Introduction

Lamivudine and emtricitabine are both widely used as a core component of the dual nucleoside reverse transcriptase inhibitor backbone in all currently preferred first-line antiretroviral combinations therapies The chemical structure of these two nucleoside analogues is very similar[1], [2]; both are prodrugs requiring intracellular phosphorylation and both are active against HIV-1, HIV-2 and hepatitis B virus.

The latest antiretroviral treatment guidelines of the US Department of Health and Human Services [3] and the World Health Organization[4] consider lamivudine and emtricitabine to be equivalent and interchangeable from a clinical and programmatic perspective. However, inferior virological efficacy of lamivudine has been suggested based on limited data from early in-vitro studies[5], [6] and this presumption of inferiority has been applied to recent cost-effectiveness analyses [7]. There is therefore uncertainly regarding the clinical comparability of these two drugs.

In order to support recommendations for future guidance for first-line antiretroviral therapy, we conducted this systematic review of available data from randomized trials to assess the comparative efficacy of these two antitretroviral drugs.

Methods

This systematic review was conducted according to the according to the criteria of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses group [8].

Results

1756 titles were screened for indirect comparisons and 1115 titles for direct comparisons. No valid indirect comparisons were identified. After excluding duplicates, 38 articles were read in full and 26 were excluded for one or more reasons, detailed in Figure 1. Among these, five trials were excluded for using non-comparable background regimens[14]–[18], one trial was excluded because all patients had high viral load (≥100,000 copies/ml) at baseline [19], and two non-randomized trials were excluded[20], [21]. One unpublished study was identified from bibliography screening[22].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Study flow chart.

https://doi.org/10.1371/journal.pone.0079981.g001

Twelve trials were included in the final review[9], [22]–[31]. In total, 15 different randomized comparisons providing data on 2251 patients receiving lamivudine and 2662 patients receiving emtricitabine. Studies were published between 2002 and 2013. Five studies were done in treatment-naïve patients. Three trials[22], [23], [32] had the same backbone regimens; the rest compared tenofovir and abacavir. Two trials included some patients with high viral loads at baseline (ie ≥100,000 copies/ml); only the results for those patients in the low viral load strata (<100,000 copies/ml) were included in the meta-analysis [9], [27]. Study characteristics are summarized in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Study Characteristics.

https://doi.org/10.1371/journal.pone.0079981.t001

Treatment success was not significantly different in any of the 12 trials. In the three trials that directly compared lamivudine and emtricitabine [22], [23], [32], the relative risk for achieving treatment success was non-significant for both trials (RR 1.03, 95%CI 0.96–1.10; P = 0.3). Overall, the pooled relative risk for treatment success was non-significant (relative risk, 1.00, 95% CI 0.97–10.2) (Figure 2). No heterogeneity was observed (I² = 0). This result was not different in any of the pre-defined subgroups (test for heterogeneity for all subgroups: p>0.1), or if random-effects methods were used to pool the data (RR 0.99, 95%CI 0.96–1.01).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Virological suppression comparing 3TC and FTC-including regimens.

https://doi.org/10.1371/journal.pone.0079981.g002

Similarly, all but one study[28] found no difference in the risk of treatment failure, and the pooled relative risk was not statistically significant (RR 1.08, 95%CI 0.94–1.22). Heterogeneity was low (I² = 3.4%), and no subgroup differences were apparent (P>0.1 for all subgroups).

Four trials provided data on the emergence of MI84V resistance mutations among virologically failing patients (n = 234) [9], [22], [23], [29]. Two of these trials genotyped all patients experiencing virological failure, and found no difference by regimen[9], [23]. The other two trials reporting resistance data did so only on a subset of virologically failing patients, and these studies reported an increased risk of MI84V resistance mutation development among patients receiving lamivudine. The overall pooled estimate, using a random-effects model, was not significant (RR 1.4l, 95%CI 0.6–3.3) but this result should be interpreted with caution due to high heterogeneity (I² = 80%) and the selective reporting of this outcome in some of the trials.

Finally, two of the three trials with identical backbone regimens provided data on adverse events [22], [23]. In trial FTC302, no difference in the incidence of any grade 3 or 4 adverse event was reported. In trial FTC-303/350, 4% of patients discontinued treatment due to adverse events in the FTC arm and there were no discontinuations in the 3TC arm.

The GRADE assessment rated the quality of the evidence overall to be moderate (Table S1). Risk of bias was judged to be low (Table S2) and there was no evidence of publication bias (p = 0.3 using Egger's test for funnel plot asymmetry). Results of all studies were consistent for the critical outcomes of virological suppression and failure. Concern was noted with respect to possible indirectness resulting from the inclusion of trials with non-identical backbone regimens[33] but the direction of this bias would be expected to favour emtricitabine[34].

Discussion

This systematic review of published and unpublished data from randomized trials found no significant differences in the efficacy of lamivudine and emtricitabine, consistent with very similar chemical structure of these two nucleoside analogues. Three of the 12 included trials in this review were identical in terms of background regimens, and the rest included regimens that are considered to be of equivalent efficacy. In the past, presumption of superior efficacy of emtricitabine has partly been based on the results of randomized trials that used different backbone regimens resulting in higher terminations in the lamivudine group due to adverse events associated with partner antiretrovirals [35], [36]. Such comparisons were excluded from this review.

Concern has also been expressed about the potential greater risk of development of MI184V resistance mutations among virologically failing patients. This review was unable to provide conclusive evidence in this regard. However, differences appear to be small, and the clinical importance of the M184V mutations is unclear; this mutation appears to be associated with reduction of viral fitness[37] and increased susceptibility to zidovudine, which is recommended as part of the preferred second-line regimen for patients in whom virological failure is confirmed[38].

The largest randomized trial included in this review, A5202, found no difference in efficacy comparing tenofovir and emtricitabine vs abacavir and lamividine in patients with low baseline HIV RNA (<100,000 copies/ml) but superior outcomes favouring tenofovir+emtricitabine at higher viral load for patients whose baseline viral load was above this level [9]. Possible reasons for this difference may include marginally superior antiviral activity of emtricitabine that becomes apparent when adherence is poor; minority species of some mutations leading to increased resistance to abacavir and enhanced susceptibility to tenofovir; and pharmacokinetic differences due to longer intracellular half lives of both tenofovirand emtricitabine compared to abacavir and lamividine. Further research is needed to understand the relative contribution of the different drugs to these findings. Another recent study compared dolutegravir combined with abacavir and lamividune against tenofovir combined with emtricitabine and efavirence and found superior efficacy and safety favouring the dolutegravir-containing regimen [39]. We did not include these results in our review as it was considered that backbone regimens were too different to determine the specific contribution of emtricitabine or lamivudine to these results.

Strengths of this review include a broad search strategy that allowed for the identification of published and unpublished trials, the restriction of inclusion to randomized trials, and the inclusion of comparable background regimens. The inclusion of outcome data from over 4500 randomizations allowed for a precise estimate of effect giving confidence in the overall result, and heterogeneity was neither detected nor apparent. The main limitation was the inclusion of studies that used background regimens that are not identical. Some studies have suggested superior efficacy of tenofovir compared to abacavir [40], although differences are not apparent in patients starting treatment with a low viral load [9]. The aim of this review was not to compare these regimens, but rather to identify studies in which the comparative efficacy of lamivudine and emtricitabine could be assessed because the efficacy of partner drugs were identical or could be considered comparable, and as such we excluded any studies in which patients started therapy with a high viral load (ie ≥100,000 copies/ml). Given the differences in safety profiles of the various background regimens, we did not report discontinuations due to adverse events as a primary outcome, but note that the frequency of adverse events was similar in the three trials in which backbone regimens were identical. We also specifically assessed differences in those studies that used identical partner drugs and explored differences formally in subgroup analysis, which found no apparent difference. Bias that may be introduced as a result of including these trials would be expected to favour emtricitabine (the drug partnered with tenofovir) and the fact that no differences are seen is therefore reassuring. Publication bias can never be ruled out, as evidenced by the non-publication of one of the few trials to directly compare lamivudine and emtricitabine (this study was terminated early by the South African Medicines Control Council and placed on clinical hold by the US FDA). We found no statistical evidence of publication bias, but such tests are poorly powered, particularly when the number of publications is small. Finally, a major limitation in the current literature is the lack of information regarding the risk of risk of 184V selection, and future research is encouraged to continue to seek new evidence to further clarify this issue.

In light of these limitations, the results of this review should not be understood as definitive evidence of equivalence. Nevertheless, the overall findings provide supportive evidence for the recommendations of current international and national treatment guidelines to treat emtricitabine and lamivudine as interchangeable and reassurance to countries that, for reasons of affordability or availability [41] have opted for lamivudine as part of first line antiretroviral therapy.

Supporting Information