Daniel P. Webster

Analysis of transmitted HIV-1 drug resistance using 454 ultra-deep-sequencing and the DeepChek®-HIV system

Next‐generation sequencing (NGS) is capable of detecting resistance‐associated mutations (RAMs) present at frequencies of 1% or below. Several studies have found that baseline low‐frequency RAMs are associated with failure to first‐line HAART. One major limitation to the expansion of this technology in routine diagnostics is the complexity and laboriousness integral to bioinformatics analysis. DeepChek (ABL, TherapyEdge) is a CE‐marked software that allows automated analysis and resistance interpretation of NGS data.

Evaluation of sequencing of HCV core/E1, NS5A and NS5B as a genotype predictive tool in comparison with commercial assays targeting 5′UTR

BACKGROUND Hepatitis C virus (HCV) genotyping is required for tailoring the dose and duration of antiviral therapy, predicting virological response rates, and selecting future treatment options. OBJECTIVE To establish whether baseline genotypes, performed by INNO-LiPA Version 1.0 (v1.0), before 2008, were valid for making treatment decisions now or whether genotypic determination should be repeated. Furthermore, to evaluate concordance between Abbott RealTime genotype II assay (RT) and genotyping by sequencing HCV C/E1, NS5A, NS5B. STUDY DESIGN Genotyping by RT and sequencing was performed on paired historic and current specimens from 50 patients previously baseline genotyped using INNO-LiPA. RESULTS Of 100 samples from 50 patients, ≥ 2 of HCV genomic target regions yielded a sequence that was suitable for genotyping, with 100% concordance, providing no evidence of recombination events. Genotype and subtype prediction based on RT and sequencing agreed in 62.8% historic and 72.7% current specimens, with a kappa coefficient score of 0.48 and 0.76, respectively. LiPA could not subtype 46% of HCV gt1 infections, and LiPA subgenotype was only in agreement with RT and sequencing in 28.6% cases, where matched baseline and historic specimens were available. Three patients were indeterminate by RT, and five patients with HCV gt1 infections could not be subtyped by RT. However, RT revealed mixed infections in five patients where sequencing detected only single HCV infection at 20% threshold. CONCLUSION Genotyping by sequencing, exhibited excellent concordance, with moderate to good agreement with RT, and could resolve RT indeterminates and subtype HCV-gt1 infections not possible by LiPA.

Prevalence of baseline polymorphisms for potential resistance to NS5A inhibitors in drug-naive individuals infected with hepatitis C genotypes 1–4

BACKGROUND The non-structural 5A (NS5A) protein of HCV is a multifunctional phosphoprotein involved in regulation of viral replication and virion assembly. NS5A inhibitors targeting domain I of NS5A protein have demonstrated high potency and pan-genotypic antiviral activity, however they possess a low genetic barrier to resistance. At present, only genotype 1, the most prevalent HCV genotype has been studied in detail for resistant variants. METHODS Utilizing a panel of genotypic-specific resistance assays, population sequencing was performed on plasma-derived viral RNA isolated from 138 patients infected with HCV genotypes 1-4 and not treated with direct-acting antiviral agents. Amino acid changes in HCV NS5A domain I at codon positions 28, 30, 31, 32 and 93, reported to confer reduced susceptibility to certain NS5A inhibitors were examined. Additionally, genotypic outcome based on NS5A sequences were compared with VERSANT HCV Genotype Assay (LiPA) 1.0 (Siemens Healthcare Diagnostics, Surrey, UK) and Abbott m2000 RealTime HCV genotype II assay (Abbott Molecular, Maidenhead, Berkshire, UK). RESULTS Amino acid substitutions associated with moderate to high level resistance to NS5A inhibitors were detected in 2/42 (4.76%) HCV-1a, 3/23 (13.04%) HCV-1b, 4/26 (15.38%) HCV-2, 1/24 (4.17%) HCV-3 and 1/23 (4.35%) HCV-4 infected patients who had not been treated with NS5A inhibitors. Genotype prediction based on NS5A sequences were concordant with LiPA and/or Abbott RealTime for 97.10% of cases. CONCLUSIONS Primary resistance mutations associated with resistance to first-generation NS5A inhibitors such as daclatasvir were observed in all genotypes, albeit at low frequencies. An excellent correlation based on NS5A genotyping and LiPA or Abbott RealTime was achieved.

Comparison of automated chemiluminescence immunoassays with capture enzyme immunoassays for the detection of measles and mumps IgM antibodies in serum

Outbreaks of measles and mumps occur regularly in the UK. Rapid diagnosis of acute infection is important for both infection control and epidemiological purposes. The objective of this study was to compare the performance of an automated platform (DiaSorin Liaison(®), Saluggia, Italy) with a manual capture enzyme immunoassay (EIA; Microimmune, Hounslow, UK) for the detection of measles and mumps IgM antibodies in serum from symptomatic individuals. Ninety sera tested previously for measles (n=50) and mumps (n=40) IgM using the manual EIA were tested retrospectively using the DiaSorin Liaison(®) and the results compared. Sensitivity, specificity, inter-assay variability and intra-assay variability of the Liaison(®) assays were calculated. Sensitivity and specificity of the Liaison(®) assay for measles IgM were 92% and 100% respectively, with inter-assay variation of 14.1% and intra-assay variation of 12.5%. The sensitivity and specificity of the mumps IgM Liaison(®) assay were 88% and 95% respectively, with an inter-assay and intra-assay variation of 13.9% and 5.3% respectively. Both the measles and mumps IgM Liaison(®) assays gave fewer equivocal results than the EIA. Neither Liaison(®) IgM assay showed any cross-reactivity with sera positive against other viruses, however the measles IgM EIA cross-reacted with parvovirus IgM. The automated Liaison(®) assays are more specific, cheaper and less labour-intensive compared to the manual EIA. The Liaison(®) assays benefit from reduced number of equivocal results compared to the EIA for both measles and mumps IgM. This allows clinical decisions to be made accurately and in a timely manner.

Mother to child transmission (MTCT) of HIV - almost a thing of the past? A cohort study of HIV positive women starting antiretroviral drugs in pregnancy

Where available, antiretrovirals (ARVs) and antenatal HIV testing and care have significantly reduced mother to child transmission (MTCT). Higher maternal viral load (VL) is linked with higher risk of MTCT, increasing risk of MTCT for women starting ARVs during pregnancy compared to those already suppressed. A single-centre, retrospective cohort of women starting ARVs during pregnancy from 2004 till 2013. Demographic, obstetric and virological data, and neonatal outcomes were collected where available. 60 pregnancies were recorded (in 56 women) in which ARVs were started or restarted from a total of 129 recorded pregnancies. 48% (27/56) were new antenatal HIV diagnoses and in these median gestational age (GA) at diagnosis was 16.1 weeks (range 5.3-37.6). Median GA at ARV commencement was 22.4 weeks (range 8-37.7) with 63% (35/56) starting before 24 weeks and 91% (51/56) before 28 weeks. HIV diagnosis during pregnancy was associated with a later commencement of ARVs (23.9 vs 19.9 weeks, p=0.009). The ARV regimen was available for 58 pregnancies. Treatment was with two NRTIs plus NNRTI (3), or PI (54) or raltegravir (1). Raltegravir was added as a forth agent in 7 patients. 87% (52/60) had resistance genotyping before (14) or during (38) pregnancy, 81% (22/27) for new diagnoses in pregnancy. The rate of any ARV resistance was 12% (6/52): 4 patients had NNRTI and 2 NRTI resistance mutations, 4 were treatment-naive. This was not associated with treatment failure. HIV VL at delivery was available in 57 pregnancies with detectable VL in 16% (9/57). Pre-treatment VL >100,000 copies/mL during pregnancy was associated with higher risk of detectable VL at delivery (p=0.016). 69% babies were delivered by Caesarean section, 32% as an emergency. There was one late miscarriage at 17 weeks. Median GA at birth was 38 weeks (range 17-42) with 21% born at <37 weeks (10/48). There was one HIV MTCT (1.7% for those starting ARV in pregnancy, 0.8% overall) in a woman who was poorly adherent with ARVs throughout pregnancy with a VL of 216 copies/mL at delivery following a period of directly observed therapy. Over 10 years of integrated HIV and antenatal care, there was only one case of MTCT. Initiating ART for prevention of MTCT is complex, requires a multi-disciplinary approach and importantly patient engagement. Antenatal practices and guidelines have changed over the period of this dataset, allowing normalisation of pregnancy when HIV is diagnosed and treated early.

The utility of different bioinformatics algorithms for genotypic HIV-1 tropism testing in a large clinical cohort with multiple subtypes

Objectives:HIV-1 tropism needs to be determined before the use of CCR5 antagonist drugs such as maraviroc (MVC), which are ineffective against CXCR4-using HIV-1. This study assessed how different computational methods for predicting tropism from HIV sequence data performed in a large clinical cohort. The value of adding clinical data to these algorithms was also investigated. Design and methods:PCR amplification and sequence analysis of the HIV-1 gp120 V3 loop region was performed on triple replicates of plasma viral RNA or proviral DNA extracted from peripheral blood monocytes (PBMCs) in 242 patients. Coreceptor usage was predicted from V3 sequences using seven bioinformatics interpretation algorithms, combined with clinical data where appropriate. An intention-to-treat approach was employed for exploring outcomes and performance for different viral subtypes was examined. Results:The frequency of R5 predictions varied by 22.6%, with all seven algorithms agreeing for only 75.3% of tests. The identification of individuals likely to fail was poor for all algorithms. The addition of clinical data improved this, but at the expense of their ability to predict success. The clinical algorithms varied across subtypes, whereas other algorithms were more consistent. Furthermore, individuals with discordant clonal and clinical predictions were more likely to fail MVC treatment. Conclusion:Eligibility for MVC varied depending on the algorithm method used. The addition of clinical parameters alongside sequence data may help predict X4 emergence during treatment. It could be that V3 loop analysis in isolation may not be the best method for selecting individuals for MVC.

Ultra-deep sequencing provides insights into the virology of hepatitis C super-infections in a case of three sequential infections with different genotypes.

The current epidemic of Hepatitis C infection in HIV-positive men who have sex with men is associated with increasing use of recreational drugs. Multiple HCV infections have been reported in haemophiliacs and intravenous drug users. Using ultra-deep sequencing analysis, we present the case of an HIV-positive MSM with evidence of three sequential HCV infections, each occurring during the acute phase of the preceding infection, following risk exposures. We observed rapid replacement of the original strain by the incoming genotype at subsequent time points. The impact of HCV super-infection remains unclear and UDS may provide new insights.

BASHH/EAGA position statement on the HIV window period.

Dear Editor, We read with interest the paper by Taylor et al. exploring the probability of false-negative HIV serological testing at various timepoints during the window period. The study provides useful data on the comparative performance of laboratory-based thirdand fourth-generation immune assays using published seroconversion panels, i.e. known HIV-positive samples from defined timepoints. The data suggest that the probability of a false-negative fourth generation test at 28 days is 8% but 0% at 50 days. This is of particular interest to countries such as the UK whose British Association of Sexual Health and HIV along with the Expert Advisory Group for AIDS has recently published a position statement, ‘a negative result on a 4th generation test performed at 4 weeks post-exposure is highly likely to exclude HIV infection.’ And further state that, ‘a further test at 8 weeks post-exposure need only be considered following an event assessed as carrying a high risk of infection’. True seroconversion window period data are difficult to obtain, requiring certainty over the timing of a single point exposure, plus the need to take account of the different modes of HIV acquisition that exist (inoculation, sexual, mother to child), infectivity of the source and host response. The paper by Taylor has the advantage of nullifying some of this uncertainty by using modelling to generate an average time from exposure to first HIV RNA positivity. The question of importance to practitioners offering HIV testing is the likelihood of missing a seroconversion following a potential HIV exposure, while those undergoing testing would like reassurance about trusting a negative result. In most HIV testing scenarios, the probability of a false-negative result if a laboratorybased fourth generation test is performed at four weeks post-exposure will be considerably less than 8%. The risk of HIV transmission is equal to the risk that the source is positive (if unknown) multiplied by the risk associated with the type of exposure. Probabilistically, if we examine a high-risk exposure, unprotected receptive anal intercourse (with a transmission risk estimated at 1.11%) occurring in a major city such as London with an HIV seroprevalence in men who have sex with men estimated at 8.5%, then the risk of transmission would be 0.094%. With a false negative assay rate of 8% at 28 days, the probability of missing a new diagnosis at this timepoint would be 0.0075% or 7.5 cases per 100,000 seroconversions. For other risk types, this figure could be a thousand-fold less. Those providing advice should be aware of the performance of the assays used (particularly as pointof-care tests are increasingly used) and consider risk when discussing a negative result. In the overall picture of HIV test counselling, the statement that after a potential exposure event, a negative test at 28 days is ‘highly likely’ to exclude HIV infection is rational and evidence-based. Further testing at eight weeks after exposure will therefore not be required in many cases but will be guided by judgement of risk to definitively exclude infection.

Baseline drug resistance mutations are detectable in HCV genes NS3 and NS5A but not NS5B in acute and chronic HIV-coinfected patients.

In 2012, Plaza et al. [1] reported the prevalence of natural polymorphisms in the HCV NS5A gene associated with resistance to daclatasvir (DCV) in 78 HIV– HCV-coinfected patients and 635 HCV-monoinfected derived NS5A sequences deposited in Los Alamos HCV database. They did not observe NS5A resistanceassociated variants (RAVs) in HCV-1a and HCV-3 NS5A sequences, whereas, major RAVs (Y93H) were detected in 7% and 13% in NS5A sequences from coinfected patients infected with HCV-1b and HCV-4, respectively, with a similar frequency of NS5A RAVs observed in HCV-monoinfected patients for these HCV genotypes (gts) [1]. Additionally, the L31M NS5A variant was observed in 7% of HCV gt-1b patients, irrespective of coinfection status and occurred in >93% of HCV gt-4 monoinfected and coinfected patients [1]. Interestingly, the presence of naturally occurring drug resistance variants in acutely HCV-infected, treatmentnaive HIV patients have been detected by population and deep sequencing in HCV NS3 in a large proportion of subjects [2]. The significance of the threshold at which these RAVs are detectable and whether these will impact on response to therapy with NS3 and NS5A inhibitors in clinical practice is not fully clear. In acute HCV amongst those who are HIV-infected, the role for new HCV direct-acting antivirals (DAAs) has not been established. Treatment with pegylated interferon and ribavirin (PEG-IFN/RBV) early in HCV infection is often successful for most gts [3]. Telaprevir, a firstgeneration protease inhibitor (PI), has been used in a small study of acute HCV infection in HIV, and results suggest sustained virological response (SVR) may be improved using ‘triple therapy’ with shortened treatment durations [4]. It is likely that in the near future DAA-based therapy will be standard of care for both acute and chronic HCV infections, with PIs forming part of this armamentarium [5,6] alongside other DAAs, including NS5B polymerase inhibitors [7] and H-CV NS5A inhibitors [8]. Ultimately, an interferonfree future is heralded, where drug regimens will consist of combinational DAAs, targeting different HCV gene products [9]. It is therefore important to establish the frequency of RAVs in all patients infected with HCV. In this study, we investigated the prevalence of RAVs by population sequencing from three groups of HIV– HCV-coinfected patients: acute HCV infections (n=25), chronic treatment-naive patients (n=20) and chronic treatment-experienced (PEG-IFN/RBV) patients who did not achieve an SVR (n=34) and compared with the prevalence of RAVs in 85 chronic HCV-monoinfected patients. Genomic regions (sites of known RAVs) were amplified from HCV RNA using reverse transcriptase PCR followed by a nested PCR. Typically, amino acids 1–181 of HCV NS3 protease (gt1 only), amino acids 1–213 of domain I of NS5A (for gt-1a, 1b, 2, 3 and 4) and amino acids 219–347 of NS5B (pan-genotypic) were included. Purified PCR amplicons were sequenced using ABI PRISM 3730 genetic analyser (Applied Biosystems, Life Technologies Ltd, Paisley, UK) and consensus sequences aligned against HCV reference sequences. Baseline RAVs were detected in all three cohorts of coinfected patients in NS3 and NS5A but baseline S282T, associated with resistance to sofosbuvir, was not detected in any of our cohorts (Table 1), possibly attributable to the low fitness of this mutation [10]. The Q80K polymorphism was the predominant NS3 variant for gt-1a conferring resistance to firstand secondgeneration PIs [11], and increased in frequency in acute Letter

Influenza immunisation: knowledge and actions taken by UK HIV-positive adults

Influenza is an important cause of respiratory illness in the general population and is responsible for exacerbations of pre-existing respiratory disease (which are more common in people living with HIV) [1,2]. Current evidence suggests that individuals with HIV who are not using antiretroviral therapy (ART) have a higher rate of complications associated with influenza and an increased duration and severity of illness – although the extent to which ART reduces this is uncertain [3,4]. The Trivalent Inactivated Influenza Vaccine (TIV) appears effective in preventing influenza in individuals with HIV [5]. It is recommended for all HIV-positive patients in the UK with a target of 95% coverage [6] although uptake may be suboptimal [7–15]. There are limited data regarding rates of immunisation in the UK HIV-positive population, or the health beliefs concerning influenza immunisation which may influence its uptake. Here we report data from a large metropolitan HIV service regarding the use and health beliefs surrounding influenza immunisation in people living with HIV infection. We conducted a cross-sectional study of HIV-positive adults presenting for routine care at a large urban HIV care service in London, UK, between September and December 2014. Individuals attending clinic appointments within the ambulatory care service during the study period were invited to complete an anonymous questionnaire detailing demographics, ART use, influenza immunisation in the previous year and intentions regarding immunisation in the 2014–2015 season. For those who had not been vaccinated, the reasons underlying this decision were sought. Questions explored each participant’s knowledge regarding the efficacy of the influenza vaccine and the clinical course of influenza in HIV-infected individuals. Two hundred and fifty-three individuals completed the questionnaire between October and December 2014: 195 participants were male (83% of those reporting their gender), 39 female and 19 did not state their gender; 177 (70%) were White, 47 (18.6%) Black African or Black Caribbean and 89% were using ART. The median age of study participants was in the 55–64 age category compared to a median age in the total clinic population of 47 years (interquartile range, 41–53 years). One hundred and sixty-three respondents [64%; 95% confidence interval (CI) 58–70%] had or planned to have an influenza immunisation in 2014–2015 and 171 participants (68%; CI 61–73%) recalled immunisation in 2013– 2014. Of these, 68 (40%) reported that immunisation was given in GP practices and 56 (33%) in HIV care services, with the remainder provided by pharmacies, supermarkets and workplaces. Immunisation rates were lower in younger patients, with 157 of 223 (70%) above the age of 45 years reporting immunisation in the 2013–2014 season compared to 14 of 30 (47%) younger participants (P = 0.018, vtest). Immunisation rates were greater in women: 34 of 39 women (87%) reported immunisation compared to 65% of men (127 of 195, P = 0.025). There was no significant difference between White and NonWhite ethnic groups. The use of ART was associated with a higher rate of immunisation uptake (70% vs. 43% P = 0.01), although only 21 participants were not using ART. Amongst 65 participants stating that they had not been immunised in 2013–2014, reasons given were as follows: 16 (25%) did not think that they needed immunisation, 15 (23%) reported concern about adverse events and 14 Correspondence: Marc Lipman, Centre for Respiratory Medicine, Royal Free London NHS Foundation Trust, Pond Street, London NW3 2QG, UK. Tel: + 44 207 317 7560; fax: +44 207 317 7561; e-mail: marclipman@nhs.net