Acute HIV infection is a critical point in the natural history of HIV disease, because of the lack of early immune responses  and the severity of clinical features associated with disease progression [2–4]. Because of the broad range of clinical features, acute HIV infection is often underdiagnosed or misdiagnosed, resulting in additional HIV transmission  and missed treatment opportunities. In addition, misdiagnosis or underdiagnosis hampers enrolment in clinical studies which are needed to evaluate new treatment strategies for acute HIV infection. It is of interest to explore whether clinical presentation differs by geographic location, which could result in bias in enrolment in studies. If such bias occurs in randomized or observational studies, it would limit national or international comparisons. The objective of this investigation was to explore clinical features, CD4 and CD8 T-cell levels within 3 months of the onset of acute HIV infection among persons diagnosed in Geneva, Seattle and Sydney. To study the early natural history of acute HIV-1 infection, we restricted the analysis to patients who did not receive antiretroviral treatment during the first 3 months after acquisition of HIV infection.
One-hundred and sixty patients with a documented acute HIV-1 infection were included in this study, as described elsewhere [6,7]. The patients from Geneva (n = 36) were 12 individuals with symptomatic seroconversion enrolled in the Geneva section of Swiss HIV Cohort Study  and 24 individuals included in the placebo arm of a randomized controlled trial (RCT) of zidovudine in acute HIV infection . The patients from Sydney (n = 81) were individuals enrolled in the AIDS Prospective Study that began in Sydney in 1987  and in a cohort of patients with acute HIV infection  (seven patients out of 81 were included in the placebo arm of the above RCT ). The patients from Seattle (n = 43) were enrolled in a prospective study of the natural history of HIV infection . A total of 129 individuals were enrolled in observational cohorts and 31 were enrolled in a RCT. The number of individuals per year of inclusion were 60 before 1992, 53 from 1992 to 1993, and 47 from 1994 to December 31, 1996.
The clinical features were reported by the patients and by their treating physician. All recent symptoms and signs occurring at or after the presumed date of HIV-1 infection were noted on case report forms in each centre. The median time after presumed HIV infection was 20.5 days (range, 5–90 days). The data collection form was almost identical in Geneva and Sydney because of a previous collaborative project . The data collection form in Seattle contained minor differences. For example, the diagnosis of fever was based on a temperature ≥ 38°C for the Geneva–Sydney cohort and > 37.5°C for USA patients. Comparisons were then based on a temperature ≥ 38°C. The diagnosis of cervical and supraclavicular adenopathy was treated as one item in Seattle and treated as two items for the Geneva–Sydney cohort. We have combined these data into one item as cervical and/or supraclavicular adenopathy. No patients received any antiretroviral therapy during the 3 months of the study.
Diagnosis of acute HIV infection and T-cell counts
Virological and immunological studies were performed in laboratories affiliated with each university centre. The laboratories at the University of Washington are certified by the American College of Pathology and the AIDS Clinical Trials Group. The laboratories in Geneva and Sydney are national reference centres and quality controls on the testing are performed periodically. In Geneva, quality control is performed every 3 months with external quality control for lymphocyte markers and serology (i.e., Nequas, UK), in Sydney quality control was carried out in accordance with Centers for Disease Control guidelines  and in Seattle, the laboratories participated in national quality control programme . The T-cell subsets were measured by flow cytometry at all three locations. In Geneva, the CD4 and CD8 cell counts were determined by flow cytometry (Coulter Epics IV, Instrumentegesellscheft, AG, Basel, Switzerland) using fluorochrome-conjugated monoclonal antibodies. In Seattle, T-cell subset analysis was carried out by flow cytometry (Coulter Epics XL-MCL) using tetraONE SYTEM which combined a four-colour fluorescent monoclonal antibody reagent. In Sydney, T-cell subset were determinate using CD4 and CD8 phosphatidylethanolamine monoclonal antibodies (Coulter Electronics, Hialeah, Florida, USA), whole blood lysis (Q-PREP; Coulter), and flow cytometer analysis (EPICS Profile I or XL; Coulter).
The criteria for diagnosis of acute HIV infection were the following: (i) positive HIV p24 antigenaemia in a patient previously HIV seronegative on enzyme immunoassay (EIA) or with a indeterminate level of antibodies (n = 93); (ii) a Western-blot having two positive bands, one of which corresponded to the env gene [glycoprotein (gp) 160, gp 120, or gp 41] before the EIA became reactive (n = 20); (iii) an interval time ≤ 90 days between the last HIV seronegative EIA and the first seropositive EIA confirmed by a positive Western-blot assay (n = 33); (iv) a seronegative EIA for HIV not more than 1 year previously, clinical features of acute HIV infection within 3 months before study entry and a seropositive EIA confirmed by a positive Western-blot assay (n = 9); (v) a patient negative for HIV on EIA 6 months previously and positive at enrolment, with clinical features of acute HIV-1 infection developing 3 months before study entry (n = 3). These criteria were applied in a uniform manner over time.
All patients were confirmed later to be HIV seropositive by Western blot assay during the follow-up (Biotech, Dupont, Geneva, Switzerland for Geneva; Epitope, Inc., Beaverton, Oregon for Seattle; Biorad laboratories, USA, for Sydney).
Plasma HIV RNA was detected using the Amplicor HIV Monitor (Roche, Basel, Switzerland) in Geneva, the Chiron (Emeryville, California, USA) branched-chain DNA assay in Seattle, and by Amplicor HIV monitor test; Roche Molecular Systems (Branchburg, New Jersey, USA), in Sydney.
Once plasma HIV RNA testing became available, plasma HIV RNA was assessed in 91 patients at an average time of 39.8 days after the onset of symptoms (median 27.0 days; range, 3–132 days). The mean was 653 652 copies/ml (median, 223 880 copies/ml; range, 1561–2 240 000 copies/ml).
Continuous variables were summarized using median, mean and SD and categorical variables using frequencies and percentages. Because of the two types of patient recruitment (observational cohort versus RCT), the sample was stratified into four groups taking account of the mode of recruitment and the location: patients from Geneva and Sydney enrolled in the RCT (group 1), patients enrolled in the Geneva observational cohort (group 2), patients enrolled in the Sydney observational cohort (group 3), and patients enrolled in the Seattle observational cohort (group 4). Multiple regression analysis was used to identify the variables independently associated with the duration of acute HIV-1 infection. The chi-square test was used for comparing proportions and one-way analysis of variance (ANOVA) for comparing continuous variables among these groups after logarthimic transformation if necessary.
To analyse the T-cell subsets, the duration of follow-up was split in to four periods of time following the onset of acute HIV-1 infection: 0–15 days, 16–31 days, 32–63 days and 64–100 days. The differences in mean CD4 and CD8 load between centres were assessed using the random-effect models for the longitudinal data . This model uses all measurements available during the follow-up. The total number of CD4 T-cell measurements was 77 for Geneva, 75 for Seattle and 180 for Sydney. The total number of CD8 T-cell measurements was the same. A two-tailed P < 0.05 was considered significant. The statistical analysis was performed with SPSS version 6.0 (SPSS, Chicago, Illinois, USA) and with the Splus statistical package, version 3.1 (StatSci, Seattle, Washington, USA).
Baseline characteristics and clinical features
Baseline characteristics of patients are reported in Table 1. The frequencies of clinical features are reported in Table 2. We observed differences in duration of acute retroviral syndrome among the patient groups (ANOVA, 3; 146 d.f.;P = 0.01) (Fig. 1). There was a longer duration for patients enrolled in the RCT versus the observational studies (t test, P < 0.0001). For the most common symptoms we observed a longer mean duration for patients enrolled in the RCT versus the observational cohort [i.e., 27.3 days versus 10.7 days for fever (P < 0.0001), 25.7 days versus 9.5 days for skin rash (P < 0.0001), 50.8 days versus 18.8 days for headaches (P < 0.0001), 31.9 days versus 10.7 days for arthralgia (P = 0.001), and 32.2 days versus 11.5 days for myalgia (P < 0.0001)]. There were no significant differences among the patients enrolled in the observational studies (ANOVA, 2; 116 d.f.;P = 0.41). The linear regression analysis adjusted for HIV risk factors (persons homosexual or bisexual versus other), age, year of HIV infection and location, showed that the inclusion in the RCT was the only variable statistically associated with the duration of primary HIV infection (linear coefficient β, 13.547; SE, 6.955;P = 0.05). The duration of several clinical features by patient groups is shown in Fig. 1 : it suggests that a bias occurred with patients with a longer duration of symptoms, and probably more severe acute HIV infection, enrolled in the RCT than in the observational studies.
CD4 T cells and CD8 T cells measurements
Figure 2 shows the CD4 and CD8 T cells measurements by groups of patients according to time of follow-up. Using a random effects model with terms for site and time period (0–15, 16–31, 32–63, 64–100 days following seroconversion), there were no significant differences between sites or according to type of recruitment (RCT versus observational) for either CD4 or CD8 values. The average differences between the centres ranged from –7 × 106/l to 52 × 106/l for the CD4 T cells and from –283 × 106/l to 188 × 106/l for the CD8 T cells (Table 3).
The objective of this study was to determine whether differences existed by geographic location in the clinical presentation of acute HIV infection and in the early T-cell changes among patients not treated with antiretroviral therapy.
There were no consistent differences in the frequency of clinical features observed between different locations. However, heterogeneity of the definitions used at the different sites must be taken into consideration for interpretation of the results. Limited differences occurred only for lethargy and can be related to minor variations in definitions rather than clinical differences. This result is reassuring from the perspective of collaborative studies as differences appear to be related more to patient recruitment than to location.
The duration of acute HIV-1 infection differed between patients enrolled in the RCT and in the observational cohorts, but was not different among patients enrolled in the different observational cohorts. This result confirms methodological reports showing that patients enrolled in RCT differ from those participating in observational studies . Duration of acute HIV infection has been associated with disease progression  and might affect outcome independently of the mode of patient selection . Thus duration of acute infection might be taken into account in clinical trials by using a stratified randomization .
The reasons for selecting a patient for a RCT can be related to the treating physician, to the patient or to both. In the presence of a long illness, if the patient has sought medical care, the physician has more time to propose a RCT to his or her patient . On the other hand, patients with a longer duration of illness may be more likely to consult their physicians repeatedly and are thus more likely to join a RCT. It would be helpful to have a detailed description of the patients who qualified for the RCT but were not included. Indeed patients with a short duration of symptoms associated with acute HIV-1 infection may not seek further medical assistance and would therefore not be enrolled. If the RCT was conducted in a sicker population, a treatment effect may be found more readily.
An important finding is that there is no significant difference between the cohorts in term of T-cell subsets. The CD4 and the CD8 T-cell responses did not differ by location or by selection criteria within 100 days of the onset of acute HIV-1 infection (P > 0.1). The estimated difference in CD8 T-cell counts between Seattle and Geneva is 283 × 106/l; this difference is not statistically significant and is probably due to the small sample size. In addition, we cannot eliminate totally a centre effect based on small differences in laboratory methods.
The demographics of the populations differed by centre. There was a higher proportion of injecting drug users in the Swiss cohort. To our knowledge, no study has demonstrated clear differences in clinical features of acute HIV infection between the patients with different risk characteristics. It is unknown if different subtypes of the virus are associated with different clinical presentations. An impact of this factor is unlikely because subtype B is predominant in Europe, North America and Australia . We did not address whether immune responses other than T cell [18,19], macrophage tropism of isolates , CCR5/CXCR4 co-receptors , nutritional , or socioeconomic status  differed by location. Nevertheless, the small differences in symptoms and signs observed in our population suggest that the clinical expression of acute HIV infection is similar in Geneva, Seattle and Sydney, although clinical features of primary HIV infection could be different in developing countries due to nutritional factors, virus subtypes and genetic factors .
A limitation of this study is a bias toward inclusion of patients with acute HIV disease; this did not represent all patients infected by HIV in the areas studied. There is probably a lower proportion of asymptomatic patients among those attending specialized consultations. The low proportion of injecting drug users, heterosexuals, and women restrict the generalization of these results to homosexual or bisexual men. Another caveat is that the follow-up in a RCT could be closer than for an observational cohort. However, the occurrence of acute HIV infection has such important implications that patients in observational studies were investigated in a manner similar to that in the RCT.
HIV antibody tests improved during the time of the study. Potentially laboratory diagnosis of acute HIV infection could be performed sooner for the more recent patients. However this would not have an effect on the time of onset of symptoms.
Plasma HIV RNA during or very shortly after the onset of symptoms was extremely variable and therefore was not analysed. It does not seem to be an accurate prognostic marker immediately after seroconversion despite the power prognostic value after 4–6 months after the onset of acute HIV infection [2,25].
In summary, few differences have been observed in the frequency of clinical features and early T-cell levels in patients acquiring HIV infection in Geneva, Seattle and Sydney. This finding is reassuring for the planning of international observational collaborative studies. However, patients enrolled in RCT differed from those enrolled in observational cohort studies as clinical features lasted longer for RCT patients suggesting that patients with more severe disease were recruited preferentially into RCT. This interpretation is also suggested by the outcome of the untreated control group in the RCT of zidovudine , where a surprisingly high number of clinical events indicative of immunosuppression were observed within 24 months of acute HIV infection. This finding suggests that generalization of results from RCT should be carried out with caution.
M. Berrey, T. Schacker, T. Shea, D. Lee of the Primary Infection Clinic (Seattle), B. Anderson, D. Baker, A. Beveridge, M. Bloch, N. Doong, C. Duncombe, R. Finlayson, V. Furner, B. Genn, J. Gold, J. Kidd, R. McFarlane, M. McMurchie, A. McNulty, H. Michelmore, A. Pethebridge, D. Quan, M. Robertson (Sydney); J. Kaldor, and the late B. Tindall at the National Centre in HIV Epidemiology and Clinical Research (Sydney); J.F. Balavoine, A. Christina, C. Junet, V. Sendersky, J. Wintsch (Geneva); S. Kinloch-de Loës (London) and the investigators who participated in the controlled trial of zidovudine in primary HIV infection (see authors and appendix of reference ), R. Gaudet and S. Robillard (University of Montreal) for technical assistance.
The member of the Swiss HIV Cohort Study are M. Bateguay (Co-chair of the Scientific Board), E. Bernasconi, Ph. Bürgisser, M. Egger, P. Erb (Chairman of the Group ‘Laboratories', W. Fierz, M. Flepp (Chairman of the Group ‘Clinics'), P. Francioli (President of the SHCS, Centre Hospitalier Universitaire Vaudois, CH-1011-Lausanne), H.J. Furrer, P. Grob, B. Hirschel (Chairman of the Scientific Board), L. Kaiser, B. Ledergerber, R. Lüthy, R. Malinverni, L. Matter, M. Opravil, F. Paccaud, G. Pantaleo, L. Perrin, W. Pichler, J-C. Piffaretti, M. Rickenbach, P. Sudre, J. Schupbach, A. Telenti, P. Vernazza.
1. Mussey L, Hughes J, Schacker T, Shea T, Corey L, McElrath MJ. Cytotoxic-T-cell responses, viral load, and disease progression in early human immunodeficieciency virus type 1 infection.
N Engl J Med 1997, 337: 1267 –1274.
2. Schacker TW, Hughes JP, Shea T, Coombs RW, Corey L. Biological and virological characteristics of primary HIV infection.
Ann Intern Med 1998, 128: 613 –620.
3. Pedersen C, Lindhart BO, Jensen BL. et al. Clinical course of primary HIV infection: consequences for subsequent course of infection.
Br Med J 1989, 299: 154 –157.
4. Vanhems P, Lambert J, Cooper DA. et al. Severity and prognosis of acute human immunodeficiency virus type 1 illness: a dose response relationship.
Clin Infect Dis 1998, 26: 323 –329.
5. Jacquez JA, Koopman JS, Simon CP, Longini IM. Role of the primary infection in epidemics of HIV infection in gay cohorts.
J Acquir Immune Defic Syndr 1994, 7: 1169 –1184.
6. Vanhems P, Allard R, Cooper DA. et al. Acute human immunodeficiency virus type 1 disease as a mononucleosis-like illness: is the diagnosis too restrictive?
Clin Infect Dis 1997, 24: 965 –970.
7. Schacker T, Collier AC, Hughes J, Shea T, Corey L. Clinical and epidemiologic features of primary HIV infection.
Ann Intern Med 1996, 125: 257 –264.
8. Ledergeber B, von Overbeck J, Egger M, Luthy R. The Swiss HIV cohort study: rationale, organization and selected baseline characteristics.
Sozial- und Praventivmedizin 1994, 39: 387 –394.
9. Kinloch-de Loës S, Hirschel BJ, Hoen B. et al. A controlled trial of zidovudine in primary human immunodeficiency virus infection.
N Engl J Med 1995, 333: 408 –413.
10. Tindall B, Barker S, Donovan B. et al. Characterization of the acute clinical illness associated with human immunodeficiency virus infection.
Arch Intern Med 1988, 148: 945 –949.
11. Centers for Disease Control. Guidelines for the performance of CD4+ T-cell determinations in persons with human immunodeficiency virus infection.MMWR
12. Kagan J, Gelman R, Waxdal M, Kidd P. NIAID Division of AIDS flow cytometry quality assessment program.
Ann N Y Acad Sci, 1993, 677: 50 –52.
13. Laird N, Ware J. Random-effects models for longitudinal data.
Biometrics 1982, 38: 963 –974.
14. Kramer MS, Shapiro SH. Scientific challenges in the application of randomized trials.
JAMA 1984, 52: 2739 –2745.
15. Chalmers TC. The control of bias in clinical trials.
In Clinical Trials. Issue and Approches.
Edited by Shapiro SH, Louis TA. New York, Marcel Dekker, Inc.; 1983:115–127.
16. Zelen M. The randomization and stratification of patients to clinial trials.
J Chron Dis 1974, 27: 365 –375.
17. Subbarao S, Schochetman G. Genetic variability of HIV-1.
AIDS 1996, 10 (suppl A): S13 –S23.
18. Lamhamedi-Cherradi S, Culmann-Penciolelli B, Guy B. et al. Different patterns of HIV-1-specific cytotoxic T-lymphocyte activity after primary infection.
AIDS 1995, 9: 421 –426.
19. Pantaleo G, Delarest JF, Schacker T. et al. The qualitative nature of the primary immune response to HIV infection is a prognosticator of disease progression independant of the initial level of viremia.
Proc Natl Acad Sci USA 1997, 94: 254 –258.
20. Zaitseva M, Blauvelt A, Lee S. et al. Expression and function of CCR5 and CXCR4 on human Langerhans cells and macrophages: implications for HIV primary infection.
Nature Med 1997, 3: 1369 –1375.
21. Michael NL, Chang G, Louie LG. et al. The role of viral phenotype and CCR-5 gene defects in HIV-1 transmission and disease progression.
Nature Med 1997, 3: 338 –340.
22. Baum MK, Shor-Posner G, Lai S. et al. High risk of HIV-related mortality is associated with selenium deficiency.
J Acquir Immune Defic Syndr Hum Retrovirol 1997, 15: 370 –374.
23. Schechter MT, Hogg RS, Aylward B, Craib KJ, Le TN, Montaner JS. Higher socioeconomic status is associated with slower progression of HIV infection independent of access to health care.
J Clin Epidemiol 1994, 47: 59 –67.
24. Bollinger RC, Brookmeyer RS, Mehendale SM. et al. Risk factors and clinical presentation of acute primary HIV infection in India.
JAMA 1997, 278: 2085 –2089.
25. Mellors JW, Kingsley LA, Rinaldo CR. et al. Quantitation of HIV-1 RNA in plasma predicts outcome after seroconversion.
Ann Intern Med 1995, 122: 573 –579.
Two patients from Seattle were enrolled based on the following reports. Patient 1 was HIV-1 seronegative in September 1994 and again in March 1995. The patient developed primary herpes simplex virus infection after a sexual encounter with unknown partner; 18 days after that encounter he presented a flu-like syndrome with fever, myalgias, lymphadenopathy, and fatigue and the first HIV-1 seropositive test was observed 2 months after the onset of symptoms. Patient 2 had a negative anonymous HIV-1 screening test on June 1996, so we were not able to document this result. Since this test, he presented with an illness consistent with a seroconversion syndrome after unprotected oral sex on September 6, 1996. The plasma HIV RNA at time of the first HIV-1 seropositive test on September 20, 1996 was 168 828 copies/mL, followed by 79 653 copies/mL 4 days later.