|Home | About | Journals | Submit | Contact Us | Français|
HIV-1 nonsubtype B variants account for the majority of HIV infections worldwide. Drug resistance in individuals who have never undergone antiretroviral therapy can lead to early failure and limited treatment options and, therefore, is an important concern. Evaluation of reported transmitted drug resistance (TDR) is challenging owing to varying definitions and study designs, and is further complicated by HIV-1 subtype diversity. In this article, we discuss the importance of various mutation lists for TDR definition, summarize TDR in nonsubtype B HIV-1 and highlight TDR reporting and interpreting challenges in the context of HIV-1 diversity. When examined carefully, TDR in HIV-1 non-B protease and reverse transcriptase is still relatively low in most regions. Whether it will increase with time and therapy access, as observed in subtype-B-predominant regions, remains to be determined.
The use of antiretroviral therapy has resulted in the emergence of HIV drug resistance [1–7]. In resource-rich settings, such as North America, Europe and Australia, significant resistance has been observed to the three main antiretroviral drug classes, nucleoside reverse-transcriptase (RT) inhibitors (NRTIs), non-NRTIs (NNRTIs) and protease inhibitors (PIs) [8–14]. Primary, or transmitted drug resistance (TDR), is defined as resistance to one or more antiretroviral drugs found in individuals with no previous drug exposure and is attributed to the direct transmission of resistant strains from treated individuals. As access to HAART is rolled out globally [15,16], the monitoring of TDR is essential, especially in areas where availability of second-line drugs is limited and the selection of first-line drug regimens is vital for effective treatment [17–19].
Evaluation of TDR is important in order to assess the efficacy of drug regimens, to determine optimal treatment regimens for a particular region and to measure the impact of risk-modifying interventions on HIV transmission. Current guidelines in resource-rich settings suggest resistance testing at initiation of care, regardless of whether HAART is planned or whether the infection occurred recently [9,20–22]. In resource-limited settings, the WHO recommends surveillance of TDR as a comprehensive approach to HIV care [23,24].
Several factors contribute to the reporting and the occurrence of TDR in a given population. Study methodology, such as the specific drug-resistance mutation list used to interpret resistance data, can greatly affect the reported prevalence of transmitted resistance and, until recently, there has been no consensus on which resistance mutations should be classified as markers of TDR. Other factors that can affect TDR occurrence include the number of patients receiving antiretroviral treatment in the population, their risky behaviors (e.g., unprotected sex and intravenous drug use), nonadherence to treatment, current or previous HAART regimens efficacy, available treatment-monitoring strategies, rates of virologic suppression and viral fitness of drug-resistant variants . In areas where HAART has been available since its development, such as North America and Europe, TDR has been reported to be as high as 25% [12,13,26,27].
The detection of TDR is dependent on various factors, such as duration of infection (mutations may disappear with time ) and sensitivity of testing methods (current common assays only detect mutations present in >15–30% of the viral quasispecies). During the course of HIV infection, certain viral populations may predominate according to widespread sequencing methods, but resistant minor variants often exist [29–36] and lead to earlier treatment failure [3,32,37,38]. The identification of TDR is further complicated by the diversity of HIV across different types (HIV-1 and -2), groups (main [M], outlier [O] and new [N]), subtypes and recombinant forms [7,39–42] that are prevalent worldwide . Group M accounts for over 99% of globally reported HIV/AIDS infections and is further classified into nine pure subtypes (A, B, C, D, F, G, H, J and K) and many circulating and other recombinant forms . Nonsubtype B infections account for approximately 90% of HIV infection worldwide, but are the least studied owing to financial and infrastructure limiations in geographical regions where nonsubtype B predominates. Subtype B is predominant in resource-rich settings where access to HAART has been widely available for over 10 years. In this article, we summarize reported data on PI, NRTI and NNRTI TDR in nonsubtype B HIV-1 and discuss its relevance and challenges.
Major challenges in the evaluation of TDR include the extensive diversity of HIV protease and RT genes , the large number of HIV-associated drug-resistance mutations  and the lack of a standard definition of which mutations are linked with antiretroviral treatment failure. Certain mutations known to be associated with drug resistance in subtype B are naturally occurring mutations (polymorphisms) in other subtypes and, therefore, should be excluded as markers of TDR . Although some of these mutations may indeed be transmitted, including polymorphisms as TDR mutations will lead to artificially inflated rates of TDR.
Various lists of drug-resistance mutations are used in different interpretation algorithms (Table 1), such as those from the International AIDS Society-USA (IAS-USA ), The French National Agency for AIDS Research (ANRS ), The Stanford University HIV Drug Resistance Database  and The Rega Institute . These are comprehensive lists of amino acid mutations that are associated with drug resistance, derived mostly from research in subtype-B HIV-1-infected individuals. These lists include mutations that were identified by in vitro experiments, drug susceptibility testing, clinical experience in patients failing therapy and genotypic studies. Inclusively, these four lists contain 100 mutations in 37 RT positions and 118 mutations in 46 protease positions that are associated with drug resistance. Although the lists seem similar, there are only 38 positions (12 NRTI, 11 NNRTI and 15 PI related; bold in Table 1) in which mutations appear in all four lists, emphasizing the importance of putting reported data in the context of which list is used, as well as potential interpretation discordances [46–49]. These methods may overestimate TDR prevalence because they are a compilation of all known drug-resistant mutations, including many that occur as natural polymorphisms in nonsubtype B viruses.
In an effort to increase the accuracy of identifying TDR mutations, Bennett et al. published the original and recently updated surveillance drug-resistance mutation (SDRM) list [50,51], compiled as a more specific method to distinguish between polymorphic variation and ‘real’ TDR mutations (Table 1). There are several noticeable differences between the SDRM list and the ‘regular’ lists. For example, M36I (methionine converted to isoleucine at protease position 36) is an accessory resistance mutation that was reported to increase viral fitness and confer resistance to ritonavir and nelfinavir [44,52]. This position is extremely variable in nonsubtype B viruses [40,45,53,54] where isoleucine is the wild-type consensus amino acid at this position in subtypes A, C, D, F and G and circulating recombinant form (CRF)01_AE. This mutation, which is present on the IAS-USA and ANRS mutation lists, is excluded from the SDRM list, as are a total of 79 protease and 47 RT mutations that are absent from the SDRM list but present on other lists. These exclusions allow for a more accurate, not inflated, estimation of TDR. Although a recent report from the UK suggests this may be negligible , in smaller populations, such as those that are recommended by the WHO resistance-surveillance protocols, as well as in areas where non-B subtypes predominate, the effect could be magnified and may lead to inaccurate overestimation of TDR.
It is difficult to adequately estimate the prevalence of TDR in non-B HIV-1 subtypes from the published literature, mainly owing to variation in study design and lack of standard definitions and classifications. Despite these limitations, most studies do report mutations in untreated individuals who are infected with non-B subtypes, which appear on resistance mutation lists and confer high-level resistance to one or more antiretroviral drugs. In resource-rich settings, where HIV-1 subtype B predominates, TDR rates as high as 16–25% in the USA [27,56,57] and 9–14% in western Europe [12,13,58–60] are measured. This is likely due to an early and gradual introduction of antiretroviral therapy in these populations.
Studies from resource-limited settings, where HIV-1 non-B subtypes predominate, usually report TDR from small populations in limited epidemiological samples. We aggregated data and reviewed 130 studies in the published literature that examined drug resistance in sequences from 9984 individuals infected with nonsubtype B HIV-1 (subtypes A: 1755; C: 2423; D: 447; F: 145; G: 420; H/J/K: 65; CRF01_AE: 1238; CRF02_AG: 1243; other recombinant forms: 1813; and unknown nonsubtype B strains that were not classified: 435). We reviewed studies that focus on, or include, more than five nonsubtype B infected, untreated, individuals that possessed resistance mutations; studies in which it was unclear whether HIV-infected patients received antiretroviral therapy were excluded. Studies that did not specifically comment on resistance mutations, but whose sequences are available in the Stanford University HIV Drug Resistance Database were also included. We compiled drug-resistance mutations as reported in the studies and included major mutations according to the IAS-USA list, which was used by many studies, and when possible, added TDR mutations based on the SDRM list. Only four (RT A62V, V75I, V108I and PR L33F) of 52 major resistance mutations according to the IAS-USA list are not considered markers of TDR per the SDRM list.
Transmitted drug resistance in nonsubtype B HIV-1 was reported from 63 countries across all continents (Africa: 25 countries; Asia: 19; Europe: 15; South America: two; Central America: one; and North America: one; Figure 1). The map in Figure 1 represents the majority of subtypes and recombinants from different geographic regions that were reported in these studies and mostly agrees with previous reports [41,43]. Notable findings include:
Country-specific subtype distribution is difficult to quantify secondary to sparse and varying epidemiological data, and diverse and, at times, unclear subtype identification methodologies.
Drug resistance in individuals infected with non-B HIV-1 subtypes was reported in 80 of the 130 studies (Table 2). In Africa, TDR was variable and 27 of 59 studies reported no resistance. Studies from South Africa reported increasing TDR with time, with almost no resistance in 2000–2001 [69–71] and high levels (21%) of resistance to NNRTIs in 2008, including major mutations K103N and V106M . This high TDR rate was reported in a small cohort (n = 14) and may not accurately reflect actual TDR prevalence. A study from Burundi reported drug resistance of 94% to PI's in treatment-naive individuals . However, this high rate is based on the prevalence of the protease M36I polymorphism, which should not be on the TDR list, as discussed previously. Additional studies that included polymorphisms as resistance mutations reported high TDR rates. The most commonly observed protease positions with high polymorphism rates were 10, 20, 36, 63, 69, 77 and 93, and rates as high as 100% were reported. Overall, TDR reported from multiple African regions was higher for RT inhibitors, probably reflecting the earlier introduction of these drugs and their inclusion in first-line HAART regimens.
The prevalence of TDR also varies in Asia [74,75]. The rate of reported TDR mutations in nonsubtype B viruses in China is low, generally less than 6% [76–79], and similar to Singapore , Cambodia , Vietnam  and Malaysia . In Thailand, reported resistance ranged from 0 [84,85] up to 24% [74,75]. This high rate was due to RT mutation M184I and protease mutations D30N, M46I, G73S and I84V, reported yet again from a limited sample size (n = 25). Although subtype B is predominant in Taiwan, the rates of resistance among non-B subtypes was reported as 8% in samples collected from 1999 to 2006 . In India, where subtype C is the dominant subtype, reported resistance is less than 5% [87–95]. In Russia and the former Soviet Union, subtype A and CRF01_AE predominate. Reported TDR ranges from 0% in Moldova  and 3% in the Republic of Georgia  to 7% in Latvia  and 8% in a multinational study of this region . The rate of 8% included A62V, which is not considered a marker of TDR according to SDRM.
Although much of central and South America is populated by HIV-1 subtype B, reports of non-B subtypes are increasing and several studies assessed the prevalence of TDR in nonsubtype B infections. Reports range from 0 to 1% in Brazil [100–102] and Cuba , to 3% in Argentina [104,105] and as high as 9% in Brazil .
In Europe, nonsubtype B viruses make up a substantial proportion of HIV-1 infections (up to 100% subtype F1 in parts of Romania ) and is increasingly prevalent in other areas of the continent [12–14,108–111]. From 1996 to 2002 in several countries in Europe, the overall prevalence of TDR among 673 nonsubtype B individuals was 5% . In the USA, although the reported population of nonsubtype B is much smaller than Europe, [65,66], TDR has ranged from 0% in 1999 in Boston (MA, USA)  to more than 20% in 2000–2004 in New York City (NY, USA) . Populations from these regions, with long-term antiretroviral access, have higher rates of TDR. Furthermore, rates of resistance to NRTIs and NNRTIs are higher than those for PIs, coinciding with the greater use of these drugs and their universal earlier introduction.
Transmitted drug-resistance mutations were reported at 37 different positions in the RT and protease genes in non-B HIV-1 subtypes and recombinants across the three classes of antiretroviral drugs (12 NRTI, ten NNRTI and 15 PI positions associated with resistance; Table 3). The most commonly reported NRTI TDR mutations were M184V/I. Additional common mutations include M41L, K219E/N/Q/R and T215F/Y/C/D. M184V, observed across numerous populations and subtypes, confers high-level resistance to lamivudine and emtricitabine. The additional reported mutations are termed thymidine analog mutations and can confer resistance to all NRTIs. The most commonly reported NNRTI TDR mutations were K103N, Y181C/I and Y188C/H/L, all conferring resistance across the NNRTI class. RT mutations A62V (commonly reported in subtype A from central Asia) and V108I (commonly reported from Africa) are not on the SDRM list, but are on the IAS-USA list.
The most commonly reported PI TDR mutations were M46I/L. Mutations at this position were not included in the 2007 SDRM list  but are part of the 2009 SDRM list (Table 1) . Other TDR PI mutations were reported much less frequently, as expected by less drug exposure owing to the inclusion of mostly RT inhibitors in first-line regimens in resource-limited settings. L33F, reported from Africa and Asia, appears on the IAS-USA list and not on the SDRM list, similar to RT mutations A62V and V108I mentioned earlier. Although major PI resistance mutations were less common than mutations in RT, prevalence of accessory protease mutations was abundant, as discussed previously [54,80,81,89,95,105,107,110,112–129]. Compared with more accepted TDR mutations, these polymorphisms appear to have minimal effect on treatment outcomes [8,128–134]; however, some HAART regimens may be less efficacious [118,135]. Although the genetic barrier (or the number of genetic changes necessary to create drug-resistance mutations) appears to be similar for different subtypes despite baseline genotypic difference , the significance of these pretherapy polymorphisms in the evolution and transmission of drug resistance is still unclear .
As demonstrated in Table 3, the number of TDR mutations reported in the studies reviewed here is small. Overall, in 9984 non-B sequences, there were 529 (5.3%) drug-resistance mutations. Upon exclusion of mutations that appear on the IAS-USA list as major mutations but not on the SDRM list (RT A62V and V108I; protease L33F), that number declines to 302 mutations (3%). When specific subtypes are examined, other than subtype F and CRF01_AE, all TDR rates are below 5%. Whether the 6.9% TDR in subtype F and 6.4% TDR in CRF01_AE (5.7% without non-SDRM mutations) are actual differences will need to be examined in larger data sets. Regarding specific TDR mutations, despite the small numbers, some observations can be made, such as the lack of M41L in 1755 subtype A sequences; the lack of K103N in subtypes D, F and G sequences; and isoleucine as the only mutation at RT position 184 in CRF01_AE. Whether these observations are significant remains to be determined. Submission of sequence data to electronic databases and clear presentations of study methodologies will allow further careful examinations of TDR in multiple HIV-1 subtypes.
Infections with nonsubtype B HIV-1 predominate globally and are on the rise in geographic areas where subtype B prevails, such as North America and Europe. The epidemiology of HIV drug resistance in resource-rich settings is complex and with increased universal access to antiretroviral therapy, we are only now beginning to examine it in resource-limited settings. TDR is important to determine as it results in longer time to viral suppression and shorter time to virological failure. High rates of TDR have been reproted from regions in which antiretroviral therapy has been available for a long time and, therefore, it is likely that it will increase in areas where treatment access is being scaled-up.
Attempts to standardize global TDR surveillance strategies are ongoing. This should include quantifying TDR and identifying mutations that accurately represent transmission of drug resistance and not merely part of HIV diversity. Methodologies to determine TDR in nonsubtype B HIV-1 vary widely across different settings, and should be carefully interpreted. Additional limitations in the ability to aggregate and interpret reported data on TDR in non-B subtypes include resistance transmission not being the main study focus, variable subtyping methods and paucity of data.
An interesting and challenging question is how long after the introduction of antiretroviral therapy should TDR be expected to develop and at what pace. Mathematical models were shown to be fairly accurate in predicting rates of TDR in certain populations in North America [137,138], and estimations regarding other global settings are ongoing [139,140]. Although prolonged access to therapy and the use of less potent regimens appear to be linked to high TDR prevalence, this observation is not always consistent. Important effects seem to include the proportion of antiretroviral treatment in the population examined, the development of acquired resistance in the population, adherence to therapy and viral fitness. How these will vary in different resource-limited settings with diverse HIV-1 subtypes remains to be determined.
Rami Kantor is funded by an NIH RO1 grant AI66922.
Financial & competing interests disclosure
The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Papers of special note have been highlighted as:
of considerable interest