Low triiodothyronine concentration in preterm infants and subsequent intelligence quotient (IQ) at 8 year follow upBMJ 1996; 312 doi: https://doi.org/10.1136/bmj.312.7039.1132 (Published 04 May 1996) Cite this as: BMJ 1996;312:1132
- Correspondence to: Professor Lucas.
- Accepted 4 January 1996
Prematurity is associated with clinically important changes in endocrine physiology. We have reported reduced plasma concentrations of many hormones in premature compared with full term infants, including prolactin,1 testosterone,2 and, notably, triiodothyronine.3 Thyroid hormones are critical signals for brain development.4 We showed a reduction of 7-9 points in scores in standardised mental and motor development tests at 18 months of age in preterm infants, even after adjustment for potential confounding factors.3 As cognitive function at 18 months may not reflect later intelligence quotient (IQ), we measured IQ in the same group of preterm infants 7.5-8 years later when scores are likely to reflect adult values.
Subjects, methods, and results
We measured plasma triiodothyronine concentrations longitudinally in 279 infants in samples taken at 1-3 days and 4-7 days and then weekly until the infant was discharged from hospital or weighed 2000 g.3 IQ was assessed in 236 infants (94% of survivors) at 7.5-8 years by means of an abbreviated version of the revised anglicised Weschler intelligence scales for children, which has five subscales and a correlation with full scale IQ of 0.96. In those children whose minimum triiodothyronine concentration was below 0.3 nmol/l3 we found substantial deficits in IQ scores, even after adjustment for birth weight, gestation, sex, and Apgar score at 5 minutes (table 1). The effect was independent of the duration of artificial ventilation.
Controls (infants with normal neonatal triiodothyronine concentrations) had a longer gestation than infants with low triiodothyronine concentrations (cases) (30.8 (SD 2.4) v 28.9 (2.0) weeks) and higher birth weights (1393 (256) g v 1189 (300) g). In a subsidiary analysis we matched two controls to each case for birth weight and gestation (28.7 v 28.9 weeks) and the cognitive deficits of the cases remained unchanged.
The adjusted deficit in IQ with a low triiodothyronine concentration was 6.6 points, with an 8.5 point deficit on the verbal scale. A causal relation is suggested by the size of the association, its persistence after adjustment for confounding factors, and its plausibility given the known relation between neonatal hypothyroidism and impaired brain development.
Our findings do not provide information about the causes of the low triiodothyronine concentrations in preterm infants. We could not measure other thyroid hormones in these infants because of low sample volume, and so we cannot speculate whether the low triiodothyronine concentrations reflect decreased tissue deiodination of thyroxine or a general reduction in thyroid hormone production by the hypothalamic-pituitary-thyroid axis. This distinction would be relevant to the type of thyroid replacement treatment tested in any future trial.
Our results should be viewed as part of the broader endocrinopathy of prematurity. We previously found that reduced plasma prolactin concentrations were associated with an increased requirement for ventilatory assistance, a longer time to establish enteral feeding, and failure to grow in length—findings consistent with the known biology of prolactin.1 We also observed attenuation of the two normal postnatal surges in testosterone concentration, at birth and around 2 months, in boys who had persistent undescended testes, which is common in preterm boys and a potential risk factor for later seminoma.2
The long term effects of this early endocrinopathy involving several further hormones need exploration. Current interest focuses on the programming effects of events in fetal life or infancy on long term health and development.5 Hormones are important biological mediators for such programming.5 Derangements of critical endocrine signals during sensitive windows in early development could have a range of long term effects, as we have suggested here for thyroxine and brain development.
Funding Medical Research Council and Farley's Health Products, a division of Heinz.
Conflict of interest None.