EPIDEMIOLOGY OF BRAIN LYMPHOMA AMONG PEOPLE WITH OR WITHOUT ACQUIRED-IMMUNODEFICIENCY-SYNDROME

Background: In recent years, brain lymphoma incidence has dramatically increased, presumably because of elevated risk of brain lymphoma amon g persons with acquired immunodeficiency syndrome (AIDS), Purpose: The objective of this study was to estimate independent incidence and sur vival rates of brain lymphoma among persons with or without AIDS and t o understand the epidemiologic features of this cancer, Methods: We li nked AIDS and cancer registry reports at nine state and local health d epartments and compared the demographics, histology, and survival of b rain lymphoma cases among persons with or without AIDS, The data were limited to people under 70 years of age, We calculated the incidence o f brain lymphoma among persons with AIDS and compared observed cases w ith those expected, The differences were statistically analyzed using the Poisson test, Epidemiologic features of brain lymphoma in persons with or without AIDS were compared using the chi-squared test, the Stu dent's t test, and the chi-squared test for linear trend, The logrank test was used to compare survival rates estimated by the Kaplan-Meier technique, All P values were two-sided, Results: We matched 50 989 AID S registry reports to 859 398 cancer registry reports (data from 1981 to 1990) and found 431 people with both AIDS and brain lymphoma, Among people with AIDS, those developing brain lymphoma versus those withou t brain lymphoma were more likely to be white (70% versus 59%; P < .00 1) and had homosexuality as their only human immunodeficiency virus ri sk factor (75% versus 64%; P < .001), Of the 431 patients, 223 develop ed brain lymphomas during 47 465 person-years of observation after dia gnosis of AIDS, The absolute incidence rate of brain lymphoma among pe rsons with AIDS was 4.7/1000 person-years (95% confidence interval = 4 .1-5.3/1000 person-years), 3600-fold higher than the base-line rate in the general population, From 1980 through 1989, overall counts of bra in lymphoma increased ninefold, Most of this increase was derived from persons with AIDS, but a substantial increase also occurred among per sons without AIDS (0.04/100 000 in 1982 to 0.28/100 000 in 1989) (chi- squared test for trend; P < .05). The median survival was shortest for persons with AIDS and brain lymphoma (2 months), was intermediate for persons with brain lymphoma without AIDS (5-7 months), and was longes t for persons with AIDS without brain lymphoma (14 months) (P < .05 fo r all comparisons), Conclusions: This analysis distinguishes the separ ate epidemiologies of brain lymphoma incidence among persons with or w ithout AIDS and shows brain lymphoma incidence among persons with AIDS to be several thousand-fold higher than that in the general populatio n, The study documents the overwhelming effect of AIDS-associated brai n lymphoma on the overall rate in the general population and demonstra tes a significantly rising trend, although of a lesser magnitude, amon g persons without AIDS, Implications: This study emphasizes a greater need to bring health care resources to this burgeoning epidemic.