Effectiveness of breath testing correction

Introduction
Breath testing of hydrogen (H2) and methane (CH4) is widely used for assessment of carbohydrate malabsorption and small intestinal bacterial overgrowth. Simultaneous breath sampling of oxygen (O2) concentration allows assessment of breath sample quality and potentially adjustment of H2 and CH4 concentration based on this. The following formula has been used for correction commercially and clinically:

H2normalised = H2measured x (O2atmospheric – O2reference)/ (O2atmospheric – O2measured)

Method
4 volunteers provided sets of 3 breath samples taken very close together: ideal, sub-optimal and deliberately bad. In total 63 pairs of samples presumed to represent the same underlying alveolar breath H2 concentration were collected. The H2 concentration in the worse sample was normalised to the O2 concentration in the better sample using the formula above and compared with the true H2 concentration actually measured in the better sample.

Results
There was very good correlation between normalised H2 concentration from the worse breath sample and H2 concentration in the better sample for the full range of O2 concentrations measured from 11.4% to 16.0% O2 (R=0.909, P<0.001). The normalised H2 concentration systematically underestimated true H2 concentration by on average 5.6% (paired t-test, t= 5.36, P<0.001).

Discussion
There is no threshold O2 concentration below which changes in O2 concentration have no effect on measured H2. The normalisation correction algorithm used is effective. For breath testing to be better accepted, widespread agreement on the correction algorithm for O2 concentration and of the reference O2 concentration to correct measured H2 and CH4 concentration is required.