The source of infection was unknown for 16 subjects. The virus donor could be identified for the accidentally infected subject 8, for subject 2, a monogamous heterosexual individual with a confirmed seropositive Zairean partner as single-risk factor, and for 1 of the 11 cohort participants, subject 17, a monogamous homosexual man. For subject 17, the origin of the virus was confirmed by comparison of the third hypervariable envelope domain (V3) sequences of HIV-I RNA in his serum and that ofhis partner, revealing the same uncommon sequence variant [II] in both (Wolfs T, personal communication). Healthy heterosexual individuals, healthy HIV-seronegative homosexual men at high risk, and HIV-seronegative individuals with acute influenza, rubella, cytomegalovirus, or Epstein-Barr virus infection were used as a control population for the immunologic parameters.
Serology. Antibodies to HIV-I were determined by commercial assay (HIV 1/2 EIA; Abbott Laboratories, North Chicago) and confirmed by Western blot. HIV-I p24 antigen in serum was detected in a commercial antigen capture assay (Abbott Laboratories).
Viral isolation. HIV-I was recovered from cryopreserved peripheral blood mononuclear cells (PBMC) as described. Briefly, PBMC of infected individuals were cocultivated with 2-day phytohemagglutinin-stimulated peripheral blood lymphocytes from a seronegative blood donor. This cocultivation procedure was repeated each week. Cultures were observed for syncytium formation 3 times a week as described [4, 14]. Twice weekly culture supernatants were tested for the presence of viral p24 [14, 15]. Virus was recovered from >95% of the PBMC samples.
Immunologic studies. PBMC were isolated from heparinized venous blood by density-gradient centrifugation on ficoll-hypaque. Lymphocyte immunophenotyping for CD4 and CD8 cells was done by flow cytometry. Anti-CD38 and anti-HLA-DR monoclonal antibodies (Central Laboratory of the Netherlands Red Cross Blood Transfusion Service; Becton Dickinson Immunocytometry Systems, San Jose, CA) were used to define the activation state ofCD8+ T cells. Median values and ranges (10-90 percentiles) ofT cell subsets relevant to this study in the HIV-seronegative heterosexuals were: CD4+ cells, 0.89 X 109/1 (0.57-1.38); CD8+ cells, 0.52 X 109/1 (0.30-0.91); CD38+/ CD8+ cells, 6%(0-14) of total CD8+ T cells. In the HIV-seronegative homosexual control group the median values and ranges were: CD4+ cells, 0.7 X 109/1 (0.45-1. I); CD8+ cells, 0.9 X 109/1 (0.45-1.3); CD38+/CD8+ cells, 10% (3-27) of total CD8+ T cells.
Results
Clinical findings during primary infection. Virologic and immunologic studies were done of 19 recently infected individuals, 8 ofwhom (patients 1-8) could be studied in detail because frequent samples were obtained. Patients 1-7 presented clinically with symptoms of acute HIV-I infection (table I). Patient 8 was accidentally infected with a small amount of blood from a patient with end-stage HIV-l-related disease during a diagnostic procedure, 14 days after a partial jejunum resection. He received zidovudine within 45 min after exposure and did not develop clinical symptoms of acute HIV-I infection. Patients 1-6 and 8 had no evidence of other recent infections, patient 7 early after seroconversion had serologic evidence ofrecent syphilis and hepatitis B. The other II individuals did not undergo physical examination at the early stage of infection. Retrospective evidence for symptoms compatible with acute HIV-l infection was obtained for 8 of these.