Longitudinal hair chromium (H-Cr) profiles in a group of patients with type 2 diabetes mellitus (n = 59; age, 62 +/- 9 years) and healthy elderly (control) subjects (n = 49; age, 59 +/- 10 years) matched by age and sex were measured by solid sampling electrothermal atomic absorption spectrometry, providing data on the magnitude of variation of Cr content along the hair length. H-Cr average (H-Cr(av)) and H-Cr proximal (H-Cr(pr))(.), relating to the average Cr content of the whole hair and the proximal 3-mm hair length, respectively, were also obtained. No significant difference between the healthy and diabetic group was found in mean H-Cr(av) or H-Cr(pr) contents (248 +/- 108 vs 247 +/- 143 and 233 +/- 98 vs 278 +/- 195 ng/g, respectively. However, women in the control group had significantly lower H-Cr values (P < .01) compared with men, but this difference was absent in the diabetic population. The distribution of log H-Cr(pr) values in the control population displayed a Gaussian shape, in contrast to the substantially wider distribution, skewed toward lower H-Cr(pr) values, observed in the diabetic group. The magnitude of variation in H-Cr content in the patient group over an interval of approximately 2 to 3 months (time of growth of the hair sampled) was found to be a factor of more than 2 larger than that in the control population (+/- 58% vs +/- 26%). A strong relationship (R = 0.656; P < .01) between log H-Cr(pr) and log fasting plasma Cr was observed in the diabetic group (n = 20). The mean fasting plasma Cr value of this group was 0.41 +/- 0.10 microg Cr per liter. No correlation between H-Cr(av.) and duration of diabetes was observed. A strong positive association was observed in the control population between H-Cr(pr) and fasting plasma insulin (n = 22; R = 0.6157; P < .01), and H-Cr(pr) and fasting plasma glucose (n = 24; R = 0.4118; P < .05), which is indicative of the interrelation of these parameters. In the control population, H-Cr(av) showed a slight decrease with age (n = 54; R = 0.2691; P < .05), which is assumed to be the result of increased insulin resistance caused by various age-associated factors including Cr deficiency. None of the above relationships was significant in the diabetic group. Evidence is presented that justifies the assumption that the longitudinal H-Cr profile resembles the variation in Cr metabolic rate over the time span of growing hair, which is not appreciably affected by external contamination. This suggests that glucose intolerance (type 2 diabetes mellitus) is an important factor that disturbs Cr metabolism.