Comparison between gastric scintigraphy and the [13C]-acetate breath test with Wagner-Nelson analysis in humans.

1. The [(13)C]-acetate breath test (ABT) quantifies gastric emptying as the half [(13)CO(2)]-excretion time (T(1/2b)), but T(1/2b) differs from the scintigraphic half-emptying time (T(1/2s)). The aims of the present study were to accurately determine the half-emptying time by ABT with Wagner-Nelson analysis (T(1/2WN)), to compare T(1/2WN) with T(1/2s) and to validate the Wagner-Nelson strategy in ABT. 2. For a comparative study, eight volunteers simultaneously underwent ABT and scintigraphy. Anterior images were acquired and breath samples were collected every 15 min for 4.0 h after ingestion of a 200 kcal liquid meal labelled with 37 MBq [(99m)Tc]-colloidal sulphur and 100 mg [(13)C]-acetate. For the validation experiment, another six volunteers underwent ABT, on two randomized occasions, using the 200 kcal liquid meal with 100 mg [(13)C]-acetate. On either of the two occasions, a gel-forming agent was stirred into the meal to intentionally delay gastric emptying by increasing meal viscosity. Breath samples were collected at regular 15 min intervals for 4 h post ingestion. 3. The Wagner-Nelson equation for ABT is F(t) = (A(breath)(t) + C(t)/0.65)/A(breath)(infinity), where F(t) is a fractional dose of the [(13)C] label emptied, C(t) is the [(13)CO(2)] excretion (% dose/h), A(breath)(t) is the area under the C(t) curve (% dose) and A(breath)(infinity) is the ultimate [(13)CO(2)] recovery in breath (% dose). The percentage gastric retention was estimated as 100 x (1 - F(t)). The time plots of scintigraphic activity and 100 x (1 - F(t)) were fitted to y(t) = 100 x e(-Kxt), K values were estimated mathematically for each plot by regression analysis and T(1/2s) and T(1/2WN) were calculated as (ln2)/K. The time versus pulmonary [(13)CO(2)] excretion plots were fitted to z(t) = m x k x beta x e(-kt)(1 - e(-kxt))(beta-1), where m, k and beta are constants; T(1/2b) was calculated as -(ln(1 - 2(-1/beta))]/k. 4. Values of T(1/2WN) were closer to T(1/2s) than T(1/2b), although T(1/2WN) and T(1/2b) yielded significant under- and overestimation of T(1/2s), respectively. The high viscosity meal significantly prolonged T(1/2WN) and T(1/2b); T(1/2WN) could detect the delayed transit of the viscous meal more sensitively than T(1/2b). 5. The Wagner-Nelson method improves the accuracy of the ABT.