Background: Although the use of HAART has led to a dramatic decline in the incidence of clinical AIDS-defining events, it is unclear whether rates of decline are similar for events of different aetiologies.

Methods: The ART Cohort Collaboration includes data on 12,574 treatment-naïve patients (pts) starting HAART from 13 cohorts in Europe and North America (Lancet 2002 360: 119–129). Clinical events and person-years follow-up (PY) in the first 3 years after starting HAART were stratified by aetiology (viral, fungal, bacterial, and other). Poisson regression was used to assess changes in incidence.

Results: The pts (79% male, 41% homosexual, median 36 years at HAART) had median CD4 count 250 cells/mm3 and HIV RNA 4.9 log copies/ml; 21% had a prior AIDS diagnosis. In 22,958 PY, 944 AIDS-defining events occurred (incidence rate [IR] 41.1/1000 PY). The IR of clinical events declined from 130.5 (95% CI 117.9–143.2) in the first 3 months to 13.8 (9.9–17.7) in the third year after starting HAART. The decline was greatest (97.9%) for events of viral aetiology and lowest (79.5%) for events of fungal aetiology (see table) and remained highly significant (p £ 0.001 in each case) after adjusting for demographics and type of initial HAART.

IR/1000 PY (95% CI), by aetiology

Months since starting HAART	Viral	Fungal	Bacterial	Other
0-3	43.3 (36.0-50.6)	23.9 (18.5-29.3)	33.1 (26.8-39.5)	30.3 (24.2-36.3)
4-6	12.1 (8.1-16.0)	9.6 (6.1-13.2)	12.4 (8.3-16.4)	14.1 (9.8-18.4)
7-12	6.9 (4.7-9.0)	5.8 (3.8-7.8)	6.7 (4.5-8.8)	11.2 (8.4-14.0)
13-24	3.4 (2.1-4.7)	5.3 (3.7-6.9)	5.6 (3.9-7.2)	8.0 (6.0-10.0)
25-36	0.9 (0.2-2.5)	4.9 (2.8-7.8)	3.5 (1.8-6.0)	4.6 (2.6-7.5)
Decline from 0-3 to 25-36	97.9%	79.5%	89.4%	84.8%

 

Events commonly occurring in the first three months were MAC (IR 21.3), Kaposi’s sarcoma (18.2), CMV (16.6) and PCP (12.1). In contrast, by the third year, events such as oesophageal candidiasis (3.2), TB (2.0), and non-Hodgkins lymphoma (1.4) were more likely to be diagnosed. Heterosexuals and IDU were more likely to develop bacterial events while homosexuals were more likely to develop viral events. Previous AIDS, lower CD4 and higher HIV RNA at baseline were associated with an increased risk of each type of event. Results were similar after excluding those with a prior AIDS diagnosis.

Conclusions: Dramatic declines are seen for events of all aetiologies, but the relative importance of different types of event changes with time since starting HAART.