Home Search Abstracts View Session E-mail Abstract Author


Session 134 Poster Abstracts
Complications in Resource-limited Settings
Session Day and Time: Wednesday, 1-4 pm
Room: Hall B


845
Cause-specific Mortality and Contribution of Immune Reconstitution Inflammatory Syndrome during the First 2 Years of ART in Uganda
A Kambugu1, Barbara Castelnuovo*1, A Kiragga1, K Kigonya1, M Kamya2, and P Easterbrook1
1Infectious Diseases Inst, Kampala, Uganda and 2Makerere Univ, Kampala, Uganda

Background:  Recent data suggest that immune reconstitution inflammatory syndrome (IRIS) may be a significant cause of death in the first months of ART in low-income countries. We determined cause-specific mortality and contribution of IRIS during the first 24 months of ART in an urban cohort in Uganda.

Methods:  A prospective cohort of 559 patients initiated on stavudine (d4T)/zidovudine (AZT), lamivudine (3TC), and nevirapine (NVP)/efavirenz (EFV) between April 2004 and April 2005 at the Infectious Diseases Institute. Patients were reviewed monthly. Cause of death was determined from medical record review and other sources. IRIS was defined as new onset with atypical presentation (unmasking) or worsening of an ongoing opportunistic infection (paradoxical) within 6 months of ART initiation. A multivariate Cox model was used to identify independent risk factors of HIV-related mortality.

Results:  Of 559 patients, 70% were female. At ART initiation, 88% were World Health Organization (WHO) stage 3-4; median age was 36 years; CD4+ count and viral load were 104 cell/mL and 5.4 logs, respectively; 74% were started on NVP and 36% on EFV-based regimens. Of the total, 22 (3.9%) were lost to follow-up; 80 (14%) died during the first 12 months (incidence was 181/100 person-years of ART), of which 58 (73%) within the first 3 months; 15 died (3%) during the second year (10.9/100 person-years of ART). In the first year, 69 (86%) of the deaths were HIV-related:  24% due to central nervous system (CNS) infections (cryptococcal meningitis, toxoplasmosis, and Herpes zoster); 14.5% to tuberculosis, 10.1% to Kaposis' sarcoma, and 7.2% to Pneumocystis jirovecii pneumonia; 2 deaths (2.5%) were due to ART mitochondrial toxicity. Of 69 deaths, 5 (7%) were attributed to IRIS, 4 to unmasking IRIS (2 extra pulmonary TB, 1 cryptococcal meningitis, 1 intracerebral mass), and 1 to paradoxical cryptococcal meningitis. In the second year, only 4 of 15 (27%) deaths were HIV related, 8 (53%) due to other medical conditions and 3(20%) to ART mitochondrial toxicity. Risk factors for HIV-related death were WHO stage 3/4 (HR 3.38, CI 1.02 to 11.7) and baseline CD4+ count <25 cell/mL (HR 2.52, CI 1.41 to 4.47). The median CD4+ count at ART initiation (24 vs 122 cell/mL) and death (36 vs 86 cell/mL) was significantly lower in HIV-related versus non-HIV-related deaths (p <0.006 and p = 0.046, respectively).

Conclusions:  Main causes of the high HIV-related mortality in the first year of ART were mycobacterial disease and CNS infections associated with advanced immunodeficiency. IRIS accounted for only 7% of these deaths. In contrast, other co-morbidities and ART toxicity accounted for the majority of deaths in the second year.