From *INSERM, U912 (SE4S), Marseille, France; †Université Aix Marseille, IRD, UMR-S912, Marseille, France; ‡ORS PACA, Observatoire Régional de la Santé Provence Alpes Côte d'Azur, Marseille, France; §Hôpital La Pitié-Salpêtrière, Service des Maladies Infectieuses, Paris, France; ‖INSERM, U897 and ISPED, Université Victor Segalen, Bordeaux, France; ¶CHU Bretonneau, Service de Médecine Interne et Maladies Infectieuses, Tours, France; #Hôtel-Dieu, Nantes, France; and **Hôpital Bichat-Claude Bernard, Paris, France.
To the Editors:
Treatment interruption (TI) strategy among HIV-infected patients on highly active antiretroviral therapy (HAART) is associated with deleterious immunologic outcomes, increased risk of virologic failure, and resistance emergence.1-3 Two major studies, Strategies for Management of Antiretroviral Therapy (SMART) and TRIVACAN, demonstrated in 2006 that TI led to a significantly higher risk of disease progression and death.4,5 Consequently, TIs are currently not recommended in patients on HAART6; nevertheless, they still occur in routine practice.7 However, it is quite difficult to determine to what extent TIs are decided on by the physician or arise from a patient's undisclosed act of treatment discontinuation. We used data from the ANRS CO8 APROCO-COPILOTE cohort to estimate the incidence of medically recorded TI and to identify their clinical and psychosocial predictors.
The ANRS CO8 APROCO-COPILOTE is a cohort of patients started on a protease inhibitor-containing regimen between 1997 and 1999, collecting prospective clinical data by standardized medical records and sociodemographic and psychosocial data by self-administered questionnaires. All visits corresponding to the HAART maintenance period (Months M12-M96) for patients with controlled viremia (plasma HIV RNA less than 400 copies/mL) during the 12 months before the then current visit were included in the analysis. All episodes of TI (including multiple episodes for some patients) lasting at least 60 days were selected from the medical records.
A history of viral rebound was defined as plasma HIV RNA greater than 400 copies/mL at least once before the 12 months of controlled viral load preceding the then current visit. The occurrence of HIV-related clinical events not related to HAART toxicity was recorded in the medical files. An extended Cox proportional hazards model for recurrent events was used to identify the clinical, sociodemographic, and psychosocial factors associated with TI.8,9 The multivariate model was built using a backward selection of explanatory variables. A period effect on TI occurrence was tested (pre-/post-January 2006, corresponding to the circulation of recommendations against TI). All analyses were performed using Stata, Version 9 (Stata, College Station, TX).10,11
Among the 832 selected patients, accounting for 9595 visits and 3592 person-years, 96 patients experienced 106 TI (incidence rate [95% confidence interval] of 2.9 [2.4-3.6] per 100 person-years) during the 7-year study period. Median (interquartile range) duration of TI was 109 (84-157) days. Mean (standard deviation) CD4 cell count before TI was 690 (326) cells/mm3. Median (interquartile range) individual CD4 loss during TI was 87 (0-204) cells/mm3.
In the multivariate model, a history of viral rebound and the number of HIV-related clinical events were independently associated with lower probability of TI occurrence (Table 1). Patients who reported having a good patient-provider relationship or no social support from their main partner were more likely to experience TI. Neither a calendar period effect (pre-/post-January 2006) nor any interaction effect between the predictors and the calendar period were found. Moreover, no change in TI incidence was observed between the two periods.
This is the first study exploring the incidence and predictors of medically recorded prolonged TI in HIV-infected patients on HAART with controlled viremia. Because short interruptions are highly inadvisable and not promoted by physicians, selecting long-term TI from medical records enabled us to focus on those TI potentially monitored by the physician and/or more likely to have resulted from an “agreement” between the patient and the physician. By choosing patients with prolonged controlled viremia during the HAART maintenance phase, we tried to select those for whom a TI would have been judged less risky by the physician. This study has highlighted a relatively small TI incidence rate among patients with controlled viremia. These individuals were usually off HAART for short periods (median duration of TI was less than 4 months). Although recommendations against TI were included in French HAART guidelines in 2006,12 no calendar period effect was observed in the present study, thereby confirming that these guidelines do not seem to influence the incidence or the pattern of TI predictors. We also tested the effect of the type of the antiretroviral regimen on the occurrence of TI, comparing currently used boosted protease inhibitor and other protease inhibitor regimens used in the late 1990s. This effect did not reach statistical significance in the multivariate analysis, showing also that no difference in the pattern of TI over time was found.
The lack of social support from one's main partner, which has already been reported as a determinant of nonadherence to HAART,7,13 remained associated with TI in the multivariate model. However, the association found between TI and both a low number of HIV-related clinical events and no history of viral rebound confirmed that these TIs occur in a population with a successful HAART history, in turn probably characterized by sustained adherence. Such long-term TIs in patients with controlled viremia are likely therefore to result either from physicians' decision or from patients' negotiations with their physicians leading to TI approval. Both situations engage physicians in an assessment of the risks and benefits associated with TI on an individual basis. Interestingly, a good patient-physician relationship is a further predictor of TI insofar as a good relationship is a necessary basis for possible negotiations about any medical decision concerning treatment.14-16 Following the identification of this pattern of predictors, we hypothesize that TIs are unlikely to arise from patient self-management without any medical supervision. For patients with a successful HAART history, “treatment holidays” may be perceived as a much-needed respite from the burden of long-term continuous therapy and its associated side effects.17 There is no doubt that physicians have to take into account patient expectations when assessing the relevance of TI, in turn leading to shared responsibility for the decision.
Some limitations of the present study have to be acknowledged. First, the reasons for granting TI have not been investigated because this information was not available in the database. On the other hand, because of the very fact that TI are officially contraindicated, any study designed to explore whether TI have been negotiated by the patients and acknowledged by the physicians will be subject to desirability bias. Nevertheless, it would be interesting to understand to what degree patient negotiation for TI could influence the physician's choice to switch to a regimen with improved tolerance instead of choosing the TI option.
This study showed that the burden of long-term therapy can be eased through more comprehensive interaction between the patient and physician. Despite not being recommended in official guidelines, the question of permitting patient TI is not over.18 An improved assessment of TI occurrence is needed to individualize treatment and counseling strategies to minimize their incidence and their potential negative clinical consequences.
We thank all patients, nurses, and physicians in clinical sites.
Camelia Protopopescu, PhD*†‡
Perrine Roux, PharmD*†‡
Maria-Patrizia Carrieri, PhD*†‡
Christine Katlama, MD§
Fidéline Collin, PharmD‖
Jean-Marc Besnier, MD, PhD¶
Geneviève Chêne, MD‖
François Raffi, MD, PhD#
Bruno Spire, MD, PhD*†‡
Catherine Leport, MD, PhD** and the ANRS CO8 APROCO-COPILOTE Study Group
*INSERM, U912 (SE4S), Marseille, France;
†Université Aix Marseille, IRD, UMR-S912, Marseille, France
‡ORS PACA, Observatoire Régional de la Santé Provence Alpes Côte d'Azur, Marseille, France
§Hôpital La Pitié-Salpêtrière, Service des Maladies Infectieuses, Paris, France
‖INSERM, U897 and ISPED, Université Victor Segalen, Bordeaux, France
¶CHU Bretonneau, Service de Médecine Interne et Maladies Infectieuses, Tours, France
#Hôtel-Dieu, Nantes, France **Hôpital Bichat-Claude Bernard, Paris, France.
1. Cardiello PG, Hassink E, Ananworanich J, et al. A prospective, randomized trial of structured treatment interruption for patients with chronic HIV type 1 infection. Clin Infect Dis
2. Fixed duration interruptions are inferior to continuous treatment in African adults starting therapy with CD4 cell counts <200 cells/microl. AIDS
3. Pogany K, van Valkengoed IG, Prins JM, et al. Effects of active treatment discontinuation in patients with a CD4+ T-cell nadir greater than 350 cells/mm3: 48-week Treatment Interruption in Early Starters Netherlands Study (TRIESTAN). J Acquir Immune Defic Syndr
4. El-Sadr WM, Lundgren JD, Neaton JD, et al. CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med
5. Danel C, Moh R, Minga A, et al. CD4-guided structured antiretroviral treatment interruption strategy in HIV-infected adults in west Africa (Trivacan ANRS 1269 trial): a randomised trial. Lancet
6. Hammer SM, Eron JJ Jr, Reiss P, et al. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society-USA panel. JAMA
7. Glass TR, De Geest S, Weber R, et al. Correlates of self-reported nonadherence to antiretroviral therapy in HIV-infected patients: the Swiss HIV Cohort Study. J Acquir Immune Defic Syndr
8. Prentice RL, Williams BJ, Peterson AV. On the regression analysis of multivariate failure time data. Biometrika
9. Lin DY. Cox regression analysis of multivariate failure time data: the marginal approach. Stat Med
10. Cleves M. Analysis of Multiple Failure-Time Data With Stata
11. Stata. Stata Base Reference Manual, Release 9
. College Station, TX: StataCorp LP; 2007.
12. Yéni P. Prise en charge médicale des personnes infectées par le VIH: recommandations du groupe d'experts. Rapport 2006
. France: Flammarion; 2006.
13. Young J, De Geest S, Spirig R, et al. Stable partnership and progression to AIDS or death in HIV infected patients receiving highly active antiretroviral therapy: Swiss HIV cohort study. BMJ
14. Carder PC, Vuckovic N, Green CA. Negotiating medications: patient perceptions of long-term medication use. J Clin Pharm Ther
15. Holt M. Agency and dependency within treatment: drug treatment clients negotiating methadone and antidepressants. Soc Sci Med
16. Benarde MA, Mayerson EW. Patient-physician negotiation. JAMA
17. Carrieri MP, Leport C, Protopopescu C, et al. Factors associated with nonadherence to highly active antiretroviral therapy: a 5-year follow-up analysis with correction for the bias induced by missing data in the treatment maintenance phase. J Acquir Immune Defic Syndr
18. Ananworanich J, Hirschel B. Intermittent therapy for the treatment of chronic HIV infection. AIDS