Home Search Abstracts View Session E-mail Abstract Author


Session 61 Poster Abstracts
Factors Influencing CD4 Counts
Session Day and Time: Tuesday, 1-4 pm
Room: Hall D


343
The Frequency and Clinical Implications of a Discordant CD4 Count and CD4 Percentage
M Gompels1, Caroline Sabin*2, A Phillips2, D Dunn3, and The UK Collaborative HIV Cohort (CHIC) Study
1North Bristol NHS Trust, UK; 2Royal Free and Univ Coll Med Sch, London, UK; and 3Med Res Council Clinical Trials Unit, London, UK

Background: Whilst some HIV-infected patients experience discordance between their absolute CD4 count and CD4%, the clinical implications of this, particularly among those starting HAART, is unclear.

Methods: Using data from the UK CHIC Study we described the frequency of CD4/CD4% discordance among untreated individuals and assessed its association with patient characteristics. All CD4/CD4% ‘pairs' (measured on the same day) from untreated individuals were grouped according to the CD4 count (in 50-cell strata); a discordant pair was one where the CD4% was <10th percentile (low discordancy; LD) or >90th percentile (high discordancy; HD) of values in that strata. Using Cox models, we assessed whether discordancy pre-HAART was associated with initial virological response (VL<50 cps/ml), viral rebound (confirmed VL >500 cps/ml), immunological response (increase in CD4 count at 6 months) or the development of a new AIDS event/death, after controlling for pre-HAART CD4 count and other confounders.

Results: 99,351 CD4/CD4% pairs (measured from 1980-2005) were included from 15,543 untreated patients (78% male; 61% homosexual, 30% heterosexual; 23% black African, median age at CHIC entry 33 yrs). Median (IQR) CD4 count and CD4% were 380 (252-535) cells/mm3 and 21 (15-28). By definition, around 10% showed LD and HD (LD:9.2%, HD:10.7%). For any CD4 count, CD4% was higher in women (p<0.001), injection drug users (p<0.001) and in more recent years (p<0.001), but lower in older subjects (p<0.001), those with a heterosexual (p=0.003) or other (p=0.01) risk and those of non-white ethnicity (p<0.001). 5,716 patients starting HAART had a CD4/CD4% pair 6 months after start. Median CD4 count increases over this period were 89, 110 and 141 cells/mm3 respectively in those with LD, no discordancy and HD (adjusted p<0.001). No significant associations were found between pre-HAART discordancy and initial viral response (LD: relative hazard 0.97 [95% CI 0.89-1.06]; HD: 0.96 [0.88-1.05]) or viral rebound (LD: 1.15 [0.95-1.41]; HD: 1.07 [0.86-1.33]). 411 patients (7%) developed a new AIDS event/died; there was no association between discordancy and clinical outcome (LD: 0.88 [0.63-1.22]; HD: 1.18 [0.88-1.60]).

Conclusions: CD4 increases on HAART were lower in those with LD compared to other patients, although the clinical implications of a discordant CD4/CD4% may be limited. If CD4 counts are monitored, there is little added cause for concern if the CD4% is lower than expected.