CD4 cell count and initiation of antiretroviral therapy: trends in seven UK centres, 1997-2003

Objectives We examined whether the timing of initiation of antiretroviral therapy (ART) in routine clinical practice reflected treatment guidelines, which have evolved towards recommending starting therapy at lower CD4 cell counts. Methods We analysed longitudinal data on 10 820 patients enrolled in the UK Collaborative HIV Cohort (UK CHIC) Study, which includes seven large clinical centres in south-east England. CD4 cell and viral load measurements performed in the period between 1 January 1997 and 31 December 2003 were classified according to whether ART was subsequently initiated or deferred, to estimate the probability of ART initiation by CD4 count and viral load over time. The effect of nonclinical factors (age, sex, ethnicity, and exposure category) was analysed by logistic regression. Kaplan-Meier analysis was used to estimate the proportion of patients who had initiated ART by a particular CD4 count among 'early' presenters (initial CD4 cell count > 500 cells/mu L). Results There was a tendency to initiate ART at lower CD4 cell counts over time in the years 1997-2000, especially in the range 200-500 cells/mu L, with little change thereafter. An estimated 34% of HIV-infected individuals having presented early initiated ART at a CD4 count <200 cells/mu L. We also found an independent influence of viral load, which was particularly pronounced for CD4 <350 cells/mu L. Use of injection drugs was the only nonclinical factor associated with initiation of ART at lower CD4 cell counts. Conclusions The initiation of ART in the clinics included in this analysis reflected evolving treatment guidelines. However, an unexpectedly high proportion of patients started ART at lower CD4 counts than recommended, which is only partly explained by late presentation