In the attempt to reduce drug-associated toxicity or delay the onset of side effects that may lead to poor adherence to antiretroviral therapy, several approaches have been explored. One of these approaches is to consider periods in which therapy is interrupted. Broadly, two main interruption strategies have been considered: (i) fixed periods of regular short interruption [1–3] and (ii) longer interruptions specifically in patients with relatively high CD4 cell counts with reinitiation of antiretroviral therapy before the CD4 cell count falls below a certain threshold [4–8]. The theoretical increased risk of resistance development [9–11] and the fact that viral rebound has been frequently observed raise concerns about the first approach . For the second potential strategy, there is relatively little evidence regarding its safety and feasibility from prospectively designed trials, some of which are ongoing, for example the SMART. However, it is clear that many patients in routine clinical practice do interrupt therapy for a variety of reasons [11–13] and it is, therefore, important to try to extract information from observational cohort studies on the possible risks and benefits of this strategy. In particular, it would be useful to identify which patients will be more likely to be able to remain off therapy with high CD4 cell counts for longer periods of time. In order to study these issues, patients were identified with CD4 cell count nadir >250 × 106 cells/l who interrupted therapy when CD4 cell count was > 500 × 106 cells/l. In this same study population, changes in lipid levels during the interruption were also assessed.
Data were derived from 139 patients enrolled in six Italian infectious diseases departments: Modena (n = 23 patients; 2.7% of the on-treatment population), Reggio Emilia (n = 13; 2.9%), Ferrara (n = 11; 2.9%), Rimini (n = 42; 8.5%), Catholic University, Rome (n = 14; 1.3%), Clinic of Infectious Diseases, Milan (n = 5; 0.5%) and one Swedish clinical centre [Gay Men's Health Clinic, South Stockholm General Hospital, Stockholm (n = 31; 8.4%)].
The percentages of patients out of the total enrolled differed among centres because the strategy of treatment interruption was not internationally accepted at the time of the analysis. Therefore, the decision to enrol patients depended on the principal investigator's willing to adopt the treatment interruption strategy at each site.
Patients, to be included in this analysis needed to have interrupted highly active antiretroviral therapy (HAART) in agreement with their treating physician and to satisfy the following inclusion criteria: they had to be 18 years or older, to have received HAART for at least 12 months, to have a CD4 cell count nadir > 250 × 106 cells/l, and a CD4 cell count pre-interruption > 500 × 106 cells/l. HIV plasma viral load at the time of discontinuation should be ≤50 copies/ml. A few previous studies [14,15] showed that patients with a CD4 cell nadir < 250 × 106 cells/l resumed therapy more promptly than those who had a CD4 cell nadir above this value, possibly because of a faster CD4 cell decline during interruption. This study was to evaluate how long a patient who starts treatment at the currently recommended threshold of CD4 cell count could safely remain off treatment, so patients were excluded if they had a CD4 cell nadir < 250 × 106 cells/l. Nevertheless, since the fall in CD4 cell count during treatment interruption can be quite dramatic even in patients with a nadir > 250 × 106 cells/l, patients were only included if they had a CD4 cell count > 500 × 106 cells/l at interruption, as their immediate risk of clinical progression is generally low. Patients treated with immunomodulatory agents or hydroxyurea were excluded. Criteria for restarting treatment were a CD4 cell count < 350 × 106 cells/l, a clinical manifestation of HIV infection, or the patient's desire to resume HAART.
In addition to the immunovirological data, fasting lipids have been routinely measured in the six cohorts in a standardized way. No data on other markers of toxicities induced by antiretroviral drugs, such as liver enzymes, were available for analysis.
The main endpoint of the study was the time (from the date of therapy interruption) to experience a CD4 cell count < 350 × 106 cells/l or to resume therapy, whichever occurred first. If none of these events occurred, follow-up was censored at the date of data freezing for the analysis (1 April, 2003). Kaplan–Meier estimates, log-rank test, and a Cox regression model were used. Predictors considered (all fitted as time-fixed covariates) were gender, age, mode of HIV transmission, HIV plasma viral load at interruption (fitted as continuous), years of antiretroviral therapy (fitted as continuous), months with a HIV plasma viral load ≤ 50 copies/ml before discontinuation (fitted both as continuous and as a categorical variable: never suppressed; 1–12 months or > 12 months), years of use of lamivudine, CD4 cell count nadir, CD4 cell count at interruption, slope of CD4 cell count decay before antiretroviral therapy (all fitted as continuous), and reason for stopping therapy (fitted as categorical). Years of use of lamivudine was included as one of the covariates because it has been shown that the reverse transcriptase mutation 184V, associated with virological failure on lamivudine, alters the catalytic core of the reverse transcriptase enzyme  and, therefore, may affect the processivity of the enzyme and its replication kinetics and could influence the decline of CD4 cell after discontinuation.
The slopes of CD4 cell count were calculated using both a standard least squares estimation approach (i.e., a standard linear regression model separately for each patient) and a robust estimator (i.e. a linear random effect model or hierarchical linear regression) . When estimating the slope of CD4 cell decay during interruption, the immunological follow-up was censored at the time of resuming therapy, assuming that CD4 cell measurements were missing at random .
For patients whose lipids values were above the normal range at the last measurement prior to the treatment interruption, the percentage of patients who, at least in one occasion, experienced a return to a normal value was calculated. This analysis was carried out according to an intention-to-treat principle, so that patients with missing values of these markers were defined as failures.
The rate of clinical events was calculated as the number of AIDS events observed during interruption and the total person-years spent off therapy. Deaths would also have counted as clinical events but none of our patients died over the study period.
Confidence intervals (CI) were constructed using a Poisson distribution.
If viral load was only known to be < 50 copies/ml, because of the technical limitation of the assay used, the imputed value of 50 copies/ml was used for all the analyses.
The analysis focused on 139 individuals who started a HAART regimen with a median CD4 cell count of 417 × 106 cells/l (range, 250–983) and interrupted this regimen with a median CD4 cell count of 800 × 106 cells/l (range, 520–1830). Forty-three patients (30.9%) had a CD4 cell count nadir of 251–350 × 106 cells/l and 96 patients (69.1%) had a nadir > 350 × 106 cells/l. At the time of interruption, patients had been taking antiretroviral therapy for a median of 3.5 years (range, 0.88–11.73).
Other characteristics of these patients at the time of starting this regimen are shown in Table 1. The majority of patients were males (72.7%), acquired HIV through sexual contacts and 58.9% (n = 82) were antiretroviral naive when they started HAART. Patients who had been receiving treatment before starting HAART had been previously exposed to antiretroviral therapy, for 2 years on average. The median calendar time of therapy interruption was March 2001: 15% interrupted in 1998 and 1999, 26% during 2000, 40% during 2001 and the remaining 19% in 2002 (the latest interruption was recorded in November 2002).
All patients were followed prospectively and, on average, CD4 cell count were measured every 3 months (median per year, 4.1; range, 0.33–13.04) over the period of therapy interruption in our study population.
The estimated median slope of CD4 cell count decay before starting antiretroviral therapy by means of the hierarchical regression was 3.87 × 106 cells/l per month [interquartile range (IQR), 3.30–4.34]. The majority of patients had started a HAART regimen containing two nucleoside reverse transcriptase inhibitors and a protease inhibitor (n = 97; 69.8%). The drugs most frequently used within these regimens were zidovudine (n = 66; 47.5%), stavudine (n = 73; 52.5%), lamivudine (n = 102; 73.4%), and indinavir (n = 62; 44.6%). The median CD4 cell increase during HAART and up to the point of treatment interruption was 366 × 106 cells/l (IQR, 232–1350).
At time of interruption, 63 patients (61.2% of 103 patients with an available measurement) had a viral load ≤ 50 copies/ml (i.e., median, 1.7 log10 copies/ml; range, 1.7–4.8). Most of the interruptions were determined by a joint decision between the physician and the patient, who decided to interrupt therapy in order to evaluate the effect of such a strategy for research purposes (n = 68; 48.9%). Other reasons for interruption were toxicity (n = 33; 23.7%), patients’ personal decision (n = 29; 20.9%) and other unknown causes (n = 9; 6.5%). During treatment with HAART, 47 patients (37.3%) never had a viral load ≤ 50 copies/ml, 34 (27.0) experienced viral suppression ≤ 50 copies/ml for less than 1 year and 45 (35.7%) for more than 1 year. For 13 patients, the number of viral load measurements was not sufficient (< 2) to estimate the time spent with suppressed viral load. During HAART, and up to the point of interruption, 17 patients (12.2%) had an increase of < 100 × 106 cells/l or a decrease in CD4 cell count; 11 patients (7.9%) had an increase of 101–200 × 106 cells/l; 37 (26.6%) had an increase of 201–350 × 106 cells/l; 30 (21.6%) had an increase of 351–500 × 106 cells/l; and 44 (31.7%) had an increase of > 500 × 106 cells/l.
Time to resume therapy
Sixty-three patients (45.3%) resumed therapy or experienced a CD4 cell count < 350 × 106 cells/l over a median duration of interruption of 75 weeks (range, 5–248). Of these, 33 patients (52.4%) experienced a CD4 cell decline to < 350 × 106 cells/l and 30 (47.6%) had restarted therapy before their CD4 cell count had fallen below this level. At the time of therapy resumption, of a total of 54 patients who restarted therapy: 24 (44.4%) had a CD4 cell count of 201–350 × 106 cells/l, 17 (32.1%) of 351–500 × 106 cells/l, 13 (24.5%) of > 500 × 106 cells/l, and for one patient the CD4 cell count was not available. Therefore, only a minority of nine patients did not restart therapy despite a CD4 cell count < 350 × 106 cells/l (27% of 33 patients who experienced the CD4 cell fall). However, the median CD4 cell count of these patients at the time of the ‘event’ was relatively high (median 330 × 106 cells/l; range, 260–344) and for 66% their follow-up was censored within 1 year from the date of the event (median, 36 weeks; range,11–206).
The Kaplan–Meier estimate of the median time to therapy resumption/CD4 cell decrease < 350 × 106 cells/l was 117 weeks (95% CI, 79–169). Moreover, on average, patients who did not experience an event during the study period had been followed-up for 21.2 months (range, 4.5–56.9); 30% were followed for < 1 year, 29% for 1–2 years, 30% for 2–3 years and the remaining 11% for > 3 years.
From fitting a multivariable proportional hazard model (including the set of time-fixed covariates described in the Methods), it was estimated that the rate of resuming therapy/CD4 cell count decrease < 350 × 106 cells/l was lower in patients with higher CD4 cell count nadir [relative hazard (RH), 0.69 per 100 × 106 cells/l; 95% CI, 0.50–0.97; P = 0.03], in those who had longer duration of viral suppression < 50 copies/ml (RH, 0.02 in those with > 12 months versus never suppressed; 95% CI, 0.002–0.20; P = 0.0008) and in patients with higher viral load at the time of interruption (RH, 0.18 per 1 log10 copies/ml higher; 95% CI, 0.05–0.63; P = 0.008). All the other factors studied failed to show a significant independent association with the outcome (Table 2). In particular, the CD4 cell count at interruption was significantly associated with the outcome in the univariable analysis (RH, 0.88 per 100 × 106 cells/l higher; 95% CI, 0.80–0.99; P = 0.03) but failed to remain significant after adjusting for the CD4 cell count nadir (RH, 0.92 per 100 × 106 cells/l higher; 95% CI, 0.83–1.02; P = 0.12) and the other covariates (RH, 1.05 per 100 × 106 cells/l higher; 95% CI, 0.87–1.28; P = 0.60). Further, the effect of CD4 cell count nadir did not seem to be different according to the CD4 cell count achieved at the time of interruption (test for interaction P = 0.20) (Table 3).
In order to investigate the issue of how long, on average, patients would remain off-therapy with a CD4 cell count > 350 × 106 cells/l if they had started HAART with a CD4 cell nadir of 251–350 × 106 cells/l and interrupted it with a CD4 cell count > 500 × 106 cells/l, the proportion of patients reaching this end point was estimated using the Kaplan–Meier method. It was estimated that, in the group of 43 patients with these characteristics, the median time to therapy resumption or CD4 cell < 350 × 106 cells/l was 61 weeks (95% CI, 47–130; Fig. 1).
When the slope of CD4 cell decay during the interruption was estimated in the whole population and in this last group, respectively, it was found that the median slope was 19.1 × 106 cells/l per month (IQR, 8.4–35.2) and 19.3 × 106 cells/l per month (IQR, 10.2–45.1) when using the least squares estimation approach and 4.1 × 106 cells/l per month (IQR, 3.0–4.5) and 3.9 × 106 cells/l per month (IQR, 2.9–4.3) when using the more conservative hierarchical model. Fig. 2 shows the average CD4 cell decline observed during the treatment interruption in patients with a CD4 cell count nadir of 251–350 × 106 cells/l estimated using a Lowess smoother curve.
Risk of clinical progression during interruption
Only one patient out of 139 developed a clinical event over the follow-up, a contemporary onset of a disseminated herpes simplex infection and a Pneumocystis jiroveci pneumonia. This event occurred 87 weeks after the treatment interruption and at a CD4 cell count of 205 × 106 cells/l. Overall, the patients spent 212 person-years off-treatment, giving an absolute incidence rate of AIDS events occurred during interruption of 1 per 212 person years (i.e., 4.72 per 1000 person years; 95% CI, 4.43–5.02). None of the patients died during the study period. As far as minor HIV-related symptoms were concerned, only six patients developed lymphadenopathy, possibly related to an acute retroviral syndrome, but none of them restarted HAART.
Cholesterol and triglycerides levels before and during interruption
For 39 patients (50.7% of those with a measurement available), the last available value for triglycerides before treatment interruption was above the normal range (i.e., 1750 mg/l). During the interruption, triglycerides returned to a level within the range of normality (i.e. 0–1750 mg/l), on at least one occasion, in 26 (66.7%) of these patients. When the analysis was repeated using the last measurement available (and defining patients with missing values as failures), this percentage was reduced to 30.8%.
Thirteen patients (16.3% of those with a measurement available) had cholesterol levels increased above normal (2500 mg/l) before interrupting HAART. In 10 (76.9%) of these patients, cholesterol levels returned to normal values (1400–2500 mg/l) on at least one occasion during the treatment interruption and eight (61.5%) were still below normal values at the latest measurement available.
Figure 3 shows the trajectories of the cholesterol values measured over the period starting 6 months prior to therapy interruption for each of the 13 patients who had an elevated cholesterol level before interrupting.
In the past few years, the view that antiretroviral therapy must be lifelong has been questioned [17–19]. The strong evidence that those patients with high CD4 cell counts are at low risk of clinical disease, and concern over the long-term toxicities and inconvenience of therapy, have led to several investigations on the effects of periods of treatment interruption in patients with relatively high CD4 cell count and with prolonged viral suppression on therapy. In order to investigate in more detail the potential consequences of therapy interruption in patients starting HAART at the currently recommended CD4 cell count, patients were only included if they had a CD4 cell nadir > 250 × 106 cells/l. As far as the reasons why patients decided to restart treatment in presence of a CD4 cell count > 350 × 106 cells/l, it should be acknowledged that these patients were starting from a CD4 cell count > 500 × 106 cells/l, and that a rapid decline to 350 × 106 cells/l during the interruption was a reason for concern. Indeed, the inclusion of ‘patient's decision’ as one of the criteria for restarting treatment was dictated by this observation.
Moreover, in order to study the association between viral load at interruption and duration of the interruption, patients with uncontrolled viraemia were not excluded. Indeed, more than one third of the patients of our study population who had uncontrolled viral load at the time of interruption never achieved undetectable viral loads. It is conceivable that this was a group of poorly adherent patients, who do occur frequently in clinical practice. Indeed, since the percentage of self-reported non-adherent patients in a large intercohort Italian study was 46.0% , a secondary aim of our study was to test the safety of discontinuing treatment in these patients if their CD4 cell count is > 500 × 106 cells/l.
Our analysis has shown that the duration of CD4 cell-guided interruptions was longer in patients who had a higher CD4 cell count nadir, as previously described [4–7,21], a higher plasma viral load at the time of interruption and a longer period of viral load suppressed < 50 copies/ml while on therapy. CD4 cell count at the time of interruption was not independently associated with the duration of interruption. Most importantly, our data demonstrate that people starting therapy with CD4 cell counts between 250 and 350 × 106 cells/l, the level recommended by the current international guidelines [22,23], can safely interrupt therapy – maintaining CD4 cell counts > 350 × 106 cells/l – for a reasonably long period of time if they had previously attained a CD4 cell count > 500 × 106 cells/l on therapy (after a median of 61 weeks). This finding supports further evaluation of interruption strategies in ongoing randomized trials. Because of our inclusion criteria, we were able to study the potential effect of having a detectable viraemia on the duration of the off-treatment period. We found that the higher the viral load at the time of interruption, the longer the patients remained off-treatment. The reasons underlying this are unclear. We can speculate that a detectable viraemia at time of interruption may indicate low adherence to therapy and that non-adherent patients were also reluctant to restart it. Controlling for adherence could have helped to clarify this issue but, unfortunately, we have no data concerning adherence in this study.
The rate of AIDS events during interruption was extremely low (5 per 1000 person-years), and the only patient who developed AIDS did not restart treatment despite a CD4 cell count largely < 350 × 106 cells/l. The only other symptoms reported were six cases of lymphadenopathy, and all patients decided to stay off-treatment. This fact could mean either that symptoms were mild or that patients were really motivated to stay off treatment.
Our data also showed, even if only with a descriptive analysis carried out on a small subset of patients, that lipid profiles improved during the treatment interruption for the majority of the patients with elevated tryglicerides or cholesterol at the time of interrupting therapy.
A recent study conducted by Tarwater and colleagues , showed results similar to those of our study, especially regarding the prognostic value of the CD4 cell nadir. Their conclusions were that, in order to decrease the total drug exposure, therapy should be initiated with a CD4 cell count > 350 × 106 cells/l and ‘pulse therapy’ could represent an alternative strategy to lifelong treatment. Our results suggest that, even in those who started therapy with a CD4 cell count > 350 × 106 cells/l, so long as viral suppression was maintained for a long period of time and a CD4 cell count > 500 × 106 cells/l was attained, the duration of interruption with CD4 cell count remaining < 350 × 106 cells/l can be remarkably long. Moreover, since previous studies have shown that the probability of virological success in patients starting the same regimen that they had interrupted is relatively high , after each interruption patients could restart the same regimen, thus potentially prolonging the viroimmunological success of the first antiretroviral regimen. Therefore, use of intermittent or pulse therapy may not infer that therapy should be initiated at CD4 cell counts > 250 × 106 cells/l.
This study has several limitations. Since only the first treatment interruption has been investigated, we cannot derive any conclusion on the efficacy and safety of pulse therapy, as the extent of immune reconstitution after several interruptions needs to be investigated.
Although in our study treatment interruptions seemed to be relatively safe, clinicians should carefully counsel patients that, during treatment interruption, (i) the risk of HIV transmission to their partners is higher as a consequence of viral rebound , (ii) HIV related-complications, such as thrombocytopenia, or dermatological problems, such as psoriasis or rosacea, can reappear , (iii) clinical events, ranging from acute retroviral syndrome, reactivation of hepatitis B , oral candidiasis and herpes zoster to major opportunistic infections can also occur  and (iiii) non-nucleoside reverse transcriptase inhibitors have a long plasma half-life and there is the risk of resistance after discontinuation .
In conclusion, our study showed that, with appropriately monitoring, patients who started therapy with CD4 cell count between 250 and 350 × 106 cells/l and who later interrupted therapy appeared able to remain off therapy with CD4 cell count > 350 × 106 cells/l and with minimal risk of developing AIDS for a median of 14 months.
Note: These data were presented in part at the 42nd ICAAC Chicago, September 2003, [abstract H-856] and at the 9th European AIDS Conference (EACS), Warsaw [abstract F11/1].
Sponsorship: This study was partially supported by Programma Nazionale di Ricerca sull’AIDS – Istituto Superiore di Sanità, Italy (Progetto Patologia, Clinica e Terapia dell’AIDS: grants 30D.56 to C Mussini).
1. Dybul M, Nies-Kraske E, Daucher M, Hallahan C, Csako G, Yoder C, et al
. A randomized, controlled trial of long cycle structured intermittent versus continuous ARV therapy for chronic HIV infection. Tenth Conference on Retroviruses and Opportunistic Infections
. Boston, February 2003 [abstract 681b].
2. Ananworanich J, Nuesch R, Le-Braz M, Chetchotisakd P, Vibhagool A, Wicharuk S, et al
. Failures of 1 week on 1 week off antiretroviral therapies in a randomized trial. AIDS 2003; 17:F33–F37.
3. Vella S, Giuliano N, Palmisano L, Bucciardini R, Andreotti M, Arcieri R, et al
. ISS-PART: a prospective, randomized, multi-centre clinical trial of intermittent therapy in HIV+ subjects with persistent suppression of viral replication. Tenth Conference on Retroviruses and Opportunistic Infections
. Boston, February 2003 [abstract 66].
4. Tarwater PM, Parish M, Gallant JE. Prolonged treatment interruption after immunologic response to highly active antiretroviral therapy. Clin Infect Dis 2003; 37:1541–1548.
5. Ruiz L, Gomez L, Domingo P, Romeu J, Tambussi G, Martinez-Picado J, et al
. A multicenter, randomized controlled trial of continuous versus intermittent HAART guided by CD4+ T-cell counts and plasma HIV-1 RNA levels. Tenth Conference on Retroviruses and Opportunistic Infections
. Boston, February 2003 [abstract 65].
6. Ananworanich J, Cardiello P, Srasuebkul P, Samor T, Hassink E, Mahanontharit A, et al
. HIV-NAT 001.4: a prospective randomized trial of structured treatment interruption in patients with chronic HIV infection. Tenth Conference on Retroviruses and Opportunistic Infections
. Boston, February 2003 [abstract 64].
7. Maggiolo F, Ripamonti D, Gregis G, Quinzan G, Callegaro A, Suter F. Effect of prolonged discontinuation of successful antiretroviral therapy on CD4 T cells: a controlled, prospective trial. AIDS 2004; 18:439–446.
8. Mussini C, Bugarini R, Perno CF, Antinori A, Borghi V, Bertoli A, et al
. Kinetics of CD4 cells after discontinuation of antiretroviral therapy in patients with virological failure and a CD4 cell count greater than 500 × 106
cells/l. AIDS 2002; 16:1551–1554.
9. Metzner KJ, Bonhoeffer S, Fischer M, Karanicolas R, Allers K, Joos B, et al
. Emergence of minor populations of human immunodeficiency virus type 1 carrying the M184V and L90M mutations in subjects undergoing structured treatment interruptions. J Infect Dis 2003; 188:1433–1443.
10. Palmisano L, Giuliano M, Bucciardini R, Galluzzo CM, Andreotti M, Fragola V, et al
. Emergence of resistance mutations during intermittent HAART. Rate, predicting factors and effect on virological response. Eleventh Conference on Retroviruses and Opportunistic Infections
. San Francisco, February 2004 [abstract 552].
11. Phillips AN, Youle MS, Lampe F, Johnson M, Sabin CA, Lepri AC, Loveday C. Theoretical rationale for the use of sequential single-drug antiretroviral therapy for treatment of HIV infection. AIDS 2003; 17:1009–1016.
12. d’Arminio Monforte A, Cozzi-Lepri A, Rezza G, Pezzotti P, Antinori A, Phillips AN, et al
. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients. ICONA Study Group. Italian Cohort of Antiretroviral-Naive Patients. AIDS 2000; 14:499–507.
13. Mocroft A, Youle M, Moore A, Sabin CA, Madge S, Cozzi-Lepri A, et al
. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS 2001; 15:185–194.
14. Boyer PL, Hughes SH. Analysis of mutation at position 184 in reverse transcriptase of human immunodeficiency virus type 1. Antimicrob Agents Chemother 1995; 39:1624–1628.
15. Laird NM, Ware JH. Random effects models for longitudinal data. Biometrics 1982; 38:963–974.
16. Hogg RS, Havlir D, Miller V, Montaner JSG. To stop or not to stop: that is the question but what is the answer? AIDS 2002; 16:787–789.
17. Deeks SG, Hirschel B. Supervised interruptions of antiretroviral therapy. AIDS 2002; 16(suppl. 4):S157–S169.
18. Lori F, Lisziewicz J. Structured treatment interruptions for the management of HIV infection. JAMA 2001; 286:2981–2987.
19. Trotta MP, Ammassari A, Cozzi-Lepri A, Zaccarelli M, Castelli F, Narciso P, et al
, for the Adherence Italian Cohort Naive Antiretrovirals (AdICONA) and Adherence Spallanzani (AdeSpall) Study Groups. Adherence to highly active antiretroviral therapy is better in patients receiving non-nucleoside reverse transcriptase inhibitor-containing regimens than in those receiving protease inhibitor-containing regimens. AIDS 2003; 17:1099–1102.
20. Poulton MB, Sabin CA, Fisher M. Immunological changes during treatment interruptions: risk factors and clinical sequelae. AIDS 2003; 17:126–128.
21. Pozniak A, Gazzard B, Anderson J, Babiker A, Churchill D, Collins S, et al
. British HIV Association (BHIVA) guidelines for the treatment of HIV-infected adults with antiretroviral therapy. HIV Med 2003; 4:1–41.
22. Yeni PG, Hammer SM, Carpenter CC, Cooper DA, Fischl MA, Gatell JM, et al
. Antiretroviral treatment for adult HIV infection in 2002: updated recommendations of the International AIDS Society-USA Panel. JAMA 2002; 288:222–235.
23. Touloumi G, Pocock SJ, Babiker AG, Darbyshire JH. Impact of missing data due to selective dropouts in cohort studies and clinical trials. Epidemiology 2002; 13:347–355.
24. Tubiana R, Ghosn J, De-Sa M, Wirden M, Gautheret-Dejean A, Bricaire F, et al
. Warning: antiretroviral treatment interruption could lead to an increased risk of HIV transmission. AIDS 2002; 16:1083.
25. Ananworanich J, Phanuphak N, Nuesch R, Apateerapong W, Rojnuckarin P, Ubolyam S, et al
. Recurring thrombocytopenia associated with structured treatment interruption in patients with human immunodeficiency virus infection. Clin Infect Dis 2003; 37:723–725.
26. Worthington MG, Ross JJ. Aseptic meningitis and acute HIV syndrome after interruption of antiretroviral therapy: implications for structured treatment interruptions. AIDS 2003; 17:2145–2146.
27. Fagard C, Oxenius A, Günthard H, Garcia F, Le Braz M, Mestre G, et al
. A prospective trial of structured treatment interruptions in human immunodeficiency virus infection. Arch Intern Med 2002; 163:1220–1226.
28. Taylor S, Allen S, Fidler S, White D, Gibbons S, Fox J, et al
. Stop Study: after discontinuation of efavirenz, plasma concentrations may persist for 2 weeks or longer. Eleventh Conference on Retroviruses and Opportunistic Infections
. San Francisco, February 2004 [abstract 131].
Writing Committee: Cristina Mussini, Andrea Bedini, Vanni Borghi, Giovanni Guaraldi and Roberto Esposito (Clinic of Infectious Diseases, University of Modena and Reggio Emilia, Italy), Enrico Barchi (Department of Infectious Diseases, Reggio Emilia, Italy), Alessandro Cozzi-Lepri and Andrew N. Phillips (Royal Free Centre for HIV Medicine and Department of Primary Care and Population Sciences, Royal Free and University College Medical School, London, UK), Patrizia Ortolani (Department of Infectious Diseases, Rimini, Italy), Göran Bratt and Lars E. Eriksson (Gay Men's Health Clinic, South Stockholm General Hospital and Department of Nursing, Karolinska Institutet, Stockholm, Sweden), Laura Sighinolfi (Department of Infectious Diseases, Ferrara, Italy), Andrea Cossarizza (Department of Biomedical Sciences, University of Modena and Reggio Emilia, Italy), Antonella d’Arminio Monforte (Clinic of Infectious Diseases, University of Milan, Italy), Andrea De Luca and Simona Di Giambenedetto (Clinic of Infectious Diseases, Catholic University, Rome, Italy), Andrea Antinori (National Institute for Infectious Diseases L. Spallanzani, Rome, Italy).
© 2005 Lippincott Williams & Wilkins, Inc.