PML) was encountered in 8% of the patients in this series, which is within the range of frequencies observed in

Autopsy in AIDS . 215

other studies [1, 3, 19, 22J. In 9 out of 21 patients, demyelinating lesions were associated with extensive tissue destruction and formation of cysts, containing numerous foamy macrophages. Similar severe histopathological changes of PML in AIDS were described in earlier reports [27, 31] and it was found that these were more pronounced than in PML due to other causes of immunosuppression such as lymphoproliferative disorders. HIV-associated encephalopathy was observed in 23% of the patients. This frequency is lower than in other series [5, 11, 22, 27J probably due to the fact that numerous patients were treated with AZT. Accordingly to some previous series [6J, the more frequent spinal cord disease in our study was vacuolar myelopathy occurring in 18% of the cases. Less frequently, necrotizing CMV (6 cases) and toxoplasmosis (5 cases) radiculomyelitis were noted. There was a high incidence (53%) of non opportunistic bacterial infections in our study. Bacterial pneumonia was seen in 65% and bacteremia in 38% of the autopsy cases. We found that a history of IVDA was associated with this high incidence. Because of induced neutropenia, bacterial infection was also higher in patients with Kaposi' sarcoma or lymphoma treated by chemotherapy. Bacterial infections were observed to increase in frequency after 1989. The causative organisms for these infections could be grouped in 2 major categories: S. aureus and E. coli. Interestingly, we observed deaths from bacterial sepsis involving organs other than the lung, such as the heart (3 cases) and the brain (3 cases). Previous studies have noted that bacterial infections are common in AIDS [10J. Bacterial pneumonia was less frequent than in our series [22J. Wilkes et al. found bacterial pneumonia in only 30% of their autopsy cases [40J. In our study deaths from KS remained essentially unchanged over the time. Malignant lymphomas were diagnosed in 14% of our cases. The increase in numbers of cases of malignant lymphoma has been previously predicted by a study in which AZT therapy was shown to increase the likelihood of subsequent appearance of this neoplasm in persons surviving with AIDS [35 J. We have diagnosed Hodgkin's disease (type II and III) in 11 patients. The association between Hodgkin's disease and AIDS is less well documented but this disease is thought to result from an Epstein-Barr virus infection [37]. We found a high frequency (18 %) of lymphocytic inflammation without association with pathogens or neoplastic diseases. The etiology of lymphocytic inflammation remains unknown among our population. A diffuse infiltrative lymphocytosis syndrome has been described in AIDS patients [20,21]. None of our patients had parotid gland swelling and sicca symptoms usually observed in this syndrome. However, in the absence of nerve or salivary biopsies performed in this series we

could not eliminate this syndrome for some of our patients. Several etiologic factors could also explain this lymphocytic inflammation: viral infection, as HIV itself, but also drug toxicity [4J. Including this latter category, we found 12 cases of pancreatitis from ddI (6 cases), pentamidine (4 cases) and ddC (2 cases) toxicity, and 11 cases of myocarditis from adriamycine and alpha interferon therapy. These findings indicate the potential for serious complications of pharmacologic therapy in AIDS. We observed increasing survival in persons with AIDS after 1989, when antiretroviral therapy and treatment, including prophylaxis for major AIDS-associated illnesses became more widespread. This was shown by a longer survival with AIDS from initial diagnosis (p < 0.05). This was due to a decreased frequency of PCP, toxoplasmosis and cryptococcosis. There is a greatest significant (p < 0.01) increase in deaths over the time of this study occurred from fungal infection as candidiasis and aspergillosis. Bacterial infections leading to pneumonia or septicemia were observed to increase. We also observed a slight increased in numbers of death from MAl and CMV dissemination. Deaths from KS remained unchanged but we noted an increase of malignant lymphoma (p < 0.05) and carcinoma (p < 0.05) after 1989. The distribution of risk factors and sex-ratio in this autopsy study was strongly different from that reported for the United-States [22J and in some other series from Europe [11]. Our series is original in including mainly drug abusers and a large proportion of females. This study confirms what has previously been reported, that there is a significant (p < 0.05) correlation between the appearance of KS and homosexuality [l1J. In contrast to some previous series where toxoplasmosis was found to be more frequent in homosexual population [5, 24J, we showed a higher correlation (p < 0.05) between toxoplasmosis infection and a risk factor for IVDA. In conclusion, evolutionary changes in the causes of death with AIDS have been a result of change in frequency of major opportunistic infections and neoplasms. The findings observed in this series have important consequences for prophylaxis and therapy to control the spread of AIDS particularly in an inner city population troubled by drug abuse. Acknowledgements. The authors wish to express their appreciation to M. Mari, S. Sadoulet, E. Selva and P. Vahala for their technical assistance.

References 1. Afessa B, Greaves W, Green W, Olopoenia L, Delapenha R, Saxinger C, Frederick W (1992) Autopsy findings in HIV-infected inner-city patients. J Acquir Immune Defic Syndr 5: 132-136

216 . P. Hofman et al. 2. Afessa B, Green W, Chiao J, Frederick W (1998) Pulmonary complications of HIV infection: autopsy findings.ChestI13: 1225-1229 3. Ambros RA, Lee Ey, Sharer LR, Khan MY, Robboy SJ (1987) The acquired immunodeficiency syndrome in intravenous drug abusers and patients with a sexual risk. Clinical and postmortem comparisons. Human Pathol18: 1109-1114 4. Barbaro G, Di Lorenzo G, Grisorio B, Barbarini G (1998) Cardiac involvement in the acquired immunodeficiency syndrome: a multicenter clinical-pathological study. Gruppo italiano per io studio cardiologicao dei pazienti affetti da AIDS investigators. AIDS Res Hum Retroviruses 14: 1071-1077 5. Bell JE, Donaldson YK, Lowrie S, McKenzie CA, Elton RA, Chiswick A, Brettle RP, Ironside JW, Simmonds P (1996) Influence of risk group and zidovudine therapy on the development of HIV encephalitis and cognitive impairment in AIDS patients. AIDS 10: 493-499 6. Budka H (1997) Neuropathology of myelitis, myelopathy, and spinal infections in AIDS. Neuroimaging Clin N Am 7: 639-650 7. Buttner A, Marquart KH, Mehraein P, Weis S (1997) Kaposi's sarcoma in the cerebellum of a patient with AIDS. Clin Neuropathol16: 185-189 8. Centers for Disease Control (1987) Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR Morb Mortal Wkly Rep 36: (suppl): 3S-15S 9. Centers for Disease Control (1992) Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly Rep 41: 1-19 10. Conception L, Markowitz GS, Borczuk AC, Factor SM (1996) Comparison of changing autopsy trends in the Bronx population with acquired immunodeficiency syndrome. Mod Pathol9: 1001-1006 11. D' Arminio Monforte A, Vago L, Lazzarin A, Boldorini R, Bini T, Guzzetti S, Antinori s, Moroni M, Costanzi G (1992) AIDS-defining diseases in 250 HIV-infected patients; a comparative study of clinical and autopsy diagnoses. AIDS 6: 1159-1164 12. D' Arminio Monforte A, Vago L, Gori A, Antinori S, Frenzetti F, Antonacci CM, Sala E, Catozzi L, Testa L, Esposito R, Nebuloni M, Moroni M (1996) Clinical diagnosis versus autopsy findings in 350 patients with AIDS. Eur J Clin Microbiol Infect Dis 15: 453-458 13. Davies J, Everall IP, Weich s, McLaughlin J, Scaravilli F, Lantos PL (1997) HIV-associated brain pathology in the United Kingdom: an epidemiological study. AIDS 11: 1145-1150 14. Davies J, Everall IP, Weich S, Glass J, Sharer LR, Cho ES, Bekk JE, Majteny C, Gray F, Scaravilli F, Lantos PL (1998) HIV-associated brain pathology: a comparative international study. Neuropathol Appl Neurobiol 24: 118-124 15. Falk S, Schmidts HL, Muller H, Berger K, Schneider M, Schlote W, Helm EB, Stille W, Hubner K, Stutte HJ (1987) Autopsy findings in AIDS. A histopathological analysis of fifty cases. Klin Wochenschr 65: 654-663 16. Guarda LA, Luna MA, Smith JL, Mansell PWA, Gyorkey F, Roca AN (1984) Acquired immune deficiency syn-

17.

18.

19.

20.

21. 22. 23. 24.

25. 26.

27.

28. 29.

30.

31. 32.

drome: postmortem findings. Am J Clin Pathol 81: 540-557 Hofman P, Bernard E, Michiels JF, Thyss A, Le Fichoux Y, Loubiere R (1993) Extracerebral toxoplasmosis in the acquired immunodeficiency syndrome (AIDS). Path Res PractI89:894-901 Hofman P, Michiels JF, Saint-Paul MC, GalibertA, Marty P, Durant J, Fuzibet JG, Mouroux J, Le Fichoux Y, Loubiere R (1993) Toxoplasmose au cours du SIDA: etude anatomoclinique de 78 cas. Ann Pathol13: 222-229 Hui AN, Koss MN, Meyer PR (1984) Necropsy findings in acquired immunodeficiency syndrome: a comparison of premortem diagnoses with postmortem findings. Human Pathol15: 670-676 Kazi S, Cohen PR, Williams F, Schempp R, Reveille JD (1996) The diffuse infiltrative lymphocytosis syndrome. Clinical and immunogenetic features in 35 patients. AIDS 10:385-391 Kibashi K, Mastri AR, Hirsch CS (1996) Neuropathology of human immunodeficiency virus infection at different disease stages. Hum Pathol27: 637-642 Klatt EC (1988) Diagnostic findings in patients with acquired immune deficiency syndrome (AIDS). J Acquir Immune Defic Syndr 1: 459-465 Lanjewar DN, Jain PP, Shetty CR (1998) Profile of central nervous system pathology in patients with AIDS: an autopsy study from India. AIDS 12: 309-313 Laing RB, Flegg PJ, Brettle RP, Leen CL, Bums SM (1996) Clinical features, outcome and survival from cerebral toxoplasmosis in Edingburgh AIDS patients. Int J STD AIDS 7: 258-264 Lemp GF, Payne JF, Neal D, Temelso T, Rutherford GW (1990) Survival trends for patients with AIDS. JAMA 263:402-406 Lyon R, Haque AK, Asmuth DM, Woods GL (1996) Changing patterns of infections in patients with AIDS: a study of 279 autopsies of prison inmates and nonincarcerated patients at a university hospital in eastern Texas, 1984-1993. Clin Infect Dis 23: 241-247 Martinez AJ, Sell M, Mitrovics T, Stoltenburg-Didinger G, Iglesias-Rozas JR, Giraldo-Velasquez MA, Gosztonyi G, Schneider V, Cervos-Navarro J (1995) The neuropathology and epidemiology of AIDS. A Berlin experience. A review of 200 cases. Pathol Res Pract 19: 427-443 Mattews DE, Farewell UT (1988) Using and understanding medical statistics. 2nd, pp 39-57. S Karger AG, New York, NY McKenzie R, Travis WD, Dolan SA, Pittaluga S, Feverstein 1M, Shelhamer J, Yarchoan R, Masur H (1991) The causes of death in patients with human immunodeficiency virus infection: a clinical and pathologic study with emphasis on the role of pulmonary diseases. Medicine 70:326-343 Michiels JF, Hofman P, Bernard E, Saint-Paul MC, Boissy C, Mondain V, Le Fichoux Y, Loubiere R (1994) Intestinal and extraintestinal Isospora belli infection in an AIDS patient. Path Res Pract 190: 1089-1093 Mobley K, Rotterdam HZ, Lerner CW, Tapper ML (1986) Autopsy findings in the acquired immune deficiency syndrome. Mod Pathol3: 34-41 Moulignier A, Authier FJ, Baudrimont M, Pialoux G, Belec L, Polivlka M, Clair B, Gray F, Mikol J, Gherardi

Autopsy in AIDS . 217

33. 34. 35.

36. 37.

RK (1997) Peripheral neuropathy in human immunodeficiency virus-infected patients with the diffuse infiltrative lymphocytosis syndrome. Ann Neurol4: 428-445 Moss AR, Bacchelti P (1989) Natural history of HIV infection. J Acquir Immune Defic Syndrome 3: 55-61 Niedt GW, Schinella RA (1985) Acquired immunodeficiency syndrome. Clinicopathologic study of 56 autopsies. Arch Pathol Lab Med 109: 727-734 Pluda JM, Yarchoan R, Jaffe ES (1990) Development of non Hodgkin lymphoma in a cohort of patients with severe human immunodeficiency virus infection on a long term antiretroviral therapy. Ann Int Med 113: 276-282 Reichert CM, O'Leary TJ, Levens DJ, Simrell CR, Macher AM (1983) Autopsy pathology in the acquired immune deficiency syndrome. Am J Pathol112: 357-382 Ree HJ, Strauchen JA, Khan PA (1991) Human immunodeficiency virus-associated Hodgkin's disease: clinico-

pathologic studies of 24 cases and preponderance of mixed cellularity type characterized by the occurrence of fibrohistiocytoid stromal cells. Cancer 67: 1614-1621 38. Welch K, Finkbeiner W, Alpers CE, Blumenfeld W, Davis RL, Smuckler EA, Beckstead JH (1984) Autopsy findings in the acquired immune deficiency syndrome. JAMA 252: 1152-1159 39. Wiley CA, Baldwin M, Achim CL (1996) Expression of HIV regulatory and structural mRNA in the central nervous system. AIDS 10: 843-847 40. Wilkes MS, Fortin AH, Felix JC, Godwin TA, Thompson WG (1988) Value of necropsy in acquired immunodeficiency syndrome. Lancet ii: 85-88 Received: July 29, 1998 Accepted in revised version: January 22, 1999