Although Brazil is one of the countries with the largest number of AIDS cases (about 600,000 patients) out of Africa, a slightly smaller number than the one estimated for the entire Central and Western Europe [34
], this is the first Brazilian study on AIDS autopsy (and also in developing countries) to compare the neuropathological lesions pre-and post-HAART era and to demonstrate that these lesions are still frequent and are often not diagnosed clinically (premortem
). Necropsy is an important tool to understand the lesions in AIDS and several other diseases. However, there has been an increasing reduction in the number of AIDS autopsy, both in our material as in several other studies [19
], probably because the complementary diagnostic and clinical experience have contributed to better management of patients and in obtaining premortem
The percentage of women in our cases has increased over the course of years (from 17.8% up to 1996 to 33% after 1996), following the world trend as also observed by Jellinger et al. [2
], whose frequency rose from 7.9% to 29.6%, explaining the differences regarding older studies, such as the Brazilian pre-HAART [9
]. The mean age of AIDS patients in this study (35.7 years), which was lower than in developed countries ranging from 38.4 to 40.5 years [2
], probably because survival rate is higher in these countries.
As for the overall frequency of neuropathological findings, we found that in 58.1% of our 284 cases there were significant morphological lesion; a higher value than that found in the Brazilian clinical study De Oliveira et al. [35
], which was 46.5%. This difference probably results from greater easiness to diagnose lesions at the anatomopathological study. However, our percentage was lower than the values obtained by Masliah et al. [3
], Morgello et al. [20
], Walsh et al. [23
], Vago et al. [22
], and Neuenburg et al. [21
], which ranged from 63% to 85%. However, if we compute in our study all cases where there was some kind of CNS lesion (relevant or focal lesions and just edema), as some authors, our percentage would be 94.4%, similar to that found by Chimelli et al. [9
] in the pre-HAART era, which was 91.3%.
shows a comparative summary of various series of autopsies studies pre- and post-HAART, regarding to neuropathological lesions.
Comparison between different studies regarding lesions to the central nervous system, in a series of autopsies of patients with HIV infection.
Assessing the frequency of all neurological lesions relevant in each group studied, a slightly higher percentage was found in the group without antiretroviral treatment (63.8%) than in the other groups (48.7% and 53.3%), although these differences were not statistically significant. This result could give a false impression that HAART did not alter the prognosis in AIDS; however, the survival time of patients clearly shows benefit of this therapy, although in the final stage, the cause of death is often by neuropathological lesions. Our percentage of lesions in groups was smaller than those of Jellinger et al. [2
], which decreased from 80% to 60%. It is possible this higher percentage elapses from the larger CNS sample assessed by these authors, or from not relevant lesions that might have been also computed. Unfortunately, in our study the number of cases of group C was small, perhaps indicating low adherence to HAART in our midst, or possibly because using these drugs makes lower mortality rates, resulting in fewer autopsies, as occurred in several post-HAART studies [2
]. On the other hand, when we began this study we thought that since 1997 almost all patients would be under regular use of HAART protocol, mainly because these medications are free in Brazil; however, only 23.2% of 194 cases of this period used HAART for 3 months or more and 44.8% of these individuals did not use any antiretroviral medication. These data are difficult to compare because many studies separated groups only for period, without reviewing the clinical data regarding to antiretroviral therapy (). Vago et al. [22
] reported to have collected data related to treatment only in the 4th period: 90.6% used HAART, but the length of treatment was not detailed. If we consider all patients who used HAART for any time, they constitute 38.7% (75 of 194) of the autopsies since 1997; however, considering the minimum duration of 3 months for the effectiveness of therapy, these cases represent only 23.2% (45 of 194) of this period.
Regarding the type of lesion found, infections prevailed in all our study groups, diagnosed in 54.2% of 284 cases, much lower than that of Neuenburg et al. [21
] and Vago et al. [22
], which were 85% and 76%, respectively; difference found perhaps due to more detailed sampling of the CNS, the routine use of immunohistochemistry for various viruses or regional differences. Considering the subtypes of infection, toxoplasmosis was the most frequent in our study as a whole (29.9%). Unlike Masliah et al. [3
], Neuenburg et al. [21
] and Walsh et al. [23
], whose global percentage of toxoplasmosis ranged from 2.8% to 18.7% and where HIV infections prevailed, it also occurred in pre-HAART studies from Anders et al. [5
], Petito et al. [17
], and Budka et al. [15
], whose toxoplasmosis frequencies ranged from 6.7% to 17% and cytomegalovirus and HIV encephalitis predominated. These differences are likely to be related to environmental, geographic and socioeconomic factors. The pre-HAART Brazilians autopsy studies [9
] corroborate these ideas, as toxoplasmosis was the most frequent infection ranging from 21% to 34.1%. This difference between studies in developed countries and ours also persisted in the isolated analysis of the group using HAART. In spite of the statistically significant reduction in frequency of toxoplasmosis among our untreated group, compared to the one in use of HAART (36.2% versus 17.8%), these values were higher than those found in studies from developed countries, in which there was reduction of 24% to 8% [2
], from 9.4% to 7.1% [21
], and 60% to 0% [23
]. The second most common infection in our study was cryptococcosis, present in 15.8% of cases and similar in the three subgroups of this study; it was more frequent in our material, even in the HAART group (15.6%) than in the corresponding groups from Neuenburg et al. [21
], Jellinger et al. [2
], and Masliah et al. [3
], whose percentage ranged from 0 to 5.6%. Those differences could be explained, at least in part, because many patients in our country seek medical help when already presenting with advanced neurological disease, affecting the diagnosis and treatment. Viral infections, which together represent more than 50% of lesions in the CNS studies in developed countries, even in the pre-HAART period [5
], were not common in our series, accounting for only 6.7% of 284 cases, while infections by protozoa, fungi and bacteria amounted to 49.3% in our material. HIV encephalitis occurred in 2.8% of total cases in our study, cytomegalovirus in 2.5% and JC virus infection in only 0.3%. This lower percentage of viral infections in relation to other series of autopsies is probably justified due to the lower survival rates of our patients, with not enough time available for the development of HIV neuropathological lesions, complicating prior by opportunistic infections, as our patients usually seek medical treatment in advanced stages of AIDS. As to the comparison of viral infections between the groups, cytomegalovirus infection was more frequent in our group without antiretroviral drugs (3.1%), while in the HAART group there was predominance of HIV encephalitis (11.1%). This change is probably related to the longer survival of patients on HAART, as it has already been observed in developed countries for several years. However, the percentage of lesion by HIV was much lower, even in our HAART group, than in other series post-HAART [2
], in which it varied between 15% and 59.5%. In some of these studies there was a slight reduction in HIV CNS infection over the periods studied [19
], unlike others who showed an increase [3
], but as mild or moderate forms. Probably the lowest frequency of HIV lesion in our HAART group, is due in part to the fact that immunohistochemistry was not performed in our study, and HIV encephalitis diagnosis was made by the direct visualization of multinucleated giant cells, unlike several of these studies that performed routine immunohistochemical as Neuenburg et al. [21
], whose percentage of lesion by HIV in HAART group was 59.5%. In addition to possible socioeconomic and geographic factors, in our material there was still a high frequency of toxoplasmosis and cryptococcosis, which may have slightly reduced the survival of this group of patients. Rare infections in the CNS as paracoccidioidomycosis, sporotrichosis, infection by free-living amoeba (Balamuthia mandrilares
), were not reported in other studies and may be related to environmental and regional factors.
Noninfectious lesions occurred in 3.5% of our cases, also no significant difference between groups (3.7%, 2.6%, and 4.4%); 1.4% represented by primary CNS lymphomas and none of these cases was diagnosed premortem
. In the series of Jellinger et al. [2
] the frequency of this neoplasm was much higher in all time periods, ranging between 6% and 11.2%, as in others post-HAART studies, which ranged from 2.4% to 18.2% [19
]. Probably because survival is higher in developed countries and frequency of opportunistic infections, lower. Important vascular lesions occurred in 4.4% of our HAART group, having been described in 10% in a corresponding group from Jellinger et al. [2
] and 19.1% in the pre-HAART study from Anders et al. [5
]. They were infrequent in our study, occurred only in groups B and C, and could not be directly related to AIDS. Such large differences between the various series can be explained, at least in part, by the way we reported those lesions, as in our percentage we considered only the relevant lesions such as extensive bleeding, while Anders et al. [5
] reported “bleeding of variable intensity.” Given these data we can conclude that neuropathological lesions, even with the use of HAART, are still common in patients with AIDS, which is consistent with the observations reported by other authors [2
Regarding the analysis of survival time after HIV infection diagnosis and use or not of HAART—which our study clearly demonstrated the benefit of this medication—no detailed assessment was found in the other series of autopsies, but only comments on the study from Jellinger et al. [2
] and Neuenburg et al. [21
], who stated that survival was increased after HAART. Our data confirm the concepts already well known clinically and are also in agreement with the reduction in the number of autopsies performed on AIDS after the introduction of HAART, although other factors may be involved in this issue, such as the possibility of an earlier-achieved diagnosis of HIV infection.
For the correlation between clinical and autopsy findings, no similar study was found in autopsies of AIDS in relation to the CNS. Wilkes et al. [14
] made a review of 101 autopsies of AIDS in pre-HAART era, comparing pre
- and postmortem
diagnoses in various organs; they reported that 74% of cases had lesions not suspected clinically. Another more recent study compared pre
- and postmortem
diagnosis in AIDS, although very detailed and classifying errors, assessed only invasive mycoses in different organs. In 36.9% of cases in which this infection was the major disease, this diagnosis was not made clinically [31
]. In our series disagreement was observed in 91 (55.1%) of 165 cases with significant lesions in the CNS (or 32% of the total 284 cases). Disagreement type 1 (important disease that if diagnosed in life could change the patient prognosis) occurred in 49 (53.8%) of 91 discordant cases, representing 17.6% of the total 284 cases studied. These data demonstrate that even serious lesions are still not suspected clinically, probably because of clinical similarities and very difficult differential diagnosis, which reinforces the importance of autopsy studies to improve understanding and knowledge of brain lesions, which are common and still prevalent, even with the use HAART and the presence of advanced diagnostic technologies.