Introduction
HIV RNA is a strong marker for the risk of HIV disease progression [1–3] [4–6] [7] undetectable viraemia . The sustainability of this phenomenon was associated with low baseline HIV DNA, but not with baseline CD4 cell count. The power of the study, however, was low since data from only 36 patients with undetectable viraemia were analysed.
Here the larger CASCADE seroconverter collaboration has been used to improve the identification of factors associated with achieving a sustainable undetectable viral load without ART. Progression of CD4 cell count during and after the period of undetectable viraemia is described and its relationship to the duration of the period of undetectable viraemia is examined. Finally, the long-term clinical prognosis of patients who spontaneously achieved undetectable viraemia during their follow-up is studied before and during the highly active antiretroviral therapy (HAART) era.
Methods
Patients
CASCADE is an international collaboration that brings together data from 22 HIV-seroconverter cohorts in Europe, Canada and Australia. Patients included in the CASCADE database are patients for whom the date of seroconversion can be reliably estimated. As of January 2004, 8332 seroconverters were included. Details can be found elsewhere [8]
This analysis included patients ≥ 15 years of age whose seroconversion date was estimated through a negative and a positive test separated by less than 2 years or through laboratory evidence of seroconversion. The patients also had to be entered into a cohort within 2 years after the estimated seroconversion date and to have at least two viral load measurments prior to starting any antiretroviral therapy.
Laboratory methods
CD4 cell measurements were performed by flow cytometry. HIV-1 RNA was quantified most of the time (54.5%) using the Amplicor HIV-1 monitor test (Roche Diagnostics, Branchburg, New Jersey, USA); the Quantiplex HIV-1 RNA (b-DNA) assay (Chiron Diagnostics, Emeryville, California, USA) was used in 20.7% of assays, Nuclisens (bioMerieux, Durham, North Carolina, USA) was used in 3.8%, in-house techniques in 2.7%, and not specified in the remaining 18.3%. Since the detection level of the unspecified method was recorded, these last measurements were retained in the analysis. For 171 (7.2%) patients, viral load assessments were performed using techniques with a detection limit > 500 copies/ml; these were excluded from analyses.
Definitions
A patient was considered to achieve undetectable viraemia spontaneously if at least two consecutive HIV RNA measurements were < 400/500 copies/ml while never having received ART. Furthermore, the time interval between any two consecutive measurements had to be at most 18 months. This time interval restriction was introduced to prevent an overestimation of the duration of the undetectable viraemia period through missing values. The period of undetectable viraemia was the period of time during which consecutive viral load measurements remained below the threshold. Multiple periods of undetectable viraemia could occur in the same patient.
The period of undetectable viraemia was assumed to have ended when any of the following occurred: (a) the end of follow-up, (b) HIV RNA above the threshold, (c) a time interval to the following HIV RNA measurement longer than 18 months or (d) the initiation of ART.
The duration of undetectable viraemia was estimated using two definitions: a restrictive definition that did not allow any increase of HIV RNA above the detection limit (definition 1), and a less-restrictive definition that allowed transient increase of HIV RNA (< 3.7 log copies/ml over a period not exceeding 4 years) (definition 2). Using definition 1, the duration of undetectable viraemia strictly corresponded to the duration of the first period, and using definition 2 the period of undetectable viraemia lasted from the beginning of the first period to the end of the last period observed.
Looking closely at the HIV RNA trends before spontaneous control was apparently achieved, 18 patients exhibited a steep decrease (median > 2 log copies/ml over 4 months) more than 4 years after seroconversion. Since these patients were suspected of having received ART without it being reported, they were excluded from the analysis.
Statistical methods
Seroconverters who achieved undetectable viraemia were compared with the other seroconverters eligible for the analysis using a logistic regression model.
Progression to AIDS and to death was compared for patients who achieved or did not achieve undetectable viraemia based on incidence rates. Incidence rates were estimated both in the pre-HAART era (i.e. before January 1996), and in the HAART era. These analyses were performed using SAS software version 8.2 (SAS Institute, Cary, North Carolina, USA).
A joint model was used to describe the CD4 cell count progression during and after the period of undetectable viraemia , and the association between the progression in CD4 cell count during this period and the duration of the period [9] [10,11] [12] undetectable viraemia was assumed to be exponentially distributed [9] [13,14] undetectable viraemia was investigated. A Bayesian approach was used to estimate the joint model parameters using the free Winbugs software version 1.4 (Cambridge University, UK).
Results
Description of seroconverters
In this analysis, 2176 seroconverters were considered, of whom 145 (6.7%) achieved undetectable viraemia during follow-up, while naive for antiretroviral therapy (Table 1 ). In these patients, the median year of inclusion in a cohort was 1993 [interquartile range (IQR) 1990–1996]. There was a trend towards slightly earlier availability of viral load measurements during the course of infection in patients achieving undetectable viraemia than in patients not achieving undetectable viraemia , 10 and 14 months, respectively; nevertheless, this difference was not statistically significant (P = 0.08). However, as could be expected, the follow-up time without ART was significantly longer in seroconverters achieving undetectable viraemia than in the others (median, 79 months versus 41 months, respectively; P < 0.001).
Table 1: Factors associated with spontaneous achievement of undetectable viraemia after seroconversion– the CASCADE collaboration.
At enrolment, seroconverters achieving undetectable viraemia exhibited higher CD4 cell counts and lower viral load than those not achieving undetectable viraemia [median, 743 versus 585 cells/μl (P < 0.0001) and 2.30 versus 4.38 log copies/ml (P < 0.0001), respectively]. Moreover, 57.4% of the seroconverters achieving undetectable viraemia had HIV RNA at enrolment below the detection limit, whereas only 2.7% of the other seroconverters did.
Among the 145 seroconverters achieving undetectable viraemia , the median delay from seroconversion to the beginning of the undetectable viraemia period was 23.3 months (IQR, 11.2–52.8). In 30 patients, the delay from seroconversion to the beginning of undetectable viraemia was longer than 60 months; in 14 of them detectable HIV RNA were observed before viraemia became undetectable (97% of the detectable viral loads < 20 000 copies/ml; maximum, 192 000 copies/ml), whereas in the 16 other patients no HIV RNA measurement was available between enrolment and the beginning of the undetectable viraemia period.
Women were more likely than men to achieve undetectable viraemia [crude odds ratio (OR), 2.35; 95% confidence interval (CI), 1.63–3.37], which explained why persons exposed through sex between men and women, compared with homosexual men, were also more likely to achieve undetectable viraemia in crude analysis (crude OR, 2.61; 95% CI, 1.74–3.90). Seroconverters with a symptomatic primary infection were less likely to achieve undetectable viraemia (OR, 0.52; 95% CI, 0.32–0.85). Seroconverters with a baseline CD4 cell count > 800 cells/μl (third quartile of the distribution) or a baseline viral load < 3.68 log copies/ml (first quartile of the distribution) were more likely to achieve undetectable viraemia than the others. Patients who seroconverted in 1986 or before were more likely to control their viraemia than those who seroconverted in 1987–1990 (OR, 2.00; 95% CI, 1.07–3.11). However, the year of seroconversion was no longer associated with achieving undetectable viraemia when the analysis was restricted to patients who had an available HIV RNA measurement at enrolment. Age at seroconversion and ethnicity were also considered but were not associated with achieving undetectable viraemia .
In a multivariate analysis (Table 1 ), sex and symptomatic primary infection were independently associated with achieving undetectable viraemia . Moreover, a high CD4 cell count at enrolment was strongly associated with achieving undetectable viraemia . No adjustment was made for viral load at enrolment since this factor was too dependent on the event of interest. Sensitivity analyses, considering only patients in whom the first available viral load was measured within 2.5 years from seroconversion, and only patients who seroconverted after 1989, found essentially similar effect estimates.
Duration of undetectable viraemia
Among the 145 seroconverters achieving undetectable viraemia , 133 had a single period of undetectable viraemia . In 12 seroconverters, two periods of undetectable viraemia were observed, separated by detectable viral loads that never exceeded 3800 copies/ml (median, 950); the median time interval between the two periods of undetectable viraemia was 9.5 months (range, 5.6–36.8).
The median duration of the period of undetectable viraemia was 9.5 months (IQR, 4.6–23.6) using definition 1, and 11.2 months (IQR, 5.2–26.7) using definition 2; the longest period of undetectable viraemia was 88.0 months for both definitions. The median delay from seroconversion to the end of undetectable viraemia was 3.8 years (IQR, 1.9–6.1) using definition 1 and 4.0 years (IQR, 2.0–6.5) using definition 2; the longest delay was 18.7 years, for both definitions. For further analyses, the duration of undetectable viraemia using definition 2 was used.
In 75 seroconverters out of the 145, the duration of undetectable viraemia ended through detection of a viral load above the limit. In the remaining 70, the period of undetectable viraemia was interrupted in nine patients when antiretroviral therapy was initiated, and in another 16 where no viral load measurement was available for more than 18 months. Viral load remained below the detection limit until the end of follow-up in 45 patients.
Progression to AIDS and death
In the 145 patients who achieved undetectable viraemia , 10 developed an AIDS-defining illness during their follow-up. In three patients, AIDS occurred during the period of undetectable viraemia . In the 2031 seroconverters who did not achieve undetectable viraemia , 298 developed an AIDS-defining illness during their follow-up. Considering the follow-up for the pre-HAART era, the incidence rate of AIDS was significantly lower in seroconverters who achieved undetectable viraemia than in the others [0.66/100 person-years (95% CI, 0.01–1.39) and 3.43/100 person-years (95% CI, 2.91–3.940), respectively; P = 0.002]. During the HAART era, the incidence rate of AIDS tended to be lower in seroconverters who achieved undetectable viraemia than in the others, but the difference was not statistically significant [1.14/100 person-years (95% CI, 0.30–1.98) and 1.87/100 person-years (95% CI, 1.56–2.19), respectively; P = 0.19]. In seroconverters who achieved undetectable viraemia , the incidence rate of AIDS remained unchanged between the two periods (P = 0.42), whereas it decreased in the other seroconverters during the HAART era (P < 0.001).
In the 145 patients who achieved undetectable viraemia , seven deaths occurred, among which two were AIDS related. In the 2031 other seroconverters, 219 patients died, of whom 68 deaths were AIDS related. During the pre-HAART era, the death rate was significantly lower in seroconverters who achieved undetectable viraemia than in the others [0.65/100 person-years (95% CI, 0.01–1.39) and 2.13/100 person-years (95% CI, 1.73–2.53), respectively]. During the HAART era, the death rate remained lower in seroconverters achieving undetectable viraemia compared with the other patients [0.63/100 person-years (95% CI, 0.01–1.25) and 1.47/100 person-years (95% CI, 1.20–1.74), respectively], although the difference was of borderline significance (P = 0.09). The death rate remained unchanged between the two periods in seroconverters who achieved undetectable viraemia (P = 0.96), whereas in the other seroconverters the death rate significantly dropped during the HAART era (P = 0.006).
Longitudinal modelling of CD4 cell count and analysis of the duration of undetectable viraemia
A joint model was fitted to the data to characterize simultaneously the CD4 cell count progression (on a square root scale) during and after the period of undetectable viraemia and the association between the CD4 cell count progression during the period of undetectable viraemia and the duration of this period (Tables 2 and 3 ).
Table 2: Joint analysis of CD4 cell progression (on a square root scale) during and after the period of undetectable viraemia , and duration of the period of undetectable viraemia in 145 patients from the CASCADE collaboration (longitudinal mixed effect model).
Table 3: Joint analysis of CD4 cell progression (on a square root) during and after the period of undetectable viraemia , and duration of the period of undetectable viraemia in 145 patients from the CASCADE collaboration (survival model).
During the period of undetectable viraemia , the mean rate of change per month in the CD4 cell count was +0.11 cell/μl on a square root scale; this rate of change corresponded to an increase of about 70 cells/months for a patient with 700 cells/μl at the beginning of the period of undetectable viraemia . Through this analysis we identified six patients with an unexpected large CD4 cell count increase whereas the other patients presented either stable or weakly decreasing CD4 cell counts during the period of undetectable viraemia . In three of these six patients, the period of undetectable viraemia began shortly after seroconversion, and the CD4 cell count increase probably corresponded to the period after the primary infection. The CD4 cell count trajectories in these six patients were widely different from the trajectories in the other patients and led to an overestimation of the CD4 cell count progression. A second analysis was conducted that excluded these patients. Of note, the other parameter estimates were not modified by excluding these patients.
In this second analysis, during the period of undetectable viraemia , CD4 cell counts remained stable on average. Nevertheless, five patients presented decreasing counts during the period of undetectable viraemia ; these decreases ranged from 5 to 15 cells/μl per month. The rate of change during the period of undetectable viraemia was not influenced by gender, age or symptomatic primary infection.
The estimated mean CD4 cell count at the beginning of the undetectable viraemia was 705 cells/μl (95% CI, 623–794). At the beginning of the period of undetectable viraemia , patients who reported a symptomatic primary infection exhibited a significantly lower CD4 cell count than the others (mean, 535 cells/μl; 95% CI, 421–663; P = 0.004). The CD4 level at the beginning of the period of undetectable viraemia did not differ according to gender or age group.
After the period of undetectable viraemia , the CD4 cell count significantly decreased with a mean rate of change per month of 0.12 cells/μl on a square root scale, corresponding to a mean rate of change of 5 cells/μl per month for a seroconverter with 700 cells/μl at the end of the period of undetectable viraemia . The rate of change after that period was not influenced by gender, age or symptomatic primary infection.
Adjusted for the individual intercept and rates of change in CD4 cell count, sex and age, the duration of the period of undetectable viraemia tended to be shorter in patients who reported a symptomatic primary infection than in others [adjusted relative risk (RR), 1.91; 95% CI, 0.90–4.11]. Sex, age at seroconversion, ethnicity and risk category were not associated with the duration of the period of undetectable viraemia . A higher CD4 cell count at the beginning of the period of undetectable viraemia tended to be associated with a longer period of undetectable viraemia (adjusted RR, 0.95 for a 1 unit increase on a square-root scale; 95% CI, 0.90–1.01). The rate of change in CD4 cell count during the period of undetectable viraemia was associated with the duration of this period; a rate of change of −0.1 unit/month (on a square root scale) was associated with an adjusted RR of 0.62 (95% CI, 0.48–0.80).
To study the effect of the delay required to control viraemia on both CD4 cell count progression and the duration of the period of undetectable viraemia , the analysis was restricted to patients in whom the delay was estimated with enough precision (period of undetectable viraemia started < 24 months after seroconversion or after the last detectable viral load). In the 107 patients selected, the median delay to control viraemia was 19 months (IQR, 9–38). The delay to control viraemia, defined in two categories according to the median, was not associated with the duration of the undetectable period, and the CD4 cell count progression was not influenced by the delay to control viraemia. Moreover, parameter estimates were essentially similar to those found in the previous analysis.
Discussion
Among the 2176 seroconverters considered in this study of a subgroup of the CASCADE collaboration, 145 (6.7%) achieved undetectable viraemia on at least two consecutive measurements while not taking ART. Women were more likely to achieve undetectable viraemia than men, which agrees with previous findings showing that women exhibit lower viral load levels soon after seroconversion [14] undetectable viraemia . This is unlikely to be explained by a change in the virulence of the strains or by problems with testing stored samples. More likely, rapid progressors were underrepresented in this earliest period [15]
As expected, during the pre-HAART era, the AIDS and death rates of 0.66 and 0.65/100 person-years, respectively, were low among the seroconverters achieving undetectable viraemia , compared with 3.43 and 2.13/100 person-years, respectively, in the patients not achieving undetectable viraemia . These rates were also much lower than the rates observed in patients with high CD4 cell counts (> 350 cells/μl) from the EuroSIDA study since 1998 [16] undetectable viraemia , the AIDS and death rates were similar to the rates estimated before the HAART era; as expected, in seroconverters who did not achieve undetectable viraemia , AIDS and death rates fell after 1996. However, during the HAART era, seroconverters achieving undetectable viraemia still kept an advantage compared with the others, since they tended to exhibit lower AIDS and death rates. In further analysis, it would be interesting to investigate whether these patients have an advantage over those who achieve undetectable viral load under HAART.
Seroconverters who achieved undetectable viraemia generally presented with a high CD4 cell count level at the beginning of the period (705 cells/μl on average), although seroconverters who reported a symptomatic primary infection had a lower CD4 cell count level. As expected, during the period of undetectable viraemia , the CD4 cell count remained stable in most patients. Once the control of viraemia was interrupted, the CD4 cell count decreased significantly, with a mean rate of approximately 5 cells/μl per month, which is similar to that described in natural history during the first years following seroconversion [2] undetectable viraemia , did not differ between men and women.
The joint model showed that a higher CD4 cell count level at the beginning of the period of undetectable viraemia tended to be associated with longer duration of this period. A symptomatic primary infection tended to shorten the period, independently of the CD4 cell count level at the beginning of the undetectable viraemia period, confirming, even in this highly selected population, the role of a symptomatic primary infection on subsequent disease progression. The joint model enabled us to investigate the association between the individual rate of change in CD4 cell count during the period and the duration of this period. Not surprisingly, a larger CD4 cell count decrease was associated with a shorter period of undetectable viraemia . We did not find evidence that a short delay to control viraemia favoured the upholding of undetectable viraemia .
In at least 10% of the patients, the period of undetectable viraemia started quite late, 60 months or more after seroconversion, and was preceded by detectable measurements. This confirms that the ‘viral set-point’, if it exists, is reached in some patients only several years after infection. The period of undetectable viraemia lasted for a median of about 1 year but ranged from 0.3 months up to 13.5 years. Considering the delay from seroconversion, the control of viraemia lasted for a median of 4 years and 14 seroconverters spontaneously controlled their viraemia without ART for more than 10 years after infection. These results from a large multicohort collaboration strengthen the evidence for the existence of a small number of patients with prolonged spontaneous control of viral replication, the HIV controllers [17]
Sponsorship: CASCADE is funded through a grant from the European Union (QLK2–2000–01431) and has received additional funding from GlaxoSmithKline.
References
1. Mellors JW, Rinaldo CR, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 1996; 272:1167–1170.
2. 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 Internal Med 1997; 126:946–954.
3. Hubert J-B, 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. AIDS 2000; 14:123–131.
4. Harrigan PR, Whaley M, Montaner JSG. Rate of HIV-1 RNA rebound upon stopping antiretroviral therapy. AIDS 1999; 12:F59–F62.
5. Garcia F, Plana M, Vidal C, Cruceta A, O'Brien WA, Pantaleo G,
et al . Dynamics of viral load rebound and immunological changes after stopping effective antiretroviral therapy. AIDS 1999; 13:F79–F86.
6. Lefrère J-J, Mariotti M, Morand-Joubert L, Thauvin M, Roudot-Thoraval F. Plasma human immunodeficiency virus RNA below 40 copies/ml is rare in untreated patients even in the first years of infection. J Infect Dis 1999; 180:526–529.
7. Madec Y, Boufassa F, Rouzioux C Delfraissy JF, Meyer L.
Undetectable viraemia without antiretroviral therapy in patients with HIV seroconversion: An uncommon phenomenon?Clin Infect Dis 2005;
40 :1350–1354.
8. CASCADE Collaboration.
Changes in the uptake of antiretroviral therapy and survival in people with known duration of HIV infection in Europe: results from CASCADE. HIV Med 2000;
1 : 224–231.
9. Guo X, Carlin BP. Separate and joint modelling of longitudinal and event time data using standard computer packages. Am Stat 2004; 58:1–9.
10. CASCADE Collaboration.
Difference in CD4 cell counts at seroconversion and decline among 5739 HIV-1 infected individuals with known date of seroconversion. J Acquired Immune Defic Syndr 1999;
32 :303–310.
11. Bhaskaran K, Pillay D, Walker AS, Fisher M, Hawkins D, Gilson R, for the UK Register of HIV Seroconverters. Do patients who are infected with drug-resistant HIV have a different CD4 cell decline after seroconversion? An exploratory analysis in the UK Register of HIV Seroconverters. AIDS 2004; 18:1471–1473.
12. Touloumi G, Babiker AG, Kenward MG, Pocock SJ, Darbyshire JH. A comparison of two methods for the estimation of precision with incomplete longitudinal data, jointly modelled with a time-to-event outcome. Stat Med 2003; 22:3161–3175.
13. Delmas M-C, Jadand C, de Vincenzi I, Deveau C, Persoz A, Sobel A,
et al . Gender differences in CD4+ cell counts persist after HIV-1 infection. AIDS 1997; 11:1071–1073.
14. Prins M, Meyer L, Hessol N. Sex and the course of HIV infection in the pre- and highly active antiretroviral therapy eras. AIDS 2005; 19:357–370.
15. Cascade Collaboration.
Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: a collaborative re-analysis. Lancet 2000;
355 :1131–1137.
16. Mocroft A, Ledergerber B, Katlama C, Kirk O, Reiss P, d'Arminio Monforte A,
et al . Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet 2003; 362:22–29.
17. Lambotte O, Boufassa F, Madec Y, Nguyen A, Goujard C, Meyer L,
et al .
HIV-controllers: a homogeneous group of HIV-1 infected patients with a spontaneous control of viral replication. Clin Infect Dis 2005;
41 :1054–1056.
Appendix
CASCADE collaboration Steering Committee: Valerie Beral, Roel Coutinho, Janet Darbyshire (Project Leader), Julia Del Amo, Noël Gill (Chairman), Christine Lee, Laurence Meyer, Giovanni Rezza.
Coordinating Centre: Kholoud Porter (Scientific Coordinator), Abdel Babiker, A. Sarah Walker, Janet Darbyshire, Freya Tyrer.
CASCADE Collaborators: Aquitaine cohort, France: Francois Dabis, Rodolphe Thiébaut, Geneviève Chêne, Sylvie Lawson-Ayayi; SEROCO cohort, France: Laurence Meyer, Faroudy Boufassa; German cohort: Osamah Hamouda, Gabriele Poggensee; Italian Seroconversion Study: Benedetta Longo, Patrizio Pezzotti, Giovanni Rezza; Greek Haemophilia Cohort: Giota Touloumi, Angelos Hatzakis, Anastasia Karafoulidou, Olga Katsarou; Edinburgh Hospital Cohort, UK: Ray Brettle; Madrid cohort, Spain: Julia Del Amo, Jorge del Romero; Amsterdam Cohort Studies among Homosexual Men and Drug Users, the Netherlands: Liselotte van Asten, Akke van der Bij, Ronald Geskus, Maria Prins, Roel Coutinho; Copenhagen cohort, Denmark: Court Pedersen; Valencia IDU cohort, Spain: Ildefonso Hernández Aguado, Santiago Pérez-Hoyos; Oslo and Ulleval Hospital cohorts, Norway: Anne Eskild, Johan N. Bruun, Mette Sannes; Royal Free haemophilia cohort, UK: Caroline Sabin, Christine Lee; UK Register of HIV Seroconverters, UK: Anne M. Johnson, Andrew N. Phillips, Abdel Babiker, Janet H. Darbyshire, Noël Gill, Kholoud Porter; Swiss HIV cohort: Patrick Francioli, Philippe Vanhems, Matthias Egger, Martin Rickenbach; Sydney AIDS Prospective Study, Australia: David Cooper, John Kaldor; Sydney Primary HIV Infection Cohort, Australia: David Cooper, John Kaldor, Tim Ramacciotti, Don Smith; Badalona IDU hospital cohort, Spain: Roberto Muga, Jordi Tor; Lyon Primary Infection Cohort, France: Philippe Vanhems; South Alberta Clinic, Canada: John Gill; Barcelona IDU cohort, Spain: Joan Cayla, Patricia Garcia de Olalla; MRC Biostatistics Unit, Cambridge, UK: Nicholas E. Day, Daniela de Angelis.
