Prevalence of Antiretroviral Drug Resistance Mutations in HIV Seropositive Patients from an Outpatient Clinic of a Large University Hospital from São Paulo,Brazil

Introduction

Despite continuous improvements in HIV patient treatment, drug resistance mutations (DRMs) remain an obstacle to both initial and second and subsequent treatment regimens. Moreover, the crescent use of antiretroviral (ARV) as prevention in both post-exposure Prophylaxis (PEP) and pre-exposure prophylaxis (PrEP) may have its success compromised by circulating viral variants harboring transmitted resistance. ^1,2

Many studies in different parts of the world have documented the frequency and risk factors for DRMs, and WHO/UNAIDS recently develop the early warning indicators, which identify risk factors for virologic failure and drug resistance. ^3,4 In general, those studies have shown increased or at least steady rates of resistance to ARV drugs; therefore, monitoring programs are important to detect increases in DRMs that may affect ARV programs. ^{2,5 –8}

Combating antimicrobial resistance, including the threat posed by drug-resistant HIV, is a major goal for the global community. Prevention, monitoring, and timely response to population levels of HIV drug resistance is critical to achieving the WHO/UNAIDS 90–90–90 targets to control the HIV epidemic. ⁴ To accomplish that, WHO instructed countries to fight DRMs by monitoring the quality of their health care and emphasizing the adequacy of measures adopted whenever any failing in treatment is detected. ⁹

The presence of DRM is an important concern both when initiating an ARV therapy and to plan modification of therapy in patients with virologic failure observed during ARV regimen. ¹⁰ Inadequate adherence and virologic failure are frequently associated with the occurrence of these mutations, ^{11 –13} but the presence of some mutations, as to the nucleoside reverse transcriptase inhibitor (NRTI) class at first-line failure, may actually be used as a proxy of success in second-line therapy, ^13,14 either as a marker of adherence of due to some, yet uncharacterized, loss of viral fitness. The major concern is the fact that resistance to HIV drugs can compromise not only therapy but also current and future prophylactic ARV prevention strategies. ^15,16

Despite the potential harmful effects of resistance emergence and evolution, limitations observed especially in low middle income countries (LMICs) lead to long delays in regime switching after a virologic failure, contributing to drug resistance emergence. ¹⁶ Moreover, social and behavioral factors contribute to high rates of resistance in children even in communities with mature treatment programs.

Not uncommonly have poor documentation about HIV infections and treatment failures in more vulnerable communities. Many current treatment programs are often insufficient to stimulate the correct use of ARV drugs, with some settings with additional logistic problems resulting in interruption of drug supplies. ^15,17 Among the measures able to minimize those tendencies are the adoption of treatments with higher barriers to resistance, but newer drug generations and new ARV classes are either in development or currently too expensive to LMICs. Technologies capable to identify early treatment failures are an important public health priority.

Our study evaluated the prevalence of (drug resistance-associated) mutations, among treatment-naive patients and ARV experienced patients from our outpatient clinic at a large university hospital from the city of Sao Paulo, Brazil, from 2002 to 2017.

Materials and Methods

We made a retrospective analysis of transmitted drug resistance (TDR) and acquired mutations, using the results of 358 HIV-1 genotyping tests for patients from the outpatient clinic at the Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (HCFMUSP).

The HIV-1 genotyping tests were performed in the Laboratory of Clinical Investigation 56 (LIM 56), HCFMUSP, or in the Adolfo Lutz Institute (IAL), according to the Brazilian Network for HIV-1 genotyping protocol (RENAGENO), organized by the Brazilian Ministry of Health. ¹⁸ Blood was collected from 2002 to 2017 for the genotyping tests. RNA was extracted and reverse transcription (RT)-polymerase chain reaction was followed by Sanger sequencing. Most sequences were obtained using a Trugene HIV-1 genotyping kit (Siemens HealthCare Diagnostics, Ireland) according to manufacturer's instructions. Some samples were sequenced using local protocols. ¹⁹

The mutations associated with drug resistance and with the determination of the subtypes were identified using the respective version for the year the sequence was analyzed (2002–2017) at the HIV Drug Resistance database of Stanford University (HIVDB): the HIVdb Program (http://hivdb.stanford.edu/). The program shows an updated list of mutations every year. ^17,20

Genotype results were noted at Excel (Microsoft Corporation, Seattle, WA). Statistical analysis was performed using the statistical package program (Stata versão 8). Proportion estimates are described along with binomial exact 95% confidence intervals (CIs). Differences between the groups were analyzed using chi-square Pearson or Fisher exact, with values of p < .05 considered significant. Resistance was considered as TDR if identified in a treatment-naive patient and as acquired DRMs for patients on ARV drug during virological failure.

Analysis of both transmitted and acquired mutations included mutations of the main classes of ARV drugs (protease inhibitor [PI], NRTI, and non-nucleoside reverse transcriptase inhibitor [NNRTI]) considered at Stanford database. Mutations are described by ARV class and as multidrug when mutations to more than one class were identified.

To define transmitted resistance, the sequences were reanalyzed using the list of mutations of Calibrated Population Resistance Tool of Stanford University (CPR Version 6.0, Stanford Database, SDRM 2009). The CPR is a program adopted by WHO to compare TDR rates across geographic regions by analyzing HIV-1 sequences according to a list of resistance-associated mutations that exclude polymorphisms such as those related to subtypes. ^21,22 Acquired resistance included mutations according to the Stanford HIV db Genotypic Resistance Interpretation Algorithm.

Results

Genotypes of 358 patients with at least one genotyping test were analyzed. From these, 249 (69.5%) were performed in treatment-naive patients and 109 (30.4%) were of genotypes performed during treatment failure of ARV-experienced patients. Overall, the average age of patients was 37 years, most males (74.3%), with 25.7% females. Many naive cases were referred from the HCFMUSP blood bank and all went through an inclusion interview to assess previous ARV treatment use.

Table 1 describes the prevalence of ARV resistance mutations to the major classes of drugs used in the treatment of HIV-1, as any mutation and multidrug, for treatment-naive patients and for those exposed to ARVs. For the untreated patients, the prevalence of any mutation was 21 (8.4%, 95% CI: 5.3–11.6) and that for those exposed to ARV was 75 (68.8%, 95% CI: 59–77).

Samples were compared considering as representative of two decades, a first period, from 2002 to 2009 (196 treatment-naive and 72 treatment-experienced patients), and a second period, for tests collected from 2010 to 2017 (53 treatment-naive and 37 treatment-experienced patients).

Table 2 describes the prevalence of ARV resistance mutations to major classes of ARV drugs for treatment-naive patients. Resistance to all classes of drugs increased from the first to the second period considering the “any mutation” criteria (p = .05).

Table 3 gives the prevalence of mutations according to the two periods of collection for treatment-experienced patients. Prevalence of DRMs for all classes of ARV tested decreased, with that for the PIs class being statistically significant.

The most common transmitted DRMs identified were 0.5% of V82A to PIs mutation, 1.1% of M184V to NRTI mutation, and 3.5% of K103N to NNRTI mutations. As for acquired mutations, PIs mutation L90M was observed in 11.3%, NRTI mutation M184V in 21.5% and the K103N mutation to the NNRTI class was present in 20.4% of the cases.

The HIV-1 subtype of the sequences was mostly B (78.2%), followed by F (7.3%) and BF recombinants (7.3%), with 3.6% HIV-1 C and 3.6% with unclear classification.

Discussion

A survey conducted by WHO reveals that in 12 of 18 countries reporting data to WHO between 2014 and 2018, levels of pretreatment resistance mutations to efavirenz (EFV) and/or nevirapine (NVP) among adults initiating first-line ART exceeded the 10% threshold considered as a warning signal for relevant resistance in the community. ²³

Our study shows a TDR rate of 8.4% (95% CI: 5.3–11.6) in 249 treatment-naive patients. Although rates at second decade are higher (15.1%, 95% CI: 6.7–27.6) than those at first decade (6.6%, 95% CI: 3.6–11.1), the CI is ample due to the small sample size. However, these results are within the CI range found in studies with a more recent sampling (2004–2016) in São Paulo, as the study of Coelho et al. who identified one or more TDR rates in 10.9% (95% CI: 8.6–13.6) of 596 sequences from newly diagnosed patients from 2014 to 2016. ²² A study analyzing samples from 1998 to 2002 in the same area showed a TDR rate of 6.3% (95% CI: 3.9–9.3). ²⁵

To evaluate whether we could observe any change in the rate of resistance, we compared the frequency of mutations at two periods. There is a significant increase in any TDR mutation among untreated individuals, mainly driven by the NNRTI class. In contrast, treatment-experienced patients showed a decreasing trend to any TDR mutation class, significant for mutations of the PIs class. Progress in ARV drugs potency and better drug tolerability may have contributed to this trend of decreasing acquired resistance. In the case of the PIs class, the decrease in the prevalence of resistance was associated with stronger drugs with better genetic barriers, as exemplified by Darunavir over the past 10 years, ¹⁸ and more widespread use of ritonavir boosted regimens. The most prevalent mutations for this class were L90M, M46L, I54V, V82A, and N88D.

Some NNRTI mutations have a major impact in drugs as EFV and NVP, ^22,26 the two NNRTI ARV drugs available in Brazil during the study. Although EFZ is still used in first-line regimen in Brazil, since 2018, the integrase inhibitor dolutegravir is recommended as preferential drug for first line regimen.

The subtypes observed are in accordance with other studies in the area, which all show a predominance of subtype B. ^{19,22,24 –27} We used the Stanford subtype classification and we did not notice associations with any DRM.

The resistance rate among treated patients in this study was really extremely high. However, these findings are similar to those in patients from Minas Gerais state, where the resistance to NNRTI increased from 74.4% to 81.6%, mainly due to K103N mutation. ²⁸ Also the high rates of transmitted NNRTI mutations limit the use of this class as first-line treatment in Brazil in the future, and validate the most recent Brazilian guideline to include as a first option dolutegravir.

There are many limitations we need to take into account in our study. First, the prevalence reported represents the specificity of this university service that includes some cases exposed to different ARV regimens. Newer classes were used in salvage therapy for some patients with advanced disease, and mutations to these drugs were not evaluated in this study. Moreover, this study included samples of only one source from a service in São Paulo; therefore, our results may not reflect the reality for other populations of HIV-infected persons in Brazil. Also, the results were noted manually, so errors such as mistyping cannot be ruled out, although another technician double checked the spreadsheet to reduce such errors. We did not access the actual sequence to confirm subtype assignments. The results are based in Stanford db and were not confirmed by more specific tools such as Rega and jpHMM.

Regarding DRMs, the widespread use of ARV therapy and the enhanced life expectancy of treated patients may increase the chances of selection and transmission of HIV-resistant strains when these patients fail therapy. Therefore, providing close monitoring of patients to decrease the period of viremia benefits the patient and also helps the control of resistance transmission. DRMs monitoring is an important tool for surveillance of drug-resistant variants, and genotyping test is very useful to define salvage therapy. ²⁹

DRMs that may jeopardize the success of HIV treatment must be carefully tracked, as ARV drug therapy is still the only feasible treatment option to suppress the virus. ^16,23,30 According to the 2018 report, the plan of action for the prevention and control of HIV and sexually transmitted infections 2016–2021 by WHO, surveillance of drug resistance continues to be fundamental to enable an effective drug monitoring of the currently available drugs and also of putative future drugs, and also to inform of policy changes in public health treatment.

Conclusions

The prevalence of transmitted DRMs in treatment-naive patients was 21 (8.4%) of 249 genotyping tests, with an increasing trend in two periods analyzed, whereas acquired resistance, especially to PIs class, may be decreasing, however, with an overall high rate of 69% of acquired resistance among treatment-experienced patients. Rates of transmitted DRMs showed an important increase (6.6% in 2002–2019 vs. 15.1% in 2010–2015). Thus, diagnosis and proper patient retention may allow continuous surveillance and improvement in health care, but special attention must be dedicated to patients failing therapy to guarantee the long-term success of ARV treatment programs.