Commentary: Do preterm infants need thyroxine replacement?

BMJ 1996; 312 doi: (Published 04 May 1996) Cite this as: BMJ 1996;312:1133
  1. C G D Brook, professor of paediatric endocrinologya
  1. a Cobbold Laboratories, Paediatric Endocrinology, University College London Medical School, Middlesex Hospital, London W1N 8AA

    The interpretation of thyroid function tests in babies, especially in premature babies, is difficult. In adults the thyroid gland secretes thyroxine which can be deiodinated peripherally to generate iodothyronines that are more (triiodothyronine) or less (reverse triiodothyronine) potent than the parent compound. Triiodothyronine has three to four times the metabolic potency of thyroxine, and 80% is derived from peripheral conversion, the remainder coming from the thyroid gland. In the fetus serum triiodothyronine concentrations are low throughout gestation because the fetus preferentially converts the thyroxine its own thyroid gland produces and that which it acquires from the mother to reverse triiodothyronine, which is inactive. Clearly there has to be a switch at birth but, as is so often the case, the preterm infant is somewhere between the fetus and the newborn infant.

    There are two further confounding variables. Intracellular deiodination of thyroxine to triiodothyronine varies from tissue to tissue so the relative concentrations of the two hormones in blood do not necessarily correspond to the biological effects of the hormones. The reverse triiodothyronine in the fetus comes principally from the liver and the placenta; the fetal brain does have the deiodinase capable of converting thyroxine to triiodothyronine, but whether it does so is unknown. The other problem is that thyroid hormones are almost completely bound to specific transport proteins so there is potential for confusion in measuring total rather than free thyroid hormone concentrations in preterm babies in whom plasma protein concentrations may vary. We know little about thyroid binding proteins (other than thyroglobulin) in preterm infants.

    The consequences of congenital hypothyroidism are becoming clearer.1 2 There is a discontinuous relation between IQ and plasma thyroxine concentration at diagnosis. A hypothyroid child with a serum thyroxine concentration less than 40 nmol/l is at risk of having a mean IQ 10 points lower than controls or a child with higher values. The IQ deficit measured at 5 years old in children with severe hypothyroidism persists at the age of 10 years.

    The transient low thyroxine concentrations in very preterm infants is associated with developmental delay, so the finding of Lucas and colleagues that preterm infants with depressed plasma triiodothyronine concentrations subsequently had reduced developmental scores tallies with all of the available data. What can be done about it?

    The administration of thyroxine to infants of less than 30 weeks' gestation does not increase plasma triiodothyronine concentrations,4 which suggests that these preterm infants have not lost their fetal habit. This would imply that any treatment to prevent the damage identified by Lucas and colleagues would need to be given in the form of triiodothyronine.

    In the management of hypothyroid patients, doctors prescribe thyroxine replacement, and this acts as a prohormone for triiodothyronine. This strategy is unlikely to work in premature babies, and the implication of the work of Lucas and colleagues is to give preterm infants with low triiodothyronine concentrations triiodothyronine replacement. However, this will not be easy and may be dangerous. Children with thyrotoxicosis do much worse at school than children with hypothyroidism, and some data suggest that intrauterine exposure to high maternal concentrations of thyroxine may have serious long term consequences.5 The construction of any trial of treatment in the management of low plasma triiodothyronine concentrations in preterm infants requires consideration of all these issues. It is good to have enough triiodothyronine but not good to have too much. Fortunately, these issues are being addressed.6


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    View Abstract