The efficacy of first-line antiretroviral therapy (ART) continues to improve, according to an analysis of outcomes in 78,000 people in 181 studies, published by Professor Andrew Carr of St Vincent’s Hospital in Sydney and colleagues at the University of New South Wales. Over three-quarters of people who started treatment between 2011 and 2015 were still on their first regimen and had an undetectable viral load three years later.
Outcomes in 2011-2015 were significantly better than those recorded in people starting therapy between 2006 and 2010, which in turn were an improvement on results seen in people who started treatment between 2001 and 2005 and between 1994 and 2000. Moreover, there was a steady and sustained fall in the proportion of individuals who stopped their first line regimen because of side-effects or due to personal choice. However, the proportion of people stopping because of virological failure flatlined at approximately 5%.
But good as these post-2011 outcomes are, the investigators believe that it’s still possible to do better. They note that even with the most modern ART regimens, a fifth of people experience treatment failure (stop or change their treatment regimen, for any reason) within three years of starting HIV therapy.
The study contained some pointers about which regimens are likely to have the greatest success: people starting treatment with tenofovir (either TDF or TAF)/emtricitabine and integrase inhibitors were each more likely to remain on first-line treatment over three years. Other findings showed that baseline resistance testing and once-daily dosing were also predictors of long-term treatment success.
Five findings of especial importance are highlighted by the authors: although initial ART efficacy continues to improve, more than 20% of post-2010 patients on integrase inhibitor-based ART failed over 144 weeks; phase 3 studies overestimate ‘real world efficacy’; there are few routinely collected patient/clinical characteristics that predicted ART failure; the rate of ART cessation for virological failure does not appear to have declined in over 20 years; and the length of study follow-up remains too short.
ART treatment guidelines make recommendations for the selection of first-line therapy based on the sequential review of individual, randomised studies. To obtain a more accurate impression of the efficacy and durability of first-line ART, Carr and colleagues conducted a systematic review of outcomes reported in 181 studies involving 77,999 people.
Follow-up was from 1994 to 2017. A similar exercise conducted in 2008 showed that therapy based on tenfovir/emtricitabine was more effective than treatment based on abacavir/lamivudine when used with either a boosted protease inhibitor or a non-nucleoside reverse transcriptase inhibitor (NNRTI). An update in 2012 showed that treatment based on an integrase inhibitor was superior to treatment including either a boosted protease inhibitor or NNRTI. It also showed that differences in outcomes according to baseline viral load were smaller for integrase inhibitors than boosted protease inhibitors and NNRTIs.
Only randomised studies and prospective cohort studies involving people starting first-line ART were included. The primary outcome was treatment efficacy at weeks 48, 96 and 114, defined as an undetectable viral load reported on an intent-to-treat basis – changing treatment for any reason was therefore regarded as failure.
Overall, the patients had a mean age at baseline of 37 years, 75% were men and 61% were white. At baseline, mean CD4 cell count and viral load were 262 cells/mm3 and 63,000 copies/ml, respectively. Patients took a mean of 4.8 pills in 1.6 doses daily.
The main treatment backbones were tenofovir/emtricitabine (44%), thymidine-based (28%) and abacavir/lamivudine (10%). The principal third drugs were NNRTIs (50%), boosted protease inhibitors (28%) and integrase inhibitors (12%).
Outcomes at week 48 were reported for almost all the participants. Mean efficacy was 71%. It improved significantly over time, from 57% in studies starting between 1994 and 2000, to 69% for people initiating therapy between 2001 and 2005, to 77% for individuals in 2006 to 2010 studies and to 84% in post-2010 studies.
Efficacy at week 96 was reported in 41% of studies with an overall rate of 64%. It increased from 52% for pre-2000 studies, to 61% for 2001 to 2005 studies, to 65% for 2006 to 2010 studies and to 80% in studies recruiting between 2011 and 2015.
Only 14% of studies reported on week 144 outcomes and these found an overall efficacy of 62%. Once again, these improved over time. Efficacy rates for the four time periods were 45%, 55%, 72% and 77%, respectively.
Predictors of efficacy included choice of drug, with outcomes favouring both tenofovir/emtricibine and integrase inhibitors. Several dosing characteristics were also associated with greater efficacy: once-daily at week 48; dosing without food restrictions at week 96; and fewer pills at weeks 96 and 144.
“The type of ART used, in particular the use of an integrase inhibitor as the anchor drug and the use of a once-daily NRTI [nucleoside reverse transcriptase inhibitor] backbone, had a far greater impact on efficacy than did patient characteristics,” comment the authors. “The number of doses per day had a stronger relationship with efficacy than whether ART was taken as one pill or more than one pill per day.”
Baseline resistance testing (p = 0.0003) and higher baseline CD4 cell count (p = 0.0003) were also predictors of greater efficacy.
Outcomes in post-licensing and “real world” studies were progressively worse than those observed in phase three clinical trials.
Generally, outcomes were about 10% better with regimens recommended in US Department of Health and Social Services guidelines, compared to World Health Organisation recommended regimens.
However, almost 30% of people with 144 weeks of follow-up stopped their initial ART regimen prematurely. The most common reasons were patient choice (13%), side-effects (9%) and virological failure (5%). Most of these cessations happened by week 48. The proportion of early treatment cessations due to patient choice or side-effects was lower in studies conducted post-2010 compared to other time periods. However, the proportion of people stopping therapy because of virological failure remained stable at approximately 5%.
“Initial ART efficacy continues to improve, but more than 20% of post-2010 patients failed over 3 years,” conclude the investigators. “Real-world efficacy is lower than in phase 3 trials. Guidelines should list non-integrase inhibitor-based initial ART as non-preferred. Strategies are needed to improve access to pre-ART genotyping and to increase early initiation of once-daily ART.”
Background: We updated a prior systematic review of initial ART efficacy through Week 144.
Materials and Methods: Studies (1994-July 2017) were drawn from PubMed, ClinicalTrials.gov, Cochrane Library, and major conferences; design, eligibility, subject and ART data were abstracted. Outcomes are expressed as group size-weighted means. Mixed-effects meta-regression was used to identify sources of efficacy heterogeneity.
Results: Within 354 groups (181 studies, 77,999 subjects), principal backbones were tenofovir-emtricitabine (TDF/TAF-FTC) (44.2%), thymidine-based (27.7%), and abacavir-lamivudine (9.7%). Principal anchors were non-nucleoside analogue (49.7%), boosted protease inhibitor (28.1%) and integrase inhibitor (INSTI; 11.5%). Mean intention-to-treat efficacy (RNA<50 copies/mL) was 71.3%, 63.5% (145 groups) and 61.8% (48 groups) at Weeks 48, 96, and 144, respectively (for post-2010 studies, 83.8%, 79.9% and 77.1%). TDF/TAF-FTC and INSTI were independent predictors of greater efficacy at Weeks 48, 96 and 144. Additional independent predictors at Week 48 were pre-ART resistance genotyping, higher baseline CD4 count and once-daily ART. Fewer pills per day predicted greater efficacy at Weeks 96 and 144. Phase-4 studies yielded progressively inferior efficacy than phase-3 studies (difference 5.1% at Week 48, 15.8% at Week 144). Cessation through Week 144 overall (29.4%) and for adverse events (8.9%) declined over time, but cessation for virological failure (5.2%) did not.
Conclusions: Initial ART efficacy continues to improve, but >20% of post-2010 subjects failed over 3 years. Real-world efficacy is lower than in phase-3 trials. Guidelines should list non-INSTI-based initial ART as non-preferred. Strategies are needed to improve access to pre-ART genotyping and to increase early initiation of once-daily ART.
Carr A, Richardson R, Liu Z