Prevalence and circulation patterns of Variant, Atypical and Resistant HIV in Los Angeles County (2007–2009)

Variant Atypical and Resistant HIV Surveillance in Los Angeles County

Variant Atypical and Resistant HIV Surveillance collected genetic sequence data from the pol region (Protease and Reverse Transcriptase) of HIV-1 from newly diagnosed persons. During the course of this activity, the efforts changed from active to passive surveillance due to insufficient remnant serum and resources to support HIV genotyping. Through active surveillance, remnant diagnostic sera were obtained from 17 collaborating laboratories that account for approximately one-third of the cases reported to Los Angeles County's enhanced HIV/AIDS Reporting System (eHARS) annually. Collected sera were sequenced at the Stanford University Clinical Virology Laboratory in Palo Alto, California. Through passive surveillance, electronic HIV-1 genetic sequence data were obtained from four collaborating laboratories that perform HIV genotyping—namely, Advance Medical Analysis LLC, Laboratory Corporation of America, ARUP Laboratory and Los Angeles County Public Health Laboratory.

The study patients consisted of Los Angeles County HIV cases reported as new diagnoses during 2007–2009, who were determined to be antiretroviral naïve and for whom HIV genetic sequence data obtained from a blood specimen drawn within 90 days of HIV diagnosis was available.

Cases with an HIV-1 subtype B, E, and D Capture Enzyme Immunoassay (EIA) result, indicating recent infection, and no laboratory or medical records signifying an AIDS diagnoses within 6 months of their HIV diagnoses, were classified as “recent” infections. Cases with an EIA result indicating long-standing infection, or cases diagnosed with AIDS less than 6 months after their HIV diagnosis, were classified as long-standing.

HIV-1 Subtypes and Mutation List

CDC methods were used for subtyping HIV-1 sequences [Pieniazek, 2011]. First, sequences were analyzed using Sierra in Perl Language, a Stanford HIV Web Service developed by the Stanford Online HIV Drug Resistance Database [Liu and Shafer, 2006]. Sequences were classified as subtype B and potential non-B variants. Assignment to subtype B was based on ≥90% similarity to subtype B standard sequences in protease and reverse transcriptase (RT). Subsequently, we used phylogenetic methods to confirm subtype classification for non-B subtypes and recombinant assignments. For the phylogenetic reconstruction, sequences were aligned using the BioEdit program [Hall, 1999] with reference sequences of Group M (including CRFs) taken from the Los Alamos Laboratory HIV database [http://www.hiv.lanl.gov/]. A neighbor-joining tree using Kimura's-2 parameter method and 1,000 bootstrap replications was constructed using the Molecular Evolutionary Genetics Analysis program [Tamura et al., 2011]. The robustness of the nodes was estimated with a bootstrap value of 80% to indicate significance.

To identify drug resistance associated mutations in Los Angeles County, we adopted a mutation list developed by World Health Organization and modified by CDC [Bennett et al., 2009; Wheeler et al., 2010]. Since subtype B is predominant in the US, the CDC mutation list includes additional mutations that are nonpolymorphic in subtype B to maximize sensitivity. Some drug resistance associated mutations were excluded from consideration in the analysis of specific subtypes, as they are likely to represent naturally occurring polymorphisms in drug-naïve HIV-infected individuals [Wheeler et al., 2010].

Statistical Analysis

Chi-square and Fisher's exact tests were used to compare study patients with all reported newly diagnosed cases on demographics and other characteristics. We also compared TDR cases with non-TDR cases. Multivariate Log-binomial regression was used to estimate prevalence ratios (PR) and 95% confidence intervals (CI). The following variables were considered: age at diagnosis, sex, race/ethnicity, HIV transmission mode, and recency of HIV-1 infection. All P-values were two-sided and the significance level set at 0.05. All statistical calculations were performed using SAS version 9.2 (SAS Institute, Cary, NC).

Study Patients

Of the 7,100 newly diagnosed HIV persons reported to eHARS in Los Angeles County from 2007 through 2009, 1,414 (20%) eligible sequences were obtained. Of these sequences, 493 (35%) were collected by genotyping remnant sera and 921 (65%) were obtained directly from collaborating laboratories.

As shown in Table I, 91% of the study patients were male, 72% were below the age of 40, 48% were Hispanic/Latino and 70% were men who have sex with men. Thirty percent were classified as recent infections and approximately 44% were diagnosed at an HIV clinic. Figure 1 displays the study patients by ZIP code and delineated by Service Planning Areas (see definition in the footnotes). Over one-third of the study patients resided in the Metro Service Planning Area at the time of their HIV/AIDS diagnosis. Study patients differed from the overall population of newly diagnosed HIV cases in Los Angeles County by age at HIV diagnosis, race, gender, mode of transmission, HIV/AIDS diagnosis facility, and country of origin.

HIV-1 Drug Resistance

As shown in Table II, evidence of TDR was found in 257 (18%) patients. No association was observed between TDR and sex, race, age, mode of transmission or recency of infection. One hundred and twenty two patients (9%) exhibited evidence of resistance to non-nucleoside reverse transcriptase inhibitors (NNRTIs), 121 (9%) to nucleoside reverse transcriptase inhibitors (NRTIs), and 76 (5%) to protease inhibitors (PIs). Forty-seven percent of patients with evidence of resistance to PIs had only one mutation, while 50% of patients with evidence of resistance to NRTIs and 66% of patients with evidence of resistance to NNRTIs had only one mutation. The mean number of drug resistance associated mutations to any drug class was 1.5 (range, 1–10). Two hundred and seven cases (15%) had drug resistance associated mutations to one drug class, 38 (3%) to two drug classes, and 12(<1%) to all three drug classes. Among these with dual drug class resistance, 13 patients had evidence of resistance to both NNRTIs and NRTIs, 9 to both PIs and NNRTIs, and 16 to both PIs and NRTIs.

The most common mutation was K103N (84/122, 69%; in Table III), which is associated with high-level resistance to first-line NNRTIs-Efavirenz and Nevirapine. The second most common drug resistance associated mutation, was L90M (36/76, 47%; in Table III), which confers resistance to Nelfinavir and Saquinavir/Ritonavir. M41L (28/121, 23%; in Table III) was the most prevalent NRTI mutation. The M41L mutation is known to be selected by thymidine analogs and confers reduced susceptibility to all approved NRTIs.

Among cases classified as recent HIV infections, 19% (55/290) exhibited evidence of TDR. NRTI drug resistance was present in 9.3% of recent HIV infections, followed by NNRTI (8.6%), and PI (5.9%). Ten (3.4%) individuals had two drug class associated mutations, and two (0.7%) individuals had three drug class associated mutations. Among cases classified as long-standing HIV infections, 16.8% (148/881) exhibited evidence of TDR. NRTI drug resistance was present in 8.6% of long-standing HIV infections, followed by NNRTI (7.5%), and PI (4.2%). Seventeen (1.9%) individuals had two drug class associated mutations, and 7 (0.9%) individuals had three drug class associated mutations. There was no significant difference in the prevalence of TDR in recent and long-standing infection.

Non-B Subtype and CRFs

While most study patients (97.2%) had HIV-1 subtype B, other subtypes and circulating recombinant strains of HIV-1 were observed. In decreasing order of prevalence, these included: subtypes C (1.2%), CRF01_AE (0.8%), CRF02_AG (0.4%), A (0.3%), and F (0.1%). Thirty eight percent of patients with non-B subtype were Black/African American followed by Hispanic/Latino (24%), White (19%), Asian (17%), and Native Hawaiian/Pacific Islander (2%). Forty-six percent of these patients identified as men who have sex with men and 60% were foreign-born. Furthermore, among Black/African American, Hispanic/Latino, and White cases with non-B subtypes, 62%, 60%, and 50%, respectively, were foreign-born.

Phylogenetic analysis of study patients with non-B subtypes revealed CRF01_AE circulating among men who have sex with men and heterosexuals in Los Angeles County. As shown in Figure 2, four of the five patients were men who reported having sex with men and one was a heterosexual female. Of the four men, two were Hispanic/Latino, one was Asian and another was a Pacific Islander. Three of the five patients were foreign-born and two were US-born.

Estimated Prevalence of Transmitted HIV Drug Resistance in Los Angeles County

Compared with the national estimates of TDR prevalence, which range from 12% to 18% [Wheeler et al., 2010], the observed prevalence of TDR in Los Angeles County in the years 2007, 2008, and 2009 (17.1%, 18.4%, and 17.4%, respectively) ranks high. This is the first population-based study to estimate the prevalence of HIV-1 drug resistance associated mutations in newly diagnosed HIV cases in Los Angeles County. By focusing on newly diagnosed persons for whom HIV genotyping was performed on blood specimens obtained within 90 days of HIV diagnosis, this study attempts to obtain a prevalence estimate that approximates TDR.

It is noteworthy that this dataset of newly diagnosed cases was largely (70%) comprised of long-standing HIV cases. Compared with recent infections, newly diagnosed long-standing HIV cases may not provide as accurate a reflection of TDR because over time in the absence of ARVs, the tendency is for resistant HIV to revert to the wild-type [Pingen et al., 2011]. This tendency to revert to the wild-type is the reason some medical providers have argued that there is little benefit in conducting resistance testing in persons for whom there is no evidence of recent HIV infection. In this study sample, however, the prevalence of TDR in recent infections and long-standing infections was not significantly different, which may indicate that TDR-associated mutations persist well beyond the period of recent infection. Other studies have suggested that TDR-associated mutations may persist for a longer period than was previously believed [Barbour et al., 2004; Pao et al., 2004]. Furthermore, certain RT mutations such as M41L, K103N, T69N, and T215 and protease mutations, such as L90M have been observed to show little reversion to wild-type over time [Pao et al., 2004; Bezemer et al., 2006]. The persistence of these resistance associated mutations underscores the utility of performing baseline resistance testing on patients even when there is no evidence of recent infection, as such testing can provide resistance information that may be useful in guiding future ARV treatment choices.

Demographic Correlates of TDR

Similar to other studies, this study did not find any associations between drug resistance mutations and basic demographics [Booth and Geretti, 2007; Palma et al., 2007; Yerly et al., 2007; Wheeler et al., 2010]. By contrast, Youmans reported a negative association between age at HIV and total number of mutations among newly diagnosed treatment-naïve individuals [Youmans et al., 2011]. Readhead et al. found a significant association between TDR and risk group and year of diagnosis [Readhead et al., 2012].

Although HIV-1 subtype B was dominant, 3.1% of study patients were infected with non-B subtypes, indicating the need for continued monitoring of HIV genetic diversity throughout the US to track the transmission and prevalence of non-B subtypes. Compared with other racial/ethnic groups, African-Americans, primarily African immigrants, comprised the largest proportion of non-B subtype HIV-1 infections. The most prevalent non-B subtype was subtype C, which is the main genetic form in southern and eastern Africa. There are, in the literature, examples of subtypes initially occurring solely in immigrants that overtime have been found to represent a major proportion of cases in the non-immigrant population. In France Couturier et al. [2000] reported that, from 1996 to 1998, 15.5% of newly diagnosed HIV cases were non-B subtypes compared with 47.7% from 2003 to 2005. Semaille et al. [2007] also reported evidence of transmission of non-B subtypes within non-immigrants living in France. Furthermore, Hughes et al. [2009] reported that in the United Kingdom, increasing immigration from Southern and Eastern Africa heralded a shift from the observed majority subtype B during the mid-1990s to subtype C (35%) and subtype A (15%) in 2000.

Distribution of CRFs

This phylogenetic analysis indicates that the HIV-1 CRF01_AE virus is currently circulating among Los Angeles County men who have sex with men [Hemelaar et al., 2011]. CRF01_AE is the predominant CRF in Southeast Asian countries. It originated in central Africa in the 1970s and was first identified in Thailand in 1989. It later spread to neighboring countries [McCutchan et al., 1992]. To our knowledge, there has been no published report on the circulation of CRF01_AE strains among Los Angeles County men who have sex with men.

In a recent study in Maryland, the presence of non-B subtype infections in US-born individuals and subtype B infections in African immigrants was shown to be an early indicator of crossover between two initially distinct epidemics [Carr et al., 2010]. Similarly, the circulation of CRF01_AE virus in Los Angeles County may be evidence of the interconnected sexual networks of foreign-born and US-born patients as well as homosexual and heterosexual patients. This circulation pattern may be indicative of crossover of non-B HIV subtypes into US-born populations from immigrants. As has previously been argued in the literature, the impact of increased HIV non-B subtypes infections among US-born individuals could have important implications for antiretroviral treatment and vaccine development, since varying subtypes may influence pathogenicity, antiretroviral drugs susceptibility, and development of resistance [Carr et al., 2010].

Limitations

This study is not without limitations. First, study results may not be generalizable as the sample differs from the overall population of all newly diagnosed cases in Los Angeles County on several influential demographic characteristics. Second, use of antiretrovirals by study participants may have been misclassified due to incomplete or inaccurate reporting of previous antiretroviral drug use. Third, current HIV resistance testing methods may not be sensitive enough to detect low frequency mutations. Finally, the HIV-1 subtypes B, E, and D Capture EIA test is best utilized, as a tool for population-based HIV incidence estimation rather than to determine whether any given case is a recent infection [http://www.cdc.gov/hiv/topics/surveillance/resources/factsheets/pdf/bed.pdf].