BACKGROUND: the number and proportion of adults diagnosed with HIV infection aged 50 years and older has risen. This study compares the effect of CD4 counts and anti-retroviral therapy (ART) on mortality rates among adults diagnosed aged ≥50 with those diagnosed at a younger age. METHODS: retrospective cohort analysis of national surveillance reports of HIV-diagnosed adults (15 years and older) in England, Wales and Northern Ireland. The relative impacts of age, CD4 count at diagnosis and ART on mortality were determined in Cox proportional hazards models. RESULTS: among 63,805 adults diagnosed with HIV between 2000 and 2009, 9% (5,683) were aged ≥50 years; older persons were more likely to be white, heterosexual and present with a CD4 count <200 cells/mm(3) (48 versus 32% P < 0.01) and AIDS at diagnosis (19 versus 9%, P < 0.01). One-year mortality was higher in older adults (10 versus 3%, P < 0.01) and especially in those diagnosed with a CD4 <200 cells/mm(3) left untreated (46 versus 15%, P < 0.01). While the relative mortality risk reduction from ART initiation at CD <200 cells/mm(3) was similar in both age groups, the absolute risk difference was higher among older adults (40 versus 12% fewer deaths) such that the number needed to treat older adults to prevent one death was two compared with eight among younger adults. CONCLUSIONS: the magnitude of benefit from ART is greater in older adults than younger adults. Older persons should be considered as a target for HIV testing. Coupled with prompt treatment, earlier diagnosis is likely to reduce substantially deaths in this group.