Objective The suitability of TSH reference limits derived from national databases to outpatient practices has not been established. We aimed to determine whether race and age influence the distribution of TSH and free T4 (fT4).
Design A cross-sectional study of an urban outpatient medical practice.
Participants Patients (n = 22 116) without clinical evidence of thyroid disease or use of thyroid-specific medications.
Measurements Comparison of TSH and fT4 distributions in specific racial and age groups including blacks, whites and Hispanics.
Results TSH was distributed at higher concentrations, without skew, in whites compared to blacks (median, 1·54 mIU/l vs. 1·18 mIU/l, P < 0·001) and in old (>80 years old) compared to young (20–29 years old) (median, 1·61 mIU/l vs. 1·13 mIU/l, P < 0·001). In all patients, blacks and whites, 3%, 8% and 5%, respectively, of those aged > 80 years were misclassified as having high TSH compared to those aged 20–29 years (P < 0·001). Using TSH limits from national databases resulted in significant misclassification of patients with raised or lowered TSH. Mean fT4 was significantly lower in blacks than whites (17·5 ± 4·38 pmol/l vs. 18·3 ± 3·99 pmol/l, P < 0·001), did not differ between young and old, but decreased progressively (average 7%) as TSH increased to > 4·5 mIU/l.
Conclusions Reference limits for TSH differ between races and with age. Use of race- and age-specific reference limits decreases misclassification of patients with lowered or raised TSH in an urban practice. The unique fT4:TSH relationships of blacks and whites may be genetically determined. The implications of the small decrement in fT4 as TSH increases remain to be explored.