In the [(13)C]-octanoate breath test, two popular parameters have been used to quantify gastric emptying rates, namely the time to the maximal [(13)CO(2)] excretion (T(max)) and the time to the half-[(13)CO(2)] recovery (T(1/2b)). Although each of T(max) and T(1/2b) is closely correlated with the scintigraphic half-emptying time, the two parameters occasionally indicate different judgments on a gastric emptying rate. In this study, to clarify which of the two parameters is more reliable, T(max) and T(1/2b) were compared to the "reference" parameters calculated using the Wagner-Nelson method, which allows accurate estimation of a time-course of gastric emptying from breath data. Ten healthy male volunteers underwent the breath test after ingestion of a muffin meal (320 kcal) containing 100 mg [(13)C]-octanoate. Breath samples were collected at 15-min intervals for 6 h. According to the conventional analytical algorithm, T(max) and T(1/2b) were mathematically calculated. By applying Wagner-Nelson analysis to the breath test, the time-percent gastric retention curve was generated and the half-emptying time (T(1/2WN)) was determined. T(1/2WN) was more closely correlated with T(max) (r=0.954, P<0.0001) than with T(1/2b) (r=0.782, P=0.008). T(max) was significantly correlated with the percent gastric retention value in the early (t=0.25 and 0.5 h), the middle (t=1.0 and 1.5 h), and the late (t=2.0 h) postprandial phase. T(1/2b) was significantly correlated with the gastric retention value in the middle and the late phase, but not with the gastric retention value in the early phase. The present results show that T(1/2b) has limited capability to reflect gastric emptying in the early postprandial period, suggesting that T(max) is more reliable than T(1/2b) as a gastric emptying parameter.