Measurement of CO2 production by the doubly labeled water technique.

Recent applications of the doubly labeled water technique to the study of human metabolism have employed multiple sampling of body water over protracted periods, rather than the more traditional method of taking only an initial and final sample for isotopic analysis. In addition fractional turnovers of the body pools have been estimated by fitting curves to the sequential log-converted isotope enrichment against time. By manipulation of data collected in the field in a study of metabolism of vespertilionid bats, it is shown the curve-fitting technique results in an accurate estimate of CO2 production only when the rate of CO2 production is constant. Biologically realistic nonsteady-state conditions result in errors in estimates of CO2 production of up to 30%. In conditions where CO2 production is known to be temporally variable, the more traditional two-sample method may provide a more accurate estimate of CO2 production.