A systematic study of a neutral meal suitable for subjects undergoing 13CO2 breath tests.

Background: Some 13CO2 breath tests require prolonged breath sampling. The study aimed to validate a meal which could be served to subjects without interference with the measured breath 13CO2 profile.

Material/methods: Eight healthy volunteers were examined on three separate days. On two occasions they took equicaloric meals of 320 kcal composed of foodstuffs known to be poor in 13CO: cooked rice (R) or a wheat bread sandwich with butter and ham (S). On the control day they fasted. Samples of expiratory air were taken every 15 min for 6 hours. The 13CO2 concentration in the samples was measured by isotope ratio mass spectrometry. Statistical analysis involved repeated measures analysis of variance (RANOVA).

Results: The intra-subject variability of breath 13CO2 concentration was remarkably low on the three study days, as reflected by mean coefficients of variation (CV) of 0.90+/-0.32%, 0.79+/-0.19%, and 0.80+/-0.34% for the control, meal S, and meal R days, respectively (RANOVA: F2;14=0.296, p=0.75). RANOVA indicated a strong effect of a meal on inter-subject variability of breath 13CO2 concentration (F2;46=390.62 p<0.000001). Post hoc comparisons revealed a mean CV of 1.64+/-0.37% on the control day, 3.62+/-0.18% after ingestion of S, and 3.81+/-0.24% after meal R (both p=0.000128 vs. the control). Nevertheless, with neither test meal did RANOVA disclose at any time a statistically significant deviation in breath 13CO2 concentration from the basal fasted value.

Conclusions: Both the test meals fulfilled the requirement of neutrality as concerns a null effect on the breath 13CO2 profile.