Background:  Carriage of Bw4 versus Bw6, 2 mutually exclusive variants within the human leukocyte antigen (HLA) system that influence HLA interactions with natural killer cells, has been associated with control of HIV infection. However, the influence of these alleles on the immunological and virological responses to HAART is unclear.

Methods:  All ART-naive, HLA-typed Caucasian adult males starting HAART in 1997-2002 with at least 12 months’ follow-up were considered, including 161 individuals from the Swiss HIV Cohort Study (SHCS) (101 Bw4⁺) and 92 from the Western Australia HIV Cohort Study (WAHIV) (50 Bw4⁺). Bw4 carriage (Bw4⁺) was assigned from high-resolution HLA typing. Trends in CD4 counts were analysed by linear mixed models and viral loads using logistic-normal models.

Results:  With respect to baseline (pre-treatment) status, Bw4-carriers had similar viral loads, follow-up time, initial HAART regimen, and mode of infection compared with individuals lacking Bw4 (all p >0.2). Baseline differences were noted within the Western Australian cohort (Bw4⁺ mean 4 years older) and within the Swiss cohort (mean CD4 cell count 90 cells/µL lower in Bw4⁺), however post-HAART CD4 trend profiles were very similar in both cohorts. CD4 percentages and counts were found to be consistently lower in Bw4-carriers, with differences established early in the treatment course and then estimated to be relatively constant 1 to 5 years post-HAART at 85 cells/µL (4 CD4%, p = 0.005) in the SHCS and 55 cells/µL (5.5 CD4%, p = 0.01) in WAHIV. The differences were particularly marked for carriage of Bw4-80T (see the figure) and remained significant after adjusting for baseline age, viral load, baseline CD4, and carriage of CCR5-Δ32. The proportion of Bw4 patients with undetectable viral loads (<400 copies/mL) was lower at 12 months post-HAART in the SHCS only (p = 0.01; WAHIV p = 0.2), with the difference dissipating by 20 months post-HAART. The differences in CD4 trends remained significant after adjusting for virological response rates (p = 0.04 and 0.002 for CD4% in SHCS and WAHIV, respectively), suggesting that Bw4 carriage has a more direct effect on CD4 cell dynamics post-HAART.

Conclusions:  Carriage of at least one Bw4 allele was associated with lower CD4 cell counts 1 to 5 years after commencing HAART in both cohorts. This finding contrasts the protective effects of Bw4 in untreated HIV-infection and might have implications for timing of treatment initiation as routine HLA typing becomes more common.