No reliable method for the estimation of chromium (Cr) status is available yet. The aim of this study is to investigate the possibility of using urinary Cr response to glucose load as an indicator of Cr status. Seventy-eight non-insulin-dependent diabetes mellitus patients, were divided randomly into two groups and given Cr supplements as brewer's yeast and CrCl3 sequentially with placebo in between, in a double-blind, crossover design of four stages, each lasting 8 wk. At the beginning and end of each stage, subjects were weighed, their dietary data and drug dosage recorded, and blood and urine samples collected for analysis of glucose and urinary chromium (fasting and 2 h post-75-g glucose load) and fructosamine. The mean urinary Cr after the glucose load was significantly higher than the fasting mean at zero time (p<0.01). However, only 52 of the patients showed an obvious increase; the others showed a slight decrease or no change. Both supplements caused a significant increase in the means of urinary Cr and a significant decrease in the means of glucose and fructosamine. Only those subjects responding to Cr supplement by improved glucose control showed an increase in post-glucose-load urinary Cr over fasting level, after the supplement but not at zero time. Therefore, it was concluded that urinary Cr response to glucose load could be used as an indicator of Cr status.