HIV Articles  
Back 
 
 
Predictive Factors for Long-Term Non-Progressors
 
 
  ".....little is known about LTNPs during the first years of known infection, that is, before they are identified as LTNPs.....In this study, the LTNP status was progressively lost in more than half of the LTNPs....In conclusion, our study shows that LTNP status may be established early during infection. Low level of HIV DNA in PBMC at enrollment was found to be a strong predictor of the duration of long-term nonprogression......Describing how virological markers such as plasma HIV RNA and HIV DNA in PBMCs evolve during the first years of infection in the patients who have been considered as LTNP after 8 years of infection could help to characterize the viral infection in this particular group of patients. These 2 markers reflect the pathophysiological mechanisms, HIV RNA as a marker for HIV replication and HIV DNA as a marker of HIV reservoirs in patients. This could contribute to understand mechanisms involved in the early and maintained control of HIV infection over time. It is also useful to be able to target future LTNPs, to undertake immunological and virological investigations early in HIV infection......Our results show that patients with baseline HIV RNA and HIV DNA below detection limits have a 60% chance of being LTNP. As a direct consequence of our selection criteria, all patients presented with CD4 cell counts >500 cells/mm3 at enrollment. However, even in this selected population, CD4 cell counts remained a strong prognostic factor for being an LTNP.....Another strength of this study is the characterization of CD4 cell count, HIV RNA, and HIV DNA evolutions in LTNPs during the first 8 years of known infection. CD4 cell counts were high at inclusion and even higher in women than in men as already shown in various populations22; they were also higher in younger patients.23 During the first 8 years of documented infection, the yearly CD4 cell count decrease was only 32 cells/mm3 in LTNPs aged less than 25 years at HIV diagnosis, and 16 cells/mm3 in older LTNPs, compared with the decrease observed in more general populations of 60 to >100 cells/mm3 per year. LTNPs presented low level of HIV replication represented by HIV RNA in plasma and low level of HIV reservoirs represented by low HIV DNA level in blood infected cells.....
 
.......In a multivariate model, the baseline HIV DNA level and the HIV DNA rate of change remained independently associated with loss of LTNP status. A baseline HIV DNA above the detection limit compared with below and a high increase (>0.103 log copies/106 PBMCs/year) of HIV DNA level during the first 8 years of documented HIV infection both increased the risk of loosing LTNP status [adjusted hazard ratio (95% CI): 2.8 (1.2 to 6.8) and 2.2 (1.0 to 4.8), respectively). It nevertheless must be reminded that this analysis is based on 60 LTNP patients, this small sample size inducing a limited statistical power......patients with a baseline HIV DNA level above detection limit were 3.4-fold more likely to lose their LTNP status than patients with an HIV DNA level below this limit....subjects with the highest HIV DNA increase (>0.103 log copies/106 PBMCs/year) were 2.7-fold more likely to lose their LTNP status.....Baseline HIV RNA tended to be lower in LTNPs who maintained their status than in LTNPs who lost their status [median (IQR): 2.86 (2.30-3.38) and 3.20 (2.30-3.70) log copies/mL, P = 0.22], although not significant. HIV RNA rate of change was not associated with the loss of LTNP status. Women tended to lose less rapidly their LTNP status than men (P = 0.07)...."
 
Early Control of HIV-1 Infection in Long-Term Nonprogressors Followed Since Diagnosis in the ANRS SEROCO/HEMOCO Cohort
 
JAIDS Journal of Acquired Immune Deficiency Syndromes:Volume 50(1)1 January 2009pp 19-26
 
Madec, Yoann PhD*; Boufassa, Faroudy MD*; Avettand-Fenoel, Veronique PharmD_; Hendou, Samia MD*; Melard, Adeline MSc_; Boucherit, Soraya MD*; Surzyn, Janina BA*; Meyer, Laurence MD, PhD*; Rouzioux, Christine PharmD, PhD_; for the ANRS SEROCO/HEMOCO Study Group
 
From the *INSERM U822, Le Kremlin-Bicetre, France; Universite Paris-Sud, Faculte de Medecine Paris-Sud, Le Kremlin-Bicetre, France; AP-HP, Hopital de Bicetre, Service de Sante Publique, France; Unite d'Epidemiologie des maladies Emergentes, Institut Pasteur, 25-28 rue du Dr. Roux, Paris, France; and _Laboratoire de Virologie, Hopital Necker AP-HP, EA 3620 Universite Paris Descartes, 149 rue de Sevres, Paris, France.
 
Abstract
 
Background: To clarify early correlates and natural history of HIV long-term nonprogressors (LTNPs) since HIV diagnosis.
 
Methods: Patients enrolled in the French ANRS SEROCO/HEMOCO cohort with CD4 count >500 cells/mm3 at HIV diagnosis. LTNP status was defined as being asymptomatic, antiretroviral free, and with CD4 cell count >500 cells/mm3 for >8 years after HIV diagnosis. In LTNPs, we modeled the biological markers' progression through a joint model. Factors associated with loss of LTNP status were identified through a Cox model.
 
Results: Sixty (9%) of 664 patients were identified as LTNPs during follow-up. At enrollment, HIV RNA was ≦2.6 log copies/mL in 24% of LTNPs and HIV DNA was ≦1.85 log copies/106 peripheral blood mononuclear cells (PBMCs) in 31% vs. 3% and 8% in others. In LTNPs, HIV RNA and HIV DNA levels increased by 0.04 log copies/mL per year and 0.07 log copies/106 PBMCs per year during the first 8 years after diagnosis. LTNP status was lost in 36 subjects; baseline HIV DNA >1.85 log copies/106 PBMCs and high HIV DNA increase were associated with an increased risk of losing LTNP status [adjusted hazard ratio: 2.8 (1.2-6.8) and 2.2 (1.0-4.8), respectively].
 
Conclusions: LTNP status is established in the first years of HIV infection, low HIV DNA level at enrollment and slow increase of HIV DNA being associated with maintained LTNP status.
 
INTRODUCTION
 
Long-term nonprogressors (LTNPs) are HIV-1-infected patients who remain asymptomatic with CD4 cell count ≥500 cells/mm3 in the absence of antiretroviral (ARV) therapy during 7-15 years, depending on the definition used.1-3 However, the duration of 8 years is the most commonly used.1-5 These patients would represent from 1% to 5% of the HIV-infected patients.3,6,7
 
Most studies dealing with LTNPs focused on the outcome of patients, once they are already identified as LTNPs.3,8 As a result, little is known about LTNPs during the first years of known infection, that is, before they are identified as LTNPs. It has been reported that LTNPs present low levels of HIV replication and HIV DNA in peripheral blood mononuclear cells (PBMCs), in association with HIV-specific T-cell immune responses.9 Describing how virological markers such as plasma HIV RNA and HIV DNA in PBMCs evolve during the first years of infection in the patients who have been considered as LTNP after 8 years of infection could help to characterize the viral infection in this particular group of patients. These 2 markers reflect the pathophysiological mechanisms, HIV RNA as a marker for HIV replication and HIV DNA as a marker of HIV reservoirs in patients. This could contribute to understand mechanisms involved in the early and maintained control of HIV infection over time. It is also useful to be able to target future LTNPs, to undertake immunological and virological investigations early in HIV infection.
 
Here we describe the natural course, from HIV-1 diagnosis, of patients enrolled in the French ANRS SEROCO/HEMOCO cohort who had CD4 cell counts >500 cells/mm3 at time of HIV diagnosis or HIV infection. The first aim was to identify factors associated with the LTNP status, this status being defined after 8 years of known infection. We also aimed at jointly modeling CD4, blood HIV RNA, and PBMC HIV DNA progression in LTNPs during the first 8 years after HIV diagnosis. Finally, we identified risk factors for loosing LTNP status after these 8 years.
 
DISCUSSION
 
This is, to our knowledge, the first study to describe the natural course of HIV infection in LTNPs from HIV diagnosis and from seroconversion for a substantial number of them. In this cohort, blood cells and blood plasma samples have been collected over a very long period, enabling us to describe markers' progression in LTNPs from soon after seroconversion.
 
In this large ANRS SEROCO/HEMOCO cohort of HIV-infected patients, 60 (9%) patients of those enrolled with CD4 cell count >500 cells/mm3 while ARV free were LTNPs as they maintained this high CD4 level without ARV for at least 8 years. LTNPs represented 3.4% of all the patients in the cohort, a proportion similar to that observed in other studies.3,6,7
 
One of the main interests of this study lies in its ability to identify early prognostic factors of becoming LTNP, knowing that 15% of the LTNPs were included <6 months from seroconversion and another 15% <24 months. Indeed, CD4 cell counts and HIV RNA and HIV DNA levels were available from enrollment, before the patients' LTNP status could be recognized. HIV RNA and HIV DNA at enrollment were found to be strong and independent prognostic factors of LTNP status. These results are in agreement with previous findings reporting that these markers are predictive of disease progression in more general populations of HIV-infected patients.15,16,25,26 Our results show that patients with baseline HIV RNA and HIV DNA below detection limits have a 60% chance of being LTNP. As a direct consequence of our selection criteria, all patients presented with CD4 cell counts >500 cells/mm3 at enrollment. However, even in this selected population, CD4 cell counts remained a strong prognostic factor for being an LTNP.
 
Another strength of this study is the characterization of CD4 cell count, HIV RNA, and HIV DNA evolutions in LTNPs during the first 8 years of known infection. Logically, CD4 cell counts were high at inclusion and even higher in women than in men as already shown in various populations22; they were also higher in younger patients.23 During the first 8 years of documented infection, the yearly CD4 cell count decrease was only 32 cells/mm3 in LTNPs aged less than 25 years at HIV diagnosis, and 16 cells/mm3 in older LTNPs, compared with the decrease observed in more general populations of 60 to >100 cells/mm3 per year.27-29 LTNPs presented low level of HIV replication represented by HIV RNA in plasma and low level of HIV reservoirs represented by low HIV DNA level in blood infected cells. Indeed, low level of HIV RNA was maintained during the first 8 years of known infection: a low increase of only +0.04 log copies/mL per year was observed, compared with 0.14 log copies/mL per year in natural history after the initial decrease in postseroconversion in more general HIV-infected populations.30 HIV DNA level also remained at a low level during the first 8 years of known infection. Overall, HIV DNA slowly increased with a mean rate of 0.07 log copies/106 PBMCs per year. It remained below the detection limit throughout the first 8 years of documented infection in 3 patients, 1 also being an HIV controller.24 This result does not totally agree with another study based on 8 LTNPs that showed no significant increase of HIV DNA over a period of 4 years after being identified as LTNPs.8
 
It is noteworthy to mention that in this study, 7 patients were not only LTNPs but were also HIV controllers,24 that is, they spontaneously controlled their viral replication in blood for more than 10 years. In these patients, the pivotal role of HIV-specific CD8+ T cells in the spontaneous viral control has recently been suggested.31 The mechanism of HIV control in LTNP seems complex however. In LTNP, a low level of HIV DNA was associated with a high level of IgG2 HIV antibodies, a high level of CD4 Th1 cell responses, and with HIV-1 Gag-specific CD8+ T-cell response.9,32 Lastly, host genetic characteristics have also been associated with nonprogression, especially some major histocompatability complex class 1 allele human leukocyte antigen B57.9,33,34
 
In this study, the LTNP status was progressively lost in more than half of the LTNPs. The longitudinal model showed that the CD4 count regularly decreased even in this very particular population. Interestingly, a steep increase in HIV DNA was associated with an increased risk of loosing LTNP status. This marker, which has already been found to be an efficient prognostic factor of HIV-1 disease progression in seroconverters and patients enrolled during primary infection,16,25,35 is therefore also a prognostic factor of duration of nonprogression. Although not significant, women tended to lose less rapidly the LTNP status than men, as were patients with low baseline HIV RNA. It must be kept in mind that the analysis was based on 60 LTNPs thus limiting the statistical power and leading to wide CIs for the OR.
 
In conclusion, our study shows that LTNP status may be established early during infection. Low level of HIV DNA in PBMC at enrollment was found to be a strong predictor of the duration of long-term nonprogression.
 
RESULTS
 
Predictive Factors for Being LTNP

 
Of the 664 patients, with >500 CD4 cells/mm3 at enrollment while naive of ARV, 60 (9%) met the LTNP definition criteria, ie, remained asymptomatic with CD4 cell counts always ≥500 cells/mm3 during the first 8 years of known infection and without ARV. The 604 other patients all progressed to a CD4+ cell count <500 cells/mm3. During follow-up, the median time interval between 2 consecutive CD4 measurements was 6.1 (IQR: 5.6-6.9) months and 5.9 (IQR: 4.0-6.5) months in LTNPs and progressors, respectively (P < 0.001). Although significant, likely due to the large number of observations, this difference was not clinically relevant. Because follow-up was longer in LTNPs than in progressors, 16 years (IQR: 11-17) vs. 10 years (IQR: 6-16), the number of CD4 measurements was also larger in LTNPs than in progressors [median (IQR): 19 (15-23) and 8 (5-13), respectively].
 
The proportion of men was not significantly different between LTNPs and progressors (Table 1) nor was the proportion of seroconverters (30.0% vs. 38.7%, P = 0.18). The median age at inclusion was lower in LTNPs than in progressors (25 vs. 28 years, P = 0.02). Although only patients with high CD4 cell counts at enrollment were selected, LTNPs still had higher baseline CD4 cell counts than progressors (840 vs. 668 cells/mm3 in median, P < 0.0001) and higher baseline CD4 percentage (35% vs. 32% in median, P = 0.0002). LTNPs also had lower baseline CD8 cell counts than progressors, leading to a higher CD4-CD8 ratio in LTNPs: more than 50% of the LTNPs had a CD4-CD8 ratio >1. HIV RNA levels were lower in LTNPs than in progressors, and the proportion of patients with HIV RNA below detection limit at enrollment was significantly larger in LTNPs than in progressors (24.1% vs. 2.6%, P < 0.0001). Similarly, HIV DNA levels were lower in LTNPs than in progressors, and again the proportion of patients with HIV DNA levels below the detection limit was significantly larger in LTNPs than in progressors (30.9% vs. 8.0%, P < 0.0001). Both HIV RNA and HIV DNA levels were below the detection limits at enrollment in 13.3% of LTNPs vs. 0.5% of progressors (P < 0.0001).
 
The ability to control viral infection was established early during infection in LTNPs; indeed, baseline HIV RNA and HIV DNA levels were similar in LTNPs included within 6-24 months from seroconversion and in seroprevalent LTNPs (Table 2). When considering subjects enrolled more than 6 months after seroconversion and seroprevalent patients, the probability to be LTNP was 57% when enrollment HIV RNA was below the detection limit, 38% when enrollment HIV DNA was below the detection limit, and 73% when both markers were below the detection limits. Although not significant, LTNPs included early after seroconversion (0-6 months) tended to present higher HIV RNA and HIV DNA levels and lower CD4 cell counts than LTNPs enrolled afterward. Likely, this result was due to the measurements taking place closer to seroconversion. Of note, in patients enrolled during this early stage, CD4 counts and HIV DNA levels were not significantly different in LTNPs and progressors, whereas HIV RNA levels were lower in LTNPs.
 
After adjustment for gender and age in multivariate logistic regression, CD4 count at enrollment between 675 and 844 cells/mm3 and >844 cells/mm3 compared with a CD4 count <586 cells/mm3 [adjusted odds ratio (aOR): 5.28, 95% confidence interval (CI): 1.37 to 20.35 and aOR: 6.35, 95% CI: 1.66 to 24.36], an HIV RNA below the detection limit [aOR: 8.60 (95% CI: 2.72 to 27.21)], and an HIV DNA below the detection limit [aOR: 3.84 (95% CI: 1.59 to 9.30)] independently increased the chance of being an LTNP.
 
Immunological and Virological Markers' Evolution in LTNPs
 
During the first 8 years after diagnosis, the 60 LTNPs presented multiple CD4 cell counts [median 12 (range 3-27) measurements], 55 LTNPs presented multiple HIV RNA measurements [median 10 (2-17) measurements], and 4 LTNPs presented a single HIV RNA measurement; finally, 47 LTNPs presented multiple HIV DNA measurements [median 3 (2-6) measurements] and 10 presented a single HIV DNA measurement.
 
The CD4 cell count at diagnosis was significantly higher in women than in men and was also higher in LTNPs who were infected before 25 years of age than in those infected at a later age. The estimated mean (95% CI) CD4 cell count per cubic millimeter at HIV diagnosis was 963 (889 to 1039) vs. 875 (812 to 940) in younger and older men and 1076 (992 to 1165) vs. 983 (893 to 1078) in younger and older women, respectively. The rate of change was not significantly different in men and women, but it was influenced by age. Over the 8-year period, the mean CD4 cell count decline was -32 cells/mm3 per year in younger LTNPs and -16 cells/mm3 per year in older ones.
 
In 16 LTNPs, the first HIV RNA measurement was below the detection limit. These 16 patients maintained an HIV RNA below the detection limit for 6.0 years in median (IQR: 2.3-7.2), and 7 of them even maintained undetectable HIV RNA (despite a single blip <3 log copies/mL in 3) for more than 10 years thus meeting the HIV-controller definition (controlled viral replication in blood for more the 10 years).24 Among all 59 LTNPs with data, the estimated mean HIV RNA at the time of HIV diagnosis was 3.21 (95% CI: 2.97 to 3.45) log copies/mL in men and 2.83 (95% CI: 2.50 to 3.16) log copies/mL in women (P = 0.034) but did not differ by age. The HIV RNA level increased at a mean rate of 0.040 log copies/mL per year (95% CI: 0.002 to 0.079), corresponding to a mean increase over the 8 first years after diagnosis of 0.32 log copies/mL. The HIV RNA rate of change did not differ by gender or age (Fig. 1).
 
In 19 LTNPs, the first HIV DNA measurement was below the detection limit. Of these 19 patients, 10 also had their first HIV RNA measurement below the detection limit. These 19 patients maintained the HIV DNA below the detection limit for 3.6 years in median (IQR: 0.6 - 7.1), and 3 of them even maintained the HIV DNA below this limit throughout the 8 first years after diagnosis; of these 3 patients, 1 also sustained HIV RNA below the detection limit throughout the first 8 years after diagnosis. Among all 47 LTNPs with data, the estimated HIV DNA level at the time of HIV diagnosis was in mean 2.19 (95% CI: 2.00 to 2.37) log copies/106 PBMCs in men and 1.75 (95% CI: 1.49 to 2.01) log copies/106 PBMCs in women (P = 0.005) but did not differ by age. HIV DNA increased at a mean rate of 0.069 log copies/106 PBMCs per year [95% CI: 0.044 to 0.095], corresponding to a mean increase over the first 8 years of after diagnosis of 0.55 log copies/106 PBMCs. Again, the HIV DNA rate of change did not differ by gender or age (Fig. 1).
 
Loss of LTNP Status
 
In the 60 LTNPs, the median duration of follow-up from the date of diagnosis was 16 years (range: 8-23 years). After the first 8 years of known infection, 36 of the 60 LTNPs lost this status due to a decline in CD4 cell count <500 cells/mm3 in 33 cases, of whom 18 subsequently initiated ARV. The other 3 patients lost their LTNP status as a B-group illness (hairy leukoplasia) were diagnosed in 2 cases, and an AIDS-defining illness (Kaposi sarcoma) was diagnosed in 1. Using Kaplan-Meier estimates, the loss of LTNP status occurred within a median time of 11 years (IQR: 9-19) after diagnosis, that is, 3 years after LTNP diagnosis. Of note, among the 7 (12%) LTNPs, who also met the HIV-controller definition,24 3 nonetheless lost their LTNP status while being HIV controller, as their CD4 count dropped below the threshold of 500 cells/mm3.
 
Among baseline factors, only the HIV DNA level was significantly associated with loss of LTNP status: patients with a baseline HIV DNA level above detection limit were 3.4-fold more likely to lose their LTNP status than patients with an HIV DNA level below this limit (Table 3). Although based on a small sample of patients, the median time to lose LTNP status, using Kaplan-Meier estimates, was 10 vs. >15 years in patients with baseline HIV DNA above and below the detection limit, respectively (Fig. 2). The HIV DNA rate of change was also associated with the loss of LTNP status: subjects with the highest HIV DNA increase (>0.103 log copies/106 PBMCs/year) were 2.7-fold more likely to lose their LTNP status. Baseline HIV RNA tended to be lower in LTNPs who maintained their status than in LTNPs who lost their status [median (IQR): 2.86 (2.30-3.38) and 3.20 (2.30-3.70) log copies/mL, P = 0.22], although not significant. HIV RNA rate of change was not associated with the loss of LTNP status. Women tended to lose less rapidly their LTNP status than men (P = 0.07).
 
In a multivariate model, the baseline HIV DNA level and the HIV DNA rate of change remained independently associated with loss of LTNP status. A baseline HIV DNA above the detection limit compared with below and a high increase (>0.103 log copies/106 PBMCs/year) of HIV DNA level during the first 8 years of documented HIV infection both increased the risk of loosing LTNP status [adjusted hazard ratio (95% CI): 2.8 (1.2 to 6.8) and 2.2 (1.0 to 4.8), respectively). It nevertheless must be reminded that this analysis is based on 60 LTNP patients, this small sample size inducing a limited statistical power.
 
PATIENTS AND METHODS
 
Since 1988, the multicentre French ANRS SEROCO/HEMOCO cohort has enrolled 1748 HIV-infected patients, with a known date of HIV infection or with a first positive enzyme-linked immunosorbent assay (ELISA) test, dating back from less than 1 year while free of symptoms.10,11 Patients who presented with a negative and positive HIV-1 antibody test or an incomplete and complete Western blot tests separated by less than 24 months and enrolled within 24 months after infection were referred to as seroconverters, in agreement with definition used in seroconverters cohorts.12,13 In these seroconverters, the date of infection was assumed to be the middle of the interval between the last negative and first positive HIV-1 test or the date of incomplete Western blot minus 1 month. Patients were followed prospectively and attended scheduled visits every 6 months completed by biological exams; moreover, serum (every 6 months) and cell (every 18 months) samples were stored. The cutoff date for this analysis was April 25, 2006.
 
Patients were eligible for this analysis if they were aged >15 years at enrollment, had a first HIV-positive test and were enrolled before April 25, 1998 (to ensure a minimum follow-up of 8 years), and had a CD4 cell count >500 cells/mm3 at enrollment while naive of ARV. In total, 664 patients met these criteria.
 
LTNPs were defined as patients who remained asymptomatic for at least 8 years after the first HIV-positive test, with CD4+ cell count being always ≥500 cells/mm3 without ARV. On this basis, 60 patients were considered as LTNPs, whereas the remaining 604 were considered as progressors.
 
CD4 Cell Counts and Virological Markers
 
From inclusion in the cohort, all patients had CD4 cell counts performed routinely at each visit by flow cytometry using standard procedures.
 
Plasma HIV RNA was measured routinely on sites at each visit since 1996. The technique used was Cobas Amplicor HIV Monitor test [Roche Molecular Systems, Neuilly sur Seine, France; detection limit of 400 copies/mL (2.60 log copies/mL)] in 78% of cases, Quantiplex HIV-RNA 3.0 (bDNA) [Bayer Diagnostics, Puteaux, France; detection limit of 500 copies/mL (2.70 log copies/mL)] in 21% of cases, and the ANRS real-time polymerase chain reaction (PCR) assay14 in 1% (11 measurements). Frozen serum samples taken before 1996 have been tested retrospectively; HIV RNA was measured using the Cobas Amplicor HIV Monitor test in 90% of cases and using the COBAS AmpliPrep/COBAS TaqMan HIV v1.5 test [detection limit of 50 copies/mL (1.70 log copies/mL)] in the remaining 10%.15
 
HIV DNA was measured in 3 batches at the same laboratory. Aliquots of frozen PBMC, prepared by Ficoll-Hypaque separation, were used to quantify total HIV DNA in PBMC. Total DNA was first estimated and then 0.5 _g was used with the ANRS real-time PCR assay. The detection limit was at 70 copies/106 PBMCs (1.85 log copies/106 PBMCs) as previously reported,16 corresponding to a detection limit of 5 copies per PCR.
 
Statistical Methods
 
Baseline characteristics in LTNPs and progressors were compared using student t test for continuous covariates and a _2 test for categorical covariates. Logistic regression was used to identify factors associated with LTNP status. Factors considered were sex, age at inclusion, seroconversion status, and immunovirological markers at enrollment: CD4 cell counts, CD4 percentage, CD8 cell counts (categorized according to the 25th, 50th, and 75th percentiles), HIV RNA, and HIV DNA [categorized as below or above the detection limits of 2.60/2.70 log copies/mL (ie, 400/500 copies/mL) and 1.85 log copies/106 PBMCs (ie, 70 copies/106 PBMCs), respectively].
 
In LTNPs, immunological and virological marker evolutions (CD4 cell count, HIV RNA, and HIV DNA) during the first 8 years of known infection were characterized using a joint mixed-effect model to take into account the correlation between a patient's measurements and to allow individual trajectories. Each of the 3 markers' evolution was modeled using a linear function of time including a random intercept and a random slope, baseline being HIV diagnosis. In this analysis, measurements within 6 months of seroconversion were not considered to study the markers' progression remote from the large variations occurring during the primary infection period. HIV RNA and HIV DNA below the detection limit were set to half the detection limit.17,18 A square root transformation was applied to the CD4 cell count to conform model assumption.19-21 The influence of gender22 and age at inclusion23 (in 2 categories based on the median) was investigated by adding interaction with the fixed slope and fixed intercept and tested using a likelihood ratio test.
 
In LTNPs, a Cox proportional hazard model was used to identify factors associated with loss of LTNP status. Prognostic factors considered were factors measured at enrollment (categorized based on the median) and the individual rates of change in biological markers during the first 8 years of known infection (introduced as baseline factors), estimated from the joint model, the reference category being values within the interquartile range (IQR). Factors associated with loss of LTNP status in univariate analysis (P < 0.20) were included in a multivariate model. The survival time was defined as the interval between HIV diagnosis and the loss of LTNP status (CD4 cell count < 500 cells/mm3 or diagnosis of an HIV-related infection) or the last follow-up visit in patients who remained LTNPs. Follow-up was censored at the date of ARV initiation when patients initiated treatment with CD4 >500 cells/mm3. For each covariate, the proportional hazard assumption was checked based on Schoenfeld scaled residuals.
 
Statistical analysis were performed using SAS software (SAS Institute Inc, Cary, NC), and results were considered significant if P < 0.05.
 
 
 
 
  iconpaperstack view older Articles   Back to Top   www.natap.org