Premature infants constitute a risk group for thiamin deficiency but only little is known about their thiamin status. The aim of the present study was to investigate the thiamin status of premature infants by determination of thiamin diphosphate (TDP) and to identify risk factors for low TDP concentrations. In a prospective, longitudinal study TDP was determined by HPLC in whole blood in the first days of life and approximately every 2 weeks. Demographical data, weight gain, type of nutrition and thiamin intake were recorded. A total of 111 premature infants were included at the Children's Hospital of the University of Cologne, Germany from May 2009 until December 2010 and 222 blood samples were analysed. TDP concentrations showed an age-dependent decline (age 0–10 d, mean TDP = 110·6 ng/ml; age 11–20 d, mean TDP = 95·4 ng/ml; age 21–103 d, mean TDP = 33·6 ng/ml). There was no significant difference between males and females. Young gestational age and low birth weight were associated with low TDP concentrations. No infant was diagnosed with thiamin deficiency. The current nutritional regimen in our hospital did not lead to thiamin deficiency in the study cohort. Further research is required to evaluate how TDP concentrations are regulated in premature infants.

The metabolism of glucose is deranged in thiamin deficiency, but once any deficiency has been corrected there is no further effect of increased thiamin intake on the ability to metabolize glucose through either pyruvate dehydrogenase (EC 1.2.4.1) and the citric acid cycle, or the pentose phosphate pathway, in which transketolase (EC 2.2.1.1) is the thiamin-dependent step. It has been suggested that the Wernicke-Korsakoff syndrome is associated with a genetic variant of transketolase which requires a higher than normal concentration of thiamin diphosphate for activity. This finding would suggest that there may be a group of the population who have a higher than average requirement for thiamin, but the evidence is not convincing. There are no estimates of biotin requirements, but either coenzyme saturation of erythrocyte pyruvate carboxylase, or the excretion of 3-hydroxy-isovalerate (perhaps after a test dose of leucine) could be used to assess requirements in depletion–repletion studies. Biotin deficiency leads to impaired glucose tolerance, but it is unlikely that glucose tolerance could be used to assess optimum biotin status, since other more common factors affect glucose tolerance to a greater extent. Plasma triacylglycerol and non-esterified fatty acids are moderately elevated in pantothenic acid deficiency. However, this is unlikely to be useful in assessing pantothenate status, since again, other more common factors affect plasma lipids. To date there are no biochemical indices of adequate pantothenate nutrition, and no estimates of requirements.