Secondary Logo

Journal Logo

Epidemiology and Social Science

The Development and Utility of a Clinical Algorithm to Predict Early HIV-1 Infection

Sharghi, Neda MPH*; Bosch, Ronald J PhD; Mayer, Kenneth MD‡§; Essex, Max DVM, PhD*∥; Seage, George R III DSc, MPH#

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: December 1st, 2005 - Volume 40 - Issue 4 - p 472-478
doi: 10.1097/01.qai.0000164246.49098.47


Prompt diagnosis of early and acute HIV-1 infection is an important goal for health care providers as well as for vaccine and microbicide trial organizers, as this is a phase of high viral replication and rapid rise in plasma viremia without a distinguishable antibody response.1-4 Identifying individuals soon after infection with HIV-1 may be important for the clinical management of patients and could also benefit public health by identifying highly infectious individuals engaging in recent high-risk behaviors.5-10 It has been postulated that if health care practitioners provide appropriate and timely antiretroviral therapy, they could alter the natural history of HIV-1 infection; public health practitioners could concurrently provide appropriate risk reduction counseling to minimize secondary HIV-1 transmission.10,11 Development of criteria to identify individuals soon after they become HIV-1 infected is also important for vaccine and microbicide trials. A recent report estimates that, in order for a collaborative and systematic global HIV vaccine enterprise to support the research and development of an effective vaccine, at least 30,000 individuals will need to be annually enrolled into phase 3 efficacy trials.12 The ability to identify, in an economically feasible manner, even a small subset of these trial participants during acute HIV-1 infection would allow scientists to investigate whether the intervention being studied had any impact on early transmission or pathogenesis of HIV-1 infection.

However, although detecting acute HIV-1 infection is important, steps necessary to do so can be challenging. Common symptoms of acute HIV-1 infection are fever, fatigue, nausea, vomiting, rash, sore throat, and headache.13-15 These clinical symptoms have been characterized as flulike, varied in severity, and often transient, lasting approximately 1 to 2 weeks.4,16-21 Because these symptoms are nonspecific and resemble many more mundane infections, patients often ignore them and do not seek medical care.15 When they do, the nonspecific symptoms may confuse health care professionals who may misdiagnose and offer inappropriate treatment.11,15,19,22-24

The goal of this study was to develop an algorithm to selectively screen individuals suspected of having acute HIV-1 infection. Specifically, our aim was to assess whether a scoring system using clinical factors, which are present in 50% to 90% of individuals with acute HIV-1 infection, could predict recent seroconversion to HIV-1 among a prospectively followed cohort of HIV-1-negative individuals at increased risk for infection. If a scoring system of self-reported clinical factors can predict recent seroconversion, then our findings could provide a valuable and inexpensive method for health care practitioners, in the clinical, public health, or prevention trial settings, to identify individuals with early and acute HIV-1 infection. These individuals can subsequently be selectively screened, counseled, and offered participation in early intervention trials.


Study Participants

The Vaccine Preparedness Study (VPS) of the HIV Network for Prevention Trials (HIVNET) was established to gather relevant baseline data necessary for the initiation and implementation of vaccine efficacy trials.25 The VPS was a prospective cohort study enrolling HIV-1-negative individuals at high risk for HIV-1 infection, either as a result of high-risk sexual behavior or injection drug use, from various sites throughout the United States.25 The study recruited a total of 4892 individuals and had a retention rate of 88% for the full 18 months of follow-up.26 HIV-1-uninfected men and women were eligible for enrollment if they were 18 years of age or older and met the definition of “high risk” for HIV-1 infection. HIV-1-negative men were considered “high risk” if they reported anal sex with ≥1 men in the past year or had injected drugs at least once in the previous 6 months. HIV-1-negative women were considered “high risk” if they reported injecting drugs within the previous 6 months; reported a current relationship with a man who was HIV-1 positive, who had sex with other men, who reported injection drug use in the past 5 years, or who was diagnosed with syphilis, gonorrhea, or chlamydia infection; or reported one of the following behaviors within the past year: exchanging sex for money or drugs, “crack” cocaine use, having ≥5 male partners, or had a diagnosis of syphilis, gonorrhea, chancroid, pelvic inflammatory disease, trichomoniasis, or an initial episode of genital herpes.25,27

Once HIV-1-negative status and eligibility were confirmed, a baseline questionnaire was administered to assess demographic information, risk behavior, and health status within the past 6 months. Participants were asked to return every 6 months, for a total of 18 months, for serologic testing, pre- and posttest counseling, and a study interview. If, at any point in the study, the participant was found to be HIV-1 infected, he or she was given appropriate counseling and referred to the Infected Participant Cohort of the HIVNET; he or she would no longer be eligible for follow-up in the VPS.


The outcome of interest was HIV-1 seroconversion, which was diagnosed using HIV enzyme-linked immunosorbent assay (ELISA) and confirmed by Western blotting. Serologic testing was conducted at screening and at 6, 12, and 18 months. Participants could also come for serologic testing between scheduled visits. Because the majority of HIV-1 testing occurred in 6-month intervals, a positive test result does not indicate the day of infection but that infection occurred at some point between the most recent seronegative and seropositive visit.

Every 6 months, clinical symptoms were assessed by providing a list of symptoms and asking participants if they had experienced any of them since their prior study visit. Sexually transmitted infections were assessed by providing a list of common infections and asking the participants whether a medical provider had diagnosed or treated them for any of the infections since their last interview. Clinical factors likely to be highly correlated with one another were combined for analysis; eg, nausea and vomiting were combined as one symptom and chlamydia infection and gonorrhea were analyzed together because of frequent misclassification and coinfection. Participants were also asked whether they believed that, since their last interview, they might have been exposed to HIV, had been hospitalized overnight for any reason, or had any illness that lasted ≥3 days and kept them from doing any of the things that they usually did.

Data Analysis

We modeled the association between self-reported clinical factors and HIV-1 seroconversion using generalized estimating equations, a technique that accounts for the correlative nature of data within individuals in longitudinal studies with repeated measures.28 Data analyses were based on pooling 6-, 12-, and 18-month risk sets. The main parameter of interest in these analyses is the odds of HIV-1 seroconversion (from previous HIV-1 seronegative to current HIV-1 seropositive), given the presence of a clinical factor since the previous visit. Individuals contributed repeated measures until loss to follow-up, seroconversion, or the final 18-month visit. We assumed that loss to follow-up was noninformative with respect to seroconversion.

Univariate analyses compared the odds of HIV-1 seroconversion between those with and those without clinical factors. Clinical factors found to be significant at P ≤ 0.20 in univariate analyses were included as candidate variables for the multivariable model, which was constructed using a stepwise procedure. Confidence intervals are based on robust standard errors, using an independent working correlation structure.28 A scoring system was developed from this multivariate model.

A second scoring system was developed to include variables that did not remain in the final multivariate model but that were deemed clinically relevant. For both scoring systems, individual factors were assigned 1 point and cutoff levels were established, depending on the number of clinical factors present. Sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) were calculated for each cutoff level. No corrections were made for the multiple factors assessed.

All analyses were conducted using SAS version 8.2.


Study Patients

Of the 4892 participants in the HIVNET VPS, 240 participants were excluded from analyses because they had only a baseline visit or a missing HIV status. Our analyses included 4652 study participants, with a total of 13,076 visits. The retention rate for this subset of the VPS was 86.8% for the full 18 months of study follow-up.

Of the study participants, 811 (17.4%) were female and 3841 (82.6%) were male. The majority of participants (39.3%) were between 31 and 40 years of age; 34.5% were between the ages of 18 and 30 years and 26.2% were 41 years of age or older. Most participants (61.3%) classified themselves as white, not of Hispanic origin, whereas 20.8% and 17.9% classified themselves as black, not of Hispanic origin, and Hispanic/other, respectively. Finally, 3120 participants (67.1%) reported that they were men who had sex with men; 721 (15.5%) were male injection drug users and 811 (17.4%) were either women at heterosexual risk or female injection drug users.

Eight-six individuals, 13 women and 73 men, seroconverted during the study, with 35, 29, and 22 individuals seroconverting at the 6-, 12-, and 18-month visits, respectively. Of these, 7 were found to be HIV infected when they came for optional testing between the 6-, 12-, and 18-month visits; the seroconversion time of these participants was allocated to their most recent previous visit. The overall annual seroconversion rate was 1.3 per 100 person-years.

Clinical Factors

Of the 23 self-reported clinical factors assessed (which included symptoms, severity of illness measures, and beliefs about participants' likelihood of being exposed to HIV), 11 were found to be significantly associated with HIV-1 seroconversion at P ≤ 0.05 in univariate analyses (Table 1). The final multivariate model included 4 self-reported clinical factors assessed since the participant's last visit: chlamydia infection, nonspecific urethritis, or gonorrhea (odds ratio [OR], 3.1), fever or drenching night sweats (OR, 2.2), self-reported belief of HIV exposure (OR, 2.1), and any illness lasting ≥3 days (OR, 2.1) (Table 2). A scoring system, as defined in the “Methods” section, then counted the number (0-4) of these key clinical factors.

Unadjusted and Adjusted Odds Ratios for HIV-1 Seroconversion for Self-Reported Clinical Factors Across Three 6-Monthly Visits, HIVNET Vaccine Preparedness Study, United States, 1995-1997
Multivariate Model of 4 Key Clinical Factors, Assessed Since Last Study Visit*

A second scoring system was also developed. As shown in Table 1, univariate analyses identified 2 self-reported clinical factors-recent chlamydia infection or gonorrhea, and recent fever or drenching sweats-to be strongly associated with HIV-1 seroconversion, with a P value <0.0001 and OR >3.5. After adjusting for these 2 factors, we chose 9 additional factors that were deemed clinically relevant and associated with HIV-1 seroconversion, either at a significance level <0.2 or an OR >1.5 (Table 1). These 9 additional clinical factors were swollen or painful glands or lymph nodes, sore throat, sores in the mouth, sore muscles or joints, headache, any illness lasting ≥3 days, belief of HIV exposure, any type of hepatitis or liver infection, and the presence of genital sores, all assessed since the last visit.

Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value

Sensitivity, specificity, PPV, and NPV were calculated for both the 4- and 11-clinical factor scoring systems (Tables 3 and 4). In the 4-factor scoring system, 32 seroconverters (37%) reported ≥2 of the 4 clinical factors; sensitivity and specificity for this cutoff level were 37.2% and 88.5%, respectively, and the PPV and NPV were 2.1% and 99.5%, respectively. Of the 86 seroconverters, 33 (38%) reported none of these 4 key clinical factors (Table 5). In the 11-factor scoring system, 24 of the 86 seroconverters reported none of the clinical factors (30%). Forty-three seroconverters (50%) reported ≥2 of the 11 clinical factors; sensitivity, specificity, PPV, and NPV for this cutoff level were 50.0%, 77.3%, 1.4%, and 99.6%, respectively. Sixteen seroconverters reported ≥5 clinical factors; sensitivity, specificity, PPV, and NPV were 18.6%, 96.9%, 3.9%, and 99.4%, respectively. In each scoring system, as the number of clinical factors in each cutoff level increased, sensitivity decreased while specificity increased. Also, the rate of seroconversion was directly associated with the reported number of clinical factors in each cutoff level (Tables 3-5).

Sensitivity,Specificity, and Positive and Negative Predictive Values for the Presence of Greater Than or Equal to 1, 2, and 3 of the Key Self-Reported Clinical Factors Assessed Since last Study Visit*
Sensitivity, Specificity, and Positive and Negative Predictive Values for Cutoff Levels of the 11 Self-Reported Clinical Factors Assessed Since last Study Visit*
Four Key Clinical Factor Score From Final Univariate Model*


Using recent HIV-1 seroconversion as a marker for early and acute infection, the goals of this study were to develop 2 scoring systems of clinical factors and to assess whether they might aid in selectively screening individuals suspected of presenting with acute HIV-1 infection. The standard ELISA is ineffective in identifying acute HIV infection because the antibodies it detects are not produced until approximately 3 to 4 weeks after HIV infection.29,30 RNA testing is better suited because HIV RNA is detectable in peripheral blood as early as 9 days after infection.29 However, because RNA testing is expensive, it is not considered a cost-effective method with which to routinely test for acute HIV infection, especially among large groups of individuals, such as those engaging in vaccine or microbicide trials. Thus, using a clinical algorithm to screen individuals suspected of being in the acute phase of HIV infection, and then selectively administering HIV RNA testing to them, may provide a low-cost solution to improving the identification of individuals in this early stage of infection.

In our first scoring system, which consisted of 4 clinical factors, the cutoff level of reporting ≥1 of the factors has a sensitivity and specificity of 61.6% and 65.2%, respectively, but a very low PPV. Hypothetically, if this cutoff were chosen to screen for acute infection among this VPS cohort, 4571 RNA tests would have to be administered, of which 53 or 1.2% would result in a positive finding for acute HIV infection. A decrease in sensitivity is subsequently observed as the number of clinical factors increases in each cutoff. This decrease in sensitivity, however, needs to be balanced against the cost-effectiveness gained by the simultaneous increases in specificity and PPV. For example, even though testing individuals presenting with ≥3 of the 4 clinical factors will result in a smaller portion of correctly identified acute infections, fewer RNA tests will need to be administered, ie, 363, of which 16 or 4.4% would positively identify acute infection. In our expanded scoring system of 11 clinical factors, the PPV for presenting with ≥7 of the clinical factors is 8.7% with a sensitivity of only 7.0% and a specificity of 99.5%. Again, even though this cutoff would capture only a very small percentage of acutely infected individuals, administering only 69 RNA tests would still correctly identify 6 acutely infected individuals. Perhaps, given the financial constraints of HIV RNA testing and the importance of identifying acutely infected individuals, the solution may be to use a scoring system of clinical factors with emphasis on high specificity and high PPV rather than high sensitivity.

The need to identify HIV-1-infected individuals as soon as they become infected has recently reemerged as an important public health priority, not only because of the public health impact of missing those in the acute phase of infection but also because a more economically feasible mechanism to test for acute infection is being proposed.11,31 In 2002, after data from a pilot study showed that some HIV-1-infected individuals were being missed from antibody-based HIV-1 testing, North Carolina became the first state to include HIV RNA reverse transcriptase polymerase chain reaction tests on pooled specimens as part of their screening program.11,32 This specimen pooling/HIV RNA testing algorithm has been suggested as a means of cost-effectively testing for acute HIV infection, especially in settings where testing volume is large and prevalence of acute HIV infection is low.11,29,30 A clinical screening system combined with a specimen pooling/HIV RNA testing algorithm may have utility in identifying acute HIV infection in vaccine or microbicide trials.

Using a specimen pooling/HIV RNA technique, a recent cross-sectional study conducted among male sexually transmitted disease (STD) and dermatology clinic patients in Malawi found a 1.8% prevalence of acute HIV infection.29 These authors reported a low prevalence of clinical symptoms typical of acute infection and did not find a significant association between these symptoms and acute HIV infection. However, the authors speculate that the low prevalence of symptoms is likely due to the study population being predominately clinic patients seeking STD care before the onset of symptomatic acute HIV infection.29 Another recent study, this one prospective and conducted among female commercial sex workers in Kenya, compared the prevalence of clinical signs and symptoms in seroconverters and nonseroconverters.33 Univariate analyses identified 11 clinical signs and symptoms to be significantly associated with HIV-1 seroconversion. Sensitivity and specificity analyses were conducted on a 6-symptom score of these clinical signs and symptoms. The sensitivity, specificity, and PPV for women with ≥2 of these 6 clinical symptoms were 51.3%, 83%, and 4%, respectively. The same cutoff level in our 11-factor scoring system had a sensitivity, specificity, and PPV of 50.0%, 77.3%, and 1.4%, respectively. As with our findings, this study also reported marked decreases in sensitivity and increases in specificity as the number of clinical symptoms in the cutoff increased.33

The VPS was designed to identify HIV-1 seroconversion at 6-month intervals to inform the development and analyses of future prevention trials and not to carefully identify acute infection. Our study is thus limited in that appropriate laboratory tests for acute HIV-1 infection were not conducted. As a result, we were unable to identify acute HIV-1 infection or to claim that clinical symptoms occurred during acute infection, only that they occurred prior to seroconversion. Another limitation of our study is that we were unable to validate our scoring system in a separate validation population. As the Malawi study was limited to male STD and dermatology clinic patients and the Kenya study to female commercial sex workers, we are also limited in our ability to directly compare the findings of those studies with ours, which included both male and female participants of a vaccine preparedness study. Finally, our sample size also limited our ability to determine whether the accuracy of symptom reporting differed by risk group and, consequently, whether different clinical algorithms for different risk groups are warranted.

Numerous strengths characterize our study. In addition to the large sample size and large number of female participants involved relative to other US-based cohorts, the prospective design of our study allowed for assessment of clinical symptoms before diagnosis of HIV-1 seroconversion, thus minimizing the possibility of recall bias. Importantly, unlike other studies, our cohort was not referral based and subsequently was not biased toward the enrollment of symptomatic patients.19,34 Also, the inclusion of different high-risk groups in our cohort allows generalizability for the different routes of transmission, both parenteral and sexual.35 Finally, even though our cohort included high-risk individuals, loss to follow-up was minimal.

Our clinical scoring system may have utility for detection of early and acute infection in vaccine and microbicide trial settings, where trial participants will already be undergoing routine serologic testing and counseling for HIV-1. Such a scoring system may facilitate diagnosis of acute HIV-1 infection by allowing trial organizers to educate trial participants to return for interim visits for testing, if they develop specific clinical symptoms or factors. By identifying these trial participants, vaccine trial organizers may optimize their ability to learn whether their intervention had any impact on early virologic events associated with acute infection, on secondary transmission, and on the pathogenesis of HIV-1 infection. Participants who are found to be HIV-1 infected can also be counseled and offered participation in early therapeutic intervention trials. In conclusion, more frequent study follow-up visits and a scoring system that incorporates clinical factors may be useful for vaccine or microbicide trials that are designed to detect early or acute HIV-1 infection.


The following institutions and persons associated with HIVNET participated in the Vaccine Preparedness Study Protocol Team: Domestic Master Contractor, Abt Associates, Inc.: G. Seage, M. Gross; Statistical and Clinical Coordinating Center-Fred Hutchinson Cancer Research Center and University of Washington: T. Fleming, S. Self; Central Laboratory-Viral and Rickettsial Disease Laboratory, California Department of Health Services: H. Sheppard, M. Ascher; Repository Contractor-Biomedical Research, Inc.: J. Leff; Denver Department of Public Health: F. N. Judson; Fenway Community Health Center: K. Mayer; Howard Brown Health Center: D. McKirnan; New York Blood Center: C. Stevens, B. Koblin; New York University School of Medicine: M. Marmor, S. Titus; and Beth Israel Medical Center, New York: D. Des Jarlais; San Francisco Department of Public Health: E. Stone, S. Buchbinder; University of Pennsylvania and the Philadelphia Veterans Affairs Medical Center: D. Metzger, G. Woody; University of Washington: C. Celum. National Institute of Allergy and Infectious Diseases: R. Hoff, M. McCauley, and Z. Rosenberg.

The authors thank the VPS participants, the efforts of VPS site staff, and the contributions of the HIVNET Community Advisory Board Members. The authors also acknowledge Sidney Atwood of the Division of Social Medicine and Health Inequalities at Brigham and Women's Hospital, and the reviewers for valuable comments that have substantially improved this work.


1. Hecht FM, Busch MP, Rawal B, et al. Use of laboratory tests and clinical symptoms for identification of primary HIV infection. AIDS. 2002;16:1119-1129.
2. Daar ES, Moudgil T, Meyer RD, et al. Transient high levels of viremia in patients with primary human immunodeficiency virus type 1 infection. N Engl J Med. 1991;324:961-964.
3. Janssen RS, Satten GA, Stramer SL, et al. New testing strategy to detect early HIV-1 infection for use in incidence estimates and for clinical and prevention purposes. JAMA. 1998;280:42-48.
4. Kahn JO, Walker BD. Acute human immunodeficiency virus type 1 infection. N Engl J Med. 1998;339:33-39.
5. Lindback S, Brostrom C, Karlsson A, et al. Does symptomatic primary HIV-1 infection accelerate progression to CDC stage IV disease, CD4 count below 200 × 10(6)/1, AIDS, and death from AIDS? BMJ. 1994;309:1535-1537.
6. Keet IP, Krijnen P, Koot M, et al. Predictors of rapid progression to AIDS in HIV-1 seroconverters. AIDS. 1993;7:51-57.
7. Dorrucci M, Rezza G, Vlahov D, et al. Clinical characteristics and prognostic value of acute retroviral syndrome among injecting drug users. Italian Seroconversion Study. AIDS. 1995;9:597-604.
8. 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.
9. Malhotra U, Berrey MM, Huang Y, et al. Effect of combination antiretroviral therapy on T-cell immunity in acute human immunodeficiency virus type 1 infection. J Infect Dis. 2000;181:121-131.
10. Rosenberg ES, Altfeld M, Poon SH, et al. Immune control of HIV-1 after early treatment of acute infection. Nature. 2000;407:523-526.
11. Pilcher CD, McPherson JT, Leone PA, et al. Real-time, universal screening for acute HIV infection in a routine HIV counseling and testing population. JAMA. 2002;288:216-221.
12. Klausner RD, Fauci AS, Corey L, et al. Medicine: the need for a global HIV vaccine enterprise [see comment]. Science. 2003;300:2036-2039.
13. 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.
14. Taiwo BO, Hicks CB. Primary human immunodeficiency virus. South Med J. 2002;95:1312-1317.
15. Kinloch-de Loes S, de Saussure P, Saurat JH, et al. Symptomatic primary infection due to human immunodeficiency virus type 1: review of 31 cases. Clin Infect Dis. 1993;17:59-65.
16. Sinicco A, Palestro G, Caramello P, et al. Acute HIV-1 infection: clinical and biological study of 12 patients. J Acquir Immune Defic Syndr. 1990;3:260-265.
17. Daar ES, Bai J, Hausner MA, et al. Acute HIV syndrome after discontinuation of antiretroviral therapy in a patient treated before seroconversion. Ann Intern Med. 1998;128:827-829.
18. 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.
19. Schacker T, Collier AC, Hughes J, et al. Clinical and epidemiologic features of primary HIV infection. Ann Intern Med. 1996;125:257-264.
20. Pedersen C, Lindhardt BO, Jensen BL, et al. Clinical course of primary HIV infection: consequences for subsequent course of infection. BMJ. 1989;299:154-157.
21. Quinn TC. Acute primary HIV infection. JAMA. 1997;278:58-62.
22. Rosenberg ES, Caliendo AM, Walker BD. Acute HIV infection among patients tested for mononucleosis. N Engl J Med. 1999;340:969.
23. Walensky RP, Rosenberg ES, Ferraro MJ, et al. Investigation of primary human immunodeficiency virus infection in patients who test positive for heterophile antibody. Clin Infect Dis. 2001;33:570-572.
24. Vidrih JA, Walensky RP, Sax PE, et al. Positive Epstein-Barr virus heterophile antibody tests in patients with primary human immunodeficiency virus infection. Am J Med. 2001;111:192-194.
25. Koblin BA, Heagerty P, Sheon A, et al. Readiness of high-risk populations in the HIV Network for Prevention Trials to participate in HIV vaccine efficacy trials in the United States. AIDS. 1998;12:785-793.
26. Seage GR 3rd, Holte SE, Metzger D, et al. Are US populations appropriate for trials of human immunodeficiency virus vaccine? The HIVNET Vaccine Preparedness Study. Am J Epidemiol. 2001;153:619-627.
27. National Institute of Allergy and Infectious Diseases, NIH, Department of Health and Human Services. Vaccine Preparedness Study, Uninfected Participant Cohort. FHCRC-HIVNET Statistical Center Version. March 2000.
28. Diggle PJ, Heagerty P, Liang K-Y, Zeger S. Analysis of Longitudinal Data, 2nd ed. New York: Oxford University Press; 2002.
29. Pilcher CD, Price MA, Hoffman IF, et al. Frequent detection of acute primary HIV infection in men in Malawi. AIDS. 2004;18:517-524.
30. Pilcher CD, Eron JJ Jr, Galvin S, et al. Acute HIV revisited: new opportunities for treatment and prevention. J Clin Invest. 2004;113:937-945.
31. Anonymous. Advancing HIV prevention: new strategies for a changing epidemic-United States, 2003. MMWR Morb Mortal Wkly Rep. 2003;52:329-332.
32. Voelker R. Detecting acute HIV infections feasible, North Carolina program demonstrates. JAMA. 2003;289:2633-2634.
33. Lavreys L, Thompson ML, Martin HL Jr, et al. Primary human immunodeficiency virus type 1 infection: clinical manifestations among women in Mombasa, Kenya. Clin Infect Dis. 2000;30:486-490.
34. Celum CL, Buchbinder SP, Donnell D, et al. Early human immunodeficiency virus (HIV) infection in the HIV Network for Prevention Trials Vaccine Preparedness Cohort: risk behaviors, symptoms, and early plasma and genital tract virus load. J Infect Dis. 2001;183:23-35.
35. Vanhems P, Routy JP, Hirschel B, et al. Clinical features of acute retroviral syndrome differ by route of infection but not by gender and age. J Acquir Immune Defic Syndr. 2002;31:318-321.

early and acute HIV-1 infection; clinical factors; HIV Network for Prevention Trials (HIVNET); vaccine preparedness study; vaccine trials; microbicide trials.

© 2005 Lippincott Williams & Wilkins, Inc.