This study determined the within-subject and between-subject variability of different ways of expressing the results of the (13)C-aminopyrine breath test ((13)C-ABT) and the effect of shortening the test duration. The (13)C-ABT was conducted on three separate occasions in 10 healthy volunteers and on a single occasion in 22 patients with established liver cirrhosis. The within-subject variability of cumulative percentage dose recovered (cPDR), using measured CO(2) production rate (VCO(2)), in the reference group over three trials was 15% over 120 min. Higher within-subject variability in cPDR would have been evident if the test was terminated at either 30 or 60 min. Substitution of predicted VCO(2) to calculate cPDR yielded comparable values at all time points. Significant differences between cirrhotics and reference group were evident after just 10 min using PDR/h, cPDR or enrichment (all P<0.05). The ABT demonstrates clinically acceptable reproducibility. Shortening of the duration may make the test more acceptable clinically, but it is associated with increasing imprecision.