Estimates of substrate oxidation obtained from appearance of 13C or 14C from tracers in breath must be corrected for retention of labeled carbon in the body. We aimed to determine the effect of a defined experimental diet and metabolic status on recovery of infused Na [13C]bicarbonate in breath. Six healthy male subjects consumed an experimental diet for 7 days before receiving a continuous infusion of formula without tracer on day 8 and received either an intragastric (ig) or intravenous (iv) infusion of Na [13C]bicarbonate on day 9 or 11 during a 4-h postabsorptive (PA), 4-h continuously fed period. A trend toward increasing PA breath enrichment during the first 7 diet days approached statistical significance (P = 0.051), whereas breath enrichments measured 3 h postbreakfast were consistently higher than PA values throughout and did not change over the 7-day period. Breath enrichments during a 4-h continuous ig infusion of formula without tracer on day 8 rose 2.0 +/- 0.5 atom percent excess (APE).10(-3) above base line (P less than 0.001, ANOVA). In the tracer studies, breath enrichments were similar for the ig and iv routes of tracer infusion. For the ig infusion the fraction of infused Na [13C]bicarbonate recovered in breath as 13CO2 was 0.74 +/- 0.02 for the PA period and 0.79 +/- 0.02 for the fed period. For the iv infusion the fraction recovered was 0.70 +/- 0.04 for the PA period and 0.82 +/- 0.03 for the fed period. Fractional recoveries were not significantly different for ig and iv routes of administration but were different for PA and fed periods (P less than 0.0001, 2-way ANOVA). The fractional recoveries for the fed period obtained here were similar to the value 0.81 reported in a number of other studies. Recovery of tracer in breath increased linearly with O2 uptake and CO2 production, suggesting that factors affecting respiratory gas exchange may alter recovery. We conclude that the primary factor determining label recovery is the immediate and recent nutritional status of the host.