Number of Doses Taken Over the Last 4 Days
Across all visits, the percentage of participants who have taken ≥95% of their medications was higher in the TDF/FTC + DRV/r arm and ranged from 91% to 93% compared with 87% to 91% in the RAL + DRV/r arm [odds ratio (OR) 1.43; 95% confidence interval (CI): 1.04 to 1.97; P = 0.029]. Results were similar when analyzing drugs taken in the 4 ordered categories <80%, ≥80–95%, ≥95–99%, and 100% (OR 1.71; 95% CI: 1.14 to 2.58; P = 0.0096). Overall adherence and the difference between the arms did not change over time (no significant statistical interaction between arm and time). When looking at individual drugs, we found no difference between the arms in adherence to DRV or ritonavir. Although in the TDF/FTC + DRV/r arm, there was no difference in adherence to DRV, ritonavir, and TDF/FTC, in the RAL + DRV/r arm, adherence to RAL was slightly lower than to DRV or ritonavir (proportion of patients with ≥95% of medication taken across all visits was 88% for RAL versus 91% for DRV/r; P < 0.001).
Adherence on the Visual Analogue Scale
The proportion of patients with ≥95% of prescribed medication taken over the last 30 days ranged from 90% to 92% without difference between the arms (P = 0.66). However, patients in the TDF/FTC + DRV/r arm had significantly better adherence across all visits when compared with the RAL + DRV/r arm (OR 1.55; 95% CI: 1.13 to 2.13; P = 0.0072). Adherence in both arms was decreasing over time (P < 0.0001): 100% adherence at weeks 04, 24, and 96 was reported in 88%, 77%, and 71% in TDF/FTC + DRV/r, and in 76%, 71%, and 67% in RAL + DRV/r.
Of note, concordance for the adherence category of ≥95% between adherence in the last 4 days and VAS across all visits was 88%, with no difference between the randomization arms.
Following the Timing-Specific Schedule Over the Last Week
The proportion of patients with no timing deviation of more than 2 hours ranged from 45% to 50%, with a trend for better adherence in the TDF/FTC + DRV/r compared with the RAL + DRV/r arm (OR 1.26; 95% CI: 0.99 to 1.62; P = 0.06), and no significant change over time (P = 0.45). Across all visits, 16% patients reported to have never followed the timing schedule, 13% some of the time, 2% half of the time, and 21% most of the time (P = 0.53 for comparison of arms).
Unplanned Treatment Interruptions After the Last Visit
Across all visits, no treatment interruption was reported in 92% with no statistically significant difference between the arms (P = 0.30); on 5%, 1%, 1%, <0.1%, and 0.6% of visits, treatment interruptions of 1 day, 2 days, 3–4 days, 5–6 days, or ≥7 days were reported. Of note, treatment interruptions were more frequently reported with longer follow-up (P = 0.008) and were 6%, 9%, and 11% at weeks 04, 24, and 96 across both arms.
Adherence by Treatment Arm and Viroimmunological Parameters
When comparing medication adherence between patients with different baseline CD4+ cell count (<200 versus ≥200 cells/mm3) or baseline HIV-1 RNA (<100,000 versus ≥100,000 copies per milliliter), there were no significant differences in any of the adherence measures, neither overall nor differential in the 2 study arms (no statistically significant interaction with treatment arms; results not shown). There were also no differences when comparing participants with CD4+ <200 cells/mm3 and HIV-1 RNA ≥100,000 copies per milliliter with the rest of the study population (Table 2).
Among participants with at least 1 adherence measurement, virological suppression at HIV-1 RNA <50 copies per milliliter was achieved at W04 in 26% of cases (12% in the TDF/FTC + DRV/r arm; 40% in the RAL + DRV/r arm), and at W12 in 63% of cases (48% in the TDF/FTC + DRV/r arm; 78% in the RAL + DRV/r arm). Thereafter, the proportion of patients with HIV RNA <50 copies per milliliter was similar in the 2 arms with overall percentages of 84%, 86%, 91%, 92%, 93%, and 91% at weeks 24, 32, 48, 64, 80, and 96, respectively. The associations between adherence and HIV-1 RNA <50 copies per milliliter at the time of assessment are summarized in Table 1, Supplemental Digital Content http://links.lww.com/QAI/B212.
Association of Adherence Measures With Virological Failure and Primary Endpoint
The analysis of adherence measures and virological failure according to the virological component of the main trial's primary endpoint definition included 770 patients with 111 virological failures (Table 3). We found that patients with ≥95% of doses taken in the last 4 days had a lower risk of virological failure (univariable hazard ratio [HR] 0.54; 95% CI: 0.33 to 0.89; P = 0.015), with no difference between the 2 arms (P value of the interaction term 0.19). Unplanned treatment interruptions (any versus none; the same for time of interruption in ordered categories, results not shown), never >2 hours timing deviations, or ≥95% adherence during the last 30 days on the VAS were not associated with virological failures. For all adherence measures, multivariable analyses gave similar results, as did analyses censoring patients after switch from their allocated regimen (results not shown).
In addition to virological failure, we analyzed the relationship between adherence measures and the primary endpoint of the NEAT001/ANRS143 trial including also clinical failures (122 failures, of which 104 were virological failures), and found similar associations.
We then repeated the analyses for the subgroup with baseline CD4+ <200 cells/mm3 and HIV-1 RNA ≥100,000 copies per milliliter, where higher failure rates had been seen in the RAL + DRV/r arm compared with the TDF/FTC + DRV/r arm. No significant associations of adherence with virological failure or the primary endpoint were observed. In the RAL + DRV/r arm, for this specific group of patients, the HR was <1 and lower for all 4 adherence measures compared with all patients in the RAL + DRV/r arm and with participants in the TDF/FTC + DRV/r arm suggesting that in this subgroup, good adherence might be more relevant than in the TDF/FTC + DRV/r arm; however, P values were not significant.
Factors Associated With Adherence
Multivariable analyses confirmed the associations described above between specific adherence measures and randomization arm and time on study (Table 4). In addition, older age was significantly associated with better medication adherence (except time deviation). We also found that nonwhites had significantly worse adherence than whites across various measures (except treatment interruptions), and that adherence significantly differed between countries. We did not find an association of any of the adherence measures with either sex, or baseline CD4+ cell count, or baseline HIV-1 RNA.
Adherence to ART, the key determinant of virological response is a very complex parameter because it reflects human behavior and as such varies from 1 patient to another and can change over time. In the NEAT001/ANRS143 study, ART-naive subjects on average self-reported high levels of adherence to study drugs in both treatment arms. More in depth analysis of adherence behaviors showed for the TDF/FTC + DRV/r group compared with the RAL + DRV/r group significantly higher adherence rates at all time points in 3 of 4 adherence measures (number of doses taken over the last 4 days, 30 days of adherence on the VAS, and deviation from timing-specific schedule). The only adherence measure significantly associated with time to virological failure or the primary endpoint of the NEAT001/ANRS143 trial including also clinical failures was <95% adherence at short-term patient recall (ie, “doses taken in the last 4 days”), with no difference between the 2-arm subgroups based on viroimmunological baseline status; we did not see significant differences in any of the adherence measures, neither overall nor differential in the 2 study arms.
The threshold of >95% adherence to prescribed doses and tablets was established as a requirement for achieving virological suppression with first-generation unboosted protease inhibitors, a class with short half-live, low genetic barrier to resistance development, and limited forgiveness.9 Further studies showed that adherence and virological suppression improved with simplified regimens and lower pill burden, although in a meta-analysis of 19 studies published through March 2013, patients on once-daily regimens did not achieve virological suppression more frequently than patients on twice-daily regimens.10 Impact of adherence behavior might also depend on the drug class considered. With ritonavir-boosted protease inhibitors, average adherence was a better determinant of virological success than was the duration or frequency of treatment interruption, whereas for non-nucleoside reverse-transcriptase inhibitors, consecutive missed doses were associated with the highest risk of virological failure.11–13 RAL is a well-tolerated and effective drug that demonstrated durable virological suppression in first-line ART through 240 weeks of therapy.14 However, because of its twice-daily formulation, RAL may be highly susceptible to various nonadherence behaviors, such as selective morning or evening dose skipping, short treatment interruptions, and suboptimal levels of adherence. These patterns are common in clinical practice.15 In a prospective study assessing the patterns of adherence to RAL-based regimens, longer treatment interruption and average adherence were both independently associated with virological failure.16 Of note, TDF/FTC is considered as the most forgiving N(t)RTI combination because of the very prolonged intracellular half-lives of the 2 drugs.17
Our study seems to confirm the importance of high adherence levels (ie, >95%) on likelihood of virological success. Keeping in mind that self-reported adherence may overestimate real drug intake attributing virological failure to differences in efficacy between treatment arms and leading to incorrect study conclusions, the only adherence measure associated with virological failure was self-reporting <95% of doses taken/prescribed in the past 4 days, with no difference between treatment arms. These results, suggesting that adherence in the past 4 days had a greater impact on virological failure than treatment interruption, can probably be explained by the fact that treatment interruptions ≥5 days were seen in less than 0.7% of visits. Indeed, over a 30-day period, missing 1 dose on 7 consecutive days (clustered missed doses) will not have the same impact as missing 7 doses on single independent days (interspaced missed doses).18 Of interest, in our study, unplanned treatment interruptions were very infrequent, increased with prolonged follow-up, but with no differences between treatment arms. This could reflect treatment fatigue or some interfering social/personal events independent of ARV regimen. In our study, given the short plasma half-life of RAL, even short-term RAL interruptions could have led to ineffective intracellular concentrations in some patients, even with high levels of average adherence.
In the subgroup of patients with baseline CD4+ <200/mmc and HIV-1 RNA >100.000 copies per milliliter, there were no significant associations of adherence with virological failure or the primary endpoint, and no significant difference between the arms. In the RAL + DRV/r arm, we noted a decreased hazard ratio in this specific patient group for all 4 adherence measures compared with all patients in this arm, and to participants in the TDF/FTC + DRV/r arm suggesting that in this subgroup, good adherence might be more relevant than in the TDF/FTC + DRV/r arm; however, P values were not significant. Based on this result, the higher rate of virological failure in the subgroup of patients with baseline CD4+ <200/mm3 and HIV-1 RNA >100,000 copies per milliliter found at the explanatory post hoc analysis was not explained by differences in adherence. However, it should be considered that numbers were small in this subgroup, and the trial was not specifically powered to conduct subgroup analyses to this regard. In fact, it is possible that, rather than an absence of effect, the sample size was not large enough to detect a significant association between adherence levels and inferiority of RAL + DRV/r in respect to TDF/FTC + DRV/r in patients with worse viroimmunological status (ie, CD4 cells <200/mm3 or CD4 cells <200/mm3 and HIV RNA >100,000 copies per milliliter).
The best way to measure adherence, from a clinically relevant standpoint, is still debated.19,20 Patient self-reported measures of ARV adherence can greatly vary in terms of item content, format, or period investigated.21 Nevertheless, the method is frequently used because of utilization ease, low costs, nonintrusiveness, and wide applicability. Overestimation of real adherence rates due to desirability bias may be of concern for the validity of self-reported adherence measure but applies in a randomized clinical trial to both treatment arms equally. An analysis of 1247 HIV-positive subjects participating in multicenter medication adherence-promotion trials showed that self-report overestimated actual medication ingestion by an average of 26% points compared with electronic drug monitoring.22 Furthermore, the patient's recall of medication intake in answering diverse adherence questions may reflect not only drug ingestion but also other aspects, such as patient's perception and/or satisfaction to treatment. In fact in our study, the rather low concordance observed between categories of self-reported number of doses taken in the last 4 days and 30-day adherence at the VAS (88%) is likely to be due to these aspects. Medication event monitoring system, the gold standard in some studies, can by itself represent an interventional bias and could be associated to poor adherence to the device that measures the adherence behavior.23 Evaluation of differences between adherence measures remains an interesting aspect to investigate, and methods for cocalibration of different instruments are needed.24 The fact that adherence worsened over time in our results is consistent with previously reported data.25
Our results show that nonwhite ethnicity was associated with lower adherence and rate of virological failure or primary endpoint occurrence. Lower adherence in nonwhite patients cared for in Australian centers has been associated with social and cultural issues, independently of regimen composition.26 Complying to a twice-daily regimen might be more problematic in patients faced with poverty, poor housing, and eating insecurity. Older age was associated with better adherence, independently of regimen allocated, and whatever the measure of adherence. Many studies have documented younger age as a relevant factor associated to lower adherence, emphasizing the need for targeted counseling on adherence barriers in this population.27–29
With regards to study limitations, the most important issue as mentioned above is that, besides missing adherence data, the study was not specifically powered to conduct subgroup analyses. Therefore, it is not possible to determine whether our findings substantiate the hypothesis of efficacy difference between TDF/FTC + DRV/r and RAL + DRV/r in ART-naive subjects with worse baseline viroimmunological status or whether they are due to insufficient power for subgroup analyses. Furthermore, disparities in clinical management across centers may have impacted adherence behaviors, and we cannot exclude the fact that some nonadherence behaviors may have been missed because of the infrequent collection of self-administered questionnaires. Finally, measurement of adherence by patient self-report can overestimate medication intake, but more objective measures such as plasma drug concentrations or medication event monitoring system have other disadvantages, such as “white coat adherence” close before clinic visit or intrusiveness in the patient's daily life. On the contrary, one of the main strengths of the study is the evaluation of adherence to ARVs within a randomized clinical trial comparing of TDF/FTC + DRV/r versus RAL + DRV/r in a large patient population that was followed up for a considerable period. Second, as the study protocol did not include adherence interventions, the results provide valuable information about adherence dynamics in ARV-naive HIV-infected persons and its relation to virological outcome.
In conclusion, in this randomized study comparing 2 strategies of first-line ART, average adherence assessed by patient self-report was high in both arms, but slightly and significantly better for TDF/FTC + DRV/r compared with RAL + DRV/r. Only adherence <95% in the last 4 days was associated to a higher risk of virological failure, with no differences between the 2 arms. Adherence levels were not different in baseline CD4+ and HIV RNA strata across arms, and there was no convincing evidence that higher failure rate in the RAL + DRV/r arm in the subgroup of patients with baseline CD4+ <200/mmc and HIV-1 RNA >100,000 copies per milliliter found at the explanatory post hoc analysis of the NEAT001/ANRS143 trial was caused by adherence differences.
The authors thank the NEAT001/ANRS143 study participants and their partners, families, and caregivers for participation in the study. They also thank the staff from all the centers participating in the trial.
1. Antiretroviral Therapy Cohort Collaboration. Survival of HIV
-positive patients starting antiretroviral therapy between 1996 and 2013: a collaborative analysis of cohort studies. Lancet HIV
2. Genberg BL1, Wilson IB, Bangsberg DR, et al; MACH14 Investigators. Patterns of antiretroviral therapy adherence
and impact on HIV
RNA among patients in North America. AIDS. 2012;26:1415–1423.
3. Saberi P1, Neilands TB, Vittinghoff E, et al. Barriers to antiretroviral therapy adherence
and plasma HIV
RNA suppression among AIDS clinical trials group study participants. AIDS Patient Care STDS. 2015;29:111–116.
4. Duran S, Spire B, Raffi F, et al; the APROCO Cohort Study Group. Self-reported symptoms after initiation of a protease inhibitor in HIV
-infected patients and their impact on adherence
to HAART. HIV
Clin Trials. 2001;2:38–45.
5. Al-Dakkak I, Patel S, McCann E, et al. The impact of specific HIV
treatment-related adverse events on adherence
to antiretroviral therapy: a systematic review and meta-analysis. AIDS Care. 2013;25:400–414.
6. Raffi F, Babiker AG, Richert L, et al; NEAT001/ANRS143 Study Group. Ritonavir-boosted darunavir combined with raltegravir
or tenofovir-emtricitabine in antiretroviral-naive adults infected with HIV
-1: 96 week results from the NEAT001/ANRSed non-inferiority trial. Lancet. 2014;384:1942–1951.
7. Reynolds NR, Sun J, Nagaraja HN, et al. Optimizing measurement of selfreported adherence
with the ACTG Adherence
Questionnaire: a crossprotocol analysis. J Acquir Immune Defic Syndr. 2007;46:402–409.
8. Liu H, Golin CE, Miller LG, et al. A comparison study of multiple measures of adherence
protease inhibitors. Ann Intern Med. 2001;134:968–977.
9. Paterson DL, Swindells S, Mohr J, et al. Adherence
to protease inhibitor therapy and outcomes in patients with HIV
infection. Ann Intern Med. 2000;133:21–30.
10. Nachega JB, Parienti JJ, Uthman OA, et al. Lower pill burden and once-daily antiretroviral treatment regimens for HIV
infection: a meta-analysis of randomized controlled trials. Clin Infect Dis. 2014;58:1297–1307.
11. Nelson M, Girard PM, Demasi R, et al. Suboptimal adherence
has minimal effect on efficacy compared wit lopinavir/ritonavir in treatment-naive, HIV
-infected patients: 96 week ARTEMIS data. J Antimicrob Chemother. 2010;65:1505–1509.
12. Parienti JJ, Ragland K, Lucht F, et al. Average adherence
to boosted protease inhibitor therapy, rather than the pattern of missed doses, as a predictor of HIV
RNA replication. Clin Infect Dis. 2010;50:1192–1197.
13. Parienti JJ, Das-Douglas M, Massari V, et al. Not all missed doses are the same: sustained NNRTI treatment interruptions predict HIV
rebound at low-to-moderate adherence
levels. PLoS One. 2008;3:e2783.
14. Rockstroh JK, DeJesus E, Lennox JL, et al. Durable efficacy and safety of raltegravir
versus efavirenz when combined with tenofovir/emtricitabine in treatment-naive HIV
-1-infected patients: final 5-year results from STARTMRK. J Acquir Immune Defic Syndr. 2013;63:77–85.
15. Ortego C, Huedo-Medina TB, Llorca J, et al. Adherence
to highly active antiretroviral therapy (HAART): a meta-analysis. AIDS Behav. 2011;15:1381–1396.
16. Gras G, Schneider MP, Cavassini M, et al. Patterns of adherence
-based regimens and the risk of virological failure among HIV
-infected patients: the RALTECAPS cohort study. J Acquir Immune Defic Syndr. 2012;61:265–269.
17. Stevens RC, Blum MR, Rousseau FS, et al. Intracellular pharmacology of emtricitabine and tenofovir. Clin Infect Dis. 2004;39:877–878.
18. Parienti JJ, Paterson DL. Number of missed doses: why 1 x 7 doses does not make 7 x 1? AIDS. 2012;26:1437–1440.
19. Chesney MA, Ickovics JR, Chambers DB, et al. Self-reported adherence
to antiretroviral medications among participants in HIV
clinical trials: the AACTG adherence
instruments. Patient Care Committee & Adherence
Working Group of the Outcomes Committee of the Adult AIDS Clinical Trials Group (AACTG). AIDS Care. 2000;12:255–266.
20. Marcellin F, Spire B, Carrieri MP, et al. Assessing adherence
to antiretroviral therapy in randomized HIV
clinical trials: a review of currently used methods. Expert Rev Anti Infect Ther. 2013;11:239–250.
21. Pearson CR, Pantalone DW, Merrill JO, et al. Self-report measures of antiretroviral therapy adherence
: a review with recommendations for HIV
research and clinical management. AIDS Behav. 2006;10:227–245.
22. Simoni JM, Huh D, Wang Y, et al. The validity of self-reported medication adherence
as an outcome in clinical trials of adherence
-promotion interventions: findings from the MACH14 study. AIDS Behav. 2014;18:2285–2290.
23. Genberg BL, Wilson IB, Bangsberg DR, et al. Patterns of adherence
to antiretroviral therapy and impact on HIV
-RNA among patients from the MACH14 study. AIDS. 2012;26:1415–1423.
24. Nance RM, Delaney JA, Golin CE, et al. Co-calibration of two self-reported measures of adherence
to antiretroviral therapy. AIDS Care. 2017;29:464–468.
25. Heterogeneity among studies in rates of decline of antiretroviral therapy adherence
over time: results from the multisite adherence
collaboration on HIV
14 study. J Acquir Immune Defic Syndr. 2013;64:448–454.
26. Siefried KJ, Mao L, Kerr S, et al; PAART Study Investigators. Socioeconomic factors explain suboptimal adherence
to antiretroviral therapy among HIV
-infected Australian adults with viral suppression. PLoS One. 2017;12:e0174613.
27. Protopopescu C, Raffi F, Roux P, et al; ANRS CO8 APROCO-COPILOTE Study Group. Factors associated with non-adherence
to long-term highly active antiretroviral therapy: a 10 year follow-up analysis with correction for the bias induced by missing data. J Antimicrob Chemother. 2009;64:599–606.
28. Shubber Z, Mills EJ, Nachega JB, et al. Patient-reported barriers to adherence
to antiretroviral therapy: a systematic review and meta-analysis. PLoS Med. 2016;13:e1002183.
29. Sheppard DP, Weber E, Casaletto KB, et al; HIV
Neurobehavioral Research Program (HNRP) Group. Pill burden influences the association between time-based prospective memory and antiretroviral therapy adherence
in younger but not older HIV
-infected adults. J Assoc Nurses AIDS Care. 2016;27:595–607.
adherence; HIV; antiretrovirals; NtRTI-sparing regimen; raltegravir; darunavir/ritonavir
Supplemental Digital Content
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.