was thus given careful consideration in deriving the Estimated Average Requirement (EAR). Urinary thiamin excretion decreases markedly as thiamin status declines and is also affected by recent dietary intake. Bayliss and coworkers (1984) reported a correlation of 0.86 between the oral dose of thiamin and urinary thiamin excretion. However, in doses of up to 1.05 mg there was overlap with baseline values. The use of a load test, in which thiamin excretion is measured before and after a test load of thiamin, helps differentiate between extremes of vitamin status (McCormick and Greene, 1994).

Erythrocyte transketolase activity has also been widely used and is generally regarded as the best functional test of thiamin status (McCormick and Greene, 1994), but it has some limitations for deriving the EAR and should be evaluated along with other indicators. In this test, erythrocytes are lysed and the transketolase activity is measured before and after stimulation by the addition of thiamin pyrophosphate (TPP); the basal level and the stimulated value (typically expressed as a multiple of the basal level, termed the activity coefficient or TPP effect) are measured. In thiamin-depleted individuals, basal erythrocyte transketolase typically is low and the incremental response after TPP addition is enhanced.

Although the test has long been used in assessing thiamin status, in one recent study (Bailey et al., 1994) it correlated poorly with dietary thiamin intake in English adolescents. Similarly, in a study population of 179 adult men, Gans and Harper (1991) found a wide range of TPP effect values (0 to 95 percent) associated with thiamin intakes that were all above 1.5 mg/day over a 3-day period. Similarly, they also found a TPP effect of 0 percent associated with a wide range of intakes (approximately 0.75 to 6.0 mg/day). Schrijver (1991) reported that the activity coefficient may appear normal after prolonged deficiency, making identification of the deficiency more problematic. From studies of the elderly, Pekkarinen and colleagues (1974) concluded that evaluation of thiamin status should consider other indicators along with erythrocyte transketolase activity.

Factors other than thiamin status, such as genetic defects, may influence the enzyme activity and thus the test results. Individuals and tissues both differ in their sensitivity to thiamin deficiency. This observation may be explained by the pronounced lag in the formation of active holoenzyme and the interindividual and cell type variation in the lag during thiamin deficiency (Singleton et al., 1995).