Variations in colonic H2 and CO2 production as a cause of inadequate diagnosis of carbohydrate maldigestion in breath tests.

BACKGROUND: Lactose maldigestion is usually diagnosed by means of the H2 breath test. When 13C-lactose is used as substrate, a 13CO2 breath test can be performed simultaneously. In an earlier publication we described the relation between both the H2 and 13CO2 exhalation in breath and the measured intestinal lactase activity after consumption of 13C-lactose. We found a discrepancy between both breath test results in 36% of the cases. To investigate the possible cause of these incongruous breath test results, we studied gas production from carbohydrate in the colon, using 13C-lactulose as a non-absorbable substrate.
METHODS: Experiments were performed in 21 subjects, by applying 5 different doses of 13C-lactulose. Repeatability studies were performed in six of these subjects, using 10 g substrate (three tests with 1-week intervals).
RESULTS: Both the H2 and the 13CO2 excretion in breath varied strongly interindividually and intraindividually after consumption of 13C-lactulose. In both cases no dose-response relation was observed. A significant positive linear relationship was found between H2 and 13CO exhalation (r = 0.45, P < 0.005). Extrapolation of these results to 13C-lactose breath tests indicates that the colonic contribution of 13CO2 production to the total 13CO2 excretion in breath varies but is on the average large enough to cause false-negative 13CO2 breath test results.
CONCLUSIONS: Excretion in breath of 13CO2 produced in the colon during a 13C-lactulose breath test correlates with the breath H2 excretion. This could explain the occurrence of false-negative 13CO2 lactose breath tests when colonic gas production is high and false-negative lactose H2 breath test results when gas production is low. It can also explain the improved sensitivity of the combined H2/13CO2 lactose breath test compared with both breath tests alone.