Home Search Abstracts View Session E-mail Abstract Author


Session 182 Poster Abstracts
ART Toxicity and Outcomes
Session Day and Time: Tuesday, 1-2:30 pm
Poster Hall


986    
CD4 Cell Count 250 Cells/mm3 Does Not Predict Rash-associated Hepatotoxicity among Women Initiating Nevirapine-based ART in Zambia, Thailand, and Kenya
Philip Peters*1, J Stringer2, M McConnell3, J Kiarie4, W Ratanasuwan5, P Intalapaporn6, D Potter7, W Mutsotso8, I Zulu7, and P Weidle1
1CDC, Atlanta, GA, US; 2Univ of Alabama at Birmingham, Ctr for Infectious Disease Res, Zambia, Lusaka; 3Thailand Ministry of Publ Hlth-US CDC Collaboration, Nonthaburi; 4Kenyatta Natl Hosp, Univ of Nairobi, Kenya; 5Siriraj Hosp, Mahidol Univ, Bangkok, Thailand; 6Rajavithi Hosp, Bangkok, Thailand; 7CDC Zambia, Lusaka; and 8CDC Kenya, Nairobi

Background:  Nevirapine (NVP) -based ART is the most common HIV treatment regimen in Asia and Africa. In Europe and North America an estimated 11% of women initiating NVP at CD4 counts ≥250 cells/mm3 develop potentially life-threatening rash-associated hepatotoxicity (RAH); however, few data exist from resource-limited settings.

Methods:  Between May 2005 and January 2007, we enrolled ART-naive women initiating NVP-based ART in a prospective observational cohort study. ART was administered according to national treatment guidelines. At 2, 4, 8, 16, and 24 weeks, we measured participants’ liver function tests (LFT) and evaluated them clinically for rash. We graded hepatotoxicity as:  grade 1/mild (transaminase [AST or ALT] elevation 1.25 to 2.49 times the upper limit of normal [x ULN]), grade 2/moderate (2.5 to 4.99 x ULN), and grade 3/severe (≥5.0 x ULN). We defined RAH as an ALT or AST elevation > grade 2 with concomitant rash. We used logistic regression to identify factors independently associated with RAH.

Results:  We enrolled 820 women (497 Zambian, 192 Thai, and 131 Kenyan) with a median age of 32 years, a median CD4 of 149 cells/mm3, and a median HIV viral load of 108,000 copies/mL. Any hepatotoxicity ≥grade 2 occurred in 109 (13%) women; in 41 (5%) it was severe. RAH occurred in 27 (3%) women (Zambia 2%, Thailand 7%, Kenya 2%):  8 (6.5%) of 123 women with a baseline CD4 <50 cells/mm3; 13 (2%) of 576 women with a CD4 of 50 to 250 cells/mm3; and 6 (5%) of 121 women with a CD4 >250 cells/mm3. In an adjusted model, RAH was significantly associated with baseline abnormal (≥grade 1) ALT or AST (adjusted odds ratio [aOR] = 3.0, 95% confidence interval [CI] = 1.2 to 7.9), and Thai ethnicity (aOR 4.2, 95%CI 1.7 to 10.4) but not with a baseline CD4 ≥250 cells/mm3 (aOR 1.7, 95%CI 0.6 to 4.4). Three participants (0.4%) died with symptoms suggestive of fatal hepatotoxicity, of whom q had preceding RAH. All 3 deaths had a baseline CD4 <100 cells/mm3 and had received concomitant anti-tuberculosis medications.

Conclusions:  Among women taking NVP-based ART, RAH was associated with abnormal baseline LFT and Thai ethnicity but not with CD4 ≥250 cells/mm3. RAH occurred at a lower rate than expected among women with a CD4 ≥250 cells/mm3, but at an unexpectedly high rate among women with a CD4 <50 cells/mm3. In resource-limited settings monitoring for RAH in women on NVP-based ART is indicated but cannot be guided by CD4 cell count.