Included Studies

Three randomized, active-controlled, noninferiority trials met the inclusion criteria for this systematic review: two double-blind (DB) trials (DRIVE-FORWARD, N = 769^5,6 and DRIVE-AHEAD, N = 728^7,8) conducted in treatment-naive patients; and one open-label (OL) trial (DRIVE-SHIFT, N = 673⁹) conducted in virologically suppressed patients on a stable ARV regimen. The DB and OL trials had a total follow-up duration of 96 weeks and 48 weeks, respectively. Treatments administered in the DB trials were DOR or ritonavir-boosted darunavir (DRV/r), each given in combination with emtricitabine/tenofovir disoproxil fumarate (FTC/TDF) or abacavir/lamivudine (ABC/3TC) (in DRIVE-FORWARD), and DOR/3TC/TDF or efavirenz/FTC/TDF (EFV/FTC/TDF) (in DRIVE-AHEAD). In DRIVE-SHIFT, patients either immediately switched to DOR/3TC/TDF to be received for 48 weeks (immediate switch group, ISG) or continued their baseline regimen for 24 weeks (consisting of a ritonavir- or cobicistat-boosted PI, or cobicistat-boosted InSTI, or NNRTI, each administered with two NRTIs) before switching to DOR/3TC/TDF (delayed switch group, DSG). The primary efficacy outcome in all trials was virologic suppression defined as HIV-1 RNA < 50 copies/mL (calculated using the FDA snapshot algorithm; all missing data were treated as failures regardless of the reasons). In DRIVE-FORWARD and DRIVE-AHEAD the between-treatment differences for the primary efficacy outcome were analyzed at week 48, while in DRIVE-SHIFT the primary analysis compared the proportion of patients maintaining HIV-1 RNA < 50 copies/mL in the ISG group at 48 weeks versus the DSG group at 24 weeks (on baseline regimen). The noninferiority margin (NIM) for the primary outcome was 10% and 8% for the DB and OL trials, respectively. Subgroup analyses were conducted to assess the effect of baseline HIV-1 RNA (< versus ≥ 100,000 copies/mL) on virologic suppression in treatment-naive patients. Secondary end points included changes in lipid levels and neuropsychiatric adverse events (AEs). Baseline patient characteristics and medical and treatment histories were largely similar between-treatment groups. The majority of the patients were male, with a mean age of 26 to 32 years (treatment-naive) and 43 years (treatment-experienced/switch). Across the trials, approximately 20% to 23% had > 100,000 HIV-1 RNA copies/mL (treatment-naive trials only), 9% to 18% had a history of AIDS, less than 5% had hepatitis B and/or C, and 2% to 7% took lipid-lowering therapy.

Limitations noted in the two DB trials are as follows. The comparators used in DRIVE-FORWARD and DRIVE-AHEAD, namely DRV/r and EFV, are less commonly used in a treatment-naive setting according to DHHS guidelines. The clinical expert consulted for this review agreed this is also the case in the Canadian context. Older ARV drugs such as EFV and darunavir are known to be associated with negative neuropsychiatric and gastrointestinal events. DOR may therefore demonstrate a favourable neuropsychiatric and gastrointestinal profile compared with EFV and darunavir, respectively. Among treatment-naive patients, the rate of discontinuation ranged between 13% and 19% at week 48, and between 18% and 29% at week 96 across trials. Notably, the discontinuation rate was higher in the comparator arms than in the DOR arms. Given that those who discontinued the study (including those who discontinued due to AEs) were considered not to have achieved the primary outcome, the comparative efficacy of DOR may be overestimated.

Several important methodological limitations were noted in the switch trial (DRIVE-SHIFT). First, the primary end point used in the switch trial was not consistent with the latest FDA recommendations for HIV drugs. According to the recommendations, the primary efficacy outcome for switch trials should be HIV-1 RNA ≥ 50 copies/mL, as the end point is focused on patients who lose virologic control as a result of switching from a stable, virologically suppressive regimen to another regimen. However, DRIVE-SHIFT was initiated before the new recommendations were published. For the primary efficacy end point, the NIM chosen for DRIVE-SHIFT (8%) was more stringent than the 10% recommended by the FDA, which was used in DRIVE-FORWARD and DRIVE-AHEAD. However, there is some uncertainty regarding whether the 8% NIM for the primary outcome in DRIVE-SHIFT was actually met, as the FDA snapshot algorithm to account for missing data (missing data = failure) was not followed properly. Instead, some patients with missing data at week 48 had their blood samples reanalyzed from other sources and the data were added to the analyses dataset post hoc. Following this modification, the NIM was met for the primary efficacy outcome. However, noninferiority was not initially demonstrated with the true FDA snapshot approach. Finally, analysis of the primary end point was based on an unequal period of exposure to the respective study drugs (DOR/3TC/TDF or baseline regimens). Patients in the ISG arm received DOR for 48 weeks whereas those in the DSG arm received their baseline regimens for 24 weeks followed by DOR/3TC/TDF for 24 weeks. Statistical comparisons were not made between the treatment arms at week 24 for most end points (including the primary efficacy end point), or were not controlled for multiplicity. Instead, results for the ISG arm at week 48 were compared with the DSG arm at week 24 in many cases.

Efficacy

All efficacy analyses were conducted in the full-analysis set, a modified intention-to-treat population that consisted of all randomized patients who received at least one dose of the study medication and had at least one measurement of the outcome (baseline or postbaseline).

Among treatment-naive patients, the primary outcome (proportion of patients with HIV-1 RNA < 50 copies/mL at week 48) was achieved by 83.8% and 79.9% patients receiving DOR and DRV/r in DRIVE-FORWARD, respectively; and by 84.3% and 80.8% patients receiving DOR/3TC/TDF and EFV/FTC/TDF in DRIVE-AHEAD, respectively. The between-treatment differences in the two trials were 3.9% (95% confidence interval [CI], −1.6 to 9.4) and 3.5% (95% CI, −2.0 to 9.0), respectively. In both cases, the pre-specified NIM of 10% was met, as the lower bounds of the 95% CI for treatment differences were above −10 percentage points. Per-protocol analyses supported the conclusion of noninferiority. The proportions of patients with virologic success at week 96 were 73.1% and 66.0% for patients receiving DOR and DRV/r in DRIVE-FORWARD, respectively; and 77.5% and 73.6% for patients receiving DOR/3TC/TDF and EFV/FTC/TDF in DRIVE-AHEAD, respectively. Results from the subgroup analyses indicated a lower virologic success rate in patients with baseline plasma HIV-1 RNA > 100,000 copies/mL compared with those having HIV-1 RNA ≤ 100,000 copies/mL in both DRIVE-FORWARD and DRIVE-AHEAD.

The proportion of treatment-naive patients with HIV-1 RNA ≥ 50 copies/mL (virologic failure) at week 48 using the FDA-defined snapshot approach was similar in both trials’ treatment arms: 11.2% versus 13.1% for DOR and DRV/r, respectively, in DRIVE-FORWARD, and 10.7% versus 10.2% for DOR/3TC/TDF and EFV/FTC/TDF, respectively, in DRIVE-AHEAD. No formal statistical testing was conducted. The proportion of patients with HIV-1 RNA ≥ 50 copies/mL at 96 weeks was 17.2% versus 20.2% for DOR and DRV/r, respectively, in DRIVE-FORWARD, and 15.1% versus 12.1% for DOR/3TC/TDF and EFV/FTC/TDF in DRIVE-AHEAD.

In DRIVE-SHIFT, the proportion of patients with HIV-1 RNA < 50 copies/mL was 90.8% at week 48 in the ISG group compared with 94.6% in the DSG group at week 24, with a treatment difference of −3.8% (95% CI, −7.9 to 0.3). Given the lower bound of the 95% CI was not less than −8%, switching to DOR/3TC/TDF was considered noninferior to continued treatment with baseline regimen. However, DRIVE-SHIFT had a number of methodological issues leading to questionable validity with respect to establishing comparative efficacy between switching to DOR/3TC/TDF versus staying on baseline regimens. The comparison of virologic suppression between groups based on different durations of follow-up is unusual and the CADTH Common Drug Review team is uncertain of the impact this had on the results. Between-treatment comparisons based on the same duration of follow-up would have been more internally valid. The between-treatment difference for the proportion of patients with HIV-1 RNA < 50 copies/mL at the same time point in each group (24 weeks) was −0.9% (95%CI, −4.7 to 3.0); statistical testing did not control for multiplicity. Further, based on guidance from the FDA, the appropriate end point for treatment-switch trials is the proportion of patients with HIV-1 RNA ≥ 50 copies/mL with an associated NIM of 4%. The proportions of patients with HIV-1 RNA ≥ 50 copies/mL were similar between the ISG and DSG at weeks 48 and 24 (1.6% and 1.8% respectively), and between the ISG and DSG at week 24 for each group (1.8% in both groups); between-treatment differences were −0.2 (95% CI, −2.5 to 2.1) and 0.0 (95%CI, −2.3 to 2.3), respectively. However, statistical testing was not controlled for multiplicity.

Among other efficacy end points, CD4 cell counts increased from baseline in all patients, irrespective of treatment arms, time points, and trials. However, between-treatment differences within trials did not reach statistical significance in any case. Resistance to any of the study medications occurred infrequently. Among patients who completed each trial, adherence to treatment was generally high, with most patients (> 85%) reporting an adherence rate of 90% or more. However, it should be noted that the overall adherence among all participants is likely lower when study discontinuation is taken into account. Health-related quality of life was assessed by the visual analogue scale of the EuroQol 5-Dimensions 5-Levels questionnaire in DRIVE-SHIFT only. The mean change from baseline between the ISG and DSG arm was −0.76 and −0.86 at week 48, respectively, and −1.23 and −0.7 at week 24, respectively. Between-treatment difference for the latter time point was −0.54 (95% CI, −3.07 to 2.00).

Harms

The frequency of AEs at week 96 was similar between the treatment arms in the DB trials: 84.6% versus 82.8% among patients receiving DOR and DRV/r, respectively, in DRIVE-FORWARD, and 88.2% versus 93.1% among patients receiving DOR and EFV, respectively, in DRIVE-AHEAD. In the switch trial, 80.3% of patients in the ISG arm receiving DOR through week 48 experienced AEs. A higher proportion of treatment-switch patients receiving DOR reported AEs at week 24. This pattern is consistent with the notion that patients switching therapies are likely to experience more AEs than those remaining on their baseline therapy: 68.9% versus 52.5% among patients receiving DOR and baseline regimens, respectively; and 60.3% of patients in the DSG arm experienced AEs post-switching between week 24 and 48.

Among treatment-naive patients, serious adverse events (SAEs) were reported by 5% to 7% of patients who received DOR, and approximately 8% of those who received DRV/r or EFV. Among treatment-switch patients, 1% to 5% of patients across treatment arms reported SAEs. The proportions of patients who withdrew from the study due to adverse events (WDAEs) were generally low, ranging from 1% to 7% in treatment-naive patients and 0% to 4% in treatment-switch patients. A total of 13 deaths were reported in the three trials, one of which (cause of death: myocardial infarction; patient was in DRIVE-SHIFT and receiving DOR) was attributed to the study drug, although no confirmatory diagnosis (by a medical professional or autopsy) was done.

DOR showed an improvement in lipid profiles among treatment-naive and switch patients at all time points. Two of the five measured lipid parameters were tested with adjustment for multiplicity, fasting low-density lipoprotein cholesterol (LDL), and non–high-density lipoprotein cholesterol (HDL). In DRIVE-FORWARD and DRIVE-AHEAD, fasting LDL and non-HDL levels were decreased in the DOR arms and increased in their respective comparator arms at week 48 in DRIVE-FORWARD and DRIVE-AHEAD; the mean differences for change from baseline in fasting LDL between the treatment arms were −14.6 mg/dL (−18.1 to −11.1) and −10.0 mg/dL (−13.5 to −6.5), respectively; P < 0.0001 in both cases. For change from baseline in non-HDL, the mean differences between the treatment arms were −19.3 mg/dL (95% CI, −23.3 to −15.3) and −17.02 mg/dL (95% CI, −20.9 to −13.2), respectively, with a P value < 0.0001 in both cases. Among treatment-switch patients, those in the ISG had a numerically greater decrease from baseline in fasting LDL and non-HDL at week 24 compared with the DSG; the mean difference was −15.3 (95% CI, −18.9 to −1.6) and −23.9 (95%CI, −28.1 to −19.6), respectively; no P value was reported in either case.

DOR was associated with fewer neuropsychiatric AEs. However, the benefits were largely seen in comparison with EFV in DRIVE-AHEAD, which is commonly associated with neuropsychiatric side effects. Statistical superiority of DOR over EFV was shown for three categories of neuropsychiatric AEs at week 48 in DRIVE-AHEAD: dizziness, sleep disorders and disturbances, and altered sensorium, with between-treatment differences of −28.3 (95% CI, −34.0 to −22.5), −13.5 (95% CI, −19.1 to −7.9), and −3.8 (95% CI, −7.6 to −0.3), respectively. Data for hepatic enzymes, cardiovascular disease, renal and bone-related toxicity, and skin disorders did not show any notable benefits in favour of or against DOR regimens or any of the comparator regimens.

Indirect Treatment Comparison

The manufacturer-submitted network meta-analysis (NMA) suggests that, with respect to virologic success (HIV-1 RNA < 50 copies/mL) ▬ ▬ ▬ ▬ ▬ The NMA did not provide adequate information on the statistical analyses plan to assess the validity of the results and NMA assumptions. The missing information, coupled with the small network size, the failure to assess NMA assumptions, and the differences in trial design and the definition used for protocol-defined virologic failure to determine virologic response, translate to a high degree of uncertainty in the presented efficacy and safety results. Other limitations include the limited scope of the manufacturer-submitted indirect comparison (IDC), where only interventions relevant to their economic model in treatment-naive patients were analyzed, without assessing relevant comparators such as ▬ ▬ ▬ ▬ ▬

Potential Place in Therapy^a

Doravirine, an NNRTI, has some positive attributes compared with the predecessors in its class, including a lack of neuropsychiatric side effects (compared with efavirenz), a lack of a requirement to be taken with food and with normal gastric acidity (unlike rilpivirine), and once-daily dose (unlike etravirine).

Its role will be limited by its late entry into the market. As a single daily-dose “third component” of an ARV combination, it has been preceded to market by rilpivirine, dolutegravir, and DRV/r, among others. As a co-formulated STR, Delstrigo (DOR/3TC/TDF) is one of almost a dozen available single-tablet options, including Atripla (and generics), Complera, Odefsey, Stribild, Genvoya, Triumeq, Biktarvy, Symtuza, and Juluca.

The most commonly prescribed antivirals for treatment-naive patients, or those switching for reasons of convenience or tolerance, are the SRTs, in particular Genvoya and Triumeq. Although each has its own idiosyncracies, most are well tolerated, convenient, and effective. The use of the DOR STR would be infrequent, as the tenofovir component of the Delstrigo STR is the TDF formulation, which is associated with renal and bone toxicities. The newer tenofovir alafenamide formulation, found in Biktarvy and Genvoya, is not associated with these side effects and is generally preferred by prescribing physicians.

As a single component of a regimen, DOR (Pifeltro) would be a reasonable treatment consideration if an STR is not available or an option for a patient. Most likely, it would be used where a tenofovir-containing regimen is not considered ideal, and where side effects of Triumeq have occurred. It would most likely be used with Kivexa (or its generic counterpart). Its use is antipcated to be infrequent.