Response to: Are all non-thymidine analogue backbones appropriate for treating antiretroviral-naïve patients?

To the Editor: The recent review by Waters and Nelson ‘Are all nonthymidine analogue backbones appropriate for treating antiretroviral-naı̈ve patients?’ (1) states incorrectly that the rates of hypersensitivity (HSR) in the Zodiac Study were significantly higher with abacavir (ABC) once daily than with ABC twice daily (8 vs. 5%, p < 0.02). In fact, the rates of suspected ABC HSR in the Zodiac Study were 9% (95% CI 1⁄4 6.6–12.7%) in the oncedaily treatment group and 7% (95% CI 1⁄4 4.9–10.3%) with the twice-daily regimen (2). In this double-blind, placebo-matched study, there was no significant difference in the clinical presentation and/or frequency of HSRassociated signs and symptoms reported between the ABC once-daily and the ABC twice-daily arms. Moreover, the rates of HSR reported in this study were consistent with those in other studies where the symptoms of HSR were solicited using a specific, detailed reporting module (3). Analysis of safety parameters, including HSR, indicated no increased risk with ABC once-daily administration, compared with twicedaily administration (2). James and Johan-Ling refer to a poster by James et al. and they are correct that the overall HSR incidence reported in nine recent clinical trials using twice-daily ABC is 8% (4). By contrast, however, an analysis reported by Brothers et al. at the same conference, involving data from more than 9000 patients in 37 clinical trials, shows an overall rate of suspected hypersensitivity of 5.4% (5). Although the result of the unplanned Fisher’s exact test reported by James et al. in Zodiac, comparing the frequency of grade 3/4 suspected HSRs in the ABC once-daily vs. the ABC twice-daily arms, is p 1⁄4 0.02 (5 vs. 2%), it is important to note that the exact confidence intervals around each of these proportions overlap (95% CI 1⁄4 3.0–7.6% and 0.7–3.7%). The interpretation of this unplanned statistical test should be made with caution. In fact, no significant differences in grade 3/4 HSRs, hospitalisations or deaths attributed to HSR were seen in the ABC once-daily vs. twice-daily arms of six large studies (6), consistent with the lack of a biological rationale for a higher rate or severity of HSR when ABC is given as 600 mg once daily vs. 300 mg twice daily.