In most HIV-1-infected patients, treatment with combination antiretroviral therapy (cART) results in sustained reduction of plasma HIV RNA levels to lower than the limits of quantification. During follow-up, however, many patients who achieve virus suppression have measurable but transient viremia while on cART.1-6 These episodes of transient viremia have been associated with low-level viral replication,2,7,8 with activation of latently infected cells and subsequent production of virus,8,9 and with a rise in target cell availability (eg, attributable to vaccination or to coinfections).9-11 In contrast to episodes of viremia during short-term therapy interruptions, transient viremia while on cART reportedly has a limited effect on treatment outcome, development of drug resistance, and clinical prognosis.2,3,5,12,13
The objective of this study was to investigate the characteristics of episodes of viremia in a large group of cART-treated patients who had experienced initial therapy success. In addition, we studied clinical, immunologic, and virologic changes that occurred during episodes of viremia and assessed covariates that were associated with the occurrence of viremia.
From the AIDS Therapy Evaluation Project, Netherlands (ATHENA) observational cohort,14 which consisted of 11,912 HIV-1-infected patients by February 2006, previously ART-naive patients receiving cART were selected. Patients were eligible for analysis from the time of achieving initial therapy success, defined by the date of the second of 2 consecutive RNA measurements <50 copies/mL at least 2 weeks and at most 48 weeks apart, allowing for changes in regimen since starting cART. Follow-up ended at the last available RNA measurement or at therapy interruption (ie, a period without any antiretroviral medication), whichever came first.
HIV RNA plasma levels after success were classified into 3 categories: <50 copies/mL, 50 to 1000 copies/mL, and >1000 copies/mL. CD4 and CD8 T-cell counts were assigned to each RNA measurement as the counts, if available, closest in time within 28 days before or after the RNA measurement. A multinomial cumulative logit model was used to assess the probability that the viral load at the next measurement would be in the same or a higher RNA category than the most recently measured viral load, adjusting for the calendar year and RNA category of the most recent RNA measurement, the time since the most recent measurement, the percentage of CD4 cells (rather than CD4 and CD8 cell counts separately to avoid problems of collinearity), age, (non)Western origin, gender, and transmission group.
An episode of viremia started when the viral load rose to a level >50 copies/mL and ended when it returned to <50 copies/mL. Analogously, an episode of viral suppression started with a viral load <50 copies/mL and ended when it rose to >50 copies/mL. By definition, all patients started with an episode of suppression. Episodes of viremia were subdivided into episodes of low-level viremia (50 to 1000 copies/mL) and high-level viremia (>1000 copies/mL). During each episode of suppression or viremia, the occurrence of 4 types of events was assessed, including therapy changes (excluding changes in dosage), the occurrence of Centers for Disease Control and Prevention (CDC) events, the occurrence of adverse events leading to therapy changes or known to be associated with antiretroviral therapy (ART), and evidence of drug resistance. Median CD4 and CD8 cell counts were calculated as the median of the averages of all CD4 and CD8 counts, respectively, measured during an episode. Resistance was assessed by finding all major resistance-associated mutations (RAMs) in reverse transcriptase (RT) and protease sequences obtained during routine clinical care.15 Incidence rates were calculated as the total number of events divided by the total follow-up during each of the 3 types of episodes.
The Poisson distribution was used to calculate 95% confidence intervals (CIs) for rates. Changes over time in proportions were studied using general linear regression modeling. Mixed-effects models with a random intercept and slope were used to analyze changes over time in CD4 and CD8 cell counts. Correlations between longitudinal measurements within patients were taken into account using a first-order autoregressive covariance structure for measurements equidistant in time and a spatial power structure for nonequidistant measurements. Generalizing estimating equations were used to account for correlated measurements, and 95% CIs were estimated using the empiric estimator of the covariance matrix. All analyses were performed using SAS software (version 9.1.3; SAS Institute, Cary, NC).
The study population consisted of 4447 patients with a total follow-up of 11,187 person-years after success. Most patients were men (3432 patients [77.2%]), originated from the Netherlands or other Western countries (2879 [64.7%]), and were infected by means of homosexual (2338 [52.6%]) or heterosexual (1623 [36.5%]) contact. At the time of initial success, median CD4 and CD8 counts were 390 (interquartile range [IQR]: 250 to 570) cells/mm3 and 924 (IQR: 647 to 127) cells/mm3, respectively, and the median age of patients was 39.3 (IQR: 33.4 to 46.1) years. For 3166 (71.2%) patients, RNA measurements were persistently <50 copies/mL, whereas 1281 (28.8%) patients had at least 1 RNA measurement >50 copies/mL and 299 (6.7%) at least 1 RNA measurement >1000 copies/mL.
During follow-up, 36,940 plasma viral load measurements were performed, corresponding to an average of 3.30 (95% CI: 3.27 to 3.34) measurements per person-year of follow-up. Of these measurements, 34,016 (92.1%) were <50 copies/mL, 1082 (2.9%) were between 50 and 100 copies/mL, 868 (2.3%) were between 100 and 400 copies/mL, 266 (0.7%) were between 400 and 1000 copies/mL, and 708 (1.9%) were >1000 copies/mL. Between 1999 and 2005, the proportion of RNA measurements between 50 and 1000 copies/mL decreased (P < 0.001) from 7.8% (66 of 849 measurements) to 5.0% (377 of 7532 measurements), although there was no significant change in the proportion of RNA measurements >1000 copies/mL (P = 0.7): 2.1% (18 of 849 measurements) in 1999 and 1.6% (118 of 7532 measurements) in 2005.
Figure 1 shows the fractions of viral load measurements observed in each of the 3 RNA categories (<50, 50 to 1000, and >1000 copies/mL), given that the previous RNA measurement was at most Δt months before. After 2 to 3 months, the fractions did not change with increasing time intervals Δt. When the time interval was <1 month, however, the probability of observing 2 consecutive viral loads >50 copies/mL was higher, albeit not significantly, compared with longer time intervals. The time between measurements was not associated with the probability of viremia (adjusted odds ratio [OR] = 1.00 [95% CI: 0.97 to 1.02] per month increase in time between measurements). The probability of viremia was associated with a higher RNA level at the most recent measurement (OR = 2.6 [95% CI: 2.4 to 2.8] per 1-log increase if RNA measured 50 to 1000 copies/mL and OR = 3.3 [95% CI: 3.1 to 3.5] if RNA measured >1000 copies/mL) and with a lower percentage of CD4 cells (OR = 0.89 [95% CI: 0.86 to 0.93] per 10% increase). Patients of non-Western origin had a higher probability of viremia (OR = 1.35 [95% CI: 1.18 to 1.55]) than Western patients.
In the total population of 4447 patients, there were 5989 (74.2%) episodes of suppression (10,612 person-years of follow-up), 1711 (21.2%) episodes of low-level viremia (440 person-years of follow-up), and 369 (4.6%) episodes of high-level viremia (135 person-years of follow-up). During the winter months (October to March), 5.2% of the 1090 patients in follow-up between 2002 and 2004 experienced an episode of viremia, whereas this proportion was 4.6% between April and September (P = 0.03). Median CD4 counts were 460 (IQR: 318 to 640) cells/mm3 during episodes of suppression, 480 (IQR: 323 to 680) cells/mm3 during low-level viremia (P = 0.03, compared with episodes of suppression), and 360 (IQR: 250 to 528) cells/mm3 during high-level viremia (P < 0.001, compared with the other episodes of suppression), whereas CD8 counts were 925 (IQR: 680 to 1235) cells/mm3, 1015 (IQR: 750 to 1354) cells/mm3 (P < 0.001, compared with episodes of suppression), and 1040 (IQR: 800 to 1400) cells/mm3 (P ≤ 0.002, compared with other episodes of suppression), respectively. During episodes of suppression, CD4 counts increased at a rate of 34 (IQR: 32 to 36) cells/mm3 per year, whereas CD8 counts decreased at a rate of 19 (IQR: 15 to 24) cells/mm3 per year. CD4 counts decreased at a rate of 87 (IQR: 54 to 120) cells/mm3 per year during high-level viremia but did not change during low-level viremia (P = 0.7). Also, CD8 cell counts did not change during episodes of low-level (P = 0.3) or high-level (P = 0.5) viremia.
Most episodes of low-level viremia were short-lasting because they consisted of only 1 (81.8%) or 2 (11.9%) RNA measurements; 1362 (79.6%) were without a clinical event or therapy change. Episodes of high-level viremia also mainly consisted of 1 (59.1%) or 2 (21.7%) measurements, but only 154 (41.7%) were without an event. During 2568 (42.9%) episodes of suppression, therapy was changed, corresponding to 0.36 (95% CI: 0.35 to 0.38) changes per person-year. Therapy was changed during 237 (13.9%) episodes of low-level viremia and 193 (52.3%) episodes of high-level viremia, corresponding to 0.63 (95% CI: 0.56 to 0.71) and 1.74 (95% CI: 1.52 to 1.98) changes per year, respectively. For 321 (4.0%) episodes of suppression or viremia, therapy was changed in the month before the episode started (P = 0.5 for differences between episodes). The incidences of adverse events were 0.41 (95 % CI: 0.40 to 0.43), 0.57 (95% CI: 0.50 to 0.65), and 0.60 (95% CI: 0.48 to 0.75) events per year during episodes of suppression, low-level viremia, and high-level viremia, respectively. The incidence of CDC events was <0.05 events per year. During 29 (1.7%) episodes of low-level viremia and 111 (30.0%) episodes of high-level viremia, a sequence was obtained. RAMs were found in 22 (76%) and 86 (77%) of these episodes, respectively (P = 0.9), or in 1.3% and 23.3% of all episodes of low-level and high-level viremia, respectively. Of the 29 episodes of low-level viremia with a sequence, 12 (41%) were followed or preceded by high-level viremia, whereas this was the case for only 208 (12.2%) of all 1711 episodes of low-level viremia.
After initial therapy success, transient episodes of viremia were found in 28.8% of our continuously cART-treated patients, whereas high-level viremia was found in 6.8%, in accordance with previous findings.3-6,16 Episodes of low-level viremia were mostly isolated measurements, as would be expected, because the sampling frequency was low and the estimated duration of these episodes is <3 weeks.4,16 The probability of observing (low-level) viremia at the next RNA measurement depended on the level of the most recent RNA measurement and was inversely correlated with CD4 cell counts but was independent of the time between the measurements if they were more than 2 months apart.4,5 These findings suggest that episodes of low-level viremia are not merely random variations but are partially explained by host-related factors such that patients have different tendencies to show blips.4,16
The probability of observing low-level viremia decreased with calendar time and, most likely in correlation, with longer follow-up (data not shown). Apparently, cART regimens have become more potent or easier to adhere to over time, yielding a more complete and more sustained suppression of viral load.7,17 In contrast, the prevalence of high-level viremia did not change over time. This indicates that the mechanism underlying high-level and low-level viremia is different and that high-level viremia is most likely the result of true therapy failure or incomplete adherence. The RNA sampling frequency also declined over time, and episodes of low-level viremia might have been missed, but this has only a limited effect on the measured proportion of viremia. Episodes of high-level viremia were probably missed to a lesser extent, because high-level viremia is less likely to decrease again spontaneously and therefore lasts longer.
It has previously been shown that the increase in CD4 cell counts in patients with short-lasting transient viremia was similar to that observed in patients with suppressed viral load but tended to be lower if viremia were long lasting.3,6,18 We did not observe changes in CD4 cell counts during low-level viremia, most likely because episodes of low-level viremia were short lasting; even when restricting the analysis to low-level viremia lasting at least 2 years, no significant changes were observed (data not shown). CD8 cell counts were higher during viremia than during episodes of suppression. They most likely represent an expanded HIV-specific CD8 T-cell response, although it is not clear whether this is caused by or is a consequence of viremia.13 The higher proportion of patients who were viremic during winter compared with summer, when diseases like influenza and common cold are less prevalent than during winter, suggested that antigenic stimulation attributable to infection with other viruses was associated with transient viremia.6,9-11
Our study showed that episodes of viremia were associated with therapy changes and clinical events. The timing of therapy changes suggested that they were more likely the response of the treating physician to viremia than the cause. Unfortunately, we were unable to determine whether clinical events were the cause or the consequence of viremia, because the date of onset of events was often accurate within 1 month. Another limitation was that resistance profiles were obtained during only a small number of episodes of low-level viremia and were often linked to episodes of high-level viremia. Also, we were not able to distinguish whether plasma virus during low-level viremia originated from activated latently infected cells or from ongoing viral replication.8,12 Further, incomplete adherence has been associated with viremia, but data on adherence were not available.6,16
In conclusion, short-lasting episodes of low-level viremia are relatively frequent, and host-related and seasonal factors are associated with the occurrence of viremia. High-level viremia seems to be frequently associated with resistance and often leads to therapy changes. The low proportion of episodes of low-level viremia that are followed by high-level viremia or have clinical events and the observation that CD4 cell counts remain similar to those observed during episodes of suppression suggest that leaving therapy unchanged during low-level viremia is a clinically acceptable strategy.
1. Zhang L, Ramratnam B, Tenner-Racz K, et al. Quantifying residual HIV-1 replication in patients receiving combination antiretroviral therapy. N Engl J Med
2. Havlir DV, Bassett R, Levitan D, et al. Prevalence and predictive value of intermittent viremia with combination HIV therapy. JAMA
3. Sklar PA, Ward DJ, Baker RK, et al. Prevalence and clinical correlates of HIV viremia (‘blips’) in patients with previous suppression below the limits of quantification. AIDS
4. Di Mascio M, Markowitz M, Louie M, et al. Viral blip dynamics during highly active antiretroviral therapy. J Virol
5. Greub G, Cozzi-Lepri A, Ledergerber B, et al. Intermittent and sustained low-level HIV viral rebound in patients receiving potent antiretroviral therapy. AIDS
6. Easterbrook PJ, Ives N, Waters A, et al. The natural history and clinical significance of intermittent viraemia in patients with initial viral suppression to <400 copies/ml. AIDS
7. Havlir DV, Strain MC, Clerici M, et al. Productive infection maintains a dynamic steady state of residual viremia in human immunodeficiency virus type 1-infected persons treated with suppressive antiretroviral therapy for five years. J Virol
8. Tobin NH, Learn GH, Holte SE, et al. Evidence that low-level viremias during effective highly active antiretroviral therapy result from two processes: expression of archival virus and replication of virus. J Virol
9. Jones LE, Perelson AS. Transient viremia, plasma viral load, and reservoir replenishment in HIV-infected patients on antiretroviral therapy. J Acquir Immune Defic Syndr
10. Gunthard HF, Wong JK, Spina CA, et al. Effect of influenza vaccination on viral replication and immune response in persons infected with human immunodeficiency virus receiving potent antiretroviral therapy. J Infect Dis
11. Fraser C, Ferguson NM, de Wolf F, et al. The role of antigenic stimulation and cytotoxic T cell activity in regulating the long-term immunopathogenesis of HIV: mechanisms and clinical implications. Proc Biol Sci
12. Cohen Stuart JW, Wensing AM, Kovacs C, et al. Transient relapses (“blips”) of plasma HIV RNA levels during HAART are associated with drug resistance. J Acquir Immune Defic Syndr
13. Karlsson AC, Younger SR, Martin JN, et al. Immunologic and virologic evolution during periods of intermittent and persistent low-level viremia. AIDS
14. van Sighem AI, van de Wiel MA, Ghani AC, et al. Mortality and progression to AIDS after starting highly active antiretroviral therapy. AIDS
15. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1: 2005. Top HIV Med
16. Nettles RE, Kieffer TL, Kwon P, et al. Intermittent HIV-1 viremia (blips) and drug resistance in patients receiving HAART. JAMA
17. Ramratnam B, Ribeiro R, He T, et al. Intensification of antiretroviral therapy accelerates the decay of the HIV-1 latent reservoir and decreases, but does not eliminate, ongoing virus replication. J Acquir Immune Defic Syndr
18. Hunt PW, Deeks SG, Rodriguez B, et al. Continued CD4 cell count increases in HIV-infected adults experiencing 4 years of viral suppression on antiretroviral therapy. AIDS
The ATHENA database is supported by a grant from the Dutch Health Minister and was set up and is maintained by the HIV Monitoring Foundation. The physicians and data analysts include (*site coordinating physicians) the following individuals: F. de Wolf (Director), D. O. Bezemer, L. A. J. Gras, A. M. Kesselring, A. I. van Sighem, C. Smit, and S. Zhang (Data Analysis Group), and S. Zaheri (Data Collection), HIV Monitoring Foundation, Amsterdam; W. Bronsveld* and M. E. Hillebrand-Haverkort, Medical Center Alkmaar, Alkmaar; J. M. Prins*, J. Branger, J. K. M. Eeftinck Schattenkerk, J. Gisolf, M. H. Godfried, J. M. A. Lange, K. D. Lettinga, J. T. M. van der Meer, F. J. B. Nellen, T. van der Poll, P. Reiss, Th.A. Ruys, R. Steingrover, G. van Twillert, J. N. Vermeulen, S. M. E. Vrouenraets, M. van Vugt, and F. W. M. N. Wit, Academic Medical Center of the University of Amsterdam, Amsterdam; T. W. Kuijpers, D. Pajkrt, and H. J. Scherpbier, Emma Children's Hospital, Amsterdam; A. van Eeden, Medical Center Jan van Goyen, Amsterdam; K. Brinkman*, G. E. L. van den Berk, W. L. Blok, P. H. J. Frissen, J. C. Roos, W. E. M. Schouten, and H. M. Weigel, Onze Lieve Vrouwe Gasthuis, Amsterdam; J. W. Mulder*, E. C. M. van Gorp, and J. Wagenaar, Slotervaart Hospital, Amsterdam; J. Veenstra*, St. Lucas Andreas Hospital, Amsterdam; S. A. Danner*, M. A. van Agtmael, F. A. P. Claessen, R. M. Perenboom, A. Rijkeboer, and M. G. A. van Vonderen, Free University Medical Center, Amsterdam; C. Richter* and J. van der Berg, Hospital Rijnstate, Arnhem; R. Vriesendorp* and F. J. F. Jeurissen, Medical Center Haaglanden, location Westeinde, Den Haag; R. H. Kauffmann* and K. Pogány, Haga Hospital, location Leyenburg, Den Haag; B. Bravenboer*, Catharina Hospital, Eindhoven; C. H. H. ten Napel* and G. J. Kootstra, Medisch Spectrum Twente, Enschede; H. G. Sprenger*, S. van Assen, and J. T. M. van Leeuwen, University Medical Center Groningen, Groningen; R. Doedens and E. H. Scholvinck, University Medical Center Beatrix kliniek, Groningen; R. W. ten Kate* and R. Soetekouw, Kennemer Gasthuis, Haarlem; D. van Houte* and M. B. Polée, Medical Center Leeuwarden, Leeuwarden; F. P. Kroon*, P. J. van den Broek, J. T. van Dissel, and E. F. Schippers, Leiden University Medical Center, Leiden; G. Schreij*, S. van der Geest, S. Lowe, and A. Verbon, Academic Hospital Maastricht, Maastricht; P. P. Koopmans*, R. van Crevel, R. de Groot, M. Keuter, F. Post, A. J. A. M. van der Ven, and A. Warris, Radboud University Nijmegen Medical Center, Nijmegen; M. E. van der Ende*, I. C. Gyssens, M. van der Feltz, J. L. Nouwen, B. J. A. Rijnders, and T. E. M. S. de Vries, Erasmus Medical Center, Rotterdam; G. Driessen, M. van der Flier, and N. G. Hartwig, Erasmus Medical Center Sophia, Rotterdam; J. R. Juttman*, M. E. E. van Kasteren, and C. van de Heul, St. Elisabeth Hospital, Tilburg; I. M. Hoepelman*, M. M. E. Schneider, M. J. M. Bonten, J. C. C. Borleffs, P. M. Ellerbroek, C. A. J. J. Jaspers, T. Mudrikove, C. A. M. Schurink, and E. H. Gisolf, University Medical Center Utrecht, Utrecht; S. P. M. Geelen, T. F. W. Wolfs, and T. Faber, Wilhelmina Children's Hospital, Utrecht; A. A. Tanis*, Hospital Walcheren, Vlissingen; P. H. P. Groeneveld*, Isala Clinics, Zwolle; J. G. den Hollander*, Medical Center Rijnmond-Zuid, location Clara, Rotterdam; A. J. Duits and K. Winkel, St. Elisabeth Hospitaal/Stichting Rode Kruis Bloedbank, Willemstad, Curaçao.