OBJECTIVE: We describe the CD4 database of the Scottish Immunology Laboratories, and its uses and limitations for making short-term predictions of a CD4 cell count less than or equal to 200 x 10(6)/l (CD4(200)) and of adult AIDS cases in Scotland. DESIGN: The date of the earlier of two consecutive samples (typically 3 months apart) both with CD4 cell counts less than or equal to 200 x 10(6)/l was taken to define when a patient had passed the CD4(200) threshold (referred to as a CD4(200) case). The CD4 database comprises HIV-1-seropositive adults in the four main risk groups [homosexual/bisexual (1), injecting drug users (IDU; 2), heterosexual contact (3), and undetermined (9)] from Scotland's three principal areas of population (Lothian, Tayside and Strathclyde) who have had a CD4 cell count of less than or equal to 500 x 10(6)/l. SETTING: Three hospitals in Scotland, the Communicable Diseases (Scotland Unit) and the Medical Research Council Biostatistics Unit, Cambridge, UK. PATIENTS, PARTICIPANTS: The CD4 database at 31 December 1990 listed 813 patients (of whom 52% were IDU): 390 were CD4(200)/AIDS cases (of whom 44% were IDU) and 192 were AIDS cases (of whom 32% were IDU). RESULTS: Individuals in risk groups 1, 2 and 3 were nearly equally represented among newly diagnosed HIV-1 infections in 1990. However, among patients with moderate immunodeficiency, IDU accounted for 50% of the total number. Co-incidence of first CD4 cell count with CD4(200) diagnosis was recorded for only 28% of IDU, but in over 50% of cases for each of the other exposure groups (57%). There was a highly significant decrease of around 80 x 10(6)/l per calendar-year-of-referral in first CD4 cell counts for patients on the CD4 database; and decreases of around 40 x 10(6)/lper decade of age at referral. Since 1988, median time from CD4(200) to AIDS diagnosis in Scotland has been approximately 2 years. Back-projection was applied to annual CD4(200)/AIDS diagnoses before 31 December 1990 and to AIDS diagnoses. From AIDS diagnoses, the central epidemic scenario underestimated past HIV-1-antibody-positive reports (up to the end of 1985). More dramatic underestimation was occasioned by back-projection from CD4(200)/AIDS diagnoses [319 inferred HIV infections compared with 445 HIV-1-antibody-positive reports to Communicable Diseases (Scotland) Unit]. CONCLUSIONS: First CD4 cell counts should complement new HIV-1 diagnoses. Past referrals for immunological monitoring were not uniform between risk groups in Scotland. Underascertainment of CD4(200) cases is a problem when CD4(200) cases are used as a basis for back-projection. More information concerning the incubation distribution from HIV seroconversion to CD4(200) diagnosis is required. It is likely that there are twice as many CD4(200)/AIDS as there are diagnosed cases of AIDS.
OBJECTIVE: We describe the CD4 database of the Scottish Immunology Laboratories, and its uses and limitations for making short-term predictions of a CD4 cell count less than or equal to 200 x 10(6)/l (CD4(200)) and of adult AIDS cases in Scotland. DESIGN: The date of the earlier of two consecutive samples (typically 3 months apart) both with CD4 cell counts less than or equal to 200 x 10(6)/l was taken to define when a patient had passed the CD4(200) threshold (referred to as a CD4(200) case). The CD4 database comprises HIV-1-seropositive adults in the four main risk groups [homosexual/bisexual (1), injecting drug users (IDU; 2), heterosexual contact (3), and undetermined (9)] from Scotland's three principal areas of population (Lothian, Tayside and Strathclyde) who have had a CD4 cell count of less than or equal to 500 x 10(6)/l. SETTING: Three hospitals in Scotland, the Communicable Diseases (Scotland Unit) and the Medical Research Council Biostatistics Unit, Cambridge, UK. PATIENTS, PARTICIPANTS: The CD4 database at 31 December 1990 listed 813 patients (of whom 52% were IDU): 390 were CD4(200)/AIDS cases (of whom 44% were IDU) and 192 were AIDS cases (of whom 32% were IDU). RESULTS: Individuals in risk groups 1, 2 and 3 were nearly equally represented among newly diagnosed HIV-1 infections in 1990. However, among patients with moderate immunodeficiency, IDU accounted for 50% of the total number. Co-incidence of first CD4 cell count with CD4(200) diagnosis was recorded for only 28% of IDU, but in over 50% of cases for each of the other exposure groups (57%). There was a highly significant decrease of around 80 x 10(6)/l per calendar-year-of-referral in first CD4 cell counts for patients on the CD4 database; and decreases of around 40 x 10(6)/lper decade of age at referral. Since 1988, median time from CD4(200) to AIDS diagnosis in Scotland has been approximately 2 years. Back-projection was applied to annual CD4(200)/AIDS diagnoses before 31 December 1990 and to AIDS diagnoses. From AIDS diagnoses, the central epidemic scenario underestimated past HIV-1-antibody-positive reports (up to the end of 1985). More dramatic underestimation was occasioned by back-projection from CD4(200)/AIDS diagnoses [319 inferred HIV infections compared with 445 HIV-1-antibody-positive reports to Communicable Diseases (Scotland) Unit]. CONCLUSIONS: First CD4 cell counts should complement new HIV-1 diagnoses. Past referrals for immunological monitoring were not uniform between risk groups in Scotland. Underascertainment of CD4(200) cases is a problem when CD4(200) cases are used as a basis for back-projection. More information concerning the incubation distribution from HIV seroconversion to CD4(200) diagnosis is required. It is likely that there are twice as many CD4(200)/AIDS as there are diagnosed cases of AIDS.