Table 4: Change over time in square root CD4 cell count by HIV-1 subtype and patient ethnicity
Slope estimate (95% CI)
Selected CD4 cell counts*Unselected CD4 cell counts†
All patients included in analysisOnly patients of African ancestry included in analysisOnly patients of other ethnicity included in analysisAll patients included in analysis
Mixed model‡
With ethnicityWithout ethnicityWith ethnicity
Change per year: patients of other ethnicity with viral subtype B
1.23 (1.27 to 1.19)1.21 (1.25 to 1.17)0.88 (1.03 to 0.73)1.23 (1.27 to 1.19)1.35 (1.40 to 1.31)
Difference in change per year compared with the change in patients of other ethnicity with viral subtype B
Viral subtype
A0.06 (0.13 to 0.25)0.21 (0.02 to 0.39)0.22 (0.04 to 0.49)0.03 (0.22 to 0.28)0.30 (0.10 to 0.49)
AE–0.03 (0.26 to 0.21)–0.05 (0.28 to 0.19)3.2 (8.6 to 2.2)§0.04 (0.28 to 0.20)0.03 (0.23 to 0.28)
AG0.17 (0.04 to 0.39)0.34 (0.13 to 0.55)0.26 (0.02 to 0.54)0.24 (0.08 to 0.55)0.23 (0.00 to 0.46)
C0.11 (0.04 to 0.27)0.32 (0.19 to 0.46)0.32 (0.12 to 0.52)0.02 (0.24 to 0.20)0.31 (0.15 to 0.47)
G0.02 (0.32 to 0.36)0.19 (0.14 to 0.52)0.46 (0.06 to 0.86)0.42 (0.92 to 0.09)0.00 (0.37 to 0.36)
African ancestry0.37 (0.25 to 0.50)0.44 (0.31 to 0.57)

Note: CI = confidence interval.
*Patients aged 16 years or older after January 1998 with CD4 cell counts likely to be within the stable phase of chronic untreated HIV infection.
†All CD4 cell counts for patients aged 16 years and older after January 1998.
‡Mixed model with a random intercept and random slope for each patient. All cohorts except Aquitaine. With intercepts for cohort and with both intercepts and slopes for age, sex, ethnicity, initial HIV RNA, and injection drug use as the likely mode of transmission.
§Few patients of African ancestry had viral subtype AE; therefore, this point estimate is unreliable.