Validation of a stable isotope gastric emptying test for normal, accelerated or delayed gastric emptying

To validate a 13C-Spirulina platensis breath test for measurement of accelerated or delayed gastric emptying, we measured gastric emptying of egg containing 13C-S. platensis and 99mTc-sulphur colloid by breath 13 CO2 every 15 min over 3 h and scintigraphy every 15-30 min over 5 h in 57 healthy volunteers. Thirty-three received no treatment, 10 received erythromycin, and 14 atropine. A generalized linear regression model predicted half-emptying time by scintigraphy (t1/2S) from breath 13CO2 (t1/2B) data. Accuracy was assessed by standard deviation (SD) of differences between t1/2S and t1/2B and by receiver operating characteristic (ROC) curves. Regression models using breath samples at baseline, and 45, 90, 105 and 120 min, predicted t1/2B (mean +/- SD) at 118 +/- 59 min, similar to t1/2S (118 +/- 67 min). Correlation between t1/2B and t1/2S was significant (r=0.88; P < 0.0001). Differences between t1/2S and t1/2B were: 18-19.2 min for t1/2 < 70-150 min, and 68.3 min for t1/2 > 150 min. Breath test detected abnormal emptying with a sensitivity of 86% and specificity of 80%. Thus, the 13C-S. platensis test measures gastric emptying t1/2 for solids, which is accelerated or delayed to mimic a range of conditions from dumping syndrome to severe gastroparesis, with high sensitivity and specificity. Additional breath samples are needed to increase sensitivity in detecting accelerated gastric emptying.