Antiretroviral drug resistance among antiretroviral-naïve and treatment experienced patients infected with HIV in Iran

INTRODUCTION

HIV antiretroviral resistance is a global problem, and the World Health Organization recommends surveillance for the emergence of HIV resistance as a component of all antiretroviral treatment programs [DiazGranados et al., 2010]. Individuals infected with HIV acquire antiretroviral resistance in one of two ways: (1) transmitted drug resistance (TDR) occurs when antiretroviral-naïve individuals are infected with viruses already harboring resistance mutations, and (2) acquired drug resistance (ADR) emerges from the selective pressure of antiretrovirals in individuals without fully suppressed viral loads either because of poor adherence or the use of a non-suppressive regimen [Toor et al., 2011]. Infected patients with either TDR or ADR are at increased risk for treatment failure. In addition, when a failing regimen is not identified and changed, further treatment options may be limited both because of cross-resistance to other antiretroviral compounds in the same drug class and because of the further accumulation of additional mutations in other drug classes [Graf et al., 2011; Hightow-Weidman et al., 2011; Raimundo et al., 2012]. Patients with TDR or ADR may also transmit their resistant virus to others.

The HIV epidemic in Iran remains most concentrated among injection drug users with an estimated prevalence of 15%. There has also been an increase in sexual transmissions, especially among sex partners of persons using injection drugs and female sex workers [National AIDS Committee Secretariat, 2012]. The Iranian government provides CD4+ T cell count monitoring to patients with HIV who are in care, and free antiretroviral medications are available for individuals with a CD4+ T cell count of less than 250. In Iran, the most commonly prescribed first line regimen is zidovudine (AZT), lamivudine (3TC), and efavirenz (EFV) [Jahanbakhsh et al., 2013]. By the end of 2012, 8% of the nearly 25,000 patients infected with HIV in Iran were on antiretroviral medications, and this improved access to care has resulted in significant declines in AIDS-related mortality [National AIDS Committee Secretariat, 2012]. However, antiviral drug resistance remains a major threat to the sustained positive health impact of HIV treatment programs in Iran.

The aims of the current analyses were to characterize the frequency and extent of drug resistance among a sample of hospitalized patients infected with HIV in Iran and to compare the prevalence of resistance mutations in a sample of antiretroviral naïve and treatment-experienced patients infected with HIV. In these analyses, TDR was assessed with antiretroviral naïve individuals infected with HIV and ADR was estimated with individuals receiving antiretroviral therapy. It can be hypothesized that because of their different genetic barriers to resistance, fewer mutations would be observed among participants taking protease inhibitors (PI) as compared to non-nucleoside analogue reverse transcriptase inhibitors (NNRTI).

MATERIALS AND METHODS

In this cross sectional study, participants were eligible for inclusion if they were admitted to the Infectious Disease Division of Imam Khomeini Hospital (Tehran, Iran) for an HIV-related illness, were older than 18 years of age, and were (a) antiretroviral naïve, or (b) had received antiretroviral therapy that consisted of one NNRTI or one PI in combination with two nucleoside reverse transcriptase inhibitors (NRTI) for at least 1 year. The study protocol was approved by the Ethical Committee of Iranian Research Center for HIV/AIDS, and written informed consent was obtained from all participants prior to blood collection.

Baseline demographic and adherence data were collected through physician interview. Blood samples were collected in sterile EDTA-containing tubes, and plasma was separated and stored at −80°C for genotyping. To screen these specimens for antiretroviral drug resistance mutations, total RNA was extracted from 140 µl of plasma by using the QIAamp® Viral RNA kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. RT-Nested PCR amplification was performed to yield a 514-base-pair (bp) product from the viral protease region and a 742-bp product from the reverse transcriptase region using primers designed by oligo7 software from HIV Refseq AC: NC-001802.1 (reverse transcriptase primers for round 1 and 2: (F RT1) ACVCCTGTCAACATAATTGG, (R RT1) CTAYTAANTCTTTTGHTGGGTC, round 2: (F RT2) ATTAAAGCCAGGAATGGATGG, (R RT2) TTYTGTATRTCATTGACAGTCCAG and protease primers for round 1and 2: (F Pr1) AGG CTA ATT TTT TAG GGAAGA, (R Pr1) GCA AAT ACTGGAGTATTG TATG, round 2: (F Pr2) TCAGAGCAGACCAGAGCCAACAGC, (R Pr2) AAT GCT TTT ATT TTT TCTTCTGTC).

The PCR products were purified by Gel Purification kit (Bioneer, Global Genomics Partner, Daejeon, Republic of Korea) according to manufacturer's instructions and sequenced in an automated DNA sequencer (ABI PRISM 3730 version 3.0, Applied Biosystems, Foster City, CA).

The reverse transcriptase and protease sequences were analyzed by BioEdit software (version 5.0.6), and the Stanford University HIV Drug Resistance Database was used for drug resistance interpretation [Rhee et al., 2003; Campbell et al., 2005; Gallant et al., 2006]. This database categorized resistance as either susceptible or as low, intermediate, or high-level resistance using a mutation penalty score based on published drug resistance and treatment outcome studies as well as in-vitro susceptibility data. Susceptible virus is comparable to wildtype in regards to anticipated drug response, with low, intermediate, and high-level drug resistance reflecting a continuum from reduced in-vitro susceptibility and suboptimal drug response (low-level resistance) to no drug response (high-level resistance). For these analyses, participants with potentially low-level resistance (mutations that may indicate prior drug exposure but no reduced susceptibility) were classified as low-level resistance.

P-values of <0.05 were used to determine statistical significance using the Chi-square or Fisher's exact tests. Statistical analyses were performed using STATA, version 12.

RESULTS

One hundred participants were enrolled in this study; 30 were ART-naïve and 70 were treated for at least 1 year. The majority (81.0%) were male, with a mean age of 37 years. Most (66.0%) reported a history of injection drug use. Self-reported adherence to ART was high (70%). All of the 70 patients receiving ART were on regimens that included two NRTIs and either EFV or lopinavir/ritonavir (LPV/r); forty-six were taking 3TC, AZT, and EFV; eight were taking 3TC, tenofovir (TDF), and EFV; and 16 were taking 3TC, AZT, and LPV/r. All participants taking EFV were on their first ART regimen, while those taking LPV/r had all been prescribed an NNRTI-based regimen previously. Sequencing analysis was successful for all participants receiving antiretroviral therapy with detectable viral loads (62 of 70 participants) and for all 30 ART-naïve participants. The inability to amplify eight of the samples was due either to low viral load or to degradation of viral RNA during sample transport.

Among the 62 sequenced participants receiving ART, 36 (58.1%) had at least one drug resistance mutation; seven (11.3%) had resistance mutations to one drug class, 27 (43.5%) had resistance to two drug classes, and two (3.2%) had resistance mutations to all three drug classes evaluated. Results from the genotypic analysis are presented in Table I. The most frequent clinically relevant NRTI mutations among the treatment-experienced participants were M184VI (19.4%) and the thymidine analog mutations (TAMs) (M41L (8.1%), D67N (11.3%), K70RT (8.1%), L210W (8.1%), T215YNS (16.1%), and K219E (8.1%)). Although no participants were on didanosine (ddI) or stavudine (d4T) at the time of study enrollment, 11.3% of participants had V75MC mutations. The two (3.2%) participants with T69S also had TAM mutations. No K65R mutations were observed. Accumulation of NRTI mutations was common, with 11 (17.7%) participants having two NRTI mutations and 14 (22.5%) having three or more mutations. The most frequent NNRTI mutations were K103N (21.0%), G190A (9.7%), and K101EQ (9.7%). The most common PI mutations were V82AI (6.5%), L90M (4.8%), and I54V (4.8%). All participants with PI mutations had accumulated three or more PI-mutations.

In contrast to those receiving ART, no major NRTI or PI drug resistance mutations were seen among treatment-naïve individuals on genotypic analysis. Two ART-naïve participants (6.7%) had evidence of transmitted NNRTI mutations, one with K103N and one with both K103N and P225H.

Table II presents phenotypic data based on imputation of participant's drug resistance mutations into the Stanford University HIV Drug Resistance Database. Among participants receiving antiretroviral therapy, high level resistance to 3TC and emtricitabine (FTC) was common for both EFV and LPV/r-based regimens (53.2% vs. 46.7%, χ² = 0.19, P = 0.77), consistent with the frequent occurrence of M184V among treatment-experienced study participants. High-level resistance to EFV was more common among participants on EFV-based regimens as compared to high-level LPV/r resistance among participants taking LPV/r (55.3% vs. 6.7%, χ² = 10.05, P < 0.0001). Although etravirine use was not reported, accumulation of NNRTI mutations among participants on EFV resulted in frequent high-level (21.3%) or low-level (40.4%) cross-resistance to ETR. The four participants with PI resistance were on LPV/r. Cross resistance to other PI was common; all four participants had either high or intermediate level resistance to the other protease inhibitors commonly available in Iran, atazanavir (ATV), and indinavir (IDV). Both of the treatment naïve participants with NNRTI mutations had high-level resistance to EFV and NVP, and one (3.3%) of these participants also had low-level resistance to etravirine (ETR).

DISCUSSION

This study of hospitalized inpatients in Tehran, Iran documented an alarmingly high prevalence of HIV drug resistance and confirmed the presence of high-level transmitted drug resistance. Consistent with prior studies in Iran [Hamkar et al., 2010; Baesi et al., 2012], more than half of the participants receiving antiretroviral therapy with detectable virus had evidence of drug resistance mutations, many with resistance to multiple drug classes or with mutations that confer resistance to other drugs in the same class.

Currently, most patients in Iran are started empirically on antiretroviral regimens without first obtaining HIV genotype data. The presence of TDR observed in this study highlights the importance of baseline genotype testing prior to initiating a first regimen both to improve clinical outcomes and to limit the emergence of further drug resistance. Transmitted mutations were documented in two (6.7%) participants, a proportion similar to prior studies in Iran that have estimated TDR in 5–15% of new infections [Mousavi et al., 2010; Jahanbakhsh et al., 2013]. In higher income European countries, the United States, Japan, and Australia, TDR is estimated to be between 10% and 17%; TDR is less common in lower and middle income countries, with only 28% of more than 70 surveys conducted between 2004 and 2010 documenting a prevalence of TDR greater than 5% [WHO, 2012]. Both of the participants in this study with TDR had K103N, a mutation that confers high-level resistance to the two available NNRTIs in Iran, nevirapine (NVP) and EFV [Shafer et al., 2000; Marconi et al., 2008]. Notably, the most common first line regimen in Iran, a combination of AZT, 3TC, and EFV, is particular susceptible to treatment failure when K103N is present. If NNRTI or NRTI drug resistance mutations are detected before treatment initiation, drugs with a higher genetic barrier to resistance such as ritonavir-boosted PIs should be included in the first treatment regimen to minimize the risk of ADR and subsequent virologic failure [Riddler et al., 2008; Wittkop et al., 2011].

These data also document the need for systematic viral load monitoring and drug resistance testing for hospitalized patients infected with HIV in Iran in order to reinforce treatment adherence, to adjust failing regimens, and to minimize the further emergence and transmission of resistance. While not directly assessed in this study, the extensive resistance seen among treatment-experienced participants likely developed during the continued use of a non-suppressive regimen or because of poor antiretroviral adherence. Unfortunately, antiretroviral treatment options in Iran remain limited, with only 3TC, AZT, didanosine, stavudine, TDF, abacavir, NVP, EFV, LPV/r, ATV, IDV, and ritonavir available. Two participants (3.2%) on ART had three-class resistance, and for these individuals, access to additional drug classes is needed to achieve full virologic suppression.

Several methodological limitations should be considered when interpreting these results. First, all study participants were hospitalized at a single hospital in Tehran, and these participants may not be generalizable to other non-hospitalized individuals infected with HIV. Hospitalized individuals are more likely to have detectable viral loads in the setting of acute illness, and those with drug resistance mutations and detectable viral loads are also more likely to have opportunistic infections and to be hospitalized. In fact, ART resistance may have caused the patients to be hospitalized in the first place. Consequently, this sample may overestimate the overall proportion of individuals on ART in Iran with detectable viral loads. Second, social desirability bias likely resulted in an overestimate of true ART adherence. Third, the small sample size limits the precision of estimates and power for detecting differences between different subgroups. Fourth, the true levels of TDR and ADR were likely underestimated in these analyses, as the genotype testing employed in these analyses was not able to detect minority drug resistant quasispecies. In addition, because baseline genotype data were not available for participants on ART, this cross-sectional study design does not enable a direct assessment of ADR. While some resistance mutations documented among ART-experienced participants may have been transmitted mutations, the large differences observed between treatment-experienced and treatment-naïve participants are unlikely to result from misclassification. Finally, detailed information about treatment-experienced individual's prior treatment regimens, including duration of therapy and time since diagnosis, was not available. However, as participants on PI-based regimens had all failed a prior NNRTI-based regimen, the observation that fewer resistance mutations were observed among those on PI-based regimens as compared to NNRTI-based regimens highlights the potential value in PI-based second line therapy.

The improved access to ART in Iran since 2003 has decreased HIV related morbidity and mortality. However, these successes may be threatened by the widespread emergence of both acquired and transmitted HIV drug resistance. In Iran, the continued success of the HIV treatment programs requires tailoring antiretroviral regiments based on the results of routine viral load and genotype testing, including at diagnosis, as well as interventions to improve treatment adherence and to expand access to potent, new antiretroviral medication from additional drug classes.