Is there really a correlation between AIDS dementia and Kaposi's sarcoma?

Abstract

Recent results of epidemiological studies on the correlation of AIDS dementia (AD) and Kaposi's sarcoma (KS) have revealed a divergent picture. We have attempted to investigate this topic further using a population of 438 autopsied AIDS patients from Frankfurt. We evaluated 355 histopathological brain examinations with regard to signs of AD, and matched the clinical and pathological information concerning the history of KS. Liestoel et al. [1] and Dore et al. [2] published letters in 1997 and 1998, respectively, about the co-occurrence of KS and AD. The authors found evidence that patients with KS are protected from AD. It is not known, however, by what mechanisms. The influence of a blocking chemokine (blocking a CD4 cell coreceptor) produced by KS-associated herpes virus has been discussed by the authors. Baldeweg et al.[3] on the other hand, did not find such a relationship. Trying to confirm the hypotheses of Liestoel et al. [1] and Dore et al. [2] we evaluated the correlation of the neuropathological findings of AD with the absence or presence of KS. We evaluated 355 HIV-infected patients from the Frankfurt University Hospital who died of AIDS between 1985 and 1997 [4] AIDS was defined according to the Centers for Disease Control classification of 1993 [5] All patients were closely followed-up in the hospital and their histories were continuously documented. Eighty-three further patients with incomplete immunological data were excluded from the study. We used neuropathological findings to determine the extent of dementia. Referring to the Consensus Report of 1991 [6] we divided our patients into four groups: (i) no AD; (ii) mild AD; (iii) moderate AD; and (iv) severe AD. Liestoel et al. [1] correlated the incidence of KS with clinical signs for AD using the clinical staging of Price and Brew [7]. We have not yet had complete neurological data in order to stage our patients likewise in this retrospective study. Like Liestoel et al. [1] Dore et al. [2] and Baldeweg et al. [3] we divided our group with reference to KS being `absent' or `present'. Contrary to the findings of Liestoel et al. [1] and Dore et al. [2] we did not find any relationship between the incidence of AD and KS (chi-square test, P 0.653). We also did not find a correlation between: (i) no AD and KS as well as between (ii) mild AD and KS; (iii) moderate AD and KS; (iv) severe AD, and AD altogether with KS. According to our data, the risk of developing AD is the same with or without KS (see Table 1).Table 1: Percentage (absolute numbers) of AIDS dementia and Kaposi's sarcoma in four cohorts from different countries. We looked at the structure of the patient groups from the four centres in order to understand the different results. Contrary to our subcohort of 355 autopsied patients, most of the 2438 patients of the cohort of Dore et al. [2] were presumably alive at the time of evaluation. Baldeweg et al. [3] gathered their data during the follow-up of a retrospective study of a cohort of patients with AIDS. These patients were attending a specialized HIV medicine centre, but it is unknown whether they were dead or alive at the time of the study. The group of Liestoel et al. [1] consisted of 229 people who died from AIDS, but their results are not comparable to ours. In Norway, the percentage of patients with AD in the cohort of Liestoel et al. [1] was greater than in all other centres (52.8% compared with 38.0, 8.1 and 7.9%). At the same time, the number of patients with KS is much smaller than those of other centres (9.6% compared with 43.7, 36.2, 27.8%). The percentage of patients with KS in our group was 43.7, the percentage of patients with KS in the group of Liestoel et al. [1] was 9.6. In the group of Dore et al. [2] 27.8% of patients developed KS and only 8.1% had AD. It is not clear whether this is related to different patient population characteristics with regard to age, sex, source of infection, nationality, number of clinical follow-ups, CD4 cell count, treatment, status (dead or alive) or other. This might be an explanation for the different results. The different definitions of AD might also play a role. In order to make reliable statements concerning the protective effects of KS, we need complete neurological, neuropathological and clinical data. Jutta K. Neuenburga Hans R. Brodtb Brian G. Herndierc Wolfgang Schloted