836    

Background:  The hypersensitivity reaction (HSR) to abacavir (ABC) is a clinical syndrome with multi-organ/system involvement, seen in about 5% of subjects receiving ABC. Rash is part of the syndrome in about 70% of suspected cases. We did an updated analysis of 37 trials using ABC once or twice daily to identify clinical and other potential determinants of the development of suspected HSR to ABC, and further analyzed a subset of 12 ABC clinical trials to characterize ABC-associated rash within and outside HSR.

Methods:  All GSK trials with ³ 24 weeks of ABC use and authorized databases by January 2004 were analyzed. Cases had a clinical syndrome compatible with HSR or HSR collection modules in the case report form (CRF). Variables analyzed for association with suspected HSR included baseline demographic data, CDC HIV-1 class, CD4⁺ cells, plasma HIV-1 RNA, prior therapy, ABC dosing, concurrent medications, and clinical lab results. After univariate analyses, 2 multivariable models with stepwise selection were constructed with and without baseline lab data. Variables with significance < 0.05 were retained in these models. Rash was further characterized in all trials using the HSR CRF module.

Results:  Among the 37 protocols, the rate of HSR ranged from 0% to 14%. Data from 9329 subjects were analyzed; 501 (5.4%) reported HSR to ABC. Of the variables, 12 yielded p < 0.10 in univariate models. In the multivariable analysis, African ethnicity (OR 0.56), male gender (OR 0.69), baseline CDC class C (OR 0.73), and use of the HSR CRF module (OR 2.08) remained significant. Frequency of ABC dosing did not correlate with risk for reporting HSR to ABC. While 3962 subjects used the HSR CRF module, 653 reported rash with no HSR. Of 304 cases of HSR; 200 (66%) included rash as part of the syndrome. Baseline demographics and characteristics were comparable between patients with and without rash, and between patients with rash with and without a diagnosis of HSR. Most rashes in both groups were mild to moderate. Median time to onset for rashes reported without HSR was 18 days, longer than the median time to onset of HSR-associated rash (11 days). Few subjects (34 of 653, 5%) discontinued ABC following rash without HSR. All subjects discontinued ABC after suspected HSR.

Conclusions:  African ethnicity, male gender, and baseline CDC class C were associated with a reduced risk of HSR, while ABC dosing frequency was not. Use of the HSR CRF module was associated with higher odds of reporting HSR. Minor differences were seen between the rash associated with ABC HSR and the rash reported outside HSR. The presence of rash is insufficient to make a diagnosis of HSR, since 95% of subjects that developed rash without HSR continued ABC. However, when ABC-related HSR cannot be ruled out, ABC should be discontinued.

Keywords: Abacavir; Hypersensitivity; Cutaneous reactions