Agreement among indicators of vitamin C status.

Agreement among three indicators of vitamin C status--serum ascorbate level, a 24-hour recall, and the frequency of fruit and vegetable consumption--was examined using data from the Second National Health and Nutrition Examination Survey conducted between 1976 and 1980. Agreement between pairs of these indicators was good when assessed at the group level but inconsistent at the individual level. These indicators, when classified as continuous measures, had moderately good agreement (r = 0.45-0.54), whereas agreement was poor when classified as quartiles (kappa = 0.17-0.23). Agreement between clinically based categories of serum ascorbate and total intake levels was poorer than expected (kappa = 0.25) as was agreement between low or deficient levels of both of these indicators (kappa = 0.3). Disagreement between low or deficient serum and intake levels was greater in participants who were younger, African American compared with white and other races, less educated, current smokers, nonsupplement users, and examined in the winter compared with in the summer or fall. These findings suggest that the indicators cannot be used interchangeably to assess vitamin C status because they distinguish between different aspects of status, intake level versus serum level, an indicator of available pool. Moreover, depending upon how these indicators are used in statistical analyses, they may classify individuals differently.