Comparison of zinc reduction with platinum reduction for analysis of deuterium-enriched water samples for the doubly labeled water technique.

OBJECTIVE Isotope ratio mass spectrometry of hydrogen and oxygen is frequently used to determine total energy expenditure (TEE) using doubly labeled water. Conventionally, hydrogen isotope ratio is determined in hydrogen gas generated from water samples using zinc reduction. We compare this with a new automated platinum method to determine the ratios of hydrogen isotopes in deuterium-enriched water samples. RESEARCH METHODS AND PROCEDURES The platinum method of sample preparation was compared with the zinc method in three ways: analytical variation in deuterium enrichment (within sample; n = 51), analytical variation in TEE estimates (within sample set; n = 10), and level of agreement of TEE estimates between both methods (n = 14). RESULTS For the zinc method, the standard deviation for multiple sets of triplicate 2H2O sample analysis was +/-4.36 per thousand and +/-2.07 per thousand for platinum. The correlation between TEE estimates when sample sets were analyzed in duplicate was r = 0.89 for zinc and r = 0.83 for platinum. The intercept and slope of the regression line were significantly different from the line of identity for duplicate TEE estimates by zinc but were not different from the line of identity for platinum. After correction for the intra-assay variation of each method, the correlation between zinc and platinum for TEE was 0.77, and the intercept, but not the slope, of the regression was significantly different from the line of identity. The mean difference between the zinc method and the platinum method was 56 kcal/day, and the 95% confidence interval was -438 to 550 kcal/day. DISCUSSION These data suggest that the platinum method is at least as reliable as the zinc method as a sample preparation technique for isotope ratio mass spectrometry of deuterium-enriched water samples. The platinum method is also less costly and less labor-intensive than the zinc method.