The virulence of microbial pathogens often changes during the course of epidemics . The extreme variability of human immunodeficiency virus type 1 (HIV-1) has given rise to several subtypes and circulating recombinant forms (CRFs) since the beginning of the pandemic . The virus is therefore still evolving. Moreover, widespread use of antiviral drugs in industrialized countries since 1996 might have applied added selective pressures, as illustrated by the occurrence of genotypic resistance to non-nucleoside reverse transcriptase inhibitors (NNRTIs) shortly following the introduction of this therapeutic class . Any increase in HIV-1 virulence could have implications for the management of the AIDS pandemic, necessitating closer clinical monitoring and more stringent prevention strategies, for example.
HIV-1 virulence and pathogenicity are complex notions that are difficult to quantify . The most direct measure of virulence, host mortality, has no longer been relevant to HIV-1 infection in industrialized countries since the introduction of highly active antiretroviral therapy in 1996 . As the CD4 cell count and viral load are highly predictive of HIV-1 disease progression [6–9], assaying these markers shortly following infection, before any antiretroviral treatment, is an indirect way of measuring trends in virulence.
Changes in the virulence of HIV-1 are an ongoing concern [2,4,10]. Studies of seroconverter cohorts have given divergent results. Most studies of patients who seroconverted before 1996 showed no temporal trends in prognostic markers or the rate of HIV disease progression [11–17]. Studies including individuals who seroconverted before and after 1996 have also given conflicting results, one suggesting a loss of virulence , some a gain [19,20] and others no change [5,21]. The most recent study, based on the Multicenter Aids Cohort Study (MACS) cohort, showed stable CD4 cell counts and HIV RNA levels measured after seroconversion from 1984 to 2005, but the number of patients who seroconverted after 1994 was small .
In these studies the date of seroconversion was usually estimated to within a very large interval (to within 12 or even 24 months), whereas marker levels may change rapidly during acute infection. In addition, temporal trends in the HIV-1 DNA level at the time of seroconversion, another informative prognostic marker, were not studied [22,23]. Finally, few potential confounders such as smoking, clinical signs of primary HIV infection (PHI), the duration of PHI, and ethnicity could be taken into account in most studies.
Since November 1996, the ongoing ANRS PRIMO cohort – a multicenter cohort in France – has enrolled more than 900 patients during PHI, allowing us to examine possible temporal trends in the CD4 cell count and HIV RNA and DNA levels measured at the time of the primary infection, taking into account changes in patient characteristics.
Patients and methods
The ANRS PRIMO cohort
The French ANRS PRIMO cohort is a prospective multicenter cohort of HIV-1 infected patients enrolled, since November 1996, during primary HIV infection . The PRIMO study protocol was approved by the Paris Cochin Ethics Committee. The diagnosis of PHI is based on a western blot profile compatible with ongoing seroconversion (incomplete western blot with absence of antibodies to pol proteins) in most patients (94%), detectable plasma HIV RNA with a negative or weakly reactive ELISA (2%), or an interval of less than 6 months between a negative and a positive ELISA result (4%). Primary HIV infection is considered symptomatic if at least one symptom associated with the acute HIV syndrome is present. Patients must be antiretroviral-naive at enrollment. The date of infection is estimated as the date of symptom onset minus 15 days or, in asymptomatic patients, the date of the incomplete western blot minus 1 month, or the midpoint between a negative and a positive ELISA result. With their written informed consent, patients are enrolled if HIV infection is estimated to have occurred less than 6 months previously [median observed interval in the cohort: 47 days, interquartile range (IQR) 35–68].
Patients are seen at one of the study sites at enrollment, then at month 1, month 3, and month 6, and subsequently every 6 months. Standardized questionnaires are completed and the patients have a physical examination and laboratory tests. The questionnaires include items on demographics (age, sex), behaviors (sexual preference, injection drug use, current smoking, and number of cigarettes/day), clinical manifestations (signs of PHI and, if present, the duration of symptomatic PHI), and other medical information such as coinfection with HCV or HBV.
Plasma and cells are collected at enrollment, every 6 months until month 24, and every year thereafter. The CD4 cell count and viral load are determined on site at each visit. The CD4 cell count is measured by flow cytometry using standard procedures. Plasma viral load is measured with the assay routinely used in each center. The Cobas Amplicor HIV-1 Monitor 1.5 assay kit (Roche Diagnostics, Meylan, France) was used in 72% of cases, the Versant HIV-1 RNA 3.0 assay (Bayer Diagnostics, Emeryville, California, USA) in 21% of cases (mainly before 2006), the NASBA QR system (Organon Technika) in 2% of cases (since 2006), and the Abbott Real-Time HIV-1 assay in 5% of cases (since 2003) .
Peripheral blood mononuclear cells (PBMCs) were isolated from fresh whole blood by centrifugation on a one-layer Ficoll Hypaque gradient. DNA was extracted from frozen PBMC and HIV DNA was quantified by real-time PCR (used from 1997 to 2007) in 76% of cases, as described elsewhere , and with the Amplicor HIV DNA Monitor kit (Roche Diagnostics) in the remaining 24% of cases (used from 1996 to 2001) . Results were expressed as the log10 number of HIV DNA copies per 106 PBMC, and the detection limit was 70 copies per 106 PBMC (i.e. 1.84 log10 copies) [22,23].
Patient samples were further characterized for resistance genotyping and genetic subtyping. Genotypic resistance tests were performed in three centralized laboratories on plasma samples before the initiation of treatment using the consensus technique of the ANRS resistance group . HIV resistance to nucleoside analog reverse transcriptase inhibitors, nonnucleoside reverse transcriptase inhibitors, and protease inhibitors, was defined according to the 2007 HIV-1 genotypic resistance interpretation algorithm of the French National Agency of Research on AIDS (ANRS) (www.hivfrenchresistance.org). The subtype and CRFs were determined by phylogenic analysis of reverse transcriptase gene.
Among the 936 patients enrolled in the cohort between November 1996 and February 2008, enrollment CD4 cell counts and HIV RNA values were available for this analysis in 903 cases (96.5%) and HIV DNA values in 775 cases (82.8%).
Characteristics of the study population and median values of the prognostic markers were described according to the calendar year of infection. Quantile regression was used to assess temporal trends . Quantile regression was introduced in econometry as an extension of linear modeling, in order to estimate the effect of covariables on various quantiles of the distribution of the response variable, and not only on its mean. As explained by Koenker and Hallock , the mean can be defined as the solution to the problem of minimizing a sum of squared residuals, whereas the median can be the solution to the problem of minimizing a sum of absolute residuals.
Univariate and multivariate linear regression models were used to evaluate temporal trends in the CD4 cell count and plasma HIV RNA and intracellular HIV DNA levels measured at the time of primary HIV infection, with each prognostic marker as the dependent variable. As the distribution of the CD4 cell count was right-skewed, square-root transformation was used to normalize the distribution of this marker. Log10 copy numbers of plasma HIV RNA and intracellular HIV DNA were used.
Temporal trends were first sought visually by using a linear regression model in which the year of infection was introduced as restricted cubic splines (RCS) . This enabled us to graphically represent the adjusted relations and to test global associations and deviation from linearity, but not to test for trends. The year of infection was then introduced as a categorical and then continuous variable to formally test for a temporal trend. The study period was divided in six intervals (1996–1997, 1998–1999, 2000–2001, 2002–2003, 2004–2005, 2006–2007), the reference period being the most recent one.
At each step, potential confounders were considered. As the relation between the three markers and both time since infection and age at PHI diagnosis is not linear, time since infection and age were introduced as RCS functions with knots at the 5th, 25th, 50th, 75th, and 95th percentiles . Indeed, our aim was not to quantify the effect of these factors on the markers but to adjust for these factors as well as possible. Other potential confounders taken into account were sex, country of birth (sub-Saharan African vs. others), symptomatic PHI, and smoking. Two virus-related factors were also considered in multivariate analysis, namely genotypic drug resistance mutations (at least one vs. none) and the HIV-1 subtype (B vs. others). We systematically verified that confounders were not modifiers before performing adjusted analyses.
Statistical analysis was performed using Stata/SE 9.0 software (Stata Press, College Station, Texas, USA).
The characteristics of the 903 patients are shown in Table 1. Mean age at HIV-1 infection was 35 years, and the mean interval between infection and the marker assays was 56 days. Most patients were infected through sexual intercourse (93.5%). Most patients were men (84%), homosexual (69% of the total population) and infected by a subtype B strain (76%), and most had symptomatic PHI (87%). Seven percent of the patients were black Africans, whereas the others were mainly from France or the French West Indies/Guyana (87% of non black Africans). Between 1996 and 2007 the frequency of symptomatic PHI increased, as did the percentage of men, mainly owing to an increase in the number of homosexual men. The percentage of individuals infected by non-B subtypes increased (mainly CRF02: 12.7% of the total population), whereas the percentage of genotypically resistant strains remained stable. There was a nonsignificant trend towards an increase in the proportion of African patients.
Table 2 shows the median values of the three prognostic markers and their IQRs according to the calendar period of infection. The median CD4 cell count was 519/μl (IQR 372; 690), the median HIV-1 RNA level was 5.09 log10 copies/ml (IQR 4.45; 5.69), and the median HIV DNA level was 3.33 log10 copies/106 PBMC (IQR 2.93; 3.67). Quantile regression showed no temporal trend in the median CD4 cell count or the median HIV RNA level (P = 0.14 and 0.31, respectively), whereas the median HIV DNA level increased significantly over the years in the crude analysis (P = 0.01).
When the year of infection was introduced as a RCS function in a regression model, we found no graphical evidence of deviation from linearity for any of the markers. Moreover, the test for deviation from linearity was nonsignificant. This enabled us to introduce the year of infection in 2-year classes, and thus formally to test for a temporal trend. The associations between the year of infection and the CD4 cell count (√CD4 /μl), HIV RNA level (log10 copies/ml), and HIV DNA level (log10 copies/106 PBMC) in crude and multiple linear regression models are shown in Table 3. We observed no calendar trend in the CD4 cell count or HIV RNA level. Furthermore, the significant trend in the HIV DNA level observed in the crude analysis disappeared in the adjusted model.
To improve the consistency of our results, we performed several sensitivity analyses. The trend results remained unchanged when we replaced sex with sexual preference divided into three classes (homosexual men, heterosexual men, and women). As most patients were infected with subtype B strains, the HIV-1 subtype was entered as a binary variable (subtype B vs. others). However, when the analyses were restricted to patients infected by subtype B strains, the results remained unchanged. Finally, we also verified that similar trend results were obtained when we adjusted for host factors alone.
As expected, all three markers were associated with the time since infection. Other well described associations were also found, including those with sex and with symptomatic PHI. Women had higher CD4 cell counts and lower HIV RNA and DNA levels than men. Patients with symptomatic PHI had lower CD4 cell counts and higher levels of HIV RNA and DNA than patients with asymptomatic PHI. The CD4 cell count was influenced by ethnicity and smoking, black African patients having lower counts than patients born in other countries, and smokers having higher counts than nonsmokers. Concerning viral factors, patients infected by subtype B strains had higher CD4 cell counts and lower HIV DNA levels than patients infected by other subtypes. Patients infected by resistant strains had lower HIV RNA levels than patients infected by wild-type strains, and the same results were obtained when the analyses were adjusted for sexual preference instead of sex.
To our knowledge this is the first study of calendar trends in three major prognostic markers of HIV-1 disease progression measured at the time of primary HIV infection. Moreover, contrary to most previous studies, the analyses were adjusted for most known confounding factors, namely sex, smoking, symptomatic PHI, and the HIV-1 subtype. We found that initial CD4 cell counts and HIV RNA and DNA levels were stable across 12 consecutive years (1996–2007) in the French ANRS PRIMO cohort, suggesting that HIV-1 virulence has remained stable.
Some studies have suggested an increase in HIV-1 virulence [19,20], whereas others have suggested the contrary . Our results conflict somewhat with those of the Italian HIV-seroconversion study conducted between 1985 and 2002 , but they are compatible with the Cascade results for similar calendar periods [18,20]. Our results are also consistent with a Swiss report of stable virulence  and with the results of a recent American study of patients enrolled in the MACS cohort . Muller et al. found no time trends in virulence markers between 1984 and 2002 among white male and female northwest European patients, using as proxies the slopes of the declines in the CD4 cell count and in the CD4: CD8 ratio, and the ‘viral set-point’. Herbeck et al. used three prognostic markers of disease progression as proxies for virulence, namely the plasma HIV RNA level and the CD4 cell count at ‘set-point’ (measured between 9 and 15 months after seroconversion) and the rate of the CD4 cell count decline during the first 3 years after seroconversion. They found no significant trend between 1985 and 2005 in this population of men who have sex with men (MSM), but only 18 of the 357 participants seroconverted during the most recent period, from 1995 to 2005. Moreover, smoking, symptomatic PHI, and the HIV-1 subtype were not taken into account in these analyses.
We chose to study the prognostic markers at the time of primary HIV infection, as we did in a recent genomewide association study . An alternative would have been to measure the markers further from infection, for instance at the ‘viral set-point’, when markers are supposed to be more stable . However, this would have meant restricting the analysis to those patients who had not yet started antiretroviral treatment. Moreover, in the PRIMO cohort, the proportion of early-treated patients fell from 88% in 1996–1997 to 29% in 2006–2007, early treatment being increasingly reserved for seriously ill patients. Finally, our decision to focus on PHI did not prevent us from observing the following well described associations. The CD4 cell count varied according to age at seroconversion, sex, and ethnicity, as described in the literature [18,31–36]. We also found that the mean CD4 cell count was about 100/μl higher in smokers than in nonsmokers [37,38]. HIV RNA levels varied according to sex and the symptomatic nature of PHI [39–41]. These associations underline the need, when assessing temporal trends in prognostic markers, to take into account population characteristics, which may also change over the years. Virus-related factors such as drug resistance mutations and the HIV-1 subtype might also influence prognostic marker values, as observed here [42–44]. In our study the proportions of injecting drug users and HCV-coinfected patients were very small (0.6 and 3.2%, respectively). We cannot exclude the possibility that diurnal variations might have influenced the CD4 cell count in some patients, but it should be noted that most samples were drawn in the morning at hospital; moreover, this did not prevent us from observing the well known influence of sex and smoking on the CD4 cell count. In keeping with French national HIV surveillance data, we attribute the increasing frequency of symptomatic PHI to a higher frequency of moderately symptomatic acute HIV-1 infections diagnosed in recent years, whereas such cases previously went unnoticed and were diagnosed only during the chronic phase .
One potential limitation of our study is that we did not take into account the different HIV RNA and DNA assay techniques used in the participating centers. Indeed, various types of assays were used to quantify HIV RNA and HIV DNA, but measuring assays were linked to the calendar period, making adjustment difficult. Furthermore, a given assay may evolve over time. These concerns have been raised in several studies [5,19,20], and can be offset by investigating temporal trends in several markers simultaneously, as was the case here.
Finally, the use of prognostic markers studied around the time of seroconversion as a proxy for virulence assumes that any changes in virulence would be observable early after infection. Furthermore, we cannot rule out the possibility that HIV-1 virulence is evolving too slowly for changes to have been observed during our 12-year study period.
In conclusion, after controlling for known confounding factors, we observed no temporal trends in the values of three major predictors of HIV-1 disease progression between 1996 and 2007, as measured during primary HIV-1 infection among patients enrolled in the French ANRS PRIMO cohort. Further studies of large cohorts of seroconverters will be necessary to detect possible future changes in HIV-1 virulence.
We thank M.O. Wehr, A. Talamani, and Y. Zittoun for data monitoring, and D. Young for editing the manuscript. We also thank all the patients and the participating physicians.
The PRIMO cohort is funded by ANRS (Agence Nationale de Recherches sur le SIDA et les hépatites virales) and P.T. received a grant for this study from Fondation pour la Recherche Médicale.
1. Ewald PW. The evolution of virulence. Sci Am 1993; 268:86–93.
2. Brander C, Walker BD. Gradual adaptation of HIV to human host populations: good or bad news? Nat Med 2003; 9:1359–1362.
3. Chaix M, Descamps D, Wirden W, Bocket L, Delaugerre C, Tamalet C, et al
. Stable frequency of HIV-1
transmitted drug resistance in patients at the time of primary infection over 1996-2006 in France. AIDS 2009; 23:717–724.
4. Van de Perre P. Viral and host determinants of HIV-1
pathogenesis. AIDS 2006; 20:933–934.
5. Herbeck JT, Gottlieb GS, Li X, Hu Z, Detels R, Phair J, et al
. Lack of evidence for changing virulence of HIV-1
in North America. PLoS ONE 2008; 3:e1525.
6. Mellors JW, Munoz A, Giorgi JV, Margolick JB, Tassoni CJ, Gupta P, et al
. Plasma viral load
and CD4+ lymphocytes as prognostic markers of HIV-1
infection. Ann Intern Med 1997; 126:946–954.
7. Hubert JB, Burgard M, Dussaix E, Tamalet C, Deveau C, Le Chenadec J, et al
. Natural history of serum HIV-1
RNA levels in 330 patients with a known date of infection. The SEROCO Study Group. AIDS 2000; 14:123–131.
8. Sabin CA, Devereux H, Phillips AN, Hill A, Janossy G, Lee CA, Loveday C. Course of viral load
infection. J Acquir Immune Defic Syndr 2000; 23:172–177.
9. Phillips A. Short-term risk of AIDS according to current CD4 cell count
and viral load
in antiretroviral drug-naive individuals and those treated in the monotherapy era. AIDS 2004; 18:51–58.
10. Quinones-Mateu ME. Is HIV-1
evolving to a less virulent (pathogenic) virus? AIDS 2005; 19:1689–1690.
11. O'Brien TR, Hoover DR, Rosenberg PS, Chen B, Detels R, Kingsley LA, et al
. Evaluation of secular trends in CD4+ lymphocyte loss among human immunodeficiency virus type 1 (HIV-1
)-infected men with known dates of seroconversion
. Am J Epidemiol 1995; 142:636–642.
12. Holmberg SD, Conley LJ, Luby SP, Cohn S, Wong LC, Vlahov D. Recent infection with human immunodeficiency virus and possible rapid loss of CD4 T lymphocytes. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 9:291–296.
13. Galai N, Lepri AC, Vlahov D, Pezzotti P, Sinicco A, Rezza G. Temporal trends of initial CD4 cell counts following human immunodeficiency virus seroconversion
in Italy, 1985–1992. The Human Immunodeficiency Virus Italian Seroconversion
Study. Am J Epidemiol 1996; 143:278–282.
14. Carre N, Prins M, Meyer L, Brettle RP, Robertson JR, McArdle H, et al
. Has the rate of progression to AIDS changed in recent years? AIDS 1997; 11:1611–1618.
15. Vanhems P, Lambert J, Guerra M, Hirschel B, Allard R. Association between the rate of CD4+ T cell decrease and the year of human immunodeficiency virus (HIV) type 1 seroconversion
among persons enrolled in the Swiss HIV cohort study. J Infect Dis 1999; 180:1803–1808.
16. Sinicco A, Fora R, Raiteri R, Sciandra M, Bechis G, Calvo MM, Gioannini P. Is the clinical course of HIV-1
changing? Cohort study. BMJ 1997; 314:1232–1237.
17. Keet IP, Veugelers PJ, Koot M, de Weerd MH, Roos MT, Miedema F, et al
. Temporal trends of the natural history of HIV-1
infection following seroconversion
between 1984 and 1993. AIDS 1996; 10:1601–1602.
18. CASCADE. Differences in CD4 cell counts at seroconversion and decline among 5739 HIV-1-infected individuals with well estimated dates of seroconversion
. J Acquir Immune Defic Syndr
19. Dorrucci M, Phillips AN, Longo B, Rezza G. Changes over time in postseroconversion CD4 cell counts in the Italian HIV-seroconversion
study: 1985–2002. AIDS 2005; 19:331–335.
20. Dorrucci M, Rezza G, Porter K, Phillips A. Temporal trends in postseroconversion CD4 cell count
and HIV load: the Concerted Action on Seroconversion
to AIDS and Death in Europe Collaboration, 1985–2002. J Infect Dis 2007; 195:525–534.
21. Muller V, Ledergerber B, Perrin L, Klimkait T, Furrer H, Telenti A, et al
. Stable virulence levels in the HIV epidemic of Switzerland over two decades. AIDS 2006; 20:889–894.
22. Rouzioux C, Hubert JB, Burgard M, Deveau C, Goujard C, Bary M, et al
. Early levels of HIV-1
DNA in peripheral blood mononuclear cells are predictive of disease progression independently of HIV-1
RNA levels and CD4+ T cell counts. J Infect Dis 2005; 192:46–55.
23. Goujard C, Bonarek M, Meyer L, Bonnet F, Chaix ML, Deveau C, et al
. CD4 cell count
and HIV DNA level are independent predictors of disease progression after primary HIV type 1 infection in untreated patients. Clin Infect Dis 2006; 42:709–715.
24. Schutten M, Fries E, Burghoorn-Maas C, Niesters HG. Evaluation of the analytical performance of the new Abbott RealTime RT-PCRs for the quantitative detection of HCV and HIV-1
RNA. J Clin Virol 2007; 40:99–104.
25. Rouet F, Ekouevi DK, Chaix ML, Burgard M, Inwoley A, Tony TD, et al
. Transfer and evaluation of an automated, low-cost real-time reverse transcription-PCR test for diagnosis and monitoring of human immunodeficiency virus type 1 infection in a West African resource-limited setting. J Clin Microbiol 2005; 43:2709–2717.
26. Avettand-Fenoel V, Boufassa F, Galimand J, Meyer L, Rouzioux C. HIV-1
DNA for the measurement of the HIV reservoir is predictive of disease progression in seroconverters whatever the mode of result expression is. J Clin Virol 2008; 42:399–404.
27. Koenker R, Hallock KF. Quantile regression. J Econ Perspect 2001; 15:143–156.
28. Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989; 8:551–561.
29. Greenland S. Dose-response and trend analysis in epidemiology
: alternatives to categorical analysis. Epidemiology
30. Dalmasso C, Carpentier W, Meyer L, Rouzioux L, Goujard C, Chaix M-L, et al
. Distinct genetic loci control plasma HIV-RNA and cellular HIV-DNA levels in HIV-1
infection: the ANRS Genome Wide Association 01 Study. PLoS ONE 2008; 3:e3907.
31. Touloumi G, Hatzakis A, Rosenberg PS, O'Brien TR, Goedert JJ. Effects of age at seroconversion
and baseline HIV RNA level on the loss of CD4+ cells among persons with hemophilia. Multicenter Hemophilia Cohort Study. AIDS 1998; 12:1691–1697.
32. Delmas MC, Jadand C, De Vincenzi I, Deveau C, Persoz A, Sobel A, et al
. Gender difference in CD4+ cell counts persist after HIV-1
infection. SEROCO Study Group. AIDS 1997; 11:1071–1073.
33. Tsegaye A, Messele T, Tilahun T, Hailu E, Sahlu T, Doorly R, et al
. Immunohematological reference ranges for adult Ethiopians. Clin Diagn Lab Immunol 1999; 6:410–414.
34. Messele T, Abdulkadir M, Fontanet AL, Petros B, Hamann D, Koot M, et al
. Reduced naive and increased activated CD4 and CD8 cells in healthy adult Ethiopians compared with their Dutch counterparts. Clin Exp Immunol 1999; 115:443–450.
35. Tugume SB, Piwowar EM, Lutalo T, Mugyenyi PN, Grant RM, Mangeni FW, et al
. Hematological reference ranges among healthy Ugandans. Clin Diagn Lab Immunol 1995; 2:233–235.
36. Bunders M, Lugada E, Mermin J, Downing R, Were W, Thorne C, Newell ML. Within and between race differences in lymphocyte, CD4+, CD8+ and neutrophil levels in HIV-uninfected children with or without HIV exposure in Europe and Uganda. Ann Trop Paediatr 2006; 26:169–179.
37. Sunyer J, Munoz A, Peng Y, Margolick J, Chmiel JS, Oishi J, et al
. Longitudinal relation between smoking and white blood cells. Am J Epidemiol 1996; 144:734–741.
38. Royce RA, Winkelstein W Jr. HIV infection, cigarette smoking and CD4+ T-lymphocyte counts: preliminary results from the San Francisco Men's Health Study. AIDS 1990; 4:327–333.
39. Farzadegan H, Hoover DR, Astemborski J, Lyles CM, Margolick JB, Markham RB, et al
. Sex differences in HIV-1 viral load
and progression to AIDS. Lancet 1998; 352:1510–1514.
40. Prins M, Meyer L, Hessol NA. Sex and the course of HIV infection in the pre and highly active antiretroviral therapy eras. AIDS 2005; 19:357–370.
41. Vanhems P, Lambert J, Cooper DA, Perrin L, Carr A, Hirschel B, et al
. Severity and prognosis of acute human immunodeficiency virus type 1 illness: a dose-response relationship. Clin Infect Dis 1998; 26:323–329.
42. Grant RM, Hecht FM, Warmerdam M, Liu L, Liegler T, Petropoulos CJ, et al
. Time trends in primary HIV-1
drug resistance among recently infected persons. JAMA 2002; 288:181–188.
43. Pillay D, Bhaskaran K, Jurriaans S, Prins M, Masquelier B, Dabis F, et al
. The impact of transmitted drug resistance on the natural history of HIV infection and response to first-line therapy. AIDS 2006; 20:21–28.
44. Hu DJ, Vanichseni S, Mastro TD, Raktham S, Young NL, Mock PA, et al
. Viral load
differences in early infection with two HIV-1
subtypes. AIDS 2001; 15:683–691.
45. InVS. Special issue: HIV and AIDS infection in France and in Europe
. Bull Epidemiol Hebd