Choices for NRTI pairs currently recommended by US and European guidelines include abacavir/lamivudine, emtricitabine/tenofovir DF, and emtricitabine/tenofovir AF. Lamivudine/zidovudine has largely fallen out of favor due to hematologic toxicity. Choice of the NRTI pair may be influenced by choice of the third agent in the regimen; other factors include patient characteristics and comorbidities, drug–drug interactions, and baseline resistance. When intolerance to all 3 recommended NRTI pairs is present, guidelines may recommend a nontraditional 2-drug NRTI-limiting regimen.¹

Fixed-dose formulations of NRTI pairs are available for all of the currently recommended NRTI pairs. In addition,

• Abacavir/lamivudine is coformulated with dolutegravir
• Emtricitabine/tenofovir DF is coformulated with cobicistat/elvitegravir, efavirenz, and rilpivirine
• Emtricitabine/tenofovir AF is coformulated with bictegravir, cobicistat/darunavir, cobicistat/elvitegravir, and rilpivirine
• Lamivudine/tenofovir DF is coformulated with doravirine

. . . to form complete 1-pill ARV regimens. These coformulations reduce pill burden and may be associated with lower pharmacy copays.

Abacavir/Lamivudine
Abacavir/lamivudine is available in a once-daily fixed-dose combination tablet and has been demonstrated to be associated with low in vitro mitochondrial toxicity. In a prospective randomized study, abacavir/lamivudine was compared with lamivudine/zidovudine in 649 treatment-naive patients, with all study subjects also receiving efavirenz. The results demonstrated similar rates of virologic suppression at Week 48 (70% of abacavir-treated patients vs 69% of zidovudine-treated patients with HIV-1 RNA < 50 copies/mL), and abacavir/lamivudine-treated subjects experienced greater mean CD4+ cell count increase (209 cells/mm³ in the abacavir group vs 155 cells/mm³ in the zidovudine group).³⁸ A randomized open-label trial of either abacavir or stavudine with lamivudine plus efavirenz found higher rates of virologic suppression and lower rates of lipoatrophy, dyslipidemia, and drug discontinuation in the abacavir arm.³⁹

Abacavir/lamivudine is part of recommended regimens in US and European guidelines only in combination with dolutegravir and only for HLA-B*5701–negative patients.^1,2,4 The guidelines advise using caution with or considering avoiding abacavir in patients with or at high risk for cardiovascular disease.

Emtricitabine/Tenofovir DF
The phase III GS934 study compared emtricitabine/tenofovir DF vs lamivudine/zidovudine, both in combination with efavirenz, in 517 treatment-naive patients.⁴⁰ At 48 weeks, this study demonstrated superior virologic activity with emtricitabine/tenofovir DF vs lamivudine/zidovudine patients (84% vs 73% with HIV-1 RNA < 400 copies/mL, respectively; P = .002) and better safety and tolerability. More lamivudine/zidovudine-treated patients required drug discontinuation (generally for anemia) than those receiving emtricitabine/tenofovir DF and, furthermore, had a greater loss of limb fat as measured by dual energy x-ray absorptiometry scanning.

Because of the results of the GS934 study, emtricitabine/tenofovir DF has become the NRTI pair most commonly used in prospective clinical trials comparing various treatment strategies.^15,41-43 It is one of the tenofovir-containing NRTI pairs suggested for use by the DHHS guidelines¹ and the EACS guidelines.⁴ The IAS-USA guidelines currently recommend emtricitabine paired with tenofovir AF over tenofovir DF due lower plasma levels of tenofovir and higher intracellular concentrations of tenofovir diphosphate, the active antiviral component, seen with tenofovir AF.² These guidelines do state that in cases where tenofovir AF is not available, tenofovir DF is an effective and generally well-tolerated option; however, they note that tenofovir DF is not recommended for patients with or at high risk for osteopenia/osteoporosis or renal impairment.²

Emtricitabine/Tenofovir AF
Tenofovir AF in combination with other ARV agents is indicated for the treatment of HIV infection.^12,44,45 It was first approved by the FDA in late 2015 as part of the once-daily, single-tablet coformulated regimen, cobicistat/elvitegravir/emtricitabine/tenofovir AF. The dose of tenofovir AF in the coformulation is 10 mg. It is also now available for the treatment of HIV in 4 other coformulations: the NRTI pair emtricitabine/tenofovir AF and the once daily, single-tablet coformulated regimens emtricitabine/rilpivirine/tenofovir AF, bictegravir/emtricitabine/tenofovir AF, and cobicistat/darunavir/emtricitabine/tenofovir AF.

Tenofovir AF is largely delivered as tenofovir AF to lymphocytes and macrophages and then metabolized intracellularly to tenofovir and phosphorylated to the active tenofovir diphosphate. This pathway is in contrast with that of tenofovir DF, which is converted to tenofovir in the blood and then taken up by cells. With tenofovir AF, this results in higher intracellular levels of the active moiety, tenofovir diphosphate, and lower plasma levels of tenofovir than are achieved with use of tenofovir DF. The distinct pharmacokinetic properties of tenofovir AF allow for lower dosing compared with tenofovir DF, providing a mechanism for reduced tenofovir-related toxicity.⁴⁶

The FDA approval of the 2 NRTI coformulation emtricitabine/tenofovir AF for use in treatment-naive patients is based on data from studies in which these agents were evaluated in the context of the single-tablet regimen, cobicistat/elvitegravir/emtricitabine/tenofovir AF. Additional supportive data on the safety and efficacy of this 2 NRTI coformulation come from switch studies involving virologically suppressed patients.

Comparison of Emtricitabine/Tenofovir DF With Abacavir/Lamivudine
Abacavir/lamivudine and emtricitabine/tenofovir DF were compared in ACTG 5202, a randomized study in which 1858 patients started ART with either efavirenz or atazanavir plus ritonavir, each combined with a blinded NRTI backbone.^29,47 Patients were stratified at baseline for HIV-1 RNA ≥ or < 100,000 copies/mL. An interim efficacy review by the data and safety monitoring board found that among patients who had baseline HIV-1 RNA ≥ 100,000 copies/mL, there was an excess of virologic failures in the abacavir/lamivudine arm vs the emtricitabine/tenofovir DF arm (14% vs 7%; HR: 2.33; 95% CI: 1.46-3.72; P = .0003) (Table 5). The subsequent full results reported that the underperformance of abacavir/lamivudine in the higher HIV-1 RNA stratum was observed regardless of whether it was paired with efavirenz or atazanavir plus ritonavir.⁴⁷ In this same patient stratum, the data and safety monitoring board also found that a significantly larger number of patients in the abacavir/lamivudine arm met study safety endpoints.²⁹ Although screening for HLA-B*5701 was not conducted, there was no association between suspected drug hypersensitivity reaction and virologic failure.

Table 5. ACTG 5202 Interim Analysis: Virologic Failure With Abacavir/Lamivudine vs Emtricitabine/Tenofovir DF Among Patients With Baseline HIV-1 RNA ≥ 100,000 copies/mL²⁹

Since the comparison between abacavir/lamivudine and emtricitabine/tenofovir DF in the high HIV-1 RNA stratum was stopped, the final 96-week primary endpoint analysis between the NRTIs focused primarily on patients with screening HIV-1 RNA < 100,000 copies/mL (as well as the comparison of efavirenz and atazanavir/ritonavir will be described in a later section. This analysis showed a similar time to virologic failure with abacavir/lamivudine vs emtricitabine/tenofovir DF, regardless of assignment to atazanavir plus ritonavir (HR: 1.26; 95% CI: 0.76-2.05) or efavirenz (HR: 1.23; 95% CI: 0.77-1.96).⁴⁷ The investigators were unable to prove equivalence as defined in study protocol, however, due to the unexpectedly low overall number of virologic failures in this stratum.

There was a significantly shorter time to first grade 3/4 safety event with abacavir/lamivudine vs emtricitabine/tenofovir DF, when each was combined with efavirenz (HR: 1.38; 95% CI: 1.03-1.85; P = .03), but no significant difference when each combined with atazanavir plus ritonavir (HR: 1.13; 95% CI: 0.83-1.54; P = .44). There was also a significantly shorter time to treatment modification with abacavir/lamivudine vs emtricitabine/tenofovir DF, regardless of assignment to atazanavir plus ritonavir (HR: 1.43; 95% CI: 1.06-1.92; P = .018) or efavirenz (HR: 1.48; 95% CI: 1.12-1.95; P = .005). This seemed to be related to an increased number of suspected hypersensitivity reactions in the abacavir-containing arms; importantly, HLA-B*5701 testing was not required for study entry and was performed in only a small fraction of patients. There were also significantly greater lipid increases with abacavir/lamivudine vs emtricitabine/tenofovir DF, regardless of assignment to atazanavir plus ritonavir or efavirenz.

Besides the ACTG 5202 study, another randomized study comparing the efficacy and safety of emtricitabine/tenofovir DF vs abacavir/lamivudine on a head-to-head basis is the HEAT trial. In this study, 688 treatment-naive patients were randomized to receive open-label therapy with one of these 2 NRTI coformulations plus lopinavir/ritonavir once daily.⁴⁸ At Week 96, the efficacy of abacavir/lamivudine was noninferior to that of emtricitabine/tenofovir DF, and 60% vs 58% of patients had HIV-1 RNA < 50 copies/mL, respectively. There were no differences in virologic suppression rates between arms when patients were stratified by high vs low baseline HIV-1 RNA (< or ≥ 100,000 copies/mL). HLA-B*5701 screening was not available before trial entry. The ASSERT study compared open-label abacavir/lamivudine with emtricitabine/tenofovir DF in 385 HLA-B*5701-negative patients.⁴⁹ Although the study was only powered to show differences between the NRTIs related to estimated glomerular filtration rate, the rates of virologic failure were higher in the abacavir/lamivudine than the emtricitabine/tenofovir DF arm, with 59% vs 71% having HIV-1 RNA < 50 copies/mL at 48 weeks. The double-blind SINGLE study compared dolutegravir plus abacavir/lamivudine with efavirenz/emtricitabine/tenofovir DF in 385 HLA-B*5701–negative patients. The dolutegravir plus abacavir/lamivudine regimen demonstrated superior virologic suppression (88% vs 81% patients achieving HIV-1 RNA < 50 copies/mL; P = .003), mean time to viral suppression (28 vs 84 days; P < .001), and increases in CD4+ cell counts (267 vs 208 cells/mm³; P < .001) at Week 48.¹⁷ The superior efficacy of the dolutegravir plus abacavir/lamivudine regimen was maintained through Week 144 with 71% vs 63% of patients achieving HIV-1 RNA < 50 copies/mL.⁵⁰

A summary of the 4 prospective randomized trials comparing abacavir/lamivudine and emtricitabine/tenofovir DF is shown in Table 6.

Table 6. Comparison of 4 Randomized Trials Including Abacavir/Lamivudine and Emtricitabine/Tenofovir DF^17,29,48,49

In addition to these results, notable differences between the A5202 and HEAT studies include sample size, different third-drug strategy, definition of study endpoints, and loss to follow-up rates. Despite these differences, major HIV treatment guidelines cite the ACTG 5202 results as a reason for avoiding abacavir/lamivudine in patients with high baseline HIV-1 RNA, except when coadministered with an INSTI.¹

Safety Considerations With Tenofovir DF and Abacavir
Tenofovir DF. Nephrotoxicity is the most important adverse event associated with tenofovir DF treatment. In general, prospective clinical trials and cohort studies have shown a low rate of clinically significant renal dysfunction in patients who begin emtricitabine/tenofovir DF therapy with normal or near normal renal function. For example, in study GS934 creatinine clearance (as estimated by the Cockcroft-Gault equation) remained within normal range for patients in both the emtricitabine/tenofovir DF and lamivudine/zidovudine arms, and there were no discontinuations because of renal complications in either group.⁴⁰

Although reassuring, these data may not apply to individuals who begin tenofovir DF with preexisting renal disease; such patients were excluded from participating in this study. Indeed, observational studies have shown that patients with renal disease or advanced HIV infection are at greater risk of developing renal insufficiency on tenofovir DF–based therapy, although the isolated contribution of the tenofovir DF therapy may be difficult to ascertain.⁵¹

An analysis of the EUROSIDA cohort demonstrated increased risk of chronic kidney disease with cumulative exposure to atazanavir, indinavir, lopinavir/ritonavir, and tenofovir DF.⁵² A similar analysis of the ANRS Aquitaine Cohort found an increased risk of chronic kidney disease in patients who had received tenofovir DF or a PI for more than 6 months, with the greatest increase in risk among those receiving concomitant therapy with tenofovir DF and a PI.⁵³

Nonetheless, clinicians should monitor serum creatinine and creatinine clearance periodically for all patients receiving tenofovir DF and should be especially vigilant in patients with preexisting renal disease or risk factors for this condition, such as hypertension, diabetes, or use of concomitant nephrotoxic agents. In general, tenofovir DF should be avoided in patients with significantly impaired baseline renal function; if tenofovir DF must be used, dose reduction is recommended, using the strategies outlined in the FDA-approved label. In addition, tenofovir DF should be used with caution or not at all in patients taking other nephrotoxic drugs.

ACTG 5224, the metabolic substudy of ACTG 5202, measured changes in bone density and limb fat among the arms in the study.⁵⁴ There was an initial loss in bone mineral density measured in all study arms that stabilized by Week 48. There were significantly greater losses in both mineral density in both the lumbar spine (difference: 2.0%; P = .004) and hip (difference: 1.5%; P = .025) among those who received emtricitabine/tenofovir DF than in those who received abacavir/lamivudine. There were no significant differences in fracture rates among the study arms. Regarding the limb fat analysis, there were similar absolute and percentage increases in limb fat with abacavir/lamivudine and emtricitabine/tenofovir DF in the intent-to-treat analysis (P > 0.1). An analysis of the Veterans Affairs’ Clinical Case Registry found a higher risk of osteoporotic fracture among patients who received concomitant tenofovir DF and a boosted PI.⁵⁵

Abacavir. Treatment with abacavir may lead to a potentially life-threatening hypersensitivity reaction (HSR). In prospective studies, abacavir HSR occurred in 5% to 8% of study subjects, generally within the first 6 weeks of treatment.⁵⁶ Continued use of abacavir in the presence of HSR or rechallenge with abacavir after discontinuation of the drug for HSR can be fatal. The risk for HSR is strongly linked to the presence of the HLA-B*5701 allele, and clinical studies have shown that a negative test for this allele virtually eliminates the risk for severe HSR.⁵⁷ As a result, all patients should be screened for HLA-B*5701 before receiving abacavir, a strategy that in one study was found to be highly cost-effective.⁵⁸

An additional toxicity potentially linked to abacavir therapy is an increased risk for cardiovascular disease. The D:A:D study is a large, multinational cohort study focusing on adverse events associated with ART. In an analysis involving more than 33,000 participants and more than 150,000 person-years of follow-up, recent use of abacavir was associated with a 1.9-fold increased risk for myocardial infarction (MI).⁵⁹ Follow-up data from this cohort also implicated cumulative use of abacavir but did not show a similar association with tenofovir DF.⁶⁰ In addition, a subsequent analysis of the D:A:D cohort showed that despite channeling of abacavir among patients with high cardiovascular risk after 2008, there was still a strong association of abacavir and risk of first MI.⁶¹ However, among people with a previous MI, there was no evidence for an association between use of abacavir post-MI and an elevated risk of a recurrent MI.⁶² Another study, a meta-analysis of 7 clinical cohorts within the NA-ACCORD, found an increased risk for MI with recent use of abacavir (HR: 1.95). However, this risk was no longer statistically significant (HR: 1.33; 95% CI crossed 1.00) when the analysis accounted for traditional cardiovascular and HIV-associated risk factors (Capsule Summary).⁶³ There have been many additional studies providing data showing a link between abacavir and risk of first MI, but it is important to note that this link has not been found in all reviews of data.

Although the link between abacavir use and myocardial infarction has not been proven, given the heterogeneity of the above study results, avoiding or using with caution abacavir-containing regimens in patients at elevated risk for cardiovascular events is recommended in current guidelines.^1,2,4

Comparison of Tenofovir DF and Tenofovir AF When Each Is Coformulated With Cobicistat/Elvitegravir/Emtricitabine
In the 2 phase III studies that compared cobicistat/elvitegravir/emtricitabine/tenofovir AF vs cobicistat/elvitegravir/emtricitabine/tenofovir DF in treatment-naive patients, pooled data from both trials demonstrated noninferior Week 48 virologic efficacy of the tenofovir AF coformulation (adjusted treatment difference: 2.0%; 95% CI: -0.7% to 4.7%).⁶⁴ A combined safety analysis demonstrated significantly smaller increases in serum creatinine, less proteinuria, and smaller spine and hip bone mineral density decreases with tenofovir AF vs tenofovir DF, although there was a greater increase in fasting lipids in patients treated with tenofovir AF.⁶⁴ The most frequent adverse events reported by patients receiving cobicistat/elvitegravir/emtricitabine/tenofovir AF were nausea (10%) and diarrhea (7%). At the 96-week analysis, tenofovir AF remained noninferior to tenofovir DF in terms of efficacy and favorable trends in bone and renal measures persisted.⁶⁵ There were also no cases of tubulopathy reported in patients receiving tenofovir AF compared with 2 cases in patients receiving tenofovir DF. Further investigation into renal outcomes at 96 weeks in these trials revealed: 1) an even greater benefit of tenofovir AF compared with tenofovir DF in patients with high risk of chronic kidney disease (ie, smaller change in median estimated glomerular filtration rate by Cockcroft-Gault, proteinuria, albuminuria, and fewer cases of incident chronic kidney disease) (Capsule Summary), and 2) fewer clinically significant renal events and discontinuations due to renal dysfunction among all patients.^66,67 At the Week 144 analysis, efficacy of tenofovir AF was superior to that of tenofovir DF (84% vs 80%; P = .021), there were fewer renal discontinuations (0 vs 12; P < .01)⁶⁸ and significantly less deleterious effects on renal biomarkers and bone mineral density vs the tenofovir DF regimen (Capsule Summary).⁶⁹ However, the tenofovir AF regimen was associated with significantly increased levels of fasting lipids over this time period.

In a similarly designed phase II study, 170 treatment-naive patients were randomized to receive coformulated cobicistat/elvitegravir/emtricitabine/tenofovir DF or coformulated cobicistat/elvitegravir/emtricitabine/tenofovir AF.⁷⁰ At Week 48, approximately 88% of patients in both treatment arms had HIV-1 RNA < 50 copies/mL. Use of the tenofovir AF–containing regimen resulted in significantly smaller decreases in estimated creatinine clearance, and significantly smaller decreases in spine and hip bone mineral density compared with the tenofovir DF regimen. Of note, there have not been any cases of proximal renal tubulopathy or Fanconi syndrome in cobicistat/elvitegravir/emtricitabine/tenofovir AF clinical trials.¹² Among patients with baseline estimated glomerular filtration rates of 30-49 mL/min who were treated with cobicistat/elvitegravir/emtricitabine/tenofovir AF in a switch trial, only 3% (2/80) required permanent discontinuation of the regimen because of worsening renal function, and one of the 2 patients was deemed to have hypertension-related chronic kidney disease progression not related to the study medication.⁷¹

Taken together, these data support a more favorable renal and bone safety profile of tenofovir AF relative to tenofovir DF, with similar virologic efficacy. However, both cobicistat/elvitegravir/emtricitabine/tenofovir AF and cobicistat/elvitegravir/emtricitabine/tenofovir DF are recommended as alternative first-line HIV regimens in the DHHS guidelines,¹ the EACS guidelines,⁴ and the IAS-USA guidelines.² Cobicistat/elvitegravir/emtricitabine/tenofovir DF should be reserved for those with estimated glomerular filtration rates ≥ 70 mL/min¹¹ and cobicistat/elvitegravir/emtricitabine/tenofovir AF should be used only in patients with estimated glomerular filtration rates ≥ 30 mL/min.¹² The EACS guidelines further recommend that cobicistat/elvitegravir/emtricitabine/tenofovir DF not be given to patients with estimated glomerular filtration rates < 90 mL/min unless this is the preferred treatment.⁴ It should be noted that many physicians (including this author) prefer the tenofovir AF–containing regimen due to improved renal and bone safety when compared with the tenofovir DF–containing regimen. In the opinion of this author, clinicians should use the tenofovir AF–containing formulation in all settings where the single-tablet elvitegravir-based regimens are selected.