Difficulties in assigning AIDS endpoints in clinical trials in resource limited settings: Lessons from DART.
- Department for International Development
- Uganda and Zimbabwe
- Document Type:
- Conference Paper
- Munderi, P., Gibb, D.M., Ssali, F., Reid, A., Burke, A. Wilkes, H., Tembo, R., Lutwama, F., Borok, M., Palfreeman, A., Winogron, D., Musana, H., Musoro, G., Taylor, K., and Peto, T.E.A.
Background: Stage 3/4 WHO criteria were designed to provide tools for HIV surveillance and staging of disease severity in the pre-HAART era. In 2002, the WHO clinical staging criteria were adapted to define clinical endpoints in the DART trial (2003-2008).
Methods: DART was designed to compare clinical outcomes under 2 management strategies in Sub-Saharan Africa. 3316 patients were recruited and followed for up to 6 years accruing 1021 reported WHO 4 events (excluding deaths) to December 2008. The protocol included definitive and presumptive criteria defining each WHO 4 endpoint. Where specific diagnosis could not be allocated to reported events, an independent endpoint review committee developed a \"syndromic\" categorisation scheme, including \"brain syndrome\" and \"lung syndrome\". Syndromes were classified as severe (severity comparable to WHO 4 events) or mild. We compared the rates of rejection of disease endpoints in DART and the large pre-HAART RCT (DELTA), completed 12 years previously.
Results: To December 2008, at independent review, 215 (21%) reported events were rejected in DART as not fulfilling the protocol criteria for trial endpoints, more than in DELTA. Lack of diagnostic investigative capacity was a major challenge in ascertaining clinical endpoints. On review of the rejected DART events it was possible to assign them into \"syndromia\" categorisations in 130/215 (60%) cases. In brain syndromes the lack of access to a CT scanner was the main reason for rejection of the presumptive and almost certainly correct diagnosis of toxoplasmosis 20/43 (47%) rejected events. In lung syndromes the lack of a typical chest X-ray with a documented response to treatment for PCP lead to most rejections (17/35 (49%) rejected endpoints.
Conclusions: Lack of diagnostic investigations is a challenge for clinical endpoint studies in resource-limited settings. Adaptation of criteria to include a syndromic categorisation for severe HIV-related events, and use of independent endpoint review committees, is recommended.
Presented at the 5th IAS Conference on HIV Pathogenesis, Treatment and Prevention, Cape Town, South Africa, 19-22 July 2009, 1 pp.