To the Editors:
A 3-drug regimen comprising nevirapine plus 2 nucleoside reverse transcriptase inhibitors is frequently prescribed for HIV-1 infection worldwide. Unfortunately, drug hypersensitivity, which can include severe hepatotoxicity during the initial weeks of therapy with nevirapine, particularly when initiated in women with >250 CD4+ T cells/mm3, have prompted changes in prescribing guidelines. Although isolated nonsevere skin rash associated with nevirapine may be managed conservatively, the major treatment-limiting side effect of nevirapine is a drug hypersensitivity syndrome comprising fever, skin rash, and/or hepatitis. This affects as many as 5% of those starting the drug.
Various human leukocyte antigen (HLA) classes I and II alleles have been associated with nevirapine rash and/or hepatitis across different populations.1–6 A population-based study from Western Australia first identified an association between HLA-DRB1*0101 and nevirapine-associated hepatitis among patients with ≥25% CD4+ T cells.1 This work was complemented by ex vivo studies suggesting a CD4 T-cell–dependent process. Associations were subsequently identified between various HLA class I alleles such as HLA-B*1402 and HLA-Cw8 and nevirapine hypersensitivity in Sardinian2 and Japanese3 populations. A case–control study in a Thai population identified an association between HLA-B*3505 and nevirapine-associated rash or hypersensitivity.4 Additional analyses from Western Australia confirmed the original association between HLA-DRB1*0101 and ≥25% CD4 T cells and identified an association between HLA-B*3501 and nevirapine-associated rash, suggesting that phenotype was particularly important for defining specific HLA associations.5 Only in the presence of hepatitis was HLA-DRB1*0101 associated with rash, whereas HLA-B*3501 was associated with rash regardless of hepatitis.5 A retrospective analysis involving 76 sites in 11 countries (none in Africa) enrolled 276 cases who had experienced severe rash and/or hepatitis within 8 weeks of nevirapine initiation and 587 nevirapine-tolerant controls.6 Based on 2-digit HLA typing, that study associated HLA-DRB1*01 with hepatitis in whites, both HLA-B*35 and HLA-Cw*04 with rash in Asians, and HLA-Cw*04 alone with rash in whites and blacks. Furthermore, CYP2B6 slow metabolizer genotype was associated with rash but not hepatotoxicity.
The above studies suggest that diverse genetic, immunological, and drug metabolism pathways contribute to nevirapine hypersensitivity syndromes and that associations may differ by specific toxicity phenotype and population. Data from Africa, where HIV prevalence and nevirapine use is high, have been scant. Herein, we used data and specimens previously collected during a clinical trial in South Africa to characterize relationships between HLA I and II alleles and nevirapine hepatotoxicity.
MATERIALS AND METHODS
We analyzed data and specimens from participants who initiated nevirapine during Gilead protocol FTC-302, a cohort previously studied for associations with CYP2B6 and ABCB1 polymorphisms.7 This randomized, double-blind trial enrolled antiretroviral naive participants in South Africa in 1999 and 2000 to compare efficacy and safety of emtricitabine and lamivudine.8 Eligible participants had >5000 plasma HIV-1 RNA copies/mL and ≥200 CD4 cells/mm3. Participants were stratified to receive either nevirapine (200 mg once daily for 2 weeks, then 200 mg twice daily) or efavirenz (600 mg once daily). Because of limited data regarding nevirapine efficacy in participants with plasma HIV-1 RNA >100,000 copies/mL, participants with plasma HIV-1 RNA ≤100,000 copies/mL were stratified to receive nevirapine whereas those with higher viral loads received efavirenz. All participants were also randomized 1:1 to receive emtricitabine (200 mg once daily) or lamivudine (150 mg twice daily), and all received stavudine (30 or 40 mg twice daily depending on weight). Only nevirapine recipients were included in the present analysis. Hepatotoxicity was defined as treatment-emergent grade 3 (5–10 times upper limit of normal) or grade 4 (greater than 10 times upper limit of normal) elevation in serum aminotransferase (alanine aminotransferase or aspartate aminotransferase), with or without associated symptoms. To preserve anonymity, the study sponsor matched each case with 2 controls on: race, black versus other; sex; age; baseline CD4; and baseline HIV-1 RNA. Fever and rash were not exclusionary. Of all patients with liver toxicity on nevirapine, 45 had grade 3/4 toxicity within 4 weeks of starting nevirapine, 1 had toxicity between 4 and 12 weeks, and 11 had toxicity beyond 12 weeks. This study complied with the Declaration of Helsinki. Institutional review boards at Vanderbilt University, Duke University, University of Witwatersrand, and Murdoch University approved this use of DNA.
Cryopreserved peripheral blood mononuclear cells underwent DNA extraction with the Agencourt Genfind v2 DNA purification system (Beckman Coulter, Sydney, Australia). High-resolution HLA-ABC and -DR typing was done with standard sequence-based typing at the Institute for Immunology and Infectious Diseases, Murdoch University.9
Among 385 nevirapine recipients on FTC-302, 66 (17%) experienced hepatotoxicity, which occurred within the first 12 weeks of therapy in 80%.8 Of these 66 cases, 63 had peripheral blood mononuclear cells available of which 57 (87%) had adequate extracted DNA available for analysis. A total of 126 controls were identified, of whom 111 (88%) had specimens with adequate DNA for analysis following amplification. Baseline characteristics of cases and controls were similar. Mean age was 32 ± 7.0 years (±SD); 113 (67%) were female; 120 (71%) were black, 8 (14%) were white, 19 (11%) were colored (mixed race), 3 (2%) were Asian, and 1 (1%) was other; median baseline CD4 T cells were 381 (302–486) cells/mm3 (interquartile ratio); and mean baseline HIV-1 RNA was 4.3 ± 0.55 log10 copies/mL (mean ± SD).
By univariate analysis, HLA classes I and II types associated with increased likelihood of hepatotoxicity included B*0705, B*1801, B*5801, C*0302, C*1203, and DRB1*0102 (Table 1, top). Conversely, HLA types associated with decreased likelihood of hepatotoxicity included A*6801, B*1510, C*0210, C*0304, and C*0602 (Table 1, top). By multivariable analysis, only B*5801 and DRB1*0102 were independently associated with increased likelihood of hepatotoxicity, whereas B*1510, B*5802, and C*0210 were independently associated with decreased likelihood of hepatotoxicity (Table 1, bottom). There was no association between the nevirapine slow metabolizing genotype CYP2B6 516 G>T and hepatoxicity, and there was also no association between this drug metabolizing genotype and HLA class I or II alleles in either model.
Various associations have been reported between HLA types and nevirapine-associated adverse events, but data from Africa have been lacking. We provide the first report in a predominantly black South African cohort that both HLA-B*5801 and HLA-DRB1*0102 were independently associated with increased likelihood of hepatotoxicity during combination antiretroviral therapy that included nevirapine. Hepatotoxicity in FTC-302, and therefore in the present analysis, was almost certainly due to nevirapine, since grade 3 or 4 hepatic transaminase elevations affected 17% of nevirapine recipients but 0 of 83 efavirenz recipients.8
The matched case–control design used in this study minimizes effects of potential confounders and may increase power to identify genetic associations. Although this prevented us from exploring associations between the matching variables such as CD4+ T-cell count/percentage and hepatotoxicity, extensive analyses of possible nongenetic predictors of hepatotoxicity have already been reported for FTC-302. Female sex, low body mass index, serum albumin, plasma HIV RNA, high mean corpuscular volume, lactate dehydrogenase, and aspartate aminotransferase levels were significantly and independently associated with subsequent risk of hepatotoxicity in nevirapine recipients, whereas hepatitis B or C coinfection, other potential hepatotoxic drugs, and traditional medication use were not. The well-established association of nevirapine-induced liver toxicity with higher CD4+ T-cell counts was not apparent in FTC-302 likely because few nevirapine recipients had lower CD4 cell counts.8
The translation of HLA-B*5701 screening into routine HIV clinical care has markedly reduced the occurrence of abacavir hypersensitivity reactions and improved patient safety. Reliable genetic predictors of nevirapine toxicity might allow HIV therapy that includes nevirapine to be initiated at higher CD4+ T-cell counts. In addition, the association of HLA class I alleles such as HLA-B*35 in multiple studies and HLA-B*5801 in this study is a strong evidence that CD8+ and not just CD4+ T cells contribute to the immunopathogenesis of nevirapine hypersensitivity. This is further supported by studies in nevirapine-hypersensitive patients showing nevirapine-specific HLA class I-restricted CD8+ cells and significant abrogation of nevirapine specific ex vivo responses with depletion of CD8+ T cells.5 HLA-B*5801 has also been associated with allopurinol, a drug used to treat gout, which causes Stevens–Johnson syndrome/toxic epidermal necrolysis and hypersensitivity in Han Chinese populations.10 The association between HLA-DRB1*0102 and hepatotoxicity in this South African nevirapine-treated population is intriguing given that HLA-B*DRB1*0102 only differs from HLA-DRB1*0101 by 2 amino acids at position 85 (val → ala) and 86 (gly → val), and HLA-DRB1*01 has been reproducibly associated with nevirapine rash-associated hepatitis in white populations.1,5,6 Class I alleles, which have previously been associated with nevirapine hypersensitivity such as HLA-B*35, B*1402, and Cw8 were not prevalent in this population, and therefore those associations could not be adequately reexamined here.
Recent studies have highlighted varying HLA associations with nevirapine hypersensitivity, particularly in white and Southeast Asian populations. Our results for the first time support the diversity of HLA classes I and II alleles associated with nevirapine hepatotoxicity in particular black African populations, suggest some HLA alleles may indeed be protective against the development of nevirapine hepatotoxicity, and underscore the need for HLA studies of drug hypersensitivity in populations where different HLA alleles may be prevalent.
The authors are grateful to the persons with HIV infection who volunteered for this study.
1. Martin AM, Nolan D, James I, et al.. Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4 T-cell counts. AIDS. 2005;19:97–99.
2. Littera R, Carcassi C, Masala A, et al.. HLA-dependent hypersensitivity to nevirapine in Sardinian HIV patients. AIDS. 2006;20:1621–1626.
3. Gatanaga H, Yazaki H, Tanuma J, et al.. HLA-Cw8 primarily associated with hypersensitivity to nevirapine. AIDS. 2007;21:264–265.
4. Chantarangsu S, Mushiroda T, Mahasirimongkol S, et al.. HLA-B*3505 allele is a strong predictor for nevirapine-induced skin adverse drug reactions in HIV-infected Thai patients. Pharmacogenet Genomics. 2009;19:139–146.
5. Phillips E, Keane N, Blyth C, et al.. Both HLA class I restricted CD8+ and class I restricted CD4+ T cells are implicated in the pathogenesis of nevirapine hypersensitivity. Proceedings of the 51st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, September 17–20, 2011. Abstract H1-1399c.
6. Yuan J, Guo S, Hall D, et al.. Toxicogenomics of nevirapine-associated cutaneous and hepatic adverse events among populations of African, Asian, and European descent. AIDS. 2011;25:1271–1280.
7. Haas DW, Bartlett JA, Andersen JW, et al.. Pharmacogenetics of nevirapine-associated hepatotoxicity: an Adult AIDS Clinical Trials Group collaboration. Clin Infect Dis. 2006;43:783–786.
8. Sanne I, Mommeja-Marin H, Hinkle J, et al.. Severe hepatotoxicity associated with nevirapine use in HIV-infected subjects. J Infect Dis. 2005;191:825–829.
9. Mallal S, Nolan D, Witt C, et al.. Association between the presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet. 2002;359:727–732.
10. Hung SI, Chung WH, Liou LB, et al.. HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci U S A. 2005;102:4134–4139.