CD4 surveillance in Scotland: perspectives on severe HIV-related immunodeficiency. CD4 Collaborative Group.

OBJECTIVE

To analyse factors that influence the following: the square root of first CD4 cell count; which individuals are severely immunodeficient at or before the start of monitoring; progression from the date of the earlier of two consecutive CD4 counts < or = 200 x 10(6)/l (termed as CD200) to AIDS.

SETTING

Scotland's HIV Immunology Laboratories and the Scottish Centre for Infection and Environmental Health.

PATIENTS

A total of 1679 adult HIV patients in Scotland to 31 December 1994 who had ever had a CD4 cell count < or = 500 x 10(6)/l or who had developed AIDS without any immunological monitoring, of whom 912 had developed severe HIV-related immunodeficiency (i.e., were CD200/AIDS cases).

RESULTS

Square-root first CD4 count was higher in women (by 2.1; SE, 0.5), in injecting drug users (IDU; by 1.8; SE, 0.5) and in younger patients (by 1.5 per 10 years; SE, 0.2), but reduced by 1.5 (SE, 0.1) per calendar year of recruitment, although it was relatively higher (by 3.8; SE, 0.8) for Edinburgh patients recruited in 1993-1994: at least 30% (nine out of 28) of new Edinburgh City Hospital patients in 1993-1994 with a first CD4 count of > or = 500 x 10(6)/l had seroconverted within the past 5 years. Two-thirds of non-IDU (67%; 348 out of 517) were already severely immunodeficient at or before the start of immunological monitoring, in contrast with only 31% of IDU CD200/AIDS cases. Overall, the odds on the CD200 date also being the date of first CD4 count have increased in recent times [log(e)(odds per calendar year of CD200 diagnosis), 0.14; SE, 0.05]. Analysis excluding patients whose AIDS diagnosis or follow-up time was within 1 month of the CD200/AIDS date supported a modest prolongation of the CD200/AIDS to AIDS interval for patients diagnosed with severe HIV-related immunodeficiency in the period 1989-1991 (log(e)[relative risk (RR)], -0.46; SE, 0.22). Similarly, year effects were evident on progression from CD500 to CD200/AIDS [log(e)(RR), -0.55; SE, 0.17) for CD500 cases diagnosed in 1989, and these year effects doubled in 1990-1992.

CONCLUSIONS

Minimal CD4 data were hypothesis-generating about region, risk group and calendar year. Lower bound for recent HIV incidence can be derived from new patients with first CD4 cell count above 500 x 10(6)/l if seroconversion intervals are available for a proportion.