1. Prospective epidemiological study of the prevalence of HLA- B*5701 in HIV-1 infected UK subjects Chloe Orkin 1, S. Tariq Sadiq 2, Felix Jackson 3, Leanne Rice 3, Bridin McCaughey 3, Shema Tariq 4, Laura Cunningham 1, Helen Pearce 3 1 Barts & the London NHS Trust; 2 St Georges, University of London; 3 GlaxoSmithKline; 4 Homerton Hospital NHS Trust 11 th European AIDS Conference, 25–26 October 2007, Madrid, Spain

2. Aim: Human Leukocyte Antigen (HLA) B*5701 is strongly associated with the risk of a hypersensitivity reaction to abacavir in Caucasians and Hispanics. To date limited data exist on HLA- B*5701 prevalence in HIV-1 infected subjects in the UK. We aimed to determine HLA-B*5701 prevalence in the general HIV-1 infected population and in specific ethnic groups such as black Africans whom, in general, exhibit greater genetic stratification. In addition, we compared HLA-B*5701 results obtained from local laboratories with those from a central provider. Methods: Multi-centre, observational study. All HIV-1 infected adult individuals receiving care at participating centres were eligible, irrespective of treatment status or prior exposure to abacavir. Subjects provided samples for HLA-B*5701 assessment by both local (blood) and central laboratories (buccal swabs). Demographic data were collected. Analysis of HLA-B*5701 prevalence was carried out for all subjects recruited (adjusted for their UK ethnicity split), and by main ethnic group. Results: This analysis includes 1494 subjects (618 (41%) white, 770 (52%) black) from eight UK centres. 89% of black subjects reported a country of origin in Africa. Mean age was 41 years (SD 9 years), with 64% (952/1494) being male. HLA-B*5701 prevalence by central laboratory analysis, adjusting for ethnicity split, was 4.77% (95% confidence interval [CI] 3.69% to 5.85%). Among white subjects the prevalence was 7.93% (5.80% to 10.06%) and among black subjects, two (country of origin: Uganda for both) were HLA B*5701 positive giving a prevalence of 0.26% (0.07% to 0.94%). Conclusions: HLA B*5701 prevalence was similar to previously reported rates in white HIV-infected subjects but considerably lower than those reported in black HIV infected subjects, perhaps due to the large proportion of black subjects that were of African origin. Abstract

3. Human Leukocyte Antigen (HLA) B*5701 is strongly associated with the risk of a hypersensitivity reaction to abacavir in Caucasians and Hispanics 1. Limited data exist on HLA-B*5701 prevalence in HIV-1 infected subjects in the UK 2, specifically in ethnic groups such as black Africans whom, in general, exhibit greater genetic stratification. HLA B*5701 prevalence rates in Africa can vary considerably 3. Introduction Objectives Primary To determine the prevalence of HLA-B*5701 in the UK HIV-1 infected population Secondary To assess the prevalence of HLA-B*5701 in the major UK ethnic groups in the study population To describe the HLA-B*5701 results provided by local laboratory methods with respect to central laboratory methods

4. Multi-centre*, observational Inclusion criteria: HIV-1 infected subjects over 18 years of age Subjects provided buccal cell and blood samples which were used to assess HLA-B*5701 status by central and local methodologies respectively. Prevalence rates were based on central lab methodology. The population was over-sampled for African American/African Heritage subjects to allow for greater precision for the lower expected prevalence of HLA-B*5701 Confidence intervals for small prevalences were calculated using the Wilson score 4 *See Acknowledgements below Ethnic sub-division Subjects reported their geographic ancestry (as categorised in previous studies) and those of their parents and grandparents as follows (multiple categories allowed): African American/African HeritageAmerican Indian or Alaskan Native Asian – Central/South AsianAsian – East Asian Heritage Asian – Japanese HeritageAsian – South East Asia Heritage Native Hawaiian or Other Pacific IslanderWhite – Arabic/North African Heritage White – White/Caucasian/European Heritage Subjects reported their country of origin and those of their parents and grandparents. Subjects were sub-divided* using all reported geographic ancestries and all countries (subjects could be in multiple sub-divisions due to differing parental/grandparental ancestry): White – White/Caucasian/European Heritage White/Eurasian – Europe, Arabic countries, Australia, Canada, Malta, New Zealand, Russia, USA African American/African Heritage NigerCongo (Bantu) – Bantu region from West and Central Africa Black Caribbean/African American – Caribbean, South American or USA Other Black African – Afro-Asiatic, Nilo-Saharan or Khoisan region Asian – Central/South Asian or East Asian Heritage South Asian – India, Bangladesh, Pakistan, Sri Lanka, Goa or China Asian – Central/South Asian, East Asian or Japanese Heritage East Asian and Oceanics – Cambodia, China, Hong Kong, Indonesia, Japan, Laos, Malaysia, Thailand Unclassifiable – Country of origin too genetically mixed for classification (e.g. Tanzania, Swaziland, Congo, Brazil, Seychelles, Africa, Asia) – Geographic ancestries and countries too diverse to classify *White, Eurasian and Asian groups by previously reported clustering by autosomal genetics; black African groups by ethnologue (linguistic classification) as a proxy for genetic clustering based on predominant tribe or language in the country of origin. Methods

5. Table 1. Prevalence of HLA-B*5701 for all subjects and main ethnic groups Demographics No. subjects recruited1502 No. subjects with a central laboratory result1494 Mean age 41 yrs (sd 9 yrs) Male 64% (952/1494) African American/African Heritage: 53%* (791/1494) White – White/Caucasian/European Heritage: 43% (647/1494) *Intentionally over-sampled; UK HIV-infected population 5 : White 53.2%, Black 37.1% n HLA-B*5701 positive Prevalence % (95% CI) Overall149456 4.77* (3.69 - 5.85) White61849 7.93 (5.80 - 10.06) Af Am / African7702 0.26 (0.07 - 0.94) * Weighted by ethnicity based on 2003 SOPHID data 5 Results

6. Table 2. Prevalence of HLA-B*5701 for ethnic sub-divisions n HLA-B*5701 Positive Prevalence % (95% CI) White/Eurasian All subjects -all grandparent/parent the same -% homogenous ethnicity 654 577 88.2% 52 49 94.2% 7.95 (5.88 - 10.02) 8.49 (6.22 - 10.77) NigerCongo (Bantu) All subjects - all grandparent/parent the same - % homogenous ethnicity 575 514 89.4% 3 1 33.3% 0.52 (0.18 - 1.52) 0.19 (0.03 - 1.09) Black Caribbean/African American All subjects -all grandparent/parent the same - % homogenous ethnicity 82 61 74.4% 0 0% 0.00 Other Black African All subjects -all grandparent/parent the same - % homogenous ethnicity 78 61 78.2% 0 0% 0.00 South Asian All subjects -all grandparent/parent the same - % homogenous ethnicity 42 20 47.6% 2 1 50% 4.76 (1.31 - 15.79) 5.00 (0.89 - 23.61) East Asian & Oceanics All subjects -all grandparent/parent the same - % homogenous ethnicity 12 8 66.7% 0 0% 0.00 Unclassifiable All subjects -all grandparent/parent the same - % homogenous ethnicity 244 102 41.8% 5 0 0% 2.05 (0.27 - 3.83) 0.00 Ethnic sub-division Subjects from 94 countries were represented in our sample. Following sub-division as described above, the prevalence of HLA-B*5701 was calculated for all subjects and for subjects whose self-reported geographic ancestry and country of origin was homogenous with those reported for their parents and grandparents (see Table 2). The percentage of homogenous subjects was calculated to provide an indication of genetic variability within each sub-division. Results (cont.)

7. Figure 1. Prevalence of HLA-B*5701 for self-reported country of origin (>50/country) No. subjects: 541 215 125 56 55 1502 subjects; 1494 with central lab; 1479 with both central and local lab 99.9% agreement between the local and central labs. Two results differed: - One false negative (local lab) was found to be a transcription error from the notes onto the case report form - One false positive (local lab) was an HLA-B*570301 reported as HLA- B*5701 (on two separate samples) by two-stage SSP methodology. Central vs local laboratories

8. HLA B*5701 prevalence in the UK was similar to previously reported rates 6 in white HIV-infected subjects but considerably lower than those reported in black HIV infected subjects, perhaps due to the large proportion of black subjects that were of African origin. 99.9% of samples tested at the local laboratories agreed with those tested at the central laboratories The prevalence rates identified in the black subjects is considerably lower than previously reported rates, perhaps due to the large proportion of black subjects that were of African origin. Sub-dividing subjects allowed clinically meaningful categorisation using self-reported geographic ancestry and country of origin. However, the 244 unclassifiable subjects underline the difficulties in basing decisions on these self-reported measures of ethnicity. Despite an overall sample size of 1502 only two sub-divisions had a sufficiently large sample to allow meaningful interpretation of the prevalence rate (White/Eurasian and NigerCongo with prevalence rates of 7.95% (CI 5.88% - 10.02%) and 0.52% (CI 0.18% - 1.52%) respectively.) Prevalence rates in different African countries is similar to previously reported rates (Zimbabwe 0.0%, Uganda 2.4%) 3, reflecting the likely greater utility of B*5701 testing in some African groups compared with others, although the small sample size may lead to sampling bias. Only one of the three HLA-B*5701 positive subjects in the NigerCongo sub-division self-reported as genetically homogenous. The false negative result reinforces the need for robust reporting mechanisms. The false positive result reinforces the need for robust laboratory quality assurance. It is reassuring that SSP methodology failed safe by identifying the patient as HLA-B*5701 positive. Conclusions Discussion

9. References Acknowledgements We acknowledge the significant efforts of clinic and research staff at: Barts & the London NHS Trust (Dr Chloe Orkin, James Hand, Carl DeSouza, Dr Rebecca OConnell, Duncan Scott, Paul Davis, Are Isaksen, Stephen Myall, Liz Spellman, Daphne Gibbs, Sai Gomez, Katie Holmes), Guys and St Thomass NHS Foundation Trust (Dr Cindy Sethi, Isabelle Jendrulek, Alice Sharp, Fiona Makia, Dr Ranjababu Kulasegaram), Homerton University Hospital (Dr Jane Anderson, Shema Tariq, Sara Paparini, Mohamed Rogers, Lorraine Muromba), Queen Elizabeth Hospital NHS Trust (Dr Stephen Kegg, Dr Sue Mitchell, Dr Judy Russell, Dr Meg Hunter, Kim Perez, Jayne Clark), St Georges Healthcare NHS Trust (Dr Tariq Sadiq, Ade Adebeyi, Muchanetta Ndorro, Marguerite Cockerill, Dr Hay, Dr Richard Lau, Dr Melanie Rosevinge, Dr Mark Pakianathan, Dr C Fernando), St Marys NHS Trust (Dr Winston, Dr Ken Legg, Norman Gariwa, Dr Simon Portsmouth), Walsall Manor Hospital (Dr Arumainayagam, Dr S Chandramarni, Helen Lathe), Whittall Street Clinic (Professor Jonathan Ross, Louise Brown, Katrina Hood) 1.Mallal S, Phillips E, Carosi G et al. PREDICT-1: a novel randomised prospective study to determine the clinical utility of HLA-B*5701 screening to reduce abacavir hypersensitivity in HIV-1 infected subjects (study CNA106030). 4 th IAS Conference on HIV Pathogenesis, Treatment and Prevention, 22 nd -25 th July 2007, Sydney, Australia. Abstract WESS101 2.Reeves I, Churchill D, Fisher M. Screening for HLA-B*5701 reduces the frequency of abacavir hypersensitivity reactions. 8 th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, 24 th -26 th September 2006, San Francisco, CA, USA. Abstract 14 3.Sadiq ST, Pakianathan M. Uncertainties of routine HLA B*5701 testing in black African HIV cohorts in the UK. Sex Transm Inf 2007 83:173 4.Wilson EB. Probable inference, the law of succession, and statistical inference. Journal of the American Statistical Association 1927; 22: 209 – 212 5.HPA. SOPHID Regional Data. 2003 6.GSK UK. Kivexa, SmPC www.emea.eu.intwww.emea.eu.int