|
Effect of WHO stage 3/4 events after ART initiation on survival of HIV-infected African adults in the DART trial
C. Kityo1, D. Ford2, A.S. Walker2, J. Hakim3, P. Munderi4, F. Lutwama5, F. Ssali1, A. Reid3, H. Grosskurth4, D.M. Gibb2, C.F. Gilks6, A.G. Babiker2
1JCRC, Kampala, Uganda, 2MRC Clinical Trials Unit, London, United Kingdom, 3University of Zimbabwe, Harare, Zimbabwe, 4MRC/UVRI Uganda Research Unit on AIDS, Entebbe, Uganda, 5Infectious Diseases Institute, Makerere University, Mulago, Uganda, 6Imperial College, London, United Kingdom
Background: In middle/low-income settings, where access to laboratory measurements is limited, WHO stage 3/4 events are often used to determine success/failure of ART. However, mortality associated with different events likely varies.
Methods: Participants initiated triple ART in the DART randomised trial of management strategies in symptomatic ART-naive African adults with CD4< 200 cells/mm3. Marginal structural models with stabilized time-dependent inverse probability treatment weights were used to estimate the causal effect of WHO stage 3/4 events on subsequent mortality.
Results: To December 2007, 3179 participants contributed 11236 years follow-up and 281 deaths. 518 participants (16%) switched to second-line (43 deaths on second-line). On first-line[second-line] ART, 123[6] participants had cryptococcosis (48 deaths), 177[9] had pulmonary tuberculosis (40 deaths), 110[8] had extrapulmonary tuberculosis (19 deaths), 178[3] had oesophageal candidasis (25 deaths) and 405[7] had oral candidiasis (59 deaths). As expected, WHO stage 3/4 events occurred more frequently among participants with lower CD4 counts, lower haemoglobin, recent weight loss and a history of other WHO 3/4 events. Overall (including first- and second-line), cryptococcosis increased all-cause mortality risk 6.01-fold (95% CI 3.65-9.87), pulmonary tuberculosis 2.84-fold (1.65-4.87), extrapulmonary tuberculosis 1.61-fold (0.72-3.62), oral candidiasis 1.98-fold (1.28-3.06) and oesophageal candidasis 0.91-fold (0.39-2.09). The risk of death was higher in the 12 weeks following cryptococcosis (OR=11.33 (6.21-20.67); 25 deaths) than subsequently (OR=4.10 (2.25-7.48); 23 deaths). Increased mortality risks in the 12 weeks following an event were also observed for pulmonary tuberculosis (OR=7.41 (4.13-13.29)), extrapulmonary tuberculosis (OR=3.56 (1.38-9.16)) and oral candida (OR=3.17 (1.86-5.40)). Risks >12 weeks after diagnosis were 1.46 (0.61-3.50), 0.97 (0.32-2.98) and 1.56 (0.83-2.94), respectively.
Discussion: Mortality rates following a WHO stage 3/4 event vary considerably with diagnosis; of note, some WHO 3 events have greater mortality impact than WHO 4 events. More work is needed to evaluate the impact of switching to second-line therapy following an event on mortality risk.
Download the e-Poster
|