Treatment Interruption Revisited with LOTTI Study: authors reported
interruptions were not inferior to continuous HAART in this study
The LOTTI Study published in April 2009 AIDS finds CD4-guided treatment interruptions within certain parameters to be equal in clinical outcomes to continuous HAART. Patients had to have >700 CD4s and no history of nadir CD4 <200. They restarted HAART if CD4 was <350 and had to remain on HAART until CD4 increased to >700 again. Besides the SMART trial, LOTTI is the largest controlled trial ever performed on CD4 cell-guided STIs.
There are 2 articles below: the editorial and the main report, the LOTTI Study.
from Jules: The Editorial neglected to raise the issue of inflammation that could be associated with detectable viral load, which over the long-term could increase rates of clinical events; it is surprising this wasn't brought up. Part of the problem with treatment interruption as presented in the LOTTI Study is the required close management of patients -- preventing CD4 counts from declining too low because patients don't return to see their provider in a timely way. Another issue is the LOTTI Study is confined to patients with >700 CD4s. Another question about the LOTTI Study is that cardiovascular events occurred more often in the interruption arm but if CVD issues are managed well these events could be prevented. The LOTTI Study excluded patients with a nadir CD4 <200. There was close monitoring of patients in the LOTTI Study: Patients were followed up after 1, 2, 4 months and every 4 months afterwards, with clinical assessment and laboratory monitoring including haematology, biochemistry, triglycerides, cholesterol, CD4+ cell counts and HIV-RNA plasma levels. Unscheduled visits and laboratory tests were performed according to the clinical needs (i.e. evidence of adverse events or viral rebound or rapid immunologic decline). Drug resistance occurred more often in the STI group than in the control group but drug resistance in the control was likely due to nonadherence. While resistance in the STI group was due to other reasons inherent in STIs. the STI strategy was significantly less costly than continuous HAART (P < 0.0001). The mean per patient daily total cost for continuous HAART was 20.29 (SD 4.73) and it dropped to 9.07 (SD 6.82) in the STI arm. This means that each patient on the STI programme induces a saving of approximately 4500 per year of follow-up.
Is it smart to continue to study treatment interruptions? Editorial Comment
27 April 2009 - Volume 27 - Issue 7 - p 757-759
Hirschel, Bernard; Flanigan, Timothy
aHIV/AIDS Unit, Division of Infectious Diseases, Geneva University Hospital, Geneva, Switzerland
bAlpert Medical School of Brown University, Providence, Rhode Island, USA.
The Strategies in Management of Antiretroviral Therapies (SMART) trial compared continuous with intermittent therapy .. Not only AIDS-defining opportunistic infections and deaths, but also metabolic and cardiovascular diseases were significantly more frequent in the intermittent treatment arm. The latter finding was particularly surprising because, before SMART, it had been widely assumed that these diseases were caused by highly-active antiretroviral therapy (HAART) itself.. SMART indicated otherwise and suggested that the spectre of HIV-related diseases was much wider than previously thought. SMART was large and well conducted, comprising only intermittent treatment study with a size sufficient to measure clinical endpoints; its results are widely accepted.
Does SMART imply that HAART, once started, should never ever be interrupted? Some have concluded as much, but careful perusal of the results of SMART leaves some doubts. For instance, most clinical events in SMART occurred when the CD4 cell count was low. Although these events were more frequent in the intermittent than in the continuous treatment arm in all CD4 cell strata, the marginal utility of treatment in the intermittent arm decreased as the CD4 cell count rose. It can be estimated that 47 patient-years of treatment are necessary to prevent one event with CD4 cell counts of 350-500 μl, and 76 patient-years with CD4 cell counts above 500 μl. Is this a reasonable investment of healthcare resources, when the event in question is, for instance, a case of oesophageal candidiasis, easily managed with a few pills of fluconazole and resumption of HAART? A formal cost-effectiveness analysis is not available, but it is clear is that intermittent treatment saves drug costs , approximately 50% in Long-Term Treatment Interruption (LOTTI) and other studies.
The LOTTI study, as reported in this issue of AIDS , included patients on HAART, with a CD4 cell nadir above 200 μl, CD4 cell counts at study start above 700 μl, and a plasma HIV1-RNA level below 50 copies per μl. Patients randomized to scheduled treatment interruption (STI) stopped treatment until their CD4 cell count dropped below 350 μl; at that time, they restarted their HAART until CD4 cell counts again reached at least 700 μl. Patients in the control group continued their HAART. Mean follow-up was more than 4 years per patient.
As expected in view of the high CD4 cell counts and the limited patient-years of follow-up (1367 years, compared with 7296 years of SMART), the frequency of AIDS-defining events was low and not statistically different between the two groups. However, there was a marked difference in events relating to cardiovascular risk, and in contrast to the results from the SMART study, the difference favoured the STI group .
Can these different results be explained? Cross-study comparisons are always tricky (but part of the fun in epidemiology!). Table 1 highlights the essential features between SMART and LOTTI. SMART was bigger, the average patient's immune deficiency much more marked, and the follow-up was much shorter than in LOTTI.
At the editorialists' request, the authors provided a life-table analysis showing when events occurred (Fig. 1).
Fig.. 1. Heart disease and new-onset diabetes in the Long-Term Treatment Interruption study. The blue line represents the continuous treatment group and the green line the intermittent treatment group. The vertical interrupted lines show median and maximal treatment duration during SMART. SMART, Strategies in Management of Antiretroviral Therapies.
As can be seen, the curves are almost superimposed out to 4 years (corresponding to the maximum follow-up in SMART) and only diverge later on. This observation raises the intriguing possibility that the cardiovascular risk of HAART may not be uniform over time but only becomes apparent after many years of follow-up. SMART, with a median follow-up of only 16 months, may have missed that signal.
Will we ever know? Registries of treated patients like the D:A:D study  may detect unanticipated toxicity, such as excess cardiovascular risk associated with abacavir , but suffer from lack of an untreated control group. Changing treatment preferences also represent a difficulty: if long-term risk is associated with a drug that is no longer used (e.g. stavudine, causing mitochondrial toxicity, or indinavir, causing diabetes and nephrolithiasis, come to mind), the findings may be relevant for the future. The Strategic Timing of Anti-Retroviral Treatment (START) study aims to determine the risks and benefits of early treatment  and will shed light on side effects in a relatively immunocompetent group of patients similar to that of LOTTI - but START has not even started yet.
Many of our patients with high CD4 cell counts want to stop treatment. The LOTTI study does not justify a recommendation in that regard, but it does give clinicians useful information that it is probably safe to stop treatment within the limits of CD4 cell counts of LOTTI. Continued vigilance is needed so that excellent adherence is maintained when patients are on HAART to prevent the emergence of resistance.
The LOTTI study adds important information to the continued question of whether there is a role for interrupted therapy. Further study is justified, particularly with newer combination therapies, which may well have less toxicity and therefore shift the balance towards continuous treatment. Clinicians will welcome the information from LOTTI because it can allay some of the concerns regarding the safety of treatment interruptions at high CD4 cell counts.
CD4 cell-guided scheduled treatment interruptions in HIV-infected patients with sustained immunologic response to HAART
27 April 2009 - Volume 27 - Issue 7 - p 799-807
Maggiolo, Franco; Airoldi, Monica; Callegaro, Annapaola; Martinelli, Canio; Dolara, Alberto; Bini, Teresa; Gregis, Giampietro; Quinzan, Giampaolo; Ripamonti, Diego; Ravasio, Veronica; Suter, Fredy
aDivision of Infectious Diseases, Italy
bUnit of Antiviral Therapy, Italy
cMicrobiology and Virology Laboratory, Ospedali Riuniti, Bergamo, Italy
dDivision of Infectious Diseases, Azienda Ospedaliera Careggi, Firenze, Italy
eDivision of Infectious Diseases, General Hospital, Monza, Italy
fDivision of Infectious Diseases, Ospedale San Paolo University of Milano, Milan, Italy.
Objective: To compare continuous HAART with a CD4 cell-driven scheduled treatment interruption (STI) strategy.
Methods: LOng Term Treatment Interruption study is a randomized, controlled, prospective trial. Patients with CD4 cell counts more than 700 cells/μl were eligible, and the immunologic threshold to resume HAART was 350 cells/μl. The primary end point was the development of an opportunistic disease, death from any cause or the occurrence of diseases, other than opportunistic, requiring hospital admission. Secondary end points were major adverse effects, virologic failures and therapeutic costs.
Results: Three hundred and twenty-nine patients were randomized 1: 1. Total follow-up was 1388 person-years (mean 4.2 years). Patients in the STI group stopped therapy for a total of 241 STI cycles, their mean off-therapy period was 65.3% of the follow-up. The primary end point occurred in 12.1% of patients on STI and in 11.6% of controls [odds ratio 1.05; 95% confidence interval 0.54-2.05]. A higher proportion of patients in the STI arm were diagnosed with pneumonia (P = 0.037), whereas clinical events influencing the cardiovascular risk of patients were significantly (P < 0.0001) more frequent among controls. Eight patients (4.8%) in the STI group and 11 (6.7%) controls developed viral resistance [odds ratio 0.79, 95% confidence interval 0.27-1.81]. The mean daily therapeutic cost was 20.29 for controls and dropped to 9.07 in the STI arm (P < 0.0001).
Conclusion: The two strategies may be considered clinically equivalent. CD4 cell-guided STIs seem a possible alternative for chronically infected individuals responding to HAART provided that CD4 cell decrements would be steadily maintained above a safe threshold.
As the current regimens cannot eradicate HIV infection , persons living with HIV are recommended to stay on life-long medication therapy. Pill fatigue, adverse effects and resistance continue to limit the effectiveness of HAART [2,3]. Scheduled treatment interruptions (STIs) have emerged as a possible strategy to ameliorate these problems [4,5]. Theoretically, STIs may decrease the cumulative time a patient is exposed to antiretroviral agents, thus improving quality of life, reducing the emergence of drug-related adverse effects and the chance of selecting for drug-resistant viruses, and may reduce therapeutic cost for patients or the healthcare system. However, the role of STIs within the modern management of HAART remains controversial . Although several experiences have documented that CD4 cell-guided STIs may be virologically [7,8] and immunologically [9,10] safe among chronically infected patients responding to HAART, the benefit and risks of STIs can best be determined by long-term, controlled studies intended and powered not only to evaluate the risk of resistance or the immunologic response, but also to compare clinical outcomes including disease progression, AIDS, major toxicities and survival .
We conducted a multicentre, controlled, prospective and randomized trial on a cohort of chronically HIV-infected individuals on effective HAART to address most of these open questions [LOng Term Treatment Interruption study (LOTTI)]. We chose to apply an individualized pulse therapy strategy, driven by CD4 cell count, and to compare it with conventional continuous HAART.
Study design and patients
The LOTTI study design was based on a previous clinical experience . It is a randomized, open-label, controlled trial that compares two different treatment strategies. The patients, recruited from several centres in Northern Italy, had to be older than 18 years, with laboratory documentation of HIV-1 infection, a confirmed (at least two consecutive tests) CD4 cell count above 700 cells/μl, a confirmed plasma HIV-1 RNA level lower than 50 copies/ml and a stable HAART defined as the same (at least three) drugs constantly assumed for at least the previous 6 months. Patients were excluded if they had ever received any immunomodulatory drug, presented a CD4 nadir lower than 200 cells/μl, were HBsAg positive and if they were pregnant or breast-feeding women. Eligible consecutive patients were randomly assigned (according to a centralized, computer generated random list) in a 1: 1 ratio to continue their ongoing treatment or to stop it. Patients randomized in the STI arm stopped their HAART drugs altogether [no differed discontinuation was implemented for patients receiving nonnucleoside reverse transcriptase inhibitor (NNRTI)] and were kept off therapy until their CD4 cell count dropped to or below 350 cells/μl (single determination), irrespective of their viral load. Once their CD4 cell count declined to this limit, they resumed HAART until their CD4 cell count was again above 700 cells/μl (single measure) and their viral load below 50 copies/ml. Another interruption of therapy was proposed at this time point. STIs were, therefore, individualized with the aim to keep CD4 cell count above 350 cells/μl with the shortest possible exposure to drugs.
Patients in the STI arm generally resumed the same regimen they had stopped.. For all patients, a modification of HAART treatment was permitted at any time after randomization in the event of drug-associated toxicity or clinical progression of HIV disease (according to the US Centres for Disease Control and Prevention 1993 revised classification system for HIV infection)  or to comply with new indications (preferred regimens) of international guidelines for the management of HIV infection .
All patients provided their informed consent, and the trial was approved by the Institutional Review Boards of each hospital.
The scope of the study was to analyse efficacy and safety of a pulse therapy strategy based on CD4 cell-guided STIs compared to a continuous, conventional HAART. The primary end point of the study was clinical and defined as death from any cause, progression of HIV infection because of the occurrence of any opportunistic disease  or development of any clinical condition (other than AIDS-defining events) requiring hospital admission. The trial power was calculated on this end point. Secondary clinical end points included symptomatic diseases requiring medical intervention, but not leading to hospital admission and the occurrence of grade 3 or 4 adverse events according to the toxicity table of the Division of AIDS of the National Institute of Allergy and Infectious Diseases (NIAID).. Data on lower-grade toxic effects were not recorded.
Secondary objectives of the study were to verify the dynamics of CD4 cell loss, HIV replication and resistance-conferring mutations selection after treatment interruption and to evaluate the economic impact of the STI strategy.
Monitoring, laboratory testing and data collection
Consecutive HIV-infected patients who met the enrolment criteria entered the study. Patients were followed up after 1, 2, 4 months and every 4 months afterwards, with clinical assessment and laboratory monitoring including haematology, biochemistry, triglycerides, cholesterol, CD4+ cell counts and HIV-RNA plasma levels. Unscheduled visits and laboratory tests were performed according to the clinical needs (i.e. evidence of adverse events or viral rebound or rapid immunologic decline).
The number of CD4+ lymphocytes was measured by flow cytometry. Plasma HIV-RNA was measured through a quantitative assay (HIV-RNA 3.0 bDNA; Chiron, Emeryville, California, USA) with a lower limit of quantification of 50 copies/ml. All other variables were measured by means of standard laboratory tests.
A genotypic analysis of the reverse transcriptase and protease gene was performed in patients of the STI arm within 2 months from each treatment interruption and in all patients with viral rebound more than 400 copies/ml while receiving HAART.
The possibility for the selection of resistance-inducing mutations was tested by means of reverse transcriptase polymerase chain reaction (PCR) of the reverse transcriptase and protease gene. PCR products were sequenced using ViroSeq-HIV-1 genotyping system and analysed by means of genotyping system software, version 2.5 (ViroSeq-HIV-1; Applied Biosystem, Foster City, California, USA) .
During the follow-up, AIDS-defining events, other clinical events and adverse events were looked for systematically at all scheduled and unscheduled visits. AIDS was diagnosed according to the Centers for Disease Control and Prevention (CDC) classification system for HIV infection. Other clinical events were classified as 'major' if requiring hospital admission or 'minor' if manageable without hospitalization.
At each visit, the current use of drugs was recorded. Costs of the therapeutic strategy were calculated by adding the costs of individual drugs (as paid by the hospital pharmacy) with the costs of additional laboratory tests required to safely monitor STI (as indicated in the protocol) or to manage clinical and adverse events, to the costs of hospitalizations. Costs of the latter two variables were calculated according to the National Health System reimbursement wages that include all costs related to the care of a defined disease according to a predefined set of diagnosis [diagnosis-related group (DRG)].
The STI and control groups were compared according to the intention-to-treat (ITT) principle.
Descriptive results are presented as means ± standard deviation (SD), medians with interquartile range (IQR) and percentages with 95% confidence intervals (CIs). Inferential statistics using either parametric or nonparametric tests was used, as appropriate for the data type. χ2 or Fisher's exact test were used to analyse all categorical variables. Time-dependent variables were analysed using Kaplan-Meier product-limit estimates.. The log-rank test was used to assess the difference between the survival curves. Analysis of variation (ANOVA) test and Student's t-test were used to analyse continuous variables. All tests were two sided, and a P value inferior to 0.05 was regarded as significant.
The trial power was calculated to assess equivalence between the two strategies under the assumption that, in the control arm, the primary end point would be observed in a proportion of patients less than 7% and that the same proportion in the STI arm would not exceed 10% with a maximum allowed 95% CI of 12%. According to the Blackwelder approach , 320 patients had to be enrolled for alpha value of 5% and 1-beta value of 80%.
All analyses were performed with the SPSS statistical software package for Windows, version 13.0 (SPSS, Inc., Chicago, Illinois, USA).
Baseline characteristics and patients' disposition
Three hundred and twenty-nine patients were randomized: 165 in the STI arm and 164 in the control group.
Table 1 summarizes key baseline characteristics. Treatment groups were well balanced at entry.
The total follow-up time was 689 person-years in the STI group and 699 person-years in the control group, with a mean follow-up time of 4.2 and 4.3 years, respectively. On average, patients in the STI group received antiretroviral therapy during 34.7% (mean 515 days) of the follow-up time; in the control group, the same value raised to 98.3% (mean 1530 days).
Patients in the STI group interrupted therapy between one and five times. One hundred and nine patients stopped therapy only once either because they never reached the CD4 cell threshold to resume therapy (45 patients) or did not achieve the baseline CD4 cell level of 700 cells/μl after restarting therapy (64 patients). In the latter case, however, nine patients presented a short (<6 months) on-therapy follow-up at the moment the analysis was performed. Furthermore, 42 patients stopped therapy twice; nine for three times, four for four times and one for five times for an overall total of 241STI cycles. The mean duration of an STI cycle was 821 days (95% CI 721-919 days). CD4 cell counts in the control group fluctuated within a 200 cells/μl range (Fig. 1) indicating a ceiling effect due to the long mean HAART period (more than 5.5 years) (Table 1). HAART interruption induced in the STI group a constant decrement of CD4 cell counts (Fig. 1). At all time points, the mean CD4 cell value in the STI group was statistically different from the analogue value in the control group (P < 0.0001, Student's t-test). As a consequence of the individualized HAART intake in the STI group, throughout the follow-up, the CD4 cell count was, on average, 206 cells/μl lower than in the control group. However, the lower limit of the 95% CI for CD4 cell counts in the STI group was never inferior to 550 cells/μl. As a consequence, patients in the STI group, on average, spent 4.1% (59 days) of the follow-up at a CD4 cell level less than 350 cells/μl and 0.5% (8 days) at a CD4 cell level less than 250 cells/μl that was reached just sporadically.
Virologic and genotypic results
STIs were invariably associated with a rapid increment of HIV-RNA blood levels; however, patients promptly responded to the reintroduction of HAART.. Two months after reintroduction of HAART, 76.3% of them presented a HIV-RNA level less than 400 copies/ml and this proportion raised to 97.4% after 4 months. HIV-RNA declined in all patients but one (see below) to less than 50 copies/ml within 8 months.
Eight patients (4.8%) in the STI group and 11 (6.7%) in the control group developed resistance-conferring mutations during the follow-up period. The 95% CI for this difference lies within 3.1 and 6.9%. Being the CI within ± the 10% limit, the two strategies may be considered, with this respect, equivalent [odds ratio (OR) 0.79; 95% CI 0.27-1.81].
As far as specific mutations are concerned, M184V/I alone was selected in one patient in the STI group and in two controls. Three patients in the STI group were selected for NNRTI resistance-conferring mutations alone (either K103N or Y181C), whereas the association of M184V/I and NNRTI mutations was observed in three patients in the STI group and in four controls. Thymidine analogue mutations (TAMs) were selected in two patients in the STI group and in four controls. The most common TAMs resulted in K70R (five cases), K219Q (three cases) and M41L (two cases). Finally, three patients in the control group were selected for primary protease mutations.
Only one of the patients, in the STI group, was selected for resistance-conferring mutations during the induction phase of a new HAART period. In the remaining seven cases, the development of mutations occurred after the complete suppression of viral replication (HIV-RNA <50 copies/ml) was already achieved, that is, during a maintenance period. All patients selecting for mutations responded to a new genotypic-based HAART.
Clinical results (primary end point)
There were 39 patients who reached the primary end point of the study (Table 2). The Kaplan-Meier estimates for the cumulative probability of occurrence of the primary end point were 0.006 after 12 months, 0.018 after 24, 0.061 after 36 and 0.074 after 48 months in the STI group, whereas the same values in the control group were 0.006, 0.055, 0.089 and 0.107, respectively. The difference was not statistically significant (P = 0.854, log-rank test). Overall, 0.5% more patients in the STI group reached the primary end point of the study; the 95% CI for this difference laid well within the ± 10% limit for equivalence (OR 1.05, 95% CI 0.54-2.05) (Table 2).
All deaths were due to fatal myocardial infarction, and the only observed AIDS-defining event (progressive multifocal leukoencephalopathy) occurred in a patient who steadily refused to restart therapy despite having reached the protocol-defined CD4 cell threshold to reintroduce HAART.
At the moment the primary clinical end point was diagnosed, 11 out of 20 patients in the STI group were off therapy. Their last CD4 cell count in all cases, but two, was above the 350 cell/μl threshold (mean 557; SD 177 cell/μl). In the control group, the clinical events were diagnosed in patients having a mean CD4 cell count of 891 cell/μl (SD 320 cell/μl).
Several other clinical events requiring medical care, but not leading to hospital admission, were observed (Table 3). The most common events were infective (namely herpes zoster, sexually transmitted infections, pneumonia and acute hepatitis), observed in 22 patients in the STI group and in 15 controls. Overall, the occurrence of infections was not different between the two groups (Table 2); however, a higher proportion of patients in the STI arm were diagnosed with pneumonia (P = 0.037). On the contrary, diabetes and hypertension (six cases each), both requiring drug therapy, were frequently diagnosed among controls and in only two patients in the STI arm. As a whole clinical events influencing the cardiovascular risk of patients, as defined by the American Heart Association guidelines for the prevention of cardiovascular diseases and stroke , were significantly (P < 0.0001) more frequent among patients in the control group (Tables 2 and 3).
Laboratory safety results
Both therapeutic strategies were well tolerated; however, a greater proportion of patients on continuous HAART (27.4%) showed a grade 3 or 4 alteration of laboratory parameters during the observational period. The same proportion among patients in the STI group was 20.6%. The difference was not statistically significant (95% CI for the difference -0.18 to 16.0)..
The most frequently altered laboratory test was γ-glutamyltranspeptidase (γGT), followed by alanine aminotranferase (AST)/apartate aminotranferase (ALT) and triglycerides. In the STI arm, a grade 3 or 4 alteration of γGT was observed in 27 patients, AST/ALT were altered in 18, whereas a triglyceride increment was detected in one patient. Among controls, 33 patients showed a grade 3 or 4 increment of γGT, seven of AST/ALT and six of triglycerides.
The economic evaluation of the two therapeutic strategies was performed calculating a daily total cost taking into account the costs of drugs, laboratory exams to monitor the patient's status or to manage adverse events and clinical events and the cost of hospitalizations. To safely monitor patients in the STI group, virologic and immunologic laboratory tests were performed more frequently than currently recommended, increasing the total costs in this group. However, savings related to drug discontinuation overwhelmed these adjunctive expenses. As a consequence, considering the whole follow-up period, the STI strategy was significantly less costly than continuous HAART (P < 0.0001). The mean per patient daily total cost for continuous HAART was 20.29 (SD 4.73) and it dropped to 9.07 (SD 6.82) in the STI arm. This means that each patient on the STI programme induces a saving of approximately 4500 per year of follow-up.
In the last few years, the view that HAART must be lifelong has been questioned [17,18].
More recently, the strategies for management of antiviral therapy (SMART) study , a very large, randomized trial, has evaluated hard end points, such as the clinical outcome, disease progression or survival in patients undergoing CD4 cell count-guided STIs. According to the findings of this study, the episodic antiretroviral strategy, guided by CD4 cell counts, was found to be deleterious exposing patients to a 2.6 times greater risk of opportunistic diseases or death.
The SMART study is a cornerstone of the clinical experience on STI strategies, and any new data on CD4 cell-guided STIs must be discussed considering the strength of its results.
The results of LOTTI indicate that the long-term clinical outcome of a CD4 cell-guided STI strategy might be equivalent to continuous HAART. The discrepancies between our results and previously published data  may find an explanation in some of the experimental features of the two trial designs.
In the drug conservation arm of the SMART trial, the CD4 cell threshold to stop HAART was set to 350 cells/μl and that to resume it to 249 cells/μl. In the STI arm of this study, the same thresholds had quite different values: 700 and 350 cells/μl, respectively. Furthermore, over time, the mean CD4 cell count difference between the viral suppression and the drug conservation arms of both studies was similar (206 cells/μl). This does not surprise, the drug exposure in patients undergoing an STI being very similar in the two studies: 33.4% of follow-up (SMART) and 34..7% of the follow-up (LOTTI). However, being the starting point different, the consequences of this reduction in CD4 cell counts were also quite different. In our study, the lower limit of the 95% CI of this difference was never below an absolute value of 550 cells/μl and patients in the STI arm spent 95.9% of the follow-up time above the 350 cells/μl threshold. In the SMART trial, patients in the drug conservation arm spent only 67.9% of the follow-up above this threshold and spent 8.6% of follow-up below 250 cells/μl . The latter level was only sporadically reached by patients in the LOTTI study who, on average, remained below the 250 cells/μl threshold only for the 0.5% of the follow-up (8 days).
These differences may explain the clinical outcomes observed in this study and in the SMART study. If viral replication and the inflammatory events linked to it are the triggering force of clinical events in the STI arm (e.g. cardiovascular diseases), the influence of immunological status on the final outcome may result determinant.
Not surprisingly, findings by the SMART researchers indicate that the interruption of antiretroviral therapy with CD4 cell count thresholds higher than those they used may result in lower risk of opportunistic infections or death from any cause. In the same study, after adjusting for the latest CD4 cell count and viral load measure, the hazard ratio for this end point was lowered from 2.6 to 1.5 (95% CI 1.0-2.1) [11,19].
Although, in most cases, not adequately powered to detect clinical differences, several other trials on CD4 cell-guided treatment interruptions may support these findings. The Trivacan study  used a CD4 cell threshold to restart HAART of 250 cells/μl (equal to SMART) and saw an increment in serious morbidity in the STI arm. Four other studies [7,10,21,22] used a CD4 cell count restart threshold of at least 350 cells/μl and did not report an increased risk of serious morbidity for patients treated according an STI strategy.
This suggests that the CD4 cell threshold selected to reintroduce HAART is important . Whether a 350 cells/μl threshold should be considered of choice or a more conservative limit for restarting therapy (e.g. 400 cells/μl) would further increment the safety of STI strategies is unknown as trials exploring the use of such a threshold  were not powered to analyse the clinical outcome.
Besides the SMART trial, LOTTI is the largest controlled trial ever performed on CD4 cell-guided STIs. The total follow-up time was greater than 1300 person-years, and the mean follow-up time of our patients was 51.3 months compared with 20.7 months in the Staccato trial, 20.4 in the Trivacan study and 16 months for SMART. A longer follow-up increased our chances of observing slow to manifest or chronic clinically relevant events.
In this study, the primary end point was a composite clinical outcome comprehensive of the occurrence of opportunistic diseases, death from any cause and diseases or manifestations requiring hospital admission. This choice was based on the fact that the STI strategy, to be considered equivalent to continuous HAART, should not just prevent death or progression to AIDS, but should ensure a good health status, too. According to this end point, we documented that the STI strategy was clinically noninferior to continuous HAART.
Virologically, STIs were invariably associated with a rapid increment of HIV-RNA plasma levels, but all patients (with a single exception) responded to the reintroduction of HAART and HIV-RNA declined less than 50 copies/ml within 8 months. Strictly related to HIV replication is the occurrence of HIV drug resistance, a major concern in STI programmes. Genotypic analysis revealed that newly selected mutations occurred in 4.8% of patients in the STI group and in 6.7% in the control group. The 95% CI for this difference was within ±10% which allowed to consider the two strategies as equivalent [24,25]. Of note, only one of the patients was selected for resistance-conferring mutations during the induction phase of a new HAART period . In the remaining cases, the development of mutations occurring after the complete suppression of viral replication (HIV-RNA < 50 copies/ml) was already achieved, suggests that treatment failure was possibly due to a suboptimal adherence or treatment fatigue, rather than strictly related to the STI strategy.
Tolerability drawbacks of pulse therapy seem to be limited. A reseeding of viral reservoirs is probably the rule during STI, and clinical consequences such as generalized lymphoadenopathy (three cases in our study) may occur, but are generally limited in time and seriousness. According to our data, the risk of grade 3 or 4 laboratory alterations was not significantly different in the two treatment arms.
Finally, STIs induced a marked and significant reduction of management costs, even after correction for additional laboratory costs due to a proper and safe monitoring of patients.
After the results of the SMART trial were made available, most researchers and clinicians thought that the natural conclusion was that once you have started HAART, you never could stop it again . The SMART trial clearly demonstrated that an STI strategy could not be applied to all HIV-infected patients. The divergent results, we obtained, reopen the question whether specific patients' subpopulation or a different management of the STI programmes [28-30] could make this option still practicable. Our results may, also, be useful in the case a clinician would not be willing to implement STI strategies, but would be forced to manage patients who decide, on their own, to stop HAART. Knowing, not only the risks, the prognostic variables, but also the safe limits within which a patient could stay, would help clinicians to advise and manage such patients.
In summary, CD4 cell-guided interruption of antiretroviral treatment seems a possible alternative, strategic option for some chronically infected individuals responding to HAART and showing a high degree of immune reconstruction. According to our results, to make this option valuable, CD4 cell-guided STI programmes should be designed to preserve the immune function when patients are off therapy and to ensure that the CD4 cell counts can constantly be maintained above a safe threshold.
STIs of antiretroviral therapy warrant further careful prospective evaluation especially to investigate virologic and clinical outcomes in the very long period.