HIV-1 drug resistance testing is essential for heavily-treated patients switching from first- to second-line regimens in resource-limited settings: evidence from routine clinical practice in Cameroon

By the end of 2017, about 70% of HIV-infected people knew their status, 77% of these were receiving antiretroviral therapy (ART) and 82% of treated patients achieved viral suppression [1, 2]. This progress led to 54% coverage of ART (14 million people) in Africa, with a coverage that is expected to double by 2020 as per the 90–90 − 90 targets (90% of all people living with HIV have been diagnosed; 90% of all people with diagnosed HIV infection are receiving ART; 90% of all people receiving ART have suppressed viral load) [2].

A major setback in achieving viral suppression is the emergence of HIV drug resistance (HIVDR), which is driven by delayed detection of virological failure and accumulation of drug resistant mutations (DRMs) in resource-limited settings (RLS). This later increases cases of new HIV-infections with resistant viruses, as well as AIDS-associated-morbidity and mortality [2, 3]. While HIVDR is gradually under control in high-income countries, the resistant patterns are rising in RLS and especially within sub-Saharan Africa (SSA) [3]. In these settings, one in every 10 adults starting ART harbour resistant virus; three in every 10 adults restarting first-line ART (i.e. women exposed to antiretrovirals for the prevention of mother-to-child transmission (PMTCT) or individuals re-initiating first-line ART after a period beyond 3 months of treatment interruption) harbour resistant virus; and five in every 10 young children diagnosed with HIV harbour resistant virus [3]. Regarding adults living in different SSA regions, resistance rates to first-generation non-nucleoside reverse-transcriptase inhibitors (NNRTIs), i.e. nevirapine and efavirenz, are already beyond 10% in East and Southern Africa regions [4], thus requiring treatment initiation with NNRTI-sparing regimens for optimal ART response in these settings, compared to West and Central Africa regions where the burden may still be bearable [4, 5]. However, the estimated incremental annual increase of 17% pre-treatment drug resistance (PDR) in West and Central Africa settings over time (p = 0.0017), the implementation of the “treat-all” strategy and the continuous use of NNRTI-based first-line ART [2], virological failure (VF) and acquired HIVDR to current first-line ART would be cumulative [5, 6], especially for countries like Cameroon where the national estimates of NNRTI-PDR are near the 10% critical threshold: 7.8% by Fokam et al. in 2017 and 8.3% by WHO in 2017 [7, 8].

With the complete phased-out of stavudine (D4T) from current first-line ART in RLS, tenofovir (TDF)-containing regimens are now widely used while patients with contraindication to TDF are generally being prescribed zidovudine (AZT)-containing regimens as an alternative [9, 10].

After VF on a standard first-line ART in RLS, an optimal selection of SLC is challenging due to limited access to HIVDR testing [2, 5, 6]. In such situations, 3TC could be recycled as part of the NRTI-backbone in SLC because it selects for the M184 V mutation, which decreases viral replication and increases susceptibility to AZT and TDF for potential use in SLC with PI/r [10, 11]. In this frame, the WHO guidelines also recommend the use of AZT + 3TC as NRTI backbone for SLC for those patients failing on TDF-containing first-line regimen; likewise, TDF + 3TC is recommended as NRTI backbone in SLC for patients failing on AZT-containing first-line regimen [9, 11]. Of note, failure on TDF leads to viral hyper-susceptibility to AZT (due to K65R mutation) while failure on AZT leads to possible cross-resistance to TDF. Moreover, in routine clinical practice, patients may experience several NRTI substitutions on first-line, which renders difficult drug recycling for use alongside PI/r in SLC. In an attempt to predict acquired HIVDR, Rutstein et al. reported a very poor sensitivity (14.7–28.0%) by using a risk-score approach [12]. Another report from South-India revealed that only half of the patients could be eligible for recycling NRTIs in the SLC [13], which somehow indicate an increasing need of implementing HIVDR testing in routine clinical practice of RLS [14]. However, Phillips et al. reported that HIVDR testing, as part of the decision to switch to SLC, was not cost-effective [15], thereby calling for the need of affordable HIVDR testing assays to the better management of ART in RLS [13‐15]. A study in 9 countries revealed varying resistance patterns at first-line failure and consistent dual-class resistance [16], which raise concerns on using HIVDR testing in deciding to switch to SLC in RLS in the frame of increasing assess to reference laboratory platforms [15, 16]. Therefore, for a switch to SLC, assessing the potential effectiveness of the current WHO recommendation in real-life would be of great benefit for ART management in RLS.

Considering the peculiarity of the national ART program in Cameroon (scale-up of ART since 2003, generalised HIV epidemic, broad viral genetic diversity) [17, 18], we sought to evaluate the patterns of DRMs according to treatment history of patients failing first-line ART in routine clinical practice and the potential susceptibility of NRTIs for selecting active SLC.

With the limited access to HIVDR testing in RLS, successful switch to SLC remains a major clinical challenge, especially for patients heavily treated on first-line ART (i.e. substitution of several NRTIs) [2, 5, 6]. Thus, implementing local strategies to ensure a successful switch to SLC is warranted [10].

With a median duration of 4 years on ART, the severe immunodeficiency (CD4 < 200 cells/mm³) and the high viral load (HIV-RNA > 10.000 copies/ml), there is a late detection of treatment failure and a substantial accumulation of DRMs in about nine out of ten patients in routine care [12‐14, 17]. This observation therefore urges the need for early viral load monitoring for timely detection of ART failure and adequate switch to SLC with limited risk of HIVDR emergence [30‐32]. Our findings are similar to several reports in Cameroon [31, 32], but with higher HIVDR prevalence compared to a study conducted at 36-months ART [33]. This is due to differences in study design (virologically suppressed and unsuppressed patients) and durations [33].

Most importantly, with only ~ 2% PI/r resistance, the use of PI/r as back bone for SLC remains standard for patients failing first-line regimens in settings with similar ART programs [2, 4, 9, 11], pending the selection of potentially active NRTIs [10‐14, 16].

In group-A (both AZT + D4T-exposure), level of HIVDR to AZT was almost two times higher as compared to TDF. This could be explained by the fact that these patients were previously exposed to D4T-containing regimens (i.e. Triomune) and were subsequently moved to AZT, most likely due to D4T-adverse events or the phased-out of D4T [34]. In the frame of treatment failure, the accumulation of TAMs would further jeopardise the efficacy of TDF due to cross-resistance mainly driven by TAMs-1 [34, 35]. Therefore, among patients exposed to both thymidine analogues, TDF still stands as the preferable option despite risks of TAMs-induced cross-resistance (~ 30%). Thus, in routine clinical practice, patients failing ART with such treatment history should either: (a) be referred for HIVDR testing or (b) be switched blindly to SLC with TDF under close viral load monitoring to detect those at risk of failure due to TDF cross-resistance [30, 34].

In group-B (exposure to TDF + a thymidine analogue D4T “or” AZT), about one-third of patients harboured HIVDR to AZT, also similar for TDF (p = 1.0000). This is due to the selection of DRMs (TAMs and/or K65R) following previous ART with a thymidine analogue- and subsequently with TDF-containing regimens [35]. Thus, because one-third of patients managed on first-line with such NRTIs substitution has lost both TDF- and AZT-efficacy [30‐34], HIVDR testing would be of great clinical relevance for selecting SLC NRTIs [14, 36].

In group-C (exposed only to a single regimen D4T-3TC-NVP), level of HIVDR was significantly higher to AZT as compared to TDF (about 4-folds), with minimal effects of TAMs inducing cross-resistance to TDF [20]. Thus, in case of exposure to a single first-line regimen, the use of TDF in SLC would have a high predictive efficacy in the majority of patients (≥80%). Thus, for such patients living in RLS, using TDF in SLC without referring to HIVDR testing might be acceptable in clinical practice [15].

As expected, K65R was only found from the group of TDF-exposed patients (group-B). Of note, K65R is a mutation known to reverse the excision phenotype of AZT resistance mutations, to increase viral susceptibility to AZT, which in turns improves the clinical efficacy of AZT [10, 37]. Thus, a wider use of TDF in current first-line regimen might improve the efficacy of AZT for subsequent use in SLC. However, the poor rate of AZT-efficacy in this group is attributed to TAMs derived from previous exposure to thymidine analogues (AZT or D4T). This implies that without HIVDR testing, AZT could be used in SLC solely for those failing on TDF-containing regimens without any previous substitution of NRTIs [8‐11].

The very high frequency of M184 V (> 80%) underscores the utility of this mutation as an indicator of therapeutic compliance in clinical practice for patients receiving any regimen containing 3TC or FTC [9]. Though FTC has an enhanced incorporation efficiency (~ 10-fold) compared to 3TC during HIV-1 RT-catalyzed RNA-dependent DNA synthesis [38], these two cytidine analogues appear as suitable alternatives without the need for programme-wide substitution of FTC for 3TC in current clinical practice [39, 40].

Phylogeny confirms CRF02_AG as the major circulating clade [41‐45]. The single case of subtype C (uncommon in Cameroon) may be due to phylodynamics from Southern/Eastern Africa [46, 47]. Even though the analysis of CRF02_AG versus non-AG showed no major effect of the local subtype distribution on emerging DRMs, molecular epidemiology surveillance merits further investigations, which include the preferential pattern of L74I as a potential signature in CRF02_AG-infected patients [48, 49]. Findings with clinical responses after switch to SLC would provide greater insights for translational application [50].

In a nutshell, first-line ART-failure exhibits variable levels of NRTI-resistance (from 17 to 62.5%) in a routine clinical setting of Cameroon, with a remarkable higher level of AZT-resistance as compared to TDF-resistance (recently introduced drug). Thus, regarding efficacy, first-line ART-failure without NRTI-substitution (i.e. exposed to only AZT or TDF) on first-line could switch to SLC following the WHO-approach. However, failure after substitution only between thymidine analogues may receive TDF as part of SLC pending a close viral load or HIVDR-testing whenever possible. Most importantly, failure after multiple NRTI-substitutions on first-line (exposed to both “thymidine-analogue and TDF”) should be referred for HIVDR-testing for selecting active NRTIs for an optimal SLC in clinical practice.