BACKGROUND: Suboptimal improvement in CD4+ T-cell count is not uncommon in HIV-infected patients with suppressed plasma HIV RNA levels, and a decrease in CD4+ T-cell count in patients with suppressed or low-level viraemia has been observed. METHODS: Our objectives were to identify the prevalence of decreasing CD4+ T-cell counts during suppressed or low-level viraemia, to determine the frequency of clinical events during and immediately after such decreases, and to examine for associations with individual variables. A matched case-control study was undertaken using the Duke Infectious Diseases Clinic database (n = 3,949). Cases had at least two consecutive significant decreases in either CD4+ absolute count or CD4+ percentage, while also having plasma HIV RNA levels < 1,000 copies/ml. RESULTS: The prevalence of decreasing CD4+ T-cell counts during suppressed or low-level viraemia was 1.22%. Only three HIV-associated clinical events occurred. The majority of cases had an increase in the CD4+ T-cell count immediately following the study period. The use of either zidovudine or stavudine was weakly associated with decreasing CD4+ T-cell counts in a multivariable analysis, but this association was not present in cases with only a decrease in CD4+ T-cell percentage. CONCLUSIONS: Decreasing CD4+ T-cell counts during suppressed or low-level viraemia are rare, typically transient, and not associated with an increase in HIV-associated clinical events.