Second International Congress Drug Therapy in HIV Infection 18-22 November 1994 Glasgow, UK

CD4 RISES DURING ANTI-HIV TREATMENT: STATISTICAL ARTEFACT?

Andrew M Hill
Department of Medical Statistics Glaxo Research and Development Greenford, UK

Int Cong Drug Therapy HIV 1994 Nov 18-22;2:Abstract No. 3.3
AIDS 1994, Vol. 8 (Suppl. 4);S3

CD4 count is a primary marker of treatment efficacy in Phase I/II HIV trials: antiretroviral treatments causing larger and more sustained rises in CD4 counts are often judged to be more effective than those causing smaller, more transient rises.

In a European Phase II trial of monotherapy with the nucleoside analogue 3TC NUCB2001), CD4 counts were assessed at screening (Day - 14, CD4 <400 determined eligibility), prior to randomisation (Day -7'to Day -3) and once on the first day of treatment (Day 0).

CD4 counts rose significantly between the screening and Day 0 assessments median rise 30 cells, p<0.001). This rise may be caused by the 'regression to the mean' effect: trials requiring low CD4 counts at screening for study entry can select patients with transiently low CD4 counts, which then rise at subsequent visits.

CD4 rises while on treatment were calculated for 68 patient with complete data, using three different methods for calculating the baseline CD4 level. The size and duration of rises in CD4 during treatment depended critically on whether screening CD4 counts were included in the calculation of baseline CD4. With screening values excluded, CD4 rose by 28 cells, sustained for 14 weeks; with the mean of screening, pre-randomisation and Day 0 values used as baseline, the rise was 38 cells, sustained for 22 weeks, and with screening alone used as the baseline, the rise was 40 cells, sustained for 48 weeks.

Reported trials with other nucleoside analogues have often been analysed with screening CD4 counts included in the determination of the baseline count. If a similar non-treatment related rise in CD4 occurred between screening and baseline in these studies, the magnitude and duration of the CD4 responses observed may have been over-estimated by an artificial lowering of the true baseline value. The results also have implications for the use of CD4 counts in decisions on the initiation of antiretroviral treatment.

Presenting author: Andrew M Hill

1994-11-18
3.3

Originally published in AIDS Volume 8, Supplement 4 and hosted with permission of the publisher Lippincott Williams & Wilkins, 250 Waterloo Road, London, SE1 8RD, UK. Tel: +44 (0)20 7981 0700 Fax: +44 (0) 7981 0701

Copyright © 1994 - Lippincott Williams & Wilkins. All rights reserved. All abstracts from the 2nd International Congress Drug Therapy in HIV Infection, appearing on the AEGiS web site, are protected by United States copyright law and may not be reproduced, distributed, transmitted, displayed, or otherwise published without the prior written permission of Lippincott Williams & Wilkins. You may not alter or remove any trademark, copyright or other notice. However, provided that you maintain all copyright and other notices contained therein, you may download material (one machine readable copy and one print copy per page) for your personal, non-commercial use only.

http://www.aidsonline.com http://www.ovid.com