Accounting for a Non-suppressed Viral Load

Based on clinical trial data, we prioritized the usage of tenofovir with XTC over ABC/3TC except if paired with INSTI

Thymidine Associated Mutations can compromise the activity of some NRTIs.  In heavily treatment experienced patients HIVASSIST conservatively considers the possibility that not all TAMS may be evident on available genotypes.  Consequently, in addition to mutation penalties, we included penalization of regimens with a compromised NRTI backbone further when part of regimens with fewer active drugs with the consideration that any observed NRTI activity may not be fully representative of underlying resistance.  We penalized such regimens more heavily if there was evidence of TAM pathway 1 mutations compared to pathway 2.

1)In the SECOND-LINE trial of 541 patients that had failed a first line NNRTI-based regimen, patients received either LPV/r + NRTI backbone (control) vs LPV/r + RAL, and both had comparable rates of viral suppression at 48 weeks (83% vs 81%)[bib]82[/bib]  

2)Similar results were observed in the EARNEST trial(Paton et al., 2014) of 1277 patients failing first line NNRTI based regimens in sub-Saharan Africa; patients received PI monotherapy vs 2NRTI + LPV/r vs LPV/r + RAL. PI monotherapy did not meet criteria for non-inferiority, but the alternative regimens were non-inferior to each other, with 60% viral suppression and 64% viral suppression in the 2NRTI+PI and PI+INSTI arms, respectively[bib]84[/bib]. 

3)Similar results were again seen in the phase III SELECT trial of 515 patients failing NNRTI based regimens, where LPV/R + RAL was non-inferior to LPV/R +NRTI[bib]86[/bib]

5)Combining NRTIs with other NNRTI’s in patients failing first line NNRTI regimens has been shown to have suboptimal responses. Ruxrunghtam et al showed that in patients experiencing NNRTI failure, receipt of NRTI + ETR had lower rates of virological responses at week 12 compared to a strategy of NRTI + PI[bib]91[/bib] 

We penalized 2 NRTI + RPV if the viral load was high or unknown

Given the low barrier to resistance, and recommendations against usage of RPV based regimens with high viral loads, we penalize regimens containing 2 NRTI+RPV for treatment naive and experienced patients with viral load>100,000

When failing a boosted PI based regimen, failure is often due to poor adherence, drug-drug interactions, or drug-food interactions.

When failing a boosted PI + NRTI regimen, several studies have shown limited resistance to XTC (ARTEMIS study[bib]109[/bib][bib]113[/bib], [bib]110[/bib]).  A systematic review has suggested that maintaining the same regimen, with improved adherence, may be effective [bib]480[/bib]. PI resistance in this setting are also relatively rare ([bib]112[/bib], [bib]113[/bib], [bib]114[/bib]).  As such, our algorithm prioritizes continuing the PI based regimen, or switching to a non-PI regimen with at least 2 fully active agents.

There is currently limited data on managing patients with first line failure of 2NRTIs + 1 INSTI.  In patients receiving 2NRTI + RAL, or 2NRTI + EVG/c, there may be emergent resistance to XTC and possibly the INSTI, but may retain sensitivity to DTG[bib]107[/bib][bib]108[/bib][bib]183[/bib].  Failing in this scenario is often due to poor adherence.  DHHS guidance suggests that one can likely extrapolate based on NNRTI failure datea, suggesting that a boosted PI plus NRTI, or a PI plus INSTI (assuming no INSTI resistance) can be considered.  In the absence of data on usage of BIC/TAF/FTC after treatment failure, we penalized this regimen in the setting of prior failure on an INSTI regimen

Also, see section on INSTI+PI below.

• In the NEAT001/ANRS143 (N = 805) trial of Naive patients, DRV/r plus RAL was noninferior to TDF/FTC+DRV/r, but had lower efficacy in patients with high HIV-1 RNA/low CD4+ cell counts.

• Efficacy in Clinical Trials

  Trial Name	Drugs Compared	Participants	Study Results
  Tivista	DRV/r+DTG (1 arm)	113 tx-experienced	At week 48, the combination of DTG and DRV/r provided a high rate of viral suppression (98.1%), with only one dropout due to drug toxicity [bib]153[/bib].

  Other Evidence in support of 1 INSTI + 1 PI:

  Trial	Regimen	Population	Result
  PROGRESS	TDF/FTC + LPV/r vs LPV/r + RAL	206 tx-naive	The LPV/r + RAL regimen was found to be noninferior to the regimen of LPV/r + TDF/FTC in both safety and efficacy. Viral suppression at 48 weeks 83% in RAL group vs 85% in TDF/FTC+LPV/r group[bib]176[/bib]
  SPARTAN	RAL + ATV vs TDF/FTC+ATV/r	94 tx-naive	Similar rates of viral suppression at 24 weeks, but higher incidence of hyperbilirubinemia and RAL resistance development[bib]177[/bib]
  A5262	DRV/r + RAL	112 tx-naive	26% virologic failure by 48 weeks, with integrase resistance in 5 participants, particularly in those with baseline viral load>100,000[bib]178[/bib]
  RADAR	DRV/r+ RAL vs TDF/FTC+DRV/r	68 tx-naive	DRV/r+RAL was less effective than DRV/r+TDF/FTC (62.5% of RAL subjects and 83.7% of TDF/FTC subjects were responders (VL<48 copies/mL) at week 48) but better for bone health. However, the proportions of patients achieving VL<200 copies/mL were similar: 72.5% and 86.0%[bib]179[/bib]
  ANRS143/NEAT 100	DRV/r+RAL vs TDF/FTC+DRV/r	805 tx-naive	DRV/r+RAL was noninferior for the primary endpoint of time to treatment failure (17.8%vs 13.8% at 96 weeks); planned subgroup analysis showed RAL arm was inferior for those with CD4<200.[bib]144[/bib]

Overall, the goal of treatment in ART experienced patients experiencing virologic failure is to establish virologic suppression (DHHS Guidelines 2019).  DHHS panel recommends that "ideally, a new ARV regimen should contain at least two, and preferably three, fully active drugs, which should be selected after considering the patient’s ART history and current and previous resistance test results and whether an ARV drug with a new mechanistic action is available (AI). When constructing a salvage regimen, it is more important to consider drug potency and viral susceptibility based on cumulative genotype data than the number of component drugs."  

In general, we prioritized drugs with higher barrier to resistance (DTG, DRV), and escalated penalties for fewer active drugs in individuals with a history of treatment failure.  We note the OPTION to further augment these penalties in individuals with intermittent adherence--when these options are selected, HIVASSIST prioritizes inclusion of regimens with more active drugs.  

SALVAGE: In situations where there is multi-class resistance, we considered the inclusion of drugs with partial or no activity. The guidelines note that partially active drugs may be used when no other options are available. In such situations, there may be benefit to even partial reduction in viral load.  Studies have suggested that even partial virologic suppression can have clinical benefit in terms of delayed disease progression.  In general, we did not consider there to be sufficient data to suggest inclusion of a resistant NNRTI.  We allowed inclusion of boosted PI's and INSTIs and some NRTIs in such situations.  

In the DAWNING trial, 627 patients failing first line NNRTI regimens (> 6 months) with at least one active NRTI were randomized to 2WHO recommended NRTIs + DTG VS. 2NRTI's +LPV/RTV.  WHO recommend ZDV + 3TC after failure of first-line TDF + 3TC or FTC; TDF + 3TC or FTC after first-line ZDV or d4T + 3TC.At Week 24, DTG+2NRTIs was superior to LPV/r+2NRTIs, with 82%  and 69%, respectively, achieving HIV-1 RNA <50 c/mL (p<0.001). The difference was mainly driven by lower rates of Snapshot virologic non-response in the DTG group. 56% (347/624) of subjects received WHO-recommended second-line NRTIs, and their response rates within each arm were higher than those for subjects who did not. Regardless of WHO-recommended NRTI use, response rates were higher with DTG versus LPV/r-based regimens (Table).  In this analysis, there were no treatment-emergent primary integrase-strand transfer inhibitor or NRTI resistance mutations in the DTG group through the randomisation phase.

In a Swedish cohort of 244 patients with preexisting NRTI's and DTG (with a control cohort of PLWH on PI/r and one or two NRTI's with matched genotype susceptibility scores) were followed.  Median observation time was 78 weeks (interquartile range 50–98 weeks) for participants on DTG and 75 weeks (50–101 weeks) for individuals on PI/r. Viral failure was detected in four individuals treated with DTG and three individuals treated with PI/r, resulting in similar success rates of 96.7% and 97.5%, respectively. No new mutations were found among participants with treatment failure. DTG in combination with one to two NRTIs was as efficient as PI/r in individuals with pre-existing NRTI mutations in this setting. (Sorstedt et al. May 2018 IJAA)

In the POWER1 and POWER2 studies, individuals with a history of treatment failure were randomized to varying PI based regimens. Approximately 83% of those with baseline viremia <20,000 suppressed, with overall 45% suppression and 61% with at least one log decline at 48 weeks when given DRV/b (with lower response rates with other PI regimens).  In particular there was a higher rate of achieving undetectable viral load when there was <=1 primary DRV mutation, but 26% also achieved virologic suppression despite >=3 DRV associated mutations. Consequently we prioritized inclusion of DRV in individuals with a history of treatment failure.[bib]482[/bib]