Hill A, Ekong T, Gompels M, Pinching A; International Conference on AIDS.

Int Conf AIDS. 1991 Jun 16-21; 7: 289 (abstract no. M.B.2429).

St Mary's Hospital, London W2, UK

OBJECTIVE: Analysis of technical and biological influences on percentage and absolute CD4 counts, to identify potential confounding variables in the use of CD4 counts as markers of progression in HIV infected subjects. METHODS: 185 CD4 counts were taken from 63 lab workers and heterosexual controls in 1990, using whole blood lysis and flow cytometry (FACSCAN). Blood samples were taken for analysis between 0930 and 1030, and only when subjects were free from acute infection; none reported needle stick injury. RESULTS: The mean absolute CD4 count was 890 cells/cubic mm (range 340-1940) and mean %CD4 was 44.0 (range 27%-62%). The machine error of %CD4 from 32 duplicate samples, measured as coefficient of variation (CV = standard deviation/mean ratio) was 2% (sd=1.7%). The corresponding CV from sets of three %CD4 counts taken from each individual within 1 month was 4.2%, significantly higher than machine error (p=0.004), but there was no significant difference between monthly and 6 monthly fluctuations in %CD4 (p=0.173). Gender and ethnicity had no significant effect on mean %CD4 or absolute CD4 in this sample. One-way analysis of variance showed that the identity of the lab worker from whom a CD4 count was taken was a highly significant parameter explaining the variance in both %CD4 and absolute CD4, and other lymphocyte subpopulations (CD8, B cells) for 20 lab workers from whom repeated blood samples were taken (p less than 0.0001). One lab worker had a mean absolute CD4 count of 546 with 11/35 counts below 500, while another had a mean absolute count of 1125, with 0/49 values below 500 cells/cubic mm. No lab worker showed a significant decline in CD4 with time. CONCLUSIONS: The fluctuation of CD4 within restricted regions of the normal range in control subjects suggests that different HIV infected individuals will have had widely differing CD4 counts before seroconversion. Variance in CD4 from both machine error and non-HIV related longitudinal fluctuations needs to be accounted for in analysis of the prognostic power of CD4 in HIV infection.

Publication Types:

• Meeting Abstracts

Keywords:

• Acquired Immunodeficiency Syndrome
• CD4 Lymphocyte Count
• Disease Progression
• Flow Cytometry
• HIV Infections
• HIV Seropositivity
• Reference Values
• prevention & control

Other ID:

• 1242991

UI: 102183349

From Meeting Abstracts