Primary infection with HIV-1 (PHI) may be associated with a clinical illness, which appears after an incubation period that averages about 14 days [1,2]. The incidence of this acute illness is not precisely known but several studies suggest that most infected people develop this syndrome [2,3]. The illness typically lasts for a few days to a week and is characterized by a sudden onset of fever, sore throat, rash, and occasionally also oral or genital aphtous ulcerations . About 1 week after the onset of symptoms, the lymph glands enlarge, CD4 cell counts decrease, CD8 cell counts increase and HIV-specific antibodies appear [1,4–6].
During PHI, extremely high levels of viremia have been observed, followed by a rapid fall and thereafter relatively stable levels of viremia [7–10]. As in many other virus infections, the virus-specific CD8 cytotoxic T lymphocyte response has been suggested to play a key role in the early viral clearance [11,12]. The importance of the cell-mediated immune response to the early virus containment was strongly supported by a recent study showing that virus replication is not controlled in monkeys depleted of CD8 lymphocytes during primary SIV infection . However, other factors, such as neutralization by HIV-specific antibodies [14,15] and limitations in the number of activated CD4 target cells , may also play a role in containing viral replication.
Progress in understanding PHI has been hindered by limited data on very early infection. Although plasma HIV-1 RNA levels have been studied in patients during early HIV infection [17–20], few patients have been sampled frequently enough during the first weeks of infection to provide detailed information on the kinetics of viremia during PHI.
In this study, a systematic search for patients with PHI was conducted in Sweden and suspected cases were immediately examined by tests for HIV-1-specific antibodies and antigens. PHI was usually confirmed or excluded within 1 day. Sequential plasma samples were collected and HIV-1 RNA levels were determined for patients without treatment and those receiving highly active antiretroviral therapy (HAART); most samples were obtained within 3 weeks from onset of symptoms. The rate of viral increase and the subsequent decline was quantified using mathematical models to explain the different phases of viremia in PHI and the declines in HIV-1 RNA following antiretroviral treatment.
Material and methods
Patients presenting with PHI have been offered antiretroviral treatment according to clinical practice in Sweden at the time of diagnosis. Apart from some early studies of zidovudine monotherapy, no antiretroviral treatment was given at this point until 1995; in 1995–1996, a combination of two nucleoside analog reverse transcriptase inhibitors (NRTI) were used; from 1996–1998, a combination of two NRTI and one protease inhibitor (PI) were used; while from 1998 on, quadruple regimens were used. The present study included all consecutive subjects with a diagnosis of primary HIV-1 infection in 1994 or later who were either not given antiretroviral treatment or given a triple combination that included a PI. All drugs have been administered at conventional doses: zidovudine 250 mg twice daily, lamivudine150 mg twice daily, indinavir 800 mg three times a day, ritonavir 600 mg twice daily and nelfinavir 750 mg three times a day.
Virological and immunological analysis
Laboratory confirmation of PHI was demonstrated by a positive test for HIV p24 antigen, using HIVAG-1 monoclonal antibodies (Abbott Diagnostics Division, Germany) in combination with the absence of HIV antibodies, as determined by Recombinant IMx HIV-1/HIV-2 3rd Generation MEIA (Abbott Diagnostics), Enzygnost Anti-HIV 1/2 (Behringwerke AG, Germany), and Diagnostic Biotechnology HIV-1 Western blot (Genelabs, Geneva, Switzerland). T cell subsets were enumerated using standard procedures.
HIV-1 subtypes were determined through phylogenetic analysis of nucleotide sequences using the PHYLIP software package . Altogether 32 plasma samples (two to six samples from each patient) were collected during the first 2 months after onset of illness from eight patients not receiving treatment during PHI. The variability of the virus populations was determined by direct sequencing in each case and by cloning of the V3 region in three samples obtained from one patient. A detailed description of methods employed and results obtained has been presented elsewhere [21,22].
Sequential plasma samples for HIV-1 RNA measurement were obtained daily, if possible, during the first 2 weeks and thereafter at intervals of 1–3 months. Samples were stored in portions at −70°C until testing For three patients, samples drawn shortly before onset of symptoms were available for analysis. Plasma HIV-1 RNA levels were determined using the Amplicor HIV Monitor test (Roche Molecular Systems, Somerville, New Jersey, USA) and negative samples were reanalyzed using the Roche UltraSensitive test (Roche Molecular Systems) with a detection limit of approximately 50 copies/ml.
For comparison of levels of plasma HIV-1 RNA, CD4 and CD8 T lymphocytes at specified timepoints during follow-up, actual values or values estimated by linear interpolation following logarithmic transformation were used. The correlation between early peak HIV-1 RNA levels and subsequent HIV-1 RNA levels as well as CD4 and CD8 cell counts were calculated using the Spearman rank correlation test. A Wilcoxon rank sum test for unpaired samples was employed to compare viral loads, CD4 and CD8 cell counts in treated and untreated patients.
In patients who do not take therapy, viral load typically increases exponentially, peaks, and then decline in two phases: first rapidly and then slowly, until the viral load levels off at a rough steady state. To quantify these patterns, non-linear least squares regression was used to fit mathematical models to patient HIV-1 RNA data. The following equation was used to model the two-phase decline from the peak viremia:EQUATION where V1 is the viral load when virus begins to decrease, t1 is the time that viral load begins to decrease, r2 is the rapid decay rate, r3 is the slow decay rate, a is the fraction of virus associated with rapid decay, b is the fraction of virus associated with slow decay, and (1 −a −b) V1 is the predicted steady-state viral load, V4 Since this model was only applied to patients who did not receive treatment, data from patients who eventually received treatment was censored at the time that treatment began. For patients for whom a steady-state viral load could not be determined, HIV-1 RNA data was fitted to a somewhat simpler model:EQUATION
For patients for whom neither the steady state nor the second-phase decay could be estimated, the data were fitted to an even simpler exponential decline model:EQUATION
For a few patients, viral load measurements were obtained prior to the peak viremia. An exponential growth model was used to estimate the rate of increase of virus prior to the peak viremia:EQUATION where V0 is the first viral load measurement prior to the peak viremia, t0 is the time of the first viral load measurement, and r1 is the rate of increase of virus during the increasing phase prior to the peak viremia.
By fitting these models to HIV-1 RNA data, numerical estimates could be obtained for r1, r2, r3, a and b. From a and b three parameters could be calculated:V4; the transition point between rapid and slow decline, t2 =t1 + ln[b/ a]/(r2 −r3); and the transition point between slow decline and the steady-state, t3 =t1 + ln[b/(1 −a − b)]/ r3. These transition points correspond to the time when 50% of virus comes from the former and 50% from the latter component of the model; i.e., when the lines for phases illustrated in Fig. 3 (below) intersect. The viral loads at these transition points, V2 and V3, are defined by putting t2and t3, respectively, into equations (1) or (2), while the peak viral load, V1, and the time at which the peak viremia occurs, t1, are defined by the highest viral load measurement. Because there are a limited number of data points before and during the peak, estimates for r1, t1, and V1 should be regarded as minimal values.
Fits to simulated data
To test the ability to estimate parameters, non-linear regression was used to fit equation (1) to simulated primary infection dynamics with known parameter values and simulated noise. Declines were simulated using equation (1) with r2 = 0.4/day, r3 = 0.03/day, a = 0.96, and b = 0.03. These values for a and b correspond to a transition time between slow and fast decline, t2, of 9.4 days after the peak and a transition between slow decline and steady state, t3, of 37 days after the peak (see equations for t2 and t3 above). To span the range of sampling schemes, the sampling schedules were used for the patient with the most samples (11) and the one with the fewest samples (seven) after the peak (patients E and F, respectively). To simulate measurement errors typical for HIV-1 RNA samples, a normal random variable with a mean of zero and a standard deviation of 35% was added to the natural logarithm of each simulated data point. (All fitting was done to log transformed values to preserve homogeneity of variance.) Each sampling scheme was simulated 50 times to gather statistics on the resulting estimates. In a few cases non-linear least squares failed to converge on the correct region of parameter space, yielding estimates (for example, r2 = 10 000/day) that were obviously incorrect. Any such estimates were excluded from the statistical analysis, as is standard practice when fitting to actual patient data.
The mean and standard deviations for the four principal parameters, after exclusion of pathological fits, for 50 simulated noisy data sets with 11 samples were:r2 = 0.42 ± 0.13/day, r3 = 0.029 ± 0.023/day, a = 0.963 ± 0.019, and b = 0.028 ± 0.018, while means and standard deviations for the 50 simulated noisy data sets with seven samples were:r2 = 0.40 ± 0.12/day, r3 = 0.031 ± 0.024/day, a = 0.962 ± 0.030, and b = 0.030 ± 0.028. The fact that none of these means is statistically different from their underlying values (t-tests not significant for all eight estimates) indicates that any bias that exists for r2, r3, a, b is small. This simulation exercise did reveal one important bias: though estimates for t2 were reasonably accurate, estimates for t3 proved to be very sensitive to variation in a and b and occasionally very high estimates for t3 were obtained, causing the average value for t3 to lie above its underlying value (11-sample simulation:t3 = 52 ± 42 days; 7-sample simulations:t3 = 84 ± 148 days). However, the median for t3, being less sensitive to occasional extreme values, came reasonably close to its underlying value in both cases (11-sample simulation:t3 = 37 days; 7-sample simulation:t3 = 34 days). In the sections that follow, therefore, median values for t3 are emphasized instead of averages.
Seventeen patients, fifteen male and two female, presented with an acute illness suggestive of PHI. Nine were exposed to HIV through homosexual intercourse, seven through heterosexual intercourse, and one through sharing injection instruments with an HIV-positive drug addict. The incubation period from HIV exposure to onset of illness in five patients with only one possible exposure to HIV was 13 days in one, 14 days in three, and 16 days in one patient. The incubation period could not be determined in the other patients since each had had several possible exposures, including one exposure about 14 days before onset of symptoms. Fourteen patients had been exposed to HIV in Sweden, whereas three were exposed during visits to African or Asian countries. Thirteen patients were infected with HIV-1 subtype B, two with subtype C and two with subtype E.
The first serum sample available for testing from each patient, drawn at a median of 5 (range 0–17) days after onset of symptoms, was positive for HIV-1 p24 antigen but negative or indeterminate (not displaying antibodies against two or more env bands) for HIV antibodies by Western blot.
Eight consecutive patients (A–H) presented with PHI in 1994 and did not receive antiretroviral treatment during the acute infection, whereas nine consecutive patients (L, O–V), who presented between 1996 and April 1998, were given a combination therapy regimen that included zidovudine, lamivudine and one PI (seven received indinavir, one ritonavir, and one nelfinavir) beginning at a median of 7 days (range 0–21) following onset of PHI symptoms.
Kinetics of HIV-1 viremia in patients not receiving treatment
A sample from patient U from 7 days before onset of symptoms tested negative, but samples from patients V and P from 6 days and 2 days, respectively, before the onset of symptoms contained 700 copies/ml and 1 × 106 copies/ml, respectively (Fig. 1). In all patients with adequate data available, peak viremia was detected during the first 11 days after onset of symptoms. HIV-1 RNA declined rapidly during the first 2 weeks after the peak viremia. During the following weeks, HIV-1 RNA continued to decrease, but at a slower rate, eventually reaching the steady-state levels shown in Fig. 2.
The data were fitted to the mathematical models described above (see Fig. 3 for a schematic) for nine patients (Table 1). The two-phase models presented above could be fitted to data from five of the eight patients studied (B, C, E, F and G). For two of the patients (A and H), the two-phase model could not be fitted to the data because of a limited number of samples. In patient D, the two-phase models could not easily be fitted to the data because viral load did not decline monotonically after the peak viremia. After falling between days 5 and 10, viral load increased slightly between days 14 and 19 and remained at a roughly constant level for another 3 weeks before falling between days 42 and 120. Viral load also failed to reach a clear steady state in patient D, with viral load rebounding once again between days 120 and 356. Patient D, therefore, was an interesting exception to the general pattern exhibited in other patients. Viral load also rebounded after 100 days in patients C and F, though in the case of patient F the rate of rebound was slow enough to justify fitting with equation (1).
The first phase could be measured in three patients (C, U, and V). In these patients, viral loads increased rapidly (mean rate of increase r1 = 0.74 ± 0.04/day) until virus peaked at a mean of 7 days after onset of symptoms at an average of 2 × 106 copies/ml plasma. The second phase, which could be measured in all of the non-treated patients, was characterized by rapid clearance (r2 = 0.42 ± 0.20/day), resulting in an average viral density of 0.125 × 106 copies/ml at a median of 21 days after onset of symptoms. During the third phase (quantified for patients B, C, E, F, and G), HIV-1 RNA levels still declined but at a much slower rate (r3 =0.030 ± 0.025/day). The transition between the third phase of slowly declining virus and steady state occurred at a median of 41 days after the onset of symptoms with steady-state viral load approximately 25 000 copies/ml plasma.
Comparison between HIV-RNA decline and viral variability
The samples examined for variations in virus population by direct sequencing were collected at a frequency enabling comparison with HIV-1 RNA levels during PHI in eight patients not receiving treatment . A change of the major sequence was detected during the first month in patients D and F, and during the second month in patient C. In patients B, E, G, and H, no changes of the major sequence were detected during the observation period. The timepoint for transition between phases of rapid and slow decline of RNA levels (t2) was determined in six of these patients (Fig. 3). The transition occurred at a mean of 14 days (range 10–19) in patients C, D, and F, in whom changes of the major sequence were detected during the first 2 months. By contrast, the transition occurred at a mean of 33 days (range 21–54) in patients B, E, and G, in whom similar changes were not recorded.
In patient D, these results were confirmed by cloning of the V3 region derived from viral RNA obtained at day 5 (20 clones), day 8 (20 clones), and day 12 (18 clones). The major sequence detected by direct sequencing at day 28 was not the same as the major sequence detected at days 5, 8, or 12. This sequence, however, was present in 5% of clones from day 5, in 0% of clones from day 8, and in 28% of clones from day 12. As mentioned above, we noted that viral load first increased and then levelled off beginning around day 14 in patient D. The viral load perturbations seen in patient D, therefore, seemed to be related in time to the appearance of a new major variant.
Correlation between recorded peak HIV-RNA level and subsequent HIV levels and T cell subset counts
Peak values of HIV-1 RNA recorded during PHI from each patient were compared with subsequent values at intervals of 50 days, from day 50 to day 600. The correlation coefficients (r) ranged between 0.73 and 1.0 (P < 0.05 for all correlations), thus showing a significant correlation between early viral load, as determined by peak level measured during PHI, and steady-state levels. Peak HIV-1 RNA levels were not correlated, however, with subsequent CD4 and CD8 cell counts (data not shown).
Comparison between treated and non-treated patients
The early peak HIV-1 RNA levels in patients receiving antiretroviral treatment did not differ from those recorded in untreated patients. At 4 weeks after the onset of primary HIV infection, however, treated patients displayed a significantly lower mean HIV-1 RNA level than untreated patients: 2500 copies/ml [95% confidence interval (CI) 450–13 800] compared with 103 000 copies/ml (95% CI 64 500–162 000;P < 0.01).
Antiretroviral treatment was started within 10 days following onset of symptoms (days 0, 3, 4, 6, 7 and 10, respectively) in six patients (Fig. 4) and HIV-1 RNA declined rapidly to < 50 copies/ml; all samples drawn day 69 or later from these patients tested negative. Three patients started treatment later (days 13, 19 and 21, respectively). Because of non-compliance and concomitant intravenous drug abuse, treatment in one patient was discontinued after 6 months, when HIV-1 RNA had declined to undetectable levels in the other two patients.
Patients receiving treatment developed milder CD8 lymphocytosis during the acute phase than patients without treatment (Fig. 5). These differences persisted throughout the study period from week 4 through 12 months of follow-up (P < 0.05). The CD4 cell counts did not differ significantly between the two groups of patients.
This study offers new insights into the time course and kinetics of viremia following onset of symptoms in PHI. The pre-illness pattern of viral dissemination has been clearly established in animal models [23–25] : genitally inoculated virus infects T cells, macrophages, and dendritic cell, followed by a stepwise progression of the infection from regional lymph nodes to distant lymphoid tissues, presumably by migration of infected cells, followed by subsequent spread to the systemic circulation about 1 week after inoculation. It seems reasonable to expect a similar pattern of HIV infection in humans, considering that our results indicate that viremia appears during the week preceding onset of symptoms and that the incubation period is about 2 weeks in most cases.
The viral dynamics during ‘typical’ PHI is characterized by the following sequence of events. (i) In the previremic phase, HIV disseminates and replicates within the lymphoid system. (ii) The first viremic phase of increasing viral density starts about 1 week after infection when the HIV virus appears in the blood at rapidly increasing levels, initially without clinical symptoms, which usually appear 2 weeks after infection. (iii) The second viremic phase of rapid decay begins at about 3 weeks after infection; virus declines from the peak viremia to about 5–10% of the peak level over about 2 weeks. Clinically this phase is characterized by lymph node enlargement, the appearance of IgG-antibodies, and CD8 lymphocytosis. (iv) The third viremic phase of slow decay occurs from about 5 weeks after infection; during this stage clearance continues but at only 5–10% of the rate seen during the second viremic stage of rapid decline. (v) The fourth viremic phase begins approximately 2 months after infection; during this period, HIV-1 RNA levels approach an approximate steady state between production and clearance but with great individual variation, both in time until, and in level of, the steady state.
Two patients that did not show this typical pattern in our study, patients C and D, both showed evidence of viral sequence change. In patient D, a brief upspike in viral load between days 10 and 20 was temporally associated with a genetic change in env that became evident by direct sequencing on day 12. In patient C, however, the change in env sequence (on day 56) occurred well before the viral load rebound between days 108 and 236, indicating that these events were not temporally associated. It is conceivable, however, that genetic changes in other loci in the virus in patient C were responsible for this rebound.
The pattern of rapid increase and decrease in HIV-1 RNA during the first 5 weeks of PHI is similar to that which is found in most acute infections. In the ‘typical’ HIV infection, however, virus starts to decline more slowly, beginning around week five, eventually hitting an approximate steady state that persists for months or years. This pattern could be quantified in our study as a result of the large number of frequently collected samples. The biphasic decay we observed in untreated patients is remarkably similar to that which occurs during initiation of potent anti-HIV treatment, though it is not clear that it results from the same mechanism. During treatment, the second phase of slower decay has been interpreted to be the decay of long-lived infected cells , which are now thought to consist mainly of resting T cells . However, if most of the virus was produced by long-lived cells during the phase of slow decay in PHI, then initiation of antiretroviral treatment would not be expected to extend the rapid decay further to approximately a 2.0 log10 copies/ml reduction in HIV-1 RNA, which was observed in our patients who received early treatment.
The lower rate of viral clearance that occurs around week 5 could indicate a rebound in target cell populations, reductions in the intensity immunological responses against HIV as antigen levels decrease, or the outgrowth of escape mutants or minor transmitted variants. The last possibility is supported by our finding, although examined in a limited number of patients, of lower clearance rates in patients in which there is an early appearance of a new major HIV sequence [see also 21,22], and also by reports demonstrating rapid selection of cytotoxic T lymphocyte escape virus during PHI [27,28]. Despite the appearance of these escape mutants, HIV-1 RNA continues to decline in these patients, although at a much slower rate than earlier, suggesting that the immune system continues to respond to the new HIV variants generated over the course of the infection. In most patients, a steady state develops between replication and clearance of virus, though the level varies considerably from patient to patient. Interestingly, we found that steady-state viral loads were related to the peak levels measured during primary infection (r = 0.73–1.00;P < 0.05), thus corroborating what has already been demonstrated in a monkey model . It should be emphasized that the peak levels measured in our patients may not be the true peak values, as these are only possible to determine if samples are drawn daily from each patient.
The natural history of PHI has become more difficult to study in recent years since it is now generally recommended that patients with PHI should be given HAART . The rationale for early treatment is that viral load reduction during the earliest phases may prevent the seeding of a latently infected cell compartment, though it is now evident that treatment must be initiated very early during infection to prevent this . Early treatment may also reduce damage to CD4 T cell lymphoproliferative responses [32,33]. Indeed, very early treatment has been shown to lead to a reduction in viral loads in HIV-2-infected macaques that persists even after therapy is withdrawn . We have found, however, that the development of HIV-specific antibodies is at least partly inhibited by early treatment (data not shown). Therefore, the possible influence of the timepoint of initiation of treatment during PHI on both the specific immune response and the reduction of HIV load needs to be further elucidated.
1. Gaines H, von Sydow M, Pehrson PO, Lundbergh P. Clinical picture of primary HIV infection presenting as a glandular-fever-like illness.
Br Med J 1988, 297: 1363 –1368.
2. Heptonstall J, Porter KP, Gill ON. Occupational transmission of HIV: summary of published reports to July 1995
. London: Public Health Laboratory Service
, 1995 [internal report].
3. Tindall B, Barker S, Donovan B. et al. Characteristics of the acute clinical illness associated with human immunodeficiency virus infection.
Arch Intern Med 1988, 148: 945 –949.
4. Gaines H, von Sydow M, von Stedingk LV. et al. Immunological changes in primary HIV-1 infection.
AIDS, 1990, 4: 995 –999.
5. Cooper DA, Tindall B, Wilson EJ, Imrie A, Penny R. Characterization of T lymphocyte responses during primary infection with human immunodeficiency virus.
J Infect Dis 1988, 157: 889 –1896.
6. Thorstensson R, Andersson S, Lindbäck S. et al. Evaluation of 14 commercial HIV-1/HIV-2 antibody assays using serum panels of different geographical origin and clinical stage including a unique seroconversion panel.
J Virol Meth 1998, 70: 139 –151.
7. Albert J, Gaines H, Sönnerborg A. et al. Isolation of human immunodeficiency virus (HIV) from plasma during primary HIV infection.
J Med Virol 1987, 23: 67 –73.
8. Clark SJ, Saag MS, Decker WD. et al. High titers of cytopathic virus in plasma of patients with primary human immunodeficiency virus type 1 infection.
N Engl J Med 1991, 324: 954 –960.
9. Daar ES, Moudgil T, Meyer RD, Ho DD. Transient high levels of viraemia in patients with primary human immunodeficiency virus type 1 infection.
N Engl J Med 1991, 324: 961 –964.
10. Graziosi C, Pantaleo G, Butini L. et al. Kinetics of human immunodeficiency virus type 1 (HIV-1) DNA and RNA synthesis during primary HIV-1 infection.
Proc Natl Acad Sci USA 1993, 90: 6405 –6409.
11. Koup RA, Safrit JT, Cao Y. et al. Temporal association of cellular immune responses with the initial control of viremi in primary HIV-1 syndrome.
J Virol 1994, 68: 4650 –4655.
12. Borrow P, Lewicki H, Hahn BH, Shaw GM, Oldstone BMA. Virus-specific CD8+ cytotoxic T-lymphocyte activity associated with control of viremia in primary human immunodeficiency virus type 1 infection.
J Virol 1994, 68: 6103 –6110.
13. Schmitz JE, Kuroda MJ, Santra S. et al. Control of viremia in simian immunodeficiency virus infection by CD8+ lymphocytes.
Science 1999, 283: 857 –860.
14. Albert J, Abrahamsson B, Nagy K. et al. Rapid development of isolate-specific neutralizing antibodies after primary HIV-1 infection and consequent emergence of virus variants which resist neutralization by autologous sera.
AIDS 1990, 4: 107 –112.
15. Moore JP, Cao Y, Ho DD, Koup RA. Development of the anti-gp120 antibody response during seroconversion to human immunodeficiency virus type 1.
J Virol 1994, 68: 5142 –5515.
16. Phillips AN. Reduction of HIV concentration during acute infection: independence from a specific immune response.
Science 1996, 271: 497 –499.
17. Kaufmann GR, Cunningham P, Kelleher AD. et al. Patterns of viral dynamics during primary human immunodeficiency virus type 1 infection.
J Infect Dis 1998, 178: 1812 –1815.
18. Lillo FB, Ciuffreda D, Veglia F. et al. Viral load and burden modification following early antiretroviral therapy of primary HIV-1 infection.
AIDS 1999, 13: 791 –796.
19. Hoen B, Dumon B, Harzic. et al. Highly active antiretroviral treatment initiated early in the course of symptomatic primary HIV-1 infection: Results of the ANRS 053 trial.
J Infect Dis 1999, 180: 1342 –1346.
20. Zaunders JJ, Cunningham PH, Kelleher AD. et al. Potent antiretroviral therapy of primary human immunodeficiency virus type 1 (HIV-1) infection: partial normalization of T lymphocyte subsets and limited reduction of HIV-1 DNA despite clearance of plasma viremia.
J Infect Dis 1999, 180: 320 –329.
21. Karlsson AC, Gaines H, Sällberg M, Lindbäck S, Sönnerborg A. Reappearance of founder virus sequence in human immunodeficiency virus type 1-infected patientes.
J Virol 1999, 73: 6191 –6196.
22. Karlsson AC, Lindbäck S, Gaines H, Sönnerborg A. Characterization of the viral population during primary HIV-1 infection.
AIDS 1998, 12: 839 –847.
23. Miller CJ, McGhee JR. Progress towards a vaccine to prevent sexual transmission of HIV (review).
Nat Med 1996, 2: 751 –752.
24. Spira AI, Marx PA, Patterson BK. et al. Cellular targets of infection and route of viral dissemination after an intravaginal inoculation of simian immunodeficiency virus into Rhesus Macaques.
J Exp Med 1996, 183: 215 –25.
25. Zhang ZQ, Shuler T, Zupancic M. et al. 1999. Sexual transmission and propagation of SIV and HIV in resting and activated CD4 T cells.
Science 1999, 286: 1353 –1357.
26. Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection.
Nature 1995, 373: 123 –126.
27. Borrow P, Lewicki H, Wei X. et al. Antiviral pressure exerted by HIV-1-specific cytotoxic T lymphocytes (CTLs) during primary infection demonstrated by rapid selection of CTL escape virus.
Nat Med 1997, 3: 205 –211.
28. Wilson JD, Ogg GS, Allen RL. et al. Direct visualization of HIV-1 specific cytotoxic T lymphocytes during primary infection.
AIDS 2000, 18: 22 –33.
29. Lifson JD, Nowak MA, Goldstein S. et al. The extent of early viral replication is a critical determinant of the natural history of simian immunodeficiency virus infection.
J Virol 1997, 71: 9508 –9514.
30. Feinberg MB, Carpenter C, Fauci AS. et al. Report of the NHI panel to define principles of therapy of HIV infection and guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents.
Ann Intern Med 1998, 128: 1057 –1100.
31. Chun TW, Engel D, Berrey MM, Shea T, Corey L, Fauci AS. Early establishement of a pool of latently infected, resting CD4 (+) T cells during primary HIV-1 infection.
Proc Natl Acad Sci USA 1998, 95: 8869 –8873.
32. Ho D. Time to hit HIV, early and hard.
N Engl J Med 1995, 333: 450 –451.
33. Rosenberg ES, Billingsley JM, Caliendo AM. et al. Vigorous HIV-1-specific CD4+ T cell responses associated with control of viremia.
Science 1997, 278: 1447 –1450.
34. Watson A, McClure J, Ranchalis J. et al. Early postinfection antiviral treatment reduces viral load and prevents CD4+ cell decline in HIV type 2-infected macaques.
AIDS Res Hum Retroviruses 1997, 13: 1375 –1381.
© 2000 Lippincott Williams & Wilkins, Inc.